Taleb, Ignorance and Action

August 27th, 2009

Views: 335

Posted by ChrisG at 3:57 pm

Over at the Guardian, Nassim Nicholas Taleb responds to judgement being passed on him as a representative of “eccentric” views – such as denying the influence of humans on climate change, and stating that burst bubbles and economic crashes tend to be “a good thing”. In doing so, he claims for himself an intrinsically conservative, even precautionary position on risk, based on the centrality of ignorance about the future to human activity.

Climate experts, like banking risk managers, have failed us in the past in foreseeing long-term damage. This is an extension of my general belief: “Do not disturb a complex system.” We do not know the consequences of our actions (this idea also makes me anti-war), and I have explicitly stated the need to leave the planet the way we got it.

As mentioned hereabouts some time ago, Taleb hasn’t written anything terribly original – or for that matter (at least to my tastes) readable – regarding the relationship between uncertainty and risk. Labour spokespersons’ opinions to the contrary, it’s hardly an “eccentric view” to point out that optimistic assumptions about the manageability of risks generated within complex systems tend to rebound on the assumers, and that low-probability, high impact events should inform sensible (and socially just, we might add) approaches to risk. On this, Taleb’s snackthoughts are in agreement with most environmental NGOs you care to mention, the IPCC, the Sustainable Development Commission, and so on and so forth.

But his reaction exposes a recurrent problem for any precautionary approach to uncertainty: how do you stop such an approach retreating into a generalised, risk-averse conservatism? Sure, we do not know the consequences of our actions – yet we must act. And here’s the paradox; we must act in concerted, organised and adventurous ways if we want to avoid tipping over complex systems, and even if we want to “leave the planet the way we got it”. So there has to be a basis for such action in the face of uncertainty, yet it cannot take its sole impetus from the kinds of naturalistic, past-based knowledge that traditional managerial approaches to decision-making tend to rely on, because – when we’re talking about problems like anthropogenic climate change, responsible technological innovation, or the right way to formulate energy policy – we’re talking about future-oriented responsibility, and therefore precisely about anticipating the near- and long-term consequences of our actions. But here, as Jean-Pierre Dupuy points out, the Kantian criterion does not hold. It is no longer a situation where ought implies can. In fact, we ought to do precisely what we cannot do.

An adequate account of the obligations, permissions and prohibitions that should apply in this sphere of action would necessarily be complex. But some of the arguments contained in this paper [PDF] might be an inspiration. Hopefully I’m going to write some more about Stefan Skrimshire’s work in more detail with respect to this action problem, over the next ten days or so.