[Date Prev][Date Next][Thread Prev][Thread Next]   [Date Index] [Thread Index] [Author Index

Simple and Complex systems



Judith,
 
I think I finally clarified it, at least for myself. I always felt it was a linguistic confustion. Here it goes.
 
Think of a specific honda civic parked around the corner. You're standing next to this material object and pondering whether or not it is complex (in Rosen's sense).
 
It belongs to a class of objects denoted by the symbol "car". The reason it belongs to that class of objects is that it is a *realization* of a specific simple/simulable/predicative organization. The class "car" is *defined* by being a realization of a specific simple organization.
 
Now there is also a Rosennean alien (evolved from a bird on another planet) standing next to this material object. He too is pondering whether this object is complex. He does not know of the class "car" or its definition. Thus the only thing he can do is to build models of this object. He cannot even guess its function (because he never had any use for a wheeled form of transportation) thus he does not know what is relevant and what is not. As he builds one of his models he focuses too much on the viscosity of the oil in the ball bearings. He concludes the system is complex.
 
So all these arguments seem to be based on the linguistic confusion between "car" as marker for a class of objects with a specific simple organization and car as a marker for a specific natural system that is an instance of that class.
 
Now the interesting question is how the word "dog" is seen under this analysis. Dog is not a realization of any known organization. Robert Rosen argued that the class dog is defined by an impredicative organization. Thus he would argue that a dog is complex in the Rosennean sens.
 
- Steve
 
 
 
 
 
 
 


Steve Johnson <***> wrote:
JR: So, having reiterated that, let's look at the issue of the system we call "car".... In order for a car to be a car, what we have to look at is the organization that makes it a car. Not the organization of the atoms or the molecules or the next larger category of components.... the organization of an atom is not what makes a car a car.
 
SJ: Judith, I struggled with this issue for a while myself when I wrote my post. As I said in it: "a car is a simple system at least as far as its car-ness (or form) is concerned." I did not put it as clear as you did but I think I had a similar idea.
 
However, one thing keeps bothering me and I don't quite know how to express it. I will try and maybe someone can pick it up. To say that X is a simple system *presupposes* that we definitively know its organization or "bauplan" (and we know it is simple/simulable/predicative etc). This is true in case of car because the word car (as you used it) *denotes* a particular mode of organization with a known form and function. So with car we do not have to build models to prove it is simple because it is simple by definition. As we analyze the minute details of viscosity of oil in the ball bearings we indeed slowly leave the "car" concept and start dealing with other complex systems outside of its car-hood.  The word "dog"  (when it is used to denote a specific dog) on the other hand does not refer to a known mode of organization and thus we do not have the luxury to declare that viscosity of blood in its veins is not part of its dog-hood. All we can do is build models of it.
 
So when you use the word "car"  it can have two meanings: 1) a  mode of organization and b) an imperative to think of a specific car. If you take the first meaning then indeed car is a simple system. If you take the second meaning (as I suspect John M. did) and you are thinking of a specific car with all the molecules of gasoline, drag coefficient, viscosity of oil etc. then you are dealing with a natural system (that was pointed to by the imperative: think of a car!).
 
To take the example even further: suppose an alien scientist who evolved from a bird like creature (and thus never had any use for wheeled form of transportation) is analyzing a specific honda civic. Unlike us it does not know which aspects are relevant to its car-hood and which are not. From its perspecitive it would be just another natural system much like a tree or a dog are two us. And thus the Rosennean alien not being able to separate the irrelevant aspects might well conclude that this object (that humans would call "car") is complex.
 
- Steve
 


Judith Rosen <***> wrote:
Hi John M.
 
One thing to always keep in mind is that "system-hood" is, itself, a relational property. It has to do with the relation of the observer to "the system" (meaning"that which is being observed"). This is one of the reasons why science can never be a truly "objective" pursuit-- I would argue that there is no such thing as objectivity in science. By its very nature, "science" is a human mental creation and exercise. However, science is what humanity has developed as a way to aid in learning about ourselves and our world/universe (because we do see both from an innately egocentric perspective). So all of these descriptions are being made from within the context of SCIENCE.
 
My father's work was not philosophy, much as he was accused of that, mostly be experimentalists who considered "theory" to be synonymous with "speculation". So you say that these boundaries of system are artificial? Yeah, they are. All human perception is a mental artifact, in some sense. And your point is??? "Reduction" is not a dirty word, unless it's the only word allowed or it's misapplied to situations where it's inappropriate. If the former is true, it guarantees the latter, doesn't it.
 
What "RR" was saying was that science can do far better than it has been doing, in how it approaches learning about various aspects of our universe.
 
So, having reiterated that, let's look at the issue of the system we call "car".... In order for a car to be a car, what we have to look at is the organization that makes it a car. Not the organization of the atoms or the molecules or the next larger category of components.... the organization of an atom is not what makes a car a car. Similarly, we needn't bother going into ever larger systems, looking at the organization of traffic or of rush hour traffic jams, etc.... that's not the "car" system, either. The main POINT: it is a significant fact that the way some system is organized contributes a causal impact on its behaviors/capabilities. In a car, that organization is simple: pretty much the sum of its parts.
 
 

----- Original Message -----
From: John M
To: ***
Sent: Monday, January 17, 2005 5:56 PM
Subject: Re: [ROSEN] Inequivalence of models

Hi, Judith,
Quote from your post:
"that simple systems are computable and have a "largest complete model" into which all others will reduce."
I have (language?) problems here. I don't believe in "simple systems" only in  'simple system models'. We use the 'car' as a good vehicle for such discussions. A car is a car is a car - wrong.
A car is only a car if we cut the model we assign to this distinction at the boundaries we consider for "a car". Otherwise the system representing that model as well, is an unlimitedly interconnected feature, from its submicrosco[pic, economical, ewsthetic, mythical,
(you name it) associations (in connected networks) all the way to the energy we receive from the sun and beyond.
A "car" is a complexity. Our limited model is simple.
As simple as that.<G>
So a car has NO largest model. Only the limited model has some boundaries within which we feel happy. The "simple car".
In our well established reductionistic thinking.
Is that what we want to perpetuate?
You emphasize the computable simple models. "...within it".
*
JR:
"Complex systems, on the other hand, have an infinite set of models, without ever exhausting all the information possible. Because there is no "largest complete model"... "
Would you like to set up a singularity with no connections outside its boundaries, what you can deem "a simple system"? A nirvana?
becuase if not, EVERYTHING is a complexity and you can stop the connectivities only by exercising reductionism: selecting the extent of your observation. When you say "single model" that means a reductionistic limited (cut) model of something that does not 'end' at those boundaries.
I for one do not want to study Rosenism restricted to reductionist models. Not even 'complex (closed) models' as they say. This is why I like better to say wholeness if I can. and I leave reductionist inequivalency to the engineers.
I think you simplified the simlicity while we realized the complexity of complexity. (Sorry, if I see a pun...)
 
Then you wrote:
and each model will then "reduce to" (fit into) that sum/largest model.
Sorry again, my non-IndoEuropean Hungarian linguistic stomach does not digest such meaning of reduction into increasing. I feel a 'reduction' makes something smaller, not into wider and more comprehensive. Not even as a word-flower. But my Inglis is poor.
Fit into is OK, refers to a larger compartment from which it was reduced. Not vice versa. I mean: the nonexistent largest model.
 
 
John Mikes
----- Original Message -----
To: ***
Sent: Monday, January 17, 2005 4:15 PM
Subject: Re: Inequivalence of models

Hi Steve, Hi Everybody,
 
Sorry for being "incommunicado" for a few days, I've been working on BioTheory. It's mostly up and running, although five papers are still forthcoming, and my own papers are among them. What can I say? By the time people get done reading all the great stuff that's there, my own will be ready to plug in. I haven't had much of a chance to take off the publishing and editing hats, much less put on the writer's hat, for a while. But they're almost done.
 
The question about "inequivalent models" isn't as complicated as you guys are making it... It has to do with the fact that simple systems are computable and have a "largest complete model" into which all others will reduce. If you go about it from the other direction; the sum of all models we can make of the system will include every individual model within it. That's what "equivalent" means. Complex systems, on the other hand, have an infinite set of models, without ever exhausting all the information possible. Because there is no "largest complete model"...
 
Do you see? Any single model is a finite thing, and there is no way to get to infinity by accretion (adding finite numbers together). Thus, inequivalence.
 
SNIP
and each model will then "reduce to" (fit into) that sum/largest model.
 
Does that make more sense?
 
Judith
 
 
Website address: http://www.rosen-enterprises.com/
My favorite discussion list (Independent-- Not part of Rosen Enterprises): ***


Do you Yahoo!?
Yahoo! Mail - now with 250MB free storage. Learn more.


Do you Yahoo!?
Yahoo! Mail - Find what you need with new enhanced search. Learn more.