Chapter 1 The Role Of Knowledge

Ideas are everywhere, but knowledge is rare. Even a so-called "knowledgeable" person usually has solid knowledge only within some special area, representing a tiny fraction of the whole spectrum of human concerns. Humorist Will Rogers said, "Everybody is ignorant, only on different subjects."

How does an ignorant world perform intricate functions requiring enormous knowledge? These intricate functions include not only such scientific feats as air travel and space exploration, but also the complex economic processes which bring a slice of bread and a piece of butter to your plate at breakfast. Anyone who has studied the actual process by which everyday food items are planned, produced, and distributed knows that the complexity staggers the mind. Many highly intelligent and highly trained people spend a lifetime studying it, and learning more all the time. Among those who speculate financially in such commodities, economic disaster is commonplace, even after they have spent years studying the market. In short, individually we know so pathetically little, and yet socially we use a range and complexity of knowledge that would confound a computer. The question is not only how given institutions (including whole societies) manage to do this, but how various institutions (and societies) differ in the manner and effectiveness with which they do it-and what do the historic and continuing changes in the way they function portend for the future?

We shall begin with the production of knowledge-with the process by which ideas are filtered and transformed into recognized knowledge, having the force to guide decisions. knowledge, having the force to guide decisions. Then we shall consider the application of knowledge in economic, legal, social, and political institutions. And finally, we shall consider the evolution of institutions, attitudes, and beliefs, and the way all these affect our ability to produce and apply knowledge in the future.

IDEAS

Physicists have determined that even the most solid and heavy mass of matter we see is mostly empty space. But at the submicroscopic level, specks of matter scattered through a vast emptiness have such incredible density and weight, and are linked to one another by such powerful forces, that together they produce all the properties of concrete, cast iron and solid rock. In much the same way, specks of knowledge are scattered through a vast emptiness of ignorance, and everything depends upon how solid the individual specks of knowledge are, and on how powerfully linked and coordinated they are with one another. The vast spaces of ignorance do not prevent the specks of knowledge from forming a solid structure, though sufficient misunderstanding can disintegrate it in much the same way that radioactive atomic structures can disintegrate (uranium into lead) or even explode.

Ideas, as the raw material from which knowledge is produced, exist in superabundance, but that is produced, exist in superabundance, but that makes the production of knowledge more difficult rather than easier. Many ideas-probably most-will have to be discarded somewhere in the process of producing authenticated knowledge. Authentication is as important as the raw information itself, and the manner and speed of the authentication process can be crucial: the surprise attack on Pearl Harbor succeeded despite the fact that knowledge of the impending attack had reached the War Department in Washington hours before it occurred. Still the bombing caught Pearl Harbor by surprise because the information had not yet passed through the authentication process established by the military institutions. Whatever the merits or demerits of those institutions as they existed on December 7, 1941, it is clear that any military organization must have some authentication process, or else any unverified idea that enters the system has the potential to set off a war. More recently, a flock of Canadian geese set off the American warning system to detect incoming nuclear missiles, and only subsequent authentication procedures prevented a "retaliatory" nuclear strike which could have ended in World War III.

Various kinds of ideas can be classified by their relationship to the authentication process. There are ideas systematically prepared for authentication ("theories"), ideas not derived from any systematic process ("visions"), ideas which could not survive any reasonable authentication process ("illusions"), any reasonable authentication process ("illusions"), ideas which exempt themselves from any authentication process ("myths"), ideas which have already passed authentication processes ("facts"), as well as ideas known to have failed-or certain to fail- such processes ("false- hoods"-both mistakes and lies).

While these various kinds of ideas are conceptually different, in reality a given notion may evolve or metamorphose through several of these states. For example, we may start with a general impression of how and why certain things happen the way they do, without having any real evidence or any logically structured argument about it. But after we begin with such a vision, we may proceed to systematically determine that if this vision is correct, then certain empirical consequences will be observable under the proper conditions. The "vision" has led to a "theory." The proper conditions may be created in a laboratory or observed in history or otherwise constructed or discovered, and the validity and certainty of the results may be more or less open to criticism. The important point here is simply to distinguish such systematic authentication procedures from decisions based on consensus, emotions, or traditions.

On the continuum of human thinking, at one end is pure science; at the other end pure myth. One is sustained entirely by systematic logical procedures, the other by consensual verification by the other by consensual verification by contemporaries, by their predecessors represented through prevailing traditions, or by posterity for those who expect historic vindication. The crucial distinction is one of procedures, not of end results. Science is no more certain to be correct than is myth. Many scientific theories have been proven wrong by scientific methods, while the great enduring beliefs which have achieved the status of myths usually contain some important-if partial- truth.

Both systematic authentication and consensual approval can be further broken down. Systematic authentication involves a testing of the logical structure of a theory for internal consistency and a testing of the theory's results for external consistency with the observable facts of the real world.

Consensual approval may mean the approval of the general public as of a given time, or the approval of some special reference group-a social class, a religious sect, an ideological movement, etc.-in the past, present, or future. Ideas which lack logical, empirical, or general consensual support may still sustain themselves as acceptable to a consensus of those who regard themselves as special guardians of a particular truth-i.e., as the consensual reference group that really matters. Sometimes the elitism implicit in such a position can be tempered by depicting the idea in question can be tempered by depicting the idea in question (religious salvation, political reconstitution, etc.) as beneficial to a broad sweep of mankind outside the group, so that the group is only a temporary surrogate for a larger constituency which will ultimately approve the idea. But, of course, this proposition is itself still another idea lacking either empirical verification or general consensual approval.

There are many variations on the two basic ways of verifying ideas, and many combinations of these variations are used-often involving combinations from both systematic and consensual methods of verification in the same argument. For example, a scientific presentation may avoid-indeed, must avoid-unlimited verification of every incidental aspect of its arguments by saying, in effect, "everybody knows" this or that, and getting on with proving the things that need proving.' Similarly, beliefs resting essentially on consensual approval- religious beliefs, for example-may also employ logical and empirical techniques, such as the scientific "proofs" of the existence of God, which were common in the eighteenth century and in the early nineteenth century, before Darwin. These more or less open combinations present no special problems. A problem does arise, however, when one method masquerades as another-for example, when the results of essentially consensual processes choose to present themselves as scientific, as in the case of much so-called "social science."

This brief and general sketch of the production of authenticated knowledge from raw, unsubstantiated ideas must be elaborated more specifically in later discussions of economic, legal, and political organizations. At this point, it is necessary to consider-in equally brief and general terms-the amount and kinds of knowledge produced, and the manner in which it is used.

THE QUANTITY OF KNOWLEDGE

It is widely believed that modern society has a larger quantity of knowledge than more primitive societies, that this quantity of knowledge is growing, and that the knowledge "required" for the average citizen to live in a modern society is also growing. Certainly the complex apparatus of modern life is beyond the grasp of most non- modern peoples, past or present. What is not so obvious, but true nonetheless, is that most modern peoples would find it equally-or more-difficult to survive individually in a "primitive" or nonmodern world. In short, it is not clear or demonstrable that the total quantity of knowledge differs as between "savage" and "civilized" man. What is more readily established is that the kinds of knowledge possessed by the average inhabitant of the primitive and the modern world are very different, and that each would be at considerable hazard in the world each would be at considerable hazard in the world of the other.

Consider a modern civilized man suddenly stranded in a primitive jungle, cut off from modern technology, and unaided by such primitive peoples as might exist in that environment. Although the civilized man might be a well educated individual, working in a complex profession such as accounting or electronics, it is doubtful whether his knowledge would be sufficient to merely sustain his life in an environment where primitive peoples have lived for untold generations. The civilized man might often have a choice of going hungry or eating wild vegetation which could prove either nutritious or poisonous. Finding a safe place to sleep at night would require more knowledge of the habits and capabilities of wild animals than he possessed. Avoiding snake bites, infected water, and predatory beasts would be among his other problems, and ordinary illnesses easily cured in a civilized community could be far more dangerous away from scientific medical knowledge and without the herbal and other folk remedies available to primitive man. In the same environment, the savage could not merely survive, but thrive, producing housing, clothing, and other amenities. But of course the primitive man's chances of survival if suddenly dropped down in the midst of New York or Los Angeles might also be bleak.

What then is the intellectual advantage of civilization over primitive savagery? It is not necessarily that each civilized man has more knowledge but that he requires far less. A primitive savage must be able to produce a wide variety of goods and services for himself, and a primitive community must repeatedly duplicate his knowledge and experience in innumerable contemporaries. By contrast, the civilized accountant or electronics expert, etc., need know little beyond his accounting or electronics. Food reaches his local supermarket through processes of which he is probably ignorant, if not misinformed. He lives in a home constructed by an involved process whose technical, economic, and political intricacies are barely suspected, much less known to him. His home is likely to be stocked with many devices working on mechanical and electrical principles which he neither understands theoretically nor can cope with as a practical matter. The chronic complaints and scandals about appliance, automobile, and other repair services testify to the civilized man's utter lack of knowledge of the everyday apparatus on which he depends. A primitive savage could never survive knowing so little about the production and use of spears, grass huts, or with such utter naivete about which berries are poisonous, which snakes dangerous, or the ways and means of coexistence in the same jungle with lions, tigers, and gorillas.

Civilization is an enormous device for economizing on knowledge. The time and effort (including costly mistakes) necessary to acquire knowledge are minimized through specialization, which is to say through drastic limitations on the amount of duplication of knowledge among the members of so ciety. A relative handful of civilized people know how to produce food, a different handful how to produce clothing, medicine, electronics, houses, etc. The huge costs saved by not having to duplicate given knowledge and experience widely through the population makes possible the higher development of that knowledge among the various subsets of people in the respective specialties.

THE MEANING OF "KNOWING"

Although the phrase "ignorant savage" may be virtually self-contradictory, it is a common conception, and one with a certain basis. The savage is wholly lacking in a narrowly specific kind of knowledge: abstract, systematized, knowledge of the sort generally taught in schools. Considering the enormous range of human knowledge, from intimate personal knowledge of specific individuals to the complexities of organizations and the subtleties of feelings, it is remarkable that one speck in this firmament should be the sole determinant of whether someone is considered knowledgeable or ignorant in general. Yet it is a fact of life that an ignorant in general. Yet it is a fact of life that an unlettered peasant is considered ignorant, however much he may know about nature and man, and a Ph.D. is never considered ignorant, however barren his mind might be outside his narrow specialty and however little he grasps about human feelings or social complexities. We do sometimes refer to a "learned fool," but the notion of a "fool" implies deficiencies in the reasoning process (so that one is easily deceived or fooled), whereas it may actually be knowledge that is lacking, so that the "learned" person has simply not learned enough outside a certain sliver of human experience.

The point here is not simply to deplore the use of certain words. The point is to avoid having our own discussion of knowledge drastically shrunk, arbitrarily, and virtually without our realizing what is happening. We need to consider the full breadth of knowledge and its depth as well. That is, we need to consider not only how much we know, but how well we know it.

We start with an idea. It may be a sense impression of some sort-something that happened to catch our eye and intrigue our curiosity. Or it may be a speculation in our mind-a daydream or a theory, for example. As the idea or theory passes through the authentication process, it may be verified, refuted, or transformed to accommodate additional and discordant evidence. But if the authentication process is doing its job, whatever authentication process is doing its job, whatever conclusion it is reaching about the idea is becoming progressively more certain (even if that means that the original idea itself is becoming progressively more dubious). Therefore, at some point in the authentication process, the probability of a mistaken conclusion is reduced to the point where we can say that we "know" this or that. Where that point is varies from person to person, so that what is "knowledge" to one is merely a plausible belief to another and only a theory to someone else. Each of us has some point-some probability levelbeyond which we will say that we "know" something. But all things fall short of absolute certainty: life itself might be a dream and logic a delusion. Still, because we act, we must decide, and how decisively we can act depends on how well we know the consequences.

How much knowledge there is depends on where we draw the line on the spectrum of probabilities. Within a given probability requirement for "knowing," how much is known varies enormously from one area of human life to another, and from one historical era to another, and of course from one person to another. Because the arena of decision making almost always exceeds the arena of knowledge, there must be belief-or at least hope-to fill in the gaps where there is no knowledge. This means that the ratio of knowledge to belief may also vary enormously from one aspect of life to another. The specific nature of the respective authentication processes available in various aspects of human life then become crucial.

To say that a farm boy knows how to milk a cow is to say that we can send him out to the barn with an empty pail and expect him to return with milk. To say that a criminologist understands crime is not to say that we can send him out with a grant or a law and expect him to return with a lower crime rate. He is more likely to return with a report on why he has not succeeded yet, and including the inevitable need for more money, a larger staff, more sweeping powers, etc. In short, the degree of authentication of knowledge may be lower in the "higher" intellectual levels and much higher in those areas which intellectuals choose to regard as "lower." A business which produces a product that the public will not buy in a sufficient quantity, or at a high enough price to cover production costs, will have its ideas validated-in this case invalidated-in a swift and painful process which must be heeded quickly before bankruptcy sets in. The results cannot be talked away. But in many intellectual areas, notably so-called "social science," there is neither a swift nor a certain authentication process for ideas, and the only ultimate validation is whether the ideas sound plausible to enough people, or to the right people. The stricter standards and independent, often conclusive, evidence in the physical sciences cannot be generalized to intellectual activity as a whole, even though the aura of scientific processes and results is often appropriated by other intellectuals.

Because what is meant by "knowing" varies enormously, according to the respective authentication processes available, it is by no means clear that there is more knowledge in civilized countries than in primitive countries or among intellectuals as compared to the less educated members of the same society. It is very possible that, as more people cease being farmers with little or no education, and increasingly acquire more schooling, that their standards for "knowing" decline while the area of their secondhand and tenuous knowledge expands. As a poet said, "we knew a million things we could hardly understand."' There may be not only a qualitative decline in knowledge, but-more important-an erosion of the very meaning of "knowing": for example, a young man might be said to know how to milk a cow if he could write an essay on that subject, and we would no longer demand that he take the pail out to the barn and come back with milk.

It is not necessary, at this point, to insist that the average amount of personal knowledge has declined over time. It is sufficient that we realize that conflicting trends are at work, and that the net result is an open question, rather than the foregone conclusion often assumed by those who depict an conclusion often assumed by those who depict an ever more knowledgeable society needing ever more years of schooling for its citizens.

The march of science and technology does not imply growing intellectual complexity in the lives of most people. It often means the opposite. Matthew Brady required far more knowledge of photoghraphic processes to take pictures with his cumbersome equipment during the Civil War than a modern photographer requires to operate his automated cameras. Science and technology lead to far more complexity in producing cameras and film today, but that growing complexity among a handful of technicians permits far more simplicity (and ignorance) in the acutal use of modern photographic equipment and materials by a mass of people. Similar trends are discernible in a wide variety of fields. Automobiles are much more complex to build, but far simpler to operate, than in the days before automatic ignition, automatic transmissions, automatic chokes, self-sealing tires, etc. The technology available in the modern home reduces not only the time but the knowledge required by a modern homemaker. Even a mere man can now perform some chores for which girls and young women were once trained for years.

The growing complexity of science, technology, and organization does not imply either a growing knowledge or a growing need for knowledge in the general population. On the contrary, the increasingly complex processes tend to lead to increasingly simple and easily understood products. The genius of mass production is precisely in its making more products more accessible, both economically and intellectually to more people. Electronic calculators enable mathematical illiterates to perform operations which only highly trained people could perform with ease in earlier times. The printing press performs daily communications miracles beyond the ability of an army of the most highly trained and dedicated scribes of the Middle Ages. Organizational progress parallels that in science and technology, permitting ultimate simplicity through intermediate complexity. An ordinary individual can easily arrange travel across thousands of miles through cities he has never seen by tapping the knowledge of travel agents and/or the American Automobile Association. Or he can weigh the relative merits of commercial products whose individual mechanisms are wholly unknown to him, by reading the (simple) results of highly complex tests conducted by general consumer magazines or by publications specializing in particular items such as audio equipment or motorcycles.

DECISION-MAKING UNITS

Knowledge may be enjoyed as a speculative diversion, but it is needed for decision making. The genesis of ideas and the authentication of genesis of ideas and the authentication of knowledge are part of a continuous process which ultimately brings knowledge to bear on decisions- when the system is working ideally. In real life, the process may, of course, fail to bring knowledge to bear, when accurate knowledge is available somewhere in the system. What matters, then, is the knowledge actually used at the decision-making point, not the knowledge in process of development or authentication, nor even the knowledge clearly apparent to particular individuals or organizations somewhere in the society. And while decisions may be thought of as made by specific individuals at specific points in space and time, the decision- making process is more usually structured so that various combinations of individuals repeatedly and habitually make certain classes of decisions, so that they form continuously functioning decisionmaking units, which may range from a married couple to a police department to a national government. A single individual may also form a decisionmaking unit for some purposes, or-more likely-he may be part of several decision-making units simultaneously, and the set of such institutions may change over time.

The emphasis on specific decision-making units is especially necessary in an era given to metaphors about an amorphous "society" deciding to do this or that: "Society" doesn't keep its air or water clean; "society" is punitive, permissive, frivolous, uptight, generous, uncaring, etc. While metaphors may sometimes be useful shortcuts, like other shortcuts they can also take us further away from our destination and delay or even prevent our arrival there.

Metaphors which suggest that "society" is a decision-making unit can be very misleading, by ignoring situations in which decisions are what they are precisely because the actual decision-making units face a particular kind of incentive structure. To ignore the specific nature of the decision- making units is to expect improvement by trying to substitute "the good guys" for "the bad guys," or by waiting for the Messiah or for the general triumph of human reason, whichever seems less improbable or less remote in time. Sometimes the metaphor of "society" is used more tendentiously, to quietly shift the locus of decision making from smaller and more numerous units to a single nationwide decision- making unit. The merits or demerits of such a change in any specific case are simply bypassed by metaphors which proceed as if "society" is doing this now and ought to do that instead-when in fact one set of decision-making units is operating under one structure of incentives now and the advantages and disadvantages of an alternative decision- making unit and the alternative set of incentives is precisely what needs to be explicitly analyzed, not covered up by metaphors about "society."

There is no one named "society" who decides anything. Even in the most democratic nations few issues are ever decided by a specific nationwide referendum. And even if they were, who could say that a bare majority as of a given instant constitutes the judgment of an organic society subsisting over the generations? Unless national laws are to vary literally from moment to moment, some decision- making units must make decisions which are binding on other units which either disagree or were not consulted. Posterity is of course never consulted.

One of the peculiarities of the American Revolution was that its leaders pinned their hopes on the organization of decision-making units, the structuring of their incentives, and the counterbalancing of the units against one another, rather than on the more usual (and more exciting) principle of substituting "the good guys" for "the bad guys"-i.e., substituting "the people" for "the oppressors," the faithful for the heathens, the Jews for the gentiles, the gentiles for the Jews, and other such substitutions based on differences of history, physiognomy, or mannerisms.

The domain of decision-making units need not be discrete or mutually exclusive. Indeed they cannot be either, or there would be no such social phenomenon as would cause us to refer even metaphorically to "society." Decision-making units seldom have complete control, even of a given seldom have complete control, even of a given segment of a society, and no decision-making unit controls the whole society, except very approximately under a totalitarian regime. Decision- making units overlap one another to some degree, and even where such units are subordinated to others in a hierarchy, the subordination is never perfect in practice. Even in the extreme case of slavery, the subordinate units took actions contrary to the general desires or specific orders of the higher units-ranging from passive or active sabotage to murdering overseers and slave owners.' The practical limitations of sheer subordination were repeatedly demonstrated by the various economic incentives which had to be resorted to under slavery, especially for getting higher quality work performed.'

In general, the ability of subordinate decision- making units to act independently of, and contrary to, the policies and orders of the higher units is based on differences in knowledge. The powers of the higher units may encompass all the powers of the subordinate units, but they almost never encompass all the knowledge. Because the powers of the higher decision-making units include the power to require transmission of knowledge, the persistence of knowledge advantages by the subordinate units implies either an impossibility or a prohibitive cost to the higher unit of independently acquiring the same knowledge as a check against the accuracy of the knowledge transmitted by the subordinate unit. In short, there are differences in their respective costs of acquiring knowledge. More specifically, there are cost differences between higher and lower decision-making units which vary according to the kind of knowledge in question.

General knowledge-expertise, statistics, etc.-is usually more economically used by the higher decision-making units. For example, a decisionmaking unit which encompasses five subordinate units can acquire a given expertise and statistical data which it applies to all five units, whereas if each unit had independently acquired the same expertise or statistical data it would have cost five times as much altogether. For this kind of knowledge, the cost advantage tends to be on the side of the larger and higher decisionmaking unit. But for highly specific knowledge-the local life style, the reliability of particular suppliers, the level of skill of a given executive, etc.-the subordinate units immediately in daily contact with the relevant facts can much more easily and more cheaply synthesize the knowledge and draw inferences.

It is unnecessary to attempt any general rule as to where the overall balance lies in comparing the respective costs of knowledge in larger and smaller decision-making units. What is important is to understand that (1) the respective cost advantages of the large and small units differ according to the kind of knowledge involved (general versus kind of knowledge involved (general versus specific), that (2) most decisions involve mixtures of the two kinds of knowledge, so that the net advantages of the larger and smaller units vary with the kind of decision, and (3) the effectiveness of hierarchical subordination varies with the extent to which the subordinate unit has knowledge advantages over the higher unit. In those cases where the subordinate unit has better information, then in terms of the whole decision-making process the knowledge is one place and the power is another; the quality of decisions suffers as a result. Moreover, subordination itself becomes illusory to the extent that the lower level unit can use its knowledge advantages to evade, counteract, or redirect the thrust of orders from its nominal superiors.

Some examples from various institutions and various societies may illustrate these crucial points. Agriculture has its general principles and statistics, but agricultural production involves much highly specific knowledge about the characteristics and contours of particular plots of land, and about the freshness, flavor and keeping qualities of specific batches of fruits, vegetables and dairy products-all of which are changing by the hour. No expert can say from 100 miles away, and sight unseen, that this year's grape crop is good, or even that last week's good grapes are still good this week. By contrast, an expert on the manufacture of steel can specify the exact quality of steel that will be produced by given combinations of iron ore and coal at given temperatures. For these reasons, steel production has been successfully centrally planned and controlled in various countries, whereas agricultural production has had such chronic problems and periodic disasters in centrally planned economic systems that even the most centralized communist governments have had to make major exceptions in agriculture, allowing decentralized decision-making of various sorts.

For similar reasons, in capitalist countries it is common to have chains of grocery and department stores selling standardized items, but there are no large chains of high quality restaurants of a sort which depend upon atmosphere and finely prepared food. Such restaurants require constant attention to the demeanor of the staff and the delicacy of the chef, and those cannot be effectively controlled by distant experts. Usually the owner and manager of a successful restaurant of this sort is right on the premises, often from the moment it opens each day to the moment it closes at night. By contrast, the top executives of Sears or Safeway cannot and need not be present at their hundreds of stores across the country, for much of the knowledge they need can be gained from statistics, experts, and accounting data.

THE STRUCTURE OF INCENTIVES

While decisions are constrained by the kinds of organizations and the kinds of knowledge involved, the impetus for decisions comes from the internal preferences and external incentives facing those who actually make the decisions. The incentives may be positive or negative-that is, rewards or penal ties. Typically, these incentives are structured in some way, so that there are gradations of rewards (or penalties) corresponding to different kinds of results. It is not just a question of being rewarded or not, but of how much reward or penalty is likely to follow from various decisions. Simple as this seems, it is a radical departure from the practice of explaining decisions in terms of "society's" choices or in terms of the official or ostensible "purpose" of an organization. An organization may make decisions which fail to achieve its assigned purpose or fail to serve society's interest, without any "failure" of understanding or ability, simply because it is responding to the actual structure of incentives confronting it rather than to the rhetoric or hopes of others.

Much criticism of "incompetent bureaucrats" implicitly assumes that those in the bureaucracy are pursuing the assigned goal but failing to achieve it due to lack of ability. In fact, they may be responding very rationally and ably to the set of incentives facing them. For example, government regulatory agencies are often very ineffective in controlling the industry or sector which they have a legal mandate to regulate. But it is a common pattern in such agencies for those in decision- making positions to (1) earn far less money than comparable individuals earn in the regulated sector, and (2) after a few years' experience to move on to jobs in the regulated sector. In short, they are regulating their future employers. Under such a set of incentives, it is hardly surprising that decision makers in regulatory agencies approach those whom they are assigned to regulate with an attitude that is sympathetic, cooperative, and even protective. The only protection of the public interest built into the incentive structure are the penalties for blatantly illegal conduct, such as taking bribes to make a particular decision for a particular company. But explicit bribes are seldom necessary in order to get the regulatory agency to adopt the general viewpoint of the regulated sector, in which many regulatory officials expect to make a more lasting and more lucrative career than is open to them in government. Morally, it is possible to deplore individual weakness or selfishness, but rationally there is little reason to expect a different outcome from a normal sample of people facing the same structure of incentives. Reform by "throwing the rascals out" seems less promising than reform by changing the structure of incentives facing whoever occupies decision-making positions.

The regulatory agency example is a case where the institutional incentive structure has to compete with an outside incentive structure that is more attractive financially. Incentive structures can have problems in themselves, aside from outside competition. The mere process of formalizing what is to be rewarded presents many complexities and pitfalls. Most problems, decisions, and performances are multidimensional, but somehow the results have to be reduced to a few key indicators which are to be institutionally rewarded or penalized: attendance records, test scores, output per unit of time, seniority, etc. The need to reduce the indicators to a manageable few is based not only on the need to conserve the time (and sanity) of those who assign rewards and penalties, but also to provide those subject to these incentives with some objective indication of what their performance is expected to be and how it will be judged. But, almost by definition, key indicators can never tell the whole story. This affects not merely the justice or injustice of the reward, but also the very nature of the behavior that occurs within the given structure of incentives. For example, one index of military success is the number of enemy killed. Clearly, it is not the only indicator, for if a major military objective can be taken while capturing the enemy, or confronting him with sufficient force to make him retreat, or bluffing him into withdrawal or surrender, this is even better than having to actually take the objective by storm, with a large loss of life on both sides. However, once the incentive structure clearly rewards the "body count" of enemy dead, it provides an incentive to more carnage than is absolutely necessary, and since enemy casualties can seldom be increased without increasing one's own casualties, it provides an incentive to needless bloodshed and loss of life among one's own troops. Again, moral condemnation without reform of the incentive structure means little. For example, continual criticism of the "search and destroy" missions of the American army in Vietnam did little to change this approach in a war where "body count" was a key indicator, used by the military high command in rewarding and publicizing its units' efforts.

TIME

Key indicators require some specified time span during which they are to be tabulated for purposes of reward or penalty. The time span can vary enormously according to the process and the indicator. It can be output per hour, the annual rate of inflation, weekly television program ratings, or a bicentennial assessment of a nation. But whatever the span chosen, it must involve some simplification, or even oversimplification, of reality. Time is continuous, and breaking it up into discrete units for purposes of assessment and reward opens the possibility that behavior will be tailored to the time period in question, without regard to its longer time period in question, without regard to its longer range implications. Desperate efforts just before a deadline may be an inefficient expedient which reduces the longer run effectiveness of men, machines and organizations. The Soviets coined the term "storming" to describe such behavior, which has long been common in Soviet factories trying to achieve their monthly quotas. Similar behavior occurred on an annual basis in Soviet farms trying to maximize the current year's harvest, even at the cost of neglecting the maintenance of equipment and structures, and at the cost of depleting soil by not allowing it to lie fallow to recover its long run fertility. Slave overseers in the antebellum South similarly overworked both men and the soil in the interest of current crops at the expense of reduced production years later-when the overseer would probably be working somewhere else. In short, similar structures of incentives produced similar results, even in socioeconomic systems with widely differing histories, ideologies, and rhetoric.

IMPLICATIONS

The broad sweep of knowledge needed for decision making is brought to bear through various systems of coordination of the scattered fragmentary information possessed by individuals and organizations. This very general sketch of the principles, mechanisms, and pitfalls involved is a principles, mechanisms, and pitfalls involved is a prelude to a fuller consideration of the use of knowledge in decision-making processes in the economic, legal, and political spheres, each having its own authentication processes and its own feedback mechanisms to modify decisions already made. Much discussion of the pros and cons of various "issues" overlooks the crucial fact that the most basic decision is who makes the decision, under what constraints, and subject to what feedback mechanisms. This is fundamentally different from the approach which seeks better decisions by replacing "the bad guys" with "the good guys"-that is, by relying on differential rectitude and differential ingenuity rather than on a structure of incentives geared to the normal range of human propensities.

The discussion thus far has emphasized premeditatedly-formed and hierarchically-structured decision-making units. These are not the only, nor necessarily the best, decision-making units, nor even the most pervasive kind of decision-making units at a given time and place. Some alternative decisionmaking units and processes include (1) trial by combat, which is seldom sanctioned today for individual decision making, but is still the ultimate decisionmaking mechanism between sovereign nations, (2) various arrangements spontaneously evolved by the participants, such as competitive bidding in economic markets or mutual benevolence in groups bound together by religious, artistic, tribal, or other affinities, and (3) premeditated arrangements in which those subordinated to the power of others in one sense are, in another sense, the ultimate arbiters of the fate of their hierarchical superiorsas with democratically elected governments, or with governments operating in the shadows of their own military forces which are both willing and able to depose them. None of these decision-making processes are mutually exclusive. A typical American, for example, lives in a family unit whose internal decisions are based on personal feelings, works in a hierarchically-structured organization whose use of inputs and volume of output are determined in a spontaneously evolved market, is subject to laws established by a government whose members are chosen and removable by the electorate and which conducts its relations with other governments in an atmosphere dominated by their respective capacities for armed combat or mutual annihilation.

The interaction of various decision-making processes makes it all the more necessary to understand the respective principles of the different individual processes. The continual evolution of decision-making units and decisionmaking processes likewise makes it all the more necessary to understand the effects of different kinds of processes, so as to know where we are headed if current trends continue.

just as decision-making units and processes vary enormously, so too do the various kinds of decisions. For example, some decisions are binary decisions-yes or no, war or peace, guilty or innocent-while other decisions are continuously variable incremental decisions: using more or less gasoline, paying higher or lower wages, living a more relaxed or more hectic life. Some decisions are once-and-for-all decisions-suicide, loss of virginity, burning a Rembrandt painting-while others are readily reversible decisions: turning off a television program that is not interesting, cancelling a subscription, ceasing to purchase a given brand of consumer goods or ceasing to use certain cliches, etc. Decisions may also be made individually or as "package deals." One can buy onions, bread, and canned goods in the same store or in different stores, but in choosing between political candidates, one must choose one candidate's whole package-his fiscal policy, environmental position, foreign policy, civil liberties views, etc.-as against the whole package of his opponent's positions on the same subjects.

The kind of decision is not tied to the particular subject matter (i.e., shoes, food, or education) so much as to the particular decision-making process: economic processes, legal processes, political processes, etc. What this means is that as certain kinds of decisions are moved from one kind of decisionmaking unit to another, it is not merely a case of a different group of people or processes making the decision; the nature of the decision itself can change. That is, what was once a continuously variable decision may become a binary decision. Prior to public schooling and compulsory attendance laws, for example, the decision a family made was how much schooling to purchase for their children; afterwards, the only decision was whether or not to obey the compulsory attendance laws. Before it became a federal crime to carry a letter in competition with the post office, the individual letter-writer could choose among various possible carriers, but afterwards the only decision was whether to communicate in the form of a letter or in some other form.

Decisions also differ with respect to whether they are instantaneous or sequential. An instantaneous decision occurs completely at a given point in time, even if a long period of consideration preceded it, while a sequential decision occurs at various points in time as reactions to previous parts of the decision entail additional adjustments, improvisations, or reinforcements. The basic difference between them is that one decision is made completely on one occasion, while the other decision occurs piecemeal over a period of time. With sequential decision making, all the knowledge which is finally available to the decision maker is not initially available when the sequence of decisions begins, and the course of action followed may be wholly different from what it would have been if all the different from what it would have been if all the knowledge had been available at the outset, or if any decision could have been postponed until after all the facts were in.

Many early supporters of the Vietnam war came ultimately to the position that it was not worth the cost, after the full cost had been revealed by time, and that early official estimates of prospective casualties and prospective outcomes were either grossly mistaken or deliberately misleading. Another contemporary example of sequential decision making in a very different area is the progression from the Supreme Court's Brown decision in 1954 that the state cannot classify students by race for differential treatment to its controversial "busing" position in which that is what it requires states to do. Years of opposition to desegregation of the public schools led to progressively tighter judicial control, designed to overcome the various strategies of opposition, delay, and evasion-ultimately arriving at a point the opposite of the court's original premise or intention. In an earlier era, British Prime Minister Neville Chamberlain conducted a foreign policy designed to avoid war with Hitler through relatively small concessions, but the ultimate result of an unanticipated series of crises and concessions was to so shift the balance of power in Hitler's favor as to make war inevitable.

None of these sequential decisions was the result of a "society" that was stupid in the light of information now available in retrospect, but rather of piecemeal decisions which acquired a momentum of their own, and of the individual decision makers who were unequal to the unfolding complexities inherent in sequential decision making. Praise or blame is not the point. What is important is to understand (1) when a situation facing us is part of a sequential decision making process, and what that implies, and (2) to understand when our own institutions set up sequential decision-making processes when there is an alternative decision-making process available. For example, Chapter 9 will analyze the criminal justice system as a series of sequential decisions presented to the young criminal in such a way as to lead more people to persist in a life of crime than would do so if all the knowledge of prospects and penalties were made fully available to them at the outset.

In addition to considering decision-making processes, we need to consider decision-making costs. These costs are not simply the salaries of decisionmaking officials during the time when they are pondering what to do. Clearly the cost of evaluating intelligence reports on Japanese intentions to bomb Pearl Harbor was not simply the pay of the military functionaries who handled these reports. The cost of those processes included one of the largest military catastrophes in American the largest military catastrophes in American history, and the loss of life and material not only at Pearl Harbor but in a series of major military defeats in the months that followed, in the wake of the crippling and near-annihilation of the American Pacific Fleet on December 7, 1941. The point here is not to condemn, or even to evaluate, the decision- making process as it existed in the military at that time. The point here is to emphasize that the cost of any decision-making process must be assessed in terms of the full consequences entailed by alternative decision-making processes. Such processes cannot be judged by narrowly conceived economic or financial criteria. As we shall see in Chapter 3, even economic decision making cannot be evaluated narrowly in money terms alone.

The chapters that follow will consider the use of knowledge in economic, legal, and political institutions, the nature of the intellectual process and the role of intellectuals as a social class in influencing trends in modern society. Some disturbing implications of those trends will then be weighed.