Re: Holistic World and Complexity

Don Mikulecky (mikuleck@HSC.VCU.EDU)
Wed, 26 Aug 1998 09:50:48 -0400


Don Mikulecky replies:

Ricardo Ribeiro Gudwin wrote:

> John J. Kineman wrote:
>
> > I understand that we can SAY (i.e., it is a philosophical option) that all
> > "real" systems are ULTIMATELY complex. This is a particular worldview.
> > Saying this ignores the concept of "levels" which I tried unsuccessfully to
> > introduce earlier. I take from the lack of comment on that topic, that it
> > is difficult to reconcile with the discussion, because they are
> > incommensurate views (which may both be valid). So it is this I want to dig
> > into a bit if you will indulge me.
> >
> > What are "levels" as I am employing the concept? I suggested that it is
> > analogous or related to "scale" -- that these are similar concepts. Are
> > levels "real?" No, in Rosen's (theoretically ultimate) sense, but yes for
> > practical purposes (which I will try to define).
> >
>
> > Our reference to something "real" in a practical sense, by this reasoning
> > MUST involve some kind of model, and therefore MUST involve the concept of
> > level or scale, which is inherent in our definition of "things" or
> > "entities."
>
> I believe that John touched two very important points, that are:
>
> a) LEVELS OF RESOLUTION - we can model reality through different
> levels of resolution - (John talked about scales ... I prefer to say
> resolution,
> it is exactly the same)

we include this when we say..."there are distinct ways of interacting with a
system
and distinct(not derivable from each other) formal systems we use in the
modeling
relation to try to capture these results

>
>
> b) KNOWLEDGE OF OBJECTS - in order to understand our surrounding
> environment, we frequently model it as a set of objects (John talked about
> things or entities - they are the same thing ... I am just using the
> terminology
> from Computational Semiotics)

and we extend "Object" to include mappings and relations

>
>
> The first thing we have to understand about reality is that we can only be
> aware of it through our sensors (remember the discussion about Umwelt !).
> So, the only kind of information we can get from environment (that is a
> synonymous to reality, if you wish) is something like ... colors,
temperatures,

yes this is what we mean when we say that, in the modeling relation, we replace
the sensory input from the "real" world (measurement, observation, experience,
etc.
with a "percept")

>
>
> contact (that can be pleasuring or paining), sounds, smells, etc. Based
> on this information, our mind INVENTS those "entities" we call objects.

ok we call them percepts...straight out of Rosen so far.....(sorry)

> For example, you are not able to sense an apple in front of you. You are
> able to sense its colour, its smell, its taste, the type of contact it has to
> your
> fingers, etc. Based on this information, you infer, by abduction, that these
> senses are due to SOMETHING, we call an object. This KNOWLEDGE
> OF OBJECTS is the essence for understanding the idea of mechanism/
> non-mechanism that is being discussed here, I believe.
> Once we created such KNOWLEDGE OF OBJECT in our mind, after
> sensing the environment, we will try to understand how it works. In order
> to do so, we use the reductionistic method of trying to guess OTHER
> OBJECTS, in a LOWER LEVEL OF RESOLUTION, that could
> explain the behavior of the original object, in a higher level of resolution.
> If we succeed in this, then we may say that our original object is
> A MECHANISM. But what happens if we try to decompose the
> behavior of this original object and we can't acquaint for the creation
> of such model ? If I open the Pandora box, and the only thing I can
> see is a whole mess ? (Just like the complex TV set given by Don).
> Then we have a COMPLEX SYSTEM, a NON-MECHANISTIC
> system.
> Other thing that maybe is important is the definition of these LEVELS
> OF RESOLUTION in terms of FUNCTIONS. I am not talking about
> mathematical functions, but to its FUNCTIONAL BEHAVIOR.

this is key in our approach!

> For
> example, let's talk about a FACTORY. This factory has a FUNCTION
> to produce goods. This is the higher level of resolution. We may decompose
> this object FACTORY, in lower resolution objects DEPARTMENTS. So,
> the factory, from the extreme functional level of resolution, is a mechanistic
> system. Then, we try to decompose and understand one of these departments,
> e.g. QUALITY CONTROL DEPARTMENT. The function of QUALITY
> CONTROL is to assure the quality of goods produced into the factory. But
> what happens now ? I can't reduce QUALITY CONTROL anymore ! There
> are a bunch of employees, that come and go, some of then working inside the
> factory, others outside the factory ... there is no such thing as QUALITY
> CONTROL DEPARTMENT PARTS, because their supposed parts are constantly
> changing or changing behavior, and we can not reduce it to simpler parts
> (in fact, WE MAY, if we consider a temporal picture of the department, but not
> the
> department as a whole phenomenum in history). So, this is the place where
> COMPLEXITY comes to help us ! If we want to fully understand the QUALITY
> CONTROL DEPARTMENT, we have to use non-mechanistic methods, in
> order to build a model for the department. There are other examples, I guess
> ... a footbal team, a nation in the world, the population into a bar, the
> traffic, etc.
> And the worst part of the story is that in some cases, we may be able to
> create
> an imperfect mechanistic model for some object, but this model does not
> acquaint for the
> FULL BEHAVIOR of the object. And this is an indication that MAYBE this
> object does not support a mechanistic model, and needs a complex explanation.
> Now, some WILD GUESSES (subject to your criticism, of course !)
> 1) To be called COMPLEX, an OBJECT must not have a FIXED STRUCTURE.
> In other words, its structure (in terms of sub-objects) must be VARIABLE.
> 2) This VARIABLE STRUCTURE is due to a constantly creation and
> destruction of objects composing the original object (remembering that
> objects are only creation of our mind .. we may see things that could
> resemble objects - not necessarily they are !)
> 3) A MECHANISTIC system always have a FIXED STRUCTURE. Its
> objects may vary its atributes, but not the number and type of sub-objects
> composing them.

this is part of what we mean when we say that the mechanistic system has a
largest
model while the complex system does not....also what is meant by analytic models
are the same as synthetic models in mechanisms, but not in complex systems

> 4) Organisms can be FUNCTIONALLY DECOMPOSED only to a certain
> level. After this level, what comes is this COMPLEX MESS, where we are
> not able to understand why some structures as CELLS, can arise from the
> simple interaction of ORGANIC MOLECULES. (Some of you may
> correctly note here that there are also those organelles inside a cell, that
> would be understood as a PART of it, but I believe that the example
> continues valid)

yes...this is what we mean by the absence of a 1:1 mapping between functional
components and physical parts

>
>
> The question now is ... these examples fully attach the problem of
> complexity, OR NOT ? If not, what is missing here ? I am not convinced
> that they fully attach complexity, in the sense we are talking here, but would
> like to hear a little from you !
>
> --
> //\\\
> (o o)
> +-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-oOO--(_)--OOo-=-=-+
> \ Prof. Ricardo Ribeiro Gudwin /
> / Intelligent Systems Development Group \
> \ DCA - FEEC - UNICAMP | INTERNET /
> / Caixa Postal 6101 | gudwin@dca.fee.unicamp.br \
> \ 13081-970 Campinas, SP | gudwin@fee.unicamp.br /
> / BRAZIL | gudwin@correionet.com.br \
> +-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-+
> \ URL: http://www.dca.fee.unicamp.br/~gudwin/ /
> / Telephones: +55 (19) 788-3819 DCA/Unicamp (University) \
> \ +55 (19) 254-0184 Residencia (Home) /
> / FAX: +55 (19) 289-1395 \
> +-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-+

I see little that differs from Rosen's dichotomous classification.....it is
not
all there...but the essence is
Don Mikulecky