Dear List,
I agree with Stan Salthe that Tainter's "kinds of complexity" are not
kinds at all but simply different circumstances in which complexity
appears.
From a anthropological point of view, it should be clear that no
scholar wisely references Wikipedia unless it is to study the
anthropological and sociological implications of its unreliable
nature and its risk to the public. It is hardly surprising that one
would find conceptual diversity there.
How does an anthropologist quantify complexity? What are the measures
that illustrate Tainter's claim that complexity has increased in
societies? It is not clear to me that these claims are true. The
numbers of individuals in societies has increased - and we certainly
appear to prefer to believe that our society is more complex than
earlier societies - but there seems to be little basis for this
intuition. These claims need to be founded upon some means of
quantification (per algorithmic complexity is, for example).
Does an individual in a hunter gatherer society, in fact, live in a
more complex society than an individual in today's society? In my own
proximity, for example, I doubt that my social relationships are
fewer than that of a hunter gatherer in a relatively sized community
of the hunter gatherer period - my family relationships are likely to
be simpler since I am disconnected from extended family - it is true
that all my relationships have a greater geographic diversity and the
medium by which I communicate has a different nature often, but this
does not seem to be enough to increase the complexity of my
individual experience.
The number of relationships that any individual can possibly maintain
is surely self limiting and this would constrain the complexity that
any individual - any single node in the complexity - can manifest. If
the nodes are bound in this way then complexity is also also bound
despite scale.
It seems likely that the complexity in societies has a natural
threshold. While the overall number of unique arrangements may
increase, the actual complexity never breaches a self-limiting
threshold. If I were to apply algorithmic measures of complexity, I
would say there is a limit to the number of steps that any given
individual can manage.
Simply enumerating elements tells us nothing about complexity.
Diversity does not equal complexity, it may be the product of
complexity but because diversity is increasing does not mean that
complexity is increasing.
For example, in algorithmic terms - if, in the example given, the
organization in which the artifacts were shipped to Africa actually
required more steps to assemble a weapon than other more orderly
organizations, then the system was indeed more complex, not merely
"complicated."
If the behavior of a society for an individual becomes simpler
because of arising diversity then the complexity is, in fact,
reduced, not increased, for that individual. Overall complexity may
remain the same.
I feel a clear definition of complexity is missing from Tainter's
discussion and I see distinct concepts being confused. I find myself,
for example, wanting a clear specification of complexity versus scale
and diversity.
I could argue that "civilization" is simply the inevitable product of
scale. Simply, just the result of the number of individuals.
Creativity has nothing to do with it *except to the degree that
solutions are kept within the bounds of the complexity threshold* and
despite scale the complexity of the system is unchanged beyond an
identifiable threshold.
In my view scale and complexity are not necessarily correlated and
problem solving efforts, in fact, do not increase in complexity -
they change and get smarter. "Smarter" or "intelligence" is a word
that seems to be missing from Tainter's discussion - intelligence
necessarily increases so that solutions live within the bounds of the
available complexity threshold.
When the complexity necessary for individuals in the system to
operate effectively has requirements that go beyond these limits then
the system remains constrained to function at the capacity of the
threshold - it simply cannot breach this threshold. Instituted
systems that require more complexity simply fail until the system is
constrained by natural selection and solutions within the complexity
threshold are re-established.
Now, in this limited response I have applied a simple algorithmic
definition of complexity - the number of steps required to do
anything - and I have avoided other characteristics of complexity -
such as decidability and termination - Tainter may be applying some
other measure and have some other way of characterizing complexity.
If he does, it isn't mentioned in his posting.
With respect,
Steven
--
Dr. Steven Ericsson-Zenith
Institute for Advanced Science & Engineering
http://iase.info
_______________________________________________
fis mailing list
fis@listas.unizar.es
http://webmail.unizar.es/mailman/listinfo/fis
Received on Thu Dec 14 17:23:42 2006
