Studying anthropology at university, you learn a lot of fairly useless stuff. If what you want to do is a fairly old-fashioned study of village life and marriage alliance in eastern Indonesia, or a dig at the site of a pre-Columbian Ecuadorian kingdom, you've got to learn about - and pass exams on - 'social theory', or the thinking of continental philosophers about human society. This is usually called, simply, 'theory'.* It can be important to know about the philosophical aspects of investigating societies, but for most of the work anthropologists do, this 'theory' stuff is basically useless. Maybe even worse than useless.
Old texts in the human sciences, including anthropology, are often surprising. Sometimes they're surprising because you wonder how anyone could possibly believe the things they say, as with a lot of racist nineteenth century works. At other times, they're surprisingly correct. (I think Freud was a bit of both.) Earlier anthropologists and psychologists were not stupid people, and they made a lot of progress in understanding key topics. Much of the data produced by earlier social scientists survives as useful data to this day. That tells me that maybe social science doesn't need the theoretical 'sophistication' represented by the bad philosophy foisted on students of anthropology nowadays. And many topics simply don't need it. At all.
Pre-industrial social structure, often studied under the label of 'kinship', is one such topic. It can be studied productively in several ways, most of which were anticipated far in the past. Alfred Radcliffe-Brown, an Oxford anthropologist of the first half of the twentieth century, gets quite a bad press in anthropology departments as someone whose theoretical outlook was perennially incorrect, but if you actually read his books and essays, you'll find that a lot of what he said at the level of description and basic explanation of pre-industrial social structure is very sensible. Bronislaw Malinowski, one of the most famous anthropologists of all time (I mean, he featured in an episode of The Young Indiana Jones), was also a bit of a poor theorist, believing that all of society results from human needs (as opposed to wants, or other kinds of pro-attitudes), which doesn't seem to be true. He had a poor general theory of society - and yet he managed to coherently address many problems in the study of pre-industrial peoples. Malinowski also produced some of the best and most interesting ethnographic works ever written, which had a real impact on how a lot of people thought about so-called 'primitive' populations. Likewise, the linguists of old, including Jacob Grimm of Grimm's Fairy Tales fame (who formulated Grimm's Law), managed to contribute a great deal to the formalisation of the discipline of linguistics without bothering too much with the philosophy of language.
It just doesn't seem necessary to talk about Pierre Bourdieu for very long if what you're interested in is the ancient Tongan empire** or Offa's Mercia, or even present-day societies like those of the Kenya/Ethiopia borderlands or modern Manila or Miami. What matters is the empirical data collected about what people do, with some reasonable attempts at explaining why they do it. I'd also say that Bourdieu is not the most reasonable thinker to employ, and that if you're going to go with higher theoretical positions, you could do a lot worse than studying some basic cognitive psychology, primatology, or analytical sociology (depending on just what it is that you intend to investigate). Continental philosophy is almost entirely bullshit, after all.***
Now, I like philosophy (real philosophy, not the continental variety). I've written here about a few prosaic topics in philosophical terms, including the workings of historical linguistics. And I think that formulating a rational general theory of human behaviour is pretty important. I also think that such a theory could have a useful application in making sense of empirical data from ethnography and archaeology, or at least in unifying our understanding of it, and that data from these disciplines should have an important role in shaping the theory. But it's clearly possible to investigate what people do and why they do it without bothering all that much with higher-level theoretical positions, and it seems as if approaching the study of humankind through the lens of poor philosophy has deleterious effects - especially if that philosophy is bad or politicised epistemology.
It doesn't matter too much how social facts exist in terms of the mental states of individual humans if what you're trying to investigate is the social structure of a Timorese village. It's interesting to think about, and a good theory of how social facts work could have significant explanatory power (while also making social science more naturalistic, an obvious step forward), but the best terminology for making sense of marriage alliance was formulated decades ago by empirically-minded theorists, from McLennan to Levi-Strauss to Leach.
It is partly for this reason that it seems important to have several levels of theory with which to approach the empirical problems of the human sciences, an idea which has encountered a surprising amount of resistance. Nobody expects biologists to phrase all of their theoretical ideas in biochemical terms, even though almost all biologists believe that their subject matter essentially reduces to biochemistry. Nobody expects car mechanics to know the chemistry and physics behind the functioning of the internal combustion engine. So I don't know why students of humankind should be expected to have a single level or kind of theory, even if they believe that all of human behaviour is fundamentally of the same kind. There seems to be so much resistance to useful abstraction in the anthropological disciplines.
Humans are interesting creatures, and they do (and have done in the past) a lot of weird and fascinating things. It is, at least in part, the role of anthropologists and archaeologists to investigate these weird and fascinating things. But when you go to study anthropology at a university expecting to read about the weird things people get up to, you find that instead much of your time is spent reading impenetrable continental philosophy - or hearing about it in lectures. It's really odd, and I just wish it would stop.
* This is different to theories that resolve particular problems in archaeology and anthropology, like the origins of agriculture or state societies. That usually isn't called 'theory'. For some reason, 'theory' refers to things that aren't theoretical at all, in the sense of falsifiable conceptual approaches to empirical problems.
**That's actually quite a poor wiki article on the topic. Maybe I'll edit it when my reading focus has shifted back to the Pacific.
*** Another pretty obvious problem is that studying this higher-level theory doesn't seem to inoculate anthropology students from bad ideas, and it doesn't stop them from holding mutually-contradictory positions. It also doesn't stop anthropologists from holding beliefs that are contradictory to known science. Some social scientists approve of the philosophy of Paolo Freire, a Brazilian education theorist, despite the fact that Freire, a Catholic Marxist, believed in an essence to humankind, an idea entirely debunked by evolutionary biology. So it seems to me like this continental theory rubbish fails on all fronts.
I haven't addressed several other features of 'theory' in anthropology, including something currently in vogue, 'ethnographic theory', which is primarily about investigating the concepts provided by ethnographic informants 'on their own terms'. There's more to it, of course, but... I'm really not sure what the purpose of this is, honestly. There seem to be much more interesting things to talk about than the nihilistic attempt to unravel the amorphous concepts given to ethnographers.