Re: [COMPLEX-M] 2nd Law and Ecosystems

["John Kineman" <***>]

Stan,

Thanks for the detailed response. You were kinder than I expected. Let
me make a few responses, then perhaps let it alone, as I think we are
each looking at the same elephant, but from slightly different
interpretations of what we see. I'll skip to the latter comments, which
seem to reveal this, but let me emphasize that I'm only trying to
clarify the thoughts I'm working on about this, not to argue than anyone
else is wrong about it or would necessarily benefit from my view. Each
has its utility. I mostly value the dialogue.

Stanley N. Salthe wrote:

>...
>
>
>John K said:
>...
>
>
>>I think this is why study of anticipatory systems is so important ...
>>
>>
>     SS: Here I will make what may seem (and it is) a radical statement:
>If we have a model of a system in the style which science privileges today,
>it would be a fully explict, mechanistic model, embodied in an equation or
>set of them.  My assertion is that once you make a model of that kind for,
>say, a biological system, only technical difficulties would stand in the
>way of applying it to abiotic systems -- the main one being the relative
>vagueness of abiotic systems, partly due to in many cases to the different
>scales of biological and physical phenomena.
>
Yes, I also agree with this radicalism. I think it is also in Rosen's
philosophy, although maybe not explicitly stated. Any description, being
formal, must simplify. Hence the only way to represent complexity in a
model is to have an infinite (or sufficient for some purpose) set of
them. I'm using this idea to design some new kinds of geographical
information systems.

>>Many physicists (I'll assert
>>apologetically) seem under the illusion that all the assumptions behind
>>physics are the assumptions of science, but that is not so.
>>
>>
>     SS: That is the source of the kind of models now used.
>
agreed, and the other fields have employed them; hence the physicalism

>     SS: Another unfortunate fact has been the "physics envy" of biologists
>in particular.  But no new style of modeling has been generated in biology
>or sociology.
>
>
I think Rosen's modeling is new. It is also quite controversial and not
common among biologists, who I think are afraid to let go of the
security of physical laws. But also, as we seem to agree (below) Rosen's
model will not replace those laws, which are essential in analyzing the
physical aspect of living systems.

>>The physical assumptions are wonderful and productive, but they
>>are a way of looking built on a limiting set of assumptions designed to
>>extract predictability, they are not the system itself.
>>
>>
>     SS: Yes. The physical is just a part of the intensional complexity of
>kinds of systems higher in the above specification hierarchy.
>
If by "intensional" you mean our way of modeling it, then we agree.
...

>     I do subscribe to a "unity of the sciences", BUT, not in its original,
>reductionist form.  I claim that the above specification hierarchy:
>{physics {chemistry {biology {sociology {psychology}}}}} would make a good
>basis for unification.
>
Hmm. Such a hierarchy seems physically reductionist to me, but perhaps
you do not mean it to be taken ontologically? In that case, I
understand. As an ontology, however, it has physics being fundamental to
everything else, with the implied reduction there-to -- so I'm guessing
you don't mean that. A different ontology could just as well have us
writing the hierarchy with psychology containing everything else. Both
are plausible.

>  Note that in this view, each new integrative level
>introduces controls, regulation, and interpretation of processes found in
>more general levels.  For example, macroscopic biology interprets diffusion
>by way of elaborate circulatory systems.
>
>
OK, and could we also say that quantum  physics interprets perception as
uncertainty, or that macroscopic physics interprets perception as
dynamic certainty? In other words, do you really envision causes
originating at the top of this hierarchy and affecting the lower
physical levels fundamentally, or is the hierarchy constructed from the
bottom up?

>     SS: I have suggested, for example, that the likes of tornadoes are
>actually the organisms of physics, and that disciplines like hydrodynamics
>would be central to a "physical biology".
>
OK, but again I  wouldn't consider this an example of biology informing
physics, but the reverse, which is traditional. When new laws or
theoretical reals are introduced in biology or sociology that then
constrain the physics (laws, not just the behavior of otherwise physical
systems), that would be an example of what Rosen is referring to .

>     SS: But -- there you go making a physics type of approach (as did Rosen)!
>
>
A very fair comment, and related to the one earlier about models being
simple not complex. I also go so far as to say that Rosen's view is
itself a reduction, as all thought is a reduction (to something, for
example language). But that it can be a different reduction than
traditional physical reduction, which defines certain kinds of
spatio-temporal reals (with extrinsic laws) and makes that view primary,
excluding elements that are not definable that way (like experience of
goals and intentions). So as far as it being physics in other cloths, I
don't think so. It is sceince, which physics is extremely good at, so
there are similarities of method, but I see the difference centering on
the formalization of goals vs. their traditional exclusion in physics
(as elements of explanatory theory).

>>I believe this must ultimately be interpreted to apply to goals,
>>purposes, intentions, and other experiences previously thought
>>(incorrectly, I believe) to be anathema to science.
>>
>>
>     SS: From this I fear you would bee just physicalizing biology.
>Somnething that you appear to think I was doing!
>
It seems to me that if a socio-ecological theory discusses the reality
of certain goals and their effect on creating structures (like Bill''s
civic design), that is introducing an element into theory that the
physical view is rather adamant about excluding. But if physics were to
include explicit representation (symbolization) of goals as causally
involved in explanations, then there is no difference. But I think that
would no longer be called physics. Ultimately, there SHOULD be no
difference, but we're not there yet.

> ...
>
>>All of the physical theory can be shown to be a
>>special case of this view (where formal and natural systems commute);
>>
>>
>     SS: In my view, not a special, but, rather, the general case, with
>increasingly more special cases as we go into higher integrative levels.
>
>
Exactly - that is the difference in our views and Rosen's, which says
the physical view can be seen as special. Perhaps there is no
resolution, just different theory structures. But Rosen seems to have a
good point that since mechanisms are a degenerative case of the modeling
relation, and there is no representation of perception, intentions or
goals (functions in their own right) in the physical view; that seems
conclusive in Rosen's favor.

If one were to say that the physical laws of nature are functions in the
Rosenean sense, then I can see calling things like the 2nd law a "goal"
of the system. But there is a very big difference between a goal that
MUST necessarily commute with the system (as natural physical laws are
defined), and one that does not but is still important (as in the
quantum case where I suppose the probability function substitutes for
some kind of goal). Its like computers following rules, vs. teanagers
following rules. It is clear in evolution that the tradition is to
violently oppose any teleological  language. Physicists are bolder with
the language, perhaps because they know their systems are better
behaved, but most physicists try to get rid of observers. I think a
Rosenean perspective would model the observer (in the nature of percepts).

>     SS: How, exacty does Rosen's approach differ from that which a
>physicist would take (say, to a tornado)?
>
I'll try: There would be no difference in viewing a "tornado," I
believe, because our model of it can be simple without much  theoretical
problem, even though that approach (dynamics) becomes impredicative as a
practical matter (and thus unparsimonious). But then the thermodynamic
view takes over and performs the same function as Rosen's theory does
for biology - it throws away the ability to be precisely predictive (of
particle motions) and describes general characteristics of the system.
But again, these general characterists fully commute with what is
observable, so it is still mechanistic (even if non-linear in the
dynamics). Only when you frame the problem as a modeling relation can
you see the possibility of something non-mechanistic. Then, if we
discuss the relationship between tornados and society, there would be a
big difference in approaches. The mechanistic view would have to look at
matter and energy and model the physical constraints on society, but not
what goes on in society mentally. Human psychological complexities would
enter the situation to cause many different interactions with tornadoes
that could live entirely within those constraints. An easier example for
me is an automobile. As a mechanism it is a commuting modeling relation
and there is no difference from the physical views (natural law is in
Rosen's Formal System box). But if we consider its ontology, which
involves artists, engineers, assembly plants, designs, marketing, etc.
there is an impossibility of producing a single model that commutes with
that system. The tornado too has an ontological complexity, but i
believe it is synonymous with the the ontology of all of material
existence, which remains consistent because it commutes with its
perceptual framework (space-time perception). So a dissipative system
that only dissipates, and doesn't utilize that energy to realize other
goals, is still a physical system. There would need to be an intrinsic
model of different goals for it to become complex in the Rosen sense.

>(Incdentally, the specification
>hierarchy is one way to exemplify Rosen's view of complexity -- that a
>system is complex if it must be viewed in more than one way.)
>
Yes, but this also would mean, more than one way of defining the hierarchy.

>     SS: My fear is that, being fully explicit, it will not escape physical
>style modeling.  (This may actually be more apropriate for higher
>integrative levels, which are less vague than physical phenomena.)
>
Yes, agreed. No model can capture all aspects of a complex systems, and
so is essentially simple and reductionistic. However, the point is to
reduce differently than in physics - to a fundamental relationship that
thus preserves and focuses on the relational aspects of nature rather
than the particulate aspects that appear when the modeling relationship
commutes (a special case). Then theories can be created about modeling
relationships that fundamentally do not commute.

>>.. In other words, it is possible that
>>something beyond physics can inform physics, as Rosen suggested.
>>
>>
>     SS: In connection with systems like tornadoes and hurricaes, I agree.
>
Well, the point would be that some new principle learned from a living
system would tell us something about physics. Rosen gave some examples,
eg., in protein folding; but they all relate to the imposition of some
organizing princple that is not emergent from the physics. I think the
main thing it will tell us is about the nature of space, time, and
perception; i.e., at a level that is general to all physical phenomena.
But I don't think it will really say much about the hurricane (IMO). For
that, I'd stick with what we've got.

>>I think the Rosen view may be most involved with studying natural design
>>principles - design being a complementarity between form and function. If
>>this is ultimately reducible to the physical view, that is not an
>>achievement that is nearby; whereas the need for functional perspectives is
>>very great and largely unmet.
>>
>>
>     SS: Tornadoes have design and funtions.
>
I would say differently - that they do not have any built-in
representation of design or function that doesn't commute with its
dynamics. All such are attributions by us, the scientists explaining
them or people affected by them. However, we do have built in
capabilities to have and vary representations of our own identity (I'm
not referring to genetics here - which is involved but not the whole)
and our relationship with our environment. These squishy aspects of
perceptual experience are what I'm associating with function (as I
believe Rosen did), not just someone else's interpretation of function
(which also occurs, but extrinsically).

> SS: I interpret ecologies as material systems.

Isn't that the definition of physicalism? The definition of ecology
involves both organisms and environments, which one is free to view
physically, but also free to view functionally.

> After all there had to be ecologies (energy flow patterns) in the
> prebiotic world that were later
>
>invaded by living systems.
>
There's no ecology without life, but aside from that, there may also
have had to be psychologies in the pre-physical world in order to get
mechanisms. The nature of a percept could be general to both.

Cheers,

John Kineman