The trend of military evolution on which the world has been set since 1939 has seen the harnessing of the most advanced technology to the elaboration of an extensive series of new weapons and weapons systems. Through the development of a large family of ballistic and non-ballistic missiles, destructive power can now be brought to bear on unseen targets over distances ranging from tens to thousands of miles; at the same time, the capacity of a single explosion to destroy has been multiplied thousands, even millions, of times. These technological developments have been associated with the emergence of new scientific processes of management (most of which are loosely grouped under the term "operations analysis"), whose purpose is to try to help the commander control the apparatus of which he now disposes. But a question which becomes increasingly urgent in our age of nuclear deterrence, and one which grows in importance as more and more technology becomes harnessed to the demands of defense, is whether these new measures of control extend or curtail the possibilities of human, as opposed to machine, judgment.

A useful approach to an answer to the question is to consider three general propositions, of which one is observable fact and the other two are unassailable deductions from experience.

The observable fact is that the amount of military input into modern weapons systems, and particularly complex strategic systems, is declining rapidly, with a complementary increase in the technical input provided by the non-military man. This change is associated with the increasing specialization of single-purpose weapons systems. By "military input" one means, of course, the fruits of actual military experience. The simplest illustration of this proposition is that no military genius or experience has gone into the conception or design of I.C.B.M.s. If one wishes to push it that far, there is no logical need for such a weapon to be deployed by the military, as opposed to some other agent of government. If the name Moscow, or New York, or London, or Paris were written on each I.C.B.M., the missiles might well be deployed and operated by the firms which produced them. The complex operations of the U.S. National Aeronautics and Space Administration (NASA) are not military operations, nor is putting a man into space a military operation, even though the men who go into space may all be military men.

Author's Note. This article is based upon a contribution to a symposium on "Science and Warfare in the 1970s" held at Supreme Allied Headquarters Europe on May 23, 1961, under the joint direction of Admiral of the Fleet Earl Mountbatten of Burma and the late Dr. Howard P. Robertson and under the chairmanship of General Lauris Norstad, U.S.A.F., Supreme Commander Allied Forces of Europe, with whose authority it is published.

Sophisticated anti-aircraft missile systems are equally automatic in their conception-and, one suspects, in their potential operation. For them to operate properly, one needs virtually complete information about what is happening in one's air space, not to mention infallible and unshakeable judgment about what to do with that information.

My first major generalization, therefore, is that the more modern technology one puts into weapons systems, and the more automated they become, the less they constitute the fruits of military thinking, and the less flexible they become in use.

My second proposition is this: However much importance one attaches to firepower, battles and wars are not necessarily won by matching unit powers of destruction, or by having a few more potential units of destruction than one's enemy. The best victory is the one where an adversary surrenders without a shot being fired; at the other extreme, the worst victory is one which is associated with a maximum of unnecessary and expensive destruction. War, of course, is a tricky game, and rather than take risks, commanders almost inevitably over-insure. Once there is no scarcity of bombs or aircraft, the tendency is to use more rather than fewer on a given task. The frequency with which the principle of economy of force has been violated in the past makes one wonder in what sense it constitutes a principle of real military action as opposed to something to talk about at staff colleges.

Space prevents my illustrating the proposition that neither battles nor wars are uniquely determined by matching units of destructive power. For present purposes, I base this assertion on an extensive series of special studies which I personally made of the "natural history" of destruction during the course of the 1939-1945 war. They all illustrate the conclusion that, while firepower is a major factor, it is only one of several vital factors which may determine the outcome of conflict. History, alas, provides no guide as to which factor will determine the outcome of some battle which has not yet been fought.

My third proposition is so obvious that it hardly bears mentioning. It is that the more vast, the more heterogeneous, the more scattered any organization becomes, and the more complicated its component parts-it makes no difference whether we consider the civil or the military sphere-the more difficult it is to control and to concert its multitudinous activities to a single common purpose. The difficulty is an ultimate limitation on the practical uses of good judgment.

II

Let us consider the first proposition-that in the present phase of military evolution military experience as such is contributing less and less to the development of complex weapons systems, and that, correspondingly, we are being committed more and more to systems which, as they grow more complex technologically than anything we could have conceived of before, increasingly limit our freedom of choice and action.

The nature of the military armory has always helped determine the character of tactics and strategy. To that extent, weapons have always set constraints on freedom of choice-which means on the exercise of judgment- when military means have been used to secure political ends. The danger today is that our complicated armory, and particularly our nuclear armory, may in the end drive us to a position in which we are committed to operate as in a set-piece. Such an eventuality would mean that if tactical warfare were ever to break out in Europe, it would simply become one part of a program of predetermined strategic action.

There is an immediate corollary to the overt recognition of this trend in weapons development. If the way in which our potential freedom of military operation and deployment is constrained becomes obvious-and in these expensive technological days these things cannot be kept secret in a democratic society-our potential enemies can adjust their plans, military and non-military, in a complementary way to suit their own convenience.

This trend in the development of tactical warfare as one part of a set- piece of predetermined strategic action may well be something we need as part of our strategic deterrence. But it is not what most people have in mind when they refer to the "proper use" of tactical nuclear weapons in a land battle in NATO Europe. It is essential, therefore, that the armory of which we dispose in order to defend ourselves in field warfare does not contain within itself trip-wires which unwittingly set into operation the whole panoply of strategic deterrence. I know that this matter is very much the concern of General Norstad today. Ten years hence his successor will have to wrestle with an even more severe problem.

Let us be quite clear about this. There are those in the West who already believe that the use of any nuclear weapons in a war involving NATO would mean the opening of a Third World War, There are indications that the Soviet leaders believe the same thing. If this view is right, all freedom of choice may be lost, and with it all possibility of exercising judgment and control in a future war between the great powers.

I am not concerned here either to discuss this view or to paint the picture of mutual suicide which it entails. I refer to it simply as an indication of a major consequence of the development of weapons systems into which one builds only a predetermined set of functions. The guidance and navigational systems of a long-range ballistic missile are magnificent achievements, but they themselves are incapable of any more judgment than is built into them. Only in a very crude way is a nuclear-headed missile of this age a substitute for a manned aircraft. Once launched, all it substitutes for is the bomb the aircraft might have dropped.

The technological panoply of set-piece warfare is the panoply of deterrence. Its weapons are not weapons in the conventional sense. Our deterrent weapons are there to make any potential aggressor realize that if he strikes he will be struck. What would happen after that, in the case of strategic war, is matter for speculation, in which scientists have no need to engage in public. They may better appreciate the significance of the facts of destructive power, but they cannot lay claim to better imaginative powers than the next man.

There is one related point, however, that I must touch on. If one assumes that such a thing as tactical nuclear warfare could occur in isolation, how are armies equipped with nuclear weapons to be controlled? This is an immense problem for the military scientist. Only someone who wishes to blind himself to the obvious facts would fail to see that our machinery for surveillance and for condensing, analyzing and disseminating military Intelligence does not match the force and range of our striking power and the potential speed of reaction of the modern weapons systems which are deployed on the battlefield.

This is an enormously important area for military research and development. Which of several towns, or how many of them, are going to be effaced from the map as interdiction targets 50 or 100 miles from the front? More important, how is one to know what the effects of their elimination are on the movement of the enemy forces? Once our own nuclear weapons start detonating, how is the news of what is happening to be flashed from one to another part of our own lines? How are we to inform ourselves of what the enemy has done with his nuclear weapons? How are all these data to be collected? And, every bit as important, how are they to be reduced quickly enough to the kind of intelligible terms in which they could be disseminated and without which neither judgment nor control can be exercised?

III

I turn now to my second proposition, which is simply that battles and wars are not necessarily won by the side with the greater number of units of destruction at its disposal. A corollary of this proposition is that the way our power to destroy has multiplied over the past 15 years may itself impose a major limitation on military freedom of action in field warfare.

I do not propose to set out figures which define the nature of modern destructive power. The yield of the two bombs which destroyed Hiroshima and Nagasaki-two moderate-sized industrialized towns-was between 15 and 20 kilotons. The NATO armory contains weapons of this caliber, and also many others with higher and lower yields. The use of any one of these- and of any equivalent nuclear weapons the Russians might use -would mean the instantaneous and total destruction or elimination of an area varying in size between, say, a large village and a large town. Depending on the way forces are disposed, the explosion of a nuclear weapon of, say, 100 kilotons could mean the elimination of one or of dozens of battle groups or of one or of several squadrons of aircraft on the ground. But if one's target information happened to be wrong, it could mean the elimination of none.

The dispersal of troops, which is a tactical consequence of the power of nuclear weapons, is one of the basic reasons why we need a better apparatus than we now possess for the reconnaissance and surveillance of potentially nuclear battlefields, and why we also need better communications. It is a reason for increased mobility of troops, since this is as much a requirement of a nuclear battle-assuming such a thing is a reality-as of a conventional battle.

There is also the problem of radio-biological hazards, which again focuses attention sharply on the need for highly sophisticated intelligence and communications systems, if any measure of control is to be exercised under conditions of chaos.

When I use the term "chaos," I am thinking not only of the direct effects of radiation and blast resulting from attacks on purely military targets, but also of the likely effects on military operations of nuclear explosions on adjacent civilian populations. For, to the best of my knowledge, there is no area in Western Europe where a nuclear battle could be fought without causing considerable damage to non-military targets.

This point can be illustrated by what is a representative scenario of such a battle, but one smaller and more confined than most that have been considered. In a war game involving just three NATO Corps, nuclear weapons were "used" against military targets only, in an area of 10,000 square miles which contained no large towns or cities. In this "battle," lasting only a few days, it was assumed that the two sides together used a total of between 20 and 25 megatons in not fewer than 500 and not more than 1,000 strikes. It turned out that 3½ million people would have had their homes destroyed if the weapons were air burst, and 1½ million if ground burst. In the former case, at least half the people concerned would have been fatally or seriously injured. In the case of ground burst weapons, all 1½ million would have been exposed to a lethal radiological hazard and a further 5 million to serious danger from radiation.

I am, as I have said, referring only to the actual battle area in which the three Corps were engaged. If we wish to obtain a more realistic impression of the circumstances which would actually prevail in such a limited battle, we need to bear in mind that no such operation, however confined, would be likely to occur without collateral, so-called interdiction attacks with nuclear weapons beyond the area of local engagement. Also, in all probability both sides would launch nuclear attacks on distant airfields or missile sites.

My purpose in trying to give an indication of some of the likely facts of a nuclear land battle-and I have described only part of the picture-is both to answer questions that have been put and to indicate that problems of judgment and control become increasingly difficult the more one's weapons automatically lead to over-hitting-even of strictly military targets.

In the ideal, the best weapon is the one just big enough to destroy the specific target against which it is directed. This is true whether looked upon from the standpoint of military practicality or plain economics. For example, both theoretical and field studies of World War II battles have shown that the smaller the units into which a given weight of anti- personnel weapons was divided, the greater the number of casualties they caused.

But nuclear weapons have changed all this. It will require the utmost judgment and control if the secondary, non-military effects-even when they are directed against assumed military targets-are not to dominate the situation once they are used, and having done so, to generate a new situation which is outside all possible military control, as it certainly now is outside all military experience. I am not suggesting that the problem of trying to operate rationally within an environment of chaos is new to military experience and command, or that civilian populations have not suffered from military operations in the past. What I am saying is that the problem today assumes such new proportions that it moves into a dimension completely different from the one in which we have all gained our experience.

There is, of course, no rule which says that a land battle in Europe would immediately "go nuclear." Equally, there is none which says that it would not. My experience of the way commanders worked in the Second World War does not lead me to suppose that, if unlimited force were available, less rather than more would be used in order to secure some objective, whether on the defensive or offensive.

The very existence of tactical nuclear weapons is thus the most urgent challenge that has ever been presented to military judgment and control. As weapons to deter aggression, they serve a very precise purpose; the context of field warfare in which they might actually be used is an entirely different matter. There can be very few military targets which are not disproportionately small in relation to the area of effect of the smallest nuclear weapon that might be used against them. Once one goes beyond, say, a yield of 20 kilotons, one is in the "town-elimination range." Because of this, one cannot make any valid comparisons between the introduction to the military scene of nuclear weapons and, say, the introduction of conventional explosives or chemical warfare at the time they emerged.

One further point. In the development of nuclear warheads, particularly those which are designed for missiles, there is a tendency to add to their yield in order to compensate for the inaccuracy of the delivery system and for inadequacies of target information. To a lesser extent, this is also true of free-falling and guided bombs. This, again, means that one builds into nuclear weapons greater destructive power than is necessary for military purposes, and that their secondary, non-military effects overshadow those which relate specifically to their military use.

IV

I turn now briefly to the third consideration which I see circumscribing the area within which military judgment and control can be exercised in this rapidly expanding technological age. It is simply that the larger, the more heterogeneous, the more complex, the more scattered any organization becomes, the more difficult it is to control its actions to fulfill a common purpose.

With the active concurrence of the military, we have moved into an era of weapons systems rather than isolated weapons, and this trend is unlikely to be interrupted. Each new system emerges as a combination of complex components; and each of these needs to be monitored and manned in an increasingly specialized way. Early warning systems, missile systems, certain forms of intelligence systems-they are all similar in that each cog is vital, each has to be complete and perfect within itself, each has to have its built-in programs, computers and comparators. Each sub-system has to synchronize precisely with the next one. Increasing specialization of systems inevitably leads to rigidity of operation-and to the danger of breakdown.

Similarly, the men who operate and service the components of the newer weapons systems, whether they be early warning radar chains or supersonic high-altitude aircraft or Polaris submarines, are working at the extremes of their physical, psychological and mental capacities. As the decade progresses, each new weapons system will place still greater demands on the judgment and control of the individual.

Today, command has to be exercised over varieties of weapons systems which interlock over vast areas and, in the case of NATO, over systems which cross national frontiers. If the ultimate limitations of human judgment and control are not to be exceeded, it is vital that every effort be made to unify and simplify those human procedures which are concerned with the collection, interpretation and dissemination of intelligence.

In examining the impact of future scientific developments on military affairs, there are two more points which I believe to be highly important. The first is that, while the sheer pressures created by technological progress themselves generate limitations to the proper and free exercise of judgment and control, the problem of maintaining this essential quality has not been made any easier by the way the military have altered the concrete nature of certain military problems by turning them into abstractions. One dangerous example of this tendency is the term "interdiction targets." Another is the idea that nuclear weapons are just a new and more powerful form of artillery, and that one exchanges "nuclear fire" like counter- battery fire. A third is that one is able "to restore a situation" with nuclear fire. There are other conventional military terms which relate properly to the pattern of warfare in pre-atomic days but which hardly fit our present technological era.

We think we know what we mean when we use such terms as "interdiction." To take a special case, let us assume a series of bridges over a river; they must be destroyed lest they be used by the enemy, even though, as is so often the case, they are in centers of population. The nodal points in a railway communications network might be another set of interdiction targets. As such, they would be dealt with as part of a set-piece offensive.

In this kind of thinking, however, the target is not a body of troops or a particular bridge. It has become an abstraction- "interdiction." The purpose of interdicting-in this case, the prevention of the movement of a particular body of men-may not be fulfilled at all in "set-piece interdiction." For all we know, the men may not have been there, or may have planned to move by another route. This is how it often was in the Second World War with one major class of interdiction targets in Northwestern Europe; and I know nothing which suggests to me that it would not be the same in the next. The use of nuclear weapons to carry out interdiction plans may or may not impede the enemy, but it is very likely to make the environment more difficult for our own commanders to exercise judgment and control.

It seems to me, therefore, that when we talk about the potential use and effects of nuclear weapons, we must avoid the conceptual framework derived from the military terminology of pre-nuclear warfare. One may fairly ask what meaning there is to the idea of using nuclear weapons "to defend our territories and peoples." One can deter with nuclear weapons. Can one defend?

V

My final point derives from the fact that the situations encountered in warfare are as empirical and as experimental as those thrown up in a laboratory, and that they do not lend themselves any better to predetermined judgments. Experience in dealing with experimental situations helps one in dealing with new ones; but the results of one set of experiments do not apply to another-unless they are identical, which is seldom the case.

Warfare, the behavior of men and nations in conflict, is a far more complex thing than the behavior of men and women in normal peaceful groupings. Experience has shown that even the latter is hardly amenable to the discipline of scientific method. History, we know, was made yesterday; and social science has not yet provided the predictive generalizations which would allow us to write tomorrow's history today.

Discussion of this point requires a consideration of military operational analysis and how it was born-or perhaps I should say, was allowed to be born-in the Second World War. It emerged as a new procedure in military affairs, as an intellectual tool for the military, so as to help the proper exercise of judgment and control. Without it, warfare was already becoming too difficult technically for those in whose hands its conduct rested. I say this with no disrespect. In the American Civil War, in the Franco- Prussian War, in the First World War, war-in the words of Colonel Nathan Forrest-was still largely a matter of getting there "fustest with the mostest." Firepower had increased over the years both in range and in the degree of physical destruction that could be caused. The tank had increased the possibilities of mobility in the face of enemy fire. Chemical warfare, in spite of its limitations, was a clear-cut and understandable procedure that nearly broke the static conditions of trench warfare. In general, the strategists and tacticians of the First World War were still operating with weapons they could understand over ranges they could, in effect, see, and in accordance with the hallowed military principle of the economy of force.

With one minor qualification, the First World War was thus historically still a period of balanced military forces and balanced weapons systems. The balance lay essentially in the control the commander could exercise over his troops and weapons, the consequences of whose use were also within his comprehension, and potentially, therefore, within his control. I can see no other possible useful meaning to the term "balanced military force."

My one qualification to this generalization was the emergence of air warfare. This was the germ which destroyed the traditional frame in which the military commander had hitherto exercised his judgment and control. Out of the first aircraft grew a weapons system whose striking power was to leap well ahead of the capacity to apply destructive force specifically for the elimination of what it is still both useful and wise to call "military targets." The aircraft was thus as much a leap into the dark as into the air.

In comparison, the submarine, when it first appeared, was a weapon whose effects could be envisaged without difficulty. Submarines certainly imposed a new threat, but it was one which was understandable and technically manageable, given a detection system, depth charges and the intelligent deployment of surface shipping and anti-submarine forces. In its traditional sense, it was-I use the word "was" deliberately, for it is different now that submarines have become launching platforms for ballistic missiles-a weapons system which was posed against specific but mobile targets. The aircraft was something quite different. Able to penetrate deeply into enemy territory, it could achieve effects the consequences of which were unpredictable, and certainly outside the framework of the airman's experience.

I do not know how it was in other countries, but we in Britain -to illustrate my point-entered the 1939-1945 war with highly unrealistic estimates of the capacity of aircraft to find their targets, and of the destructive power of the bombs they dropped. Equally, we either underestimated or overestimated the resilience of people under attack, or the economic consequences of our strategic attacks, or the functional significance of different target systems. We had to learn the hard way. We had to learn to treat air operations as an empirical experience, the results of which had to be analyzed with a strict respect for scientific discipline if the next set of plans was not to perpetuate the errors of those preceding it. Because the results of our air plans were not analyzed in this way during the first half or so of the war, at the start they did, in fact, often merely perpetuate errors. Submarine warfare, too- on the scale it was waged during World War II- became a matter for analysis of a kind different from what had gone before. In short, war was becoming extremely complicated technically. Scientific analysis, in consequence, became as powerful an intellectual tool as the more traditional kinds of military intelligence. And sometimes it was far more valuable and reliable.

Operational analysis was thus born out of the need generated by the interaction of a variety of complex problems resulting from the intrusion of modern technology into the military environment. In essence, and at its simplest, it was no different from the methods by which industry had improved its operations over the years. It came late into the military sphere, because the military world until then was a more conventional world, one less amenable to directional change than the continuously competitive world of commerce and trade. Then three varieties of operational research emerged.

The first category can be described as the analysis-observational and experimental-of the technical and military problems relating to the introduction of new weapons and techniques. The advent of radar is the classic example of this category of operational research. If radar was going to work, civilian scientists had to be brought into the military machine. And if the civilians were to discover how to work radar, they had to become operational researchers, or operations analysts, in order to see how the new apparatus could be fitted into a military context. Other examples of this category of operational research can be found in the field of electronic counter-measures and in the logistics of supply.

The second category of operational research dealt with the tactical consequences of the interaction of new and complex weapons systems and technologies. It concerned itself with the scientific appraisal of new situations in warfare resulting from the threat imposed by these weapons systems. Anti-submarine warfare is the classic example of this kind of operational analysis.

The third category of operational research consisted in the analysis of the way military plans-both strategic and tactical- unfolded in actuality, as opposed to how they were supposed to work. It was in this kind of operational analysis that I was particularly interested, and which I had the opportunity to develop as a result of my association first with Admiral Mountbatten, at that time Chief of Combined Operations in the United Kingdom, later with General Norstad, then on the staff of General Spaatz in the Mediterranean Area, and then with Air Chief Marshal Tedder in northwestern Europe.

No opportunity of analyzing operations in relation to their stated purpose was afforded us until early 1943, when the Axis forces were driven out of the Western Desert and North Africa. As the British armies advanced, studies were made of Tripoli and of other targets of air operations. The bomb damage as seen on the ground, and as recorded by the civil authorities at the time, was compared with the operational planning of, and intelligence reports on, the attacks which had caused the damage.

Essentially, what was being done was to treat each operation as one might an experiment of a very crude kind. How closely did what was achieved correspond to what we had set out to do? Why were intention and effect not always the same? A much more direct experimental approach was followed in planning and in executing the operations which led to the capture of the island of Pantellaria in 1943 and in analyzing them afterward-work in which I was closely associated with General Norstad. The lessons learned from this operation were then applied in 1944 to the pre-D-Day offensive against the coastal defenses of northern France. Similarly, the lessons learned from a detailed analysis of the functional effects of attacks on the railway system of Sicily and southern Italy provided the basis for the 1944 plan to destroy the communications network of northwestern Europe. Subsequent analysis fully confirmed all the expectations that had been based on the experience gained in the Mediterranean. However, there was little opportunity for this kind of analysis of field warfare-largely, I think, because it moved too fast for the results, when they became available, to be applied; also, the situations in field warfare were infinitely more varied than in either air or sea war.

I am emphasizing this aspect of operational research because its purpose is also that of the new school of operational analysis which is trying to achieve its objectives by theoretical and abstract reasoning, aided by recourse to various mathematical treatments of the concept of probability.

But before I discuss this new aspect of modern operational analysis, let me refer to another and simpler postwar variant of operational research as it applies in the military field. This is the kind of systems operations analysis that is carried out before one starts developing a weapons system for some definable purpose. It considers the operation of component sub- systems not only within themselves but also in their relation to other sub- systems. For example, a sophisticated air-defense system requires ground radar and associated data-handling systems. It needs command links. And these in turn have to be tied in with the active or passive elements which are deployed to deal with the actual attack.

Obviously, the only thing that matters is: "Can the whole system deal with the attack?" If the strike component of the defense is too slow, it does not matter how efficient the radar system is- and so on. Within the limitations imposed by such factors as range and response time, extent of cover, and-dare one say it?- financial cost and lead time, each part of the whole must be designed for optimum efficiency in order to achieve one end result.

Compared to what goes on in an actual battle, this should be a relatively simple problem for the operational analyst. The requirements seem precise enough; the parameters in the equations seem measurable; everything is as it should be for the mathematical and statistical analysis of an extremely complicated situation. In fact, one cannot do without this kind of extrapolated, theoretical operations analysis. But let us be honest with ourselves. The number of times this kind of systems analysis has come up with the wrong answers-by which I mean encouraged the development of the wrong systems-provides a warning that theoretical operational analysis is only an aid to, and not a substitute for, human judgment.

VI

Now let me turn to the more abstract form of operational analysis which has emerged over the postwar years. It deals with what has been called "the mathematizing of thought processes," and is based upon "game theory," which burst upon the world a few years ago. Protagonists believe that this form of operational analysis can contribute materially to "decision making" in the fields of strategy and tactics. The issues with which it deals, then, are among the most important which face Western man today- deterrent strategy and nuclear warfare in particular. Upon its conclusions-so we are told-are based certain of the most important decisions which have been taken in determining the defense posture of the United States in the strategic field, as well as some of the predetermined decisions which would be taken if deterrence ever failed.

I do not pretend to understand the intricacies of either the mathematics or the logical symbolism one finds in the writing of the game theorists. As I understand it, game theory is based upon the interaction of sequences of probabilities. It is assumed-to oversimplify the basic case-that you can calculate from a move made by an opponent at chess which of several possible moves open to you is the best-and so on to victory. In the next remove, it is assumed that you can make this best choice even when one randomizes one's opponent's moves. As a result, decision makers, we are told, can "optimize their solutions" even in the most complicated situations. This should undoubtedly be true in some cases-at any rate in theory-for example, in such problems as choosing where best to build a road, or which of several ways is the cheapest and best to build a nuclear reactor of a given type.

Certain parts of the very wide areas with which the calculations of the strategic game theorists deal are undoubtedly amenable to the most rigid treatment by formal logic and by means of probability statistics or mathematics, and the conclusions which emerge are fully valid over the part of the field to which they relate. What worries me is the fact that the total situation with which the theorists deal also contains extremely broad parameters of so qualitative a nature that no one could attribute numerical values to them. Some of these parameters are among the most important issues with which the game-theory strategists pretend to deal-if not the most important. For example, they include such matters as the enemy's intentions, as well as his strength and capacity; the resolution of our people; the capacity of a country to restore itself economically when it has suffered a degree of devastation well beyond anything that lies within human comprehension-let alone experience-and other matters equally vague. These are vitally important issues. But they are not numerical issues, and probably never could be made such, even if they were ever to come within our experience.

I have already said that some of the more empirical results of the analysis of operational planning in the Second World War were subsequently used in the planning of new operations. What I did not say was that this took place only after considerable energy had been spent in persuading people that what had happened had indeed happened, and that experience was as useful a basis for the elaboration of future plans as was wishful thinking. Since this was so 20 years ago in a less rapidly moving technological age, what I, as a scientist, would like to know now is what weight would a commander really attach-when the moment of decision came-to the new kind of abstract operational analysis as an aid to his judgment?

Mathematical conceptions of probability clearly have real value in experimental situations when they relate to precise issues. But the more qualitative and non-numerical-the more human-the factors that are played into the system, the less precise and, indeed, the less meaningful become the estimates of probability that are churned out of the machine. One has to remember, too, that the social and political significance and weight of the different qualitative factors concerned in these strategic situations vary not only between themselves but also from moment to moment. Strategy, no more than politics, can be static.

The fact that the considerations of the new school of operational analysis do not violate the laws of physics or engineering seems unimportant to me. What is important is that, if our sophisticated preparations for possible tactical warfare link the latter more closely with the set-piece strategic deterrent, game-theory operational analysis may not be able to help the tactical commander in his hour of need. As I have said, estimates of probability become less valuable the less precise numerically the matters with which they deal become. Can this kind of operational analysis serve any purpose at all in the field of military judgment and command? Does It now? Can it ten years hence? These are vital questions to which commanders should-so it seems to me-address themselves.

I believe that some commanders are over-optimistic, and over-generous, in what they imply about the scientific nature of the social sciences. I know of no philosopher of science, no student of scientific method, who would claim that the generalizations of the social sciences could ever be characterized by the same attributes of predictive value which apply to the natural sciences. The processes of war are essentially sociological phenomena-and like other such phenomena they certainly do not constitute the material out of which one can yet see a predictive science being born. It is conceivable that it might become such one day-as conceivable as the idea that "the thought processes can be mathematized." But if this involves, as I imagine it must, a prior understanding of the biochemical and biophysical nature of thought processes, that day is far away.

The validity of probability methods when applied to single events is an academic matter which has not yet been sorted out. The conclusions of the new school of theoretical military analysis are in essence based upon probability laws which apply to repetitive situations. If one decides wrongly about the use of nuclear weapons, we shall be in a situation which may never repeat itself, which may end the leadership of the Western world- win or lose-in one critical shot.

Before game theory burst upon a too-generous military world, Whitehead, a world-renowned philosopher and mathematician, had written:

There is a curious misconception that somehow the mathematical mysteries of statistics help us to evade the proper limitations of the observed past. But statistics tell you nothing about the future unless you make the assumption of the permanence of statistical form. . . . Mathematics can tell you the consequences of your beliefs. For example, if your apple is composed of a finite number of atoms, mathematics will tell you that the number is odd or even. But you must not ask mathematics to provide you with the apple, the atoms, and the finiteness of their number. There is no valid inference from mere possibility to matter of fact, or in other words, from mere mathematics to concrete nature.[i]

If ever there was a world in which situations do not repeat themselves like some mass production model, it is the military world. If we are to avoid the imposition of arbitrary limits to the exercise of judgment and control, let us be careful not to create in a mathematical vacuum situations which are based neither on past experience of affairs, nor on any conception of the innumerable variables and factors that determine social decision either today or tomorrow. The human brain, human values, human judgments, are still superior to the mechanics and processes of electronic computers or guidance systems. The day this ceases to be true there will probably be no human brains. But until then, let us use true scientific method as an aid to human judgment- and not as a hindrance. Science is human experience; it is not an alternative to judgment, and it is certainly not something that can operate outside human experience.

[i] Alfred North Whitehead, "Adventures of Ideas." New York: Macmillan, 1933, p. 161.

You are reading a free article.

Subscribe to Foreign Affairs to get unlimited access.

  • Paywall-free reading of new articles and a century of archives
  • Unlock access to iOS/Android apps to save editions for offline reading
  • Six issues a year in print, online, and audio editions
Subscribe Now