You're right, that is something we do not directly address. But the implications of what I think is going on--which we highlight in the conclusions--are relevant to those Afghans as well. Ultimately it is all a question of co-opting, or getting buy-in. Because we see the insurgency as fundamentally a local thing, the Coalition's level of popular support is going to fluctuate from one village to the next. And if that level is the center of gravity--which I certainly think it is--then an Afghan who decides to throw in with the Coalition, in whatever capacity, is going to be very exposed. The militants in his village will know him and his family, and where to find them. You have to make the benefit of cooperation greater than the penalty of that Afghan potentially losing his family. Not an easy thing to do, but not impossible either. And so you want a strategy that is flexible enough to allow our COIN measures to look differently in different parts of the country. That is the main thrust of our argument.