"Is there Progress?"

I work on the question of evolution, not as it exists in Nature, but as a formal system which enables open-ended learning. Can we understand the process in enough detail to simulate the progress of biological complexity in pure software or electronics? A phenomena has appeared in many of my laboratory's experiments in learning across many different domains like game playing and robots. We have dubbed it a "Mediocre Stable State." It is an unexpected systematic equilibrium, where a collection of sub-optimal agents act together to prevent further progress. In dynamical systems, the MSS hides within cycles of forgetting that which has been already been learned.

When a MSS arises, instead of achieving creativity driven by merit based competition, progress is subverted through unspoken collusion. This occurs even in systems where agents cannot "think" but are selected by the invisible hand of a market. We know what collusion is: the two gas stations on opposite street corners fix their prices to divide the market. Hawks on both sides of a conflict work together to undermine progress towards peace. The union intimidates the pace-setter, lest he raise the work standards for everyone else. The telephone company undercapitalizes its own lucrative deployment of broadband, which might replace toll collection. Etc.

As a scientist with many interests in High Technology, of course I know there is progress. I am witness to new discoveries, new technologies, and the march of Moore's law. Clearly, the airplane, long distance communication, and the computer are revolutionarily progressive in amplifying human commerce, communication and even conflict. But these scientific and technological advances stand in stark contrast to the utter depressing lack of progress in human affairs.

Despite the generation of material wealth, health breakthroughs, and birth control methods which could end want and war, human social affairs are organized almost exactly the way they were 500 years ago. Human colonies seem — like ant colonies and dog packs — fixed by our genetic heritage, despite individual cognitive abilities. In fact, it is difficult to distinguish anymore between Dictatorships, Authoritarian Regimes, Monarchies, Theocracies, and Kleptocracies, or even one-party (or two party oscillatory) democracies. When labels are removed, it looks as if authority and power are still distributed in hierarchical oligarchies, arranged regionally. Stability of the oligarchic network is maintained by complex feedback loops involving wealth, loyalty, patronage, and control of the news.

Of course, I'm not against stability itself! But when patronage and loyalty (the collusion of the political system) are rewarded more than competitive merit and excellence, progress is subverted.

The 90's really felt like progress to me, especially with visible movement towards peace in certain regions of the world and an unparalleled creative burst in our industry. But now its like we've just been memory bombed back to the 1950's. The government is printing money and giving it to favored industries. We are fighting an invisible dehumanized enemy. War is reported as good for the economy. Loyalty to the fatherland must be demonstrated. One Phone Company to rule us all. An expensive arms race in space. And law breaking secret agents are the coolest characters on TV.

Havent we been here before? Haven't we learned anything?

Jordan B. Pollack is a computer science and complex systems professor at Brandeis University who works on AI, Artificial Life, Neural Networks, Evolution, Dynamical Systems, Games, Robotics, Machine Learning, and Educational Technology.

John Brockman, Editor and Publisher
contact: editor@edge.org
Copyright © 2002 by
Edge Foundation, Inc
All Rights Reserved.