[From: Jeff Dooley 920323.1100]
(Bill Powers 920322.1700)
Since becoming a subscriber to this net some months ago I
have learned a great deal, and many concepts of control
theory which had eluded me in the past have become, if not
clear, at least (apparently) less confusing. For having
facilitated that learning I want to thank you all.
I'd like to comment briefly on two issues that popped out at
me from Bill's post: (1) the scientific/philosophical
difficulties surrounding the concept of "purpose," and (2)
the phylogenetic transmission of "control machinery." I'm
beginning to think of these as potentially related issues.
Talk of "purpose" is routinely dismissed among scientists
and philosophers, as Bill suggests in his post and in his
article "On Purpose" from _Living Control Systems_. But why
does the concept have such a bad reputation? Further, the
question arises, can the notion of "purpose" be
decontaminated for sci/phi discourse?
A possible response to the first question centers on the
general anti-intentionality of logical empiricism and,
later, of philosophical behaviorism. These traditions
seemed to assume that "purpose" was cognitively meaningless
since it was assumed that such a property was intentional--
that there was no possible observable phenomenon to which it
could be connected. This appears to be a fundamental
assumption which PCT calls into question. According to PCT,
as I understand it, purpose may be inferred, inductively, as
the maintenance a controlled phenomenal variable at or near
reference. Find a phenomenal variable apparently controlled
under perturbation and you have located evidence of purpose.
As a response to the second question, could not the PCT view
of purpose as a matter of observation in this sense help
decontaminate the concept for philosophy and science?
Meanwhile, what's worse for the notion of "purpose," it has
got entangled with the Aristotelian concept of teleology, by
which, as Ernest Nagel has derisively quipped, future events
are seen as agents of their own realization. To linear-
thinking, cause-effect, (S-R?) philosophers of science, this
sounds like laughable illogic. While under a linear
model of causality (development, progression, etc.) such a
notion as "purpose" appears a paradigm example of bad
science, reconsidering it under a circular, equilibrating,
control model may cast the whole situation in a new and
possibly fruitful light. This brings me to my second issue:
the idea of the phylogenetic transmission of the
reorganizing system, and to how I see these two issues as
related.
I'd like to suggest the idea, blending concepts of control
theory purposiveness with a thesis of biologist Stuart
Kauffman's: that species (complex, adaptive systems) evolve
(drive themselves) to the edge of chaos and maintain
themselves there. For Kauffman, "the edge of chaos" is the
phase-transition space at which order melts into chaos. Too
much order and the system can't adapt to perturbations; too
much chaos and development can't get a foothold from which
to progress. If we postulate an "organizing function" of a
species (or any cocontrolled variables.
Could not a homeostasis at the edge of chaos be considered a
controlled variable? Some current work in biology does
consider the idea of trans-individual organizing functions
working on the phylogenetic level, (in fact, on all
hierarchical levels from RNA to species!), but they have
not, to my understanding (not being a biologist),
figured out a way of modeling the dynamics of progression.
It is precisely in the reintroduction (decontamination)
of the concept of purpose in scientific discourse, under PCT,
that we might begin to progress in our learning about the
possible organization functions of complex adaptive systems,
on whatever recursions of organization they become apparent.
Finally, a footnote on different ways of "seeing" the world:
The two paradigms of S-R, open-loop, etc, and PCT, closed
loop, etc. seem incommensurable in a Kuhnian sense. Workers
in the different traditions really do seem to think, speak,
operate in, or at least "see," different worlds. This could
be serious, since theoretical development, experimental
design, and other activities influencing the direction of
science are informed, one could say, as a function of which
world one belonged to. Could incommensurability in this
sense help explain the apparent fact that S-R types and
PCTers seem to just talk past one another?
jeff dooley
dooley@well.sf.ca.us