----- Original Message -----
Sent: Thursday, April 22, 2004 12:07
PM
Subject: Function, evolution, emergence,
and modeling...
Hi Folks,
I've been swamped with work, but as I've been working, I've
also reread some areas of my father's work that I think are worth
bringing into discussion, particularly as they bear on the "Rosennean Modeling
Project". In the
General Introduction to "Anticipatory Systems", right after a short
discussion of modeling and side effects, he begins a new discussion:
"The theoretical principle underlying this analysis of failure in
anticipatory control systems is not wholly negative. In fact, we shall argue
later that it also underlies the phenomena of emergence which characterize
evolutionary and developmental processes in biology. It may be helpful to cite
one more excerpt of a paper originally prepared for the Center Dialog, [Robert
Hutchins' "Center For the Study of Democratic Institutions"]. which
dealt with this aspect:
"It may perhaps be worth noting at this point that the above
phenomenon is responsible for many of the evolutionary properties exhibited by
organisms, and many of the developmental characteristics of social
organizations. Consider, for example, the problems involved in understanding,
e.g. the evolution of a sensory mechanism such as an eye. The eye is a
complicated physiological mechanism which conveys no advantage until it
actually sees, and it cannot see until it is complicated. It is hard to
imagine how one could even get started towards evolving such a structure,
however valuable the end-result may be, and this was one of the major
kinds of objections raised by Darwinian evolution. The response to this
objection is essentially as follows: the proto-eye in its early stages was in
fact not involved in the function of seeing, but rather was primarily involved
in carrying out some other functional activity, and it was on this other
activity that selection could act. If we now suppose that this other activity
involved photosensitivity in an initially accidental way (simply because the
physical structure of the proto-eye happened to also be photosensitive), it is
easy to imagine how selection pressure could successively improve the
proto-eye, with its accidental sensory capacity, until actual seeing could
begin, and so that selection could begin to act on the eye directly as an eye.
When that happened, the original function of the eye was lost or absorbed into
other structures, leaving the eye free to evolve exclusively as a sensory
organ.
"This "Principle of Function Change" is manifested even more
clearly by evolution of the lung as an organ or respiration. Many fish possess
swim bladders, a bag of tissue filled with air, as an organ of equilibration.
Being a bag of tissue, the swim bladder is vascularized (possesses blood
vessels). When air and small blood vessels are in contact, there will
necessarily be gas exchange between the blood and the air, and so a
respiratory function is incipient in this structure, designed initially for
equilibration. It is easy to imagine how successive increases in
vascularization of this organ, especially in arid times, could be an
advantage, and thus how selection could come to act on this structure as a
lung. This Principle of Function Change is thus one of the cornerstones of
evolution (and indeed of any kind of adaptive behavior), and it depends
essentially on the fact that the same structure is capable of simultaneously
manifesting a variety of functions."
Thus the basic problem of avoiding infinite regresses in
anticipatory control systems could be reformulated as follows: Can we design
systems which are proof against a principle of function change?
This was the circle of ideas which I was led to place on the
table at the Center. The response elicited thereby could perhaps best be
described as restrained, but encouraging. I received some comments to the
effect that my approach was logical and mathematical, and hence fundamentally
inapplicable to politics. In particular, there seemed to be no room for
perversity, a major factor in human behavior. Indeed, I had often noted that
almost the only way for man to prove that he is truly free is to deliberately
do himself an injury; to deliberately act against his obvious best interests.
But I did not feel that this sort of objection was insuperable. Mr. Hutchins
himself made few direct comments, except at one point to remark that one
of my conclusions was the most outrageous thing he had ever heard. I took this
as a high compliment.
--snip--
"For my own part, I continue to believe that the properties of
Anticipatory Systems raise new questions for the scientific enterprise of the
most basic and fundamental kind. These questions have led me to reformulate
and refocus all my previous work in the foundations of theoretical biology,
and in the relation of biology to the physical and human sciences. Indeed,
there is no aspect of science which can be untouched by so fundamental an
activity as a reconsideration of the reactive paradigm itself. The results of
this reconsideration, and its implications, are the basic subject
-matter of the developments which follow.
Now that we have reviewed the genesis of the theoretical problems
with which the present volume [Anticipatory Systems] is concerned, we may turn
to a consideration of the problems themselves. The first and most basic of
them is simply this: What is a model? What is the nature of the relation
between two systems which allows us to assert that one of them is a model for
the other? The basic property of the modeling relation is that we can learn
something about a system in which we are interested by studying a model of
that system. "
--snip--
I think the discussion on the subject of
models and side effects is also useful, so I will post that in the next
couple days.
Cheers,
Judith