Re: Re[2]: Philosophy underlying LO? LO334

Jim Michmerhuizen (jamzen@world.std.com)
Sat, 4 Mar 1995 23:33:40 +0001 (EST)

If I made my living at this I might be able to do justice to Fred's post,
which is exciting. I can only cover a couple of points. I've done this
by annotating Fred's text. I've done this without niceties of phrasing,
and my comments consequently may sound a little abrupt. For that I
apologize in advance.

Regards
jamzen@world.std.com
---------------------------------------------------+---------------------
- - - - There are far fewer things in heaven and earth, Horatio, - - - -
- - - - - than are dreamt of in your philosophy... - - - - -

On Fri, 3 Mar 1995, Fred Reed wrote in LO298:

> Replying to LO298 --
>
> jamzen@world.std.com wrote:
> (stuff deleted)
> >Yes. I believe that the shift of perspective afforded to us by the new
> >understanding of chaos and nonlinear systems is more encompassing than
> >either quantum theory or relativity. The present revolution dwarfs
> >anything since Newton.
>
> >And frankly I'm welcoming it. There were a lot of things I never liked
> >about "science", back when "science" meant "finding Newton-type laws for
> >whatever you happen to be looking at" and "Newton-type laws" meant "slap
> >numbers on everything and then find at least a couple of plausible
> >closed-form algebraic expressions with time as an independent variable
> >that look like they might have some predictive power". Newton did that
> >for the planets, and from then until now everybody's been trying to do
> >that for economics and sociology and psychology and biology and a lot of
> >other *logy's. And then the "philosophers of science" get PhD's for
> >wondering about the reversibility of time! The closed solution is the
> >real conceptual villain in here, and it's the closed solution -- the
> >analytical algebraic expression -- that is the _FIRST_ casualty of
> >nonlinear systems study. Chaotic systems _DON'T_ have closed-form
> >solutions with time as an independent variable; LaPlace goes down with a
> >stake in his heart.
>
> >And the glorious paradox in all of this is that it's accomplished without
> >giving up ordinary causality. Even chaotic systems don't have "free
> >choice"; each momentary state is continously connected to the one before
> >and the one after. (For centuries after Newton, it was commonly argued
> >that unpredictability was inconsistent with causality. Tee-hee.)
> (more stuff deleted)
>
> in response to John Conover discussing Rudy Rucker's book:
> (stuff deleted)
> > (which is a very good book, BTW,) and presents some very formidable
> > arguments in support of your premiss. Such systems that exhibit this
> > phenomena are usually called fractal or non-linear dynamic systems. It
> > does indeed appear that social institutions are such a system. As an
> (ditto)
>
> IMHO, chaos/fractal/nonlinear dynamics are *not* the real revolution.
> They are only qualitative changes in an unstated theory that the world can
> be accurately described in formal/mathematical terms of any level of
> complexity. For example, the heart of chaos theory is the "discovery"
> that even simple systems (let alone complex ones) can be extremely
> sensitive to initial conditions.

This is not the heart of chaos theory.

> Left unstated is the assumption that the
> even chaotic systems can be described in formal/ logical/mathematical
> terms.

This is _not_ an "unstated assumption". _This_ is the central discovery,
the heart.

> Chaos only says that our job of prediction is only much (near
> impossibly) harder. This assumption is fine for mathematicians (as long
> as they don't actually claim the math applies to a real system) and
> computer scientists (whose "ground truth" is the logical computation going
> on inside their box). But to say that chaos/fractals/etc. are the key to
> understanding the dynamics of *real*, particularly, "living" systems such
> as organizations and other social systems is to continue the long
> tradition of assuming that such systems follow the same system of logic
> that mathematics addresses.

No doubt some people are saying that stuff about "key". But I'm not, and
neither are a lot of other people. With regard to those who _are_, you
and I are in the same army: they're wrong, and you and I are right.

With regard to nonlinear systems theory in the physical sciences, it's
already proved itself. In very many instances in the life sciences, it's
also proved itself, regardless of whether you or I know or care about that.
In the human areas, my current convictions may be fairly close to yours.
My position is that fractal imagery and concepts from nonlinear systems
are a source of powerful new _paradigms_ and _models_ of my human
experience of my fellow humans and of the structures we share. I'm not
sure that this is science, or even knowledge, but I'm certain that with
these models I can bring far more experience into my thinking than I
could under the old "Newtonian" regime.

Every really _big_ scientific revolution is always followed by a strong
current of "popularization" as the culture _around_ the science works at
absorbing whatever the new concepts and perspectives are. But this work
of assimilation is not itself science; by and large, the people who
take part in it are not doing scientific research. James Gleich, whose
book did so much to spread awareness, is not himself a researcher in any
of the fields he describes so well. Neither am I, and (I assume) neither
are you. We are part of the rest of humanity, finding out whether these
new concepts can become useful paradigms in areas other than those where
they have a specifically scientific application.

> On other words, I think the
> chaos/fractal/etc. *revolutionaries* are confusing difficulty in
> predicatablity with *inherent unpredictability*.

That may be true. But the researchers aren't. The unpredictability of
chaotic physical systems is perfectly real and "inherent". It is not
merely an empirical "difficulty".

> Living systems such as
> people and their social systems are inherently unpredicatable because the
> *create* things in the real physical world. Any formal description of a
> physical system can only describe the "important" aspects of reality *that
> have already been discovered*, which are very small compared to all the
> possible descriptions of that same reality. But living systems are able
> to use the implications of *physical*, not just logical, features and
> relationships to produce inherently unpredictable (via any formal means)
> future realities. It is the impications of this absolute unpredictability
> and role of creativity in science that will be the *real* revolution.
> George Kampis, in his book "Self-modifying Systems in Biology and
> Cognitive Science" (Pergammon 1991), presents this argument in
> excruciating detail. It has been some years since I have read it, but I
> seem to remember an example of physical implication that goes like this:
> If someone were to ask the information capacity of a particular
> computer tape, they might look at its length, its magnetic properties, the
> bit per inch capability based on these magnetic properties, and any other
> number of "formal" properties thought to have anything to do with its
> information capacity. Such analysis would allow me to "predict" its
> capacity *in the frame of the formal system I have chosen to represent
> it*.
> Now lets say this tape is lying around the computer room, not being
> used. I might agree with my fellow programmers that if I tie a knot in
> the end of the tape when I leave the room, it means I intend to return and
> not to shut down the system. I am now conveying information with the tape
> that was not predicted in my formal analysis because my formal
> representation (based on already discovered ways of using the tape for
> carrying information) did not account for all possible physical
> implications, in this case, the fact that this particular tape is flexible
> enough to tie into knots. This inability to predict is a different *kind*
> of unpredictability than the sensitivity of initial conditions (or brute
> complexity of a formal system) addressed by chaos/fractals/non-linearity.

Well of course it is.

> The "-ologies" that jamzen@world.std.com referred to (e.g., sociology,
> psychology, biology) will undergo the *real* revolution when the own up to
> the inherent limitations of formalization *of any kind* (including these
> "new sciences of complexity"), the loss of predictability, and ultimately
> (especially so for organizations and management) the loss of the ability
> to *control* based on this predictability when dealing with people and
> living things.

Well, we got through this one, after all, shoulder to shoulder. Sort of.
There is however a fractal dimension to our agreement that could bring us
to a duel sometime in the future.

For one thing, I believe that [a] the loss of predictability and control is
already here, on the grounds I gave above; [b] that formalization is
independently useful and achievable nonetheless.

One last footnote: in my original post, the absence of "closed-form"
solutions for things corresponds quite closely, I think, to the loss of
predictability that you refer to. In any case, it's something we can both
exult about, and (to bring the whole thing full circle at last) is clearly
one of the central facts for organizations and management, as you pointed
out.