Bob Myers said:
Why? How are you distinguishing "system properties" from
"part details."
For example, a 747 is a fairly complex system.
Not really. In comparison to real world complex dynamic
systems it's comparatively simple. The definition of the
word complex, as defined by complexity science, still
escapes you. A 747 is a very complicated system
on a linear scale from few to many.
Complex, the new meaning, is neither few or many.
It has too many variables for a Newtonian or classical
solution, but too few variables for a statistical or quantum
like solution. Complex behavior requires ...both...realms
of mathematics at the same time to fully describe the
behavior. That is why a behavior is called complex, since
both opposing realms of math is required. Simple now
becomes defined as that which can be solved with
just ...one or the other... realms of math.
Perhaps the simplest example of a complex system
would be a cloud. Where molecules are randomly
changing from liquid to vapor and back again.
A continuous sequence of step changes.
Even a tiny change in temp or pressure can have
cause the system to dramatically change forms.
One nut
taken from that same 747's gear assembly is clearly a
simple part. If I drop either the 747 or the nut, I will gain
the data necessary to describe how objects fall within a
gravitational field. The complexity of the 747 makes no
difference at all here; it would be relevant only if I wanted
to learn something about how the internal workings of
that system behave, as opposed to its gross externally-
visible properties. Moving up a few levels, I can also
observe/demonstrate the basic principles of aeronautics
through either watching that 747 in operation, watching
a Cessna 172 - although Cessna is undeniably a far simpler
system.
Just the opposite is true with real world systems. Again
this science is about how nature creates. Are the properties
of a forest or a society best displayed by the Congo or
your back yard? Is public opinion better judged by
large or small samples?
The basic difference between a man-made system and
natural systems which this science is about can be
seen this way. A man made system is one where the
goal or final product is envisioned ahead of time, and
the parts are constructed to fit the goal. Reverse
engineering nature.
A natural system allows the final product to emerge as
it will from the interaction of numerous autonomous agents.
A 747 and a society are opposites of each other.
One would think, from considering most examples that
come to mind, that the best way to "derive laws" would
be through the observation of the SIMPLEST system which
provides sufficient data to do so.
For man made systems sure, for natural systems not at all.
Or, in other words, one
should not bring in unnecessary complexity into an
experiment without having a good reason for doing so.
In natural systems, the ...complications..are added by
reducing in scale causing a proportional increase in
variables and interactions. You have to properly envision
what complexity now means. Imagine two variables
that are being pulled in opposite directions at all times.
And just so that neither direction wins, but a persistant tie.
The variable behavior becomes complex as it's future
behavior depends on the most delicate of future
changes. In short, the future of the variable cannot
be determined, in fact complexity is where there's
the least ability to determine the exact state of
an object or behavior.
Precisely; but you do not need the full system to demonstrate
or derive ALL of its properties, only those which might be
considered to be "emergent" at that level of complexity.
For anything less than this, you're better off studying simpler
systems, since they have less potential for generating
irrelevant problems which would complicate the study of
these more basic properties. For example, if I want again
wished to study the BASIC principles of flight, I may be
better off with a much simpler aircraft than a 747, for the
simple reason that it will be easier to operate, more reliable,
and therefore will more readily get me the necessary data.
I would need to turn to the 747, or something similar, though,
if I specifically wanted to study, say, the behavior of autoland
systems in multiengine jet airliners. The complexity of the
system being studied must be sufficient to provide the
relevant data, but certainly should be no more complex than
that. So we still have not justified a belief that we should
always seek the most complex system possible for study.
Certainly - and if you wanted to study those, you would
certainly need to look at a sufficiently complex market.
If, on the other hand, you merely wanted to study the very
basics of economic transactions, watching a child's lemonade
stand doing business may suffice.
Are you going to see panic behavior in a lemonade stand?
Or the effects of rumors? Emergent behavior is that which
is only a property of the whole. Behavior that is only seen
in a mob for instance, and nowhere else.
Yes - which again argues only for the need to study
complex systems in those cases where we are investigating
properties seen only in those systems, and not in simpler
cases.
Which "system properties" do this, and on what grounds do
you make the above assertion? Until you can identify
them, I am also not sure how you can get to the conclusion
that:
Tell me how much a market force weighs? How large is it?
Where can I find them? A market force only exists as a
part of the functioning whole. Not a single part of the entire
system will tell you what they are? As only the combined
interactions of all the parts produces them. We know
markets find a way of increasing efficiency, stability and
creativity. And with Adam Smith like invisible hands
these forces ...emerge...from the interaction of autonomous
agents and provide the long term guidance or direction
of the whole. It is the self tuning properties of nature that
created us and intelligence. And the physical universe.
First of all - "most important" in what context? On
what scale? As judged by whom?
Based on what is most responsible for the visible order
in the universe.
An example - the behavior of the force of gravity is
clearly a very important aspect of our universe, as
it controls the paths of moons, planets, stars, and even
entire galaxies; it literally shapes the universe. Yet the
"laws" which describe this behavior are actually rather
simple, and may to a very high degree of accuracy
be derived from observations of much simpler structures
than the universe as a whole. This, then, would seem to
be at least one example which contradicts the above
assertion. It may not be the case that ALL "important
behaviors" can be understood from the simpler cases,
but it is clear that at least one can, and I believe there are
other examples that will readily come to mind if you think
about it.
Lets compare two entirely different fields/things using
a system approach. Using the output of the whole
as the primary information instead of the part details.
Gravity and biological fitness.
One is a pervasive force for order in the physical
universe, and the other for the living world.
A gravity well vs a fitness landscape.
When systems become complex as described
above, standing persistantly poised at its
phase transition point, fitness peaks then
tend to clump together, and higher fitness
peaks have a larger basin of attraction.
A living system interacting randomly with its environment
is more likely to 'fall' into a region of higher fitness
than lower. Just like objects tend to fall together.
On the....input...side the two are as different as can
be...material vs living. On the output side they follow
essentially the ...same laws of organization.
Using the part details of each system as the first source
of knowledge leads to two entirely different sciences
each so unique such commonalites cannot be seen.
And this commonality is the primary guiding force
for the system, as you said with gravity. So it is
with biology. Just two vastly different scales of
complexity.
By looking first at the output side, the system properties
we see what is common between two systems.
By looking at the part details first we see what is
different. And very different, so different the two
fields can't really compare anything between them.
Physics and biology.
OK, so if this is true - what IS the single unified
description of the universe offered by this science?
If it is too complicated to outline here, then I would
submit that this strongly suggests that it is NOT a
fundamental behavior or "law," but instead is itself
derivable from simpler principles.
All order in the universe, physical or living, is the result
of a system that's persistantly poised at the phase
transition of its own opposite extremes in possibiliy.
Or, evolution of the physical and living worlds gets
it's impetus from complex behavior. Where the parts
cannot be precisely determined, and only the output
is predictable and repeatable. As with the relationship
between a forest and its components. The parts are
quite often acting randomly, so much so as to defy
prediction. While the output, the system properties
are stable and resilient...predictable and repeatable.
The laws and certainty all of us instinctively look for
only reside in the whole...the output of a large collection
of interacting autonomous agents. The larger the more
predictable...the more certain...the more informative
of the future.
In that case - how has "complexity science"
explained life?
What do you mean by the symbol "god?"
That which is as far above us, as we are above animals.
You have described your desired process, but I am
afraid this doesn't give us much to go on in terms of
why it should be preferred, or even what it actually means.
Why SHOULD "subjectivity replace objectivity"?
Again, it goes to the meaning of the word complexity.
The duality of light for instance. Light act as both a particle
and a wave, so it is a complex system. To objectively
measure it, it is reduced to ...either.. a particle or a wave.
It becomes simple when objectively observed.
But our existance and reality is based on the properties
of light when it's in a natural....complex...state.
It's the properties of the ...system...that define our reality.
Not the parts. Whenever you reduce or simplify to allow
objective measurements, the most important properties
the emergent properties, cannot be seen.
Only by observing the whole can the guiding self tuning
system properties responsible for the constant process
of evolution be seen. They cannot be measured, but
only known subjectively.
Unless you are using these terms differently in this context,
are you implying that there is NOT such a thing as a single,
shared, objective reality?
I'm stating there is no objective reality. Think of a ecosystem.
The components and the environment are coevolving.
Each constantly changing around each other.
Nothing stands still long enough to exactly quantify as
everything is constantly changing. It is only our
very arbitrary decisions on when and what to quantify
that allow any objective comparisons.
Objectivity measures the past and the simple.
Subjectivity measures the present and the complex.
And it is the complex that is the source of all order.
If that is the case - if the
universe fundamentally subjective rather than objective -
there is very little sense in discussing any of this in the
first place, since we can't possibly be assured of any
common ground on which to base descriptions which
would be useful to more than a single person.
Complexity science has found a way around that problem
as well. As this is a relativistic approach that includes
not excludes the observer.
In complexity science two people can easily make the same
subjective observations. As each system is compared not
to some independent yardstick, not to some other system.
But each system is compared to...itself.
It's very easy. The first thing to do when looking at a system
is to define it's opposite practical extremes in possibility space.
For instance, with a complex dynamic system such as the
stock market, thousands of autonomous agents interacting
in highly random and unpredictable ways. To find where a
variable becomes complex one merely observes the system
in operation and asks; "what are the practical, not theoretical
range of operation for that variable in the system being observed?"
Here's how it works. Systems display universal behavior when
the primary driving variables are all 'complex' at the same time.
Using the new defintion of complex as being midway between
opposite extremes so that the value has the least possible
predictive use. Which is when the variable or part displays
the least amount of certainty or ability to quantify.
With a stock chart you only have three variables, price, volume
and time frame. The simplest example possible.
The task is then to ...subjectively...determine the range of values
for each that defines the most complex state.
What is a complex price?
Observing market behavior, in general any price below a dollar
is a stock either entering or rising from oblivion.....one extreme
in possibility. And price over...say..ten dollars is considered a
stable company or a blue chip. The other extremes in possibility.
Another person may set the range of one extreme at below two
dollars and the blue chips at twenty. This subjective difference
doesn't matter much once one further constrains the analysis
with the other two remaining variables of volume and time.
What is a complex time frame?
Observing we find the day traders operate in spans of minutes and hours.
The opposite extreme, the longs, operate in spans of months and years.
So a complex time frame is that which is neither short or long.
Days and weeks is a complex time frame for the stock pattern to play out.
Five or ten day chart behavior defines the transition point between
opposite practical extremes in possibility.
And so on. When all the primary system variables are complex
the systems all behave in the same way. Or, when component
uncertainty is at it's highest, system output is at its simplest
and most predictable.
And btw, as with all evolving systems, this critical behavior brought
on by complexity is where predictability and volatility converge
to simulaneousl maximums. As in a thunderstorm, highly dynamic
yet, once we've seen enough cycle, very predictable.