> There's some insight into design principles that I can offer from
> explorations at the Santa Fe Institute. Stu Kauffman has been doing
> work to discover how to optimise the possibility of solving a problem
> by dividing the agents working on it into cells (he calls them
> "patches"). If the problem is simple - that is it doesn't require a
> great deal of information processing or creative input - then a
> relatively centralised system is frequently the best. If the problem
> is a difficult one requiring much information processing, creativity,
> experimentation, etc. then the agents should be broken into _self
> optimising_ teams and left to sort out their own solutions [snipped here]
> --
> Michael McMaster <Michael@kbddean.demon.co.uk>
>
So far, I have a good record for agreeing with Michael, but here is
something I want to argue with. Two point are involved. The first one
expresses my bias. The second one expresses my firm belief.
SELF-SEALING ORGANIZATIONS. I think the term "self-sealing
organizations" came from Chris Argyris. In any case, the Santa Fe
Institute is such an organization. I can think of several others that
also have this capacity. I can also think of numerous professional
societies that have that attribute, along with some "semi-pros".
One of the biggest reasons to be very unhappy with self-sealing
institutions is that, when they have considerable prominence, i.e., are
high profile, they tend to be regarded as experts and innovators by
people who are too busy to research the whole literature, and who may
believe that no high-profile organization would ever be so foolish as to
pretend to be the world's leader in what they are doing.
THE CONTEXT-CONTENT-PROCESS triad. One of the most common mistakes in
working with groups in organizations, often carried out by social
scientistw who ought to know better is to assume one or both of the
following:
(a) Because a person has an excellent repertoire of content knowledge
about some field or issue, they automatically have the power to invent
or choose effective process on the spot.
(b) Because a person has an excellent repertoire of content knowledge
about some field or issue, they automatically have the power to envisage
the broad context within which they are working.
I can tell you from very extensive study, both in the literature and in
practice, that both of these assumptions are almost always wrong.
Unfortunately, it is not that they are wrong that is the big difficulty.
The big difficulty is that, when people behave as though these
assumptions are correct, the consequence is almost always the same in
essence: no progress is made, and when a decision is taken it makes the
situation worse. A further consequence often is a mistaken belief that
progress has been made and less energy needs to be expended to reach a
useful result. A global consequence is that the search for much better
approaches is foregone, so even when such approaches are available, and
case after case shows that they are better, they continue to be largely
ignored. Adding up all the instances of this kind of situation gives a
very large sum.
To put it another way, "Superficial Intelligence is much more
dysfunctional than Artificial Intelligence".
Aggressively yours,
-- JOHN N. WARFIELD Jwarfiel@osf1.gmu.edu