Friday, December 02, 2005

The limits to knowledge

Cicero attended a very important discussion last night. In the rather stuffy surroundings of the Travelers Club, he heard a revolutionary: Nassim Nicholas Taleb. Taleb has a profound insight: that humans perceive their behaviour as part of a symmetrical world. In fact it is highly non symmetrical. Partly as a result, humans are extremely bad at making accurate predictions and the more detailed the prediction, the more likely it is to fail. The implications of this are very profound indeed.

Taleb rails against the "scandal of prediction"- and is particularly fierce in his denunciation of politicians. Politicians draw up detailed plans for conditions that are inherently unpredictable- he suggests that this is little better than a fraud against the electorate. He is right. Cicero does not believe that detailed government policies can change outcomes in predictable ways. Macmillan's famous comment on the primary challenge to political leaders being "events" refers to the fact that a successful political leader, almost by definition, can only react to the conditions that they find in office. Rarely do policy proposals actually achieve their expected goals in a highly predictable way. Taleb's collaborator, Benoit Mandelbrot, has pointed out that there is a mathematical state of chaos- the so called fractal geometry- that takes the consequences of given events in unpredictable directions, and no where is this more true than in the social sciences and especially politics. Politics is almost a stereotype of what a fractal system might look like.

This is why the system of political discourse matters more than specific policies. The collective of a system is far more predictable in aggregate than any individual feature in a system. The only predictable thing about political outcomes are that they are highly unpredictable. The Liberal commitment to retrenchment reflects that fact that it is not only desirable the state should be limited- it is ultimately inevitable.

The work of Hayek and Popper contains some of this recognition of the centrality of uncertainty and that is why they are philosophers more than economists. There are critical limits on what information may be accurately inferred (predicted) in detail. This is a systemic problem. From the perspective of Economic Liberalism it is one of the key insights that informs our ideology. Taleb is bringing to the science of the self, the same revolution that Mandelbrot has brought to the science of causation. Essentially if we are limited to systemic prediction and that detailed prediction is structurally inaccurate, then the key for political leadership is going to be the question of structural limits to the system and not the false god of detailed policy prediction.

The link above will lead you to Taleb's website, but as far as detailed plans by politicians by concerned, then the last word should belong to Yogi Berra:

"You gotta be very careful if you don't know where you're going, because you might not get there".


dearieme said...

It is because "events" matter so much that one should disincline to elect politicians of low character - I am thinking of Clinton, W and Blair specifically.

chris said...

Interesting, people try to see symmetry even when there is none. This does explain one factor in politics. Why the 'left' is automatically considered to also be liberal.

In most developed world countries most politicians views are to the 'right', that is using markets where possible, and there is a tendency for them as they go further 'right' to get more authoritarian as well.

Seeking symmetry it is natural to infer from this that the further left economically a politician the more liberal. However, with some exceptions, you find a similar thing happening on the 'left' as the 'right', the more extreme they get the more authoritarian they get.

maswey said...

thank you nice sharing

cep program