Anthropology with a big-A is the study of human beings. I'd say that this Anthropology includes studies of everything humans do and every aspect of how they do it, ranging from what we call political science to cognitive psychology to criminology to economics to paleoanthropology to Assyriology. Fundamentally, all those sub-disciplines are about the same thing, even if they disagree about it. Despite studying basically the same thing, there are nevertheless a lot of sub-disciplines in existence just for studying people. That should make it easier and lighten the workload for scholars working in these fields, but to focus on even only one of these sub-disciplines of Anthropology requires an Atlantean burden.
Social anthropology is more or less the study of people in pre-industrial societies and how they live, although that isn't much more than a core anymore, around which a plethora of other interests have caked themselves (both beneficially and maleficently). And social anthropology is really hard. All of the human sciences are. The reason for this is that whatever you say about people, it will always be more contentious to more people than whatever you say about ants or magnetism. In order to study people, you've got to master a lot of extra subjects and ideally keep up with the major works of the literature on them because people like to argue about people.
Because the questions of whether or not studies of people constitute "science" or whether people have access to the world outside their heads are frequently brought up in attempts to discredit the scientific enterprise as applied to humans, you've got to know something about epistemology, which is a philosophical discipline concerned with the justification of belief.
I re-read the first chapter of Bob Parkin's book, Kinship: an introduction to the basic concepts, recently, and in it he claims that beliefs are only justified by other beliefs - hence science is primarily a social phenomenon. This is news to epistemologists; the idea that beliefs are only backed up by other beliefs is called coherentism, and is an out-dated theory in epistemology (its inverse is foundationalism, the view that there is an absolute foundation on which all claims to knowledge can be based). Neither coherentism nor foundationalism is all that reasonable, and most epistemologists today accept some form of foundherentism (ugly, ugly neologism), which Susan Haack devised about two decades ago (in Evidence and Inquiry) to overcome problems with the two prior theories. This seems like an arcane point - it is, in fact, despite its enormous importance - but arcane points are always being dredged up because arguments about humans get so fierce, and any number of things can be used to claim that someone is full of shit when they're just stating reasonable things. You could be forgiven for believing that a book on a subject like human kinship could avoid philosophical technicalities, but this is very much the norm.
Parkin isn't opposed to science, so this is not a problem peculiar to relativist attacks on the human sciences. It's just a problem of having too much to know about in order to pursue the fundamental questions of how humans work.
Anthropologists also have to know a lot about ontology, another aspect of philosophy that most people overlook (including anthropologists, of course). Ontology addresses bullshit-sounding questions like, "is a 'fleet herd of deer' the same as a 'herd of fleet deer'?" That example actually comes from biology, and is often used in arguments about group selection (which is predicated on the frankly silly view that a 'fleet herd of deer' is more than just a 'herd of fleet deer'). The question of whether society is the same as or reduces to the individuals that constitute it is an ontological question. How money exists is an ontological question.
Place a foot wrong on these questions and you will be criticised by somebody, but given that there is a great diversity of ontological views in anthropology departments (from the Durkheimian, unnaturalistic, non-reductive view of society as having properties that don't reduce to the individuals that make it up, to free-will-based so-called "existentialist" anthropology that doesn't recognise the existence of any kind of social fact), you can't put really a foot right. Endorse the most reasonable and scientific position and you'll be accused of being bad at thinking and a political rightist determined to destroy social science or humanism or life itself, despite that often being the precise opposite of the truth. The only way to be right all the time is to write appealing nonsense.
You also have to know a lot of facts, and a lot of ways of approaching problems. If you want to look at the prehistory of human civilization in the Pacific Ocean, then you have to be familiar with the tools of historical linguistics, archaeology, comparative ethnology, and possibly even pollen analysis and genetics. You also have to know a lot of arcane detail from your studies of these things, which involves reading literally hundreds of books and thousands of articles, and retaining the data somehow so as to be able to correlate it all for the purpose of formulating answers to your research questions. Not only that, but you should be familiar with the languages of the region, perhaps to the point of being fluent in one or two of them. Needless to say, each of these things could take up a specialised degree in itself (as could the philosophical topics). And once you think you've grasped all these things, when you decide to publish about them you have to get them checked by experts in the relevant fields (ideally). That takes time. Looked at this way, it's quite easy to see the appeal of "critical theory" and writing literal nonsense, if only as a time- and effort-saver.
Anthropologists also have to be aware of a series of problems relating to studying people whose experiences and assumptions are different from the anthropologist's own. These include ethical considerations and practical ones, as well as the epistemological questions of how to verify ethnographic information and how to assure yourself that you know what others are referring to, amongst others. Add to this the fact that fieldwork, whether ethnographic or archaeological, depends on being away from your usual abode, and can easily end up being in a place without internet or a library (of course) - not to mention the fact that many remote parts of the earth have not eradicated malaria or leishmaniasis. This can be dangerous for your health, and it also takes time away from hitting the books - which is necessary for scholars with so much stuff to read and so many areas of study to become familiar with.
There's also a great deal of specific technical knowledge about pre-industrial human societies that social anthropologists need to learn. The study of human kinship is one such technical subject, one with lots of engaging intricacies that are the bane of many students. There's a lot to learn; it turns out that pre-industrial social structure is just as varied and weird as industrial and post-industrial social structure, but it relies to a greater extent on the vagaries of sexual intercourse, reproduction, and physical relationships than does the social structure of, for instance, the United Kingdom in 2012. This presents many tricky problems. Moreover, if you view humans naturalistically and believe that behaviour results fundamentally from the brain and nervous system, then you can hardly ignore psychology and neuroscience. Which aren't easy topics.
It is hardly surprising that, overwhelmed by the necessity to produce work with some measure of expertise in all of these things, anthropologists end up being perpetual amateurs or jargon-spouters.