(Warning: this is going to be very rambling and unfocused. I'm thinking as I type, here.)
There are too many interesting takes on these questions out there for me to comment on more than a tiny fraction of them, but I do want to express a very general concern I have about the terms of the debate.
Take this excerpt from the brilliant Max Sawicky:
The politics of war should be difficult for the aggressor. There should be a high bar to clear for its advocacy. It should not be enough for some other nation to be an enemy, for it to have nuclear weapons, for it to be a tyranny, for there to be idle U.S. troops not engaged in some other war, for it to abuse its subjects or its neighbors, for it to be universally despised, for the U.N. to vote for its demise. My three exceptions would be 1) self-defense (in the face of an imminent, manifest, tangible threat, or act of aggression), naturally; 2) the threat of genocide, or 3) the near-guarantee of very great benefits at very low cost.
I agree very strongly with what I believe to be the underlying premise here: a case for war needs to be stronger than a case for just any policy. It's not enough to make the argument that there's a good chance that a war will make things better in some way, to some degree. This would be enough argument on behalf of, say, a health care plan, but war is different from health care. One might hope that this would be obvious, but the evidence of our national experience in 2002/3 shows that apparently it isn't obvious enough.
I do worry a bit about the "exceptions" Max outlines. The third exception is particularly tricky, if taken literally--though I doubt Max intends it in this way--because it would seem to allow superpowers free reign for military adventurism against smaller and less powerful countries. Seizing the Panama Canal or the oil fields of Kuwait would probably be possible at very little cost, and might be thought to yield very great benefits, but surely shouldn't be considered legitimate.
The three exceptions neatly line up with those specified by the Nuremburg Principles, as helpfully laid out by our friends at Runes:
1. When a government wages a genocidal war against its own population;
2. When a government attacks another country without provocation; or,
3. When a country is faced with a humanitarian crisis of such proportions that its own government is unable to respond due to internal instability or civil war.
(I say that these line up because I think the third sort of situation is the kind that Max was imagining for his third exception--as in Rwanda, for instance, where thousands of people were slaughtered with machetes.)
As M.J. O'Brien at Runes puts the point, the Nuremburg Principles are supposed to disallow "wars of choice" in general. These exceptions to principled pacifism are supposed to specify the cases in which a war is necessary and not a matter of choice.
I worry about this way of putting the point, though, because strictly speaking even the most clear-cut cases of defensive war involve choice. Surrender is always an option, and sometimes the best. (Surely, for example, it would have been better if Germany had surrendered instead of defending itself when invaded at the end of World War II, even though by that point it was clearly a case of defensive war and hence "necessary".) Certainly, any non-defensive foreign intervention, even to stop a genocide, is also a matter of choice; it might be the best choice, but it's still not strictly necessary.
More generally, though, I worry about this method of creating systematic exceptions itself. Just about any war can be sold as self-defense. The Vietnam war was sold in those terms, and certainly Bush tried to convince us that Iraq posed an immediate threat. Just about any war can be sold as humanitarian. Just about any war can be sold as easy and cheap.
I'm concerned that even if these principles were to gain wide acceptance, we would be just as likely to get suckered into yet another unjust war. The principles are too vague and too easily stretched. I'm not sure they can be made more precise, though, since they have to apply to situations as-yet-unimagined.
The trouble with allowing any exceptions is that it's alarmingly easy to convince people that your pet war fits the exception. The trouble with not allowing any exceptions (principled pacifism) is that it allows genocide to go unanswered. Which is worse: allowing the Iraq war, or allowing Rwanda?
Certainly it's possible for reasonable people to come to a middle ground position here, but I'm not sure it's possible to enshrine this position in terms of general principles that will actually achieve the desired result.
I'm inclined to think that if we have to choose a set of principles, they should be wholly pacifistic. Allowing for particular exceptions seems far less dangerous to me than allowing for general exceptions. The problem with this, of course, is that it means that we have to consider every potential war as a particular case. But we already do this.
So how do we avoid the next war? Try not to elect idiots, and try to keep idiots from acting out their war fantasies when we do. Take it to the streets, and take it to the ballot box. That's the best we can do.