Yesterday, I had an interesting discussion with a couple of my fellow Environmental Anthropology grad students about different theoretical approaches within the field. We had read chapter 3 from Arturo Escobar's book Territories of Difference - one which I have found very helpful in sorting out different theories of nature/culture. In this discussion, the three of us explored the different approaches listed, and tried to make some sense of how they relate to one another. I'll not go into Escobar's schematic - if you're interested, you can pick up the book or find one of the many articles Escobar has written on the subject. What I want to talk about instead is the practice of categorizing and arranging different theoretical approaches.
In my opinion, there is nothing wrong with trying to categorize theories. We do it all the time, and it is a helpful practice for those who are just learning or for those who are only trying to get a general sense of the field. Once we begin to delve deeper, and really try to understand the different theoretical approaches, however, the categories tend to get in the way. The reason, I think, is that categorization is the wrong method for trying to truly understand theory. Aside from a few iconic individuals and/or papers (i.e. Levi Strauss and structuralism, Steward and cultural ecology, etc.), the categories simply don't fit nicely. People borrow from a number of different theoretical frameworks, theories are misused or misunderstood, different people have different names for the same theory or the same name for different theories, and so on. It becomes just a jumbled mess of confusion. That's fine in practice - it seems to work much of the time - but for trying to make sense, it becomes very difficult.
The way I want to approach it - instead of viewing theories as categories into which we fit people and projects, we can understand theories as assemblages of practices that people can borrow from and utilize in different projects. A person will borrow certain ideas from Foucault, or Bourdieu in a project - these will not embody the whole of those theorists' thought, but only fragments which she then hangs together to form her overall practice (at least in this project). Even those theorists who are said to embody a particular theory will be doing the same. In this sense, a theoretical framework (structuralism, phenomenology, constructivism, actor-network theory, etc.) is really a heterogeneous mix of ideas and ways of thinking - not bounded and homogeneous as the categorization method assumes or suggests. Of course certain ideas fit well with one another, while others do not, so there is some sense of structure in that, but not in the sense of over-arching categories.
This is not to say that the categories don't matter. The categories are very real, and have a great effect. When you look up phenomenology in an encyclopedia or online, you'll no doubt see Hussrl and Heidegger listed as primary philosophers. You may see others listed as well. As a result, these categorizations tend to assemble certain people and ideas in a particular way. Without these categorizations, there may not be a "phenomenology" to speak of. So the categories themselves play a role in the assemblage of ideas.
I think this is a far more useful way of conceptualizing theory, though I know it is by no means the only way. Does anyone want to try to fit this theory of theory into a nice category of meta-theories? It's theories all the way down!