I've been away for a week and already I'm having a hard time catching up. On the question of a textbook, there wasn't one I could use, nor would I have recommended such a thing. Even if someone had achieved the near impossible task and bring out a generally satisfactory book on the subject I'd not have recommend it as an overall guide. The problem with recommending a textbook, or of preparing one's own reader, were that the chances were that that's all a student would read. Of course, I write this with the luxury of having had soas library available to my students; but it's not the only comprehensive library around and it didn't have everything (it doesn't have Magiciens de la terre for example: in fact my copy is the only one I know of 'in captivity' in the UK - but perhaps given the misjudgements it spawned perhaps that's just as well!).
African Arts, the journal, is the nearest one gets to a single publication; but even so, if one recommends articles across as many publications and journals as one can there is the hope that an idle student will browse, and come across other things, developing their interests thereby and discovering for themselves a familiarity with the wider literature, both regional, disciplinary and theoretical, within which any study of art has to subsist. That is not to say that the literature giving a more comprehensive survey is of no use, and Colin has already listed some of the obvious things; but I wouldn't have wanted to give any one book textbook status.
As to boundaries, given that they may exist as much in one's own imagination as in any substantial reality, they constitute one of the problems that must be dealt with. The kaleidescope of identity-and-difference is a never ending engagement, and in drawing that out for a student, the only boundaries were what did I know (or rather: what could I get away with as if I knew), and how long did each lesson (class, lecture, seminar, etc) last, and how did one build up students' access to the available data through a reading list interpreted across a set of classes in a 24-week teaching schedule; and I'd like to think that no matter how narrowly focussed my classes were at the outset, they did broaden out as I became more confident as a teacher, and, indeed, as the material available expanded. The "contemporary/classical" dichotomy was among the 'border disputes' that had to be deconstructed, following upon the demolition of the old-hat ethnic categories. (This was not to suggest that ethnicity was of no relevance to art but rather to demolish its status as the one-size-fits-all paradigm.) I could go on ... and on ...
-John Picton, Emeritus, SOAS