Man and Computer: Uneasy Allies of 25 Years

In the 25 years since the late John von Neumann de signed the first fully modern electronic computer, tens of thousands of the controver sial machines have reached into nearly every realm of human endeavor to alter dra matically the shape of to day's society.

Yet however commonplace, indispensable and obedient the computer has become, the accommodation between man and this most manlike of machines remains uneasy.

This was strikingly evident at a recent three‐day sympo sium at the Institute for Ad vanced Study in Princeton, N.J., where an invited group of about 50 leading scien tists and scholars gathered to honor von Neumann and his landmark contributions to the computer.

Many of those who spoke displayed a fear that insofar as the computer simulates thinking, it threatens the primacy of man. And, to the extent that the computer de mands new, more rigorous methods of thinking or ap pears to lend undue credulity to the information it spews out, there was also some con cern that the machine was displacing the quality of human elegance in science.

The subject of the confer ence was the impact of the computer in the last quarter century—not on commerce, where the machine is well known to every credit card holder and telephone customer —but on selected fields of science and learning, where its impact has been far more profound.

By the meetings end, there was a deepened impression that, for all the problems solved and data processed, the less tangible conse quences of the computer might have been the most important. Thus, it is not so much the answers the com puter has given us that are remarkable as it is that, with so many earlier problems be hind us, we are now able to ask questions of unprece dented boldness.

Basic Questions Now Routine

The computing machine has propelled science through so many once insurmountable problems in such a short time that “hard” scientists are now routinely expecting an swers to the most fundamen tal questions about the na ture of the universe and the meaning of life.

Dr. Sydney Brenner, a biol ogist at the Medical Research Council Laboratory of Molec ular Biology in Cambridge, England, for example, said that biologists are using the knowledge of how computers work as a guide in their search for possible parallels in the way living organisms work.

A living organism, Dr. Bren ner said, begins with a set of instructions, a program writ ten in the genetic code, and somehow the “chemical com putations” specified by the program produce not just a blob of new chemicals but a complex, well‐organized liv ing creature.

“Now that we know how the genetic code is translated into molecules of various sorts,” Dr. Brenner said, “we want to go further to ask not how do you make a protein but how do you make a fin ger or a nervous system.

“We want to know how that happens and, quite frankly, I'm prepared for the idea that it happens in a pure ly mechanistic way and that I, as a biologist, am really only a natural engineer.”

A Similar Impact

The computer's impact on other disciplines has been similar.

“The real importance of the computer in psychology today,” said Dr. George A. Miller, professor of psychol ogy at Rockefeller Univer sity, “is that it has created a new and pervasive state of mind. Psychologists have come to take for granted in recent years that men and computers are merely two different species of a more abstract genus called ‘infor mation processing systems.’”

Dr. Miller said that what he calls “the computer meta phor” has led many psychol ogists to a revival of purely mechanistic views of life in which mental functions are thought of as analogous to the various functions of com puters.

Not all the conferees ac cepted the metaphor, how ever. Anthony G. Oettinger, professor of linguistics and applied mathematics at Har vard University, held that merely because computers simulate certain kinds of mental activity, it does not mean that computers are brains or that they work the same way. Nor were all con vinced that computers had led to fundamental changes in science.

“Both the miscroscope and the telescope led to dramatic advances in our understand ing of nature when they were first introduced and it is doubtful whether comput ers have yet done as much,” said Keith V. Roberts, a phys icist at England's Culham Laboratory, an atomic en ergy research center.

Nonetheless, the computers are helping many scholars cope with growing complexi ties.

Michael Atiyah, professor of mathematics at the Insti tute for Advanced Study, said that his field was getting so complicated that the working out of formal proofs of new theorems was being turned over to computers while the mathematicians stuck to the more intellectually satisfying job of theoretical discussion.

Already, he said, the step‐ by‐step checking out of proofs is so lengthy and diffi cult that when a mathemati cian reaches the middle of a proof, he has lost his grasp of the early part. When he sight to the problem that he started with.

“lt may be,” Dr. Atiyah said, “that the computer is the only one who really un derstands the kind of mathe matics being done today.”

To many at the conference, the remark had a serious point, particularly for the mathematicians. There was a sizable contingent present— including the logician Kurt Gödel of the Institute and Mark Kac of Rockefeller Uni versity. They devoted con siderable discussion to wheth er a computer that can handle the mathematics of complex physical phenomena that are beyond the compre hension of the human mind actually “understands” the phenomena. finishes he has no further in

Indeed, It seemed that the very possibility of under standing all of nature—which is, after all, the loftiest goal of all science—was in ques tion because, through the computer, many disciplines have discovered hitherto un known strange phenomena.

These phenomena, said Stanislaw Ulam, senior scien tist at Los Alamos Scientific Laboratory, “seem to make matters more difficult to un derstand than had been ex pected.”

The Nineteenth Century hope that just a little more investigation would reveal the ultimate truths about the workings of the natural world has now virtually gone. Each layer of the onion that is removed reveals not the irreducible core but yet an other layer for which whole new theories and mathemati cal techniques must be de veloped to say that it is understood.

Like the long mathematical proofs, many of these phe nomena can be studied only by using very difficult mathe matics that only a computer can handle efficiently and many appear to require mathematical manipulations that are beyond today's com puters.

What hope is there then, the conferees asked, that the mind of man can understand the workings of the universe from the most distant galaxy to the most ephemeral atomic particle?

Never far from the sessions was the memory of von Neu mann himself—a warm, out going mathematical genius who, though he remains little known to the public, already stands as one of the giants of Twentieth Century scien tific thought.

He was, as Carl Kaysen, director of the Institute and chairman of the conference, said, “one of the most daz zling intellects of his time.” Von Neumann (pronounced “von noy‐man”) did not in vent the computer. No one person can be said to have done that Computing ma chines of one sort or another have been built for centuries.

Von Neumann's contribu tion was a major improve ment on the design of the electronic computer that con verted it from a special pur pose machine to a flexible, all‐purpose device that could be used in a variety of ways.

The computers that existed before von Neumann's entry into the field relied on slow mechanical methods of stor ing the program, the set of instructions that direct the computer's operations. These included the Mod 1, devel oped in 1938 by George R. Stibitz at the Bell Telephone Laboratories, and the 1944 Mark I, designed at Harvard University by Howard Aiken.

The ENIAC (Electronic Nu merical Integrator and Cal culator), which began oper ation in 1946, was the first completely electronic com puter, replacing relays with vacuum tubes that speeded the computation rate a thou sandfold. It was designed by J. Presper Eckert and John W. Mauchly and their associ ates at the University of Pennsylvania.

Still, however, the ENIAC depended on a human oper ator to plug in hundreds of wires to connect one part of the machine with another. The way the wires were plugged in determined the operations to be performed when the machine was turned on.

Still the Computer Pattern

Von Neumann developed a concept that did away with the plugboard method. He de signed a machine in which the program could be written in numbers and stored in the memory exactly as if they were numbers to be manipu lated in the computation.

The internally stored pro gram made it possible to have a general purpose com puter that, almost instantane ously, could electronically reorganize its internal cir cuits to meet special needs simply by feeding in a set of numbers.

This design, which includ ed a new organization of the computer subassemblies — once referred to as the “von Neumann architecture” or the “von Neumann machine”— remains today the basic pat tern on which nearly all digital computers are made.

More than any other tool in the history of science, the computer has found applica tions in almost every field of research. Geologists use com puters to analyze tremors for patterns that may some day forecast earthquakes. Chemists watch TV screens as the machines, display slow motion pictures of dia grammed molecules combin ing. Space scientists use dozens of computers to process streams of data from satellites, stations on the moon and deep space probes.

In these and scores of other applications, the com puter is doing work that otherwise would be beyond the capacity of scientists either because doing the same work by hand would take so much time it never would be attempted, or be cause the computer does things that could not be done at all by other means.

Because the computer deals only with information that can be expressed in mathematical terms and can deal with it only in a rigor ously defined way, the con ferees said, the computer is forcing many disciplines to re‐evaluate the quality of in formation they work with.

Von Neumann himself once predicted that the computer would keep scientists honest because it required that a sci entific problem be formulated in terms meaningful to a computer—terms that would therefore be logical.

A Different Impact

Several participants at the conference observed that while the computer might have forced the older scholars to construe their problems in new ways, the impact on younger persons who had grown up with the computer had been different.

“Some of these younger fellows know Fortran [for mula‐translation, a computer language] better than Eng lish,” said Lawrence R. Klein, professor of economics at the University of Pennsylvania. He said that because comput er time had become so cheap, students could afford to use clumsy trial‐and‐error meth ods of problem solving rather than thinking hard to come up with the simpler and more elegant insights necessary to solve the problem without the prodigious capacity of the computer.

Despite its shortcomings, said Keith V. Roberts, the British physicist, the compu ter has established itself as an important new way to ap proach a scientific problem. Heretofore scientists could seek answers through direct experimentation or theoreti cal investigation—the extrap olation of logical conse quences of observed phenom ena.

With the computer, Dr. Roberts and others noted, there is now the computa tional method. Phenomena that cannot be observed di rectly may be studied in the computer if enough is known to write certain mathematical descriptions of the phenome na. The machine works out mathematical descriptions of the phenomenon at various intervals throughout its course of change.

Far more common than using computers to simulate natural phenomena is the use of the machines simply to collect and process large masses of data generated by conventional experiments. Be cause the computer can han dle the volume of numbers, more complex experiments are now being done in many laboratories.

The various disciplines that rely heavily on statistical methods of research, such as sociology and psychology, use computers heavily to compile into a coherent form the results of, for example, thousands of questionnaires, each with scores of questions having several possible an swers.

One important application of the computer in the “soft” sciences is in economics, where researchers are now able to study the flow of wealth almost as physicists study the flow of heat.

Yet another way in which computers are aiding science is in the actual carrying out of experiments. In several disciplines complex equip ment must be started, stopped and procedures changed many times and data must be recorded at each point.

The impact of the computer on science has been far reaching and the machine has taken over much of the effort that researchers once had to make themselves. “The com puter has reduced Nobel work to the M.Sc. level,” Dr. Bren ner remarked.

For all its concrete applica tions, however, the computer still carries a more abstract burden—its similarity in cer tain respects to the human mind and, therefore, its threat to the primacy of man.

Marvin Minsky, who is working on “artificial intel ligence” machines at the Mas sachusetts Institute of Tech nology, said that he believes the computer to be the best analog of the human brain that has yet been developed. He predicted that the future would bring new types of computers that would rival man's brain in many ways.

An intriguing point of view was offered by Philip Morri son, professor of physics at the Massachusetts Institute of Technology. He forecast that in the foreseeable future, man would have to recognize three new forms of life in ad dition to the kind of human, plant and animal life already on earth. He said these would be life forms created in the test tube by biologists using off‐the‐shelf chemicals, living organisms on other planets— simple forms on Mars or in telligent forms in more dis tant places whose radio sig nals we receive—and a form of machine life that will be the outgrowth of today's computer science.

Dr. Morrison predicted that although the machines would have many of the qualities that are attributed to existing life forms they would not be the equal of man.

“We ought to see an arti ficial, synthetic device pos sibly not designed from the beginning by any human pro gramer but only begun at some [lower] hierarchial level by such a programmer and evolving subsequent [more sophisticated] hierarchies by its own internal directions, which, when complete, would behave in a way simulating the behavior which we're used to. I think this is all likely to happen in the fore seeable future.”

Dr. Morrison emphasized, however, that it would be its own class of life.

To think of a highly com plex machine as a form of life is not so unthinkable as it once was, for the traditional definitions of life are now quite shaky.

Few biologists would argue that an essential ingredient of life is that it be built of proteins, fats, and carbohy drates. More important now, and quite in keeping with the computer metaphor, the es sentials are functions and not forms. If wires and magnets and transistors can be made to function in ways analogous to the function of nerve fibers and protoplasm, are they not just as much alive?

Would such a machine, liv ing or not, be the product of some new diabolical force in men? No, Dr. Morrison con tended, it would be the logical culmination of an ancient tradition—including even the humanistic Greeks.

“I think it is fair to say,” Dr. Morrison said, “that there is no theme in intellectual life so persistent as the theme of the mechanical reproduc tion of life and the effort to reflect the cosmic motions.”

That theme, he contended, is present today in the work of computer scientists who are striving to make the ma chine still more life‐like.