Air University Review, January-February 1986

Policy, Intelligence, and
The Billion-Dollar Petroglyph

Lieutenant Colonel G. Murphy Donovan

THE relationship between intelligence and policy is complex and frequently difficult to understand. In an ideal world, good intelligence serves power and truth with equal integrity. In practice, this is not always the case. Many political paradigms, especially those with a strong ideological base, are impervious even to the best intelligence.1 On one hand, good analysis does not guarantee good decisions or policy. On the other, the potential for sound policy is ill-served by the alternative. Nevertheless, the relationship between intelligence and policy is essentially symbiotic. Policy looks to intelligence for a logic of evidence, and significant analysis, looks for policy to serve.2

Intelligence and policy, their institutions and systems, are often viewed as separate activities. At their worst, they probably are. At their best, institutional and systemic boundaries are blurred. The twilight zone that separates intelligence and policy is infinitely more difficult to describe than the institutions themselves. This "no man's land" seldom becomes the object of scrutiny except after an intelligence failure. Yet most intelligence failures are political failures waiting for the jury to come in.3 Institutionally, policy is the ultimate power broker, and realistically, it is in the position to be lion to the intelligence ram. The danger is one of proximity. As P. T. Barnum once observed: "If you put a ram in a cage with a lion, you need plenty of rams in reserve."4

Institutions and definitions of their activities are both plagued by oversimplification. For the sake of convenience, we often speak of "the" policymaker or "the" intelligence analyst, but policy and intelligence seldom have single authors. Each is fashioned in separate institutions (for reasons prudent and convenient), and the players who influence output are numerous.

Within the intelligence community, the need for multiple views is often cited as the rationale for multiple institutions. However, the output of the corporate intelligence "system" is more often characterized by consensus than institutional originality. The pressure for corporate intelligence consensus is as great as the pressure for corporate policy consensus.

The policy and intelligence processes are different but not separate. Intelligence is defined through analysis, and policy is defined through implementation.5 New policy can focus intelligence analysis, and new intelligence may influence policy changes. The relationship is dynamic, and exchanges are not necessarily sequential but invariably interactive. This interaction is not always harmonious; indeed, often it is a troubled road characterized by the need for reduction, the intrusion of bias, and the vagaries of a vast collection and processing subculture.

Too often the policy/intelligence relationship, particularly in the defense establishment, is viewed idealistically, and this romantic view undermines the very process of effective interaction. Ideally, policy and intelligence are collegial partners in pursuit of larger national security goals. In practice, intelligence is somewhat of a junior partner with, what may be, a self-imposed image problem.

Traditional suggestions for improving the quality of military intelligence support to the national security debate have focused on resource augmentation. Improved outcomes are inexorably, and often inexplicably, tied to more dollars and more sophisticated collection technology. However, what would improve intelligence most in the defense arena are three shifts in emphasis that require little or no new resources: a better understanding of the corporate personality of policymakers; a recognition of the role that bias plays in policy formulation and intelligence analysis; and a change in the image of the intelligence process, coupled to an upgrade in the stature of intelligence managers.

The Corporate Personality
of Policymakers

If the intelligence ram is to lie down with the policy lion and survive, he should, at the outset, understand the nature of the beast. There are a number of attributes that are common to most successful policy players. Primary is the possession and use of power. Power is the fuel that fires the political furnace. Intelligence can only influence the decision process, while the power to drive policy lies in other hands. Bad intelligence can be embarrassing or inconvenient; bad policy can be fatal.

Power brokers do not suffer fools gladly. They are decisive, confident, sure of their ideology, and, not uncommonly, convinced that they are their own best analyst. The policy lion also has a vested interest in his policy.6 He thrives on optimists and boosters--and often finds it difficult to quarrel with their good judgment. If he makes policy, he has been successful. More often than not, he views his success as a confirmation of his way of doing things. In short, the policy lion is a formidable beast.

the definitive policymaker

The casual political science artist paints the typical policymaker as a member of the executive branch.7 The usual stereotypes are cabinet officers, department heads, or military commanders. Even select members of the intelligence community occasionally enter this elite group.8 Such formulations would give the framers of the Constitution collective gout.

The stated intent of our founding fathers was not to vest such sweeping authority in a single elected or, worse still, appointed official. The original design of the (republican) government vested the authority to make policy (or law) with Congress--the representatives of the people. The executive branch was chartered only to enforce policy. Nonetheless, over time, Congress has delegated much of its authority to the executive branch, which, in turn, has passed much of its policy charter down to appointed officials. Today, the popular myth sees the policymaker as anyone but a member of Congress.

Modern policy, and even intelligence, apparatchiks commonly remonstrate against the "meddling" or "intrusions" of Congress in the policy process. Their protests are mostly heat and smoke . The fire in the political furnace may smolder occasionally, but the policy oven is still up on Capitol Hill.

the predispositions of policymakers

The task of influencing policy is, like lionbaiting, often uncertain and always dangerous. There are a number of policy predispositions that will invariably give the intelligence analyst fits.

Fearing the unknown and uncertain. Decision makers don't like to see untidy intelligence, even if it does accurately reflect a complex and ambiguous world.

Forecasts or estimates are especially bothersome because they illuminate variables and expand uncertainties and also tend to be windy and wordy. Who has time to read in the midst of a policy brawl where time is short, the stakes are high, and the relevant intelligence is buried in a 500-page tome? If the decision maker has not seen and understood the estimate long before the crisis, the estimate will have no influence during or after the key events and decision making. Even with the best of estimates, a smug "I told you so" from the intelligence corps will do little save hasten the transition of ram to sacrificial lamb.

Policymakers understand the difference between a forecast and a prophecy, yet, given a choice, most would still prefer a prophecy. Unfortunately, intelligence, unlike religion, seldom provides elegant solutions.

Wanting viable options. Power brokers seldom care for uncertainty, but they do like options. Unfortunately, intelligence often reminds them of their limited influence on events. Intelligence that limits choices corners the beast.

Recent events in Lebanon illuminate a policy environment where choices were limited severely. In such narrow confines, policy often becomes an ally of the problem. In such cases, even the most objective assessments may serve only to remind power brokers that things can get worse.

Disliking that which undercuts established policy. The political world is awash with pet paradigms, conventional wisdom, and vested interests. All of these at times find their way into policy. Intelligence that questions policy, often in the form of protracted divergent views or new insights, is seldom welcomed. Policymakers frown on continuous disagreements and absolutely abhor surprises, especially those that challenge policy. Worst of all, disagreements and surprises provide ammunition to opponents. Bad news can be correct but seldom will yield good effects--especially for the messenger.

The infamous Pentagon Papers revealed that there were a number of protracted, divergent views on Vietnam policy within the intelligence community--for more than a decade. Later, during the Carter administration, the sudden discovery of a significant increase in the North Korean order of battle was a good illustration of new intelligence that undercut a plan to withdraw the U.S. Second Infantry Division from South Korea. The more recent discovery of a Russian brigade in Cuba is another example. In the Korea instance, it is still not clear whether or not Eighth Army Chief of Staff, Major General John K. Singlaub, was speaking for command intelligence when he publicly disagreed with the Korean withdrawal policy. In any case, his message was bad news, and he was the first casualty.

Avoiding public controversy. The policy lion purrs with constancy and cringes from controversy. Controversy is another form of bad news. When intelligence analysts cannot agree on the range of a bomber, the value of civil defense, or the level of Ivan's defense spending, these issues are likely to be settled by fiat. Controversy and uncertainty often provide the much for the garden of asserted conclusions and worst-case scenarios.

Persuading the public. Policymakers are vested with uncommon authority. There is an element of mystery or magic about what they do. Yet they still look to intelligence for the logic of evidence that occasionally argues for public confidence.

In the early 1960s, President Kennedy disclosed sensitive intelligence to argue his Cuban policy. Unfortunately, the long-term impact of such a dramatic public gesture was not well appreciated. If intelligence could be used to argue for "good" policy, then policy opponents reasoned that it also should be used to argue against "bad" policy. Thus "leaks" for and against all manner of national security issues became the order of the day.

Public disclosure now has all the charm of Pandora's box. The Cuba-related performance, for which the intelligence community is still oddly taking bows, may have done more to encourage security breaches than the KGB.9 The policymaker's conflict is between expedience and prudence. Usually, putting the intelligence system at risk for transient political gains is a costly practice in the long haul.

The Intrusions of Bias

Policy thrives on certainty and sureness of purpose. Intelligence seeks to extract certainty from the uncertain.10 In the process, intelligence must oversimplify reality to some degree. Unfortunately, analysis based on too few variables is certain but can be erroneous. On the other hand, too much ambiguity in intelligence analysis opens the door for the nabobs of bias to play their role in policy formulation.11

Intelligence analysts and their policy clients are alike in their uniform tendency to see bias as a disease that infects someone else. Actually, bias is the common cold of all intellectual processes. None is immune.

I recall a roadside lunch in Vietnam where a lieutenant, a naive seeker of truth, inquired about the ingredients in a tasty stew--after he had consumed it. When told that the meat used in the recipe came from a small dog, the young officer became acutely ill immediately. His perceptions about the edibility of dogs were more important than the reality of a good lunch.

Biases have a tendency to overpower reality. Therefore, they need to be illuminated and controlled. A bias properly recognized can be used as an assumption. Such assumptions are more significant than methodology. Once stated, assumptions are often regarded as reality. The difference between an analyst and an advocate is not strength of logic but how each deals with assumptions and probabilities.

Bias in the policy community is likely to be personality-dependent. Bias in the intelligence community has institutional roots. Policy chefs also tend to overpower their ingredients,12 while intelligence ingredients tend to overwhelm their cooks.

The institutional roots of bias in the intelligence community are varied and can be attributed to the following: the focus of collection, the sheer volume of data, security paranoia, and analytical inertia.

Intelligence collection systems, especially sophisticated sensors, are focused on things quantifiable--or only on that which can be heard or seen. Technical means of collection can discriminate, count, measure, and catalog, but they don't qualify very well. The empirical strength of technical collection--an unprecedented ability to quantify accurately and rapidly--overshadows the systems' limited capabilities in qualifying or in retaining context. Technical collection tends to extract the measurable (e.g., weapon capabilities) from context and simply illuminate it in isolation. Although some might argue that analysis reimposes context, in practice, it is very difficult to reconstruct the context from which an elect piece of evidence has been drawn. (Genuine contextual analysis is an art lost by technology.) The very focus of collection has a great influence on analytical outcomes.13

Further, the sheer volume of raw data reported by collection systems often saturates, and frequently overwhelms, the analytical process. The wealth of unevaluated data encourages selectivity, not all of which is wholesome. It often forces analysts to ignore that which is difficult to process, and it encourages analysts to choose only evidence that supports their arguments. The volume problem also tends to obscure the distinction between reporting and analysis. A hard-pressed analyst frequently finds it expedient, and safer, to regurgitate data rather than to divine its significance. The system is further constipated as each intelligence headquarters feels compelled to publish "summaries" of the same summary reports published by other headquarters.

Similarly, security paranoia often excludes nonintelligence data and eliminates competing evidence. A premier Air Force intelligence facility, for example, prohibits analysts from bringing unclassified periodicals to their offices.14 Security is the stated purpose of the policy. How security is improved by banning incoming literature escapes most observers. Nevertheless, the net effect sends a clear message to analysts: Don't mix open sources with classified sources. Such practices reinforce the common, often erroneous belief that classified data are inherently more credible than that which are not. Misguided security is not much of a tradeoff for isolated analysis.

And finally, even blessed classified evidence is often abused by common inertia. Intelligence, like other disciplines, tends to do most often what it does best: quantify a few formulae with a few criteria. This type of analysis is safe, unambiguous, and, in many cases, useful, if not valid. Threat analysis, based on military capabilities alone, is an example of the mischief that this bias leads to most often. Analytical inertia tends to favor oversimplification.

One of the common outcomes of bias (and of ensuring longevity in the analysis business) is the production of worst-case scenarios. Although it may have something of an undeserved reputation as the exclusive distributor for worst-case scenarios, intelligence is seldom accused of wishful thinking. Pessimism has always been a safe course for analysis. If you predict the worst and nothing happens, your clients might raise a brow or two, but privately they breathe a sigh of relief. If events confirm gloomy forecasts, no one is happy, but your credibility and theirs are still intact. However, if your predictions are optimistic and things take a turn for the worse, stand by for a witch hunt. Columnist William Safire has pointed out, pessimism is a kind of no-lose hedge for all sorts of analyses.15

Traditional Images
of Intelligence

A significant number of the problems of the intelligence community and the policy/intelligence relationship stem from traditional images that have failed to keep pace with and reflect both the intelligence clients' desires and their needs for sound, relevant current information that may be useful.

the category problem

Traditionally, intelligence that comes to the policymaker is categorized as one of three types: term, current, or estimative.16 (See Figure 1.) A term intelligence product contains historical and encyclopedic information. As a rule, it looks to the past. Current intelligence primarily deals with that which is new or changing. It, for the most part, focuses on the present. Estimative intelligence tackles the toughest questions: forcecasts and futures.

Figure 1. Traditional Categories of Intelligence

These categories have remained unchanged since they were originally defined, but they have been bypassed by new requirements. Today's needs argue that the basic list of three should be revised or expanded to reflect a world that is both more complex and more rapidly changing than in the past. These new categories should include deception, warning, and military threat intelligence. (See Figure 2.)

Figure 2. New Intelligence Categories


By any measure, deception analysis is the most obscure and most unsavory intelligence task. It alone raises uncertainty to a threshold of pain. The mere possibility of deception assaults the policymaker and intelligence manager with equal vigor. Intelligence doesn’t want to be reminded that it can be led to bad judgment, and policy doesn't want to hear that it can be fooled--especially in retrospect. Unless it stands alone, deception analysis tends to be suppressed.

The policy lion and the intelligence ram both have a subjective tendency to dismiss evidence that doesn't fit their pet paradigms as "disinformation." Neither have expressed much interest in establishing a separate analytical discipline or criteria that might umpire the balls and strikes.


In the past decade, the complexity of weapon systems has expanded, while warning times have been compressed. Simultaneously, the spectrum of potential conflict has grown in ways that allow little time for the more deliberative, or traditional, intelligence methods to operate effectively. These changes have spawned the need for the new collection/analytical specialization of indications and warning intelligence. Traditionalists might argue that indications and warning intelligence is merely another facet of current intelligence. However, current intelligence often only addresses information needs, whereas tactical and strategic warning are matters of survival. The importance of timely warning coupled with specialized analytical and reporting needs suggests that this category of intelligence is unique.


Military threat intelligence is another modern category that could also stand alone for reasons of significance. It is difficult to overstate the complexity and ambiguity of threat intelligence. (See Figure 3.) The components of threat analysis not only are varied but require very different methodologies. The concerns of military threat are characterized by a rank order of understanding and analytical difficulty, compounded by an inversion of significance.

Traditionally, military threat has been treated as a subset of estimative intelligence. The military dimension of threat now contains such unique lethality, however, that it is a curiosity not to see it in a class by itself. Further, the art of threat analysis has long been a victim of oversimplification. Collection and analysis for the more uncertain components of military threat could surely benefit from special illumination.

Figure 3. Components of Military Threat

the process problem

The "category" inadequacies are further aggravated by the rather bizarre image of the intelligence process. The traditional symbol for this process is a circle,17 or cycle, which begins with requirements that generate collection that provides grist for processing and analysis, which produce outputs that are then disseminated to clients who finally close the loop by generating more requirements.

This cycle metaphor, depicted in Figure 4, is among other things an oversimplification. Surely there are different weights of effort and investment of treasure in the various slices of the pie. But the sense of the metaphor is accurate. Intelligence is portrayed as a closed system which, some critics uncharitably suggest, feeds on itself. A circular matrix that appeals to intelligence engineers may not be a view that charms senior managers and policy analysts. The policy lion does not jump through hoops gladly, nor does he like to think about occasions where he might chase his tail. An intelligence system with a circular image may be an unfortunate choice to serve a policy system with a sense of purpose. Policymakers thrive on linear images that lead them toward goals--preferably their policy. They are not fond of circular logic that appears self-serving and directionless.

Figure 4. Intelligence  Processing Cycle

The circular image of intelligence also perpetuates some unfortunate mythology about the role of requirements. It suggests that the policymaker, after viewing the intelligence product, closes the intelligence loop with some definitive statement of satisfaction--or identifies new requirements. This loop-closing simply does not occur in most cases. For the most part, intelligence managers alone play the requirements "game." In practice, they are often unaware of needs of policy, yet the requirements flow continuously. More frequently than they would care to admit, intelligence personnel are kept in the dark by hidden agendas, security considerations, or the more understandable discontinuities of changing administrations.

There are times when intelligence is simply not informed of a policy action until after the fact. Worse still, as Hans Heymann reminds us, the stated objectives of policy, when known, are often not the real objectives.18 Further, the security walls around policy are often more impenetrable than those of the intelligence community. Policy cliques are purposely kept small and exclusive. Even edicts of record are highly classified and/or sparsely circulated. The problem is regularly aggravated between administrations, when classified and unclassified policy papers are uprooted, dispersed, or buried in presidential libraries.19

As a system, intelligence usually has more continuity than policy, Yet, that intelligence anticipates the needs of policy at all is probably due more to intelligence archives and good guesswork than formal feedback. Still, the intelligence ram traditionally plays the goat in disputes over requirements. In 1973, for example, one of the stated reasons for disbanding the Board of National Estimates was that it was unresponsive to policy requirements.

Needed: A New
Image of Intelligence

The circular image of intelligence is no doubt an insider's perspective, the result of a fascination with mechanics and the sophisticated gadgetry of collection and processing. Unfortunately, this image fails to distinguish between the necessary components for production and the desirable components of outcomes. A focused image of the system more compatible with client concerns and current needs might be an open-ended linear matrix, of inputs and outputs, which conveys a sense of direction. Such a new image could reflect the needs of clients and the substance of useful intelligence: relevant expertise, sound analysis (coupled with integrity), and effective communication. (See Figure 5.)

Figure 5. The Intelligence System: A Client's View


On the input end of the intelligence process, the only truly relevant substance (for analysts and clients) is new information, the context in which it is set, and any background that history might provide. Intelligence purists might argue that the powerful roles of collection and data bases are slighted in such a reduction.20 Not really. The esoterics of collection are to the analyst and policymaker what logistics are to the battlefield commander. Neither worry much about the minutiae of acquisition as long as the material is sound, enough is available, and it gets where it needs to be on time. The question here is not one of significance but emphasis. Collection and data bases support the adequacy of expertise. Expertise is the first major threshold that intelligence must cross for the client. How one gets there is more a question of mechanics than substance.


Assumptions and methodologies are the substance of analysis. Both must be explicit and defensible. Analysis is, in turn, the head and the heart of the intelligence process. The head addresses the rational needs of inductive or deductive logic, and the heart speaks for intuition and integrity. Many rationalist methodologies ignore the nonrational and moral elements of analysis, but they do so at the risk of excluding precipitous insights and inviting suggestions of duplicity.

Intuition and integrity play their strongest role in any statement of assumptions. Clearly, assumptions fall into the intuitive realm, as they are notions, or suppositions that something is true. If they were proved, they would merely be grist for the rational mill. How assumptions are used defines the real integrity of most arguments. Unstated or imbedded assumptions are the most worrisome. No inductive or deductive logic can overcome the mischief of unwarranted intelligence assumptions.

What an analyst or policymaker does not know will never cause as much trouble as that which he thinks he knows but which isn't so. This dilemma is standard issue when bias weds unwarranted assumption.

On the rational side of analysis, there is a striking similarity in the relationships between intelligence and induction and between policy and deduction. Intelligence is concerned with inducing evidence, or reasoning from parts to a whole. Policy is concerned with generalities, or reasoning from a whole to its parts. If the assumptions are used with integrity, intelligence fairly looks for policy to serve, while policy fairly looks to intelligence for evidence.


The tough part of intelligence analysis is trying to do the task with integrity on policy that the policymakers regard as clearly their turf. Most intelligence functions come under the explicit control of powerful policy lions. The situation is acute in the military. Here the intelligence ram is outgunned, outflanked, and outranked.

Service headquarters are illustrative. Most support staff elements under the Chief of Staff are "deputies," but the intelligence element is invariably only an "assistant." Rank disparities also reinforce subordination. Intelligence flag officers are commonly allocated at least one star less than other support counterparts.

At lower echelons, intelligence functions are understandably subordinate to troop commanders. In many cases, they are further subordinated to other staff elements, such as operations. Here the stature and rank disparity is likely to be even greater. Intelligence is often a junior officer, while other staff sections are led by field-grade ranks. Intelligence managers often compound the problem by manning their higher headquarters at full strength with the best analysts and letting operational units fend for themselves. Traditionally, combat units have the greatest number of junior and inexperienced intelligence officers.

Policymakers frequently admonish intelligence to be independent.21 However, they do little to underwrite integrity in bureaucratic hierarchy, eminence of position, grade structures, or manning equity. In a hypothetical conflict between policy advocacy and intelligence objectivity, the military intelligence officer not only is outgunned, outflanked, and outranked but also faces a spectrum of choice that runs from bad to awful. When evidence and argument fail, he has four choices: resignation, public confrontation, capitulation, or bureaucratic subversion. Resignation and public confrontation are similar for their probable outcome, career suicide. Neither is very likely or realistic, as each represents a choice between integrity and livelihood. In a real impasse, capitulation and bureaucratic subversion become more attractive and less wholesome. The purity of analysis, under fire, may rely more on policy temperance than intelligence integrity.

Testimony during the Westmoreland versus CBS trial provided some insights on this problem. In late 1967, General William C. Westmoreland's Chief of Intelligence in Saigon tried to surface evidence that would have increased enemy strength figures above those generally accepted in Washington. Westmoreland apparently rejected the new figures at the time as "politically unacceptable." Order-of battle bookkeepers, in turn, were told to trim their figures. Commanders, like Westmoreland, often see assessments from intelligence in the same vein as reports from other staff elements--i.e., simply papers to be accepted, rejected, or revised. Unfortunately, intelligence judgments that are changed at the whim of commanders are often accomplished at the expense of truth. For General Westmoreland, the truth became apparent on 30 January 1968 when the Communists launched an unprecedented countrywide offensive that was to change the course of the Vietnam War.22

The lion also intimidates the ram in less subtle ways with an occasional direct assault on analytical institutions or analytical criteria themselves. As cited earlier, in 1973, the Board of National Estimates and its staff was disbanded by William Colby, presumably on orders from the Nixon/Kissinger White House.


If analysis is the head and heart of the intelligence process, then communication is the voice and also the cutting edge of intelligence. Here policy is served well or not at all. The best intelligence, poorly communicated, is worthless. Communication, in the best sense, is a verb (action), not a noun (medium). The methods of intelligence communication are writing or speaking. The purpose of communication is impact--always a noun. Writing and speaking are the actions of intelligence communication; impact is the desired accomplishment.

The written word. As the intelligence officer seeks to influence the policymaker with the written word, there are probably only three axioms worth remembering. First, all that will ever be known about any report's routing is what office received it, not who read it. Second, the readership of any report is probably inversely proportionate to its length. Finally, if the report is written the way most government reports are, it hasn't got a prayer of having impact. Easy reading demands painstaking care in writing (wordsmithing for precision, conciseness, and clarity), but few intelligence authors take the trouble.

A National Security Council staffer once observed that there were only two, possibly three, types of reports that are read on a regular basis in Washington: point papers, the editorial pages of select dailies, and cartoons. It would be difficult to confirm the observations on the first two categories of reports, but the third is a cinch. The walls of the offices of most bureaucracies are papered with cartoons. Staffers not only read cartoons but cut them out and hang them in a place of honor to be savored indefinitely. No one has ever been observed nailing a 500-page national estimate to the wall of any office.

It is no accident that intelligence reports are "distributed" and "disseminated," while intelligence briefings are "presented." The difference is all the difference.

Policy lions are always on the move and have little time for sedentary pursuits, least of all lengthy reading. Nevertheless, their offices are awash in paper. The sheer volume of reporting represents a problem for analyst and decision maker alike.23 As the volume of available paper goes up, so does the likelihood that briefings will play a larger part in the decision process.

The spoken word. The pile of written analyses and options gets reduced prudently or arbitrarily. This reduction most often comes in the form of a briefing, which is red meat to the real policy lion. Here he and the intelligence ram will be face to face in the same cage. In this arena they will be at their symbiotic best or fratricidal worst. It is possible to move the policy broker with a phone call, a shout in the hall, or even with a quiet chat over lunch. But when all those are done, he will probably still need or request a briefing. In the briefing room, rams are separated from lambs and, as Barnum observed, some of the lambs are lost.

For intelligence, briefings may be ubiquitous, but they are also inevitable. Briefings, like no other form of communication, have captive audiences. Here the interaction of policy and intelligence is assured, for good or ill; and intelligence never gets off this treadmill. Even when there is nothing to say, the briefings go on. If no other reason for them exists, they will demonstrate and justify continued funding of collection technology.

The Intelligence Phalanx

The personalities that inhabit the intelligence world often mirror the system itself. The principals, especially those on the cutting edge, usually fall into one of four categories: managers, experts, analysts, or communicators. In the policy corridors of the Pentagon, it is common to see such a group advancing as a unit toward the den of some policy lion.

The intelligence phalanx is an impressive sight. The wedge is usually led by some bemedalled panjandrum who confidently strides in the point position.24 The manager is usually flanked by a somewhat junior (but always impeccably groomed) briefer. The communicator, irreverently known to peers as "talking dog," invariably carries a chrome divining rod strikingly like an antenna off a 1959 Buick. It is, in fact, a collapsible pointer--the omnipresent baton that binds formal arguments.

This trio is followed by one or two nervous experts. Experts are those intelligence specialists who command a high degree of knowledge in some specific discipline. Through long years of experience or study, they either know much about a specific topic that has little breadth, or, equally important, know what data base to tap. These data specialists are often mistaken for analysts and seldom volunteer to correct the confusion. Experts are known in the trade as "backup. " They are trained to respond to nothing save direct questions.

Analysts seldom travel with the phalanx. It is too dangerous, and they are too valuable to put at risk. Analysts usually know both sides of the impending argument. Worse still, the best among them are inclined to volunteer relevant but ambiguous judgments at just the wrong moment.

The rear of march is usually brought up by a couple of acolyte briefers who are affectionately known as "flippers." They are burdened with heavy satchels, not unlike ammunition boxes of old. The function of flippers is to force-feed their precious cargo of visual aids into projectors on command from the briefer.

As the lights dim in some cool and windowless Pentagon inner sanctum, an eerie chill of collective déjà vu sometimes sifts through the silence. The colors that dance across one wall in the dark are reminiscent of what? Plato's cave shadows? Ancient rock drawings etched by flickering candlelight? The sketches and pictures are the final reduction of the best and most complex technology known to man. The output, the cutting edge, the billion-dollar petroglyph of unprecedented collection and analytical empires, is an eight-by-ten-inch acetate cartoon! Upon such, the fate of nations rests.

Such observations do not trivialize or demean the wealth of energy and treasure invested in the intelligence system. They merely recognize the primitive reality in that final and crucial step in the process: after thousands of years of socialization and technological achievements, mankind's modus operandi remains unchanged. Power and persuasion still retreat to darkened caves to put the handwriting on the wall.

Perhaps thousands of years hence, archaeologists will sift through the debris of our civilization. Certainly they will recognize the bones of the policy lion and the intelligence ram, and they are likely to view the computer as the definitive artifact of our generation. But when they come upon some faded vu-graph in the pile, will they recognize it as the billion-dollar petroglyph?

OVER time, policy and policymakers will change. However, the characteristics of a successful policymaker will probably remain fairly constant. Intelligence officers cannot afford to be unable to distinguish between the flux of policy and those bedrock predispositions that always characterize the corporate policy environment. Solid evidence and strong argument alone will not win the day. Effective intelligence must also understand and overcome those fixed obstacles to persuasion that exist independent of specific policy and intelligence support.

Intelligence must also recognize and explicitly deal with a thicket of biases, many of them generated by the structure of the intelligence system itself. Today the system is front-loaded with a complex and prolific collection technology that warps focus and frequently overwhelms the very clients it is designed to serve. This volume and complexity problem is compounded by a dangerous tendency to view unclassified data and analyses as inherently inferior and, hence, to ignore them.

The great irony of the national security decision process is that intelligence will always have the potential to make itself irrelevant. Nonetheless, with or without sound intelligence, the policymaking process is likely to march on. Thus, the initiative for improving the impact of analysis lies with intelligence. Unfortunately, in recent years intelligence managers have emphasized inputs, the gadgetry of collection, and technical processing at the expense of outputs.

As a consequence, the output of intelligence may not be as effective as it could or should be. The categories of intelligence products are outdated--a factor that may also skew the focus of collection and analysis. In the warning arena alone, the continued categorization, functionally and institutionally, of indications and warning intelligence as a subset of current intelligence misplaces the emphasis. Placing "warning" under "current" is like classifying nuclear war as another form of friction.

The image of the intelligence process is sorely in need of repair also. The self-serving cyclical metaphor needs to be replaced with a more dynamic image that conveys a sense of direction. Beyond image, the communication phase of the intelligence process needs to be streamlined and revitalized. Written products are both too numerous and too voluminous, creating logjams within the intelligence system and encouraging indifference without.25 Simplistic briefings may be inevitable, but the billion-dollar petroglyph need not be the only currency between the lion and the ram.

Finally, the stature of intelligence functions and managers needs to be elevated. For intelligence, the war is now! If warning and threat analysis is effective, the actual shooting may never start. If the voice of the intelligence warrior is to be heard, the speakers must enjoy some stature beyond that of supply officer,

Washington, D.C.


1. Yehoshafat Harkabi, "The Intelligence-Policymaker Tangle," Jerusalem Quarterly, Winter 1984.

2. Thomas L. Hughes, "The Fate of Facts in a World of Men," Foreign Policy Association, Headline Series No. 233, December 1976.

3. Richard K. Betts, "Analysis War and Decision: Why Intelligence Failures Are Inevitable," World Politics, October 1978.

4. Apologies to Barnum. His original homily spoke of lambs.

5. Hans Heymann, Jr., "The Intelligence-Policy Relationships: From Arms Length to Love Hate," an address at the U.S. Air Force Academy, Colorado Springs, Colorado, 6-7, June 1984. Many of the observations in this article on the nature of the policymaker were extrapolated from this address and a series or lectures presented by Heymann at the Defense Intelligence College during the summer of 1984.

6. The masculine pronoun is used here and elsewhere in recognition, however unfortunate, that most government policymakers are men. Were they not, their characteristics might, hopefully, be different.

7. Lincoln P. Bloomfield, The Foreign Policy Process (Englewood Cliffs, New Jersey: Prentice-Hall, 1982), p. XIV.

8. The director of CIA, while not a member of the cabinet per se, has been elevated to cabinet rank recently.

9. Committee for State Security (KGB) of the Soviet Union.

10. Betts, op. cit.

11. Ibid.

12. Former national security advisors Henry Kissinger and Zbigniew Brezezinski are two of the most prominent in recent memory.

13. Collection, in a sense, randomly manufactures puzzle pieces that analysis is charged to assemble. Unfortunately, there are always more pieces than pictures. Intelligence analysis labors inductively, reasoning from many parts to too few elusive wholes.

14. Newspapers are generally forbidden at the Foreign Technology Division of the Air Force Systems Command. Other unclassified periodicals need to be specifically deemed (by supervisors) as "work-related."

15. William Safire, "On Wishful Thinking, " New York Times, 30 August 1984.

16. Sherman Kent, scion of the old Board of National Estimates, defined original categories of "modern" intelligence.

17. Amos A. Jordan and William J. Taylor, Jr., American National Security Policy and Process (Baltimore, Maryland: Johns Hopkins University Press, 1981), p. 131.

18. Heymann, op. cit.

19. National Security Council staffers must frequently trek to the presidential libraries to search for documentation that provides the threads of policy continuity.

20. Today, it would be very difficult to slight the role of collection. The collection cart has been pulling the analytical horse for at least two decades. Clearly, there is more reputation and treasure to be made in buying and selling collection gadgets than there is in improving the outputs of analysis.

21. Senator Daniel Patrick Moynihan, "Only the Brave Risk Intelligence," Commencement Address at the Defense Intelligence College, Washington, D.C., 15 June 1984.

22. Connie Bruck, "The General v. CBS: Rush to Settlement,"Washington Post, 7 April 1985, p. Fl. 23. In the defense intelligence system, there is some control over production through the Delegated Production Program but no similar program controls reporting.

24. Moynihan, op. cit.

25. Roberta Wohlstetter of the Rand Corporation identified the intelligence need to separate the "relevant" from the "noise" more than two decades ago in her analysis of the Pearl Harbor surprise attack. Thus far, intelligence has managed only to increase the volume of its own noise.

It would be absurd to Claim that technological development by itself could rid the world of nuclear weapons. The two primary agents for abolishing nuclear weapons must be international negotiation and the aroused conscience of mankind. But the success of negotiation and moral indignation in bringing about nuclear disarmament will also depend upon technical factors. We will have a far better chance of achieving nuclear disarmament if the weapons to be discarded are generally perceived to be not only immoral and dangerous but also obsolescent. An intelligently conducted arms race, leaving nuclear technology further and further behind, could help mightily to sweep nuclear weapons into the dustbin of history.

Freeman Dyson
Weapons and Hope, p. 41


Lieutenant Colonel G. Murphy Donovan (B.A., Iona College; M.S., University of Southern California) is Chief, Intelligence Research Division, Air Force Intelligence Service, Washington, D.C. In earlier assignments, he has served with Tactical Air Command, Strategic Air Command, the United Nations, and the Defense Intelligence Agency. He also has been a Senior Air Force Research Fellow at the Rand Corporation and has had overseas tours in Vietnam and Korea. Colonel Donovan is a graduate of the Defense Intelligence Postgraduate School and Air War College.


The conclusions and opinions expressed in this document are those of the author cultivated in the freedom of expression, academic environment of Air University. They do not reflect the official position of the U.S. Government, Department of Defense, the United States Air Force or the Air University.

Home Page | Feedback? Email the Editor