A country must first have exhausted all other means of defending itself. The attack should be proportionate to the objective. And it must stand a reasonable chance of achieving its goal. On all three of these tests Israel is on shakier ground than it cares to admit.
I'm inclined to agree with The Economist that Israel is on "shaky ground" with respect to tests 1 and 3, but I really wonder about test 2--proportionality. Chris Dierkes of Indistinct Union has posted a thoughtful entry arguing that the criterion of proportionality cannot be meaningfully applied to the Israeli-Hamas situation. He quotes Shmuel Rosner:
But such is the tricky nature of modern warfare: How do we measure proportionality without reducing the concept to an impossibly pedantic tit-for-tat? (How would it work? For every rocket launched into an Israeli town, Israel would retaliate by launching a similar rocket? And even then, how could we achieve proportionality without making sure that Palestinians in Gaza have the same alarm systems and comparably effective shelters?) How do we measure “success” in a situation where no side is likely to bring real closure to a volatile situation?
Dierkes answers as follows:
Easy answer: you don’t and can’t. The doctrine of proportionality (which has its roots in classical just war theory) was announced–as Rosner correctly points out–in the 1907 Hauge Conventions. That defined the era of Clausewitz’s Trinitarian Theory of Warfare: government, armies, and populations. The first two fight the war, the third does not. In the post-nuclear, post WWII era, of Fourth Generation Guerrilla Insurgencies, propotionality is a meaningless term because what you have is the equivalent of a swarm of bees stinging an elephant. What would be a proportional response of an elephant to the bees? It doesn’t make any sense. Proportionality only works as a guiding principle within the bounds of a nation-state with a professional army built for conventional warfare. The other side in this conflict does not have that edifice of social organization and therefore is not going to abide by those rules. Meanwhile for the side that does (Israel in this case), the technological difference is so vast, there is no way (as Rosner himself points out in the article) for there ever to be a proportionate response. All the elephant can do is step on some bees. There is no proportionate elephant equivalent to a sting.
I'm inclined to agree with Dierkes that we cannot reasonably apply the concept of proportionality to the situation at hand and to other instances of asymmetric warfare. So, what do we do instead? Use only tests 1 and 2. or add a new test to the mix?