Logo DBNL
On Growth (1974)

On Growth

(1974)–Willem Oltmans–rechtenstatus Auteursrechtelijk beschermd

Vorige Volgende

41. Robert M. Fano

Professor Robert Fano was born in Torino, Italy, in 1917. In 1939 he came to the United States. He received his doctorate in electrical engineering at the Massachusetts Institute of Technology in 1947. In 1956 he was nominated professor of electrical communications, and in 1962 he was appointed to the Ford Chair in the school of engineering at MIT.
Professor Fano has worked and published works in the field of network theory, microwaves, electromagnetism, communications and computers. He has made special studies on the social role of computer-communication systems.

During my tour d'horizon, with various personalities around the world, I met with many opinions about the usefulness of computers. Dr. Edward Teller1 felt them to be extremely dangerous and misleading. Herman Kahn2 stressed that the first eight computers in the United States were under his direction and that he knew enough about computers to consider them untrustworthy. Do you feel computers are the most useful means to tackle problems of future global management, like the Forrester3 team has done?

 

Computers have roles both in making studies about the operation of

society, as well as in the operation of society itself. They are fairly distinct roles. Particularly in the first role computers are excellent instruments for answering questions that are formulated. The real difficulty does not lie in the solution but in the formulation of the question. That's where the arguments about the use of computers in studying growth really lie, in the formulation. The formulation remains entirely in the human domain. The computer will answer what is being asked and not what ought to be asked.

 

Edward Teller said, for instance, a computer cannot tell me who is a friend.

 

Here we arrive at an aspect of computers which is quite important. If we want to tackle problems that are too complex in structure for a human to be able to deal with alone and without help, we must be sure that we make possible a very intimate collaboration between computer and man. On the other hand, we know fully well, as I have stated in all my papers, that the process of problem-formulation and the process of problem-solution are essentially concurrent. One never knows whether one has formulated a problem correctly without exploring some of the consequences of the formulation, which really means exploring the solution of the problem. The process involves, first of all, formulation, observing what is the nature of the solution, then one checks whether something has been left out of the formulation, reviews human aspects that have not been taken into account, for instance, restrictions or limitations that have not been put in the formulation. Then one reformulates, solves again, sees what comes out and keeps formulating, solving, formulating, solving until one obtains a formulation that, as far as the person knows, appears to be correct. One no longer sees any consequences of the formulation that are in conflict with what instinctively one knows to be true. The ability of interacting very closely with computers, the ability to mix what I may call ‘human information processing’ with ‘computer information processing’ is extremely important.

 

Could the computer - what Marshall McLuhan calls ‘the ground of the brain’4 - be viewed as an extension of the brain?

 

I'm always wary of this terminology, but I do believe that one can in a sense call the computer an intelligence amplifier, in the same sense that we can think of some of the power tools that we have today, an

electric drill or a washing machine, which are muscle amplifiers, as devices that increase, augment power. Not only power, but also the precision and skill of our muscles. It's very true. One can do with a power tool precision work that just the unaided hand could not do. Well, in the same sense computers can be regarded as having the power to augment the power, the precision, the skill of our mind. That's, indeed, the way I look at it.

But I think there is another aspect of computers from a social point of view, which is even more important, namely, to facilitate a meaningful intellectual communication between people. That will be in the future the greatest importance of computers in society. If we use computers in the communication process - we think of communication as telephone, television - all that these devices do is to eliminate distance between people. One can talk on the telephone over thousands of miles as if it were next door, but they help in no way to relate the intellectual process of communication. How can I really convey better what's in my mind to your mind. They don't help in that respect, and that's where we need help.

 

You have written that a book contains knowledge which does not become available to a person until he reads the damned book. Could computers be brought in to speed the process of reading and learning?5

 

There are two aspects to that question. First of all, is there any way of storing knowledge in directly usable form, i.e., a person is able to utilize the knowledge without himself absorbing it, and practicing the intellectual skills involved. The answer seems to be a preliminary yes. There exist programs today that store knowledge in a way that people can utilize directly. For instance, the particular example of work done here at MIT by some colleagues of mine, particularly Professor Moses and Professor Martin. They have created a computer system that has the knowledge of an extremely competent mathematician. As a matter of fact, I heard one of them claim that he cannot think of one mathematician who has all the knowledge that is in that system.

I think that the answer is yes. It seems to be possible to store knowledge in computer programs in directly usable form, that's one aspect. But I was talking about the process of facilitating communication between people. Let me give you an example of what is involved. Two people are interested, generally speaking, in the same sort of thing; but, as often is the case, one point of view is different from the other. If I am

a good teacher, I try to understand your point of view and explain it from your point of view. That's the skill of a teacher. But it is not very easy for me to know exactly your point of view. If you are going to really understand what I say, a translation process has to take place. That is, you have to translate this knowledge into your point of view, or in other words fit it into your frame of reference, fit it and tie it to the rest of the knowledge that is in your mind. This process of translation is extremely difficult and complex, and this is - in my view at least - one of the barriers to human communication. People talk with one another but don't communicate. I think that computers eventually will be able to help people in this respect.

Let me give you some trivial examples of this: Suppose that in describing the computer - let's say some object - I think of that object - let's say by looking at it from the front - that's my interest. (This is of course a trivial example.) Your interest is to look at it from the back. Now, a computer can very easily turn it around and change the display. I describe it from one point of view; the computer translates the point of view - literally speaking, the direction from which you look at it - and you see it from the other end. In more general terms you can think of a computer as having the knowledge stored, perhaps acquired from a person, but then be able to present it to another person in a different way, as the other person requests. In effect the computer may converse with me and absorb knowledge from me. The knowledge is stored. Another person converses with the computer and extracts the knowledge in a form and from a point of view that suits him. This is for the future, but if one looks at what is being done today, we will be able to do that in the future. One begins to see examples of that sort already appearing. Small examples, trivial, but we are on the way. That's one form of communication.

Another form of communication is what I call communication in the presence of data. When we sit down together to discuss something, we may need pages and pages of papers, of data. We are forced to look at it together. This is awkward. Computers can facilitate this a great deal. They can have the information stored, and we can interact with one another and with this information through a computer. Consider a meeting - this experiment in fact has been carried out - like a committee meeting in which each person sits in front of a computer console, computer display. The communication between them takes place through the computer with the result, for instance, that one hears or sees somebody talking about something. It involves information that perhaps we talked about

before or which is background information. I want to see it. While the other person is talking, I ask my computer and there it appears, the details of what the other is saying appears for me selectively. The others are not bothered unless they want to. There is a wealth of information that is put into the conversation this way.

 

- On a kind of TV screen?

 

That is right. This screen is at the fingertip of each individual as an individual, and he does not need to bother other people. One just finds out what one wants. Someone else finds out something else.

 

Professor Delgado of Yale University speaks in his interview about the transmission of emotions by computers from one part of the brain to a certain part of another person's brain;6 now, there is a very wide new range of computer utility.

 

 

I am reluctant to comment upon it without knowing what he means about emotions.

 

The computers that were sending the Apollo 17 up in the air gave some trouble, and scientists spoke of computer hypochondria. Technicians had to adjust these computers like psychiatrists, to prevent a further delay. Do we need now computer psychiatry?

 

Those words are mysterious and out of the world. Let me debunk them a bit. The story is this. We are characterizing human behavior with certain expressions, and really we are faced with an extremely complex organism. We are observing certain behavior characteristics. We have no idea how they come about. We ask ourselves what are the mechanisms internally that result into that external behavior? We characterize this with expressions such as the one that you've used. Computer systems are becoming so complex that it is often difficult to understand where certain unexpected behaviors result from. Computer behavior? Yes. There are tendencies to characterize general patterns of behavior in an anthropomorphic way. It's natural and not surprising that a certain behavior of the computer could be associated with certain words.

For instance, I remember years ago people were asking me whether a computer could show preference for certain people. In a trivial sort of way, yes, you could say that you can apple-polish a computer, just

like the old story about apple-polishing a teacher. All it boils down to is this: If you know how the computer behaves, you can fit your behavior to suit the computer, in which case the computer will preferentially do what you want.

 

Let's assume the Chinese will be capable of designing computers within the next decade. Then of course we would need a new science, because Chinese computers would be reacting to Chinese behavior. Would Chinese computers be able to talk to American computers or European computers? Next we will see the consultation of behaviorists versed in Oriental behaviorism to adjust our computers.7

 

No, no, let's not go into science fiction. On the other hand, there may be a grain of truth in what you say. Since computers are designed to interact with people, they're fitted into a certain culture. It isn't that Chinese computers will be very different from, let's say, American computers, but rather that it will be fitted in a certain respect to the Chinese culture rather than another - in the same sense furniture does. The furniture you will find in a Japanese home is different from what you find in an American home, since Japanese have a different culture, different habits of life, and the furniture is designed accordingly.

 

Yes, but human emotions interact entirely differently even between Japanese or Indonesians - let alone between Japanese and French. If human emotions, human behavior, interact, come out in these computers, you are going to get a different kind of interaction between man and the computers in the Far East, for instance, as compared to in the West.

 

It could be. But let me talk about this question in a more general sense to characterize it. The fact is that there is a rather strong coupling between computer systems and the community of people that is affected by their use. By strong coupling I mean that the characteristics, the initial characteristics, of a computer system have effects on the community around it, possibly social effects. The characteristics of the community around them influence the evolution of a computer system. Let me make clear that a computer system is not a static thing. New programs are written. Things are reorganized. It evolves in time just like a community evolves in time. The two evolutions influence each other very strongly. There is evidence to that effect. In those terms you can see social phenomena being very closely related to the characteristics of computer

systems and vice versa, i.e., you can see initial characteristics of a computer system that may have very strong influences on the evolution of the community around them, including relationships between people and even affecting the values of the community.

 

Would there not be a danger then that some evil power would get hold of the programming of the computers and therefore program the community according to possible vicious ideologies? We get into Skinnerian philosophy of programming the environment in order to program the individual.

 

I am not particularly worried about that. It's a kind of a detail, in a sense. What I am speaking about is something much broader and in a sense much more dangerous than that. Basically the way you would utilize computers in the operation of society depends also on the structure of the computers themselves, what is available. For instance, to draw things to the extreme, there are computer systems that do not interact with the users directly, and there are the computer systems that can fairly easily interact with individuals. Depending on which type of computer system you use in a particular operation of society, you force the community to operate in a different way. An example has to do with the seat of control. There are certain computers that force a mode of operation - or I would say not force, but induce - further and further centralization of control. Another way of utilizing computers is the easing of decentralization of control, a computer that maintains an overall coordinating activity. This can have a big influence on the entire behavior of the community. There are subtle things like computers that can protect individual privacy or that won't.

 

Do you feel that computers in the immediate future will assist man by accepting the fact that the planet is finite and that we have to come to a form of managing this planet?

 

There is no question in my mind that computers can be very helpful in that respect. I would say even more. I am seriously afraid that advanced societies such as the society in the United States, may fall under their own weight of complexity without the help of computers, which is going to be an essential tool for survival of a complex society. It is not just a question of planning. It is a question even of operation. Because what you can plan depends on what you can actually implement.

When one talks about managing the globe as a whole, one does not talk only about planning but implementing workable solutions. These workable solutions imply much more collaboration in close interaction between all parts of the globe, that I cannot expect to be feasible except through extreme utilization of computers. This is why I was talking about facilitating human communication. It is not enough to make a plan. One has to be able to execute it. The execution requires tight interaction between people about many different matters. People need better tools to cope with this greater complexity, otherwise it's going to fall apart.

 

To come down to earth for a moment. I noticed outside your office some signs and recommendations. Universities are asking for computer experts. The Rand Corporation8 asks for them. Yale University is running a Ph.D. in computer technology. Tell me in a few words how are young Americans reacting to the science of computers? Are you hopeful that you can educate a generation that can tackle problems?

 

There is no question that the interest among students in what is called computer science and engineering is growing very fast.

 

Both boys and girls?

 

As a matter of fact I regard computer science as a branch of engineering. I would say that is the particular branch in which substantial interest from girls keeps growing.


Vorige Volgende