Heinz von Foerster Festschrift

Mere Tuna

In the early days of Cybernetics, there was a certain brash but shy young student of Physics who became aware of such things as information theory, as formulated by Claude Shannon, Norbert Wiener's little book about cybernetics, which gave a name to the field, and the little automata by Gray Walter. Together, these suggested that purely mechanical things might truly be made to emulate the perceiving, goal-seeking and self-maintaining capacities of living things.

Now, most students of Physics tend to view the world as if looking over the shoulder of God, so the prospect of being able to emulate in some degree His highest acts of creation drew this young student irresistibly. When Heinz created the Biological Computer Lab, it was simple destiny that our student become a part of it, and so he did. And so was opened to him the door to his few minutes of fame, to which, by conventional wisdom, he was personally entitled.

In the BCL, there was much thought given in those days to the problem of understanding how living things manage to perceive other things, living or not. This appears to be done with the sensory machinery at their command, such as eyes and ears, and the nervous systems to which those are connected. The problem is something like this: just how is it, in fact, that, as my neighbor's child goes running past my yard, I see one child running, and not simply a vast number of different distributions of light and color one after the other flashing over my retina? The latter alternative couldn't really be called "seeing" at all, of course.

How, indeed, can a machine be made to recognize an abstract quality in its surroundings, some property which may be common to an infinite number of different patterns on its sensors? Such abilities were at that time taken to be mysterious and unique to the higher living organisms. This was the starting point of the theory of neural networks (which have been rediscovered in the last decade, and put to some practical use).

Much less significantly, this abstraction problem led also to the creation of the NumaRete, a machine consisting of an iterated network of identical elements, which could instantly report the number of areas of shadow, i.e., "objects," cast on its photocell "retina," utterly independently of their shape or position. That is, it appeared capable of the abstraction of number. This was the brainchild of our young student, who by this time was a student of Engineering.

Naturally, if the actual mechanism of the machine were revealed, it might be called mere trickery. This has been the fate of nearly every achievement in what has come to be called Artificial Intelligence. Suffice it to say that the mechanism was disarmingly simple, but functioned entirely convincingly to those who saw it work and tried to confuse it. Skeptics often put small objects inside of holes in larger ones, for instance. No such trick ever confused NumaRete, and skeptics were made true believers.

The machine's audience grew from the students at the Urbana campus, to a science exposition at Springfield, the state capital. There were some newspaper write-ups, and finally (with some strategic input from Heinz) there was an invitation to go to New York City, to show NumaRete on national television, hosted by none other than Walter Cronkite. Ah! Fame at last!

By a cruel twist of fate, NumaRete's hour of glory came posthumously. The Rete had acquired by that time a traveling case of fiberglass and aluminum, and had an attractive face designed by a commercial artist employed in the Engineering College at the Urbana campus. When it left Champaign with Heinz and its creator to go to the NBC television studios the Rete was a well-accoutered professional traveler. It rode in the luggage hold of the plane with professional dignity.

At the studios there was shock and dismay to find that the Rete was comatose after its trip. The magnificent travel case had been dropped by some callous baggage handler, dealing the network inside a lethal blow. Frantic attempts at neuro-surgery were inadequate; the Rete would count no more! It was Heinz, as ever ingenious and resourceful, who found the way to preserve the NumaRete's moment in the eye of the nation. Although the network was brain-dead, its numerical display could still be operated. A cord with an attached switch was run unobtrusively out of the case, and Heinz acted as the brain while the unconscious Rete lay before the cameras.

By a further curious turn of events, the scheduled broadcast was preempted by coverage of Winston Churchill's funeral. It was shown at a later time, though, and the most respected reporter and commentator, Walter Cronkite, put the work of the NumaRete's builder before the eyes of the nation.

This was the pinnacle of that engineering student's public visibility, and contributed to his future happiness by immeasurably raising his stature in the eyes of his father- and mother-in-law, who were startled to hear their son-in-law's name coming from the television program they were casually watching. He went on to a completely undistinguished career in the halls of academe, and is now very happy in retirement, indulging with his wife of forty-three years their mutual interests in their granddaughter, music, and world travel.

The body of the NumaRete lay for many years in storage in the domain of the BCL, and has now passed entirely out of sight.

Paul Weston
Electrical and Computer Engineering
University of Illinois, Urbana, Retired

Heinz von Foerster Festschrift