On the psychology of defence


There are two bits of Clausewitz (just two?) that have long puzzled me. The first is where he says that a statesman should know what sort of war he’s embarking on, before getting into it. In modern times, that line has often been contrasted with John Reid’s unfortunate quote about Helmand, in which he said the British would be perfectly happy to leave without having fired a shot. The second is where Clausewitz argues that the defence is stronger than the attack.

I still don’t understand the first. Prediction in complex social systems is impossible – and any foresight more a matter of chance than genius. Philip Tetlock gets into that in his last two books – on the value of expert political judgment (negligible) and the possibility of ‘superforecasting’. If we read Clausewitz’s remark as a cautionary note against overconfidence, then fair enough. But retrospectively expecting elites to have a firm grasp on the character of a future war seems a bit unfair. Strategy is in part improvisational; responding creatively to the unexpected in situations that are always in flux.

On the second bit of Clausewitz, I have an answer! For the real aficionados, the full paper, presented at the ISA conference this week, is online here. The short answer is here – and it comes with a heavy dose of psychology.

In brief, I argue that Clausewitz is right, and that Daniel Kahneman and Amos Tversky’s Prospect Theory can help us see why. This theory suggests we are profoundly loss averse. We are prone to gamble more when we perceive ourselves to be losing than we would if winning. You can demonstrate this very neatly in lab based psychology experiments which set up precisely the same (cash) payoffs and probabilities associated with two scenarios, but depict one scenario as the choice between a gamble on avoiding a loss, or the acceptance of a certain, but smaller loss; and the other scenario as a gamble on a large gain, versus the possibility of a smaller, but certain gain. We tend to gamble in the loss scenario, but not in the gain.

Why? A bird in the hand is worth two in the bush. Evolutionary psychology provides a compelling rationale – in marginal environments, where survival is rather finely balanced affair, losing valuable resources could be disastrous; but at the same time it is profoundly dangerous to risk everything on the possible acquisition of new resources. And so in ‘wars’ between groups of chimpanzees overwhelming odds are required before the chimps will attack. The tactics of raid and ambush are favoured over pitched battle – as they often are in wars between hunter gatherer communities.

Some important caveats: In the real world, of course, the payoffs and probabilities associated with them are rather messier than in the psychology laboratory. That’s certainly true in war, as Clausewitz noted with his extensive discussion of uncertainty and chance. There are also plenty of real world examples of attack dominating defence, in battles and campaigns alike. How are we to square this with Clausewitz’s assertion? It’s hard to control for the many particulars of historical examples – terrain, technology, numbers, cohesion and so on. Sometimes it’s hard even to know which is the attacking force, and which the defender, as attacks culminate, and defence gives over to the counter-attack.

Fair enough. And yet, at some meaningful level of abstraction, Clausewitz’s remarks chime with Prospect Theory. In a struggle for possession (Clausewitz thought mostly about possession of territory, but there may be other plausible currencies in which to measure appetite for risk), there is a tendency to gamble in order to avoid the prospect of losing. The attacker, meanwhile, will be increasingly reluctant to gamble on marginal gains, especially if his initial objectives have been satisfied.

What we have, we hold.

For more on Prospect Theory, check out Daniel Kahneman’s excellent popular psychology book, Thinking Fast and Slow.

Image via flickr.