Norm talks a bit about humanitarian intervention and whether recent experiences in Libya and Syria are causing anyone to rethink their positions. I tend to agree with Norm that the decision to intervene is always going to be idiosyncratic to the case at hand, and trying to make general rules out of the collection of experiences is both descriptively and normatively bad: people are going to respond or fail to respond for reasons that have little to do with past experience.
But having paid close attention to this issue for ten years now, and having paid some attention to it since Bosnia/Somalia dominated the news cycle, I do believe there's an underlying dynamic behind these shifting responses: a theory that puts forward an obligation to help in trying circumstances is widely disseminated and accepted; a case arises that would test this obligation; states act or do not in response to the problem; if they act, they encounter unforeseen but knowable complications and eventually lose the desire to follow through; that lost desire means subsequent cases are neglected, and the theory of obligation is judged a failure; if they don't act, the theory is judged a failure. Consequently, theorists get together and create a new theory of obligation, but this theory just repeats the same process over and over. Post-Cold War humanitarianism fails in Rwanda (and Bosnia and Somalia, but never mind that), therefore the doctrine of liberal internationalism is proposed, which is rejected as illegal and fails in Afghanistan or Iraq, therefore the doctrine of Responsibility to Protect is proposed, which fails in Darfur and elsewhere. It's a strange conclusion, but I think it's the correct one: when it comes to intervention, we have cases but can have no theory.