Historical Origins of Chaos


Isaac Newton formulated his laws of motion as systems of differential equations, with initial conditions. Mathematically, it can be shown that solutions to such systems are in most cases unique.

Given a particular set of initial conditions,
the solution is determined uniquely.

(Check here for a rigorous mathematical statement of uniqueness for initial value problems.)

Another way of stating this is to say that

Given initial conditions are known (perfectly),
the solution is determined for all future tiimes (perfectly).

This reinforced the concept of "determinism" - that the future is an inevitable consequence of the past.

The subtlety in this interpretation of a valid mathematical theory is the assumption that initial conditions can ever be known "perfectly". Another assumption is the implicit assumption of "continuous dependence" - small changes in initial conditions lead to small changes in the solution.

The fact that large changes in the behavior of a complex system can depend on infinitesimal changes in the initial state led to the discovery of "chaotic systems".


One of the earliest physical systems in which chaos was "uncovered" was the so called N-body problem. The actual application was the solar system. The problem, as posed by King Oscar II of Sweden in 1887, was to solve the equations governing the motion of N arbitrary planets. In fact, the goal was to answer the question

Is our solar system stable for all time?

The French physicist and mathematician Pierre Laplace (1749-1827) was one of the earliest scientists who studied the stability of the solar system. He showed that in fact it was indeed stable for short times (ignoring the effects of tidal friction).

The French Mathematician Henri Poincare considered a simple case (N=3) and showed that the solution of 3 bodies could be extremely complex (non-periodic). He submitted a memoir Sur le probleme des trois corps et les equations de la dynamique to the Journal Acta Mathematica in 1890, and won the prize. He showed that the solution of the 3-body problem was in general unstable as time progressed.

The saving grace is that any instabilities in the solar system will appear only after millions or even billions of years.


In 1927, Balthazar van der Pol discovered that the equations governing vacuum tubes could exhibit "noisy" behavior. Although it was not realized at the time, this was another example of chaotic behavior.
Finally, in 1962, Edward Lorenz discovered that a simple climate model demonstrated extreme computational sensitivity. Restarting the computational model with initial data rounded to several decimal places, resulted in completely different behavior in time! This led to the first large scale investigations into "deterministic chaos".