So as soon as you want something to happen you begin skewing the data to support it. Our stuff is invaluable to decision-makers precisely because we have no ax to grind.

We cannot anticipate the moral circumstances under which we will live. And therefore [I refuse] to say that I will not live in a moral circumstance that requires violence.

I think the personal and psychological aspects of war remain the same. War is about killing and dying. A man or woman stands at the post and there is a very real possibility of dying in the next five minutes. Whether he dies or not depends partly on him and partly on luck, and yet he must continue to function.

War and sex are what I call the two abysmal - by that I mean deep - parts of the human condition.

The Chinese central government will slowly and steadily lose authority while regional armies [gain power]. The Western powers are going to take sides to protect their investments - they have put billions of dollars into Shanghai. Their fear is that [these investments] are going to be expropriated by a warlord from the interior who will sweep down on Shanghai. They will try to form alliances with warlords to protect their concessions, and there will be a huge flow of weapons into China.

But every time new powers emerge they have to find their balance. New powers are emerging, old powers are declining. It's not that process that's dangerous, it's the emerging position that's dangerous.

Constraint theory argues a number of things. First, that the impossible has to be identified. Second, that the actor is then constrained by circumstances to act a certain way. For example, should we invade ISIS? Can we invade ISIS? What would it take to invade ISIS? Once you ask that question you discover the price of that option and then you take a look at American politics and see that the country is probably not prepared to invest the 2 to 3 million people that it would take to defeat ISIS and the insurgency afterwards. All right, so that's not going to happen.