Re: Isn't spontanious order overrated?

Dan Fabulich (
Mon, 07 Sep 1998 16:08:55 -0400

Anders Sandberg wrote:
>"Max M" <> writes:
>> Especially in politics and economics spontanious order and the invisible
>> hand is hot. But if the free market is the answer to most economical and
>> political problems, then it's a bit like saying that intelligence has
>> value. If simple interconnected systems can perform wel in complex
>> enviroments why the have intelligence done so well until now.

I missed the original post here, but it's a false dichotomy to place human intelligence on one end of the stick and spontaneous order on the other. Ultimately, human intelligence runs on spontaneous order; no one runs their brain from the top-down.

>> Then my question is, at what level will spontanious order work best?
>Exactly. How do we find out?

You impose a block on your thinking if you view it this way... Ultimately all of our solutions to problems will involve spontaneous order if we solve them with our brains.

Perhaps a better question might be to ask when it is best to find all of the information and find the optimal solution, compared to when it is best to remain "rationally ignorant," as economists often put it. We do not have enough information to find the optimal solution to every problem which presents itself. In some cases, we can try and guess how likely it is that one solution or another will be correct, but even that isn't possible in a wide variety of cases.

In these situations, a strategy which relies on the adaptive behavior of systems will tend to work better. To go back to the economic analogy, if we could know everything everyone knows at no cost, we could easily make centrally-planned economies work. However, the cost of acquiring and using this information is so great that this strategy is unfeasible. Since we must remain in rational ignorance, the best strategy is to employ a system which does not require anyone to know everything about the whole system: capitalism is one such solution.

In comparison, we *can* know all or most of the relevant variables for some problems, and in those cases we can find the optimal solution from known principles rather than depend on something like a genetic algorithm which may get stuck in a saddle-point.

My $0.02.