October 10, 1968
This report might well have been called The Quest for Simplicity. For that is the great gift of information automata for mankind; they enable the human mind to focus on simplicity in the midst of engulfing complexity. This is the greatest option we know, for through it the personal man can feel, indeed, that he controls at least a little the choices from which his actions and perhaps his fate will come. His ultimate freedom of choice, and thus his freedom itself (not only as a manager but), as a citizen and a self -determining human being, will be based on a conviction that he has significant, even if not complete, control of an access to the information which guides his doings. Thus each man must have a belief that he can simplify and pick meaning from confusion and complexity. The infinite variety of human reactions to this compulsion can range from the detachment of Thoreau at Walden Pond, to knob twirling on TV receivers (to cut out some parts of the intricacy of daily information injections into every man’s home). Some sense of the intensity of this issue nowadays is reflected in the words of one of my friends who was long a co-worker in the Washington welter of information exchange. He asserted that the “user is becoming the hunted, not the hunter for information,” as has been man’s earlier classic exercise.
Just as the historic function of the telephone has been to provide the individual with selective communication with one other person among the multitude, so can digital machines give the human mind ways to deal with prime and central thoughts, with ideas and concepts amid the welter of knowledge, data, facts, in which the workings of society and business and government are today immersed. The merciful capability of talking with one desired person through a switching system, reaching at least a hundred million separate stations, is a boon that every human can have. So, in the virtual infinity of human knowledge, of all the data in the world, we can see the heroic role of the digital machine system in letting the teacher’s, or the scientist’s, or the economist’s mind deal with something like one thing at a time.
Notice that these devices are not called, at this stage, thinking machines--they are knowledge machines, in which endless information can be stored and processed. Thus, the human mind is allowed to exercise in its marvelous modes of association, of intuition, of invention, of concept formation with relief from a suffocation of the complexity of detailed information alone.
So let us examine, and illustrate now, some of the levels of complexity which can be reduced by the knowledge machine. Thus will be opened a new era of analytic and imaginative thought and action for the conduct of business, science, government, and all human affairs.
We can think, for instance, of the dimensions of time and space, which are, indeed, the fundamental limitations for human action. Already, telecommunications have yielded almost real-time connections on a nearly global scale with any two minds and personalities which seek to communicate. No longer is some mechanical relay system of word passing, by face-to-face mouth or by some printed form, necessary for the detailed interaction of people in touch with such a communications net. But as the questing mind seeks knowledge, the rate of perceptual absorption averages but 40 bits per second for any person. His output, in terms of spoken or written words, may be even at a slower speed. Since even a single picture, which he may wish to describe, can contain, as we know from the analyses of information theory, millions or even billions of bits, the input and output of the information processing which can be done by the individual is still limited in time, indeed. In this regard, digital computers are a prime time-saving machine. Nowadays they regularly process at a rate 3 to 10 million times a second, and thus several million times faster than man, the logical operations of arithmetic, letter identifications and transposition, symbol arrangement, and the like. A typical reaction time for a person to make a mark with a pencil, to speak a syllable of a word, or to effect a logical choice in calculation is one-tenth to eight-tenths of a second, at best. A machine can treat all these matters in a few microseconds at worst, and may also give access to a 200 million word (36 bits each) memory in half a second, or to a 262 thousand word (each of 36 bits) memory in a microsecond. We often think of speed of travel as a major time saver and virtual multiplier of life activities, as indeed it is; but in human history we have gone from walking at about 3 or 4 miles an hour to supersonic flight at about 2000 miles an hour or occasional rocket propulsion at 4,000 miles an hour, an increase of a factor of a thousand at best. In the handling of information and the mental operations necessary to process it, the present digital machine can exceed the human speed by a factor of at least a million, and we know how to increase that by something like another thousand fold. In this sense, is not the digital knowledge machine and its associated communication net (within which the speed of travel is the speed of light (186,000 miles a second)) the greatest life extension method yet encountered since the Fountain of Youth?
As to space, we do not know how to think or to visualize in more than three dimensions--nor to travel or move in less than three. So even moving as fast as thought leaves us still space-limited. Interrelationships among the many variables, with which a 20th Century citizen must deal every hour of the day, cannot really be depicted, as mathematics has long illustrated, in more than three-dimensional plots, and mostly we depend on two. Even this use of graphics is relatively recent in economic and industrial affairs, but digital computers can easily be programmed to deal with many dimensions beyond three and are being programmed to display results containing relationships of even hundreds or/thousands of interacting variables whose spatial representation goes far beyond human visualization. So already we have suggested a reduction in complexity which as the poet has said gives man anew the chance
“To follow knowledge like a sinking star,”
“Beyond the utmost bounds of human thought.”
Yet such a quest remains a most human matter, for it is the mind which creates and programs the machine, eventually to master such complexity. And of all our qualities, the mind is the most human. Indeed, as you will know, the very word mind and the word man came from the same linguistic root, and in the earliest Sanskrit “man” meant “to think”. Through the endless corridors of the edifice of language, men and mentality were long spoken as the same, and memory, too, was said with the same sounds. This whole sweep of evolution has been symbolized in the mystery and wonder of the human mind by poetry, as in Davis’ “Immorality of the Soul”:
“God first made angels, bodiless, pure minds”;
“Then other things which mindless bodies be”;
“Last He made man.”
Now we see, after these millennia of the strivings of mankind, a new stage of evolution, in which the long accepted limitations of the mind for enfolding and reacting to abstraction and complexity are challenged. These limits may be moved by the help we can get from the knowledge machine.
The thrust of thinking about the philosophy of knowledge machines carries us directly into their operational qualities. Namely, they must be a tool of the individual, and accessible to the quest of the single mind. In this deeper sense, their role cannot be realized as some detached institutional equipment, like a vending device which delivers a hot TV dinner of neatly wrapped and packaged data, obtained by pressing the right button (following, of course, insertion of the right coins). Like the telephone, they are an extension of the individual personality and mind if they are to act, indeed, as a supplement to human intelligence. For creative coherent thought does not come in pieces partly from one person or partly from another, despite the effectivity of teams. Rather, we are now getting to that phase of the use of digital machines through communications so that each person, very personally, can deal with his particular burden of complexity and his particular quest for insight.
Hence as you can see, in this report we shall not dwell on the wonders of machines themselves or even their software, despite the marvels of modern solid state electronics--of transistors, diodes, and magnetics in their modern micro-thin, integrated circuit forms. It is true that these enabled the Computer Age, for the original electrical digital machine of Stibitz in 1940, based on telephone relays for logic operations, and the subsequent triumph of Aiken and Eckert and Mauchly, with their vacuum tube electronics, sophisticated memories, and program storage, could hardly have provided the 30,000 major diverse digital machines now used in the U.S.A. But, mindful of this, and the future changes that physical and biosciences may present, we shall pursue our theme of the reduction of complexity by the knowledge machine.
To get us in the right frame of mind let’s think for a moment about algebra. (Many of us may not have thought about algebra for decades, but it was something which opened our minds to a new train of thought and abstraction at some stage in our lives.) Now, Dr. W. S. Brown has created an ALPAK system and an ALTRAN language for the working of symbolic algebra. With it, rational functions of several variables, polynomials in several variables, truncated power series with rational function coefficients, matrices of rational functions can be treated by addition, subtraction, multiplication, division, and integral exponentiation. But it also enables substitution, differentiation, computation of lowest common divisors, simplification through auxiliary relations. With it, one man-hour by a competent, professional mathematician is equaled by one second on even the IBM 7094 computer, and in that machine the memory will hold up to 8,000 polynomial terms. It represents a polynomial not in some literally coded form of numbers, names, symbols, and parentheses, but rather as an array of coefficients and exponents so that a term of a multi-variable polynomial can be stored in two machine words. Here is the tip-off--the user has made the machine go beyond mere symbol manipulation, or even the marvelous versatility of Fortran II or Fortran IV.
But maybe people don’t want algebra, or at least don’t want to think they are using it. Let’s go to geometry, or more broadly the perception of shape and form and its analytic expression. This is what plots and graphs are, and anyone who hasn’t tried those may now be excused from the room, “What happens here is that the digital machine has been highly instructed, to make graphs. It selects the x and y scales, adjusts them automatically to encompass the available data (for instance over an easily variable ratio of height to width of the finished graph from 0.25 to 2.5} and communicates all this to an electron beam, such as in a cathode-ray tube. The examples we show clearly accent the view that quantities of data, quite inaccessible and incomprehensible to the decision seeker in manual form, now can be displayed for relatively simple use in his mind.
But you may ask what kind of card-punching, tape-holing horde must be assembled to feed this machine for its elegant graphics display. Are we in for a new era of punch-hunkies? Fortunately, no; the keyboard console and attached TOUCH-TONE telephone are already at the fingertips. But even this can be reserved for the most personal and individual interactions with graphics or graphing from the machine, for the raw data input is easily achieved, automatically, from the original measurements either by straight number code or in some analog signals by efficient analog-to-digital converters. My associate, Dr. J. E. Kruskal, has collected many examples, of how graphical assimilation of complex data in digital computers reveals relationships completely intractable to correlation coefficient, polynomial regression, or other formal statistical analyses. Such would be the case in the scatter diagram of points forming, for instance, a simple five-pointed star. But often the information relevant to business operations and decisions includes a host of concurrent variables. These might be end-products of a large diversified machinery factory, such as is necessary for automobile fabrication. Each of these products may be subject to R measured qualities apiece. The qualities can be exceedingly numerous, such as costs, shape and storage qualities, deterioration, weight for floor loading and transportation, packing qualities, reliability, tool demand, labor input, and so forth and so on, going into scores or even hundreds of measurable features. The data thus available could be regarded as N-points, one for each of the products, in R dimensional space, a dimension for each of the qualities. We said in the Introduction that visualization of these relationships, even for R = 3, in ordinary three dimensions, is difficult and rarely attempted. Yet methods now being developed by Dr. Roger Shepard, Dr. J. D. Carroll and Dr. Kruskal, based on the use of digital computers (and their associated graphics), permit astonishing reduction in dimensions and consequent comprehension of surprising interactions in the system. For instance, the interpoint distances in the R dimensional space containing the original N-points can be massively processed in various important ways. One would be by Shepard’s least squares monotone regression, which would arrange a monotonic sequence of the square -root -of -the -sum-of -the -squares separations of the points in multidimensional space. A newer and more sophisticated form developed by Carroll and Shepard can take the scattered points, say distributed on a sphere, and display them on a flat graph. This is called continuity scaling and can be accomplished without human intervention. This makes the relationship between the new coordinates and the original ones as smooth as possible. The illustration we have comes from our usual limitation to thinking in two or three dimensions, but can easily be applied to large values of R. Such matters as market research in determining the preferences of consumers for a diversity of circumstances and types and, indeed, determination of social attitudes generally will be revolutionized by this kind of graphing, derived from the knowledge machine.
But there are still other forms of expression from the knowledge machine which aid in new ways the reduction of complexity for the questing mind. We should like to see complex situations portrayed not only as graphs or line drawings reducing spatial relations to the easily visual, as we asserted at the beginning, but also those which indeed compress or reduce the flow of time itself. This of course implies moving displays in which the sequences of data or conditions extending over weeks or years or decades can be treated and displayed in convenient condensed dynamic form by the knowledge machine. Again, depending on the elegant control of electron beam movements directly by the digital computer, my associates Dr. K. G. Knowiton, Dr. E. E. Zajac, Dr. F. W. Sinden, and severs! of their co-workers, and also concurrently workers in other computer centers, such as at Livermore Scientific Laboratory, Boeing Aircraft, M.I. T. , and the Systems Development Corporation, have created just such moving pictures and animated displays. Indeed, some of the basic principles of our report today, that these information and communication systems do represent the new mode of simplification, have been tested by widespread experiments in educational use of representation of complex mechanics (such as the motion of celestial bodies), of intricate physics (such as the creation and diffusion of charged electrical plasma in crystals), and of various forms of analytic mathematics based on computer-generated films. Results have indeed confirmed our postulate that this is a new option for simplicity in the search for knowledge and understanding.
Also, in keeping with our theme that the man must be master, not only of information but of the knowledge machine processing it, there are various input schemes for computer graphics which are also, themselves, graphical in movement. One of the earliest is the Rand tablet based on the Teager tablet, devised by Professor Herbert M. Teager, then at M. I. T. The tablet provides coordinate input, by x and y position lines etched in copper on either side of an insulating film which contains also an encoding section. The position lines and encoding sections react capacitatively by a Gray code pattern determined by the position or coordinates of the lines. A stylus of high impedance then permits the input, to a cathode-ray tube, of various patterns traced on the tablet.
Another somewhat more elaborate system uses a voltage pickup stylus responding to voltages applied on a transparent but conducting coating of stannic oxide on the tube surface itself. Tracking balls, joy sticks and other kinds of flexible position-generating inputs have also been developed. Among all, the so-called light pen or photo sensor which is adapted to the appropriate display languages which produce images directly on the face of the cathode-ray tube (light from those very images can be picked up and fed back through the sensor into the machine cycle again) seems to be the most satisfactory manual input at present. This feeding back and taking heed of the position of the light sensor or light pen is best done by a separate smaller computer, which else administers the major signal processing eventually necessary from the large machine. While at present this system is operative mostly for technical purposes, such as circuit design, body structures, and mechanical design, and various civil engineering education applications, such as done at M.I. T. , it too, in the long run, affords a new opportunity for personal use of the manager, organizer and leader. He can, in a graphical display sense, talk back to the plots of data and effects which the knowledge machine is able to present in either static or dynamic form. Again, this can be done with reduction of complexity and enhancement of his own expression.
In all the cases like it, and in fact, in the whole activity with the knowledge machine, a close match in space and time between human reactions (requiring say a few hundreds of milliseconds typically) and machine reactions in terms of graphical output, data reduction or acoustic responses must be assured. This requires ready access and flexible participation of the communications net, since the inputs and outputs must surely be where the person wants to be and the peak in processing must of course occur at the most efficient part of the machine. So far, literal time-sharing of large central processors, in the sense that great communications network channels are multiplexed or even that a system like TASI is used in a transatlantic cable transmission, has not yet actually been achieved, despite strong ambitions by such enterprising sponsors as ARPA in the Department of Defense and the very ingenious scientific contributions of performers like M.I.T. and S.D.C. Nevertheless, progress is appreciable and . ARPA intends to have an experimental system of as many as 1500 remote consoles linked in a communication net to about 35 major processors at 16 locations around the nation.
So, indeed, we look forward to the further liberation of man from his own and nature’s complexity by use of the knowledge machine. Also, this will extend the presently time- and space-limited dimensions of individuality. Let us accent this by the reduction of an ultimate mode of complexity--that of natural language and its symbolic representation in conventional printing. Here the story is only begun for, of course, reduction we are discussing like that of the reduction of masses of physical behavior and of mathematical manipulation must eventually involve an interpretation of meaning in language, in extracting the simplicities. This we are far, far from, but certain provocative steps have been taken. For instance, Dr. Lee McMahon, among others, has had encouraging results in structuring a form of English, in his case called FASE (Fundamentally Analyzable Simplified English), which at least in the formalities of parsing the actual sentence structure yields a consistent composition and decomposition by machine. Also in a graphical sense, Dr. M. V. Mathews and his associates have produced a system of creating English letters, such as formed by the movable type of the Gutenberg era, by computer writing with an electron beam. Thus printing enters a new era, in which as many as 200,000 characters a. second have been generated this way on a vacuum tube screen without any master type or mechanical movement. There has been produced also automatic editing and justification so that the actual format of recorded language can be fully governed by machine. In this mode, the Rugged Individualist may reach his summit, by being able to compose and to publish his autobiography without other (or some would say-any) human intervention! What more individual freedom can be imagined? And you can see that the future of book and other printing and publishing is in for some changes!
But this use of the electron beam controlled by the computer leads us to the role of time again. The mode of use of the digital machine and communications compresses time, or stretches out time, as we have said, manages time in a new way, in the continuum we have referred to. So knowledge becomes not only visible, but playable, to match the timing of the mental quests of the user. And this is what we shall try to illustrate. This happens by the enormous flexibility of the knowledge machine in processing the digitized knowledge and in displaying it. In each case, as we’ll see, the display is by the machine manipulation of an electron beam which is directly photographed or otherwise displayed, and there is no analogue in between, there is no mechanical linkage, or human intervention.
Case 1 is an example of the guidance of research and development in the current state of microwave technology from simple crystals. Dr. Gunn, now at the IBM Laboratories, some years ago made the astonishing observation that microwave radiation could occur from single crystals of gallium arsenide, and he thereby baffled semiconductor and electronics engineers all around the world. And this remained so until my associates Chynoweth and McCumber finally found that a computer analysis would reveal the cause of this astonishing effect. I cite this chronology because some very able people, of course including Gunn himself, all extremely competent physicists, attempted to understand this remarkable effect. You just take a crystal. It doesn’t have a junction or anything. You put a potential on it, and you get microwaves radiating. And so here was a worthy challenge.
Now you will see a computer-generated plot of very simple quantities, among all this complexity of microwave generation, simply of electric fields on the vertical axis, and a distance through the crystal which is two-tenths of a millimeter, in this case, on the horizontal axis.
Here is the field, here is the distance, the critical thing is to watch this field change as it migrates through the crystal, and the critical element that you’re seeing of these analyses done by machine is this “catastrophe” that just occurred, this upsurge in field. But you are seeing events, each of which is separated by 2. 5 trillionths of a second. And therefore, the physicist was able to analyze an event which had been displayed to him as a result of complex equations, occurring in steps of 2.5 trillionths of a second. He could easily distinguish the physical situation over that distance.
Now the next example is quite a different but equally pervasive variable of a series of company earnings which are presumably related to business trends, although as you’ll see this may be debatable, as you see the result. These, however, bear on this “time-axis” problem again, because they are shown in a way that even a president, to say nothing of a chairman, would have time to look at over a period of 20--they’re shown over a period of 20 years. We found this very sobering, incidentally, to a couple of presidents, but the plots are the result of course of a computer doing all the calculations, doing all the assembly of the data, which are very extensive in this case; I’m simply showing you, with the time-axis changing as it goes, what happened to the earnings. They are not for a conglomerate, but they may qualify--or at least, conversely, a conglomerate I should think would qualify for this treatment in time, if one lasts twenty years.
But the results can be extended of course to any sort of information processes of that kind. Now the next little section of the movie will show actually some of the techniques of making these lines. This is Knowlton’s Beflix method, and you see it’s enormously magnified so that you see the digital formation by the electron beam. The beam produces each one of these squares. It’s an enormous magnification, and it’s a step-by-step process preserving the digital quantized nature that we spoke about in the first place.
Now there is another one. The engineer and the scientist always have had this classic problem of the behavior of atomic or molecular particles, or objects, they can be particles of smog, or atmosphere, bumping into each other, they can be molecules of gasoline in a combustion chamber, or they can be, of course, reaction molecules in a chemical process unit. But in all cases, the kinetic behavior of interacting particles has been one of the great challenges of our age.
Zajac and Sinden have made a new start on this by teaching the machine, as in each of these cases we’ve seen. The machine is doing all of the equations. This is no simulation, you see, this is a complete analytic solution to the situation. But they have taught the machine in this case to do multiple collision problems, to express its results, not in an infinite series of charts or equations that you have to shuffle around, or tables of data, but a direct picture of the particles bumping, with vectors appearing on them, which represent the energy and direction which they acquire from collision. Now, the number of particles is still limited, but it’s given some new insights about kinetic processes.
In still another area of somewhat larger scope, the weather is always with us. And this has become of critical importance in some detail for our own research and planning, not only because it’s nice not to have all the telephone poles blown down (but we’re getting used to that), but rather that it’s been said that the moisture envelope of the earth is too wet to permit the successful propagation of radio microwaves much above ten or fifteen gigacycles.
Now this has been a. legend which is pretty widely accepted. Of course there have been communication legends before, that you couldn’t propagate radio waves over long distances and science has rather modified that. So we’re encouraged to modify the idea that the envelope of this planet has, to a certain degree, a prohibitive amount of humidity. The stupidity may be unlimited, but the humidity may have useful zones.
So the first thing we do of course is to find out what the actual precipitation is over a significant area, and particularly what its variations are, to see whether of course these average figures that the weather is always represented by are true, or whether there is a fine structure, which means that the averages are misleading, and that the averages in fact give you much too pessimistic a figure.
Well, we’ll see indeed that’s the case. These results are derived from a network in the eastern part of the country here, connected into the telephone system. A new kind of rain gauge has been invented by Tillotson and his people at our laboratory at Crawford Hill, and this is connected into a network, fed into a computer directly, which puts out the results in this way.
Now what you see in this way is a chart, a plot, of rainfall as measured, by a grayscale in this case, and it gives one a very useful depiction of the tremendous variation in precipitation during very heavy rains, also during light rains, so that our chances for radio microwave propagation, particularly by diversity paths, is very much better, both by satellite and by earth radio at these high frequencies, than has been presumed, and than has been acted on. But the data processing as well as the depiction of what’s happening here would not have been accessible without the communications and computer net. This sort of thing will undoubtedly be useful in all sorts of commercial operations which depend on the weather.
So we have in technology a number of examples of how these systems are revising our thinking. Now, what about the human perception element? These cases are all things in which we’ve learned something about physical science or engineering or technology. But shouldn’t we, if this is a great supplement to the mind of man, be able to learn something about man himself from them? Well this is probably the best opportunity. But the little case we’ll show now is one in which we found a new element of man’s sensing ability. In spite of having looked at how he saw and heard for a great many years, we find there are new aspects of this because of the things you can do with the digital knowledge machine.
Namely, this is, I suppose, particularly timely for us, for we have seen a new human capacity for perception in confusion, or in noise. The remarkable temporal sense of the human sensor appears to be behind this. Julesz has been able, because of the machine methods we have cited, to create completely noisy pictures, as completely random dot patterns to look at. We’ve never had these before, but now the computer will generate a pattern which nobody has ever seer; before. Therefore, anything you see in that is not the result of your past experience, or the result of cues of past experience which we have always had in earlier visual studies. And then, if you do see something under some particular set of conditions, you can be quite confident that it is a reflection of your basic perceptual ability.
This is a rather amusing example of that, that Julesz has found. The next little projection shows the random dot pattern, the perfect noise in graphics prepared by the computer. And you see, in fact, all these dots. Now if we stopped the picture, which we don’t want to do, but if we did it would look also randomly prepared. But as you are looking at it, you will see something appearing in the center which is not moving, though actually the whole thing is moving, and if we stopped at this time, you wouldn’t have seen that separation, you wouldn’t have seen that separate section at all. It’s a super-imposition which demonstrates the ability of man’s mind and visual system to perceive and correlate certain things in a random pattern which we never would have suspected could be done. And it depends entirely on motion. You wouldn’t see it at all if it weren’t moving.
Well, we have barely begun to see, we think, the horizons which these knowledge machines provide. A great role for them, as we have said, will be in human and social science studies, and those will probably be greater roles than any we have been able to touch on this evening.
We said however that the computer and communications system had indeed modified the age-old limitations of man in time and space, and let us see the one little start we are making on the spatial representation. We said that the machine knows no limitation to three-dimensional processing of information, and Noll and his associates have prepared numbers of projections which go into higher dimensions.
Now we don’t know how to look at them really to understand them, but we will show now a projection of a title, it’s just a kind of far-out title, which is rotated and projected by the machine through the fourth dimension, perfectly rigorously. I think you will observe some interesting psychological feelings as you look at it, although it will appear without the use of even Polaroids or other three-D projection aids. It will appear as just kind of a curious convolution to you. I think that you will perceive something more than a 3-D quality.
Here is the 4-D hyper cube; these elements are being convolved through the fourth dimension. There they go, and they follow rigorously the geometricians’ specifications of that. Now you can’t do that of course with animators or something of that sort. But the machine with its beam knows how to convolve through that dimension.
And I think perhaps the freedom, the option of the individual to find ways to simplify the intricacies of science and technology, of data and measurements through the use of knowledge machines and communications, stands out as one of our great gains in this Century. So we can believe with the poet, that
“Not in vain the distance beckons. “
‘‘Forward, forward let us range.”
‘‘Let the great world spin forever down the ringing grooves of change.’’
So, altogether we have tried to show not that computers can do what you already somehow get done, but that they may do things that lie beyond the present exercise of knowledge, and that indeed they do extend the mind’s quest when they are, through communications and electronics, seen by the mind’s eye.
THE ACTION IS WITH PEOPLE AND KNOWLEDGE MACHINES
W. O. Baker
Bell Telephone Laboratories, Incorporated Murray Hill, New Jersey
A new and delicate balance between man’s mind and body, his intellectual and material systems, is being found in this Century. This balance is the humanizing quality of mankind. Science has so fortified our material world, and especially in the freedom of the Western nations, that now the spiritual and mentally humane qualities cry for emphasis that cannot be and must not be put aside. But universities and students and faculties long ago have built a spiritual and intellectual concept for humanity, having a vast potential for joining with the Age of Science, not in Lord Snow’s two cultures, but in one unity of living creation.
Especially in this University, with its distinguished tradition of spiritual study, we could expect a particular interest in what Huxley has called the Evolution of the Mind, in Teilhard de Chardin’s noosphere, so that some new capacities for learning and acting can arise to balance the immense power of natural science. Other speakers in this series will describe examples of what this power can mean in the preservation and regulation of biological life. Already it has also given us, if we use it right, the assurance of freedom and national security which must come first (as any Czechoslovakian can tell), before any gains in social justice and the spiritual liberty of the person can go forward.
But I shall report to you another phase of the joining of science and society, which is the most direct promise of a new level of human understanding and expression since the creation of the written word. This is the ability to use digital computers, with their fabulous speeds of logical processing of knowledge, as adjuncts to the mind and spirit. Whatever happens to people and societies, it will be nice also to know what it was as well as what it could be. For the first time since writing was evolved, a new way to accumulate and to use knowledge is appearing. For science, with its endless conveniences and resources through technology, has cost society the price represented in the complexity which seems to threaten and dehumanize the individual. Now we are beginning to see that this need not be so, and that knowledge can be handled at its various levels, so that there will be time and understanding to achieve some of the compassion for mankind besought, but often only vainly groped for, in the noble religious ethics of history.
But is it not bold to think that man may be more good, just because some scientific machines may make him more knowing? Well, here experience seems to be on our side, for wise men and learned men, not always, but most often, have been good, and I hope you will decide when we conclude that this is what can come from the action of people with knowledge machines.
With these machines and their associated electronic communications, we are going to be able to carry forward the world’s revolution of expectations in a way that is organized for the common good, but polarized to the sanctity of the person. This is what the collectivist’s revolutions have so cruelly failed to do, but what, in the Free World, given time and patience and work, we can achieve. Of course, digital machines are only one of the resources which are essential, but I thought you would like to know that they are one; they are here, and we understand them, just as we do much of energy generation and antibiotics and motor vehicles and modern agriculture end rockets and moonships. But, in addition, they couple directly with the human mind and spirit, and we shall be writing new Essays on Human Understanding because of them. So we can take heart and have joy that, in the long upward toil of civilization, these new things have come from our society, from our education, indeed from our nation, that will enlighten the generations to come and give us great and worthy things to do. Our education is not an empty rerun of things past or a hollow formalism on the upper crust of society. Computers are the case of action that hits right at the compelling issues of disadvantage and underprivileged and civil rights and free societies. They are no magic solution to everything, but they give near magic aids to many things. They demonstrate that the strengthening and growth of man’s mind has in no way gone out of the world, and that learning and knowing how to use learning are the action.
 A host of examples of this flood of information occur to each of us. Photo copying has mounted by 15% a year for the last five years and reaches now over 12 billion copies annually. About 6,000 commercial radio and television broadcast stations operate in the United States, along with 1,700 daily newspapers and journals. Thirty thousand new books appear a year, and about 6 million students in colleges and universities provide a doubling of use of academic libraries every six to eight years. The libraries themselves are doubling in size about every 15 years at present rates, and the 50 million youths in elementary and secondary schools, with about another 25 million in adult education, will surely push upward this trend of invention, of concept formation with relief from a suffocation of the complexity of detailed information alone.