The fundamental theories of science have been the products of free human minds. The unorthodox mind, asking questions directly opposed to all current thought, has produced our major scientific advances. These require repeated observations to establish their validity, but the fact remains that scientific progress is absolutely dependent on freedom of thought. And a vital part of this freedom is the right of a scientist to communicate with his peers.
In America the vital nature of this right to science is well publicized, but apparently not well understood. The developing picture of science in the U.S.S.R. seems to indicate that the Russians may be in a better position in this respect than we are. In the popular view they have an “Iron Curtain” around their borders which prevents leaks in security to the outside but permits freedom of communication within. By contrast, our democratic form of government prevents the establishment of such a wall and forces us, if we are to prevent the transmission of information, to build a wall around the individual, thus preventing communication either internally or externally. To the extent such a view is true the Russians hold a distinct advantage. We cannot overcome this advantage by establishing a similar dictatorship, for to do so would be to sacrifice the battle to win the skirmish. But the “Iron Curtain” cannot be idea-tight. Witness the large number of scientific papers which we are receiving from Russia, much of which covers material which would be considered classified in this country, and little of which is being translated into English. (My authority for the italicized statement is a theoretical scientist who reads and speaks Russian.) To me this means that there must be considerable trust placed in the Russian scientist by his government. After all, the scientist produces the ideas for military developments; these cannot be hidden from him to any great extent. Yet the Sputniks seem to indicate that the vital secrets have been very well concealed, facts which tend to show that the Russian scientist plays an active part in security as well as in scientific development. This line of thinking leads one to ask, “Isn’t it possible that our security system is a major crippling factor in the development of our science? Isn’t it possible that- the Russians have achieved a higher degree of scientific interchange by a better solution of the science- security problem?”
Obviously, for the maximum efficiency of our science, a complete freedom of interchange of information is desirable. Equally obvious is the fact that we cannot afford to hand the enemy a military advantage by giving information on weapons, war materials, and methods of their use. Since the latter are almost entirely the products of the physical sciences, we face a fundamental conflict to which there is no perfect solution but to which we must find an optimum solution if we are to survive. In the past we have approached this problem as if it had a perfect solution by almost ignoring the requirements of science and attempting to classify almost everything of the remotest value. The resulting over-classification, I assert, has been a major contributing factor to our lag in satellites and missiles, as well as in other fields.
To approach an optimum solution we need first some understanding of the distinction between basic science and applications, between basic and applied research, while recognizing that the distinction is not clear-cut and that there is a wide area of overlap. The basic researcher takes the point of view that by learning as much as possible about the characteristics of the medium he can efficiently build the best possible equipments to satisfy many different requirements. Conversely, the applied researcher starts with equipment and tries to improve it, or he starts with a known principle and tries to build equipment to exploit it. Thus it may be said that basic research deals primarily with ideas as an end product, applied research with equipment. The “grey” area between arises most frequently when the basic researcher builds a tool for an experiment which becomes a prototype for operational equipment.
This is not an argument for one type of research over the other. Both are required for a balanced team. The point here is that applied research, being concerned with equipment, has a much simpler relationship to the classification problem, for equipment must generally be classified. This article will, then, be primarily concerned with basic research (and the “grey” area) in which areas our classification procedures are unnecessarily hamstringing a vital part of our research effort.
One of the least appreciated characteristics of basic research is the widespread interaction of the various fields of science. Non-technical people have a strong tendency to assume that a chemist can only be a chemist, an atomic physicist only an atomic physicist. While it is perfectly true that an atomic physicist has a specialty, he is required to have a working knowledge in many other fields of physics. As a sample list, he might easily, in a given problem, be required to use information from the fields of:
1. Electricity, magnetism, and electromagnetic theory
2. Electronics
3. Hydrodynamics
4. Mechanics
5. Metallurgy
6. Chemistry
He cannot possibly be skilled in all of these fields, and he must go seeking information. Immediately his task becomes many times more difficult since he must get over the hurdles of classification and “need to know.” And if the source is classified, how is he to locate it? And too many times, in the minds of military personnel responsible for classification, atomic physics is one subject having one classification and “need to know”; metallurgy and chemistry are others. In addition, mathematics is the physicist’s principal tool. But mathematics is no respecter of the divisions of physics or science. Its methods apply across all fields. It is common to find that the mathematical method devised to solve a problem in one field of science is applicable to many other fields. It is also common to find the mathematical solution to a classified problem expressed in pure mathematical symbolism and requiring no classification. It is because of this broad nature of science that the problems of scientific classification are aggravated by the concept of need-to-know. Originally designed to limit the information the enemy could get from a captured soldier, and still valid in this context, it is a concept which has no place in scientific inquiry.
When we carry this idea over to applications and applied research, the picture is less complicated. The required information is becoming more and more concentrated as the development proceeds. As the finished product begins to emerge, the need-to-know becomes a real and workable concept—an undesirable concept even here, from the point of view of distribution of knowledge. But in dealing with operational devices, classification becomes a necessary evil.
Another facet of the problem is the amount of mis-information which is reaching our classified files as valid information. This is primarily the result of our failure to appreciate the system of checks which science has developed to insure integrity and to apply these to classified science. In the classical system of science, the technical journal serves the primary purpose of presenting work for criticism by interested and qualified persons in the field. Papers submitted for publication are checked for accuracy by leading scientists in the field before publication and are open to criticism by any reader. It is this criticism which is a vital part of the growth of a reliable scientific literature and which permits the layman to take the conclusions without his own critical review with a high degree of confidence. This criticism is the foundation of modern confidence in science. Authors are constrained to be exceptionally careful in the preparation of papers and in the preliminary experimental work, to avoid the scathing attacks which they know will result from misinformation and errors. This review system is almost entirely vitiated in the case of our classified articles. Primarily these are not seen by a sufficiently large audience to elicit adequate criticism, nor does the criticism, if made at all, reach a sufficient audience to enhance the reputation of the critic. The result is that our classified files contain an importantly large number of papers which have not passed through the crucible of criticism. The problem is worsened by the human tendency to avoid the hard labor of precise thinking and exacting experimental work without this goad of potential criticism, which leads to some cases, and I believe an important number of cases, of mis-information reaching the classified files. From here it passes to military personnel who cannot be expected to evaluate it fully since their training is in other directions. They tend to take it on faith because of the reputation which the critical method of review has established for science in the mind of the layman, but they do not realize that this piece of paper has not had the review to guarantee its integrity. The consequences of such misunderstanding can be most serious.
Another aspect of the security problem arises from the fact that scientifically unqualified personnel are dictatorially making many, but by no means all, of our classification decisions. To the extent that scientifically unqualified personnel are making classification decisions, there is a real danger that some vital “building blocks” in the scientific structure will be lost in the security system. Each classified item may not be important in itself, but it may be an important part of an important structure. Only a qualified scientist can reliably judge the individual item’s importance and properly evaluate it. Another important effect is the duplication of effort which results from the loss of such “building blocks.” One of the major functions of the open literature is to reduce this duplication to a minimum, a function which is nearly blocked by the classification system. Many scientists maintain that we are not short of scientists and technicians, we are merely using what we have inefficiently.
Many scientists are withdrawing entirely or partially from classified work. After all, publication is the “bread and butter” of a scientist. The phrase, “publish or die,” is heard often at scientific meetings, civilian or service sponsored. A scientist is paid primarily to produce ideas. He has nothing else to sell, and his entire reputation is based primarily on the quantity and quality of his published work. He either publishes or can expect to be discharged in many cases. But the reduced audience makes publication of classified work of far less value to him than publication in the open literature. Lacking adequate review, the classified work cannot have the stature of an article in the open literature. Thus many of our scientists cannot afford to engage full time in classified work; yet if they do not, they cannot get the financial support in the form of services and equipment which they need. A few flatly refuse to do any classified work at all. Others accept it reluctantly because the contract will provide some support for their unclassified work.
The absence of papers in the open literature can reveal vital information. To illustrate the point, imagine yourself in a boiler factory. Your guide shouts in your ear, but you can at best only understand the general drift of what he is trying to say. Yet you could understand easily and clearly when he spoke in a low voice from across the room, back in the office. An acoustician calls this factor the signal-to- noise ratio, and recognizes that this ratio, rather than the loudness of the sound, is the important factor in understanding. What has this to do with security? Examples are barred by security restrictions, but we have in our open scientific literature areas in which the “noise” level is so low that the slightest bit of information carries a tremendous amount of intelligence, and the silence, by itself, fairly shouts, “We are vitally and militarily interested in this particular area of science.” Any good scientist in that field can draw some very shrewd conclusions from the mere silence. On the other hand, when there is a large volume of literature, the enemy is faced with the evaluation of all of it to determine for himself the military value of any one piece of information.
Finally, the fundamental objection to the classification of science: science is not classifiable, particularly if one restricts the definition of science to the measurement of natural phenomena. This definition accurately describes the function of the physical sciences and basic research as discussed earlier. The phenomena cannot be altered; they can only be measured and used. The uses are clearly classifiable, but we need to take a long hard look at the wisdom of classifying the measurements. After all, anyone with the ability and ambition, the money and the tools, can make these measurements. We would commit a major military blunder to assume that the Russians cannot make them. In fact the Sputniks indicate that we have been guilty of just that military error. Many scientists believe that the most significant result of the Sputniks may be to make this point clear to the American people. Then granting that the enemy has just as good access to the secrets of nature as we have, are we not better off to know that he knows than to be in doubt as to whether he knows or not? Particularly if we balance the possible gains to the enemy against the damage that we do to ourselves.
The foregoing statements do not prove the inefficiency of our present classification system. The system itself prevents obtaining or quoting adequate evidence, dividing and isolating the components until an objective study is nearly impossible. There is, however, sufficient evidence, not all presented here by any means, to indicate that the classification system is a drag on our scientific progress and a major factor in the inefficient employment of our available scientific and technical manpower. There is an area here in which a correct decision can result in an immediate and significant improvement in promoting our scientific efficiency.
The elements of this decision are:
1. A thorough understanding of the difference between basic and applied research. Basic research should be completely unclassified: applied research, in general, classified, with decisions required in the “gray” area between, decisions which can only be made from a thorough scientific background.
2. A drastic reduction in the amount of classification, particularly in the fields of basic research, to provide a maximum of review and a minimum of loss of “building blocks.”
3. A drastic reduction in the application of the need-to-know to science to provide a broader review of material which must be classified, with consequent improvement in the incentives and efficiency for scientists.
4. A decision in each case as whether classification hurts us more than the enemy.
5. An evaluation of the possibility that the absence of information may provide military intelligence.