COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	comple[f86,jmc]		Dealing with complexity
C00005 ENDMK
C⊗;
comple[f86,jmc]		Dealing with complexity

Suppose a phenomenon is too complex for a given person to understand.
Indeed even if it isn't too complex for a person to understand,
it may be too complex for a group to jointly understand, especially
if it has political aspects.  By having political aspects I mean
that power relationships within the group are substantially affected
by whether certain aspects of the phenomenon are recognized.

Let's assume that the phenomenon has to be dealt with.  In this case
the individual or group will take an oversimplified view of the
phenomenon and base policy on this oversimplified view.  The oversimplified
view will often consist of regarding one dimension of a multi-dimensional
situation as most important, and the policy will involve optimizing
this dimension, not taking into account the effects of the policy on
the other dimensions.  The policy may seem to work for a while,
but if the view is genuinely oversimplified, the policy will turn out
to have some unwanted consequences, and these consequences may
become more apparent over time.  Specifically, one of the other
dimensions of the phenomenon may assume a value far from optimal.

Attention then switches to this dimension, perhaps on the initiative
of different people than those who set up the first policy.  A new
oversimplified view becomes prominent emphasizing the new variable.
A new policy is adopted with short term benefits and which also has
some long term unfortunate conseqences.  The cycle continues.

The questions are the following.  Under what conditions will this
process converge to a reasonable policy?  Under what conditions
will it lead to increasingly violent oscillations?