Interesting discussion, folks and, since I know how much Wilf loves social sciences, I thought I'd toss in my$.0195.
Well, words, in and of themselves, have no meaning, only that which is ascribed and agreed to by those who use them. That, BTW, isn't just playing with semantics; it's a root behind semantics.
On "civil war" vs. "insurgency", I tend to think of them as overlapping circles of meaning, with a fair amount of overlap. In some cases, civil wars have absolutely nothing to do with policy or governance structure (the dynastic wars of the 11th - 16th centuries are examples), while in other cases it is, IMHO, possible to have a "civil war" with little or no violence and, certainly, no open warfare (think along the coup d'etate line or, at the other extreme, the economic lawfare of the Byzantine bureaucrat faction against the aristocracy).
Personally, I've found that it it really useful to assume that the boundary condition of conceptual terms gets defined by identifying central characteristics and then assigning membership values for specific instances to each of those core characteristics and seeing what clusters develop. Sometimes when you do that, what you think is one or two competing terms actually shows up four or five clusters.
Cheers,
Marc
Bookmarks