I agree with most of Randy's and Ilya's comments on libertarianism and war. I'm just writing to silghtly disagree with Ilya about the nature of libertarianism.

As Ilya writes, the idea that "nondefensive humanitarian military intervention is potentially compatible with libertarianism" may seem to be "a paradoxical claim," but really isn't paradoxical at all. It only seems paradoxical if you think libertarianism, in its essence, is necessarily opposed to government action. But the variety of libertarianism I subscribe to isn't a set of answers; it's a way of asking questions. (Libertarianism is, of course, large; it contains multitudes. "The variety of libertarianism I subscribe to" is quite different than, say, anarcho-capitalism and many other varieties of libertarianism. However, I claim, my variety is consistent with libertarianism, and moreover, a better variety than others!)

And the questions one should ask about any government intervention is whether it will, on balance, increase the protection of rights (where "rights" are defined in a libertarian way -- an important debate but not relevant to this post).

Now -- step 1 -- many government interventions make no pretense at increasing the protection of rights. So that's sufficient reason to oppose them right there. (Adducing examples of this is left as an exercise to the reader.)

But some government interventions, in particular all recent U.S. interventions, make at least a claim that they will increase rights protection in some way. Then -- step 2 -- it's probably a good idea to be deeply suspicious that such claims will pan out, because governments usually do a bad job at this sort of thing. But this is an empirical claim and, as Ilya says, "cannot be deduced from the nature of libertarianism itself on a priori grounds."

Therefore, while I agree with Ilya that "libertarians oppose all or most large-scale government programs in the domestic realm," I disagree that "if they didn't, they would no longer be libertarians." I think libertarianism requires ruling out certain things at step 1, but once the government intervention comes with a claim of rights protection, it can only be rejected on the merits at step 2. Libertarians can disagree at step 2, and many of them will turn out to be incorrect, but such disagreements are fully consistent with the framework of libertarianism.

Thus, one can imagine a libertarian who's either exceptionally optimistic about the chances of success of most potentially-rights-enhancing government programs, or exceptionally pessimistic about the state of rights without the government intervention (even an incredibly bungled one). This libertarian might support most potentially-rights-enhancing government programs (and so you could potentially have a super-hawkish, high-tax libertarian, or similarly a super-environmentalist, high-regulation libertarian). Others, more pessimistic about the programs or optimistic about the state of the world without the program, might take the opposite view. But all of these can properly be called libertarians.