From: Robin Hanson (hanson@hss.caltech.edu)
Date: Fri Mar 07 1997 - 18:01:10 MST
Lee Daniel Crocker writes:
>Much of our day-to-day activity can be accomplished purely by
>subconscious pattern-matching. Even something as complex as driving,
>once learned, becomes a self- reinforcing pattern stored and brought
>up when needed, but not actively reasoned. ...
>What more controlled, quantitative analyses are needed for is the
>formulation of rules that are likely to hold for completely new
>situations, and that can be communicated meaningfully to people
>who have not shared your specific experiences. Social predictions
>fall into this category, as do scientific hypotheses.
>... they extrapolate how to deal
>with these new things. This is probably a useful function of the
>brain for dealing with life, but it is a primitive, incomplete, and
>purely heuristic way of doing it. More trustable decisions demand
>more rigorous methods. Especially when the minds of other free
>individuals enter the picture. ...
>I am guilty of using such extrapolations sometimes. ... to imply that
>governments, as a rule, aren't a good idea.
I fully grant that the more explicit you can make your reasoning, the
better you and others can trust it. On the other hand, formal
explicit reasoning takes more time and is more costly. The choice
between them depends on the extra costs and benefits in any given
situation.
I just don't agree that the costs of formal explicit reasoning always
outweigh the benefits as soon as you deal with social issues or try to
persuade other people. Things are really much more complex that
that. Nor do I agree that the predictive value of informal
analysis falls to zero in such contexts.
Our ordinary lives are full of social issues and persuasion contexts
where we appropriately use just as much hueristic analysis as you do
when driving. No need to feel guilty about driving heuristically nor
about heuristic social analysis. But of course one's confidence in
any one prediction should take into account the type of analysis used,
the solidity of the data it was based on, and the strength of
connection perceived between the data and related models and predictions.
Robin D. Hanson hanson@hss.caltech.edu http://hss.caltech.edu/~hanson/
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:44:14 MST