Re: Kay & thermodynamics

["Tim Gwinn" <***>]

John K.,

Here is the other branch of the thread regarding Kay and thermodynamics.

> -----Original Message-----
> From: ROSEN Forum [mailto:*** Behalf Of John
> Kineman
> Sent: Friday, August 15, 2003 5:43 PM
> To: ***
> Subject: Re: Greetings
>
>
> Tim et al.,
>
> Thanks for reading so thoroughly! I'll try to respond, with the caveat
> that I don't want you to form too early an opinion about the ontological
> question. It is something I expect will require considerable discussion
> and may not have an easy answer.
>
> By the way, I assume by MR you mean modeling relation, not
> metabolism-repair, which perhaps can be designated as M-R
>
>
> Tim Gwinn wrote:
>
> >Hi John,
> >
> >Welcome to the group!  Thanks for including your bio and
> interests regarding
> >Rosen and his ideas.
> >
> >Your remarks bring up some points for possible discussion:
---snip---
>
> >2) Kay & complexity
> >I tried to read up on Kay, but alas, it seems the links from his page
> >(http://www.jameskay.ca/) to his papers are down. However, from
> this page on
> >thermodynamics and ecology (http://www.jameskay.ca/about/thermo.html), it
> >seems that he is deeply entrenched in the idea of thermodynamics as being
> >key to self-organization: "The emergence of self-organizing structures
> >provides an avenue for dissipation that otherwise would not
> exist". The idea
> >being that dissipation toward equilibrium is not only
> inevitable, but that
> >self-organizing structures are generated naturally by this dissipative
> >tendency towards optimizing the rate of dissipation. At least,
> that is what
> >I gather from it.
> >
> >But this tendency toward equilibrium is based upon a thermodynamics of
> >closed systems as *the* standard, in which equilibrium is an
> attractor for
> >the dynamics of *the* system. The use of the properties of
> closed systems as
> >a paradigm for all systems is, however, a poor one. To quote Rosen: "The
> >entropy function, which was a Lyapunov function for the
> autonomous dynamics
> >when the system was closed and isolated, will in general have no special
> >relation to the new dynamics [of the opened system], and loses
> thereby all
> >its original theoretical significance." [EL p. 251]
> >
> >
> What Kay does and what he believes may not fully commute. In other
> words, his thinking is more complex but he knows what to survive on. He
> has been successful in adapting analogies from chaos theory and thermo
> to ecosystems, identifying for example the phenomenon on "flips" using
> the concept of attractors of various stable states. I spoke with him and
> his knowledge of Rosen is limited, yet he is open and interested. The
> problem is linking with the "mainstream" so he sticks with the
> traditional stuff. We recommended Kay as an advisor to NSF's complexity
> program as a means of getting it one step out of a fully deterministic
> view. They accepted and Kay has been lecturing NSF staff. At the same
> time I recommended to Margaret Leinen (head of NSF complexity program)
> that she look into Rosen's ideas as a "next step." She bought a copy of
> "Essays..." and read it, but I'm not sure to what avail. Kay is more
> careful not to go so far that he loses their trust (a talent I have not
> learned).


I guess I am thankful that I am outside of academia and professional
science. :)  I am not sure I could operate in that way.


>
> >There are in nature no ecosystems that are closed thermodynamically and
> >therefore, those ecosystems will not have equilibrium as an attractor the
> >dynamics within it.
> >
> Its an interesting point and I'm sure it is technically true of
> thermodynamically open systems. But ecosystems are arbitrarily defined,
> so one can perhaps define the boundaries in such a way as to focus on
> conforming properties. The important question to me seems to be how open
> vs. how closed are ecosystems, as variously defined? Not sure it can be
> treated in a boolean manner (open OR closed).


I must defer to Rosen in Essays ch. 12 & 16. He argues the case in ways that
are out of my league. I can only refer to one small but significant mention:
that when a closed system is perturbed (that is, opened in some arbitrary
way to some arbitrary degree) the resulting system behavior could be almost
anything. Closed systems are so degenerate that they do not provide any
information about how the resulting open system will behave. The way in
which the system was opened will determine much of the resulting behavior.
[EL 184-185] If this is true (which I think it is), then proximity to being
thermodynamically closed will not generally coincide with proximity of
behavior to that of a closed system, so that 'how' open or closed a system
is will not be a useful guide to its behavior or attractors.


> It is clear that
> ecosystems, and certainly many of their components, do exhibit stable
> modes and certain kinds of equilibria. They incorporate many physical
> mechanisms. While equilibriom theory was clearly over-emphasized in
> ecology of the 50's and 60's (or was that a necessary step in its
> development?), and now system pattern-process dynamics are the rage,
> ecosystems nevertheless have both aspects. Perhaps it is a case, as
> Rosen described, where a fundamentally complex system can behave as a
> mechanism at times??
>
> > So, the use of entropy (based on closed-system
> >thermodynamics) as a causal agent for promoting
> self-organization within any
> >given ecosystem is invalidated.
> >
> I  tend to agree, but again perhaps not as an absolute. Certainly I
> believe that living systems must not be entirely subject to the 2nd law,
> otherwise there could be no "creative act." But to the extent that they
> involve mechanical sub-systems, 2nd law dyanmics would seem to apply.


Well, but if organisms are Rosennean complex, then their thermodynamic
qualities cannot be analyzed in terms of analyzing the thermodynamics
qualities of their parts, or their sub-systems. That would be
reductionistic, I think.


> Any given example seems to be a mixed bag. Perhaps it is the case that
> there are very few examples of a fully realized living system, with no
> components whatsoever that behave mechanically (except perhaps for
> quantum systems and superconscious Swamis?).


The components may well behave as mechanisms. But, once those components are
assembled into a complex system, the system will have complex behavior.


> It was a point that became
> somewhat tiresome in prior discussions, as one group would argue for
> complete mechanism and another for complete absence of mechanism,
> whereas any casual observer can see both.


A complex system will have complex behavior. However, importantly, a complex
system can be modeled temporarily and locally as a mechanism. If this were
not so, we could never have been doing all the science we have done
heretofore. It is simply that the complete description of the complex system
via models will always exceed any finite set of mechanistic (simple) models.


> But your point regards
> self-organization. To what extent can that be mechanical and to what
> extent not? Does this question translate into this one: To what extent
> is a "creative act" needed? If one says that a whirlpool in a bathtub is
> an example of self-organization, then there are self-organizing systems
> that can be explained mechanically (assuming we ignore the person taking
> the bath). But I certainly think more is involved in a living system. My
> tendency is to see it as a combination of both mech. and non-mech.
> processes.
>
> I am tempted to say that the 2nd law applies to all systems for which
> the ontology (origin - agreeing with RR's usage) is not considered part
> of the system. So the universe is said to conform to 2nd law entropy
> except for the big bang itself. Organisms may be analizable as
> dissipative systems conforming to 2nd law thermo, except for their
> evolutionary history and present day capabilities to involve "creative
> acts."


I think I disagree. As I understand it, the 2nd law applies only
statistically and only to the closed system as a whole. If we posit the
universe as a thermodynamically closed system (which I am highly dubious
of), then we can say that, statistically, the universe as a whole will tend
to dissipate toward equilibrium. This to me prescribes no specific
conditions toward dissipation upon any given small spatial region, such as
that occupied by an organism, or an organism within some ecosystem.


=========
On a broader level, one could take Kay's notion that (roughly) "tendency of
the ecosystem toward dissipation is a driving force for self-organization
within it" ought to be restated more generically as "tendency of the
ecosystem toward some attractor(s) is a driving force for self organization
within it".  Stated this way, free of thermodynamic language, it seems
clearer to me that Kay's proposal is deeply Newtonian: 'force' is something
that comes from the environment and determines how the system (the organism)
behaves.

But complex system arise not from externally imposed dynamics but from what
Rosen calls "gravitational" aspects within the system. It is possible to
freeze some organisms to near absolute zero and essentially remove all
dynamics, then thaw them and have the organism resume functioning. This, to
me, in itself obliterates any argument for a Newtonian approach toward
organisms. And likewise, obliterates any argument for some external driving
force toward dissipation (or any other attractor) as the basis for either
the arising or maintenance of self-organized systems like living organisms.
=========



> These are both ontologies.  Perhaps the entire mechanistic
> paradigm can be said to be built on the separation of an observable
> system from its ontology, hence the escape from mechanism would seem to
> requiring adding that ontology back in. I hear this loudly when I read
> Rosen.
>
> My view is that Rosen-complexity (to distinguish it from all the
> varieties) is general, and that it accordingly explains the appearance
> of mechanisms, including all "classical" matter and apparently classical
> processes. To do that it must be considered ontologically. Having thus
> explained the appearance of mechanisms, it does not follow that they
> have been eliminated from consideration. There will still be many
> questions and situations where a mechanistic answer is adequate and less
> confusing, although I fully agree that explaining life, consciousness,
> ecosystems, etc. are not such cases, exclusively, but contain many such
> cases within the circle of questions that may be asked of them.


I agree that recognition of Rosennean complexity in no way invalidates the
use of mechanistic models in many situations. If anything, complexity only
adds the opportunity to widen the paradigm of available models so that
mechanistic ones become a subset of a more complete set of models.


Regards,
Tim