Ultimately, no matter how good your ideas may be, or how much you know about a region, it doesn't mean anything if you can't convey them. Facts have to be interpreted, ordered, and presented. (This is assuming, of course, that there are such things as 'facts', but that's another issue.) I have found two thinkers – both non-anthropologists – useful in talking and thinking about writing and explanation. The first is the German philosopher Jürgen Habermas. The second is the Italian semiotician and novelist Umberto Eco. (Some of these issues have also been covered by anthropologists – see, for example, Clifford and Marcus's book Writing culture .) Let me talk about Habermas first.
Jürgen Habermas has what I think has to be the clearest, briefest discussion of explanation and writing anywhere. In talking about discourse and explanation, he says that what matters is the "unforced force of the superior argument." What I take him to mean (at least in this context) is simple: you can not convince people your understanding or your point of view is correct through force. You must convince them through a combination of reason and presentation. In the end, people will accept your argument because they want to, because you've managed to convince them to do so.
This does not mean that all explanations are equally valid, and simply the better writer will prevail. Nor does it mean that you should be able to get away without solid data to back you up. (Unfortunately, I've seen this happen way too often.) What it does mean is that you need to be able to present ideas and arguments coherently. You need to be able to tell a story. You need to be able to weave a narrative that makes sense, and is coherent. You need to be able to marshal observations and data and present them in a meaningful way. Even the best argument is going to fall flat if it isn't well supported and well crafted. (Well, once or twice I've read papers that have succeeded in spite of the author, but they are extremely rare.)
When you are writing an article or book, or giving a paper at a conference, in effect you are saying "Here's what I think is going on, and why." What explanation and writing is about is being able to convey what you think is going on, and at the end, have people sit back and say "Yes, that makes sense," or "That seems right to me." Beyond this – how you actually go about writing well and creating a coherent argument – is really beyond the scope of this little essay.
That brings me to my next writer – Umberto Eco. Perhaps best known in the US for his novel The name of the rose, Eco is actually an academic. His work deals with symbols, meaning and representation. (This is also clearly a concern of his in his novels.) As such, he has written about interpretation – reading and understanding a text. I won't get into the whole history of literary theory and what is known as post-modernism here, but Eco's argument is that while you may not be able to conclusively prove there is a single, correct meaning to a text, you can rule out certain interpretations as wrong. You can, in other words, bound your interpretation and understanding of a text.
A text – especially fiction – does not ultimately have one "true" meaning. We all bring baggage to our encounters with the text, and this effects how we read and think about them. You can not, Eco would say, prove beyond a shadow of a doubt what Proust's Remembrance of Things Past means, but you can say with complete confidence what it is not about. It is not – to pick a banal example – about the American Civil War. Nor it is about Trotsky's relationship with Lenin. But within the confines of what everyone would agree it is about, there is room for disagreement. This is where Habermas comes in – convincing people that your understanding within that space is the correct one.
For those of you wondering what an anthropologist is doing citing literary theory, there are two answers. The general one is that anthropologists (and most academics and intellectuals, I think) are magpies – we take theories and ideas that work for us from where we find them. (The French anthropologist Levi-Strauss used the term bricoleur, but I think it sounds a bit pretentious if you don't speak French.) The second, more specific one, germane to this essay, is to go read Paul Ricoeur's work on action as text, which can be found in his collection Hermeneutics and the human sciences. He argues, convincingly, I think, that we can see action as a form of text – something which can be "read" and interpreted.
Note: I'm pretty sure Habermas himself has made a similar point. I'm sure he's talked about the "unforced force of the superior argument" in his book On the logic of the social sciences. Alas, my copy is buried in a box somewhere, and I haven't been able to dig it back up to double-check. If I can find it, I'll post more information.