THE INTELLIGENCE PROCESS A DIGEST FROM STRATEGIC INTELLIGENCE BY SHERMAN KENT

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78-04718A000600100003-3
• Release Decision: RIFPUB
• Original Classification: K
• Document Page Count: 27
• Document Creation Date: December 9, 2016
• Document Release Date: May 18, 2000
• Sequence Number: 3
• Publication Date: January 1, 1949
• Content Type: REPORT

Approved For Release 2000/08/28: CIA-RDP78-04718A00060010
INTELLIGENCE
PROCESS
A Digest from
Strategic Intelligence
by Sherman Kent
To: TS S C
Copyright 1949
PRINCETON UNIVERSITY PRESS
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
Chapter 1
In the language of the trade, the word intelligence is used not merely
to designate types of knowledge. It is also used as a synonym for the ac t t v i t y
of an intelligence organization. I will discuss in this paper, intelligence as
activity, or better, as process. My primary concern will be the large num-
ber of methodological and other problems to which this process gives rise.
But before discussing these problems, I shall briefly discuss the intelligence
process itself.
The knowledge, which I call strategic intelligence, serves two uses:
it serves a protective or defensive use in that it forewarns us of what other
powers may be hatching to damage our national interests; and it serves a
positive or outgoing purpose in that it prepares the way for our own foreign
policy or grand strategy. Here the important thing to grasp is that, no mat-
ter what its diversity, this knowledge is produced by the process of research.
Sometimes research is formal, highly technical, and weighty; some-
times it is informal, untechnical, and speedily arrived at. Sometimes a re-
search project requires thousands of man-days of work, sometimes it is done
in one man-minute or less.
This research process, in strategic intelligence, is initiated in two
chief ways. When the policy people or planners of our government begin
formulating something new in foreign policy they often come to intelligence
and ask for background. (They should do this more than they do.) In their
request for this or that block of knowledge, they stimulate the intelligence
force to embark upon a piece of research and a course of specially directed
observation or surveillance. There is, however, a second way in which this
process is initiated. This is through the systematic and continuing surveil-
lance activity of the intelligence staff itself. So important is this surveillance
that it is often conceived of as separable from research. I do not think it
should be.
Surveillance, as I am using the word here, is the observation of what
goes on abroad and the deliberate attempt to make sense of it.
Thus, in foreign countries we carry on through a multitude of open-
and-above-board personnel -- some civilian, some military -- whose duty
is to keep eyes and ears alert and report what they learn. They are the for-
eign service officers and attaches. Each of them has his field of special in-
terest and competence, whether itbe political, military, commercial, or cul-
tural, etc., and each is supposed to keep himself and his principals at home
posted within this specialty.
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
Some foreign governments supplement the work of their overt officers
with espionage activities; that is, they send out secret agents,- or undercover
recruiters of secret agents, to discover and report on matters which would
be difficult to discover overtly. If you would like a sample of how such ac-
tivities are established and operate, read Richard Hirsch's "The Soviet Spies",
or the "Report of the Royal (Canadian) Commission"...upon which it is large-
ly based.
Not all surve'.illance activities take place abroad; some very impor-
tant ones take place at home. Queer as it may seem to observe a foreign
country from a home observation post, there are several reasons for this
paradox.
First, there must be surveillance at home purely and simply for con-
venience. For example, what the official French radio beams on the rest of
the world is of considerable interest to us; we should like to know the content
of its political news and commentary. But this can be done easier domesti-
cally. Hence, tht extremely important surveillance organization known as
the Foreign Broadcast Information Branch is established at home. Its moni-
toring; stations pick tip the most significant programs; the home office tran-
scribes them, translates (and sometimes abstracts them), reproduces them,
and sends them around to officers of the governments. Departmental, i.e.,
Army, Navy and State intelligence organizations are, of course, the chief
beneficiaries.
Sharp newspaper foreign correspondents though they have no con-
nection with intelligence work, are important observers of foreign affairs
and important, though inadvertent, contributors to the surveillance activity.
Wise is the government that does not intercept their dispatches at point of
origin, but lets them land in the home cable rooms of our domestic papers,
there to put the content to official use. Doing business this way means that:
intelligence engaged in overt surveillance must have some small force at
home which follows the best foreign news.
There is a second reason for home surveillance. It is based upon the
proposition that anything being hatched abroad to our detriment is conspira-
torial;, i.e., it is hatched in secret; but there are several people or groups
party to it. In the world of international relations these parties to the con-
spiracy may be residents of half a dozen countries, and the story of what they
are up to, must be pieced together from fragments supplied from the half-
dozen different sources. For example, what Franco was considering at a
given moment might be less available from Madrid than from Mexico City,
Buenos Aires, Lisbon, Bayonne, and Rome. This is not to argue that Wash-
ington is the only place where surveillance should take place, but it is to
argue that given the complicated nature of the modern world, there must be
a listening and observation post and clearing house in a central spot.
However conducted --overtly or clandestinely, abroad or at home --
surveillance, as it goes about uncovering a policy or an action hurtful to our
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
national interest, stimulates our desire and need for knowledge. Thus it stim-
ulates the production of defensive-protective knowledge and tells us what we
must know about affairs abroad if we are to implement our own active outgoing
policies. It is surveillance, then, which produces that wide range of phenom-
ena without which strategic intelligence would have little content of current
importance.
In talking of surveillance there is always the danger of portraying
something passive. Surveillance sounds like sitting back and awaiting the
impression. But surveillance worthy of the name must be vigorous and ag-
gressive; aggressive in that the observer covers as much ground as possible,
seeking to expose himself to a miximum number of phenomena; and more im-
portantly, it must be aggressive in that the observer does a maximum amount
of following up his impressions of these phenomena.
So long as I use the imprecise term "following up" I am on safe ground
with the general reader and intelligence brotherhood. It implies checking the
accuracy of sources, comparing divergent accounts, and gaining perspective
by broadening the field of inquiry, finding new leads- -out of which emerges a
proposition which seems the truest of all possible propositions. Now I would
like to call this process of following-up by the more precise term of "re-
search" and say that it must accompany the surveillance activity. Such re -
search consists of a systematic endeavor to get firm meaning out of impres-
sions. Lacking it, surveillance will produce spotty and superficial informa-
tion.
Research has greater importance than merely supplying the cutting
edge to surveillance. It has a role entirely its own. In wartime, research
produces the knowledge of enemy strategic capabilities, enemy specific vul-
nerabilities; it produces knowledge of the political and economic strengths
and weaknesses of the enemy; and knowledge of the physical plant which the
enemy is using. On such knowledge our own offensive military plans are
based. In peacetime, research produces that knowledge of foreign lands which
you would wish if you had to decide whether to sponsor a European economic
recovery program; and then defend it before Congress and your fellow coun-
trymen.
Research is the only process which we of the liberal tradition are
willing to admit is capable of giving us the truth, or a close approximation
to truth. We insist, and have insisted for generations, that truth is to be ap-
proached, if not attained, through research guided by a systematic method.
There is such a method in the social sciences, which largely constitute the
subject matter of strategic intelligence and in which, for the purposes of this
paper, I include the science of military strategy. The method is much like
that of the physical sciences. It can be described, for instance, by paraphras-
ing the discussion of physical sciences as set forth by President Conant of
Harvard. One could say that the method of the social sciences involves the
development of new concepts from observations and that these concepts in
turn indicate and lead to new observations. But perhaps this is less useful
Approved For Release 2000/08/28 CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
than to spell out seven steps or stages in the research process specifically
designed to meet the requirements of strategic intelligence. They are:
1. The appearance of a problem requiring attention of-a strategic in-
telligence staff.
2. Analysis of this problem to discover which facets' of it are of im-
portance to the U. S. and which of several lines of approach are most likely
to be useful to its governmental consumers.
3. Collection of data bearing upon the problem as formulated in stage
2. This involves a survey of data already at hand and -available in the librar-
ies of documentary materials, and an endeavor to procure new data to fill in
gaps.
5. Study of the evaluated data with the intent of finding some sort of
inherent meaning. The discovery of such meaning can be called the moment
of hypothesis. In reality there is rarely such a thing as one moment of hy-
potheses nor can it be said categorically at what moment hypotheses appear.
One would wish that they appeared here at respectable -stage 5, but in actual
practice they begin appearing when the -first datum is collected, and may
appear long after the -project is closed out.
6. More -collecting of data along the lines of the more promising hy-
potheses, to confirm or deny them. -
7. Establishment of one or more hypotheses as truer than others and
thus as the best present approximations of truth. This is the last stage and
is often referred to as the presentation state.
At each of these stages two problems arise. One is characteristic of
all systematic research in the social sciences, the other derives from the
peculiarities of intelligence's research activities. My principal concern in
the next chapter will, be with these latter methodological problems.
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
Chapter 2
Before proceeding with an analysis of the problems in intelligence, I
would like to clarify my use of the adjectives methodological and substantive.
By "methodological" I mean a problem characteristic of the method of trying
to establish a new approximation to truth. By "substantive" I mean a problem
in the subject matter of strategic intelligence. As an example of a "sub-
stantive problem" consider the strategic stature of the Chinese Communists;
as an example of a "methodological problem" consider the means you would
employ to get the basic data on the Chinese Communists' military establish-
ment.
1. Stage One, the appearance of the substantive problem
The substantive problem in strategic intelligence can emerge in three
principal ways.
a. It may emerge as a result of the reflections of a man employed to
do nothing but anticipate problems. In actual fact, the intelligence business
employs too few such men. Their job is to ask themselves the hard, the
searching, and the significant question and keep passing it on to professional
staff. An intelligence operation should be bedeviled by such questions, and
a substantial part of its work program should be concerned with getting an-
swers. A Pearl Harbor maybe ascribed in no small measure to the absence
of that unpleasant and insistent person, who, conscious of the growing animus
of Japan, would have kept asking when is the attack coming, where is it coming,
and how is it coming?
The methodological problem involved here is a slight one, on the sur-
face. It consists of devising means by which the question-askers will be
sure of formulating good substantive problems. The only answer lies in
picking a man who knows the area in which he is supposed to ask questions,
who has an inquiring mind, and then see that he has ready access to every
scrap of incoming evidence, access to everyone who knows about it, and free-
dom from other burdensome duties.
b. The substantive problem may emerge when surveillance makes
one aware of something unusual. For example, suppose the people watching
Great Frusina (my hypothetical country) learn it is expanding its Christian
mission program in the Belgian Congo and that it has named Brother Nepomuk
as aide to the new director. If surveillance is sharp enough to recognize
this shift in a minor Great Frusinan policy it has initiated a substantive prob-
lem which may be very important when followed-up; or be of no importance
at all.
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
The methodological problem here is similar to that just touched upon;
how can surveillance be sure of putting the finger on the three unusual and
potential things per week out of the thousands it observes and-the millions
that happen? The answer is again: procure wise men, wise in the subject,
and pray that they w'.ill produce hypotheses of national importance.
c. The last way in which the substantive problem may emerge is at
the request of the consumer. For example, let us suppose that the policy
people, who are the prime intelligence consumers, are facing a revision of
the established China policy. They summon some of the control and profes-
sional staff of intelligence to a meeting where the problem is put on the table.
In this meeting aspects of the. China question will appear which the policy
people have not considered before. Let us assume that they have to do with
population. A prospective change in policy has caused a substantive problem
to emerge.
Here there is no methodological problem. Intelligence feels things
have gone just as they should. True, the assignment is so large and general
that serious difficulties are presented but since intelligence was at the meet-
ing, it may assume further guidance from the consumers in shaping the sub-
stantive problem to their needs. But what usually happens is that decision
to revise policy is taken without intelligence. Weeks later, when the policy
people are up against a deadline, they discover they need a new population
estimate and that at once. They pose a. problem all right, but they pose it
to the consternation of intelligence, which is asked to do a month's work over
night?
2. Stage Two, the analysis of the substantive problem;
The substantive problem has at last emerged in rough form. It must
now receive some close and searching analysis. The aim is not merely to
discard what is irrelevant or unimportant, but to shape the problem toward
solution.
For example, to return to Brother Nepomuk, the surveillance people
have many courses of observation opened to them by their discovery of Great
Frus:ina's new missionary zeal. They can begin watching the church-state
relationship looking for new angles; they can start an observation of the
Great Frusina-Belgium relationship; they can skip over Great Frusina,
Belgium, and the Congo, and -start chasing after developments in the general
field of missions to find new church policies therein. They are almost cer-
tain to turn up interesting leads no matter what lines they pursue. But the
question is, what particular line of observation is likely to prove most im-
portant to the security of the United States?
The research people back from the policy -on- China meeting are posed
a somewhat similar problem. They were asked to come up with some popu-
lation data; no more explicit than that. Obviously there are dozens of kinds
of population data. But only one or two will have bearing on the task of the
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
policy people. What are these data, and in what detail should they be worked
up?
As the surveillance and the research people search for the most
fruitful line of attack they will seek guidance. This guidance should come
both from their own inner selves and from the policy, planning, or operating
people whom they are endeavoring to serve. Let me take the problem of
guidance as it appears to the surveillance man.
He discovered that Great Frusina was enlarging its Christian mis-
sions program in the Congo; he knows that the Congo has large uranium de -
posits; he asks himself, is there a connection ? Then his foray into research
reveals that Brother Nepomuk won a Nobel prize for work in geology. He
now sees a connection and has uncovered the most frut t fu l l the of attack. He
now has a hypothesis that Great Frusina is trying to get uranium from the
Congo and that Brother Nepomuk is a Great Frusinan agent. At this point
he must get outside guidance. What other lines of attack will the people
whom he serves designate as fruitful; what do they propose to do if such and
such a line confirms an ill-intentioned activity on Great Frusina's part?
The sequence maybe exactly reversed with the research people work-
ing on the population of China. They will promptly go back to policy and ask
advice about lines of attack. They will also ask how the policy people see
the task shaping up, and what their aim is in revising the old policy. If they
get answers they can state the substantive problem and answer it in a way
which will have practical utility to policy. Moreover, as research advances
their study will get useful hypotheses which spring from familiarity with the
subject matter, and which the policy people might never have got on their own.
The methodological problem here is not that of inner guidance but
guidance from the users of the knowledge. Here is one of the critical prob-
lems of intelligence, i.e., the relationship between its producers and con-
sumers. Intelligence often finds it impossible to get that guidance which it
must have if its product is to be useful. One of the places where this lack
of guidance produces its most disastrous results is right here at stage 2 of
the intelligence process. Unless the intelligence organization knows what use
its product is designed to serve, and what sorts of action are contemplated
with what sorts or implements, the analysis and proper formulation of the
substantive problem suffer.
3. Stage Three, the collection of data
The collection of data is the most characteristic activity of intel-
ligence. There can be no surveillance nor research without the collection
of data. Accordingly, an intelligence organization cannot exist until it does
a broad and systematic job of collecting. But in this very task lie metho-
dological problems so tough that they are a perpetual source of inefficiency.
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
a. One such problem is that which a member of the professional staff
encounters when he embarks upon a piece of research. He has blocked out
his substantive problem. His nextstep is to see what data bearing upon the
subject exist in his own and other intelligence organizations. Let us assume
that his own files a2re in good shape and that his outfit has a centralized
library of properly indexed documents. The materials from his own sources,
so to speak, indicate, as will also his horse sense, that there are other kindred
materials in neighboring intelligence organizations. He must reach these, but
there are real difficulties if (1) he must work through a third person in his
own organization who has an exclusive mandate to -collect data, and (2) if the
other organizations possess no central library of indexed documents. Unfor-
tunately, many intelligence organizations raise these considerable barriers.
b. Now, let us assume that the staff member discovers that even after
canvassing every resource in his headquarters city there are -still a number
of unanswered substantive questions which he must explore. He must com-
municate with the field; he must try to explain to someone in a foreign capital
what he wants. Now if the man on the other end of the wire was formerly a
worker in the home office, has a feel for home-office functioning, and per-
sonally knows the staff, he then will more readily understand what he is be-
ing asked to do and will do it with efficient good grace. He will quickly
grasp the instructions (which can be given in office shorthand) and will act
pretty much as an overseas projection of the home staff. But if he has not
served in the home office, and instead has gone to his foreign post improperly
briefed on home problems then there may be difficulties.
It is not easy to explain in a letter or cable precisely what is desired
to someone who starts from scratch. Requisitions of this 'sort must be
spelled out in detail and to achieve results they must communicate in their
substance a sense of urgency and importance. Such a requisition is time-
consuming. If it is no more than a blunt command it is likely to be handled
in a perfunctory fashion.
The problem increases when the recipient is a stranger to the sub-
ject. The home office may wish to have a foreign official interviewed on a
technical demographic matter or wish to have someone audit and report on
a scientific congress. But the men in the field may have had the wrong kind
of professional training or no professional training at all, and thus be totally
incompetent to handle such a request. Or, most likely, the field staff is com-
pletely engulfed in making good on a previous request which seems to them
to be of highest importance.
The foregoing problems of collection are not too formidable because
simple good sense can probably beat them. But there are other problems not
so easily disposed of and they are inherent in the surveillance phase of intel-
ligence.
The surveillance force is supposed to watch actual, fancied, or po-
tential ill-wishers or enemies of the United States and report on their
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
activities. It is also supposed to procure information which, though possibly
less dramatic, is none-the-less calculated to forward the success of our
own policies. In both areas the surveillance force must work clandestinely,
or it could not deliver on a small but extremely important part of this task.
Thus a certain fraction of the knowledge which intelligence must produce is
collected through highly developed secret techniques. Herein lies the major
methodological problem of the collection stage of intelligence.
It begins with the segregation of the clandestine force. This segre-
gation is dictated by the need for secrecy. The minimum of people must
know anything about the operation, and the greatest caution and dissimulation
must attend its every move. But unless this clandestine force watches sharply
it can become its own worst enemy. For if it allows security to cut it off from
all guidance, it destroys its own reason for existence. Guidance, in the nature
of things, should come from the ultimate consumer directly or indirectly, i.e.,
through the overt part of intelligence to which the consumer has gone for help.
As this relationship is stopped down (as it may have to be for long periods);
as it becomes formalized to the point where communication is by the written
word only; as it loses the informality of man-to-man discussion, some of its
most important tasks become practically impossible. Requisitions become
soulless commands. The consumer may ask for something the organization
is not set up to deliver, or he may ask for so wide a range of information that
the resources of the organization would be fully deployed for months, or he
may ask for something which the clandestine force knows is not worth the
effort. With a high wall of impenetrable secrecy the consumer has great dif-
ficulty in not abusing the organization, and the organization has equal dif-
ficulty in shaping itself along lines of greatest utility for the consumer. It
is constantly in danger of collecting the wrong and not collecting the right in-
formation.
Clandestine intelligence involves highly complicated techniques: the
correct approach to a source, its "development," its protection once de-
veloped, the security and reliability of its own communications, and so on.
Isolated by the security barrier, the perfecting of these techniques sometimes
becomes an end in itself. One can understand the technician's absorbed in-
terest in the tricks of his trade, but it is hard to pardon him when he gets his
means and ends confused. There are records where clandestine intelligence
has exploited a difficult and less remunerative source while it has neglected
an easy and more remunerative one. This kind of mis-collection would be
less likely to occur if the operation were less free to steer a course behind
the fog of its own security regulations.
4, State Four, the evaluation of data
If the language of intelligence were more precise it might use the
word "criticism" in place of the word "evaluation", and if "criticism of
data" were permitted we might move forward with a little more certainty
and speed. The word criticism means the comparison of something new and
unestablished with something older and better established. The best critic,
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
in these terms, is the man who has the greatest number of established some-
things in the right sort of mind, for he will be able by comparison to appraise
the validity of the new somethings as they come in. Thus he rejects a report
which puts Great Frusina's steel capacity at 45 million tons because he knows
from other evidence of unquestionable reliability that her capacity is 36
million tons.
In intelligence research the collected data bearing on the substantive
problem must be criticized before they can become the stuff from which a
hypothesis emerges. If incorrect data are not rejected the emergent hy-
pothesis will be incorrect, and thus the final picture incorrect. The methodo-
logical problem boils down to the expertise of the critic, the breadth of his
understanding, and the freedom he is permitted in arriving at his appraisal
of the data. Maybe, as in an earlier problem, this is as much a problem of
administration as of methodology. But the point is, that intelligence which
tries to run itself on an assembly-line basis and tries to substitute admin-
istrative techniques for high-class professional personnel is all too likely
to fall[ down on its all[-important criticism of the data. This is just another
way of saying that intelligence is a pursuit which cannot get along without men
of knowledge and wisdom.
There is, however, a problem in the area of evaluation which can prop-
erlybe called methodological. It arises because of the two ways in which the
produce of the surveillance operation is distributed to the consumers. The
first way of distribution is through the finished digest, report, or daily or
weekly summary. The new stuff is put on the expert's desk; he criticizes it,
judges its importance, mixes it with other data he received yesterday and
the week before, gives it background and point, and sends it on-to the consumer.
This activity may be called "reporting", but as can be seen it contains all
of the elements of research.
The second way in which the produce-of the surveillance operation is
distributed is in a much less finished form. The collectors pass to a sort of
middleman what -they have picked up. The middleman grades the data for re-
liability of source and the accuracy and reliability of content. He may then
distribute direct to-the consumer or to the research staff of his own organi-
zation or to other intelligence organizations. There is reason for the ex-
istence of this middleman, i.e., he is handling data which have been collected
clandestinely and his organization must protect its sources. But the middle-
man --regardless of his reason for existence --often does far more than ob-
literate the sources's identity. He attempts to grade the reliability of the data.
In doing so he is sometimes guided by some strange patterns of thought.
The middleman, according to standard practice, is restricted to a
very narrow language in making evaluations. He is permitted to grade the
reliability of the source according to the letters A, B, C, D, and the content
according to the numbers 1, 2, 3, 4. Thus A-1 would designate a report of
unvarnished truth straight from the horse's mouth. Data from less dependable
sources, and less accurate, might be B-2, C--4, etc. If the data happen to have
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
come from a document, a newspaper or press release, one school of evalu-
ators designates their value with the single word "documentary". Middle-
men have insisted on not amplifying their comments beyond this elementary
code and have done their best to see that others who might well amplify were
prohibited from so doing. They cling to this procedure on the ground that
they are purveyors of a raw commodity and that it is their duty to distribute
the commodity in the rawest state possible.
If this argument has force the middlemen themselves at times do much
to negate it by distributing the commodity in a state anything but raw. They
edit it, abbreviate it, or otherwise obscure its import, frequently losing the
point-of -observation or slant of the information: Was it a French Communist,
Socialist, or Rightist source which told the number of machine guns on the
headquarters of the Communist newspaper, L'Humant te; or told of new politi-
cal instructions from the Vatican? When the information lands on the con-
sumer's desk, it is no raw commodity but a semi-finished one.
Evaluation of the source by the middleman may be a valid and valuable
service. If the source is known to be a good one and if it must be protected
at all costs, to label it as grade A is helpful. But it is helpful and valid only
in so far as the middleman knows what he is talking about, or in so far as
the validity of the source has bearing on the content. But middlemen if they
lack independent line on the reliability of the source have been known to
grade the source based on the apparent reliability of the content. This is
neither helpful nor valid, particularly as the ultimate intelligence consumers
often tend to use the data without further and systematic criticism. Accepting
the evaluation at face, they are accordingly misled.
Middlemen have at times been people who neither directed clandestine
operations nor sat in a place where they were forced to view all incoming
materials. By all incoming materials I mean those collected overtly from
newspapers, government reports, transcriptions of foreign radio broadcasts,
etc., as well as those collected clandestinely from other secret sources.
Middlemen so placed were insulated from both the field experience of the
operator and the desk experience of the research man who is constantly and
aggressively working at a specialty. I can understand how a man living in
Rome and spending all his time collecting information on Italian politics can
develop a high critical sense. I can understand how a research man in Wash-
ington who immerses himself in the data of his specialty and every moment of
his professional life runs an obstacle race with his own and other people's
hypotheses must have a high critical sense and ability. It is less easy to
understand how a man who passively reviews a wide range of material wt th-
out doing anythtng about i t except grade i t, can possibly have the necessary
critical faculties.
To illustrate further: During the war a document graded as A-3 was
circulated which told of the American failure to take care of the inhabitants
of the city of Oran, Algeria, in the winter of 1943. The source was given an
A rating because it appeared to be someone familiar with Oran; the content
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
was graded as unreliable because the evaluator knew conditions there
were not as bad as represented. One recipient of this document poked around
until he identified the source as none other than an important French official
and the document as the text of one of his off-the-record speeches. Now the
official was unquestionably an A source; his knowledge stemmed from first-
hand informants or even his own experience. But what he said about Oran
under the Americans was of relatively little importance even if it had happened
to be correct. The importance of this document was that the source -a
man who was allegedly a friend and close ally-had voiced violent adverse
criticism of Americans. Yet this, its real value, had been completely ob-
scured by the encoded evaluation. To serve the more important use, the
evaluation should have called attention to the authorship of the document. If
the document had fallen into American hands through the work of a secret
agent whose identity had to be protected, the evaluation would have required
four or five sentences instead of one. But suppose that these sentences
could not be written without compromising the agent, is this adequate reason
for misleading the consumer through the A-3 evaluation? I would say not.
I would say that the middlemen should think up some other method of handling
the problem or get out of the business.
The crowning peculiarity that is met at times in this sort of evalua-
tion is that of removing the name of the newspaper from the reproduction of a
newspaper clipping and substituting the word documentary. What purpose this
can serve has always eluded me. Without the name of the newspaper the reci-
pient is deprived of perhaps the most useful piece of information in making
his own evaluation. For example, would you not like to know whether the
New York Times or the Dat ly Worker was responsible for an estimate that
Henry Wallace would poll ten million votes for President in 1948? Or would
you settle for the attribution "documentary" ?
5. State Five, the moment of hypothesis
What is desired in the way of hypotheses, whenever they may 9ccur, is
a large numberof possible interpretations of -the data, a large number of in-
ferences, or concepts, which are broadly based and productive of still other
concepts.
There are two things an intelligence organization must-have in order
to generate more aril better hypotheses: (1) professional staff of highest
competence and devotion to the task, and (2) access to all relevant data.
There were many men who lived contemporaneously with Mahan and
Mitchell, with Darwin and Freud, with Keynes and Pareto who could have
made these men's discoveries, for to a very large extent the facts were there
for anyone. But the great discoveries of the race are the result of rigorous,
agile, and profound thinking. The many failed because they lacked the -brains
capable of such thinking and the stamina to face up to an intellectual re-
sponsibility. This all, points a moral for intelligence: worthwhile discoveries
Approved For Release 2000/08/28 : CIA-RDP78-04718AO00600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
are not made by a lot of second-rate minds, no matter how they may be jux-
taposed organizationally. Twenty men with a mental rating of 5 put together
in one room will not produce the ideas of one man with a mental rating of 100,
and you cannot add minds as if they were so many fractional parts of genius.
But even if intelligence recruited its professional staff from among
the nation's most gifted people it does not follow that there are no problems
other than those which face any university researcher or journalist. Even
gifted people would not produce the good hypotheses unless they had access
to all the relevant data. This is by no means easy to arrange in intelligence.
One of the things that gets in the way is again security.
The worker in intelligence is dealing with state secrets upon which the
safety or well-being of a nation may rest. On the theory that the secrecy of
a secret is in inverse ratio to the number of people who know about it, a
highly important secret cannot be too widely known. But a man cannot pro-
duce the good hypothesis in any area if he does not know as much as there is
to know. It is interesting to speculate on how far Lord Keynes would have got
if libraries withheld large blocks of economic data on the ground that they
were operational, or how far Dr. Freud might have progressed if mental
clinics sealed their records against him on the ground that they were too
confidential. Yet intelligence people are constantly confronted with this
very argument which seeks to restrict the scope of knowledge on the grounds
that an important secret is involved. Security here is bought at great cost in
terms of results. Secrecy should be allowed to interfere only so far as ab-
solutely necessary.
I am skipping stage 6 (i.e., more collecting and more testing of hy-
potheses) in the intelligence process because it contains few, if any, problems
not covered in stages 2 and 3. The last stage, the one in which the established
hypothesis is presented as a new and better approximation to truth, contains
at least two important problems.
The form which the finished product must take is one of unadorned
brevity and clarity. To be sure, intelligence produces long reports - some
may reach many hundred pages -- but there are few studies, reports, or
monographs which do not also furnish a one or two-page summary. This
limit forces the intelligence producers to be clear in their thought and con-
cise in their presentation, and it enables the hurried consumer to digest while
he runs. The result is by no means an unalloyed good. There is such a thing
as a complicated idea; one that cannot be expounded in 250 words, or in two
pie-charts, an assemblage of little men, little engines, and three-quarters
of a little cotton bale. The consumer who insists that no idea is too com-
plicated for the 300-word summary is doing himself no favor. He is re-
quiring the impossible and paying heavily for it in two ways: he is kidding
himself that he really knows the subject, and he is contributing to the de-
moralization of his intelligence outfit.
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
The intelligence people who spend weeks of back-breaking work on a
substantive problem and come up with an answer whose meaning lies in its
refinements are injured at the disortion that may occur in a glib summary.
Next time they go at such a problem they will have less enthusiasm for ex-
haustive work, will turn in a poorer study with a still poorer summary-tacked
on the front. This is not a plea to the harassed man of action to read all the
hundreds of pages which come his way, but, it is a plea for the middle ground.
If he lets it be known he will read nothing longer than one double-spaced page,
many of his most loyal and hardest workers will lose some of their fervor in
serving him.
A second problem of the presentation stage is that of footnote re-
ferences. Intelligence consumers, unlike most serious and critical readers.
have not demanded footnotes; in fact, they have often condemned-footnoting
as mere evidence of .an academic mind. Thus in those intelligence organiz-a -
tions where rules of styling are made by men who do not understand the
methods of research there is opposition to the reference note. -Even in or-
ganizations where the value of citing sources is fully understood, many
sources must be concealed for the reason of security. Thus on both sides
there are reasons for skimping on citations and citations are skimped.
-I know of no formula for evil any surer than sloppy research unfoot-
noted. Sloppy and footnoted is not good, but sloppy and unfootnoted multiplies
the danger in a way the-layman can hardly imagine. The following example is
in point.
The military staffs of two countries, X and Y, had some pre-war con-
versations about the airfields which Y had in one of its colonies. Y told X
that it had some airfields built, some about to be built, -and a third group to
be built when the land had been purchased. On the outbreak of war the con-
tent of these conversations became an important item of intelligence, and
one of Country X's intelligence outfits distributed a report which accurately
named and located the fields, noted that some were ready, others not yet
built, and others only planned. It cited its source and gave the dates of the
conversations. So far so good.
A few months later another intelligence outfit in another country, Z,
got out a report on the colony. The report had a section on airfields. The
information which it contained came from the earlier study, but it was changed:
those airfields the land for which had not yet been bought were not so indicated,
and the citation of source was omitted. We now have a report in which three
categories of airfields have been reduced to two, i.e., those in operation and
others soon to be completed.
A little later a second intelligence outfit of Country Z took the -second
report and entered the airfield data on cards. These cards were printed
forms which had no appropriate box for noting that an airfield was in operation
or merely in the process of construction. The cards carried no footnote re-
ferences. All three categories of airfield thus dropped into one category.
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
Taking information from the cards you would have thought that the area had
some fifty more airfields than it in fact possessed.
It was about this time that a third intelligence outfit of Country Z came
into being and inherited the card file. It developed a technique of presenting
airfield data on maps with symbols to indicate length and type of runway.
Now back in the original document no length was given for the runways of
fields to be, but it was noted that the areas to be purchased for development
were to be one mile square. This datum had been repeated in all the suc-
ceeding reports. But when the map-makers landed upon it they found it
inconvenient. They did not wish to do the unrealistic thing of depicting a
square runway one mile by one mile, so they compromised. They reasoned
that the runways would be of maximum length, hence must follow the diagonal,
and hence be something over a mile, say 7,000 feet, in length. This point
decided, they made their maps and assigned a symbol indicating a 7,000 to
8,000 foot runway to fields, some forty-eight of which were never completed.
This sort of error is not entirely ascribable to a lack of footnote, but
I would say that the lack: considerably enhanced the chance of error. Further-
more, the lack of the footnote made correction more and more unlikely as the
data went through the producer -consumer -producer -consumer chain. By
the time the map was made a discovery of the error demanded hours of time
from the most studious and professionally competent man who might have
had the hours to spend. Even so the damage was irreparable, for his more
correct and cautious appraisal of airfields in Y's colony could not possibly
reach all the consumers of the erroneous reports, or convince all those whom
it did reach that his was the truer picture.
The methodological problems which I have discussed above appear to
be the most vexing ones, but my catalogue is not exhaustive. There are other
problems and there are other facets. Taken together they make the calling
of intelligence a difficult one, and cause the results of the intelligence process
often to fall below necessary standards of quality.
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
Chapter 3
PRODUCERS AND CONSUMERS OF INTELLIGENCE
There is no phase of intelligence more important than the proper re -
lationship between producers of intelligence and the people who use its prod-
uct. Oddly enough, one would expect this relationship to establish itself
automatically, but it does not do this. It only results from persistent and
conscious effort, and is likely to disappear when effort is relaxed.
Proper relationship between intelligence producers and consumers
is one of utmost delicacy. Intelligence must be close enoughto policy, plans,
and operations to have the greatest amount of guidance, and must not be so
close that it loses its objectivity and integrity. To spell out this meaning is
the task of the next pages.
One of the main propositions of this paper may be summarized as
follows: Unless the kind of knowledge herein discussed is complete, ac-
curate, and timely, and unless it is applicable to a specific problem or one
coming up, it is useless. In short, intelligence is knowledge solely for the
practical matter of taking action, and this requires that the intelligence staff
know a great deal about what is under discussion in other units of, say, the
department charged with policy, plans, and' operations, and that it have all
the guidance and cooperation from them which can be afforded. The need
for guidance should be evident, for if the intelligence staff is sealed off from
the action to be planned and carried out, the knowledge which it produces
may not fill the bill.
Let me be precise about my use of the word guidance. To be properly
"guided" in a given task intelligence must know almost all about it. If you
wanted to find out from a road contractor how big a job it was to build a
particular piece of road, you should not go to him and ask: "How hard is it
to make a road?" Before you could expect any meaningful answer you must
stipulate what two points the road was to connect, what volume of traffic
you wished to run over it, the axle loading of your heaviest vehicle, and so on.
After you had made your specifications clear the contractor might still be
unable to give you the final answer. He might give a rough estimate but re-
fuse to commit himself until he had investigated the terrain to be traversed,
the weather with which he would have to contend while putting in the road, the
local. labor force, etc. When he had made these investigations-he might come
up with a figure answering all preliminary specifications but which was pro-
hibitively high in cost. At this point he must return to you to begin conver-
sations on compromises. Will you accept two lanes instead of three of -four ?
Will you accept a more circuitous route with fewer cuts, fills, and difficult
grades? Will you accept a less expensive surface? As you talk these matters
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
over with him you find yourself, although you are not a professional road-
builder, batting up suggestions on how he can avoid this or that technical
difficulty, and he, though no professional transportation man, begins asking
questions about your problems. If things go well, you fetch your technical
people in to the discussion, and he does also. Before you are done, your
two organizations are working together straight across the board and a com-
munity of interest and understanding emerges that produces a workable plan
and a smooth operation. Naturally and unconsciously the guidance which was
mandatory for his (and your) success has been brought into being.
Now this guidance is essential at all levels of strategic intelligence.
Intelligence does not formulate objectives; does not draft policy; it is not the
maker of plans; nor does it carry out operations. Intelligence, to use the
dreadful cliche, performs a service function. Its job is to see that the doers
are generally well-informed; to stand behind them withthe book opened at the
right page, to call their attention to the stubborn fact they may be neglecting,
and--at their request--to analyze alternative courses without indicat-
ing choice. Intelligence cannot serve if it does not know the doers' minds; if it
has not their confidence. It cannot serve unless it can have the kind of
guidance any professional man must have from his client. The uninitiated
will be surprised to hear that the element of guidance present in the full at
the lowest operational levels becomes rarer and rarer as the job of intelligence
mounts in augustness.
Without proper guidance and the confidence which goes with it, the
surveillance operation, while relatively certain to keep its eye on obvious
foreign problem areas may well neglect the less obvious though significant
ones. There will be a playing of hunches: "Watch Bolivia, they'll be screaming
for information on it in a month", "Isn't it about time we began watching for
unrest in Madagascar or Soviet activities in India"; "Say, how about the
Spanish underground, how about West African nationalism?" There will be
plain and fancy guess work on what is to be watched and what can be left to
cool off. There will be differences of opinion as to what is and is not impor-
tant; on where this, that, and the other matter belongs on the priority list.
This striving to anticipate the trouble spot is not to be discouraged, but it
certainly should be supplemented continuously by the very best advice that
intelligence consumers can offer.
Research in intelligence suffers even more than surveillance when
improperly guided. In the first place the knowledge which it purveys may
be inapplicable to the use it is to serve, incomplete, inaccurate, and late.
It is not reasonable to expect otherwise, when through lack of guidance intel-
ligence is asked to do in a week's or a day's time what may be simply beyond
human competence. To be able to deliver would demand a research staff
large enough to codify and keep up to date virtually the sum-total of universal
knowledge. Even then it is doubtful if what was required would emerge unless
intelligence had had some advance warning.
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
In the second place, the want of-sharp and timely guidance is chief
contributor to the worst sickness which can afflict intelligence -- that of
irresponsibility. When intelligence knows little or nothing of what lies behind
a request, it loses desire to participate in the thing to be accomplished; it
loses the drive to make exactly the right contribution to. the. united effort.
When this stage is reached, men cease to be either intelligent or sensitive;
they begin behaving as dumb and unhappy automatons who worry, if at all,
about the wrong thing. What they. hand on of knowledge is strictly non-
additive; it must be worked over by someone else up the line, less well-
informed, before it has value. It may even be out.of date or inadequate be-
cause long ago they quit caring. .
There are a number of reasons why intelligence producers and con-
sumers have difficulty in achieving the proper relationship. Some of these
are perhaps less typical of civilian departments than of the armed services.
The services, of course, are organized on the well-known staff pattern. This
is composed of six divisions responsible respectively for: personnel,tntel-
ithence, organization and training, service-supply-procurement, plans and
operations, and research and development. Now it is to be expected that
loyalties, as they jell in this organizational structure, will jell first up and
down the vertical administrative line of the division. Only secondarily, will
loyalties spread horizontally to coordinate divisions under the commander.
Thus there is reason inherent in staff structure why Intelligence might ex-
perience difficulty in ;etting the proper guidance on plans, projected opera-
tions, the strength of one's own forces, etc., from its coordinate staff as-
sociates. Some critics of staff organization then go further and point to a
doctrine buried deep Ln service formulae called "The Estimate of the Situ-
ation." They assert that herein lies something which further adds, and in no
small way, to an unsatisfactory relationship between intelligence producers
and consumers.
The estimate of the situation is what a military commander must
make before he decides upon a course of action. In the preparation of this
estimate each staff officer has a clearly defined role: personnel, operations,
and logistics tell the commander precisely about his own force; intelligence
tells him about the physical environment and the enemy force, etc. However,
the degree to which intelligence is permitted knowledge of his commander's
own forces and the courses of action which the commander may be mulling
over are not spelled out in the formula. It has been argued, however, that the
G-2 (intelligence officer) should approach his job of estimating the enemy with
complete objectivity, and that if he has full knowledge of his own forces and
how they may be employed, his thought may jump ahead to the showdown of
strength. He will secs his side about to win or lose, and his elation or fear
will effect his estimate of the enemy. If he sees his side the easy winner,
the argument runs, he will tend to underrate the enemy; if 'the loser, to over=
rate the enemy. The commander, having enough difficulties conquering his
own subjective self, may not wish to complicate his task by having to screen
out that of his intelligence officer. He may feel justified in keeping his whole
intelligence arm in ignorance; or he might tell his G-2 everything but only on
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
the expressed condition that the information be withheld from all subordinate
members of. the G-2. staff.
It must be said, however, that no matter how good this reasoning may
appear to the commander, it rarely seems good or compelling to his intel-
ligence officer. The latter will always be miffed at the thought that his chief
doubts his ability to overcome his subjective self, or that his chief holds him
or his organization as a poor security risk. To top it all, he will be even
more than miffed feeling that no matter how hard he works, he runs the risk
of turning out a useless product.
Those who argue that staff structure and the doctrine of the estimate-
of-the-situation have within them the means of stultifying a free give and take
between intelligence producers and consumers have a point. I would be more
impressed if this doctrine were the only discernible cause and if civilian de-
partments which have inherited no such doctrine did not also have their dif-
ficulties in the producer-consumer relationship. There are other causes and
the first of these arises as a psychological by-product of intelligence practice
itself.
This practice separates out the thing called intelligence from all other
elements necessary to accomplish an end, and then bestows upon one group of
men, to the formal exclusion of others, all contact with the various steps
necessary to the intelligence process. Deep in their subconscious selves,
then, the excluded may well harbor the feeling that someone has told them
they are not-quite bright- -has said, in effect, "Now don't worry, your thinking
is being done for you. We've arranged to give you an external brain. When-
ever you want to know something, just ask Intelligence."
If Intelligence were staffed with supermen and geniuses who promptly
and invariably came up with a correct and useful answer, the sting might
wear off; intelligence might come to be revered by its users as a superior
brain. But so long as intelligence is not so staffed (what is?) the relation-
ship between producers and consumers will continue a troubled one.
Another cause for a not too happy relationship is again that of "se-
curity" which I have discussed in other contexts heretofore. Policy makers
and planners will, in the nature of things, deal with secrets of state. The
disclosure of such secrets would amount to a national calamity. (What if
one month before the Allied assault on Normandy or the American landing at
Leyte, the enemy learned the exact time, place, and magnitude of the pro-
jected attack?) Likewise, intelligence must have tts secrets. Apowerful
intelligence organization can develop sources of information of value beyond
price. They themselves can even become the points of departure and the
guarantors of success for a policy, a plan, or an operation. The revelation
of such sources or even a hint of their identity will cause their extinction.
Their loss can be likened to the loss of an army or all the dollars involved
in the Marshall Plan, or, upon occasions, the loss of the state itself.
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
The stakes being what they are, security and its formal rules are an
absolute essential and, as I have said before, the first rule of security is
to have the secret known by as few people as possible, all of established
discretion. What is the effect of this rule in the intelligence producer-
consumer relationship ?
When the rule is rigidly applied the consumers are entitled to a le-
gitimate doubt as to the validity of the producers' findings. Suppose you,
as a planner, were told something which was contrary to all previous know-
ledge and belief and contrary to the laws of common sense? Would you
accept it blindly and stake a policy or a plan upon it? What would be your
emotions, your considered judgment, and your final decision if, after re-
ceiving such information, you went back to the producer and got "Sorry,
but I cannot say more than I put in the memorandum" ?
Likewise, when -the consumers--the policy people and planners--
rigidly apply the rule, they give the intelligence producers good cause for
non-compliance; or the production of useless knowledge. Suppose you were
an intelligence producer and one of your consumers appeared with a re-
quest for everything you could find out about Java. Suppose the request was
phrased just this way. Suppose your entire staff were occupied on other
high-priority jobs and that you. could not put any of them on this request
without some justification. Suppose you told him this. It might be that he
would feel he could not give you the justification without a breach of security.
You are at cross-purposes. Possibly the consumer would drop the matter
there. But then again he might carry his request up through two echelons
and see to it that it came -back to you through two higher echelons of yours.
You would be given your orders to get to work on Java.
The chances are excellent that a request thus routed is one in which
security is paramount. The consumer does not really want to know -all about
Java; he wants to know merely about some tiny fraction of it. But he dares
not stipulate the fraction for fear of revealing his intent. So he asks for all
of it, hoping to get his information out of one paragraph or chapter of your
encyclopedia. He has no guarantee that this paragraph or chapter is not the
very one you consider unimportant and accordingly leave out. Nor have you
any guarantee that if you write the paragraph or chapter you will write it in
the way that will best, serve his interests.
Now what I have said above is the extreme. When the issues are of
highest importance both producers and consumers go to all permissible
lengths to help each other forward the success of the common task. But
this very leaning over backwards merely confirms the basic problem which
security throws in the way of a perfect relationship. Furthermore, when
the substantive issue is of some lower order of importance no one,may lean
over backwards and something akin to an impasse can easily develop.
- 20 -
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
I am not playing down in these paragraphs the importance of security
regulations and their observance. I am concerned with the point that se-
curity is like armor. You can pile on armor until the man inside is absolutely
safe and absolutely useless. Both producers and consumers of intelligence in
safeguarding their secrets can so insulate themselves that they are unable to
serve their reasons for being. This problem is critical and it deserves the
continuing study of a high-powered board. It cannot be met by the earnest,
informal but sporadic efforts current today. Nor do I believe it would vanish
with the passage of an official secrets act. Such an act would help enor-
mously, but it would not be the all-powerful panacea its proponents would
have it.
A final reason for misunderstandings between intelligence producers
and consumers is the understandable reluctance of consumers to embark
upon a hazardous task on the basis of someone else's say-so. After all, if
anyone is going to be hurt it probably will not be the producers. I will war-
rant that the Light Brigade's G-2 was high on the list of survivors in the
charge at Balaclava. The casualties, in both the literal and figurative senses,
will be to the intelligence users first, and to the producers late down the line.
Hence it is easy for the users to adopt the attitude expressed in the rhetorical
question: "Why should intelligence worry about doing a perfect job, after all
it's not their neck?" From this there can emerge a disrespect, perhaps even
a derogation, for the opinionof those who do riot carry the weight of operational
responsibility. Let intelligence make any mistake from which a penalty
follows and relations are likely to worsen.
One last word: intelligence is bound to make mistakes. Some of the
questions it must answer demand a divine omniscience; others demand more
painstaking work than can be accomplished in the time allotment; still others
canbe had only with the most elaborate of undercover preparations which have
never been made., But let intelligence make a mistake or come up with an in-
adequate answer and all too often the reaction of the consumers is on the
bitter side: "I wouldn't ask those geniuses to tell me how many pints there
were in a quart." When intelligence errs there seems to be less tolerance
than for the error of other specialists. For example, when a dentist pulls
out the wrong tooth (as the best dentists have done) or a lawyer loses a case,
the client's reaction -is not that he, himself could have done a better job, and
that hence forth he will do his own dental'and legal work. Yet in intelligence,
pardonably wrong diagnosis and understandably inadequate presentation very
often arouse just such reaction. For good reason or bad, an intelligence fail-
ure seems to rankle out of proportion to its importance, and to justify the con-
sumer in doing his own intelligence henceforth.
Thus there are a number of reasons why the relationship between pro-
ducers and users may at times be extraordinarily difficult. The result is that
the all-important element of guidance is lost. Once this occurs, intelligence
must remain innocent of the consumers' requirements, and the consumers in-
nocent of intelligence's capacity to contribute to their problems. In wartime
the closer to the fighting front and the smaller the operating unit, the better
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A0006001000.03-3
the relationship and the keener the guidance; the more remote and the larger
the unit, the worse the guidance. There are few situations even comparable to
the fighting front, and where they exist they lack that element of common phy-
sical peril which makes all men of one side friends and brothers. Thus in
peacetime, top-level intelligence must function in the very worst area of
wartime relations without the leaven of what you might call front-line tol-
erance. The relationship is likely to remain poor. This danger of intelligence
being too far from the users appears to be a lesser danger than that of being
too close. But what of this other danger?
The Problem of Objecttutty and Integrity
The danger of being too close to the consumers is not to be readily
dismissed. In a moment of intense exasperation intelligence producers and
consumers might agree to knock down the administrative barriers and move
intelligence piecemeal into the policy, plans, or operations section, or to
break up intelligence into its regional and functional units and disperse them
among appropriate parts of the operating organization. If this were donet in-
telligence would likely acquire all the guidance it could wish, perhaps even
more -than it could legitimately stomach. There will be great and obvious
advantages; there will also be costs, some of them prohibitive.
Intelligence is likely to be diverted from its essential task. I mean
this in. its most crude sense: the intelligence personnel who are profession-
ally studious yet possessed of some of the talents of the doer are going to
find themselves asked to share a-non-intelligence burden of office. Personnel
raids of this sort are very familiar to intelligence people; practically every-
one not in intelligence has a way of fancying the best intelligence staff as a
pool of unencumbered and -elite manpower ready to be tapped at will. Fight-
ing off such raids is a well-known necessity. Once the intelligence man has
crossed into operations he will have much difficulty arranging his return.
He will soon find himself engulfed in the day-to-day business of the new job.
Soon the intelligence staff is whittled -down to its least valuable members and
has thus lost its identity and its functional integrity. This very thing has
happened enough times to be worthy of serious -consideration.
Intelligence, it brought too close to its consumers, is likely to be di-
verted in-a less crude sense, but scarcely a less, damaging one. For instance,
the detailed problems of an operating office can be many and compelling. A
great many of them require the production of "spot intelligence." The tend-
ency will be to divert :intelligence personnel to this kind of work. This is not
to argue the work's unimportance, but it is to argue that too much diversion
of this kind makes poor use of an intelligence staff and deprives it of those
long stretches of uninterrupted time required to carry out important long-
range projects.
Finally, if intelligence is brought into and dispersed among appropri-
ate planning or operations sections, the substantive integrity of intelligence
can be seriously injured. Under these conditions the components (regional
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
or functional) of the former intelligence organization become relatively free
agents responsible solely to the planning or operations chief. They are then
under no compulsion to coordinate their views with other intelligence groups
with cognate interests. It thus becomes increasingly difficult to be sure that
all resources are brought to bear on the problem. Indeed, it is highly prob-
able that it will be dealt with solely by those expert in but a single sector of
the subject.
This, and the want of substantive give and take which it implies, is
not the only disadvantage. In addition there are the contrasting standards
of performance as a price of dispersal. An intelligence outfit, which is
administratively separated from its consumers and unified within itself, is
able to strive for a uniformly excellent product. The best work will inevitably
become the scale against which other work is measured. Destroy that cen-
tralization and unity and you destroy the best and most natural method of
competition and the good deriving from it.
To all the foregoing, there may be administrative remedies. They
will not be wholly effective, but they maybe able to meet the worst objections.
There is, however, one high-order disadvantage in bringing the producers
and consumers too close together which will elude the most ingenious ad-
ministrative devices.
Almost any group of men confronted with getting something planned
or done will sooner or later hit upon what they consider the single most
desirable course of action. Usually they will try to reach this solution as
quickly as possible. It is not unlikely, therefore, that they may have arrived
at this solution in ignorance of many relevant and important facts; with their
prejudices and cliches of thought discriminating in favor of the facts which
they do use. This kind of off -the-cuff solutiontends to harden: Their "view"
is thus and so; their "position" therefore, thus and so; their "line"in sup-
port of the "view" and "position" thus and so. Now add the. ingredients of
time and opposition and you have something which may be termed "policy."
Now it's rather easy for an intelligence staff itself to fall into this
error; to have its own difficulties with that view, position, slant, and line
which tends to harden into policy. In spite of the fact that professional schol-
ars, i.e., intelligence officers are supposed to have acquired a technique in
guarding against theirown intellectual frailties; they are by no means always
successful. Intelligence is full of battles between the pro-Mihailovitch and
pro-Tito factions, between the champions and opponents of aid to China, be-
tween defenders and detractors of the Jewish nation home in Palestine.
The fact that there have been such differences of opinion among supposedly
objective and impartial students who have had access to substantially the
same material, is evidence of someone's surrender to his irrational self.
If intelligence separately administered and thus under the best of
conditions finds itself guilty of hasty and unsound "policy", it is likely to
find itself doing more of this sort of thing when it is under the administrative
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
control of its consumers in plans or operations. I do not see how, in human
nature, it can be otherwise. Nor under these conditions do I -see how intel-
ligence can escape, every so often, from swinging behind the upolicy" of
the operating unit and prostituting itself in the production of what the Nazis
used to call kam.p fende Wissenchaft . (roughly, knowledge to further aims of
state policy.)
I cannot escape the belief that where intelligence is under the ad-
ministrative control of consumers, it will find itself right in the middle of
policy, and that upon occasions it will be the unabashed apologist for a given
policy rather than its impartial and objective analyst. As Walter Lippmann
sagely remarks, "The only institutional safeguard (for impartial and ob-
jective analysis) is to separate as absolutely as it is possible to do so the
staff which executes from the staff which investigates. The two should be
parallel but quite distinct bodies of men, recruited differently, paid if pos-
sible from separate funds, responsible to different heads, intrinsically un-
interested in each other's personal-success." (quoted from Public Opinion,
The Macmillan Co., New York, 1922 with the kind permission of the pub-
lisher. Chapter XXVI, Section 2).
For these reasons, when intelligence is too close to operations what
is unquestionably gained in ~utdance may well be lost in tnte~rtty and ob-
jectivity. The absorption of intelligence producers by the consumers may
prove to be too heroic a cure for both disease and patient.
The only way out of the dilemma seems to lie in the very compromise
that is usually attempted: guarantee intelligence its administrative and sub-
stantive integrity by keeping it separate from its consumers; keep trying
every known device to make the users and the producers familiar with each
others organization.
Intelligence must not be the apologist for policy, as I have said, but
this does not mean that intelligence has no role in policy formulation. Intel-
ligence's role is definite and simple and might be described in two stages:
(1) the exhaustive examination of the situation for which a policy is required,
and (2) the objective and impartial exploration of all the alternative solutions
which the policy problem offers.
It goes without saying that intelligence can skew its findings, so that
one alternative will appear many times more attractive than the others. It
is not heartening to reflect that just this has been done, though it would be
hard to prove that each such crime was one which intelligence embarked
upon entirely on its own responsiblity. For instance, during the war British
intelligence could prove at the drop of. a hat that there was such a thing as
a soft underbelly and that compared to it all other portals to fortress Europa
were as granite. Merely because intelligence is capable of getting off the
- 24 -
Approved For Release 2000/08/28 : CIA-RDP78-0471.8A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
beam is not sufficient reason to exclude it entirely from policy considerations
or to condemn it as unprincipled. As long as its complement of professional
personnel is of high intellectual and moral caliber, the risks which the policy-
making users run in accepting its analysis of alternatives are far less than
those they would run if intelligence is excluded from councils.
The Problem of Intellt~ence (the Product) and its Acceptance
What intelligence desires above all else is that its findings prove use-
ful in making decisions. There is, however, no universal law which obliges
policy, plans, and operations to accept and use these findings. If intelligence
is guilty of poor method or errors in judgment, there is nothing to coerce its
consumers into acting upon its advice. This fact has its benefits and its evils.
The benefits are almost too obvious to mention: for example, no one would
advocate taking a course of action which evidence, not considered by intel-
ligence, indicated to be suicidal. Just because an intelligence aberration
happens to indicate the law of gravity is inoperative in Lent does not constitute
sufficient reason to jump off a high roof on Good Friday. But in this very
laudable liberty to discount intelligence lies a source of danger. Where is one
to start and stop discounting intelligence ?
In one of the books for children written by James Willard Shultz there
is a story of some Indian tribes readying themselves for the warpath. The
combined chiefs met to discuss the operation and instructed headquarters
G-2 (a medicine man named White Antelope) to give an estimate of enemy
capabilities. In a couple of days' time White Antelope, having gone through
the necessary professional gyrations, came back to the combined chiefs. It
seems that the gods had favored his ceremonial by granting him a vision in
which he saw a lone raven seated on the carcass of a dead deer. As the raven
feasted he did not notice a magpie who slipped into a tree overhead and, took
some observations, nor did not notice that the magpie gave the signal for the
concentration of his deployed force. When the magpies' build-up in strength
was sufficient, they dropped down upon the raven and attacked. The raven
put up a game fight, but as things moved from bad to worse decided to re-
treat to prepared positions. If White Antelope were an irresponsible G-2 he
might have left it at that, but being responsible and feeling that he should
make his contribution to the common cause, he hazarded an interpretation.
To him the raven was the allied force and the magpies were the enemy--
the facts would justify such an interpretation--and plainly the enemy's cap-
abilities were more than adequate. The allies were in for a licking. But Bull
Head who was supreme commander spoke up and said in effect, "What you
tell us is not much more than that the expeditionary force will be in danger.
This we already know. As to the raven and the magpies, it is my belief that
we are the magpies, and the enemy, the raven, We start tomorrow." The
G-2's estimate had not been accepted.
It is important to notice that White Antelope had done his best accord-
ing to a method which was standard operating procedure. Bull Head himself
would have admitted as much. Bull Head did not override his G-2 because of
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
 Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3
a reasoned distrust of -his data or a rational doubt of his objectivity; he over-
rode him on the basis of -a hunch and probably a wishful one -at that.
Now I -do not wish to reject all -hunches and intuitions as uniformly
perilous, for there are hunches based upon knowledge and understanding
which are the stuff of highest truth. What I do seek to reject is intuition
based upon nothing and which takes off from the wish. The intelligence com-
sumer who has been close to the problem of the producer, who knows it
inside out, may have an insight denied the producer. His near view of the
broad aspects of the problem and his remoteness from the fogging detail
and drudgery of surveillance or research may be the very thing which per--
mits him to arrive at a more accurate -synthesis of truth than; that afforded
the producer. But let the consumer beware. If he overrides the conclusions
of his intelligence arm, and makes a correct estimate, let him deeply ponder
why this came about. Let him not get the notion that he need only consult his
stars to outdo his G-2. If he gets that notion, he will destroy his intelligence
organization- -its members will not -seek truth if a mere soothsayer may
negate their conclusions.
Adolf Hitler was such a consumer. There is every reason to think
that his intelligence at both surveillance and research levels was technically
adequate, and that his general staff was technically competent. There is
every reason to believe that he got accurate knowledge from his intelligence
and good advice from the staff which based its judgments upon this know-
ledge. But Hitler had his hunches and the first few of them were brilliant.
Because of luck, or because of n profound and perhaps subconscious knowledge
of what was at issue, he called the turn correctly and in opposition to his
more formal advice. But the trouble was that he apparently did not try to
analyze the why of his successful intuition. He went on as if it were a natural,
personal, and infallible source of truth. When he ordered -a cut-back in
German war production in the fall. of 1941 he began to reap the penalties for
his own errors, i.e., overestimating the Luftwaffe's capabilities and under-
estimating the capabilities of the Soviet Union. Thus, he not only took some
direct and positive steps, but he also took indirect and equally hurtful steps
to lose the war. He succeeded in damaging severely the continuing utility
of his staff and intelligence service.
Where intelligence producers realize that there is no sense in for-
warding knowledge which does not correspond to the consumer's precon-
ceptions, there ceases to be intelligence. The consumer is out on his own
with no more guide than the tea leaf and the crystal ball. He may _do well
with them, but for the long haul I would place my money elsewhere. Without
discarding intuition as a false friend, I would urge the consumer to use it
with full knowledge of its frailties. When the findings of the intelligence arm
are regularly ignored because of intuition, the consumer should recognize
that he is turning his back on the instruments by which western man has,
since Aristotle, steadily enlarged his horizons--those of reason and scientific
method.
Approved For Release 2000/08/28 : CIA-RDP78-04718A000600100003-3