Crutchfield, James P., J. Doyne Farmer, Norman H. Packard, and Robert S. Shaw. “Chaos.”
This previously published
paper by James P. Crutchfield, J. Doyne Farmer, Norman H. Packard, and Robert
S. Shaw is reprinted here to give a broad introduction and background to the
science of chaos and complexity.
Until recently scientists
assumed that natural phenomena such as the weather or the roll of the dice
could in principle be predictable given sufficient information about them. Now
we know that this is impossible. “Simple deterministic systems with only a few
elements can generate random behavior.” Though the future is determined by the
past, small uncertainties are amplified so radically by chaotic systems that, in
practice, their behavior rapidly becomes unpredictable. Still there is “order
in chaos,” since elegant geometrical forms underlie and generate chaotic
behavior. The result is “a new paradigm in scientific modeling” which both
limits predictability in a fundamental way and yet extends the domain in which
nature can be at least partially predictable.
The article acknowledges the
challenge to Laplacian determinism posed by quantum mechanics for subatomic
phenomena like radioactive decay, but stresses that largescale chaotic
behavior, which focuses instead on macroscopic phenomena like the trajectory of
a baseball or the flow of water, “has nothing to do with quantum mechanics.” In
fact many chaotic systems display both predictable and unpredictable behavior,
like fluid motion which can be laminar or turbulent, even though they are
governed by the same equations of motion. As early as 1903, Henri Poincaré
suggested that the explanation lay in the exponential amplification of small
perturbations.
Chaos is an example of a
broad class of phenomena called dynamical systems. Such systems can be
described in terms of their state, including all relevant information about
them at a particular time, and an equation, or dynamic, that governs the
evolution of the state in time. The motion of the state, in turn, can be
represented by a moving point following an orbit in what is called state space.
The orbits of nonchaotic systems are simple curves in state space. For
example, the orbit of a simple pendulum in state space is a spiral ending at a
point when the pendulum comes to rest. A pendulum clock describes a cyclic, or
periodic, orbit, as does the human heart. Other systems move on the surface of
a torus in state space. Each of these structures characterizing the longterm
behavior of the system in state space  the point, the cycle, the torus  is called
an attractor since the system, if nudged, will tend to return to this structure
as it continues to move in time. Such systems are said to be predictable.
In 1963, Edward Lorenz of
MIT discovered a chaotic system in meteorology which showed exponential
spreading of its previously nearby orbits in state space. The spreading effect
is due to the fact that the surface on which its orbits lie is folded in state
space. Such a surface is called a strange attractor, and it has proven, in
fact, to be a fractal. The shape of a strange attractor resembles dough as it
is mixed, stretched, and folded by a baker. With this discovery we see that
“random behavior comes from more than just the amplification of errors and the
loss of the ability to predict; it is due to the complex orbits generated by
stretching and folding.”
The essay closes with some
profound questions about scientific method. If predictability is limited in
chaotic systems, how can the theory describing them be verified? Clearly this
will involve “relying on statistical and geometric properties rather than on
detailed prediction.” What about the assumption of reductionism in simple
physical systems? Chaotic systems display a level of behavioral complexity
which frequently cannot be deduced from a knowledge of the behavior of their
parts. Finally, the amplification of small fluctuations may be one way in which
nature gains “access to novelty” and may be related to our experience of
consciousness and free will.
Email
link  Printerfriendly  Feedback  Contributed by: CTNS/Vatican Observatory
571e
