Physicists and engineers have given a name to an exceedingly useful concept: transient behavior. A "transient" is a temporary wiggle, an instability in a system on the way to settling down into equilibrium. In an electrical circuit, flipping a switch may cause power surges or voltage spikes for a few milliseconds. In an economy, monetary policy changes may cause a rise in unemployment or in the general price level for a few years before long-term effects appear. In an encounter between two galaxies, colliding gas clouds may cause a burst of star formation lasting a few million years.

Transients, by definition, will pass --- unless the input tweaks that provoked the transients are themselves repeatedly changed. If we see a short-term transient signal and respond aggressively to it, we can easily get into a vicious cycle of reacting to the reaction to our own actions --- and so cause instability where we were trying to cure it. This happens often, particularly when significant time delays are involved in a system.

Flying an airplane, when we see the ground approaching our natural tendency is to pull back on the stick, increasing the angle of attack of the wings. For a few seconds, that gives us a transient altitude gain. Then, the plane slows and resumes its descent. We pull back more and more, until we stall and crash. What we really should have done was the opposite: get the stick forward to gain airspeed while feeding more gas to the engine. This counterintuitive feedback loop --- with the throttle controlling altitude and the stick controlling speed, once transients have passed --- is beautifully described in Wolfgang Langewiesche's classic aviation book Stick and Rudder.

Similarly, in medicine many patients get sicker the more they are treated; they are prescribed drugs to counteract the side effects of earlier drugs they have been given, and then more drugs to remedy the problems caused by those drugs, in an accelerating cascade. In government, politicians get the blame (or credit) for the long-term consequences of their predecessors' actions, and in turn pursue short-term policies which result in worsening conditions when they have left office. In business, a vigorous new CEO comes in, "cleans house", declares victory, and then moves on before the corporate wreckage comes tumbling down.

How can we see through the veil of transient behavior to the ultimate consequences of our choices? One way is to develop mental models of systems, with explicit recognition of feedback loops and delays between cause and effect. Such "systems thinking" forms the core of Peter Senge's highly readable book The Fifth Discipline. Working with systems models can help us avoid foolish and self-defeating "decisiveness" and instead develop mature, productive decision-making skills.

Tuesday, May 11, 1999 at 06:00:49 (EDT) = 1999-05-11

(see FifthDisciplinarians)


(correlates: Theory of Flight, BeatingExpectations, AlteredStates, ...)