optimal social structure and the study of history

Eric Watt Forste (arkuat@factory.net)
Tue, 30 Jul 1996 18:53:43 -0700

At 10:14 AM 7/30/96, Robin Hanson wrote:
>>The question of the optimal social structure is explorable primarily by the
>>study of history, a fact that futurist technophiles are often prone to
>I take issue with "primarily". Data about current social structures
>can tell us a great deal - you don't have to go back in time for data
>to be useful. Laboratory experiments also tell us a lot. And theory
>is very important for giving one concepts and hypotheses with which to
>understand empirical data. Without patterns to look for, history
>doesn't say much.

Certainly history in and of itself doesn't do much to supply us with
theoretical structures. Theory is a completely different issue. I'll take
good theory from whereever I happen to find it.

The thing that really sucks about history is that there's never a control
group. But history is still an incredibly rich source of data about
structures that are too large to study in a laboratory, and, because they
involve human beings who sometimes messily and unpredictably invent new
theories to guide their own actions, too complex to simulate. If laboratory
data and simulation data conflict with historical data, my own cautious
preferences lead me to side with the historical data. If a computer
simulation or a laboratory experiment were to lead to a conclusion that
Diocletian's policies would lead to a thriving and wealthy Roman empire, or
that Stalin's policies would lead to a thriving and wealthy Soviet Union,
I'd be *inclined* to discount the laboratory/simulation data on that basis
alone. Before I read Hayek I was a lot more cavalier about these things.

Eric Watt Forste <arkuat@pobox.com> http://www.c2.org/~arkuat/