1. It is too highly centralized. Detailed decisions get made at too high a level. Detailed budget control is exercised at too high a level. There is too little delegation to the contractor and to the official on the spot.

2. Precise paths with far too little flexibility—are dictated by administrators to researchers.

3. There is too little duplication, especially at the level of exploratory development, where it is cheap.

4. There is not enough competition among laboratories and contractors, or even among the services. There are not enough competitive purchasers of new ideas, or enough competition in exploiting them. Where there is competition among services or agencies, the record shows that we tend to make rapid progress. Where we suppress it, trying to decide the optimal path and contractor in advance, we do not do well.

But almost all the proposals for improving military research and development seem to be aimed at strengthening central planning and coordination, adding new layers of authority, and getting rid of what little desirable duplication and competition we have left. They are justified on the grounds that they will eliminate duplication and competition. They run the risk, I fear, of creating neat, dead, bureaucratic monopolies.

That some of the strengths of the competitive economy can be emulated by governmental research organizations has been demonstrated time and again. Two of the best examples, ironically, are the Manhattan District Project and the wartime Office of Scientific Research and Development— both hailed as ideals by some contemporary centralizers. The MDP typically solved its problems by letting each of half a dozen different firms or university laboratories try a different path to the objective, competing vigorously and in secrecy from each other. The wartime OSRD was perhaps the most loosely knit, decentralized, uncoordinated species of research organization that ever existed-and spectacularly successful.

If we would let it, our competitive economy could teach us that in all research and development, including that sponsored by the military services:

1. The quickest-and frequently the cheapest-way to achieve many research objectives is to try multiple paths-some of which, in retrospect, are inevitably cul-de-sacs and apparently wasteful.

2. The greater the payoff and the greater the uncertainty, the more duplication and apparent waste make sense.

3. The cheaper the multiple paths, the more it makes sense to try. More should be tried in research than in development; more in exploratory development than in weapons systems development; more in weapons systems development than will eventually be procured and made operational.

4. The most expert predictions of the results of research and development are highly unreliable. Frequently the only way to learn whether an idea will work is to test it.

5. The person best qualified to choose a path (as opposed to an objective) is the person doing the research-or someone very close to him.

6. There is no incentive quite as effective as the competitive spur. The problem in managing governmental research and development is not how to suppress competition, but how to divert it into more productive channels.

Discussion

M. H. TRYTTEN

Director, Office of Scientific Personnel

National Research Council, National Academy of Sciences

Chairman Neal: You have just heard three most distinguished papers. You are now going to hear commentaries on certain aspects of these papers from three distinguished gentlemen. In place of J. Douglas Brown on your program, we have as our first discussant, Dr. M. L. Trytten, Director of the Office of Scientific Personnel, National Research Council, National Academy of Sciences.

IN

́N THINKING over the papers, a brilliant succession of papers, there are two things that occur to me of major interest that I want to add. Both of them relate to the future of research, particularly industrial research in this country: 1) what will be the eventual level of research and development that we must look forward to, or what will be the rate of increase of industrial research and research in general, and 2) what will this level and rate of increase mean in the way of training of personnel.

We have heard this discussed largely, I think, from the economic point of view. I would like to comment, however, on the noneconomic forces that are working toward this increase in volume and importance of research. I might cite Dr. Slichter's paper, which also took these other motivations into consideration. I think the future levels of research and development activity will be only partly determined by economic considerations, as is the case at present. This is in spite of the fact that there seems every evidence that industry in our country is becoming very much research minded. Although Dr. Slichter found that the returns on investments in research are very good, there are a number of other factors one needs to realize that point to a higher level of research activity as an essential necessity in our country. These factors to a large extent arise out of the very nature of research and development itself, as well as out of the results of research and development in our economy, where, because of the fact that our technology has reached the stage where it is, it is continuously opening up new areas in which new problems arise. We are in the eras of new speeds, new

pressures, new temperatures, new altitudes, and many expanding boundary conditions. Boundaries or limits of activity heretofore are being exceeded. This always introduces new dimensions to be explored. These dimensions are increasing in geometric ratio and for that reason will involve accelerated activity in these many lines in the near future.

This is not a new observation at all for us. It has been made by many people, and it has been made over a rather long period of time. But I think it is becoming a little more dramatically evident at the present time and has been noted in quite a number of places. The most recent one I think of is the document put out by Dr. Harrison Brown and his associates at California Institute of Technology, in which they look forward over several decades to try to get a perspective on what is going to happen in research and development and, more particularly, in the resources of our country.

I want to mention particularly a point that Dr. Slichter made with respect to the development of technology as a new frontier to replace the geographic frontier. You will recall that he raised the question of whether there was a cause and effect relationship here or whether this technological growth might have been an accident of history.

I think, as a matter of fact, that it is an accident of history, and for a number of reasons. And I think this idea is worth commenting on because it bears directly on the conclusion to which I want to call attention.

In the first place, the whole development of technology in the world, and particularly in the United States, has fitted into a time schedule which does not admit of much change. That is, technological development rises out of the circumstances and out of context of past scientific developments. It is, after all, only about 300 years ago since the major framework of the physical sciences was laid down in the time of Newton, Galileo and some of the others. It took a couple of hundred years or more following that before any very great substantial development of applied science was possible. And most of that applied work happened in the past century, that is, with the development of the dynamo and the motor, the development of the great technological industries in the latter part of the last century, and ultimately with the development at about the turn of the century of the consciously organized research laboratory, notably, those like General Electric and the Bell Telephone Laboratories. All of this was a fairly steady growth, each part depending on what had gone before in a time schedule which grew out of the preceding conditions and discoveries.

Consequently, it was the twentieth century before the massive movement of today was possible. Now, I would like to call attention to the other side of the picture, that is the training of personnel for these scientific and technological activities. Here, too, there is a very definite historical pattern of development, set by the growth of schools and colleges in the United States. In this develop