In a blog post written last March (which I linked at the time but did not comment on at any length), Michael Walzer rehearsed J.S. Mill's argument about non-intervention, an argument Walzer had also summarized in his Just and Unjust Wars (Basic Books, 1977), pp. 87-91. With the debate about the Libyan intervention, sovereignty, and R2P continuing to simmer (in the IR blogosphere and elsewhere), and with Gaddafi still at large and one or two cities in Libya still resisting the rebels (or revolutionaries, or anti-Gaddafi forces, whichever label you like), it may be worth going back to Walzer's post. The question whether the U.S. and/or NATO should intervene in Libya is now of course moot, but the broader issues will likely recur (and have already recurred in a way in the case, e.g., of Syria).
Mill's position was basically that oppressed peoples had to struggle for their own freedom without outside help; if they failed to secure freedom that proved they didn't deserve it, weren't "fit" enough for it. In his blog post of last March, Walzer wrote that if the Libyan rebels were on the verge of defeat he would not be willing to go all the way with Mill, i.e. to declare the rebels "unfit" for liberty and leave them to their fate after a Gaddafi victory. But Walzer said that when intervention became necessary -- and he wasn't sure exactly when that point of "necessity" would occur -- it should be done by neighbors, by the Egyptian and Tunisian armies, rather than by the U.S. and NATO.
Even though he was not willing to go all the way with Mill in the Libyan case, Walzer clearly has a lot of sympathy for the view that oppressed peoples should do their own struggling, with outsiders intervening only in cases of real "necessity" (however defined). In Just and Unjust Wars [JUW] (pp. 90-91), he wrote: "We need to establish a kind of a priori respect for state boundaries; they are, as I have argued before, the only boundaries communities ever have. And that is why intervention is always justified as if it were an exception to a general rule, made necessary by the urgency or extremity of a particular case."
It is perhaps unfair to focus on something Walzer wrote 30-plus years ago, ignoring his more recent writing on these issues; still, the sentence just quoted shows a weakness, in my view, of his approach in JUW, namely the attachment of too much moral value to state boundaries. He recognized the (in some cases) "arbitrary and accidental character of state boundaries... [and] the ambiguous relation of the political community or communities within those boundaries to the government that defends them" (JUW, p. 89), but his basic position was that boundaries enclose communities which should be left to work out their political fates for themselves. There is definitely something to be said for this view but it is also necessary to acknowledge that the ways in which state boundaries are routinely penetrated or breached by outsiders, whether they be governments, corporations or NGOs, make the issue somewhat more complicated [note: some, e.g. Robert Jackson, would deny this]. Moreover, it is not the case that state boundaries are "the only boundaries communities ever have." Students of international relations have spilled much ink writing about all sorts of boundaries (ethnic, zonal, tribal, etc.). State boundaries retain a special place in international law and practice, but they are not the only boundaries communities have.
So where does this leave matters? Intervention should still be an exception to a general rule, and R2P, at least as I understand it, does not alter that. But in a world that some see as being full of cross-boundary 'networks' and transnational communities, the principle of non-intervention, assuming one wants to keep it, perhaps needs an updated justification, one that does not rely quite so heavily on a picture of self-enclosed national communities, each working out its own political destiny in isolation from the world outside. I'm not sure exactly what that updated justification of non-intervention might look like; perhaps political theorists and IR types have already produced one and with a little research I could find it. But laziness being the blogger's prerogative, I'm not going to bother searching, at least not now.