Re: playing God

DON MIKULECKY (MIKULECKY@VCUVAX.BITNET)
Fri, 7 Apr 1995 10:06:21 -0400


Don Mikulecky, MCV/VCU, Mikulecky@gems.vcu.edu
Reply to Bruce

> In response to "DON MIKULECKY", who wrote:
>
>>> We would then reserve the most complex
>>> category for self-referential systems that have a non-simulable
>>> semantic component. Will that fly?
>
> Cliff Joslyn writes:
>
>>Sounds good to me. Now try to convince THEM.
>
Bruce wrote:
> Does this mean that the most complex systems are ones which can (1) map
> their own internal models, as well as (2) keep some track of external
> events on which they do not (yet) have fully adequate information ?
>
Yes. This is what Rosen calls "Anticipatory Systems" (the title to
his next to the last book).
> If this is what a complex system means, more or less, what is the problem
> in acceptance ?
The main problem is that these systems are basically not simulable.
This is because there are non-computable aspects to them (The halting
problem, Goedel, semantics vs syntax, etc.) Yet the main proponents of
the new Complexiry research, Artificial Life, etc. (represented in
large part by the Santa Fe Institute group and its affiliates) mostly
do simulation!!! If you look at what they call complex systems, they are
really very sophisticated mechanisms all comutable.
>
> On the other hand, _the *most complex* category_ might need to add some
> capacity for goal-directed reorganization of its own self-referential
> systems - i.e. creative restructuring in the pursuit of increased freedom
> (e.g.less dependency upon the contingencies of *non-simulable semantic
> components*.
>
This sounds reasonable.
This is already part of Rosen's scheme. He really requires final cause
(a form of goal directed behavior based on internal self referential models).
In his and Casti's treatment of M-R (Metabolism-Repair) systems there
is some discussion of which components are modifiable.

I begin to see what must come next. The M-R system is Rosen's
relational model for the organism. However, organisms do not exist
as independent entities. What we need to do now is to formulate
the relational aspects of ecosystems and societies. It begins to
look doable!
Best regards, Don Mikulecky