Merging at the edges

A (hopefully measured) rant about sociobiology / pyschology / behavioural modelling, in response to an email on the SIMSOC list, rehashing themes covered on this blog:

James Scott's "German forest-death" parable is a perfect way of getting to the nub of this problem, and I'll try and paraphrase it in a few sentences if you're not familiar with it: C18th Germany wants to know how much revenue in trees they have, so they send people through the forests with bags of various sized nails. Counting the remainder, they know the number and size of trees and can plan their budget accordingly. Some time later, after much felling, the next logical step is to plant trees of a similar number and size. This works for a little while - but slowly yields drop, until eventually the forests are entirely unproductive. The reason - of course - is all the aspects of the forest's ecosystem not accounted for by how many nails they had left. They were measuring one aspect of the ecosystem, but when they applied a view of the world derived *from* those measurements back onto reality, the ecosystem bit them in the ass.

How is this relevant? Anyone modelling - especially anyone aiming for 'policy relevance' - is doing something equivalent. I'm not arguing that makes it invalid, or that we should all be Oakeshottian quietists: clearly, e.g. the British census only gets a tiny sliver of info, but it's useful info. However, we have to be very self-aware when we're talking about the consilience of modelling, social science, evolutionary biology and psychology.

An example of why I think this is so problematic: here's Tyler Cowen reviewing Shermer's "the mind of the market":

... the original environment in which people evolved, namely the small groups in hunter-gatherer societies, helped people develop altruism and cooperative behavior. At the same time, our biological heritage drives human discontent. Our evolution on the East African plains did not equip us to cope emotionally with the large-scale and often impersonal nature of a modern market society. Small tribal groups are used to sharing, not to extremes of wealth. That is why envy is rife. In other words, he says, you may feel that modern capitalism is unfair if you apply the outmoded moral code of the tribe. Shermer believes we nonetheless should look to the market as the dominant mode of organizing social affairs.

Hayek's argued exactly the same and had been since the 30s: human society had evolved from an atavistic altruism into the "Great Society", where prices worked through the Catallaxy for the good of all. There was no going back - and anyone suffering from pangs of conscience should be taught to understand: it was not their place to challenge an evolved system they couldn't truly comprehend. Here's Hayek:

By a process which men [sic] did not understand, their activities have produced an order much more extensive and comprehensive than anything they could have comprehended, but on the functioning of which we have become utterly dependent.

Here's where it starts looking like Scott's parable: Hayek argues that we are even now 'born socialists' - evolutionarily inclined towards altruism and solidarity, as Shermer says - but for the continued functioning of the Great Society we need have our solidarity educated out of us. Lord Bauer argued a few years back - "economic development rests on people having the right desires and aptitudes, and on a political and legal system that allows people to act on them." It's these desires and aptitudes that Hayek suggested we needed to learn (whilst unlearning contrary ones like the 'envy' Cowen mentions.)

A long while back I did some work looking at the World Bank's projects in Africa: the Bank's aims were to reform institutions (where institutions can include "behavioural norms") by placing them into two categories - either "market functional" or "market dysfunctional" and then "supplant or amend" the latter. This is only a slightly more politically correct version of Senator Henry Dawes, after a visit to the Cherokee in 1885 -

The head chief told us that there was not a family in that whole nation that had not a home of its own. There was not a pauper in that nation, and the nation did not own a dollar. It built its own capitol, and it built its schools and its hospitals. Yet the defect of the system was apparent. They have got as far as they can go because they own their land in common. It is Henry George's system, and under that there is no enterprise to make your home any better than that of your neighbours. There is no selfishness, which is at the bottom of civilisation. Til this people will consent to give up their lands, and divide them among their citizens so that each can own the land he cultivates, they will not make much more progress.

There's a peculiar twist of logic going on here: 1. we are the way we are because we evolved this way; 2. but our evolved altruism is atavistic and should be socialised out of us if we're to function in modern society.

I'm not saying anyone on this list subscribes to that view - but I am arguing that, unfortunately, there's absolutely no way you can put your slippers on, put your feet up and say, "we know the science about human behaviour is correct: we are this way." I'm also not trying to re-state that old tired argument: "norms are nothing more than a social construct". I'm all for a strong empiricism about human nature; there's some excellent work on the commonalities we have as a species - the best I've read thus far is still:

J Henrich et al., "Economic man" in cross-cultural perspective: Behavioral experiments in 15 small-scale societies," Behavioral and Brain Sciences 28, no. 6 (December 2005): 795.

But the models we make are like Scott's nails in the trees. We use theories about ourselves not just to describe, but to reproduce society, to educate ourselves and - e.g. through the World Bank - to carry out well-funded state- and infrastructure-building projects. We should all have Scott's parable stapled to our foreheads, so we don't forget: we change the systems we describe, when those descriptions become part of how we enact policy, and reality is always more complex than we think.


And after we've stapled

And after we've stapled Scott's parable to the heads of all the agent-based modellers we could count the staples left and know how many modellers we can cut down and burn each year...

But seriously, amen Dan, amen