SILENT WARFARE

UNDERSTANDING THE WORLD

OF INTELLIGENCE

Abram N. Shulsky
Foreword by
Roy Godson

BRASSEY’S (US), Inc.

A Division of Maxwell Macmillan, Inc.
Washington • New York • London • Oxford
Beijing • Frankfurt • Sâo Paulo • Sydney • Tokyo • Toronto
All rights reserved. No part ot this book may be reproduced, stored in a retrieval
system, or transmitted in any form or by any means—electronic, electrostatic,
magnetic tape, mechanical, photocopying, recording, or otherwise—without per­
mission in writing from the publisher.
Brassey’s (US), Inc.
Editorial Offices Order Department
Brassey’s (US), Inc. Brassey’s Book Orders
8000 Westpark Drive % Macmillan Publishing Co.
First Floor Front and Brown Streets
McLean, Virginia 22102 Riverside, New Jersey 08075
Brassey’s (US), Inc., books are available at special discounts for bulk purchases for
sales promotions, premiums, fund-raising, or educational use through the Special
Sales Director, Macmillan Publishing Company, 866 Third Avenue, New York,
New York 10022.
Excerpts from Strategic Intelligence for American World Policy by Sherman Kent.
Copyright 1949 by Princeton University Press. Copyright © renewed 1976 by
Princeton University Press. Reprinted with permission of Princeton University
Press.
Excerpts from Secrecy and Democracy: The CIA in Transition by Admiral Stans-
field Turner. Copyright © 1985 by Stansfield Turner. Reprinted by permission of
Houghton Mifflin Company.
Library of Congress Cataloging-in-Publication Data
Shulsky, Abram N.
Silent warfare : understanding the world of intelligence / Abram
N. Shulsky.
p. cm.
Includes index.
ISBN 0-08-040566-5
1. Intelligence service. 2. Intelligence service—United States.
I. Title.
JF1525.I6S49 1991
327.12—dc20 90-49399
CIP
British Library Cataloguing in Publication Data
Shulsky, Abram N.
Silent warfare : understanding the world of intelligence.
1. Intelligence services
I. Title
327.12
ISBN 0-08-040566-5
Published in the United States of America Designed by Nancy Sugihara
10 9 8 7 6 5 4 3 2
To the memory o f
my parents
 CONTENTS

Foreword ix

Preface xi

Introduction: Writing about Secrets xiii

1 What Is Intelligence? 1

The Scope o f Intellignce 3

The Elements o f Intelligence 7

2 Spies, M achines, and Libraries: Collecting the Data 11

Human Intelligence Collection 11

Technical Intelligence Collection 20
Comparison o f Humint and Techint 30
Open-source Collection 33

3 What D oes It A ll Mean?

Intelligence Analysis and Production 37

What Is Analysis? 37
The Intelligence Product 53
Intelligence Failure and Surprise 59

Vll
 4 Working Behind the Scenes: Covert Action 73

What Is Covert Action? Some Definitions 73

Types and Examples o f Covert Action 76
Covert Action and Secrecy 90
Covert Action and Intelligence 95

5 Spy vs. Spy: Counterintelligence 99

The Classification o f Information 100

Security 105
Counterespionage 109
Multidisciplinary Counterintelligence (MDCI) 115
Deception and Counterdeception 118
Counterintellignce Analysis 128

6 Guarding the Guardians:

The Management o f Intelligence 131

Secrecy and Control 131

Expertise and Policy 136
Intelligence and Democracy 144

7 Tw o V iew s o f Intelligence 159

Historical Development o f the American View 161

Intelligence and Moral Issues 168

8 Toward a Theory o f Intelligence 171

The Dual Nature o f Intelligence 177

Notes 181
Index 217
About the Author 223
 FOREWORD

The literature on intelligence, security, and statecraft is large and continues

to grow. These “ serious’’ studies are matched in the popular literature by a
large volume of espionage fiction. Sifting the wheat from the chaff, how­
ever, is not easy. The Intelligence & National Security Library is intended
to make available important studies and firsthand accounts by Western and
non-Western intelligence practitioners, scholars, and others who have
amassed interesting information and analyses. Individual titles are reviewed
by an editorial advisory board of scholars and former senior intelligence
officials from Western Europe and North America. Though every volume
stands on its own merit, the series as a whole is intended to improve our
understanding of intelligence and its relationship to security.
This volume provides a unique introduction to the world of intelligence
and fills the need for a single, comprehensive, understandable book on
this subject. The author, Abram N. Shulsky, has extensive knowledge and
experience with intelligence both in academia and in the legislative and
executive branches of government. In the early 1980s he was minority
0 -e., Democratic) staff director of the U.S. Senate Select Committee on
Intelligence, which is responsible for overseeing the entire U.S. intelli­
gence system. He then served as the director of strategic arms control
policy at the U.S. Defense Department, and later was a consultant to the
president’s Foreign Intelligence Advisory Board in the White House.
In this book, Shulsky explains the major missions and products of intel­
ligence agencies, as well as the procedures and activities used to achieve
diem. He also explores the complex tensions and relationship between a

IX
secret intelligence agency and the democratic government and society that it
serves. Silent Warfare breaks new ground in intelligence studies by propos­
ing a framework that focuses on the dual nature of intelligence— as both a
search for knowledge about political, military, and economic matters (with
similarities, therefore, to social science) and as an element of the struggle
among nations (with similarities to war). Shulsky brings together the knowl­
edge and insights gained in the past decade as intelligence studies mush­
roomed in the United States, Canada, and the United Kingdom. This book
codifies “ intelligence” for the professional and the serious student, while
also providing a realistic framework for those fascinated by fiction about
intelligence agencies and their operatives.

R oy G odson
General Editor
Intelligence & National Security Library
Washington, D.C.
 PREFACE

This book has its origin in a course on intelligence that I taught in 1985 as
a visiting professor at the University of Chicago’s John M. Olin Center for
Inquiry into the Theory and Practice of Democracy. Returning to the aca­
demic world forced me to impose some order on my thoughts concerning
intelligence, a subject with which I had become familiar in the public policy
world of Washington. I am grateful to the Olin Center’s director, Allan
Bloom, for this opportunity and for much, much else besides.
While working on this book, I enjoyed the hospitality of the National
Strategy Information Center, where I held the title of senior fellow, an
enviable position with maximum freedom and a minimum of responsibili­
ties. I am grateful to the center’s president, Frank R. Barnett; to the director
of its Washington office, Roy Godson; and to Jeffrey Berman, factotum, for
their support.
Through this affiliation, 1 remained in close contact with the work of the
Consortium for the Study of Intelligence and the Intelligence Studies Section
of the International Studies Association. This book owes a great deal to the
participants in the colloquia and conventions of these two organizations, and
I hope that it repays its debt to them by providing an introduction to intel­
ligence studies that they will find useful in their own teaching.
I am particularly grateful to the Lynde and Harry Bradley Foundation for
,ts generous and patient financial support and to its vice president for pro­
grams, Hillel G. Fradkin, for his personal interest and sponsorship. I also
wish to thank the John M. Olin Foundation for its support of the project’s
initial stages.

xi
 My work on this book was assisted by information and comments pro­
vided by the following friends and fellow students of intelligence: Eliot A.
Cohen. Diane S. Dornan, Kenneth E. deGraffenreid, Sam Halperin, Carnes
R. Lord, and Kenneth G. Robertson. I wish to express my gratitude to them
and to Gary J. Schmitt, whose careful critique of the manuscript, to say
nothing of his friendship and encouragement, contributed greatly to the final
result. I am. of course, solely responsible for its errors and omissions.

A bram N. S hulsky
Washington, D.C.
 INTRODUCTION
WRITING ABOUT SECRETS

A n oth er book about intelligence. So much has been published on this topic
in recent years that we have almost lost the sense that there is anything
paradoxical or even controversial about the public discussion of a subject
that has secrecy as an essential characteristic. Of course, in any democracy,
government business is the people’s business, and there is strong sympathy
for a policy of publicizing rather than concealing the government’s actions.
Nevertheless, it is also generally understood that safeguarding national
security in a dangerous world requires that many government actions be kept
secret from foreign adversaries. This, in turn, requires that the domestic
public be kept in the dark as well. Indeed, even democratic governments
(that of the United States included), to say nothing of undemocratic ones,
impose regulations requiring almost complete secrecy about most aspects of
intelligence operations and information. Therefore, those outside the gov­
ernment who write about intelligence are sooner or later faced with the
questions of why are they writing about a secret subject and, perhaps more
important, how are they able to do so.
In reviewing the extensive literature on intelligence, you can see certain
ways in which the question of secrecy has been addressed, although this is
more often done implicitly rather than explicitly. According to one view,
intelligence has, or should, become less of a “ cloak-and-dagger” affair and
rnore like a branch of the social sciences. In fact, intelligence should be a
universal social science that seeks to understand, and ultimately to predict,
all sorts of political, economic, social, and military matters. As such, it need
n°t be an inherently secretive endeavor. While some secrecy may be nec­

xiii
essary to protect the sources of important bits of confidential information, in
general, the most important facts— concerning the political, economic, so­
cial, technological, and demographic trends that shape a country’s behavior
in the long run— will not be secret.1
According to this view, as intelligence becomes more mature— in the
sense of becoming more scientific and more concerned with underlying
trends and causes— it relies less on specific secrets and can become more
open. As a result, there should be no inherent difficulty in writing about it
publicly. In fact, the public discussion of intelligence may help it achieve its
proper goal of becoming more like a social science by demystifying intel­
ligence and encouraging the flow of ideas between the intelligence and
academic communities.
Other, primarily American, publications on intelligence want to expose
alleged misdeeds of a country’s intelligence agencies and to bring about
some sort of change for the better in the way they operate. Typically, such
critiques regard the secrecy with which intelligence agencies operate either
as the source of the misdeeds or, at any rate, as a precondition for them.
In this view, the publication of books or articles about intelligence, which
destroys that secrecy, is an essential element of any reform program. What­
ever damage may be done by publicizing legitimate secrets is regarded as
relatively small compared to the damage that could be caused by the intel­
ligence agencies operating in continued secrecy. From this perspective,
there is nothing problematic about disclosing information concerning an
“ unreformed” intelligence agency. In fact, since secrecy is seen as a major
cause of, or crucial precondition for, wrongdoing, it is not clear whether
even a reformed intelligence agency would be entitled to secrecy.
One of the earliest and best-known books of this sort begins with the
assertion that “ there exists in [the United States] today a powerful and
dangerous secret cult— the cult of intelligence.” Publicity is an essential
weapon in the fight against “ this secret fraternity of the American political
aristocracy” :
The aim of this book is to provide the American people with the inside
information which they need— and to which they without question have
the right— to understand the significance of this issue and the importance
of dealing with it.2
Silent Warfare belongs to neither category. I believe that intelligence is
primarily about the discovery and protection of secrets and that it will
remain so in the future. I also believe that the discovery of other nations’
secrets and the protection of one’s own is, and will remain, so vital to
national security that the release of secret information that would impede
that work would not be justifiable, despite the catalogue of past misdeeds of
the U.S. Central Intelligence Agency (CIA) or allegations that might be
made concerning the intelligence agencies of other democratic states. Thus,
1 face the issue of secrecy and publicity in a more complex manner.
It is possible, by relying on public sources and “ leaks,” to write a book
containing a great deal of apparently classified (i.e., officially secret) infor­
mation about U.S. intelligence agencies and activities.3 1 have not chosen to
do this. This is in part because I have had authorized access to classified
intelligence information in connection with my various positions with the
U.S. government (not, however, with any intelligence agency). In the
course of that employment I promised not to disclose such information.4
Any attempt to claim that classified information presented in this book was
derived solely from public nonofficial sources could be reasonably viewed
as a subterfuge. More important, I regard secrecy as essential to intelligence
operations and, in any case, do not see the importance of putting into the
public domain various esoteric details about the operations of U.S.— or
other— intelligence agencies, the vast majority of which appear not to raise
significant issues of public policy.
Rather, in this book, while steering clear of classified detail, I attempt to
discuss basic concepts in intelligence. My purpose is to enable the reader to
think about the general issues of intelligence policy in a way that does justice
to the subject’s complexities and ambiguities. Thus, my approach is funda­
mentally theoretical; specific details that are publicly available are provided
to illustrate general points rather than to be comprehensive, either with
respect to the history of any aspect of intelligence or with respect to its
current condition.
While intelligence is becoming a recognized field of academic study at
least in the English-speaking world, the theoretical treatment of it remains
undeveloped.5 In this book, I attempt to contribute to that endeavor. At the
same time, I recognize that the book’s relatively heavy reliance on the
Anglo-American experience precludes it from reaching a truly general “ the­
ory of intelligence.” This reliance is due in large part to the fact that most
of the publicly available information about intelligence, not surprisingly,
concerns the intelligence agencies of those countries that have the most open
political systems. Nevertheless, where possible, I refer to intelligence ac­
tivities of other societies and try to consider the question of how intelligence
agencies vary depending on the nature of the regimes they serve.
This book does not provide readers with any secrets or “ inside informa­
tion,” either spectacular or humdrum. I hope it provides a framework for
understanding intelligence as well as the many revelations that will, most
likely, be forthcoming in the future.
 1
WHAT IS INTELLIGENCE?

In popular fiction, and to the public, intelligence has often been synony­
mous with espionage and skulduggery, with the sexual blackmail of a Mata
Hari and the cloak-and-dagger exploits of a James Bond. While activities of
this sort have their place within the world of intelligence, the full concept is
much richer. Let’s begin by looking at the phenomena to which the term
“ intelligence” is applied; these phenomena include certain kinds of infor­
m ationactivities, and organizations. 1
Intelligence refers to information relevant to a government’s formulating
and implementing policy to further its national security interests and to deal
with threats to those interests from actual or potential adversaries.2 In the
most obvious and often most important case, this information has to do with
military matters, such as an adversary’s plans for military action. Potential or
actual enemies typically do their best to keep this type of information secret.
Of course, other types of secret information may be equally important— for
example, information about another country’s diplomatic activities and in­
tentions, as well as information about its intelligence activities.
In addition to information of this sort, many types of information about an
actual or potential adversary may be useful to know, even though the ad­
versary may not attempt to keep them secret. These could include informa­
tion about internal political affairs and societal developments as well as
economic and demographic statistics. How much material of this sort is
actually published depends on the nature of the political regime. In a dem­
ocratic society, this type of material is almost always publicly available. A
totalitarian system, however, often strives to conceal any information about

1
might find helpful. Regardless of whether publicly available information is
to be considered ‘‘intelligence,” clearly there must be some process by
which it is systematically made available to government officials in a usable
form. An intelligence service often performs this function.
Finally, intelligence information typically includes not only the “ raw
data” collected by means of espionage or otherwise, but also the analyses
and assessments that may be based on it. It is this output, often referred to
as the intelligence “ product,” which is typically of direct value to policy­
makers. To what extent this intelligence product should strive to present a
comprehensive evaluation of a situation, based on all available data, both
public and secret, may vary from one intelligence service to another.
As an activity, intelligence comprises the collection and analysis of in­
telligence information— information relevant to the formulation and imple­
mentation of government national security policy, it also includes activities
undertaken to counter the intelligence activities of adversaries, either by
denying them access to information or by deceiving them about the facts or
their significance.
Intelligence comprises a wide range of activities. For example, there are
various methods of collecting information, such as espionage, photography,
intercepting communications, and research using publicly available docu­
ments and radio and television broadcasts. There also are different tech­
niques for analyzing the information that has been collected: some of these
may be similar to the methods the social sciences use, while others, such as
the decryption of coded messages, are peculiar to the intelligence world.
Similarly, denying information to others involves various activities, some of
which are similar to law enforcement work, such as investigating suspected
foreign intelligence agents to learn whether they are in contact with gov­
ernment officials. Others are more esoteric, such as using encryption to
protect communications. Finally, various means of deceiving adversaries
exist, such as “ double-agent” operations and deceptive radio signals.
Looking at this wide variety of intelligence activities, it seems difficult to
see any common thread running through them. They all, however, have to
do with obtaining or denying information. Therefore, intelligence as an
activity may be defined as that component of the struggle between adver­
saries that deals primarily with information (as opposed to economic com­
petition, diplomatic maneuvering or negotiations, or the threat or use of
military force, for example).
Finally, the term “ intelligence” also refers to the organizations that carry
out these activities. One of the most notable characteristics of such organi­
zations is the secrecy with which their activities must be conducted. Many
rules concerning access to information, derive from this requirement. Since
intelligence agencies are organized to enhance their capacity for secrecy,
they also may be given, along with their information-obtaining or -denying
functions, the responsibility of undertaking secret activities to advance their
government’s foreign policy objectives more directly.
Such activity, which in the U.S. intelligence lexicon is referred to as
“ covert action,” may range from the mundane, such as covertly providing
critical assistance to a friendly foreign government, to the spectacular, such
as orchestrating the overthrow of a hostile one. Whether it should be as­
signed to the same organizations that collect and analyze intelligence infor­
mation has occasionally been a controversial question. Even if, for practical
reasons, intelligence organizations are given the responsibility for covert
action, the more fundamental question of whether covert action should be
considered a part of intelligence from a theoretical, as well as a practical,
viewpoint would remain.

The Scope of Intelligence

Because not only governments but many other types of organizations oper­
ate in an environment characterized by a competitive struggle with adver­
saries, the concept of intelligence might be applied to them as well. For
example, some researchers try to extend the concept to business corpora­
tions, treating intelligence as “ organized information . . . designed to meet
the unique policy-making needs of one enterprise.” 3 Similarly, we could
speak of the intelligence function of a political party or campaign in trying
to figure out what the opposition is up to.
1 do not consider these possible extensions of the term “ intelligence” ; I
limit its meaning to the traditional scope, to information and activities rel­
evant to the national security concerns of governments.
This limitation does not adequately define the scope of intelligence, since
the term “ national security” is itself unclear. Its core meaning has to do
with protecting a nation against threats, ultimately military, emanating from
foreign nations. When a nation is or is about to be invaded, its national
security concerns clearly center on preventing or defeating the invasion and
securing itself against a similar situation arising in the future. In less omi­
nous circumstances, it may be much less clear which foreign nations or
events can threaten national security and therefore require attention by the
nation’s intelligence.
Ihe problem is further complicated by the fact that national security
interests and threats to them cannot be considered independently of the type
of government the nation has (the regime) and its ideological outlook.
Although adherents of realpolitik would argue that a nation’s interests are
determined by the objective factors of the international system, ideological
views affect how government perceives them. For example, a regime’s
ideological character may determine whether it views a given foreign coun­
try as a threat. In particular, status quo and revolutionary powers are likely
to have different views about what constitutes a threat to national security.4

Domestic Intelligence
An even more important area in which the nature of the regime affects the
scope of intelligence is what is called “ domestic intelligence.’’ Any gov­
ernment must be concerned not only with purely external threats (such as
military invasion) but also with threats against its ability to govern, or its
very existence, that arise from individuals or groups within the nation’s
borders. Such threats could come from groups that seek to overthrow the
government by illegal means, that seek to use violence to change govern­
ment policies, or that seek to exclude from the body politic members of a
given ethnic, racial, or religious group. How a government defines such
internal threats depends heavily on the type of government it is.
A regime in which a dynasty or a single political party has a monopoly of
power, for example, is likely to regard any domestic political dissent as a
security threat, and its intelligence service will focus a great deal of attention
on detecting and thwarting it. In the most extreme case, the government of
a totalitarian state may regard all nonmembers of the ruling party as actual
or potential enemies. By contrast, the notion of a “ loyal opposition,” as
found in parliamentary and other democratic systems, implies that the gov­
ernment’s domestic political opponents do not pose a security threat and
hence are not a suitable concern of intelligence.
In addition, there may be many possible types of connections between
domestic groups and foreign powers. At one end of the foreign-domestic
spectrum would be the activities of an individual or domestic group that acts
on behalf, and at the direction, of a hostile foreign power. Then there could
be groups or individuals who share common objectives with that power and
cooperate in the pursuit of them. Finally, there also could be groups that are
seen as subversive of the nation’s constitutional order or disloyal to it but
without any ties to foreign governments.

Intelligence and Law Enforcem ent

A related question deals with those transnational threats that do not emanate
primarily from a foreign government— for example, narcotics trafficking,
 international terrorism, or certain types of organized crime. These can be
serious threats to a nation’s well-being, but they appear to be problems that
come within the ambit of law enforcement rather than intelligence. Never­
theless, intelligence is often involved in the fight against them for several
reasons. First, they involve activities in foreign countries, where domestic
law enforcement agencies have no jurisdiction and where the local law
enforcement agencies may be unable or unwilling to be of much assistance.
Intelligence may be called upon for information about the foreign aspect of
these activities that would otherwise be unavailable.
Second, and more important, the law enforcement approach typically
involves waiting until a specific crime has been, or is about to be, committed
and then attempting to solve that particular crime and arrest the perpetrators.
This may not be an acceptable approach toward international threats. A
single incident, such as blowing up a passenger airplane in flight, may cause
so much harm that it is necessary to prevent these crimes rather than merely
solve them. Furthermore, a specific crime, such as smuggling narcotics
across an international border, may be part of a criminal enterprise’s oper­
ations; if the enterprise is large and well organized, arresting the perpetrators
of a single crime may not have much impact on it. Such occasional arrests
may even be a tolerable cost of doing business from the perspective of the
kingpins. Finally, even if agencies were content to wait until a specific crime
occurred, the chances of solving it would depend heavily on their having
a great deal of background information available about the organizations
involved.
For these reasons, an intelligence approach is often adopted with these
types of activity. Instead of waiting for a specific criminal act on which to
focus, agencies gather information over a long time concerning various
individuals and groups, their motivations, resources, interconnections, in­
tentions, and so forth. Often it is necessary to use informers who penetrate
the groups involved and who operate like spies. It also may be possible
to intercept communications or use other technical methods of collecting
information.
Thus, intelligence agencies are often involved in the fight against such
groups. Depending on the regime, agencies’ involvement may be limited to
the foreign aspects of these activities while domestic aspects remain in the
purview of law enforcement agencies. Even in these cases, however, law
enforcement agencies often must resort to some intelligence techniques to
deal with organized criminal groups. For example, with respect to domestic
law enforcement activities, the U.S. Federal Bureau of Investigation (FBI)
distinguishes between criminal intelligence investigations and ordinary crim-
*nal investigations. The former type of investigation is described as “ broad-
criminal act but rather on an ongoing criminal organization, whose size,
composition, past acts, intended criminal goals, and capacity to do harm
must be determined.5 The dividing line between law enforcement and in­
telligence, while not a sharp one, generally depends on whether the focus is
on punishment of a given criminal act or on an ongoing struggle with an
organization engaged in criminal activity.

Economics and Intelligence

Finally, there are questions concerning the function of intelligence agencies
with respect to economic issues. Again, much depends on the regime’s
nature and its economic system. In a system under centralized government
control (a “ command” economy), intelligence would be concerned with the
economic aspects of the government’s relations with foreign governments,
just as it is with all other aspects of its international relations. In addition,
intelligence can be used to enhance the state’s economic well-being. Ac­
quiring advanced Western technology, for example, is an important function
of the Soviet KGB (the Committee for State Security). This activity not only
increases the technological level of the Soviet military, but also saves Soviet
industry the great expense and difficulty of developing such technology on
its own.6
In a market economy, however, it is much less clear which economic
issues have a national security dimension that justifies or requires the in­
volvement of intelligence agencies. In addition, while many developments
in the international economy may have important national security ramifi­
cations, it is not obvious that the government should or, in any case, will do
very much about such developments; thus, the government may not have a
use for information about them. In a democratic society, economic policy is
much more likely to be determined by the interplay of domestic economic
interests than by any coherent view (whether or not based on intelligence
information) of the future world economic environment. For this reason, it
is not clear whether the government in such a society would be an important
consumer of such intelligence. Private economic interests could probably
put it to much greater use, but it is not clear that information gathered at
government expense would be distributed to individuals or corporations to
further private interests.
Furthermore, most of what an advanced industrial country would find
significant would be economic information about the other advanced indus­
trial countries that play a major role in the world economy. Currently, these
countries are not political rivals; it is not clear to what extent intelligence
erwise friendly nations.
In general, the relationship between intelligence and economic informa­
tion depends critically on the extent to which a nation sees its economic
situation and relationships in national security terms. For some countries, in
particular those with command economies, viewing economics in national
security terms may be almost axiomatic. For others, it depends on particular
circumstances.
For example, in wartime, intelligence means could be used to gain access
to strategic materials (or deny the enemy access to them), a key national
security goal. Similarly, in the course of negotiating with a foreign country
on economic matters, a nation might use its intelligence capability to learn
about its partners’ negotiating positions. More generally, the appetite for
economic intelligence on industrial, commercial, and financial activity in
other countries would probably depend on whether a government has an
“ industrial policy” bureaucracy that could make use of it.

The Elements of Intelligence

Whatever its scope, intelligence activity can be divided into four parts, often
referred to as the elements of intelligence: collection, analysis, coven ac­
tion, and counterintelligence. Since these elements are treated in detail in
subsequent chapters, they will be touched on only briefly here. The short
discussions that follow are meant only to elucidate the nature of the intel­
ligence activity included under each of the four headings and to sketch the
relationships between them.
Collection refers to the gathering of raw data, through espionage, tech­
nical means (photography, interception of electronic communications, and
so on), or in any other manner; thus, it comes closest to what is popularly
considered intelligence activity. While collection is obviously fundamental
to intelligence work, opinions differ regarding the relative importance of the
various methods. For example, students of intelligence have debated the
importance of collection from “ open sources” (such as publications and
radio and television broadcasts) as compared with using methods unique to
intelligence services, or the importance of collection through espionage as
compared with technical collection.
No matter how good the collected information is, however, it almost
never speaks for itself. In other words, some analysis of the information is
necessary for it to be useful when formulating and implementing foreign
policy or military activities. In the vast majority of cases, the collected
information is fragmentary, ambiguous, and susceptible to widely divergent
make judgments about the capabilities, intentions, and actions of another
party is a vital part of the intelligence process. Even more difficult is the
process of forecasting ( “ estimating,” in American intelligence jargon) the
future capabilities, intentions, and actions of a foreign government or po­
litical organization.
Conceptually, covert action differs from the other elements of intelligence
in that, while the others are concerned with seeking and safeguarding knowl­
edge, covert action seeks to influence political events directly. In terms of
intensity, covert action can range from persuasion or propaganda to para­
military action; it has been described as “ an activity midway between di­
plomacy and war.” 7
While the techniques for exerting this influence are many, they have the
common characteristic of anonymity, that is to say, the role of the govern­
ment conducting the activity is not readily apparent or publicly acknowl­
edged. For this reason, an intelligence agency’s ability to act secretly often
brings it the assignment of carrying out covert action as well. But because
it involves implementing policy rather than informing policymakers, there
have been occasional suggestions in the United States that covert action not
be a function of the same agency (the CIA) that collects and analyzes
intelligence. On the other hand, both the United States and Great Britain
have discovered through experience that having two organizations involved
in running clandestine operations (one for espionage and the other for covert
action) can result in energy-sapping rivalries, duplication of effort, or mu­
tual interference.8
The proper scope and nature of counterintelligence is less well defined
and more controversial than the other intelligence elements. In its most
general sense, counterintelligence refers to the protection of a society and
particularly its own intelligence capability from the actions of hostile intel­
ligence services. In the first place, counterintelligence means denying cer­
tain information to adversaries. This protection is accomplished by programs
of security— actions taken to keep information away from those not autho­
rized to have it; and by counterespionage— actions taken to apprehend or
otherwise neutralize foreign agents to prevent them from acquiring and
communicating secret information. In addition, a society could attack its
adversary’s intelligence analysis as well as its collection capability; this is
done using deception operations that provide false or deliberately misleading
information to the adversary to induce him to reach an incorrect analysis.
However, protecting a society or oneself against an adversary’s intelli­
gence capability, understood broadly, would require other activities as well.
It may be necessary to take steps to ensure that you are not deceived by
terintelligence must safeguard the integrity of the collection and analytic
functions. You would also want to know about a foreign government’s
covert action aimed at influencing your society and government. Ultimately
then, the breadth of counterintelligence activities is determined by the threat
an adversary’s intelligence activities pose.
 2
SPIES, MACHINES, AND
LIBRARIES
COLLECTING THE DATA

W e begin our examination of the elements of intelligence by looking at the

collection of raw data, which can then be correlated, analyzed, and reported.
The various methods of collecting these data are usually referred to as
“ disciplines” and can be broadly characterized as (1) collection from hu­
man sources (variously referred to as espionage, human intelligence collec­
tion, or, in U.S. intelligence jargon, humint), (2) collection by technical
means (technical intelligence, or techint), or (3) noncovert collection via
diplomatic contacts or generally available sources of information such as
newspapers or radio broadcasts (open-source collection).1

Human Intelligence Collection

Human intelligence collection, or espionage, is what the term “ intelligence”
is most likely to bring to mind. Its essence is in identifying and recruiting
into your service someone who has access to important information and who
is willing, for some reason, to pass it to officers of your intelligence service.
Typically, such people have access to this information by holding positions
of trust in their governments. In some cases (especially in wartime), the
person providing the information may not be a government official but a
private individual who has the opportunity to observe something of interest,
such as ships’ arrival in and departure from a harbor.
Ordinarily, individuals in two different roles are involved: an intelligence
officer, who is an employee of the foreign intelligence service, and a source,
who provides the officer with information for transmission back to the

11
intelligence service’s headquarters. The intelligence officer, or “ handler,”
maintains communication with the source, passes on the instructions coming
from the intelligence service’s headquarters, provides necessary resources
(such as copying or communications equipment), and, in general, seeks to
ensure the continuing flow of information.2

Types o f Intelligence Officers

Since they must avoid the attention of the government of the country in
which they are posted, intelligence officers cannot simply hang out a shingle
advertising their willingness to pay cash for secrets. They require what in
intelligence jargon is called “ cover,” that is, a plausible reason for being in
the country, visible means of financial support, a pretext for meeting people
with access to sensitive information, and so forth.3
In current U.S. parlance, a distinction is made between “ official” and
“ nonofficial” cover. Official cover refers to disguising an intelligence of­
ficer as a diplomat or some other kind of governmental official who would
be posted abroad. This kind of cover has several advantages. Most obvi­
ously, it can provide the intelligence officer with diplomatic immunity. If
his espionage activities are detected, international law limits the host gov­
ernment to declaring him persona non grata and expelling him from the
country.
In addition, posing as a diplomat improves the intelligence officer’s ac­
cess to potential sources; as a diplomat, he would, without raising suspi­
cions, meet with host-government officials in the course of his ordinary
business, as well as with other countries’ diplomats stationed in the same
capital. Indeed, since other countries also will use official cover for their
intelligence officers, he will have innocent opportunities for meeting them as
well.
Also, stationing intelligence agents under official cover in an embassy
guarantees that, if a national of the host country approaches the embassy
with sensitive materials or an offer to provide them, the matter can be
handled by intelligence professionals. In this sense, the existence of official
cover intelligence officers eases attempts by host-country nationals to make
contact with the intelligence service; such positions serve as a useful and
perhaps necessary “ mailbox,” especially in countries that strictly regulate
or prohibit their nationals’ travel to or communication with the outside
world.
Finally, official cover has certain administrative conveniences. The of­
ficer can be paid and other personnel matters can be handled through regular
government channels, and communication with the intelligence service’s
headquarters can be maintained through the intelligence “ station” (the
group of intelligence officers under official cover) located in the embassy.
Alternatively, official cover has several drawbacks. Most important, be­
cause of the relatively small number of officials posted to a given host
country, that country’s counterespionage service may be able to determine,
fairly precisely, which “ diplomats” are intelligence officers and which are
not. This may be done by the obvious, but laborious, methods of maintain­
ing surveillance on each official and noting his movements and contacts,
tapping his telephone, bugging his apartment, and so forth. The practice of
hiring nationals of the host country to work in embassies in various support
capacities probably facilitates such surveillance, especially in those coun­
tries, such as the Soviet Union, where it must be assumed that anyone
allowed to work in a foreign embassy has agreed to cooperate with the host
country’s intelligence service.4
Actually, less-expensive methods may be able to accomplish the same
goal. Materials published by a country might be used to trace the careers of
those serving as foreign service officers and to identify patterns that indicate
an intelligence connection. For example, it has been claimed that biographic
materials published by the U.S. Department of State made it relatively easy
to identify intelligence officers under official cover.5
Furthermore, while official cover may provide easy access to some po­
tential sources (primarily other diplomats and host-country foreign policy
officials), it may block access to others who might be hesitant to deal with
foreigners, either in general or with those from a particular country. In any
case, potential recruits are immediately put on notice that they are dealing
with an official of a foreign government, and that may make them more
cautious. In addition, if diplomatic relations are broken off, as might happen
in case of an intense crisis or war— when good intelligence may be most
necessary— official cover officers must leave the country, thereby disrupting
the operation of any network of sources they established.
The advantages and disadvantages of nonofficial cover are, for the most
part, the obverse of the considerations already discussed. On the one hand,
since they can pose as members of a wide variety of professions or strata of
society, nonofficial cover officials (NOCs) can have access to a different,
and perhaps wider, spectrum of potential sources. Similarly, they can pose
as nationals of the country to which they are posted or of some third country;
they thus obscure their connection with the government for which they
work, which may help them gain access to information or make contact with
potential sources. If diplomatic relations are broken off, they can remain and
continue to operate. In general, NOCs also should be much harder for the
host government to identify.
 limit its use. The expense ana auminisirauve umicuny mvuivcu m ^uviuiu6
nonofficial cover is much greater than in the case of official cover. One
method is to persuade a corporation or other private organization to allow an
intelligence officer to pose as a member of its staff. (In the case of U.S. or­
ganizations, the perceived risks of doing this— in particular, the possibility of
public disclosure that will harm the organization’s own relations with the host
government— have greatly increased in recent years, and it is often difficult
for the government to obtain their cooperation.) Alternatively, the officers
may themselves establish businesses or engage in activity that provides plau­
sible explanations for their presence in the target country. Not only may this
latter course be expensive, it also requires a good deal of time if the cover is
to be persuasive, which reduces the time and effort the officer can spend on
the primary task of intelligence collection.6 Communications are likely to
be more difficult to maintain, since an NOC cannot make regular use of com­
munications facilities in the nation’s embassy without raising some of the
very suspicions that nonofficial status is intended to avoid.
A clear distinction between official and nonofficial cover would exist only
in the case of a country whose citizens can and routinely do travel abroad for
their own reasons and without special government permission. Thus, until
the Gorbachev-era easing of Soviet restrictions on foreign travel, any Soviet
travel to the West could in some sense be seen as official because permission
to travel abroad was a rare and highly prized privilege. In this sense, all
Soviet visitors to foreign countries had some sort of official status, and an
intelligence officer could equally well take the disguise of a journalist for the
Novosti Press Agency as that of a diplomat.
With respect to Soviet and allied intelligence services, in the place of the
official cover-NOC distinction, a related one is made between “ legal” and
“ illegal” officers. The reference is not to the officer’s status in the host
country— whether he is there legally or not— but to his means of commu­
nicating with the intelligence headquarters. An illegal officer is one who
does not communicate via one of the “ legal” establishments in the host
country (the embassy, consulate, trade office, or so forth) but directly with
the headquarters. In this way, an illegal is similar to an NOC.
Obviously, it will be easier to insert illegal officers into a country that
routinely receives a large number of immigrants and visitors and is relatively
casual about controlling its borders than into one that does not receive
immigrants, generally keeps a watch on visitors, and guards its borders
carefully. Thus, it is not surprising to find many more references to Soviet
or Eastern European illegal officers operating in the United States and, until
1990, West Germany than vice versa. Judging from instances that have
LIU*'-' •*•*** ~ — r — ---- o — 7 — — o----» -- --- --- ©—
intelligence officers.
A case that illustrates this point involved Ludek Zemenek, a Czech na­
tional who became a KGB officer. Given the identity of a Rudolf Herrmann
(the real Rudolf Herrmann was a German who died in the Soviet Union
during World War II), he lived in East Germany for about a year. Then, at
the end of 1957, he left for West Germany, where he appeared to be an
ordinary East German refugee with an East German wife (who had also been
recruited by the KGB) and an infant son.
After four years in West Germany, Herrmann was instructed by the KGB
to emigrate to Canada, where he eventually established a small business
producing advertising and commercial films. He fulfilled various minor
tasks for the KGB, such as filing “ personality reports’’ on politicians and
journalists he met through his business and maintaining communications
with a Canadian professor who was a KGB agent. His most important
mission, however, was to preserve his cover so that he would be able, in
case of a break in diplomatic relations between the USSR and Canada, to
take control of the KGB’s network of Canadian sources from the legal
“ resident” (chief of the KGB station at the Soviet Embassy in Ottawa).
After six years in Canada, he was instructed to move to the United States,
where he performed similar tasks for the KGB as he had in Canada. When
his son Peter was seventeen, Rudolf explained who he was and what he was
doing; the son agreed to work for the KGB as well. Since Peter had been
brought to West Germany as an infant and raised from the age of four in
Canada and the United States, his background would not give rise to any
suspicion; he was to prepare himself for a career in the U.S. government
where he could operate as a long-term Soviet source, or “ m ole.” Presum­
ably he would have been able to do so had the FBI not confronted Rudolf
Herrmann several years later and, by threatening his arrest and the arrest of
his wife and son, obtained his cooperation.7

Types of Intelligence Sources

Just as we may classify intelligence officers as official cover or NOC,
legal or illegal, we may make some distinctions among types of intelli­
gence sources. One basic distinction is between recruited sources (whom
an intelligence officer, after preparing the ground, “ pitches,” or asks to
become a source) and “ walk-ins” (who volunteer to assist the intelligence
service of a foreign country, sometimes literally by walking into that
country’s embassy).
 While walk-ins have often been sources of extremely vaiuauic mu,m
gence, they are inherently suspect, since there is always the possibility that
the volunteer has in fact been dispatched by his own country’s intelligence
service to pass false or misleading information, to inform his country about
the opposing intelligence service’s methods of operation, and/or to entrap
one of the opposing service’s intelligence officers, leading to the officer’s
arrest or expulsion. On the other hand, an intelligence service that is too
suspicious of walk-ins may fail to obtain information that it easily could
have. Intelligence lore and history contain stories of walk-ins who were at
first ignored or spumed by the country to which they volunteered their
services but who turned out to be valuable intelligence sources.
One of the most famous cases of this sort involved Colonel Oleg Pen-
kovskiy, a Soviet military intelligence (GRU) officer who served as a British
and American intelligence source during 1961-62. Because of his position
as an intelligence officer working in the state committee charged with ob­
taining Western technology, as well as his close ties to the upper echelons
of the Soviet military establishment, Penkovskiy has been regarded as one
of the most important Western spies since World War 11. It comes as some­
thing of a shock, then, to realize that his initial attempts in 1960 to approach
the U.S. Embassy in Moscow were rebuffed. It seemed inconceivable to the
intelligence officers there that someone of his high rank and position would
volunteer to work for the United States; it appeared much more likely that
his offer was designed by Soviet counterintelligence as a provocation. It was
only in the following year, when Penkovskiy explained his interest in work­
ing for the West to Greville Wynne, a British businessman with whom he
conducted official business in Moscow, that his espionage career began.8
In addition to classifying sources, we might distinguish among the rea­
sons why sources provide information. Sources may be attracted by the
ideology or way of life of a foreign country or repulsed by that of their own.
They may be greedy; they may be somewhat unbalanced people who wish
to bring some excitement into their lives; they may desire to avenge what
they see as ill treatment by their government; or they may be being black­
mailed. The relative predominance of these motives depends on the char­
acteristics of the society involved, as well as the tactics of the opposing
intelligence services.
For example, the history of Soviet human intelligence collection against
the United States and Great Britain since the 1930s shows a substantial shift
from ideology toward greed and revenge as the motives for Americans or
Britons to betray their country. In the 1930s, the Soviets found that the
appeal of communism to some Cambridge University students and instruc­
tors, including those from prominent families, made the ideological atmos-
1----
became major Soviet agents within the British govemmeni w eic uuv um-
gess, Donald Maclean, and Harold ( “ Kim” ) Philby.9 On the other hand,
Americans and Britons who have been arrested for committing espionage on
behalf of the Soviet Union in the late 1970s and 1980s appear to have been
motivated mostly by greed.10 In other cases, most notably that of Edward
Lee Howard, the former CIA officer who gave the Soviets important oper­
ational details concerning the agency’s activities in Moscow, the motive was
revenge against the CIA for having fired him .11
In general, the popular imagination probably overestimates the prevalence
of blackmail as the reason for espionage, although the potential for black­
mail may enable an intelligence officer to keep active a source who had
become a spy for some other reason but who, later on, wants to quit. Even
so, some cases involving blackmail of the sort featured in popular spy novels
have come to light. A recent book recounts the blackmailing of a French
homosexual by the Czech intelligence service on the basis of photographs of
him with some Czech youths who were, of course, indirectly supplied by the
service. Following his recruitment, the Frenchman worked successively for
Czech, Romanian, and Soviet intelligence until his connection with the
Soviets was detected by the French secret service (SDECE), and he was
tried, convicted, and sentenced to seven years in prison.12

Problems of Human Intelligence Collection

Many of the problems encountered in human intelligence collection are
inherent in the nature of the enterprise, while others are more specific to the
nature of the intelligence target. Among the former is quality control— the
difficulty of ensuring that the information sources provide is genuine.
Sources may, for example, simply fabricate information or imaginatively
repackage and embellish publicly available material to make it appear that it
came from highly placed inside sources in order to sell it (creating, to use
intelligence jargon, a “ paper mill” ). The history of intelligence contains
occasional examples of clever paper mill operators who successfully bilked
their clients of large sums of money.
Such paper mills flourished in the late 1940s and early 1950s, taking
advantage of the Western intelligence services’ difficulty in operating be­
hind the iron curtain. Often, they were run by émigrés from Eastern Euro­
pean countries who were barely surviving as refugees in the West. They
soon discovered they could make a living by selling to Western intelligence
service information they claimed to be receiving from acquaintances among
1 eir former countrymen who had risen to important positions in the new
Communist governments. As many of these émigrés were well educated and
politically sophisticated, they were able to embellish and interpret publicly
available information to produce convincing intelligence reports.13
A potentially more serious quality-control problem arises from the pos­
sibility that an agent has been “ doubled”— that he is secretly working for
his supposed target and that the information he is providing to his sup­
posed employer is intended to deceive. Such doubling can occur when an
intelligence source is apprehended and chooses to cooperate with his cap-
tors to avoid punishment. Alternatively, the source could have been a dou­
ble from the beginning, never intending to serve the interests of his
supposed employer.
Some of the most interesting and remarkable stories in intelligence history
concern this situation. For example, in what was called the “ Double-Cross
System,” the British succeeded in gaining control of, and running, the
entire German human intelligence collection effort in Great Britain during
World War 11. From almost the beginning of the war, the British controlled
all intelligence reports that flowed to Germany from its supposed agents in
Britain. Among other achievements, these reports helped deceive the Ger­
mans about the location and nature of the D-Day landings in Normandy;
even three days after D-Day (on June 9, 1944). a message from a British-
controlled source was instrumental in retaining a German panzer division in
the Calais area (to repel the supposedly imminent landing of a nonexistent
main force), thereby helping the real invasion force in Normandy secure its
foothold.14
Other problems derive from the nature of the target. The more effective
and pervasive a target country’s internal security apparatus is, the more
difficulties it poses for human intelligence collection in that country. By
maintaining close control over international travel and communications, as
well as over the movements, communications, and economic activity of its
people generally, the government of a totalitarian country can make it ex­
tremely difficult for nonofficial cover, or illegal, intelligence officers to
operate. Official cover agents can be subjected to intensive surveillance,
making it hard for them to meet with any citizen of the target country
without being observed.
Allen Dulles, the U.S. director of central intelligence under President
Dwight Eisenhower, was referring to the effectiveness of such controls
when he said, in 1954, that “ it’s the toughest job intelligence has ever
faced— getting good information from behind the Iron Curtain.” 13 The re­
sult is what is termed, in U.S. intelligence jargon, a “ hard target” or a
“ denied area”— a country in which intelligence activities can in general
proceed only under official cover and then with great difficulty.
 Other targets pose particular troubles as well. For instance, in attempting
to collect intelligence about terrorism, we are hampered by the relatively
all, secretive, and tightly knit nature of the terrorist groups. To the extent
that membership in these groups depends on long acquaintanceships, family
ties, or a history of previous criminal acts, it becomes very difficult to insert
an intelligence agent into such a group (in intelligence jargon, to “ pene­
trate” it). Similarly, the loyalties existing within such a group (to say
nothing of the discipline it can impose on its members) make it very difficult
to persuade an existing member to betray it.
Even when an agent has been inserted into a terrorist group, serious
problems continue. To remain a member in good standing of the group, the
agent must provide material support for, or participate in, terrorist actions.
Yet an intelligence agency must put some limit on actions its agent can take
to preserve his bona fides. At the same time, using the information provided
by the agent to warn against or otherwise thwart a planned terrorist action
runs the risk of exposing the presence of an agent among the terrorists, thus
endangering the agent’s life. As with organized criminal groups, it is often
not until a member has been apprehended that an agency gets an opportunity
to look into the inner workings of the group.

Tradecraft
The particular methods an intelligence officer uses to operate and commu­
nicate with sources without being detected by the opposing intelligence
service are known collectively as “ tradecraft.” Part of the problem an
intelligence officer faces is to defeat the opposing side’s surveillance efforts
to be able to meet with sources or potential sources without giving away
their identity. An officer uses a wide variety of countersurveillance tricks to
determine whether he is under surveillance and to attempt to escape it.
For example, an officer may spend several hours traveling to a meeting
by a circuitous route, taking several different forms of transportation. If he
notices that the man who sat next to him on the westbound subway also
happens to be on his eastbound bus, he may reasonably conclude he is
being followed. The surveillance team may try to avoid discovery by us-
lng a relay system so the same individual is not tailing the officer all the
time. The game of surveillance and countersurveillance can be compli­
cated almost indefinitely.
Tradecraft also includes techniques for communicating with a source
without having to meet with him at all. For example, an officer may unob­
trusively hand off a package or piece of paper to a source as he passes by on
the street (a “ brush pass” ). If done correctly, the maneuver may be unob-
 it are fewer. One cannot rely solely on reports from journalists, travelers,
diplomats, or attachés; in general, press coverage will be restricted, borders
will be closed, and travel within the target country will be restricted, even
for the country’s own nationals and nationals of neutral countries. At the
same time, the international law prohibition against overflying another coun­
try without its consent no longer poses a political obstacle.
Even so, U.S. experience in trying to conduct human intelligence collec­
tion activities against the Soviet Union during the troubled peace of the years
right after World War II was so unsatisfactory that attention again turned to
photoreconnaissance. This interest was reinforced after the beginning of the
Korean War in 1950, when the West feared that the North Korean invasion
of the South foretold a Soviet invasion of Western Europe. It was believed
that the Soviets could secretly prepare a massive invasion force that would
then be able to attack Western Europe with very little warning. In addition,
basic intelligence information about the size of Soviet military forces was
often lacking.
Various attempts were made to overcome this deficiency in knowledge
about the Soviet Union. One project, code-named Moby Dick, involved
launching balloons equipped with downward-pointing cameras in Western
Europe. The plan was that these balloons would then drift across the Soviet
Union on the prevailing westerly winds until they reached Japan or the
Pacific Ocean. At that point, their camera pods would be released in re­
sponse to a radio signal and recovered. In reality, many of the balloons came
down over the Soviet Union, enabling the Soviets to assess the camera’s
technology. In any case, the random movement of the balloons over the vast
land mass of the Soviet Union probably meant that the photographs devel­
oped from the film that was recovered yielded minimal intelligence.22
On July 21, 1955, at a Geneva “ Big Four” (the United States, the Soviet
Union, Britain, and France) Summit, President Eisenhower proposed an
“ Open Skies” plan for mutual aerial surveillance of the United States and
the USSR. According to this plan, the two countries would “ give to each
other a complete blueprint of [their] military establishments, . . . from one
end of [the] country to the other” and would “ provide within [the] countries
facilities for aerial photography to the other country.” Each country would
have the right to conduct aerial surveillance over the other’s territory as it
saw fit.23 The purpose of the plan was to inhibit any attempt at launching a
surprise attack, which Eisenhower judged would be impossible if the po­
tential attacker was subject to unlimited aerial surveillance. This mechanism
was also seen as a precondition for various disarmament possibilities.
 roved the development ui a new reconnaissance piane, tne u-Z, which
could fly above the reach of Soviet fighters and surface-to-air missiles
sA Ms). Furthermore, it could probably have been foreseen that photo-
' hjc reconnaissance from satellites would become feasible within the
next several years. Thus, Eisenhower’s plan was an attempt to obtain by
aareement an opportunity to conduct aerial surveillance that technological
advance was going to produce (at least for the United States) in any case.
Soviet leader Nikita Khrushchev, however, rejected the scheme, character­
izing it as an attempt to legalize espionage. On June 21, 1956, Eisenhower
secretly approved the first U-2 flight over Soviet territory.24
During the next four years, the U-2 flights, both over Soviet territory and
along its periphery, yielded a rich intelligence return, providing the United
States routine coverage of important Soviet military test facilities and bases
for the first time. Although the Soviets lodged confidential diplomatic pro­
tests against the overflights, they were at first unable to do anything about
them.
It was just a matter of time before the Soviets developed some means to
attack the U-2. When the first successful attack occurred, on May 1, 1960,
however, the Eisenhower administration was unprepared to deal with the
situation, the difficulty of which was increased by the Soviets’ capture of the
plane’s pilot, Francis Gary Powers, and retrieval of its cameras. After U.S.
spokesmen had put forward an untenable cover story (that the U-2 was
engaged in weather research and had accidentally strayed from its intended
route over Turkey), Eisenhower personally took responsibility for the over­
flights of Soviet territory, which he then ended.
Even before the U-2 incident, however, the U.S. had begun work on a
satellite reconnaissance capability. In January 1958, an Air Force major
general publicly testified about the ongoing development of an advanced
reconnaissance system to be launched by the Atlas booster, as well as about
some other reconnaissance capability, with a recoverable capsule (for re­
turning the film to earth), that he optimistically predicted could be launched
in the spring of 1959.25
Despite being openly predicted, the existence of this satellite photorecon­
naissance capability was, during its early years, regarded as classified. Al­
though it could be assumed that the Soviets would understand that some
Sllch reconnaissance capability existed, it was not thought expedient to call
Public attention to the fact. At the time, the Soviets, consistent with their
Ejection of the Open Skies plan, took the position that space-based recon­
naissance constituted espionage just as much as did aerial reconnaissance
and was just as illegal. They maintained,
 From the standpoint of the security of a state it makes absolutely no
difference from what altitude espionage over its territory is conduct­
ed. . . . there is absolutely no ground for alleging that espionage at a high
altitude, with the aid of artificial Earth satellites, is quite lawful under the
existing rules of international law. Any attempt to use satellites for espi­
onage is just as unlawful as attempts to use aircraft for similar purposes.26

Given this Soviet position, the U.S. government feared that public dis­
cussion of its capabilities and actions in this area would rub salt in the Soviet
wounds and prompt them to make a political issue of it. Thus, after some
public discussion of space reconnaissance at the end of the Eisenhower
administration and seeing no reason to goad Khrushchev by openly asserting
the right of the United States to conduct such reconnaissance, the Kennedy
administration took “ steps . . . to turn the fledgling space reconnaissance
program from medium gray to deep black.” 27 This policy remained in effect
until 1978, when President Jimmy Carter, in a speech designed to assure the
American people of the verifiability of the almost-completed SALT II
Treaty, publicly confirmed that the United States possessed such a capabil­
ity; its existence had been, by that time, an open secret for many years.
However good a photographic reconnaissance capability may be, it de­
pends on sunlight and the absence of cloud cover. To get around these
limitations, other types of imagery, using radar or infrared waves (emitted
by hot objects), may be envisaged. For example, a radar can use the elec­
tromagnetic waves it emits not only to detect metal objects, but also, under
some circumstances, to obtain a kind of image of the object analogous to an
ordinary photographic image; experts use this image to determine some of
the object’s characteristics. (This differs from the techniques involved in
signals intelligence, discussed below, in that it uses the reflection of the
electromagnetic waves it emits rather than those emitted by its target.)
Similarly, another part of the electromagnetic spectrum, the infrared fre­
quencies, can be used to detect objects that are hotter than their back­
grounds. For example, in the 1930s, the British developed both infrared
sensors and radar as means of detecting enemy aircraft. At that time, radar
was much more feasible, and the infrared sensors were not developed. In the
future, radar and infrared alternatives to conventional photography may be
expected to become more important.28

Signals Intelligence
Signals intelligence (or sigint) is the generic term given to the process of
deriving intelligence from the interception of electromagnetic (radio) waves,
 •I referred to as signals. It may be subdivided according to the type
of dectromagnetic wave being intercepted:
. The interception of, and derivation of information from, foreign com-
* munications signals (radio messages) by other than the intended recip­
ients is known as communications intelligence, or comint.
• The interception, processing, and analysis of foreign telemetry (radio
signals that relay information from sensors on board a test vehicle to the
test engineers concerning the vehicle’s flight and performance charac­
teristics) are known as telemetry intelligence, or telint.
The interception, processing, and analysis of noncommunications elec­
tromagnetic radiations coming from a piece of military equipment (such
as a radar) while it operates are known as electronic intelligence, or
elint.29
In principle, any electromagnetic wave, emitted either as a necessary part or
as a by-product of the functioning of a piece of electrical equipment, is
subject to interception by a receiver that is properly placed and sufficiently
sensitive.30

Communications Intelligence (C om int)

Of these varieties of signals intelligence, the oldest is comint, which is
practically contemporaneous with the use of radio for military and diplo­
matic communications.31 Through the end of World War II, comint (com­
bined with cryptanalysis— the breaking of the codes and ciphers in which the
valuable messages are transmitted) was more important than any other
source of intelligence for the major powers, both in peace and in war, for
which we have adequate information to judge.32
One of the first uses of comint for which a record is publicly available is
the British Navy’s use during World War I to give it advance warning of any
hid by its German rival to venture out into the North Sea. In addition, at the
beginning ot the war, the British severed German-owned underwater tele­
graph cables that linked Germany with the Americas, Africa, and Spain. As
a result, Germany’s diplomatic “ cables” to those areas were sent by radio
and were intercepted by the Royal Navy’s monitoring stations as well.33
Whatever its earlier successes, it is World War II that marks the heyday
°1 British, as well as American, communications intelligence. Under the
rubrics “ Ultra” and “ M agic,” vast amounts of accurate and timely infor­
mation were made available to British and American political and military
eaders; measuring the precise extent of its impact on the course of the war
remains a huge historical task.
 While we would expect that a country’s most sensitive radio messages
would be encrypted to protect their confidentiality, some areas of commu­
nications intelligence are still independent of cryptanalysis. Because of the
expense and difficulty involved, some radio traffic is not encrypted; for
example, tactical voice communications, such as those between airplanes
and ground control stations, are often broadcast “ in the clear.” In other
cases, message traffic is not encrypted because it is believed, rightly or
wrongly, that an adversary does not wish to, or cannot, intercept or (because
of the large volumes of information involved) process it. Finally, we should
not underestimate the importance of simple inertia: for example, the United
States has made only belated responses to its discovery that the Soviet Union
can, and does, intercept long-distance telephone calls that are transmitted by
microwave from one tower to the next.
In addition, a technique known as “ traffic analysis” seeks to derive
useful information from fluctuations in the volume and other external char­
acteristics of radio communications, even when the content of the messages
cannot be understood. For example, if an army headquarters and its subor­
dinate command posts are exchanging an unusually large number of mes­
sages, we might conclude that an important operation is about to take place.
Similarly, by means of “ direction finding,” a technique for determining the
geographic origin of a radio signal, we can determine the location of the
ship, plane, or command post that is transmitting it.
While most comint involves interception of radio messages, messages
being passed by wire also can intercepted. This requires physical access to
the wires and so is of a less general application than comint involving radio
traffic, although in particular cases, it may be important. For example,
wiretaps may readily be maintained on telephone lines to an adversary’s
embassy or consulate in one’s own country.
On occasion it may be possible to gain surreptitious access, even outside
one’s own country, to telephone or telegraph wires an adversary is using.
For example, in the divided cities of Vienna and Berlin, in the early and
mid-1950s, British and American intelligence managed to tap such wires
used by the Soviet military authorities. In the case of the Berlin operation,
this involved elaborate secret tunneling from a point in the American sector
of the city, under the border between the American and Soviet sectors, to
intersect cables running entirely within the Soviet sector.34
In addition, through the cooperation of the private companies or gov­
ernment agencies that run the international telegraph service, nations have
been able to obtain the texts of other nation’s diplomatic telegrams; of
course, these messages were typically encrypted, so the benefit from ob­
taining them depended on a nation’s cryptanalytic capabilities. For exam-
 • rr the 1920s, the U.S. Black Chamber (a secret cryptanalytic
1*3 HilfjIIK 11
P1 ■ sUbordinate to the Department of State) worked on encrypted dip-
office s ^ tp,e texts of which it presumably obtained from the private
l0T'international telegraph companies.35 At present, of course, radio is
^ e likely to be used for communicating between a government and its
embassies abroad.

Telemetry Intelligence (Telint)

The other forms of sigint are newer and reflect the expanding military use of
electromagnetic phenomena. Telemetry intelligence (or telint) is similar in
concept to comint except (1) that the communications on which one is
eavesdropping are between a test vehicle (such as a missile) and a ground
station and (2) that they consist not of words but the readings of various
sensors and other on-board equipment. These measurements of such test
characteristics as the acceleration the vehicle is undergoing, the temperature
at various points within the vehicle, the rates of flow of fuels, and so forth,
taken together, give the engineers on the ground a picture of what is hap­
pening on the test vehicle. Just as in the case of comint, the country con­
ducting the test may seek to deny others access to the telemetric information
by encrypting it on the missile before broadcasting it back to the ground
station.

Electronics Intelligence (Elint)

Elint involves monitoring and analyzing noncommunication electromagnetic
emanations from foreign military equipment. By intercepting a radar signal,
for example, we can determine various operating characteristics of the radar,
such as its beam width (how much space it can scan at one time). Similarly,
the maximum operational range of a radar that emits discrete pulses can be
determined from the pulse repetition rate (the number of discrete pulses, or
bundles,” of radio waves emitted per second).
British scientific intelligence during World War 11 supplies an illustration
° f this point. From the fact that the German Frey a air defense radar operated
at a pulse repetition rate of 500 pulses per second, the British concluded that
'ts maximum range was 300 kilometers. (The calculation is made as follows:
t e radio wave emitted by a radar travels at the speed of light, or 300,000
alorneters per second. Hence, it can travel 300,000/500 = 600 kilometers
e ween pulses. Therefore, if the wave is to reach the target and return to the
j'adar before the next pulse is emitted, the target cannot be more than 300
kilometers from the radar).36
The possibilities for technical intelligence are limited only by the laws of
physics and human ingenuity. In addition to photint and sigint, other tech­
niques have been used to collect intelligence. For example, special sensors
have been developed to detect and characterize nuclear detonations. These
include seismometers, which measure the shock waves associated with un­
derground nuclear tests; devices to detect the radioactivity associated with
nuclear materials or the fallout of an above-ground nuclear test; and sensors
for the remote detection of the flashes of light produced by an above-ground
nuclear test. A different kind of sensor, sonar, uses another natural phe­
nomenon, sound waves, to detect submarines under the ocean’s surface;
aside from its location, other characteristics of the submarine may be de­
terminable from an analysis of the sounds it emits or reflects.37

Platforms fo r Intelligence Collection

In public discussion and in our discussion so far, the emphasis has been on
satellite-based systems to collect intelligence. Satellites have the obvious
advantage of being able to overfly any part of the world, without respect for
international boundaries. However, they have certain drawbacks as well:

• They are, in general, much more expensive than other reconnaissance

platforms.
• They cannot get very close to their targets (satellites in low-earth orbit
generally remain at altitudes well above 100 miles, while those in geo­
stationary orbit must be at an altitude of 22,400 miles).
• They are relatively useless a large part of the time even if their orbits are
calculated to maximize the intelligence they gather (in other words, a
photographic reconnaissance satellite that earns its keep by taking pic­
tures of the Soviet Union nevertheless must, because of the laws of
orbital physics, spend a lot of less productive time over the Southern
Hemisphere as well).
• Their orbits are predictable, so that the adversary may be able to warn
sensitive installations when a reconnaissance satellite will be in range.38

Consequently, the development of intelligence satellites has not meant

that other kinds of platforms are not used to carry intelligence sensors. For
example, airplanes can fly missions along the periphery of an adversary’s
territory, they can intercept communications that cannot be intercepted from
farther out in space, they can take pictures of areas near the territorial
d°fcnse radars, some of which may have become active precisely to chart
the airplane’s course.
Ships, such as the ubiquitous Soviet “ fishing trawlers,” can intercept
dio signals from off an adversary’s coast. Seaborne or ground-based ra­
dars if they can be located close enough to a site where test missiles are
launched or where they reenter the earth’s atmosphere, can track an adver­
sary’s missile flight tests and determine some of the missile’s characteristics.
Ground-based sites, depending on location, also can be used to intercept
various radio signals.

Legal Issues with Respect to Overhead Reconnaissance

As we have noted, the chief advantage of satellites is in their ability to
overfly an adversary’s territory without his consent. The U.S. position has
been that since a nation’s sovereignty does not extend into outer space
(beyond the earth’s atmosphere), a nation is as free to conduct activities in
outer space as it is on the high seas. At the beginning of the space age, the
Soviet Union took the position that satellite reconnaissance was as illegal
under international law as aerial reconnaissance involving the unapproved
overflight of another nation’s territory. The Soviets argued that, although
outer space did not fall under any nation’s sovereignty, it was not a legal
vacuum: the presumed illegality of espionage in general applied to space
reconnaissance as well.
The Outer Space Treaty in 1967 did not directly address the issue of space
reconnaissance. It is often claimed that in the Strategic Arms Limitation
Talks (SALT) the Soviets accepted the legality of photoreconnaissance and
other forms of satellite surveillance. However, the wording of the relevant
SALT provisions (the ABM Treaty, the SALT I Interim Agreement on
Offensive Arms, and the SALT II Treaty are the same in this regard) is
vague:

hor the purpose of providing assurance of compliance with the provisions

ot this Treaty [or Interim Agreement], each Party shall use national tech­
nical means of verification at its disposal in a manner consistent with
generally recognized principles of international law.39
rhe term "national technical means” (NTM) is not defined. While U.S.
0 ficiaIs have stated that NTM includes reconnaissance satellites, Soviet
cmfs have been silent about this interpretation. More important, the
Provision refers to the use of NTM “ in a manner consistent with generally
gnized principles of international law .” Thus, the provision is some-
11 gnUllS 11 Olliy WI1CI1 SUU11 reCOIUiaibSaillC 15 t u n u u t i t u 111 u m a iu iu L iu citu y
regarded as consistent with international law principles and, hence, presum­
ably already permissible.40 In short, the provision neither legalizes space
reconnaissance nor recognizes its legality.
Nevertheless, as a practical matter, the SALT provisions have inhibited
any Soviet challenge to overhead reconnaissance; post-Sait writings have
ignored the issue. In other regards, however, these Soviet writings have
argued against the right to use satellites for direct television broadcasting or
for surveying natural resources without the target nation’s permission.41
Thus, the theoretical basis for an objection to space-based reconnaissance
still exists. As one Soviet work puts it, “ It should be noted, however, that
permission to use space objects for functions of verifying compliance with
[the SALT agreements] cannot serve as grounds for an extended interpre­
tation of the verification provisions.” 42

Comparison of Humint and Techint

In the past several years, something of a debate has been going on in the
United States concerning the relative importance of humint and techint. On
one side are those who view technical intelligence as primary and as likely
to keep increasing in relative importance as technology advances; on the
other are those who think that the United States has overemphasized tech­
nical intelligence at the expense of traditional espionage and that it is nec­
essary to right the balance.
Perhaps the most striking recent contribution to the debate has been the
spirited defense of techint’s primacy in the memoirs of Admiral Stansfield
Turner, President Carter’s director of central intelligence (DCI):

Now that we have technical systems ranging from satellites traveling in

space over the entire globe, to aircraft flying in free airspace, to miniature
sensors surreptitiously positioned close to difficult targets, we are ap­
proaching a time when we will be able to survey almost any point on the
earth’s surface with some sensor, and probably with more than one. We
can take detailed photographs from very long distances, detect heat
sources through infrared devices, pinpoint metal with magnetic detectors,
distinguish between barely moving and stationary objects through the use
of Doppler radar, use radar to detect objects that are covered or hidden by
darkness, eavesdrop on all manner of signals from the human voice to
electronic radio waves, detect nuclear radiation with refined Geiger
counters, and sense underground explosions at long distances with seis-
 I tlllU O V * —O
engines, the magnetism in their armor, or photographs. A nuclear weap­
ons plant emits radiation, has a particular external physical shape, and
receives certain types of supplies. One way or another, we should soon be
able to keep track of most activities on the surface of the earth, day or
night, good weather or bad.43

When we consider the sheer technological prowess involved in creating

these sorts of technical collection devices, it is indeed breathtaking. Nev­
ertheless, admiration for the technology can easily lead to an overestimation
of what techint can accomplish and a concomitant depreciation of human
intelligence. In the United States, this view was particularly prevalent dur-
ing the mid- and late 1970s; since that time, however, humint’s importance
has received renewed recognition.
The strengths of techint are well described in the passage cited above. For
the United States, technical intelligence has been indispensable for collect­
ing the major part of the information it obtains about the Soviet Union. It
should be noted, however, that a large part of this information is publicly
available in many other countries. Information about the size and general
composition of the armed forces and the development of major new weapon
systems, for example, is available in most countries from official (especially
budgetary and parliamentary) documents. With respect to the Soviet Union,
however, intelligence methods are required to collect such information be­
cause of the regime’s pervasive secrecy on military issues, a secrecy that has
continued, in attenuated form, into the era of Gorbachev and glasnost.
Thus, much of the technical collection effort is devoted to countering the
effects of this secrecy; with respect to many of the categories of information
all countries keep secret— such as political and military intentions and
plans— human intelligence collection still may be required. Proponents of
humint stress that, despite technological progress, old-fashioned espionage
remains necessary to provide information about the intentions, political
activity, and strategic concepts of an adversary’s leadership. Indeed, under­
standing the adversary’s intentions, his strategy, and his perception of the
situation in which he finds himself are often the most important intelligence
information we could have.
Interception of communications among the highest levels of an adver-
sary s political and military leaders (high-level comint) can also provide
,ntelligence of this sort; it would be similar to having a human agent with
excellent access to the adversary’s leadership. Obtaining this information
r°ugh comint, however, is much more difficult than the Turner passage
tive communications; as noted in the next chapter, it appears that techno­
logical advance (in terms of increasing computer capabilities) aids those
trying to develop more secure codes and ciphers rather than those trying to
break them.
In this connection, note that human intelligence collection can provide
codebooks or other information that facilitates the decrypting of messages.
For example, in the early 1930s, the French intelligence service obtained
from an agent in Germany important information about the German coding
machine, Enigma, including instructions for its use, some details of its
construction, and the keys (or daily settings) indicating the position to which
the machine was set on a given day.44 This information was shared with the
Polish intelligence service, which achieved the first major cryptanalytic
breakthroughs that formed the basis of the eventual British success during
World War II.
Second, most comint capabilities can be defeated by keeping messages
off the air, that is, transmitting them via wire instead. Intercepting such
messages requires gaining access to the wires, which is likely to be difficult
in most cases. Furthermore, a new means of transmitting signals, fiber
optics, uses coherent light (laser beams) to carry messages and is relatively
invulnerable to wiretapping.
In addition, all types of technological intelligence collection, including
comint, suffer from a potential “ embarrassment of riches.” Despite the
global reach of the technical collection systems, as described by Turner,
there must be some method by which they are told which targets to observe
and which to ignore. While it may be correct that “ we [soon] will be able
to survey almost any point on the earth’s surface with some sensor,” no one
could survey them all at the same time, and even if one could, one would
not be able to process all the data the sensors could collect.
Thus, from the existence of technical capabilities such as described by
Turner, we cannot conclude that nothing of importance on the earth’s sur­
face can escape the notice of an intelligence service with an up-to-date
technical intelligence collection capability. Everything depends on the abil­
ity to target the sensors appropriately. Human intelligence collection can
provide the essential first indication that something of interest is occurring
or will occur at a given location. This makes it possible to target the tech­
nical systems on that area. Without such clues, the technical systems would
be less efficient and might miss important developments either for long
periods or altogether.
With respect to some kinds of intelligence information, the problem of
correctly targeting the technical sensors may be virtually insoluble. For
about nongovernmental targets (such as terrorist organizations) that lack the
fixed facilities or communication networks that are vulnerable to technical
collection. In other words, the fewer known locations that a group can be
associated with, the harder it is to target technical sensors on it. Intelligence
collection against such groups is likely to depend heavily on the ability to
infiltrate the group or to recruit its members as informants. If groups of this
sort become important intelligence targets, human intelligence collection
will become more important as well.
A human intelligence source can also provide the clues to interpret the raw
data gathered by a technical collection system. Even with a good picture of
a building, for example, an intelligence analyst may not be able to determine
its function; a human source familiar with it may be able to explain that the
presence of a certain detail, not otherwise remarkable, in fact indicates that
the building was designed for a specific purpose connected with a specific
military program. Without the human source, it might have been unlikely that
the detail would have been noticed or its significance understood. But once
this “ signature” is recognized, pictures of similar buildings can be examined
to see if the same detail is present.
As these examples illustrate, humint and techint can serve complementary
roles. The intelligence services of a global power such as the United States
will continue to require both. Regardless of any technological advances that
enhance technical intelligence collection capabilities, human intelligence
collection will still be necessary.

Open-source Collection
Publications and Broadcasts
No discussion of intelligence collection would be complete without refer­
ence to the gathering of information from open sources, that is, newspapers,
books, radio, and television broadcasts and any other public source of
information. Even in the case of a society as secretive as the pre-glasnost
Soviet Union, a lot of valuable information could be gleaned from such
sources; with more open societies, even greater amounts of information are
openly available.
^le lmP°rtance of open sources in the intelligence process is a matter
0 dispute and is ultimately tied to some basic questions about the nature
«intelligence. One view, expressed by Sherman Kent, is that the bulk of
'g -level foreign positive intelligence . . . must be through unromantic
° pen ancbabove-board observation and research.” 45 The more traditional
 view, on the other hand, has held that while open sources may provide
important context and background, the key facts, such as an adversary’s
specific intentions, must be obtained primarily, if not solely, from nonpublic
sources by means of espionage or technical collection.
This question is discussed at length in chapter 7. At this point, the focus
will be on the various uses to which open-source information can be put. For
example, a standard task of military intelligence officers has been to prepare
vast compendia of information concerning countries in which future military
operations may have to be conducted. Such compendia should include all
information that might be useful to a military staff preparing such an oper­
ation. Information concerning the country’s geography, its communications
and transportation networks, key military and economic facilities, and so
forth would all have to be included in sufficient detail to allow the planner
to make many decisions, such as determining how quickly troops and sup­
plies could be moved from point A to point B.
Except in the case of the most secretive countries (such as the Soviet
Union, where it was admitted in 1988 that previously published maps had
been falsified as a security measure46), much of this information will be
available in open-source literature: road and railroad maps and timetables,
newspaper and magazine articles, government economic and statistical re­
ports, and even old travel guides. Collecting and cataloging it are other
matters. Huge amounts of information are involved and the resources to
collect it for every relevant country are unlikely ever to be available. If, as
discussed in the next chapter, the U.S. military force that invaded Grenada
in 1983 was not supplied with adequate maps and other information about
the island, this was not due to any difficulty in collecting the information but
rather to the failure to allocate sufficient resources to what probably seemed
like a small, out-of-the-way island.
Similarly, a vast profusion of statistical sources publish economic data
throughout the world. Deciding what should be collected, however, requires
some sense of what intelligence requirements are likely to be. Economic
data may be needed to support the formulation of policies on international
trade or economic sanctions, to deal with the effects of shortages and inter­
national cartels (such as the Organization of Petroleum Exporting Countries
[OPEC]), and so forth. Other kinds of economic data, such as might result
from the monitoring of cash flows, could shed light on the international drug
trade. Finally, economic data inay be needed to support political and mili­
tary analyses; for example, predictions of Soviet behavior in the late 1980s
were dominated by questions of how the Soviet economic situation would
affect other areas of policy. Each of these intelligence requirements calls for
different sorts of economic data to be collected.
 In addition to these obvious types of open-source materials, intelligence
analysts require the same sorts of information— speeches by prominent po­
litical figures, texts of laws and resolutions, census and other demographic
data—-that would be required for any academic analysis of the political or
social conditions in a foreign country; indeed, the end product may be
similar as well.
Depending on the available resources, an intelligence agency may be able
to collect specialized data to a degree that would overwhelm an academic
researcher. For example, an important technique of kremlinology (the study
of Soviet leadership politics) involves tracing the careers of midlevel Soviet
officials to determine which midlevel officials are the protégés, or “ cli­
ents,” of which Politburo members or other senior officials. Once these
patron-client relationships have been identified, an analyst can, by noting
the promotions and demotions of the clients, determine the relative stand­
ings of their high-level patrons.
To do this, however, requires the collection and maintenance of an ex­
tensive biographic data base. It would involve scanning all major Soviet
newspapers and news broadcasts (those originating from the capitals of the
various republics as well as those from Moscow, including inter alia the
governmental, party, trade union, and military organs); recording every
news item that indicates the promotion, demotion, or transfer of an official
or that links two officials; and computerizing the resulting data base to make
it useful for researchers. A scholar, or even an academic institute, might be
hard-pressed to undertake such a task.47

Diplomatic and Attaché Reporting

Another means of collecting information, which might be regarded as a
composite of humint and open source because it is obtained by human agents
but in a straightforward and aboveboard manner, would be the reports
diplomats and military attachés file concerning events in the country to
which they are posted. Indeed, the line between diplomat and intelligence
0 lcer ^as not always been clear-cut. As one historian of diplomacy primly
pats it, Ambassadors sometimes readily crossed the nebulous line between
Ultimate gathering of information and espionage and other ill-reputed
activities. *8 But even today, when the two functions are separate, diplo-
tantIC rep0rtS on the Political situation in the host country can be an impor-
P mput t0 any P°litical analysis. A diplomat who has good access to major
in tItlCa* é tir e s in a country should be able to provide insights into the
crnal political situation that would not be found in the media.
lfnilarly, military attachés, to the extent they have access to military
officers of (heir host country, can gain insights into the host country's
military establishment, such as the personalities of the leading military
officers and their competence, characteristic ways of thought, views on
military doctrine, and relations with the civilian leadership. In addition,
attachés may be invited to observe military exercises or attend reviews or
ceremonial occasions where new military equipment is paraded or flown by.
They also may be able to travel in the host country and thus observe air­
fields, harbors, and other military and civilian installations of interest.49

Other O vert Hum an Sources

Diplomats and attachés are not the only foreigners who visit a country and
report their observations. For example, the 1987 treaty between United
States and the USSR on the elimination of intermediate-range and shorter-
range missiles (the INF Treaty) provides that each party send inspectors to
specified military facilities of the other party to ascertain whether the coun­
try is complying with the treaty’s terms. The inspectors have greater access
to those facilities than attachés or other visitors would have. Obviously,
their reports will contain valuable information concerning their host’s mil­
itary forces.
Similarly, businessmen, scientists, and other travelers learn information
about the countries they visit that, while not officially secret, is nevertheless
not available in the public media. Whether this information makes its way
to their country’s government depends, in a country whose citizens do not
need special permission to travel abroad, on the willingness of such travelers
to report it. The notoriety achieved by the U.S. Central Intelligence Agency
has probably meant that, since the 1970s, it has benefitted less from this sort
of voluntary cooperation. But the general increase in travel to and from the
Soviet Union has meant that the opportunities for such overt human collec­
tion have grown tremendously.50
 3
WHAT DOES IT ALL
MEAN?
INTELLIGENCE ANALYSIS AND
PRODUCTION
What Is Analysis?
Analysis refers to the process of transforming the bits and pieces of infor­
mation that are collected in whatever fashion into something that is usable
by policymakers and military commanders. The result, or “ intelligence
product,” can take the form of short memorandums, elaborate formal re­
ports, briefings, or any other means of presenting information. This section
indicates the breadth of analytic techniques. In describing these activities,
standard intelligence terminology is not always used; indeed, analysis does
not have any standard categories. Furthermore, the categories that do exist
are neither precise nor mutually exclusive. An attempt has been made to
arrange the techniques from the most technical, such as the decrypting of
coded messages, to the most speculative, such as the predicting of future
social and political trends. A subsequent section discusses the variety of
intelligence products and the functions they serve.

Technical Analysis
I use the term “ technical analysis” here to refer to analytic methods that
transform highly specialized data, totally or virtually incomprehensible to
everyone but the specialist, into data that other intelligence analysts can
Use‘ The examples discussed are cryptanalysis (which transforms seem-
tngly random strings of letters or numbers into the text of a message in a
novvn language), telemetry analysis (which transforms a radio signal into
group of time series describing the performance of a missile or other test
lc e '» and photo interpretation (which identifies and measures objects
In a Photograph).

37
Cryptanalysis
Cryptanalysis refers to the solving, or “ breaking,” of enemy codes and
ciphers, thereby enabling analysts to transform an intercepted encrypted
message into its original, meaningful form. In most cases, the interception
involves the reception of radio signals by someone other than their intended
recipient. However, the cryptanalytic problem is the same in the case of an
encrypted message carried by a captured courier, an encrypted letter opened
by postal censorship, or an encrypted telegram obtained by tapping a tele­
graph cable.
In technical usage, the terms “ code” and “ cipher” refer to different
methods of encryption. In a code, a word or phrase (signifying a thing,
concept, or location) is replaced by the group of digits or letters (which may
or may not form an actual word) that is found opposite that word or phrase
in the codebook. Thus, the message “ attack on Saturday” would be en­
crypted by finding the code group for “ attack” (say fg hj ) and that for
“ Saturday” (say adfk ), thus producing the encrypted message fghj adfk .
If a word is not found in the codebook, the sender can either reword the
message to use words and phrases in the codebook or spell out the missing
word using special code groups of individual letters or syllables.
In a cipher, on the other hand, each letter in the original message— called
the “ plaintext”— is replaced, following some formula or algorithm, by
another letter, thereby forming the “ ciphertext.” (In a transposition cipher,
the letters of the original message are retained, but they are transmitted in a
jumbled order, according to some scheme.) For example, a cipher might
consist of the rule that each letter be replaced by the letter following it in the
alphabet. In this extremely simple cipher, the message in the example used
above would be enciphered as buubdl po tbuvsebz .
A more complicated cipher (still much too simple to offer any security)
would replace the first letter of the message by the letter following it in
the alphabet, the second letter, by the letter two places down in the al­
phabet, and so forth. (For the purposes of such a cipher, the alphabet
would be envisaged as written in a circle, with the letter A immediately
following the letter Z. Thus, the twenty-sixth letter of the message would
be unchanged, while the twenty-seventh, like the first, would be replaced
by its immediate successor, and so on.) The message would now read
BVWEHQ VV BKEGERPO.
These techniques can be combined in various ways. For example, a coded
message may then be enciphered in the same way a plaintext might be.
Thus, our example message encoded as before and then enciphered by the
first method discussed would yield the following “ superenciphered” text:
GH*K B E O L . lilt flu v am ag v , W i O U V sll u y iw v u w iv to vitv OWUI 11
o^erS__to recover the original message, an adversary must break both the
code and the cipher. Typically, the code would remain in effect for a long
time, while the cipher algorithm would change frequently. The reason is that
while the former is quite difficult to change (each change of code means new
codebooks must be distributed to each embassy, headquarters, or post that
might send or receive messages), it is a simpler matter to change the cipher.
Cryptanalysis refers to the solving or breaking of both codes and ciphers—
that is, to reconstructing the adversary’s codebook or figuring out the method
he is using to encipher messages— using primarily the encrypted messages
themselves. (Decryption, on the other hand, usually refers to the intended
recipient recovering the plaintext.) Cryptology refers to the more general
study of these things, both for cryptanalytic purposes and for devising more
secure codes and ciphers. Cryptography is sometimes used to refer specif­
ically to the latter activity (the devising of codes and ciphers) but is some­
times used synonymously with cryptology.
The typical raw material with which cryptanalysis begins is a collection
of encrypted messages. Cryptanalysis’s task is to discern whatever patterns
exist in the apparently meaningless jumble of letters or numbers and to relate
those patterns to the known patterns that exist in the language in which the
messages were presumably written. For example, in ordinary English text,
the letter £ appears more frequently than any other; therefore, in a group of
messages enciphered in the simple substitution cipher discussed above (in
which each letter of the plaintext is replaced by one immediately following
it in the alphabet), the letter F would probably be the most frequent.
Noticing this fact, the cryptanalyst would begin by assuming that F stands
tor E. He then might notice that U1F appears more frequently than any other
three-letter word and assume that it stands for THE. At this point, he might
well guess that each letter in the ciphertext stands for the letter immediately
preceding it in the alphabet and, trying it out, would discover that this was
indeed the case.
As may be seen by even this simple example, the amount of raw material
a cryptanalyst can obtain is an important factor in determining whether a
solution can be achieved. In a short message, E may not be particularly
requent: in our example ( “ attack on Saturday” ), it does not appear at all.
tak* ^°n®er messa2e >or series of messages, however, the laws of probability
e over, and it would be extremely unlikely for the frequencies of the
various letters (or small groups of letters) in the plaintext to differ much
r°m the Irequencies with which they appear in ordinary English text,
into C .^St0ry cryptology is a fascinating one, and it is impossible to go
0 tt in detail here.1 For our purposes, it will be sufficient to divide that
history into three periods, according to the technology that was available for
encrypting messages.
In the first period, running from antiquity to the 1930s, messages had to
be encrypted by hand. Various techniques were available including code­
books and substitution and transposition ciphers. By the end of this period,
codebooks had grown to be cumbersome affairs, with thousands of entries.
However, since they tended to be used for long periods, cryptanalysts could
eventually reconstruct them. Similarly, cryptanalysts were helped by the
fact that the complexity of ciphers was limited because all the steps involved
in producing the ciphertext had to be performed by hand. Because the
possibility of an encryption or decryption mistake increases as a cipher
becomes more complex, it was not always practical to introduce more
difficult ciphers.
Of course, the easiest way to break a code is to steal, capture, or other­
wise obtain a copy of the codebook, without letting the enemy know about
it. Since producing and distributing a new codebook to everyone who would
need it is not an easy matter, a code was likely to remain in use for a long
period of time, thus enhancing the usefulness of the captured book.
Before and during World War I, all the major European powers had
special offices for decrypting foreign diplomatic and military messages. At
the beginning of World War I, for example, the British managed to obtain
(from captured or sunken ships) copies of the major German naval code­
books. Although the Germans took the precaution of further enciphering the
coded messages, the British decrypted them nevertheless.2
However, probably the most important British cryptanalytic success in­
volved German diplomatic, rather than naval, codes. On January 17, 1917,
British cryptographers decoded a message, known to history as the Zim-
mermann telegram, in which the German foreign minister directed the Kai­
ser’s ambassador to Mexico to propose a German-Mexican alliance against
the United States, in case the United States reacted to German initiation of
unrestricted submarine warfare (scheduled for February 1) by declaring war
on Germany. Subsequent publication in U.S. newspapers of the telegram,
which envisaged Mexico recovering Texas, New Mexico, and Arizona,
created a fire storm that did much to make President Woodrow Wilson’s
decision to go to war inevitable. The release of the telegram was handled so
skillfully that the Germans did not suspect that their code had been broken.'
The story of American cryptography during World War I and the follow­
ing decade illustrates how, despite the immense intellectual challenge in­
volved, it was still possible at that time for an individual, working by
himself or with a small group, to discover the basic principles of the art and
to solve codes and ciphers used to encrypt the most sensitive messages.
 d Herbert Yardley relates how he was able, in several hours on his
ln tQ cryptanalyze a message to President Wilson from Colonel Edward
u ^ s e Wilson’s personal representative then on a diplomatic mission to
Germany.4 Yardley’s account of his leadership of the U.S. Army’s crypt-
analytic unit once the United States entered the war and of the State De-
' rtment’s Black Chamber after the war’s end indicates the extent to which
cryptography was, at that time, an art rather than a science and how it
depended on the insights and intuitions of individuals working essentially by
themselves.
However, in the case of cryptology, as in many other areas of military and
intelligence activities, World War I marked a crucial stage. According to a
historian of cryptography,
The First World War marks the great turning point in the history of
cryptology. Before, it was a small field; afterwards, it was big. Before, it
was a science in its youth; afterwards, it had matured. The direct cause of
this development was the enormous increase in radio communications.5

Increased radio traffic meant a greater advantage to be gained by break­

ing an enemy’s code or cipher. At the same time, the increased amount of
traffic meant that, other things being equal, codes or ciphers of the same
complexity became more vulnerable to cryptanalysis, since the more raw
material (encrypted in the same code or cipher) available, the greater the
chance that the cryptanalyst will detect the repeating patterns that make a
solution possible.
More radio traffic also meant that manual systems for encrypting and
decrypting messages were more and more overloaded. This, plus the desire
tor more secure (more complex and thus harder to break) encryption sys­
tems, led to a second period, in which mechanical or electromechanical
devices, such as the American Hagelin, the German Enigma, and the Jap­
anese ‘‘Purple” machines, were used to encipher texts.6
Such machines incorporated a much more complicated enciphering algo­
rithm or formula than that used by previous cipher systems. More compli­
cated cryptanalytic techniques and procedures were therefore needed to
S°lve these ciphers. Breaking them involved the development and use of
primitive computers as well as advances in statistics and other branches of
e°retical mathematics.7 The story of solving electromechanical cipher
niachines such as Enigma and “ Purple” is contained in the literature on
^rld War II cryptography.
both 6 arC ^ Per^ ’ which *s characterized by the use of computers
p encipher and decipher messages and to support cryptanalytic efforts.
vious reasons, this subject is very sensitive, and not much information
is publicly available about it. The rise of “ public cryptography”— the study
and use of cryptography by independent scholars and businessmen— has
meant that some information about it has come into the public domain.8
However, it is not possible, from unclassified sources, to determine how the
sophistication of public cryptography’s methodology compares to the real
thing.
Nevertheless, a few things can be said about the current situation and the
relative cryptologic strengths of the United States and USSR. Given the
centrality of computers, it would seem likely that the U.S. lead in that area
gives it an important advantage; to some extent, this may be balanced by
Soviet strengths in the area of theoretical mathematics.9
At the same time, it would appear that advances in computer power (in
terms of the speed with which the basic arithmetic operations may be per­
formed) favor the defense. In other words, although an increase in com­
puter power enables both the encrypter to use ciphers of greater
complexity and the cryptanalyst to solve ciphers of greater complexity, the
net effect is to favor the encrypter. A given increment in computer power
allows complexities to be introduced into the cipher that the same incre­
ment in computer power cannot unravel. If this is so, then the long-term
future of cryptanalysis, at least as it attempts to break the ciphers of major
powers using the most advanced technology, is in principle not very good.
In practice, of course, because of some error in its construction, a given
cipher may not take full advantage of the available computer technology
and is therefore weaker than it should have been. In addition, mistakes in
using the cipher may allow cryptanalysts to solve it.10 Finally, espionage
may bring solutions to cryptologic systems that could not have been
solved by cryptanalysis.

Fragility o f Cryptanalysis
Cryptanalysis is among the more fragile of intelligence methods, since a
country usually changes its cryptographic system if it realizes that an ad­
versary can read its encrypted messages. Thus, it is not surprising that the
American and British cryptanalytic successes during World War II were
considered to be among the most vital secrets of the war, worthy of the most
careful security arrangements that could be devised. In the case of the
British success in mastering the German Enigma machine, the secret was
kept for about thirty years, despite the large number of people involved in
the operation.
The American case was far different. In what might be the worst security
breach of the war, the Chicago Tribune published, on June 7, 1942, an
 • 1 on the U.S. victory at the Battle of Midway. Headlined “ Navy Had
Word of Jap Plan t0 Strike at Sea,” it cited materials derived from decoded
nese messages. In particular, it included the names of not only the
P nese carriers involved, but also the light cruisers that supported the
would-be occupation forces. Furthermore, it asserted that when the Japanese
fleet moved toward Midway, “ all American outposts were warned” and that
U S naval intelligence had been able to predict “ a feint at some American
base [Dutch Harbor, in the Aleutians], to be accompanied by a serious effort
to invade and occupy another base [Midway Island].” 11
A Japanese intelligence expert would probably conclude from the article
that the United States had been reading coded Japanese naval messages.
Although the matter was referred to a grand jury in Chicago, however, no
indictment was brought. It appears that the key U.S. officials ultimately
decided against prosecution on the grounds that a trial would only call
attention to the article and probably require a public confirmation of the fact
that the United States was reading coded Japanese naval messages. (Ironi­
cally, on August 31, 1942, this very fact was mentioned by a congressman
in a floor speech castigating the Tribune's “ unthinking and wicked misuse
of the freedom of the press.” 12)
On August 14, 1942, the Japanese Navy introduced a new version of its
code (dubbed JN-25d by U.S. cryptanalysts). Although the appearance of a
new version of the code so soon after the previous version(JN-25c) had been
introduced (on June 1) was suspicious, postwar researchers have not dis­
covered any compelling evidence that the change was motivated by the
article.13 Furthermore, the new version of the code “ retained the charac­
teristic of the broad JN-25 formula,” thus allowing the United States to
regain access to the Japanese message traffic more easily than if the Japanese
had instituted an entirely new system.14 In general, the historians who have
written on the issue have believed that the existence or, in any case, the
S1gnificance of the Tribune leak escaped the notice of the Japanese altogether
and that the August code change was merely a routine one.
Nevertheless, it did take the United States about four months to become
as familiar with the JN-25d code as it had been with previous versions.
urthermore, the difficulties involved in developing a new cryptographic
system and distributing it to the fleet could easily explain why the Japanese
faction was limited to merely advancing the date of an already planned
C°de change. Recent research has uncovered a reference in Japanese diplo­
matic traffic to a request to the Japanese Embassy in neutral Lisbon for
newspapers, particularly the antigovemment ‘Chicago Tribune,’ with as
any ^ack issues as possible.” 15 Although far from conclusive, this sug-
Sts ^at Japanese authorities were aware of the Tribune article on Midway
and at least concerned enough about its possioie sigiuucain^ w — w% ,
it for themselves.
A more clear-cut example of the fragility of cryptanalysis, involving
Britain this time, occurred in 1927. During the preceding years, the British
cryptanalytic bureau, the Government Code and Cypher School, had been
able to decipher messages between the Comintern headquarters, or the So­
viet government in Moscow, and the various Soviet representatives in Lon­
don (the Soviet Embassy as well as a trade delegation and a trade office of
Arcos, the All-Russian Cooperative Society). By this means, the British
government tracked the activities of the Soviet diplomats and trade repre­
sentatives, activities that included espionage and involvement in trade union
affairs. In 1927, when the British broke off diplomatic relations with the
Soviet Union, these departures from diplomatic decorum were cited as the
reasons for the decision. However, the British ministers, when pressed in
Parliament to defend their action, were indiscreet in discussing the matter
and revealed publicly their ability to read encrypted Soviet messages. The
predictable result was that the Soviet Union introduced new ciphers, which
the British were unable to break.16
A similar case, in which a democratically elected government revealed
sensitive cryptanalytic capabilities in order to build support for its policies
and thereby lost access to encrypted messages, seems to have occurred in the
United States in 1986. Apparently to justify its eventual decision to bomb
government facilities in Tripoli, Libya,

President Reagan [the Washington Post reported] and his top advisers
made an extraordinary disclosure of sensitive intelligence information
. . . to demonstrate that United States has hard evidence that Libya . . .
was directly responsible for the bombing of a West Berlin nightclub. . . .

The specifics cited by the president . . . sources said, will make it clear
that the United States has the capability to intercept and decode Libya’s
sensitive diplomatic communications.17

The predictable result in fact occurred: “ The public disclosure of decoded

Libyan diplomatic cables has caused American intelligence analysts to lose
a valuable source of information that may take weeks or months to replace,
Administration officials said today.“ 18
These examples demonstrate a general point about intelligence that ap­
plies to other collection and analysis techniques as well: the better the
information analysts have available, the more inhibited they arc about using
it and perhaps alerting the adversary to their intelligence capabilities. This
 J'lplTlUlti AW
*- ■ —
0 j t0 reveal detailed intelligence information to persuaue me puuiiL uuu
Pre' k licies are reasonable. But a form of the same dilemma exists for any
^vernment, under any circumstances, whenever it possesses specific intel-
faence information on which it wishes to act.
^Thc problem is most likely to come up in wartime, when the incentive for
acting on the intelligence may be very great. For example, during World
War II the British knew, from intercepted and decrypted communications,
the precise schedule of the German ships bringing supplies across the Med­
iterranean to General Erwin Rommel’s forces in North Africa. Nevertheless,
to prevent the Germans from becoming suspicious about the security of their
communications, the British adopted the rule that no ship could be attacked
before it was overflown by reconnaissance aircraft, thus providing the Ger­
mans with an alternative explanation of how the ship came to be located and
attacked.
This point is also illustrated by the story that Churchill knew in advance
about the particularly destructive German air raid on the English city of
Coventry, but did nothing to alert its air defenses or emergency services for
fear of compromising the intelligence source. (The source in this case was
the ability to cryptanalyze Luftwaffe messages encrypted on the Enigma
machine.) Historical research, however, seems to have established that this
story is not true: the available communications intelligence did not allow the
British to identify Coventry as the raid's target.19 Nevertheless, the story
provides a particularly dramatic illustration of the moral and strategic di­
lemmas that a government may face in deciding whether to risk an intelli­
gence source by acting on the information it provides.

Telemetry Analysis

Telemetry refers to the radio transmission of information from a test vehicle,

su°h as a missile being test-fired, to a ground station. The information may
concern the vehicle’s performance (such as the amount of thrust being
generated by its engine or its acceleration), its condition (for example, the
emperature of various components or the amount of vibration it is under-
th at^ ° r ° l^er m*ss**e characteristics. The purpose is to provide the data
ation^ lneers conducting the test would need to assess the vehicle’s oper-
inf 0r'. ln case ° f failure, to determine the reasons for it. Using this
th e n v ^ 1011, —e teSt en8^neers can reconstruct the vehicle’s flight and assure
The^ VCS reliability or make modifications.
bgence^arne ra<^ ° s*®na*s’ however, can be intercepted by a foreign intel-
service, a type of intelligence collection known as telemetry intel-
used to deduce tne oasic characteristics or me test vemcie. course, when
the test engineers begin to interpret this data, they already know, from the
design of the telemetry system, precisely to which missile characteristic
each part of the data stream (or telemetry “ channel” ) refers. The telemetry
analysts working for a foreign intelligence service, on the other hand, begin
with none of this information; they must figure out, from the raw data
themselves, what characteristic of the missile is being measured by each
channel.20 Once the analysts have made these determinations, they can
attempt to use the data to create a computer model of the missile, from
which such characteristics as initial weight (launch weight) and payload
(throw-weight) can be derived.
Since the mid-1970s, U.S. access to Soviet telemetric information has
become an important issue in the area of strategic arms control. During the
debate over the ratification of SALT 11 in 1979, it became clear how im­
portant access to telemetry was to the United States’s ability to verify Soviet
compliance with the treaty’s provisions. Several aspects of the question
figured centrally in the debate.
First, the Iranian revolution in 1979 meant that the United States no
longer had access to important “ listening posts” (sites favorably located to
intercept radio transmissions) in Iran. As a result, the question was raised
whether the United States would be able to intercept the needed telemetric
signals from Soviet ICBM missile test flights. Second, the treaty explicitly
addressed and partially prohibited the Soviet practice of encrypting some
telemetric signals, to conceal from the United States the information they
contain, but the provision was particularly vague; the question was raised
whether it constituted an adequate protection of U.S. access to telemetric
information.21
In actuality, Soviet encryption of telemetry increased tremendously dur­
ing the SALT II period.22 This eventually led to President Ronald Reagan
declaring in 1984 and 1985 that
Soviet encryption practices constitute a violation of [SALT II obliga­
tions]. The nature and extent of such encryption of telemetry on new
ballistic missiles, despite U.S. request for corrective action, continues to
be an example of deliberately impeding verification of compliance in
violation of this Soviet political commitment [to abide by SALT II].23
It is not clear what the prospects are for any attempts to cryptanalyze
encrypted telemetry. In any case, aside from encryption, other methods can
be used to deny telemetric information to an adversary’s intelligence ser­
vice. For example, the data may be recorded on a tape inside some kind of
 14 f that can survive reentry into the atmosphere and the impact ot the
a*rCra a and be recovered. Similarly, telemetric data from a cruise missile or
tendmg^entai pjane may be broadcast at very low power to a “ chase” plane
expcH ancj recor(ied on board the plane; if the radio signal is suffi-
* 3 . weaj^ it might prove impossible to intercept it outside the test range.

photo Interpretation
Despite the sophistication of the equipment that can take pictures deep
within otherwise inaccessible territory, no substitute has been found for the
human eye when it comes to interpreting what those photographs show. This
is not so simple a task as might be thought: while it is often said that
photographs are a particularly persuasive form of intelligence (in the sense
that senior officials feel more confident about the intelligence they are get­
ting when they can see for themselves), the average photograph is likely to
be unintelligible to the layman. It is only after the photo interpreter (PI)
points out and labels the interesting items that ordinary viewers can under­
stand what they are seeing.24
A photograph’s quality is typically measured by what is called the
“ ground resolution distance.” While the precise meaning of this term is
quite technical, it may be thought of as the minimum-sized object (in terms
of its length and width) that is distinguishable from neighboring objects or
from the background. While many other factors must be taken into account
to determine how well a photograph can be interpreted, ground resolution
distance is “ nevertheless . . . a convenient measure, useful in making gross
comparisons and evaluations.” 25
The actual ground resolution distances U.S. photographic surveillance
satellites can obtain are classified, because this information would tell ad­
versaries what details (of military equipment, facilities, and so forth) the
United States can, or cannot, see. Planners of any deception efforts also
would find this information useful, since it would tell them how closely a
decoy would have to resemble the real object for the two to be indistin­
guishable to photographic surveillance.
However good the resol a, all sorts of important information cannot be
earned directly from a ph ^ograph. Pis must often work from what are
ed signatures,” that is, specific details that, on the basis of experience
^lc’ are associated with a certain piece of equipment, the carrying out
f o n - ain activity, the intention to take a certain step in the future, and so

For
example, the lengthening of an airfield runway would indicate the
experience, a Pi might know that a given runway lengin was correlated with
a given type of airplane and could project that aircraft was to be stationed at
that airfield. Similarly, in a lighter vein, it might be possible to distinguish
a military base for Cuban soldiers from one for Soviet troops by the presence
of a baseball diamond, as opposed to a soccer field.
One method of using signatures that the United States has relied on goes
by the fanciful name of “ crateology”— the correlation between military
equipment and the type and size of crate in which it is usually shipped. Once
it has been determined that, for example, MiG-21 s have been shipped in
crates of a certain size, crateology suggests that other boxes of the same size
whose contents have not been observed directly may be assumed to be
MiG-2 Is.26
The key to using such signatures is the fact that most organizations—
military ones in particular— tend to follow routines, to do things by the
book. Thus, when the Soviets decided to deploy intermediate-range ballistic
missiles (IRBMs) in Cuba, they first prepared the missile base defenses by
deploying SA-2 antiaircraft missiles around it in a standard trapezoidal
pattern. By observing this pattern, and identifying it as being the same as
that used at older 1RBM bases in the Soviet Union, analysts could have
predicted the future emplacement of IRBMs in Cuba.27
The use of signatures can be very productive. But if it becomes known
what signatures Pis are using, an adversary can use their dependence on
them as a means of deception. In the Cuban missile crisis discussed above,
it seems apparent that the Soviet Union did not realize that the trapezoidal
pattern of SA-2 emplacements was known to the United States as a signature
of an IRBM base; given the effort the Soviets invested in keeping the IRBM
deployment secret, their failure to alter the pattern in which SA-2s were
emplaced seems simply a mistake.28 The absence of the typical pattern
might have suggested to U.S. intelligence analysts that IRBMs would not be
deployed in Cuba.

Data Banks (Basic Research)

As already noted in the section on open-source collection, a major function
of an intelligence service is often the assembly and maintenance of large
data bases covering many topics of interest to a government’s foreign and
military policy.
In one aspect, this work takes the form of compiling encyclopedias of
relevant information on all countries where military activity could conceiv­
ably take place.29 Such compendia are designed to support a military plan-
 contain detailed miormauon aoout military lorces in nemg, transport
heIlCe ilitary facilities, and military geography as well as information about
and country’s politics, demographics, industry, agriculture, and so forth.
l*lC\Vhile the analytic problems confronted in compiling such an encyciope-
. are minimal, it is a major task to try to sort out what information is
^tant and what is not. Decisions on which countries to cover and to
'vhat degree of detail can be very difficult to make yet have important
consequences if the information is needed to support a military operation.
For example, the plan for the invasion of Grenada in October 1983 by the
United States and the countries of the Organization of Eastern Caribbean
States was developed very rapidly. One consequence was that adequate
military maps were not available. According to Admiral Wesley McDonald,
commander in chief of U.S. forces in the Atlantic, “ The Army, particularly
the troops on the ground, were operating initially from roadmaps or other
types of maps which made it very difficult for them to determine in [s/c] their
grid coordinates. That is one of the lessons learned. . . ,” 30
Other types of data banks serve the needs of intelligence analysts them­
selves. For example, as already noted, collating and analyzing vast amounts
of unclassified data can provide insight into a closed society such as the
pre-glasnost Soviet Union. Biographic data files (containing all newspaper
references to party and government officials, their positions, promotions,
travel, and so forth), for instance, can be analyzed to shed light on the
relationships among officials.

Production of Finished Intelligence

The types of analysis described above are the necessary initial steps of the
analytic process; for the most part, the results of this work serve as sources
lor other intelligence analysts but do not go directly to the policymaker or
military commander (the ultimate consumer of intelligence). This section
iscusses t^ie kinds of finished intelligence produced. The analytic tech­
niques used are neither as technical as cryptanalysis or telemetry analysis
nor as unique to intelligence work. In some cases, such as the production of
economic or political intelligence, the techniques are not distinguishable
I0m t"ose of the corresponding social science.

! len,ific and Technical Intelligence

new s years immediately before and during World War II, the pace at which
scientific and technological principles were incorporated into weaponry
increased substantially. A nation’s ability to compete militarily uuam t
dependent on its technological level, and its ability to manufacture weapon
systems embodying that technology, as on its overall productive capacity,
the size of its military forces, or any other measure of military strength. This
was particularly true of air warfare, which saw the development and intro­
duction of air defense radar, sophisticated navigational systems for guiding
bomber aircraft to their targets, jet aircraft, ballistic and cruise missiles, and
finally, the atomic bomb. Understanding new weapon systems that the en­
emy was developing thus became an important objective of each nation’s
intelligence agencies. In many cases, it was important to obtain fairly de­
tailed information about the way the system worked in order to develop
methods of countering it.
This imposed a new task on intelligence systems: to predict the emergence
of new weapons and to understand them well enough to defend against and
to counter them.31 This required that the intelligence agencies have access
to specialized scientific and technical knowledge and that this capability be
integrated into the intelligence process.
Collection requirements for scientific and technical intelligence have to be
more precise than for other forms of intelligence; it is not obvious to some­
one unfamiliar with the new technology what is to be looked for and where.
This in turn means that close coordination is needed between the analysts
and the collectors. To some extent, the process must be similar to the
scientific method in which hypotheses (formulated by the analyst) are tested,
albeit by directed observations rather than by experiments. For example,
R. V. Jones, the head of the British Air Ministry’s intelligence section
during World War II, recounts how his theory concerning the nature of the
first system of radio signals the Germans used as a navigational aid for
bombers (Knickbein) was confirmed when an aircraft— fitted out with spe­
cial receivers and flying in an area designated by him— picked up the radio
signals he had predicted would be there.32
Analysis in the area of scientific and technical intelligence requires the
blending of intelligence and scientific or technical expertise. Although the
scientific principles themselves are, of course, universal in nature and known
to experts throughout the world, the way in which they are put to use in
technological developments can differ widely between countries. Thus, this
type of analysis requires enough familiarity with scientific and technical
developments to understand what an adversary is doing, combined with
enough imagination to realize that he might solve technical problems that the
analyst has not been able to solve or that he might reach a different, but
nonetheless valid, solution to them.
Military inieuig*'
. . intelligence deals with information about the military establishment
^ foreign country for planning one’s own military forces in peacetime or
° conducting military operations against it in time of war. The most
dementary military intelligence is what is known as the “ order of battle,”
ttie basic information about a nation’s military forces— amount of man-
Wer numbers and types of weapons, organizational structure, and similar
data (Characterizing this information as elementary does not, of course,
imply that it is easy to get or even that it is a simple matter to know which
types of auxiliary or reserve troops should be counted and which should
not ) Not strictly speaking part of the order of battle, but a necessary com­
plement to it, would be information about the qualitative aspects of a foreign
military establishment: how good the training is, what the quality of the
leadership is, and so forth.
One step up from this fundamental level would be information about how
the forces could be expected to fight: what tactics they have adopted and
trained for, how they envisage the nature of a future war (what in Soviet
terminology is called “ military doctrine” ), and what their strategy would be
for fighting it. The raw data for this intelligence can come from open sources
(such as military publications), from attachés’ or diplomats’ overt contacts
with the adversary’s military, from observing (either directly or via techni­
cal collection systems) deployments and exercises, or from human intelli­
gence sources with direct access to military plans. As with scientific and
technical intelligence, this sort of analysis calls for blending the intelligence
perspective (with its concentration on all available bits and pieces of evi­
dence about the adversary and what they tell about how he thinks) with that
ot the military specialist (with its understanding of weapon systems’ capa­
bilities and the kinds of warfare that can be waged with them).
But above all, it requires the open-mindedness to be able to imagine that
the adversary has adopted different solutions to common military problems
and that his solutions may well be appropriate for his circumstances and
resouices or even superior to one’s own solutions. In summary,

strategic planners need intelligence on the otherness o f the enemy—

>nte lgence that will reveal the enemy’s methods of operation, internal
jsputes, and ways of doing business, as well as the ways in which the
enemy differs from [oneself!.33

place^llf W^en mtBtary operations appear imminent, or are actually taking

ere must Be information about the disposition and movement of
known as “ indications and warning,” which addresses the task ot avoiding
strategic surprise.

Political Intelligence

Political intelligence consists of information concerning the political proc­

esses, ideas, and intentions of foreign countries, factions, and individual
leaders. The analysis that produces this intelligence is similar to that under­
lying all academic and journalistic public writing and speculation on both
international politics and the domestic politics of foreign countries. One
obvious difference, of course, lies in the existence of secret intelligence
sources, human or technical (primarily communications intelligence), that
provide information not available to the public commentator.
How important these additional sources are depends on both the nature of
the regime of the country being studied and the number and quality of
intelligence sources one has with respect to that country. The more demo­
cratic and open a society is, the easier it is to study its political life without
recourse to intelligence sources. Even in the case of a closed society, how­
ever, there may be open-source data that, properly analyzed, reveal much
more about its political life than would appear at first glance.
A more important difference between political intelligence analysis and
academic or journalistic work has to do with their different purposes. This
difference shows itself in various ways aside from the content of the work.
For example, intelligence analysts must provide answers in a timely fashion;
their work is useless unless it reaches policymakers before they have been
forced by circumstances to act. Academic authors, by contrast, generally
have the luxury of taking as much time as they feel is necessary to reach a
conclusion. Similarly, academics typically work alone and on subjects of
their own choice; intelligence analysts, meanwhile, are guided in their
choice of topics by the requirements of others, and their work must be
coordinated with others before it is completed.
The different purposes the works serve also affect the contents of intel­
ligence and academic analyses. For one, intelligence analyses must empha­
size the aspects of a situation that underlie the immediate issues facing the
policymakers; the focus in an academic work is presumably determined by
those aspects the researcher deems the most fundamental. This question is
discussed at greater length in chapter 6 on the overall relationship between
intelligence and policy.
 c and social intelligence deal with the same phenomena as do the
^ U11V" social sciences, at least as far as they are concerned with major
various jssues. As with political intelligence, the focus of the intel-
c°nte analysis should be determined by the information policymakers
ligence intelligence analysis in these areas is often less sophisti-
req^l^theoretically and mathematically— than corresponding academic
ca . t^iS does not seem to be due to differences between either the
circumstances under which intelligence and academic work are conducted or
their goals.
Not only is the method the same (at least in principle) but the academic
scholar is more likely in this area than in any other to have access to the
same data as the intelligence analyst, since little of it is acquired by intel-
liaence methods. Nevertheless, in some areas, intelligence methods may be
necessary. Much important economic data about the Soviet Union has been
and, despite glasnost, is still kept secret; this includes, for example, the size
of Soviet gold reserves and the Soviets’ annual sales of gold on the world
market. While intelligence analysts are often forced to estimate these num­
bers in the same way that any other economist might, an intelligence source
might be able to provide important clues or the official Soviet numbers
themselves.
In general, however, the contribution of intelligence sources to economic
and social analysis is small. Furthermore, the intelligence agencies cannot
expect to attract the highest quality analysts in these areas. The freedom of
academic life proves more attractive to talented social scientists, even if the
salaries are lower. In addition, multinational corporations, banks, and bro­
kerage firms can easily outbid intelligence agencies for the services of highly
qualified economists. In general, there is no reason to expect that analyses
by intelligence personnel of economic issues not involving societies that
keep basic economic data secret would be superior to those of economists in
|he business world.34 Of course, much of the economic analysis done in the
^usiness world will not be available to policymakers in the government, but
east some of it could be via consulting services and newsletters.

The Intelligence Product

reporTufUCt *nte^gence process can be any means, from a formal
analyzed * ^Urr*ec* conversation, by which an intelligence analyst transmits
information to the policymaker or military commander who needs
it and can use it. In this section I discuss the various forms of intelligence
product and the functions they serve. The categorization is general rather
than precise; the terminology reflects ordinary usage among U.S. intellU
gence agencies but is not completely standardized.
Broadly speaking, I divide intelligence output into three groups: current
intelligence, basic intelligence, and intelligence estimates. In this, I follow
Sherman Kent’s lead; he speaks, using his own coinages, of the current
reportorial form of intelligence, the basic descriptive fo rm , and the
speculative-evaluative form and describes them as relating to information
about the present, the past, and the future, respectively. 35 I use primarily
U.S. examples to illustrate the different intelligence products because more
detailed information is available. The distinctions, however, are generally
valid, although the emphasis placed on the different products may vary from
one intelligence service to another.

Current Intelligence
Intelligence agencies have a variety of products designed to inform policy­
makers of major new information that might affect policy. In this regard,
agencies serve a function similar to that of the news media.36 The range of
information that should be watched for and reported is as large as the entire
scope of the nation’s intelligence interest. Obviously, some priorities must
be established on the basis of what is, in Kent’s words, “ positively germane
to national problems which are up now and other problems which appear to
be coming . “ 37 While a formal system of priorities may exist, it must be
supplemented by the intelligence producers’ own judgments about what is
important to their consumers. This judgment, in turn, depends not only on
the intelligence analysts’ common sense but also on their being in close
contact with the policymakers.
In the United States, the best-known product of this type is the National
Intelligence Daily, or NID, a sort of classified newspaper containing a series
of short (several-paragraph-long) articles on the previous day’s events. Pre­
pared by the CIA in coordination with the other major intelligence
agencies— the Defense Intelligence Agency (DIA), the National Security
Agency (NSA), and the State Department’s Bureau of Intelligence and
Research (INR)— it is sent to several hundred top national security officials
in the government.
A more sensitive daily report is the President's Daily Brief, which has a
much more restricted circulation. According to Cord Meyer, a former career
CIA official, writing in 1980,
 • , d e s i g n e d to be read in ten or fifteen minutes by the President at the

'inning of each working day. It does not attempt to recapitulate what

h news media have reported in the last twenty-four-hour period, but
1 , t0 SUInmarize the significance of what secret sources have reported
that bears on current world developments.38
From time to time, the view is expressed that U.S. intelligence agencies
d too much time and effort fulfilling this “ current intelligence” func-
. t0 the detriment of their ability to analyze situations in greater depth.
F o r example, a Senate committee report dealing with the CIA’s analytic

branch, the Directorate of Intelligence (DDI), made the following com­

ments under the heading “ The ‘Current Events’ Syndrome” :
The task of producing current intelligence— analyzing day-to-day events
for quick dissemination— today occupies much of the resources of the
DDL Responding to the growing demands for information of current
concern by policymakers for more coverage of more topics, [sic] the
DDI has of necessity resorted to a “ current events” approach to much
of its research. There is less interest in and fewer resources have been
devoted to in-depth analysis of problems with long-range importance to
policymakers. . . .
According to some observers, this syndrome has had an unfavorable
impact on the quality of crisis warning and the recognition of longer term
trends. The “ current events” approach has fostered the problem of “ in­
cremental analysis,” the tendency to focus myopically on the latest piece
of information without systematic consideration of an accumulated body
of integrated evidence. Analysts in their haste to compile the day’s traffic,
tend to lose sight of underlying factors and relationships.39
The very persistence of this criticism over the years probably indicates
that the pressure toward a current events approach— the desire of policy-
makers to be kept informed of the latest hot news— is inherently strong and
. 1 °ften predominate over the intelligence analyst’s desire to conduct
,n~depth studies.40 The same pressures are probably at work in the intelli­
gence agencies of other countries as well.

lndications and Warning

tone most important functions an intelligence agency can perform is

the^r° V^ e Ume,y warnmg ° f hostile military action; in the nuclear era, it is
m°st important. And, in fact, the character of the post-World War II
U.S. intelligence system was probably determined more by the failure of the
prewar intelligence organizations to provide effective warning of the Japa­
nese attack on Pearl Harbor than by any other factor.
Because of its importance, and because judgments might have to be made
quickly, this intelligence function has been (in the United States, at least)
systematized to a much greater extent that any other. The system, known as
Indications and Warnings (I&W), is based on an analysis of the steps an
adversary would either necessarily or probably take to prepare for an armed
attack. These hypothetical events— the calling up of reservists, forward
movement of military forces, changes in communications patterns, and so
forth— are called “ indicators” ; when the event actually occurs and is ob­
served, it is then referred to as an “ indication.” Analysts determine how
great the threat is by the totality of indications and issue warnings at various
threshold levels .41
Oleg Gordievski, a KGB officer who worked for British intelligence from
1975 to 1985, has provided a glimpse into Soviet l&W procedures. Ac­
cording to him, in 1981, the Soviet Politburo ordered the KGB and GRU to
cooperate on a new worldwide monitoring system to watch for signs of a
possible Western nuclear attack. The GRU was responsible for watching the
purely military indicators, while the KGB was to watch for signs that the
political decision to launch a nuclear attack had been taken. Among the
indicators that the KGB residency in Britain (the KGB officers working out
of the Soviet Embassy in London) was responsible for monitoring were:

• the pattern of work at 10 Downing Street (the prime minister’s resi­

dence), the Ministry of Defense, the Foreign Office, the U.S. Embassy,
and the headquarters of the British intelligence and security services (for
example, were the lights on at night?);
• the frequency with which couriers traveled among those establishments;
• the comings and goings of the prime minister and other key ministers
(for example, was the prime minister making an unusual number of trips
to Buckingham Palace?); and
• the existence of any unusual civil defense measures, such as stockpiling
food or preparing emergency blood banks.

Routine reports were to be made every two weeks, but any particularly
striking events were to be reported immediately.47
This type of formalized system, in which indictors are determined in
advance and can thus guide the collection process to some extent, exists only
with respect to the most important questions, such as the possibility of
hostile military action. The formal system also ensures that intelligence is
 the policy-making process and is not ignored. In most other areas,
fed int0 SyStem is set up to guarantee that the available intelligence is
n° 3yht to bear when decisions are made.
broU*jl wjnp what was widely regarded as the CIA’s failure to provide
warning of the Iranian revolution of 1978-79, proposals were
3 d to establish a political I&W system to warn of political instability in
ma ' n countries. This would require establishing indicators comparable to
hose used to warn of military attack. This is a much more difficult propo­
sition with politics than with military matters; it is unlikely that the current
state of the social sciences allows this to be done sufficiently to be of real
use Nevertheless, a more formal system would clarify the extent to which
intelligence is responsible for warning of potentially serious events, even
when the policy-making community is not focusing on that country or region
of the world.

Basic Intelligence
Another general type of intelligence product may be termed the “ basic
intelligence report.” Such a report tries to provide as full a picture of a given
situation as possible, drawing on publicly available data and relevant infor­
mation from all intelligence sources ( “ all-source” intelligence). In the mil­
itary arena, for example, such products might be an order of battle, the
description of a nation’s armed forces in terms of numbers and types of
combat units, and their manpower, equipment, organization, and subordi­
nation of units to higher-level commands. Similarly, a basic intelligence
report on a nation's political system could include an account of all the
major political forces and personalities, their traditional views and interests,
the ways in which they have related to each other, and so forth.

Periodic Reports

Just as l&W represents a systematized form of current intelligence to deal

W]th a specific question of great importance, there may be special series of
reP°rts to deal with specific issues. For example, since the SALT 1 agree-
ments of 1972, U.S. intelligence has produced semiannual reports on Soviet
best^lianc.6 strategic nuclear arms agreements; these reports present the
v lnteUigence judgments concerning Soviet strategic nuclear forces rele-
n to the agreements’ provisions.

Intelligence Estimates
The most ambitious type of intelligence product is that which not only
describes! he current situation but also attempts to predict how it will evolve.
particular, a national intelligence estimate (NIE) represents the most au­
thoritative statement on a subject by the U.S. intelligence agencies, all of
which contribute to its drafting and review.
These estimates are supposed to take the broadest view of their subject
and to project the current situation into the future. In the United States, they
are produced by a special staff, with the support of analysts throughout the
intelligence agencies. The estimates are supposed to incorporate the views
of all the agencies in the intelligence community.43 A good deal of effort is
devoted to finding a consensus position, but, when this is not possible, a
dissenting view may be expressed in what is traditionally called a footnote,
even though the dissent is sometimes included in the text.
NIEs on some topics of major importance, such as Soviet strategic nuclear
offensive and defensive forces, are produced periodically; other topics are
covered in response to specific requirements, either self-generated (that is,
the idea for the estimate arose within the intelligence community itself) or
from elsewhere in the government, most often the National Security Council
(NSC). Shorter, more topical estimates, known as Special NIEs (or SNIEs),
may be produced in response to more urgent requirements.
In contrast to the U.S. practice, producing estimates in Israel during the
period before the 1973 Yom Kippur War was a monopoly of Military
Intelligence, a branch of the armed forces. Although other intelligence
services existed— the Central Institute for Intelligence and Security (Hamos-
sad), operating covertly in foreign countries; the Shin Beth, responsible for
counterintelligence and counterterrorism; and the Foreign Office research
department— only Military Intelligence had access to all intelligence infor­
mation and was responsible for analyzing and distributing it. Furthermore,
the cabinet’s staff did not have any intelligence expertise, and there was no
mechanism by which the political leadership tasked Military Intelligence to
provide evaluations of specific subjects.44 This concentration of power was
criticized as contributing to the failure of Israeli intelligence to warn of the
Egyptian-Syrian attack on October 6 , 1973, which is discussed in the “ in­
telligence failure” section of this chapter.

Variations Among Intelligence Services

While current and basic intelligence functions seem essential to the work of
any nation’s intelligence services, how integral is the estimative function to
the intelligence process? Its importance varies from nation to nation and
time to time. For example, in the United States, preparing estimates is seen
as the peak of the intelligence process. In the Soviet Union of the 1930s, by
 jsjj^vD ( a predecessor of the KGB) not to send him estimates or
c o n tra s ,
ordered theAof the foreign situation, but to confine itself to reporting the
assessmen of high-level informant reports) it had obtained.45 This
secrCtt to do with the relationship between the intelligence and policy-
point ’a^unctions 0f a government, discussed further in chapter 6 .

Intelligence Failure and Surprise

Types of Failure
Surprise Attack
The possible failure of intelligence to assess a situation correctly is a danger
coeval with intelligence itself. The most dramatic, and potentially most dam­
agin'’, intelligence failure occurs when there is no warning of an attack, so
that a nation’s military forces are taken by surprise. Indeed, the U.S. intel­
ligence community owes its current configuration to an effort to remedy the
intellieence deficiencies thought to have contributed to the United States be-
ins surprised by the Japanese attack on Pearl Harbor on December 7, 1941.
The literature on successful surprise attacks is extensive. The German
attack on the Soviet Union on June 22, 1941, and the Egyptian-Syrian attack
on Israel on October 6 , 1973 (the Yom Kippur War) are, aside from the
Pearl Harbor attack, among the most frequently discussed cases, but others
have happened in this century.46 Even during a war, when military forces
are presumably much more on the alert than during peacetime, substantial
surprise can be, and has been, achieved. Among other examples are the
German attack in the Ardennes in December 1944 (Battle of the Bulge) and
the Chinese entry into the Korean War in November-December 1950.47
Closely related to the failure to anticipate an attack is the situation in
which a nation expects an attack but, because of a significant misestimation
of where or how it will occur, responds disadvantageously. An impressive
example of this kind of intelligence failure is the German reaction to the
Allied landings in Normandy on June 6 , 1944 (D-Day); the Germans cer-
tamly expected an attack, but they were so convinced that the main thrust
would be against the Pas de Calais region that they treated Normandy as a
■version and passed up a good opportunity to counterattack in force while
1 e Wellheads were most vulnerable.48

O tha Kinds o f Surprise

2111 unanticipated attack may be the most damaging surprise a

Can suffer, it is not the only one. For example, an unexpected political
event (such as a shift in alliances or a coup d’état that overthrows a friendly
foreign ruler) also may be a serious blow to a nation’s foreign policy inter
ests. The American underestimation, in 1978, of the Iranian shah’s political
troubles and of the seriousness of the opposition to his rule is perhaps the
best-known recent example of this type of intelligence failure. It was con-
sidered sufficiently damaging at the time to prompt a blunt handwritten note
from President Carter to Cyrus Vance, Zbigniew Brzezinski, and Stansfield
Turner (secretary of state, assistant to the president for national security
affairs, and director of central intelligence, respectively) that stated, with a
directness uncharacteristic of government bureaucracy: “ I am not satisfied
with the quality of our political intelligence . ” 49
Similarly, no warning of a sudden, major economic change, such as the
failure of U.S. intelligence agencies to foresee the 1973 “ oil shock”—the
rapid increase in the world price of oil— also may be accounted an intelli­
gence failure, depending on one’s opinion of the field of view and respon­
sibilities of intelligence agencies. As has been noted, intelligence’s role with
respect to economic questions is less settled and more complicated than with
respect to political or military matters. In any case, the 1973 oil shock is a
complicated example of the problem, since the underlying causes were
economic (the previous low price of oil had encouraged rapidly increasing
oil consumption and discouraged exploration and development) while the
immediate cause was political (the Arab oil states’ support for Egypt and
Syria in the October 1973 Yom Kippur War) .50
While the disasters that can flow from being surprised on the battlefield
are obvious, the harm caused by the failure to foresee political or economic
events is more difficult to assess. In large part, it depends on whether the
government would have been prepared to take action had it been warned and
whether there were strategies that would have averted the event or mitigated
its unfavorable consequences.
For example, it is not clear what, if anything, the U.S. government would
have done had it been warned of the 1973 oil shock. Similarly, correctly
assessing the shah’s political troubles is a vastly different task from deter­
mining what, if anything, to do about them; success in the former would not
have necessarily guaranteed success in the latter.

Other Kinds o f Failure

Most cases of intelligence failure sooner or later involve surprise, since a
mistaken view of the external world is likely to cause unexpected misfortune
at some point and in some manner. (One exception would be a case in
 through faulty intelligence, an agency is unaware of an adversary’s
vvhichi ty s0 that it loses the opportunity to take advantage of a favorable
^ j however, in many cases the notion of surprise is perhaps not the
situatioHpfui ^ j-or understanding the harm intelligence failure causes.
* * Rey intelligence failure is a misunderstanding of the situation, which
an agent to take actions that are inappropriate and counterproductive
his own^interests; whether he is subjectively surprised by what happens is
t0 important than the fact that he is doing or continues to do the wrong
th^n^ Thus, the German intelligence failure with respect to the Normandy
D Day landings resulted not so much in the German Army’s being surprised
' in its misestimation of the Normandy landings as feints meant to distract
attention from the “ real” invasion force targeted on the Pas de Calais
region.51 Similarly, the primary cause of the Chinese rout of the U.N. forces
in Korea in December 1950 had more to do with a fundamental misunder­
standing of Chinese doctrine and tactics than with surprise.52
Furthermore, the examples of surprise, involving failures to foresee what
action a hostile force would take or with what success it would meet, apply
particularly to nations that are on the defensive, strategically or tactically. A
nation on the offensive, however, also can suffer through misassessing the
opponent’s strengths or misunderstanding some other relevant factor. For
example, the Allied attack on the bridge at Arnhem, Holland, the “ bridge
too far,” in September 1944 was unsuccessful in large part because of an
intelligence failure— the failure, due to overconfidence and a negligent anal­
ysis of the available information, to realize that two German armored divi­
sions were present in the immediate area.53
Finally, the intelligence failures discussed above involve a misunder­
standing that is revealed quickly once one side or the other takes action.
Another sort of failure involves the misestimation of a continuing process or
condition in which, because nothing occurs to reveal the truth of the situ­
ation, it is possible to remain in error for a longer time.
^ F°r example, in estimating the size of the Soviet strategic arsenal, the
’ ’ intelligence community has made both major overestimations and
erestimations that colored American strategic views over periods of
ars. The major overestimations of Soviet strategic weaponry, which pro-
" n i .^ mer*can Perceptions of a “ bomber gap” in the mid-1950s and a
kno: ile *rom the late 1950s until 1961, are perhaps the better
t^ro nh° f ^ eSe errors- But at *east equally important was the tendency,
estimat°Ut ^ anc* ending only in the early 1970s, to under-
f0 rce^ 5 4 extent to which the Soviet Union was building up its strategic
Aside from instances in which relevant information cannot be obtained at
all, intelligence failure refers to a disorder of the analytic process, which
causes data to be ignored or misinterpreted. Therefore, it is similar to a
mistake in any other intellectual endeavor, for example, a mistake by a
historian in interpreting an ancient text, leading to an incorrect description
of some aspect of antiquity, or a mistake by a meteorologist in assessing the
importance of a low-pressure system, leading to an incorrect prediction
about the next day’s weather.
In addition, however, some peculiarities of the intelligence analysis proc­
ess introduce further sources of error. While intelligence analysis is an
intellectual activity, it is one that, unlike the work of an individual aca­
demic, takes place in an institutional setting and according to standard
procedures so that the final result is more the product of a system than of any
individual. As such, it is vulnerable to certain pathologies that can be
addressed in institutional or bureaucratic terms. This section addresses some
of the causes of intelligence failure, looking first at institutional ones and
then at those that relate more directly to the intellectual content of the
intelligence work.

Subordination o f Intelligence to Policy

The possibility that intelligence judgments will be made to produce the

result superiors wish to hear instead of on the evidence is perhaps the most
commonly discussed source of error or bias in intelligence analysis. It is,
however, extrinsic to the intelligence analysis process itself (assuming that
the analyst realizes he is going against his own best judgment). As rather an
issue in what may be called the management of intelligence— the way in
which intelligence relates to the higher-level policymakers whom it serves—
it is dealt with in the discussion of intelligence and policy in chapter 6.

Unavailability o f Information When and Where N eeded

Given the large size of the organizations involved in collecting and analyz­
ing intelligence, a possible source of problems is the unavailability of in­
formation that is in the system to those who need it when they need it. This
unavailability has various causes: for example, security regulations that
restrict the circulation of sensitive information, bureaucratic jealousies a n d
power struggles in which control over information becomes one of the
weapons, or a simple unawareness in the office possessing the data o
another office’s information needs.
 Mil* ~do analysis in a given area and access to an mrormation relevant to
bility t0 different offices may work on different parts of the problem,
it. *n t!llS
k e V C.
p i;e ^ nfr linformation
ce O n i U I I i u u i u i i may
* i * o y be
u * missed
1 1 1 1 . ^ ^ if
xx its
x to significance becomes

bUt 3 nt only in the context of all the available data

apparen the Pearl Harbor intelligence failure is eexplained in terms of a
Typically
of communication between the various intelligence and operational
lack • volved. However, while there may have been some communication
UnIhl1 is between the Army and Navy commands in Hawaii, it is not clear
Pr0 ' corresponding difficulty existed between the two services in Wash-
*^atf \ A more important difficulty stemmed from the lack of a centralized
analytic office with the resources and the responsibility to analyze all rele­
vant intelligence. Instead, intelligence seemed to be handled on the fly by a
few people who worked directly for the high-level policy-making officials—
an effect of the small size of the U.S. military establishment at that time as
much as anything else.

Received Opinion
The success of any intellectual endeavor (scientific discovery, for example,
as well as intelligence analysis) can be compromised by the force of re­
ceived opinion, often referred to as “ conventional wisdom”— those opin­
ions about a subject that are generally regarded, without sufficient
investigation, as true. While, in one sense, we obviously cannot do without
it (we could hardly afford the time or effort to reinvestigate and rethink all
our opinions every day), relying on received opinion poses the real danger
that we will either misunderstand or perhaps ignore evidence that suggests
the truth is otherwise.
However pernicious the effect of received opinion on an individual’s
thinking, its impact is probably heightened by the bureaucratic environment
m which an intelligence service carries out its analysis. While an individ­
ual s thought can be influenced by particularly striking evidence that does
n°t fit the preconceived pattern, or even by a chance insight into another
mterpretation of the evidence, it is unlikely that many people will be affected
ln same way at the same time. Thus, individuals challenging the received
Nlew are likely to encounter the resistance of the much larger number who
remain comfortable with it.
individuals economize on time and effort by not rethinking all their
°Pinions all the time, the organization may be said to do the same. Given the
^2e anci complexity of the organizations involved and the consequent need
0 °btain the consent (or at least the acquiescence) of many people to the
judgments reached, fundamental réévaluation of the evidence with a view
toward rethinking all the basic assumptions is a very difficult and time-
consuming process.
Thus, if an analyst has to write a report, he is likely first to review what
was written the previous year and then merely update it as necessary. This
not only saves labor but also offers a degree of security. Because last year’s
report was approved, if the analyst limits himself to the changes indicated by
new evidence, the underlying reasoning does not have to be duplicated, and
the sources of potential controversy can be minimized.
Clearly this creates a tremendous opportunity for error in situations in­
volving gradual change. Because intelligence evidence is almost always
spotty and incomplete, the accumulation of various bits and pieces of new
information may not force a change in the underlying analysis. This may be
true even though the preponderance of evidence at some point would suggest
to someone starting afresh that a different framework would make more
sense. In this manner, an older way of viewing the situation may survive
much longer than it should.55

M irror-Imaging

The two problems discussed above apply primarily to intelligence analysis

conducted in a bureaucratic setting. Another source of error in intelligence
analysis, however, reflects a more common intellectual failure not tied to the
setting in which the analysis is conducted. It may be termed “ mirror­
imaging,” by which is meant the judging of unfamiliar situations on the
basis of familiar ones.56 In the case of intelligence, it typically means
assessing or predicting a foreign government’s actions by analogy with the
actions that the analyst feels he (or his government) would take were he (or
it) in a similar position.
For example, the failure of Israeli intelligence to foresee the Egyptian-
Syrian attack that started the October 1973 Yom Kippur War appears to be
due in part to this phenomenon; Israel’s intelligence services did not imagine
the Arabs beginning a war that they seemed sure to lose.57 For example, less
than two months before the outbreak of war, Moshe Dayan told the Staff
College of the Israel Defense Forces that “ the balance of forces is so much
in our favor that it neutralizes the Arab considerations and motives for the
immediate renewal of hostilities.” 58
In retrospect, however, it appears that Egyptian President Anwar Sadat
wanted the war to break the stalemate that had developed in Israeli-Arab
relations, and he did not, in fact, count on winning it. The initial Arab
victories (in particular, the Egyptian crossing of the Suez Canal and pene-
 of the Israeli defensive line on the other side), although reversed on
» a U -u n d later in the war, provided the psychological backdrop for the
the gf al regaining of the Sinai Peninsula. Discounting the possibility of this
e - f aggressive risk-taking, foreign to Israeli and U.S. military experi-
ce is an example of mirror-imaging.
^Another striking example may be found in the report of the Church
mittee, the special Senate committee formed in 1975 to investigate the
• °t Hi степсе community. That report cites the following paragraph from a
draft of the 1969 National Intelligence Estimate on Soviet offensive nuclear
forces:
We believe that the Soviets recognize the enormous difficulties of any
attempt to achieve strategic superiority of such order as to significantly
alter the strategic balance. Consequently, we consider it highly unlikely
that they will attempt within the period of this estimate to achieve a
first-strike capability, i.e., a capability to launch a surprise attack against
the U.S. with assurance that the USSR would not itself receive damage it
would regard as unacceptable. For one thing, the Soviets would almost
certainly conclude that the cost of such an undertaking along with all their
other military commitments would be prohibitive. More important, they
almost certainly would consider it impossible to develop and deploy the
combination of offensive and defensive forces necessary to counter suc­
cessfully the various elements of U.S. strategic attack forces. Finally,
even if such a project were economically and technically feasible the
Soviets almost certainly would calculate that the U.S. would detect and
match or overmatch their efforts.59
Despite the context in which the Church Committee cites this passage—
the Nixon administration in general, and Defense Secretary Melvin Laird in
particular, are being taken to task for successfully pressuring DCI Richard
Helms to delete this paragraph from the final version of the NIE— its report
does not indicate that the judgments in the passage were based on inside
information about strategic thinking in the Kremlin. Indeed, the wording
0r example, “ would almost certainly” ) suggests instead that the para-
£raph contains analytic judgments about Soviet thinking rather than direct
eVl^ CnCe ^rawn fr°m documents or high-level informants.
s such d is an example of mirror-imaging not only in the three major
but ri10nS ^ast ^ ree sentences cited) it offers to support its main thesis,
j r S° *n entife way in which the question is framed. The first assertion
sider^8 ^ ^ intelligence understood what costs the Soviets might con-
ima ' ^r° ^ k ^ ve’ a judgment that rests essentially on a sort of mirror-
® United States had concluded that this goal was not econom-
orists of the U.S. strategic community but not necessarily believed by Soviet
military thinkers. Finally, the third attributes to the Soviets a mechanistic
apolitical view of the arms race (as an action-reaction cycle) that is common
in the United States but not necessarily in the Soviet Union.
Perhaps even more important, the entire manner in which the question is
framed in the first two sentences— that is, the equating of “ strategic supe-
riority [that] significantly alterfs] the strategic balance” on the one hand
with “ first-strike capability” on the other— reflects U.S. mutual assured
destruction (MAD) doctrine rather than any attempt to understand what the
Soviets might consider significant. It implies that the Soviets would see
partial solutions to the problem of countering U.S. offensive strategic forces
as insignificant and not worth developing and deploying.
But since this view is merely implied rather than stated, it is not examined
on the basis of the evidence. (For example, a careful evaluation of this view
would have to deal with the emphasis the Soviets place on strategic air
defenses and the Moscow antiballistic missile system, even in the absence of
a nationwide ballistic missile defense.) Instead, its attribution to the Soviets
seems to be a case of mirror-imaging by analysts who believe that the logic
behind the U.S. view is so strong that the Soviets must have accepted it.
Another common form of mirror-imaging, one that affects the ability to
assess correctly scientific or technological developments within an adver­
sary’s military establishment, is often referred to as the “ not invented here
syndrome.” This syndrome is characterized by an unwillingness to take
seriously the possibility that an adversary has developed a weapon or device
that one’s own military-technological establishment has not thought of, has
dismissed as infeasible, or has given up trying to develop. We would expect
this syndrome to be more common with respect to an adversary who is
regarded as technologically inferior.
R. V. Jones discusses this phenomenon in his account of British scientific
and technical intelligence during World War II. In an epilogue to an intel­
ligence report concerning the V-2 ballistic rocket, which the Germans de­
veloped toward the end of the war and first fired against London on
September 8, 1944, Jones goes through the four possibilities that can exist
with respect to a technical development: the British either do or do not
succeed with respect to it and the same for the Germans. The not invented
here syndrome obtains when

our experts either fail or do not try, [and] the Germans succeed. This is
the most interesting Intelligence case, but it is difficult to overcome the
prejudice that as we have not done something, it is impossible or foolish.
 1 ^ b u t I lO V lC C b , CUIU l l i a j ' U ll> ltsiV /l ^ m u x v v * » u u v /» UAW»

*°n^rpence [i.e., intelligence analysts who are not as such experts in the
e t technology], which at least has the advantage of closer contact
releva 60
with the enemy.
s *s solution to the problem is for the intelligence analyst to be wary
f the technical expert's views while trying to make use of his expertise:
PI m an Intelligence point of view, it must always be borne in mind that
the advice comes from a British, and not from a German expert. If this
difference in background is not continually appreciated, serious misad-
iustments can be made. In the tactical field, Napoleon knew this danger
well: he called it, “ making pictures of the enemy“ . In the technical field
the same danger exists. . . .61

“Solutions” to the Problem o f Intelligence

Failure

What Is Failure?
More frequently than the popular myth of an omniscient intelligence service
would lead us to expect, intelligence reports or estimates contain erroneous
statements, and important events occur without intelligence agencies having
predicted them. Determining whether such situations constitute intelligence
failures requires a standard against which it is reasonable to measure intel­
ligence achievements.62 Often, it is assumed in discussions of this sort that
intelligence performance should be measured against an ideal of clairvoy­
ance. Typically forgotten is that, for the most part, intelligence involves not
a metaphorical struggle with nature (as is the case with scientific research
that seeks to force nature to reveal its secrets) but a real struggle with a
human adversary.
For example, the victim’s failure to anticipate a surprise attack is the
reverse side of the coin of the attacker’s success at achieving surprise. In
mteliigence work, one opponent is most often trying to frustrate his adver-
Sary s attempts to understand the situation accurately. To do so, he will use
a whole array of intelligence techniques, including deception, as is dis-
Cussed in chapter 5 on counterintelligence.
Thus, we cannot compare progress in intelligence techniques to, for ex-
amPle, progress in chemistry. The consequence of progress in chemical
^ tarc*T and the dissemination of the knowledge and insight gained thereby,
an be that chemists throughout the world achieve better results in their
 4 _____ — v/K\/ ui intelligence oinciais to achieve better
results in its work, this means, in general, that its adversaries’ intelligence
officials are not performing their own tasks as well as in the past and, in a
sense, have fallen down on the job. This is obviously true with respect to
questions of deliberate surprise.63
Taken to an extreme, speaking of intelligence failure is similar to speak­
ing of “ chess failure,’’ defined as the failure to win chess games. Obvi­
ously, to improve our chess-playing abilities, it makes sense to critique
styles of play, as well as individual moves, as thoroughly as we can. The
result should be better individual chess play and, if we share the insights we
have gained, by others as well. It cannot be, overall, an increase in the
number of games of chess won per number played. Of course, as citizens
we are not concerned with better intelligence in the abstract; we seek to
improve our own country’s intelligence capabilities and, in so doing, de­
value those of our adversaries.

Institutional Solutions

Among the most frequently discussed institutional solutions— solutions that

require the restructuring of the institutional framework in which intelligence
analysis is carried out— are competitive analysis and the establishment of a
devil’s advocate.
Competitive analysis refers to the deliberate fostering of separate analytic
centers within a government, each of which has the right to formulate and
distribute its own intelligence assessments. In principle, each center would
have equal and comprehensive access to the raw intelligence data, regardless
of which intelligence agency collected it. (In practice, because of the ex­
treme sensitivity of certain types of intelligence information— for example,
information that might identify a human source— this result may well be
difficult to achieve.)
Typically, some or all of these centers are constituent parts of the major
departments and agencies of government or the military services and exist
primarily to serve the specialized intelligence needs of their parent organi­
zations. For example, in the United States, the Defense Intelligence Agency
is a part of the Department of Defense, reporting to both the Joint Chiefs ot
Staff and the secretary of defense; each military service has its own intel­
ligence command or service; and the State Department has its Bureau of
Intelligence and Research. To some extent, the resulting competition among
them reflects the fact that they are subordinated to different government
departments, which arc themselves in competition.
 jc^iy to compete, most likely on an ad hoc basis to examine a specific
eXP American example for which information is publicly available was
*SSUA B Team exercise of 1975-76, in which a separate team of experts (the
^T eam ) was established to review evidence concerning Soviet strategic
® lopments and determine if it supported conclusions other than those of
h writers of the previous years’ National Intelligence Estimate (the A
Team) . .. , . .
In either case, the virtue or competitive analysis resides primarily in
allowing differing points of view to be expressed at high levels, thereby
sharpening the debate and focusing attention on whether available evidence
unambiguously supports one or the other position. In this way, competitive
analysis can act as an antidote to the problems caused by the easy acceptance
of conventional wisdom. With competing analytic centers, it is more likely
(although not guaranteed) that at least one of them will be skeptical of the
received opinions on a subject and will be able to force the other agencies
to mount an explicit defense of them. It is an attempt to imitate, within the
limiting confines of the government and its regulations concerning security
classification of information, the free marketplace of ideas that exists in a
democratic society and that presumably furthers the advance of knowledge
generally.
From this perspective, the possibility that the competing centers may take
positions that are biased in favor of their parent agency’s policy preferences
or budgetary requests is not such an important drawback. Vigorous debate
among the multiple centers will, according to this line of argument, expose
the invalidity of positions the evidence doesn’t support, while providing a
greater chance that new, unconventional ideas will receive a serious hearing.
Of course, it could happen that the conventional wisdom pervades the
different parts of the government to an equal degree. In that case, we could
not expect the mere existence of competing analytic centers to challenge the
conventional wisdom since they would all be infected by the same ortho-
d°xy. This consideration has led to proposals to establish a devil’s advocate,
that is, an analytic organization whose purpose is to challenge accepted
views.
The B Team is an example of an ad hoc devil’s advocate established to
* at one particular problem and motivated by an already-existing sense
at the previous understanding of it had been lacking. The devil’s advocate
Proposal is an attempt to institutionalize this type of practice and, in so
g Protect it from some of the political controversy that surrounded the
cam, which some critics regarded as illegitimate outside (i.e., political)
Ure on the intelligence analysis process.
 ----------------------- -------------------------------------------------------------- --------------- ---------------- ------------------------------ © ~v. U U 1U IV Jf

them, neither can be regarded as a panacea, or even as necessarily a step in

the right direction. In an article provocatively subtitled “ Why Intelligence
Failures Are Inevitable,” Richard K. Betts has argued that neither institu­
tional solution discussed above, nor any other institutional solution for that
matter, can prevent intelligence failure.6"5
Competitive analysis, for example, may air dissents to the conventional
wisdom that would otherwise be suppressed; at the same time, however, it
presents the policymakers with a spectrum of views, which may leave them
confused or may lead them to ignore intelligence altogether and follow their
own prior opinions instead. Similarly, the dissents of an institutionalized
devil’s advocate may lose any real power to force others to rethink their
positions, precisely because, in dissenting, “ he is just doing his job.”
Indeed, the perverse result may be that otherwise powerful challenges to the
orthodoxy are ignored because they are seen as routine and predictable.
I am not saying, however, that institutional arrangements do not matter or
that changing them may not be beneficial. Rather, I am suggesting that the
optimal bureaucratic organization of intelligence cannot be determined until
we have a sense of what the characteristic problems or deficiencies of a
nation’s national security apparatus are; the organizational structure should
be determined by what would best counteract those bad tendencies. The
same solution— for example, competitive analysis— may be helpful or harm­
ful, depending on whether the system suffers from too much conformity
with the conventional wisdom (in which case competition might be able to
introduce needed new ideas) or from domestic political concerns dominating
national security questions (in which case a strong, unified intelligence
voice may be necessary to bring international realities to the fore).

Intellectual Solutions
Institutional solutions have the advantage that they can be implemented
top-down through managerial decisions and directives; they do not, how­
ever, attack the heart of the problem of intelligence failure, the thought
process of the individual analyst. Improving thought processes, however, is
not a matter that administrative means can deal with directly. Indeed, when
we try to be specific about what would be involved, it is not even clear we
know what “ the improvement of thought processes” means, aside from a
general and not very useful admonition to be “ smarter.”
While it may not be possible to lay down rules that will inevitably guide
us to correctly analyze intelligence information, it is nevertheless useful to
try to identify intellectual errors or deficiencies that may be characteristic of
lilts
place,~ in- so doing, we can become aware u i w c a c I t l l U U l W I V O «**V*
quently try to correct them.
The previous discussion of the intellectual causes of intelligence failure
lists mirror-imaging as an important error that must be guarded against.
While any intelligence analyst is vulnerable to this error, cultural reasons
may make it a particular problem for the U.S. intelligence community.
Americans are more open to a belief in the basic similarity of people through­
out the world, perhaps because of America’s experience in successfully
absorbing and assimilating immigrants from diverse cultural and religious
backgrounds. Thus, the U.S. intelligence analysts risk being more likely
than other analysts to understand and predict the actions of others on the
basis of what they would do under similar circumstances.
Thus, an emphasis on expertise in foreign societies and cultures is an
important corrective to this error.66 This expertise can be fostered by a study
of the language and history of the country, by an awareness of its religious
and cultural traditions, and so forth. A deliberate attempt must be made to
see the international situation from its leaders’ point of view, rather than our
own.
 WORKING BEHIND THE
SCENES
COVERT ACTION

Covert action— especially in its more sensational forms, such as the over­
throw of a government or the assassination of its leader— looms large in the
public’s view of what intelligence is; for professionals and students of
intelligence, however, it is something of a question whether covert action
should be considered a part of intelligence at all. This issue arises because
covert action involves taking action to implement a nation’s foreign policy,
while collection and analysis are limited to providing the information on
which that policy may be based.

What Is Covert Action? Some Definitions

Covert action, in the U.S. intelligence lexicon, refers to the attempt by one
government to pursue its foreign policy objectives by conducting some
secret activity to influence the behavior of a foreign government or political,
military, economic, or societal events and circumstances in a foreign coun-
lJ*y- As the term implies, the defining characteristic of covert action is that
e 2overnment conducting the activity do so in a secret or covert manner,
owever, what secrecy means precisely can vary according to the particular
Clrcumstances.
wiij1 some cases, the need for secrecy may be absolute, and a government
trY to act so that the details and even the existence of the activity do not
actio- o r a l l y known. This would be true, for instance, if the covert
fore' » lnvo^vec* hoping a group of conspirators prepare a coup d’état in a
lgn country by smuggling weapons into the country for it. In other cases,

73
involvement in it. An example would be a government’s secret financing of
a newspaper or radio station that supported government policies or was
hostile to its adversary’s. Finally, there may be cases in which a good deal
of information becomes public, but for diplomatic or other reasons, the
government involved still avoids officially acknowledging its connection
with it.
Covert action (euphemistically referred to as “ special activities’’ for
bureaucratic purposes) is currently defined by the U.S. government in the
most recent presidential Executive Order on U.S. Intelligence Activities
as follows:

Special activities means activities conducted in support of national foreign

policy objectives abroad which are planned and executed so that the role
of the United States Government is not apparent or acknowledged pub­
licly, and functions in support of such activities, but which are not in­
tended to influence United States political processes, public opinion,
policies, or media and do not include diplomatic activities or the collec­
tion and production of intelligence or related support functions.1

As part of the executive order on intelligence, this definition is valid

within the executive branch; for example, it is the definition that would be
used in the intra-executive branch process used to evaluate and approve
covert action proposals. The term “ covert action,” however, is also com­
monly used to refer to a similar but not identical set of intelligence activities
about which the president is required by statute to provide “ timely” noti­
fication to the Congress. This set of activities was defined in 1974 by
legislation known as the Hughes-Ryan Amendment as “ operations [involv­
ing expenditure of funds by or on behalf of the Central Intelligence Agency]
in foreign countries, other than activities intended solely for obtaining nec­
essary intelligence.” 2
By either definition, covert action is clearly distinguished from the col­
lection and production of intelligence information, which is to say, from
what in the United States is typically regarded as the heart of the intelligence
function. The distinction is that the goal of covert action is not knowledge
but is instead the direct furthering of national foreign policy objectives. In
this respect, it more closely resembles the nation’s other foreign policy
tools, such as diplomacy or military force, than the rest of intelligence. On
the other hand, insofar as it is activity undertaken covertly, involving the
secret employment or support of agents or allies, it resembles human intel­
ligence collection more than it does the overt policy means a government
might use.
 not absolutely precise. For example, although the executive order ex-
afe tlv excludes diplomatic activities from the ambit of covert action, it
P .. ' defines diplomatic activities nor specifies the dividing line between
ne , maCy and covert action. Diplomacy also involves attempting to influ-
P the ^ behavior of other governments and
gflCv § # , their officials and is often
arried on secretly; for example, it is often important to keep negotiations
between two countries secret from other countries and the public. It may
even be that conversations with some officials of a foreign government are
not shared with other officials of that same government.
Historically, diplomacy and intelligence have not always been distinct
categories. In Renaissance Italy diplomatic envoys, aside from their
i n f o r m a t i o n - g a t h e r i n g function, which might involve clandestine methods of

doubtful legitimacy,
often became involved in intrigues far from ambiguous. The Venetian
secretary and the Milanese commisarias at Genoa in 1496, for example,
served as intermediaries between their governments and an adventurer
who offered to burn up two of [sic] three French ships either in the port
of Villefranche or at Nice. The two governments agreed to the proposal
and to splitting the cost of 400 ducats.3
Even today, there are cases that illustrate the artificial nature of the line
between diplomacy and covert action. For example, the unpublicized at­
tempts in 1984-86 by State Department and National Security Council
personnel to persuade foreign leaders to contribute money to support the
Nicaraguan resistance come within the usual scope of diplomatic
exchanges.4 Had they been conducted by intelligence agency personnel,
however, they could just as easily have been regarded as covert action.
Another example, less well known, arose from the U.S. response to the
kidnapping of the Italian statesman Aldo Moro by the Red Brigades terrorist
group in 1978. The Italian government requested the services of a CIA
psychologist who had specialized experience in the counterterrorism area.
This was considered to be covert action under the Hughes-Ryan amendment
definition; delays in preparing the required paperwork led the Italians to turn
to State Department, which, unimpeded by covert action approval pro­
cedures, provided a psychologist from its staff. For practical purposes, the
lne between diplomacy and covert action appears to be drawn in ambiguous
CciSes according to the departmental affiliation of the personnel involved
rat er than the activity’s nature.
ican S^°U^ n°ted that the term “ covert action” appears to be an Amer-
lnvention, not necessarily used in the lexicons of other intelligence
 uvt 11 anu covert lecnniques ot influencing events and behavior in, and
the actions of, foreign countries.” 6 As such, it does not fall entirely within
the sphere of intelligence. Instead, it includes other foreign political activ­
ities of both the Soviet government (for example, diplomacy and the use of
official media such as Radio Moscow) and the Communist party of the
Soviet Union (conducted, for example, through foreign Communist parties
or front groups such as the World Peace Council). The term focuses atten­
tion on the goal sought— political influence— rather than on the secrecy or
openness of the means used.7
Regardless of the terminology used, it is important to keep in mind that
what is called covert action in the United States is only one tool of foreign
policy among many. Because of its unique features and flamboyant past, it
has attracted a great deal of attention, and it does pose some particular
problems in terms of its secrecy for a democratic government. Nevertheless,
the term may be unfortunate in that it emphasizes a characteristic of the
means used (their secrecy) while obscuring the fact that the ends sought arc,
or should be, those of foreign policy as a whole. Because of this focus on
the form rather than the substance of the activities involved, the debate in the
United States about covert action has emphasized questions of legality and
propriety at the expense of more fundamental questions concerning foreign
policy goals and strategy.
This chapter will treat covert action as an element of intelligence that is
also a foreign policy tool; in chapter 6, the specific issues connected with the
use of covert action by a democratic government will be addressed. Because
of the ambiguities created by the existence of two nonidentical, official U.S.
definitions of covert action, the discussion in the remainder of this chapter
deals with activities more inclusive than the covert action category in order
to clarify its limits. A more theoretical definition of covert action would not,
to be sure, be tied so closely to the vagaries of U.S. law and practice; such
a definition, however, remains to be achieved.

Types and Examples of Covert Action

In its most general sense of the secret influencing of foreign behavior,
events, or circumstances, the term “ covert action” covers a wide spec­
trum of activities, running from the most pedestrian, such as secretly pro­
viding technical assistance (such as security or communications equipment
or training) to a friendly foreign government, to the most spectacular,
such as assassination or supporting or fomenting a coup d’état. It is there-
 u ^ v n v iv o l i t * ▼v v v /lJ L U U V lV U < XXI ClUVJlllUll, II1C U .O . liu e lll-
e agencies have not developed a standard typology of covert action,
^eIlChove those who study them. This section attempts to develop such a
nor nav
rategori^^i*^ •
The purpose of covert action is to influence the actions of foreign gov-
ents or events or circumstances in foreign countries. These attempts
C be directed at the government of a country, at the society as a whole, or
at*a particular segment of it. In democratic countries, influence successfully
b r o u g h t to bear on the society may be felt quickly by the government as

well* but such influence also can have important effects in the long run in
nondemocratic or even totalitarian countries.
Influencing political behavior, circumstances, or events in foreign coun­
tries is the very stuff of foreign policy; diplomacy is as directed toward this
iioal as is covert action. The development and deployment of military forces
also may be carried out to influence foreign behavior; this is common in
peacetime but may occur during war as well. Thus, it is not surprising that
many of the covert action techniques discussed below can be, and are,
carried out in a noncovert manner as well. What is or is not considered
covert action, for U.S. government purposes, often depends less on the
nature of the case than on the peculiarities of the applicable legal and
administrative definitions of covert action.

Covert Support of a Friendly Government

In many cases, covert action is carried out through a tacit alliance with
individuals or groups with whom one shares common objectives. It is least
difficult to do this when the ally in question is a friendly foreign government.
In this instance, covert action can be limited to such unsensational activities
as secretly assisting it with personal security for its leader or equipment for
secure (encrypted) communications.
There is some reason to call such an activity covert action. The support
provided to a friendly government helps it stay in power and, in that sense,
influence the course of events in that country. In addition the action takes
ace in secret, in that information is not released to the public concerning
e support nor would there be any public acknowledgment of it if word
0U out- But, we could just as easily consider the action a form of
‘Cciet diplomacy. In fact, the term “ covert action” would probably be
le<- to this sort of activity only if it were undertaken by an intelligence
CV’ suPplying secret aid via diplomatic or military personnel might not
e so considered.
 Such activities are often small scale and relatively noncontroversial, and
hence they are unlikely to make their way into the newspapers. They may
however, be important both technically (by providing capabilities that the
friendly government could not provide for itself) and psychologically (as an
indication to the receiving government of the reliability and effectiveness of
its patron’s support). However, this absence of visibility and controversy
need not always be the case.
First of all, the same disputes about which governments should be con­
sidered friendly and worthy of assistance would be likely to occur as in the
case o f noncovert foreign aid. The secrecy involved does not change the
nature of such disagreements, although it may lead opponents of a program
to suspect that it was kept secret primarily to avoid any challenge to it.
Secondly, a dispute might arise from the nature of the assistance, espe­
cially if the recipient has an unsavory reputation.x For example, in 1974 the
U.S. Congress banned the use of foreign aid funds for police training; this
came about because of controversy concerning the human rights records of
Latin American regimes that were recipients of such assistance.9

Intelligence Support

One common type of support rendered to a friendly foreign government may

be noted in particular: the providing or sharing of intelligence information.
In part because this activity is routine, it would not typically be considered
covert action. In cases involving an exchange of information of equal value,
and the purpose of providing the other government with information is to
receive valuable intelligence in return, it seems more reasonable to view the
activity as intelligence collection than as covert action.
In other cases, however, where the exchange is one sided, the major
motivation of the party providing the bulk of the intelligence may well be the
desire to help the other government. In this case, it seems reasonable to
consider it covert support of a friendly government, since it enables another
government to achieve objectives it could not on its own, and it is, ot
course, carried on secretly.
While intelligence would typically be shared only with friendly govern­
ments, cases crop up in which a government would pass specific information
to another government it did not view as friendly to induce or enable the
latter to take a specific step in the interest of both. This possibility, where the
motivation is less to help the other government than to make it take a specific
action, is discussed below as a method of influencing how a foreign gov­
ernment perceives its situation in the world.
 Influencing Perceptions
1*. ijjg situation discussed above where one is interested in supporting a
U ejan government whose policies are already favorable to one’s interests,
rnore common and more characteristic covert action tries to change
jf16 . behavior, events, or circumstances to further one’s interests. This is
done either by influencing the foreign government’s actions or by influ­
encing the foreign society, or groups or sectors of it, independently of its
government’s actions and often against its government’s wishes. The ulti­
mate °oal is to influence the government’s policy or to create the conditions
for a change of government. The next two sections discuss how actions,
events, or circumstances can be affected by influencing a foreign govern­
ment’s perceptions or those of elements of a foreign society about political,
military, or economic matters.

Influencing the Perceptions of a Foreign

Government
Agents o f Influence
The simplest and most direct method of affecting a foreign government’s
actions is to use an agent of influence— an agent whose task is to influence
directly government policy rather than to collect information. Since an agent
who is a high-level official in another government would be in a position to
do both, this distinction could be more a theoretical one than a practical one.
In such cases, using an agent whose prime function was intelligence col­
lection to influence its government’s policy might not be considered covert
action. However, an intelligence service might not wish to jeopardize a good
source by having that person try to influence his government’s policy if, by
so doing, he might call unfavorable and potentially dangerous attention to
himself.
An agent of influence could be an official of the target government or a
prominent member of the target country’s political class (an insider, so to
speak) who would have good access to government officials, opinion lead-
ers, and the media. If an agent of influence is a high official in the target
^ ern m en t, he might, ky himself, be able to take actions that benefit
Mother government. The more normal situation in a large, bureaucratic
overnment, however, is that he will be most useful in persuading col-
vi | S t0 ac^°Pt P°licies congenial to another government’s interests. Ob-
ben cannot openly advocate policies on the grounds that they are
Cla 1° some foreign, presumably unfriendly, power; he must seek to
agent he is.
A recent example of an agent of influence was Pierre-Charles Pathé
politically well-connected Frenchman who, in 1979, was convicted of es
pionage for the Soviet Union and sentenced to five years in prison. As a
political insider, he knew a great deal about high-level politics in Paris as
well as about the personal lives of major political figures, although he did
not have access to classified information. It appears, however, that his real
crime was less espionage than being an agent of influence.
In 1976, partially funded by the KGB, he started a political newsletter
called Synthèse (Synthesis). During the next three years, this newsletter
which reached a large part of the political elite in France (at one point, 70
percent of the members of the Chamber of Deputies were among its sub­
scribers), denigrated and attacked Western interests and policy; exaggerated
differences of interest and policy between France and other members of the
North Atlantic Treaty Organization, or NATO (in particular, West Germany
and the United States); and defended the USSR and its allies. Pathé was
arrested after he was observed receiving some money at a clandestine meet­
ing with a KGB agent in the suburbs of Paris.10
The nature of the relationship between an intelligence service and an
agent of influence varies according to the circumstances. In the case of
someone like Pathé, a sophisticated journalist and political insider whose
sympathies for the Soviet Union predated any clandestine connection with
the KGB (it is likely that he first came to its attention when he published an
article favorable to the Soviet Union in 1959), the relationship was probably
a flexible one based on shared interests rather than a strictly disciplined one
based on detailed instructions.
Soviet terminology recognizes a spectrum of agents of influence, depend­
ing on the degree and type of control exercised by the intelligence service.
Pathé would be an example of a trusted contact, that is, someone who is
willing to work with a foreign government to advance goals he shares with
it but who is not receptive to detailed instruction and would typically not be
paid. (The funds Pathé received helped defray the costs of publishing his
newsletter but were probably not seen by either party as compensation lor
services rendered.) He may be contrasted with a “ controlled” agent, who
receives and executes precise orders and who would normally be compen­
sated monetarily. At the other end of the spectrum would be someone who
is manipulated (for example, via his aides or his social contacts) to act in a
way that serves the foreign government’s interests but who is unaware of
this manipulation.11
 h r method of influencing a government’s actions is by providing it
-ts 0|- information (or misinformation) designed to induce it to act in
vv't . Jj manner. A recent example of this is the reported case in which the
a & d States, in 1983, passed to Ayatollah Khomeini specific information
“ Soviet agents and collaborators operating in Iran.” This resulted in
^ ^
two Ihundred rexecutions and the *“* r*closing
t v a z - 'i i t i A n e n n r l ta 1
down [of] the ccom-
i n rto u m o m
as many as
st Tudeh party,” which “ dealt a major blow to KGB operations and
Soviet influence” in Iran. The United States reportedly obtained this infor­
mation when a Soviet “ diplomat” in Teheran, Vladimir A. Kuzichkin,
defected to Britain in 1982; Kuzichkin was in fact “ a senior KGB officer
whose job it had been to maintain contacts with the Tudeh party.” 12
Although, in this case, the CIA reportedly participated in the activity,
there is no inherent necessity for intelligence channels to be used to pass
such information. Absent any intelligence service involvement, such an
activity would probably not be considered covert action, although its nature
would remain essentially the same. A historically significant example of this
type of activity, which would not be considered covert action at all, would
be the unsuccessful effort of British Prime Minister Winston Churchill, in
April 1941, to warn the Soviet Union of the impending German attack.
Under the circumstances of the time, it is easy to see why Stalin might
have suspected Churchill's motives. Stalin, probably possessing an exag­
gerated sense of English duplicity to begin with, certainly understood that it
was in Britain’s interest for relations between the Soviet Union and Ger­
many to worsen. Churchill might well have hoped that Soviet defensive
measures taken as a result of his message would be misinterpreted by Ger­
many as threatening, thereby provoking further German military measures
along the Soviet border, and so forth. As a result, Stalin regarded this true
warning as a deception effort and ignored it.13
While the two examples cited above involve the passing of true infor­
mation, there is no reason why misinformation, as long as it is plausible,
cannot serve the same purpose. One technique of making such misinfor­
mation plausible is called “ silent forgery,” which is a forged document
ttat is passed privately to a foreign government but is not made available
|° roedia. In the best case, the target government (the government the
orgery is designed to influence) accepts the forged document as genuine
and does not investigate the matter. The government whose document the
£ery purports to be never learns about it and is hence unable to deny
lls authenticity.
n example of this technique, which failed because the target government
sciicc agcm. n e claims mat in 1v / / tne soviets prepared a forged memo
randum, purportedly from the U.S. Embassy in Teheran and arranged for *
to be sent anonymously to the Egyptian Embassy in Belgrade. The memo
randum, stamped “ Top Secret,” discussed a supposed plot by the United
States, Iran, and Saudi Arabia to overthrow Egyptian President Anwar
Sadat. The covert action failed when the Egyptian government, in response
to an official inquiry to Washington, received information that persuaded it
that the memorandum was a forgery.14
This sort of disinformation effort may be directed specifically against the
target’s intelligence service, rather than the government as a whole. In this
case, the misinformation, instead of being given to the target government
is, so to speak, left lying around for its intelligence services to discover; for
example, deliberately misleading conversations may be held over telephone
lines the target government’s intelligence is known to have tapped. This type
of disinformation campaign is usually referred to as a “ deception opera­
tion” and considered part of counterintelligence rather than covert action.
To some extent, this distinction is arbitrary, since the operation’s basic
purpose and method remain the same. On the other hand, a deception
operation is a means of defeating the target’s intelligence collection and
analysis activities and depends on, and makes use of, a detailed knowledge
of the target intelligence service’s sources and methods. In these respects, it
is reasonably considered a part of counterintelligence.

Influencing Perceptions in a Foreign Society

As opposed to the techniques already discussed, those discussed below are
predominantly directed at influencing currents of thought within a foreign
society rather than its government. This is a more amorphous task, and it is
harder to measure what effect some of these techniques have in any given
case. While many of the examples may seem insignificant taken one at a
time, the cumulative effect on public opinion over a long period may be
large. These techniques, then, are more suitable for use by a government
that understands it is engaged in a long-term struggle than by one that deals
with each problem or crisis as a separate event.

Agents o f Influence
Although this technique would ordinarily involve agents who can more or
less directly influence governmental perceptions and policies, one could
image an agent whose main task was to affect a foreign country’s media. To
 e jjte character ot nis newsletter s reauersmp.
e examples, involving the Soviet use of agents of influence in Japan,
U been recounted by Stanislav Levchenko, a KGB major who defected to
^ U n ite d States in 1979.15 They include the following: the use of a prom-
-he member of the Japanese Socialist Party (JSP) to prevent another party
^ernber whom the KGB considered to be a Chinese agent, from achieving
^leadership position; the use of a senior correspondent of the Tokyo news-
a c Yomiuri to promote the publication of an article to obtain the release
of a Soviet military intelligence (GRU) agent who had been arrested in a
double-agent operation; and the use of a young American stringer for the
Associated Press to surface a letter purportedly from the wife of a Soviet air
force pilot who defected from Siberia to Japan in his MiG-25 that implored
him to return to the Soviet Union.16

Unattributed Propaganda
One of the more direct ways of attempting to influence a society is by
disseminating opinions, information, or misinformation through the avail­
able media, that is, by propaganda (as it is pejoratively called). For exam­
ple, nations with active foreign policies usually have radio stations (such as
the Voice of America, Radio Moscow, and so forth) that openly express
their views on international questions, much as newspaper editorials express
the views of the newspaper’s editor or publisher.
At times, however, a government may wish not to be officially associated
with the material contained in its propaganda. In these cases, it may put
certain opinions or facts into circulation in a manner that does not make their
origin apparent. This may be accomplished either by planting them in news
media it does not own or control or by means of media that appear to the
public to be independent but that are in fact controlled by the government.
There are two major reasons a government might resort to such unattrib-
uted (or “ black” ) propaganda.17 First, the target audience may be more
disposed to believe the propaganda if its origin is disguised and the ulterior
motive ot the propagandist is not evident. For example, anti-Soviet infor­
mation is more likely to be believed if it appears to come from an indepen-
ent source than if it comes from the U.S. government. For that reason, as
RadUSSe^ tIie U.S. government established Radio Free Europe and
a_ to Liberty as apparently private organizations.
cia te d ° ‘ *°F ^^*omatlc reasons> a government may not wish to be asso-
aud' Wlt^ certain opinions it nevertheless wishes to propagate to a given
,ence. For example, during the Iranian hostage crisis of 1979-81, the
 mg me iiusiage-utKing ai me u . in., wime us mac*. rauio station, the
National Voice of Iran, implicitly approved it and sought to inflame anti
American opinion in Iran.18
In addition to “ black” propaganda media, whose origin is meant to be
concealed, one may also speak of “ gray” propaganda, whose origin
while not totally or effectively concealed, is nevertheless not publicly ac’
knowledged. For example, the U.S. government established Radio Free
Europe and Radio Liberty in 1949 and 1951, respectively, to broadcast to
the peoples of Eastern Europe and the Soviet Union. Unlike Voice of
America, these stations were not to convey official American views but
were designed to provide the target populations with information about
their own countries that was not available in their own government-
controlled media, as well as with information about the West. They were
set up as private U.S. organizations; to support this cover, they even made
public appeals for contributions.
in fact, the stations were run by the CIA. According to Ray Cline, “ The
CIA organized this effort . . . because it was thought the broadcasts would
be more effective if their connection with the U.S. Government would be
concealed.” 19 But while the U.S. government connection was unacknowl­
edged, it was clear that the radios were an American operation. Following
the public confirmation by Senator Clifford Case that they had enjoyed
covert CIA funding, Congress decided in 1973 to support the radio opera­
tions openly via the Board for International Broadcasting, which was es­
tablished as an independent federal agency.20
A similar Soviet “ gray” propaganda technique is the use of Novosti Press
Agency. Unlike TASS (whose full name literally means the “ Telegraph
Agency of the Soviet Union” ), Novosti is said to be a nongovernmental
institution. This is seen as allowing it somewhat greater latitude in dissem­
inating views with which the Soviet government does not want to be asso­
ciated, although it is unlikely that its “ unofficial” status is taken very
seriously by anyone outside the Soviet Union.21
Another method of conducting propaganda in an unattributed fashion
involves planting stories in independent news media or arranging for books
to be written and published by authors and publishing houses that have no
visible connections with the government or its intelligence agencies. For
example, during a covert action campaign directed against Communist in­
fluence in Western Europe, the CIA used unattributed propaganda, includ­
ing secret subventions for publishing books and planting newspaper articles
via agents of influence working for wire services or newspapers.
One of the most famous CIA activities of this sort was the publication m
“ secret speech" attacking Stalin s •cun 01 persuuamy. 11K, aF^ vll nuo
also published in the New York Times, which received a copy from the
pepartment of State but which was presumably aware that the speech had
been originally obtained in a clandestine manner.22 Another example is the
c jA’s support for the writing and publication of The Penkovskiy Papers, an
account, based on the actual case materials, of the CIA's premier spy in the
Soviet Union in the late 1950s and early 1960s.23
A related technique is the use of front groups as a propaganda medium.
These groups, while ostensibly broadly based, are in fact under the control
o f a government and can be relied on to take positions consonant with that

government’s objectives. They thus serve the same function as unattributed

propaganda— they express views that serve a government’s interest but in a
format more likely to make them acceptable to the intended audience
Among the numerous Soviet front groups in operation, the World Peace
Council is the best known; others include international organizations of
lawyers, students, and women.

Forgeries
Another technique for putting material into circulation without taking any
responsibility for it is the preparation and circulation of forged documents.
In recent years, many forgeries purporting to be official U.S. documents
have come to light.24 While in most cases it is difficult to prove their origin,
the indicators point to the Soviet Union.
All of these techniques serve the same general purpose: to influence a
target audience’s perceptions so it will take desired actions. The material
used to do this— the arguments advanced by an agent of influence; the
content of the propaganda, attributed or unattributed; or the text of the
forgery— must be plausible to those whom one is attempting to influence. In
general, therefore, one would expect the false material to be mixed in with
some truths to enhance its plausibility. The resulting amalgam, which has
toe effect of misleading the target audience in some important respect, is
°ften referred to as “ disinformation” (from the Russian word dezinformat-
which is used to refer to the active measures technique of misleading an
audience to induce it to act in one’s own interest).25
course, it is possible that a totally true message (or one that was
a levec* t0 be so by those propagating it) could have the desired effect on
(in r^et autoence. The most effective part of the Khrushchev secret speech
Eur 6 SCnSe rectocing communism’s prestige in Western and Eastern
°Pe ) consisted of the revelations concerning Stalin’s crimes and the
VAAV/M^il VIIV tVAi HUI) UUUlVlUlV« ^ i l l l l l U l IJ 9 U1V L /llllÜ ll VS1 HjÇ
autften-
tic Zimmermann telegram, discussed in the previous chapter, was
a mas-
terstroke of unattributed propaganda.

Support fo r Friendly Political Forces

Another way of influencing events in foreign countries is the provision of
material support to friendly political forces, such as political parties, civic
groups, labor unions, and media. While this can be, and sometimes is, done
openly, covert aid may be more palatable to the recipient groups and is less
likely to lead to politically damaging charges of foreign interference in the
target country’s internal affairs.
Traditionally, the Soviet Union facilitated this type of activity by a fic­
titious distinction between the Soviet government and the Communist party
of the Soviet Union. While the former could pursue correct diplomatic
relations with other states, the latter, through such organizations as the
Comintern, the Cominform, and the International Department of Central
Committee, maintained relations with, and provided support for, “ progres­
sive” forces throughout the world. In response to Soviet activity of this sort
in Western Europe after World War II, the United States engaged in a covert
action program of aid to democratic political and cultural groups and trade
unions.
The immediate impetus was the Truman administration’s fear (in late
1947) that the Communists would win the Italian elections scheduled for the
following spring. To support U.S. overt propaganda efforts, the CIA con­
ducted such activities as providing financial and other support (for example,
training in electoral campaign techniques) to the non-Communist political
parties. It also attempted to influence the Socialist party against cooperation
with the Communists. As Ray Cline, former deputy director of the CIA,
explained,

These kinds of financial and technical assistance to the Christian Demo­

crats and other non-Communist parties, as well as the efforts to split off
Socialists from the united front group dominated by the Communists, had
to be covert. Italian party leaders could not afford to let Communists
obtain evidence that they were supported by foreigners because it would
blunt public anger at the Communist Party for its own financial and policy
dependence on the Soviet Union. Hence CIA got the job of passing
money and giving the technical help needed to get out the vote and win
the election.27
t Wa that would negate Communist efforts to expand Soviet political in-
acti°n \y e stern Europe.” 28 As described by the Church Committee, the
flucnce ram involved “ subsidies to European ‘counterfront ’ labor and
political organizations’’ that were
• tended to serve as alternatives to Soviet- or communist-inspired groups.
fThere were e]xtensive . . . labor, media and election operations . . . in
the late 1940s, . . . Support for “ counterfront” organizations, especially
the areas of student, labor and cultural activities, was to become much
in 29
more prevalent in the 1950s and 1960s,
As Ray Cline has emphasized, much of this funding went to organizations
of the non-Communist left that were not necessarily sympathetic to specific
U S foreign policy positions. He cites specifically the Congress of Cultural
Freedom and political journals such as Encounter in England and Der Monat
in West Germany.30 This strategy was presumably dictated by the view that
the major ideological battleground in Europe was on the left and that the
more conservative organizations would, in any case, have fewer problems in
obtaining funding domestically. It points out, however, a typical ambiguity
in covert action: it is often possible to work with people with whom one has
major disagreements; the embarrassment the revelation of such an arrange­
ment would cause becomes another reason for secrecy.
A more recent, and more controversial, U.S. program of this sort was
conducted in Chile following the inauguration of the Marxist Salvador
Allende as president in 1970. This program was designed to provide oppo­
sition political parties and media with the financial resources to survive the
country’s economic chaos and hostile regulatory action by the government.
The controversy surrounding it focused on two issues (aside from the larger
question of what the U.S. policy toward the Allende regime should have
been).
First, the program was condemned as unnecessary, since, according to the
Church Committee, CIA national intelligence estimates during the Allende
Period indicated that “ despite attempts to harass and financially damage
°Pposition media,” the opposition press (and in particular El Mercurio [the
nati°n s largest newspaper and the most important anti-Allende medium])
Remained tree and that “ the traditional political system in Chile continued to
°monstrate remarkable resiliency.” 31 However, the committee noted that
ose responsible for the estimates were unaware of the covert action pro-
KaiTL Thus, there was no estimate of whether those sectors would survive
Qbsent U.S. money.” 32
 * -- -*** muse
groups that were actively seeking to overturn the government (such as the
trucker’s union, whose long, damaging strikes provided the occasion for the
military coup) and to limit it to opposition parties and media. According t0
Gregory Treverton, a critic of covert action who was a staff member of the
Church Committee, “ The pattern of deliberations within the U.S. govern
ment suggests a careful distinction between supporting opposition forces and
funding groups trying to promote a military coup. The attempt to draw that
distinction was, so far as I can tell, an honest one. . . .” 33 Nevertheless as
Treverton concludes, this distinction was fundamentally artificial, one that
existed “ in the minds o f Americans, not Chileans.” 34 The opposition groups
chose their own strategy for achieving their objectives and picked their
tactics and allies accordingly; if the various anti-Allende forces wished to
cooperate with each other, then U.S. support for some groups helped their
allies as well.

Influencing Political Events by Violent Means

In the public mind, covert action is most often associated with violent
methods of influencing political events. As the discussion so far has shown,
many nonviolent techniques can be covertly employed to influence political
behavior, events, and circumstances in foreign countries. In addition to
these activities, which make up the vast majority of covert action operations,
there are techniques that involve the supporting or use of violence, to which
we now turn.

Support fo r C oups y “Wars o f N ational L i b e r a t i o n a n d “Freedom

F ighters’9
Support for political groups need not be limited to the peaceful opponents of
a foreign government but can be extended to groups that seek either to
influence that government’s policy through violent means or to overthrow it.
This could involve the support of an existing group or the creation of a
“ puppet” group to carry out these activities.
The precise type of support would depend in part on the group being
supported and the strategy it followed. It could take the form of military aid
for an insurgent fighting force, operational support for a coup d’état, exter­
nal political and economic pressure and internal subversion directed against
the target government, or training and economic support for the develop­
ment of cadres for a long-term guerrilla struggle. The two large, basically
- a ( 7AT\ a U I I U llia u c iw v A io v o w p p v y jiv iu i u*w £ ,c* v .x x x ix tx u u u g g i v a ^ ttiih M H it

afl^' t supPortec* Afghan regime and support for the contra resistance in
S,0V1raeua) exemplify this type of covert action.
№Ca ^rtjCUlar form of this sort of activity that has received a great deal of
^ ? n is patron-state support for international terrorism. The support
attCernments provide to terrorist groups can vary greatly in magnitude and
f ° v rtance At the low end of the spectrum, a patron state might provide
imP . facilities or a haven where terrorists would be safe from arrest while
tr -nCTtheir attacks or to which they could flee afterwards. At the higher
end this support could include money, weapons, genuine or forged pass­
es or use of the diplomatic pouch to send weapons or explosives into the
target country. The amount of control the patron state exercises over the
terrorist group’s activities can also vary; in some cases, it may be fairly
detailed, while in others almost nonexistent, as long as the general nature of
the activities is consistent with the patron’s interests.

Paramilitary
At the borderline of real covert action are cases in which a government uses
irregular (or volunteer) forces in a military conflict, either alone or in alli­
ance with similar groups or indigenous forces. Because a relatively large-
scale effort of this sort can hardly remain secret for long, we might argue
that this type of activity should be excluded from the category of covert
action. On the other hand, since intelligence services are likely to be tasked
to carry it out, it is usually grouped with more covert types of action. In any
case, since the activity would not be publicly acknowledged by the govern­
ment carrying it out, it remains covert action in that sense, regardless of how
transparent the pretense of noninvolvement becomes.

Specific Acts o f Destruction or Violence, Including Assassination

Finally, covert action can take the form of specific acts of violence, directed
a£amst individuals (such as the assassination of foreign government offi-
£!als’ ^ey political figures, or terrorists) or property (such as the French
°wing up the Greenpeace ship Rainbow Warrior).
i >e s bungled attempts to assassinate Fidel Castro, as well as its
vement with the political forces responsible for assassinating Patrice
^umumba of the Belgian Congo (now Zaire) and Rafael Trujillo of the
in minican ^ p u b lic, were among the main issues that surfaced during the
Rations of the agency during the mid-1970s and were minutely stud-
\ inclusion in President Gerald Ford’s executive order on intelligence of an
explicit ban on “ engag[ing] in, or conspiring] to engage in, political
assassination.” 36 This provision was retained in the executive orders on
intelligence promulgated by Presidents Carter and Reagan.37 This issue
resurfaced in the late 1980s in connection with U.S. policy toward Pana­
manian strongman Manuel Noriega, when it was alleged that the provision
prohibited the CIA from assisting anti-Noriega members of the Panamanian
Defense Force in planning a coup (for fear that it might involve or, at any
rate, result in Noriega’s assassination). From this point of view, the provi­
sion thereby contributed indirectly to the U.S. invasion of Panama in De­
cember 1989, which achieved the same result but at a much higher price.
One of the better-known recent covert actions involving violence against
property was the blowing up, by French intelligence agents, of the Rainbow
Warrior in 1985. This ship, which was attacked while at harbor in New
Zealand, was owned by the environmental organization Greenpeace, which
intended to use the ship to protest and interfere with French atomic weapon
tests in the South Pacific. A Portuguese crew member on board was killed
in the explosion, and two of the French agents involved were caught before
they could leave New Zealand.

Covert Action and Secrecy

Although secrecy appears to be at the very heart of covert action, its actual
importance and function, as noted at the beginning of this chapter, is not at
all a simple matter; it depends very much on the type of covert action and
the circumstances in which covert action is undertaken. The degree of
secrecy required, and the motivation for it, can differ greatly from case to
case.
In some cases, secrecy may be essential to the operation’s success. Ad­
miral Stansfield Turner, director of central intelligence under President
Carter, in the course of congressional testimony, provided an example of
such a covert action. The operation involved sending a CIA agent to Teheran
to facilitate the escape of six Americans who had taken refuge in the Ca­
nadian Embassy after the American Embassy was seized in November
1979.38 Any revelation of the operation (even if the identities of the CIA
agent and the six in hiding were not revealed) would have endangered the
safety of the seven people involved by heightening Iranian vigilance, by
leading Iran to invade the Canadian Embassy, or by complicating the proc­
ess of extracting from Iran the Americans using false passports.
Similarly, keeping certain details of an operation secret may be needed to
P ^ n ^ a l l y known. For example, we might consider U.S. assistance to the
&e ieen jn Afghanistan. A specific channel used to transfer supplies
^ h t be vulnerable to sabotage or diplomatic pressure by the Soviet or
governments if they knew about it. Thus, with respect to this detail
^ the covert action, absolute secrecy would be necessary, regardless of the
0t blicity the operation as a whole received.
P ^ s e s such as this, where secrecy about either the entire operation or a
ecific part of it is crucial to carrying it out at all, are probably rare. In
many other cases, however, secrecy is very important for the operation’s
effectiveness. In this category, we could consider the kinds of unattributed
ropasanda or support for political activity that have already been discussed.
These activities arc carried out covertly on the grounds that public aware­
ness of the foreign source of propaganda or of funding for political activities
would diminish the effect they could have. The concept of “ black” propa­
ganda, for example, is that information that would be questioned or rejected
were it known to come from an adversary or from some other foreign source
with an obvious ulterior motive will be more acceptable if its origin can be
hidden or falsely identified. (In some cases, secrecy would be required
because the funding of political organizations would violate the law of the
target country.)
In addition, “ black” propaganda can be used to sow dissension in the
adversary’s ranks by falsely attributing to him provocative words or actions.
In this case, secrecy concerning the true source of the propaganda is nec­
essary to support its credibility. Sefton Delmer, who ran British “ black”
propaganda operations in World War II, claimed that “ the simplest and
most effective ‘black’ operation is to spit in a man’s soup and cry ‘Heil
Hitler!’ ” 39 Similarly, many Soviet forgeries of recent years have involved
fabricating U.S. government memorandums that evince a hostile attitude
toward a third country.
In other cases, secrecy may be necessary to secure the cooperation of a
third party that is willing to help only if the activity is concealed, thus
enabling that party to avoid damaging its other foreign relationships. For
that reason, even in a case in which the contours of a covert activity could
e safely made known, keeping it secret might be important to gain the
cooperation of a third country.
A final category deals with cases where public awareness is avoided less
cause it would endanger those carrying out the operation or harm its
ctiveness than because diplomatic reasons make it desirable not to ac-
owiedge governmental involvement.40 For example, the lack of official
n°wledgment of an operation may inhibit an adversary’s response, even
 if the adversary is aware through intelligence information of the involve­
ment. This is particularly true if the adversary is a democracy in which the
government must obtain public approval or understanding for any response
it might make; the adversary may be inhibited from trying to obtain public
support by the fear that publicizing the incriminating evidence will jeopar­
dize the means by which it was collected. Thus it makes sense for a state that
supports international terrorism, such as Qadhaffs Libya, to disavow any
connection with a specific terrorist event, even when its complicity other­
wise becomes known.
For example, the U.S. government’s ability to respond to the Libyan-
supported terrorist bombing on April 5, 1986, of a discotheque in West
Berlin, in which an American serviceman and another person were killed,
ultimately depended on U.S. willingness to divulge sensitive intelligence
information proving Libyan involvement. While the U.S. government was
willing to do this, it paid a heavy price; the Libyans were made aware of the
extent of U.S. intelligence capabilities directed against them and hence were
able to reduce the effectiveness of those capabilities.
More generally, although governments are understood to engage in covert
action (as in espionage), covert action is frequently illegal under the laws of
the country in which the covert action is taking place. It also may be said to
be contrary to the norms of international law, for which nonintervention in
the internal affairs of sovereign states is a basic premise.
For these reasons, it is less provocative and less disruptive to diplomatic
relations not to acknowledge an operation even if the country adversely
affected by it is well aware of one’s involvement. The target country, either
in the interests of good relations, or because it cannot effectively prevent it,
may ignore the covert action; it is much harder for it to do so if the gov­
ernment conducting it publicly acknowledges what it is doing.

Plausible Denial
Closely related to this is the doctrine, long considered axiomatic, that
even if a nation’s involvement in covert action becomes known, the chief
of state should be able to deny that he authorized or even knew of the
action. He should be able to assert, with some plausibility, that it was
carried out by subordinates who acted without his knowledge or authority-
In the post-World War II period, this doctrine came to public attention
most forcefully when President Eisenhower disregarded it and publicly ad­
mitted that he had authorized overflights of the Soviet Union by the U-
reconnaissance plane. Eisenhower’s action came in May 1960, after a U-
plane, whose high-flying altitude had previously made it invulnerable to air
 was brought down by a Soviet surface-to-air missile while on a
nnaissance flight across that country. (Although not covert action, this
re- t involving an apparent violation of international law, raised the
^ m /issu e of plausible deniability.)
S In his memoirs, Soviet First Secretary Nikita Khrushchev claims that it
Eisenhower’s admission of responsibility, rather than the flight itself,
'hat caused him to scuttle the Paris “ Big Four” summit meeting scheduled
for later that spring.41 Eisenhower, however, felt that he could not deny
knowledge without suggesting that he did not effectively control U.S. mil­
itary forces, especially given the large infrastructure necessary to support the
flights. While it is possible to disavow the actions of a small group of secret
agents, who can be portrayed as renegades acting without or contrary to
orders, it is another matter to disavow an operation that obviously required
the cooperation of many people at various air bases around the world.
An example of plausible denial in recent years arose from the case of the
French intelligence agents who blew up the Rainbow Warrior. The French
government maintained that official involvement in the affair went no higher
than Defense Minister Charles Hemu, who was responsible for the DGSE
(Direction Générale de la Sécurité Extérieure, the French foreign intelli­
gence service subordinate to the Ministry of Defense) and who resigned
three months after the attack. President François Mitterrand escaped essen­
tially unscathed.42
In the United States, the doctrine of plausible denial was in effect abol­
ished by Congress in connection with its investigations into intelligence
matters in the mid-1970s. At the end of 1974, following revelations about
a covert action program in Chile, the U.S. Congress passed the Hughes-
Ryan Amendment, which prohibited CIA covert actions
unless and until the President finds that each such operation is important
to the national security of the United States and reports, in a timely
fashion, . . . to the appropriate committees of the Congress, including
[the Senate Foreign Relations Committee and the House Foreign Affairs
Committee].43

Congressional consideration of the amendment focused on the part that

squired the executive branch to inform Congress about covert actions,
co^r V e r a r n e n c *ment,s requirement that the president find that proposed
abT actlons are important to the national security destroyed the president’s
beo ^ t0 P^ausibly deny involvement in any covert action operation that
became Publicly known.44
In the
C|aj^ c°urse of the Iran-Contra investigation, Admiral John Poindexter’s
at he did not inform President Reagan about the diversion of the
proceeus ниш ию 11t t l l l u i l U l l l l k i и м к г » __
sented an attempt to revive the plausible denial doctrine. In particular
Poindexter testified that he did not tell the president about the diversion in
order to “ provide some future deniability for [him] if it ever leaked out.”45

The Abandonment o f Secrecy

Because of the increased congressional role in controlling covert action
among other reasons, the United States has, in effect, abandoned secrecy in
recent years with respect to some operations that would ordinarily be (and
often, still are) called covert action. For example, the question of aid to the
Afghan and Nicaraguan insurgents was openly discussed and debated in
Congress and in the press. The United States even asserted, in connection
with the 1988 Geneva agreement on Soviet withdrawal from Afghanistan, a
right to continue aiding the rebels as long as the Soviets provided aid to the
Kabul government. Despite the fact that these activities are called covert
action, they are overt in many respects (although some details may remain
secret).
In other cases, the United States has abandoned secrecy altogether for
activities that previously would have been considered to require it. In 1973,
for example, after CIA funding of Radio Free Europe/Radio Liberty (RFE/
RL) had been publicly confirmed, Congress replaced the covert funding
mechanism with overt government funding via a newly created independent
federal agency, the Board for International Broadcasting. While this change
was not undertaken voluntarily, it may not have had much effect on RFE/
RL’s credibility in the target countries. Regardless of how the initial dis­
guising of sponsorship enhanced RFE/RL’s effectiveness, it would seem
that, by 1973, the radio stations already had a long track record by which
their audiences could judge them.
Similarly, following President Reagan’s speech in London on June 8,
1982, in which he called for strengthening the infrastructure of democracy
on a global basis, the National Endowment for Democracy (NED), a fed­
erally funded private organization, was established in 1983. It provides
overt support (directly and via affiliated institutes of the AFL-CIO, the U.S.
Chamber of Commerce, and the Democratic and Republican parties) for
institutions— trade unions, business associations, other civic organizations,
media, and political parties— that can make possible or contribute to dem­
ocratic life in foreign countries where democracy is weak, threatened, or
nonexistent.
This effort was to some extent modeled on the international activities о
the political foundations associated with the West German political parties,
 similarly minueu punucai W i M “ * “ — ---------------
through
democracy. They are generally credited, for example, with providing
mote oport for the non-Communist parties in Portugal during the turbulent
rnU7
Vita dangerous period following the overthrow of the Caetano dictatorship in
31174 a|[n certain respects, this type of overt support for democratic institu-
V naiiially replaces the covert support the United States provided to
democratic forces in Western Europe right after World War II.

Covert Action and Intelligence

Havim7 reviewed the wide range of activities that come under the heading of
covert action, we now look at its relation to the rest of intelligence. Two
questions suggest themselves. First, should covert action be conducted by
the intelligence-gathering and analysis agencies, or should it be lodged in a
separate bureaucratic structure? Second, should covert action, from a purely
theoretical perspective, be considered part of intelligence?

Should Covert Action Be Separate?

Arguments Against a Separate Covert Action Agency
As noted in the first chapter, both the United States and Britain have dis­
covered through experience that having two organizations involved in run­
ning clandestine operations— one for intelligence collection (espionage) and
one for covert action— can create serious practical problems. During World
War II, the British Special Operations Executive (SOE), charged by Prime
Minister Winston Churchill with “ setting Europe ablaze,” was separate
from the Secret Intelligence Service (SIS, or MI6) and reported to the
minister for economic warfare. This gave rise to the inevitable interdepart­
mental rivalries as the two agencies competed for scarce resources.46
Similarly, in the United States, a special CIA component called the Office
ot Policy Coordination (OPC) was set up in 1948 to conduct covert action.
Although administratively part of the CIA, it operated on the basis of policy
guidance provided by the Departments of State and Defense, and its head
^ as appointed by the secretary of state.47 This office was separate from the
A s Office of Special Operations (OSO), which collected intelligence via
the1011^ 6 Eventu*Uy’ arrangement was found to be unworkable, and
for pV° °^48CeS Were cons°tidated in 1952 into the newly created Directorate
the anS 8 ^ ccortitng to the Church Committee, this step was due to (1)
anomalous position of the director of central intelligence in having an
was funded through the CIA) that nevertheless took policy guidance directly
from the Departments of State and Defense and (2) the competition between
the OPC and the OSO for the services of the same agents and the difficulty
given the rivalry, of ensuring operational cooperation between them where
necessary.49
Although conceptually different in function, covert action and human
intelligence collection rely on similar means, especially the secret cooper­
ation of agents able to operate effectively in the target country. In many
cases, the agents who would be useful for one function can perform the other
function as well; two separate offices could easily find themselves compet­
ing for the services of the same individuals. In addition, many of the support
functions, such as assuring communications between officers and agents or
making clandestine payments, are similar or identical and require similar
skills, contacts, and resources. Close coordination would be necessary to
ensure that the operations of one office in a given country did not inadver­
tently interfere with those of the other; such coordination might be difficult
to achieve, especially if it required that one office’s interests be subordinated
to those of the other. Thus, a strong case can be made for combining these
functions in a single organization whose operations in a target country or
region of the world would be controlled by a single chief.

Argum ents fo r a Separate Covert Action Agency

Despite the operational convenience of consolidating these functions, it has
been argued that the different nature of the tasks involved makes consoli­
dation inadvisable. In particular, the argument is made that giving the same
organization responsibility for implementing policy (via covert action) with
respect to a given country, on the one hand, and collecting and analyzing
intelligence about that country (including assessing the results of the covert
action), on the other, jeopardizes the objectivity of the analysis. In essence,
this is the same argument that is made for the existence of a central intel­
ligence agency— that is to say, one that is not a part of a policy-making and
policy-implementing ministry or department of the government.

Is Covert Action a Part of Intelligence?

Aside from this organizational question, the more theoretical issue remains
of whether covert action should be considered a part of intelligence,
including covert action among the elements of intelligence, I have noted t a
it is typically a function of intelligence agencies; this alone does not address
tllC tical purposes."'
the°re ^ er t|1js question, let’s get a better sense of just what covert action
2 already noted, the term itself is taken from the U.S. intelligence
*S ^ and does not necessarily have a clear counterpart in the usages of the
lexicon e otjjer countries. Furthermore, the various U.S. def-
.nte s of covert action were developed in the context of attempts to reg-
101 * it by law or executive order; they are not necessarily useful for the
present theoretical purpose.
P In reviewing the types of covert action, some (such as secret support for
friendly regime) seem almost indistinguishable from diplomacy, and
others (such as paramilitary activity) shade off into guerrilla or “ low-
intensity’’ warfare. What seems to be uniquely covert action is a middle
around between diplomacy and war that involves the secret manipulation
of the perceptions of others to induce them to take actions that one sees
as in one’s interest. As such, is covert action reasonably considered a part
of intelligence?
On the one hand, it has been argued that covert action— a means of
implementing policy— is so different from the rest of intelligence, which
deals with obtaining the information on which policy should be based, that
it is a mistake to consider it a part of intelligence. From this point of view,
it is more reasonably categorized as one of many foreign policy tools— such
as diplomacy, economic aid, “ most favored nation” tariff status, overt
propaganda, exchange programs, Peace Corps-type programs, public diplo­
macy, military aid, and the threat or use of military force— that a nation uses
to advance its interests.
On the other hand, it may be countered that defining intelligence as the
obtaining of information is too narrow: if obtaining information is so im­
portant, then the denial of valuable information to one’s adversary also must
fie an important part of intelligence. As becomes clear in the next chapter,
an effective and particularly profitable way of preventing one’s adversary
from learning the truth is to deceive him into believing a falsehood. From
ls perspective, covert action and counterintelligence look similar: both
involve affecting the adversary’s behavior by manipulating his perceptions,
tied 6 ^Ues^0n wfiether covert action is a part of intelligence is intimately
j UP wuh ffie question of what intelligence is. I will return to this question
n f e concluding chapters of the book.
 SPY VS. SPY
COUNTERINTELLIGENCE

O f all the elements of intelligence, counterintelligence is probably the

hardest to define. In its most general terms, counterintelligence refers to
information collected and analyzed, and activities undertaken, to protect a
nation (including its own intelligence-related activities) against the actions
of hostile intelligence services. Under this definition, the scope of counter­
intelligence is as broad as the scope of intelligence itself, since all manner
of hostile intelligence activities must be defended against.1 In a narrower
sense, however, counterintelligence often refers specifically to preventing
an adversary from gaining knowledge that would give him an advantage. I
examine this notion of counterintelligence first.
In conceptualizing what is involved in such a defense against an adver­
sary’s intelligence collection, let’s first distinguish between passive and
active measures. The most important passive measures, which seek to deny
the adversary the information he is seeking simply by blocking his access to
|t by, as it were, building a wall around it— are usually referred to as
security.’*-1 discuss the nature of the wall— prohibiting access to sensitive
form ation by individuals whose trustworthiness has not been ascertained,
the protection of documents and communications in safes and by encryption,
and so forth— in the section on security, below. First, an organization must
ccide what information is important enough to protect— what information
must be kept within the walls.

99
 m e t^iassincauou ut iiu u r iiia iiu u

Levels of Classification
A classification system categorizes information according to its sensitivity
which is to say, the amount of damage its revelation to a hostile foreign
power could cause and, hence, the importance of protecting it. In the United
States, the first such system for protecting national security documents was
promulgated by the War Department in 1912; it and the Navy Department
set up their own classification systems during World Wars I and II.3 pres_
ident Harry Truman established the first governmentwide system of classi­
fying information in 1951.4 The system is currently governed by an
executive order President Reagan promulgated, setting out the definitions
rules, and procedures for operating the system.5
The U.S. classification system attempts to classify information according
to the degree of harm to the national security its unauthorized release to an
adversary would cause. The more sensitive the information, the more care­
fully it is to be protected and the fewer the people who arc “ cleared” for it
(authorized to have access to it). Under the current system, the basic levels
of classification are confidential, secret, and top secret, which are defined in
terms of the damage to national security their unauthorized disclosure rea­
sonably could be expected to cause:

• top secret: “ exceptionally grave damage“

• secret: “ serious damage“
• confidential: “ damage“ 6
In addition, numerous other classifications are used to restrict further
access to information. For the most part, these classifications refer to infor­
mation about technical intelligence collection techniques and capabilities
and are imposed by the director of central intelligence by virtue of his
responsibility, under the National Security Act of 1947, to “ protect intel­
ligence sources and methods from unauthorized disclosure.“ 7 The years
have seen, however, a proliferation of these special classifications (or “ com­
partments“ ) as various entities within and outside the intelligence commu­
nity have tried to protect information they see as particularly sensitive.
Periodically, attempts are made to systemize this process, and DCIs have
contended that they should have sole authority to authorize these special
compartments. It seems, however, to be an unending struggle, and the
repeated admonitions not to create new classifications testify to the streng
of the incentives to do so.
Heed to Know
tp i»

a5 ove, the more sensitive the information (in terms of level of

As note the fewer are (he Ilulllber of people authorized to have access
classihca cj je> however, mere clearance is not enough to obtain access
to *l- Information. People seeking access also must establish a need to know
to the lnton , ____c___-----------------------TkJn
tD f mation in order to perform their official responsibilities. This part of
the ^ f° rmation-control system is much less formal than that dealing with
the m Ces* in general, anyone controlling classified information is respon-
/bhTfor ascertaining a requester’s need to know before providing the
information. . . .
It is questionable whether this requirement constitutes an effective barrier
the unnecessary dissemination of classified information. In 1985, a De­
triment of Defense (DoD) Commission to Review DoD Security Policies
and Practices, headed by retired Army General Richard Stilwell, observed
that “ the principle that a cleared individual is authorized access only to that
information he ‘needs-to-know’ is generally not enforced.” 8

What Should Be Classified?

The basic answer to the question of what should be classified, at least as far
as the U.S. government is concerned, is contained in the definitions of
confidential, secret, and top secret given above. These may seem straight­
forward in principle, although the key terms “ damage,” “ serious dam­
age,” and “ exceptionally grave damage” are vague, both in themselves and
as applied to a concept as general as national security.
In fact, the real situation is much more complicated. The very way in
which the problem has been posed— what information should be walled off
trom public access and discussion— presupposes a fundamentally liberal
view of government and society in which the free flow of information is the
rule and its denial, by means of classification, the exception. This view is
Piobably the exception rather than the rule as far as history goes. The more
°mmon tradition has been that of governmental secrecy, broken when the
government itself sees some advantage in disclosing information.9
lish °nS^ er’ f°r example, the fact that democratic countries routinely pub-
of |ar8e arn°unts of data about their defense budgets. Although the notion
assitying a democratic country’s entire defense budget is out of the
iud 10n’ k°th the United States and the Soviet governments have believed,
the^ln^ ^ actions, that publishing their defense budgets is harmful to
e Ves and useful to their adversaries. With respect to the Soviet gov-
secret.1” U.S. intelligence, meanwhile, devotes consiaeraoie resources
determining the size of the Soviet defense budget and its major elements r
this information were not considered helpful to the U.S. government thes
intelligence resources would presumably be employed on other tasks Si C
ilarly, if the United States did not disclose its defense budget, the Sov' "
intelligence system would at the very least have to redirect some resources
from their present tasks to that of determining its size; Soviet intelligence
would also be less certain about American military developments and less
confident in its assessment of them.
What this example is meant to show is not that a democracy should
classify its defense budget but that the reason it does not do so has more to
do with internal political considerations than with abstract criteria for
classification.11 The key dilemma can be stated as follows: almost any
organization— political party, business corporation, football team, or
whatever— that is competing with other organizations will want, other things
being equal, to keep secret most information about its strategy for carrying
on that competition, the resources it has available, and so on. Thus any
information that is released damages its competitive position.
The same is true of the government’s national security activities. The
question is always one of balancing the potential harm to national security
against the requirements of the domestic political order. No natural harmony
exists between these two standards, and no unambiguous line can be drawn.
Despite the tendency to secrecy, any government, but especially a dem­
ocratic one, often finds it useful or even necessary to release information to
explain its actions and build public support for them. As noted in the
previous chapter, the Reagan administration, to justify and secure public
support for the 1986 bombing raid on Libya, released information derived
from decrypted Libyan cable traffic implicating Libya in the terrorist bomb­
ing of a West Berlin discothèque. In alerting the Libyan government that its
cable traffic was being read, the administration risked losing an important
intelligence capability.

Overclassification
It has been a recurring complaint that the U.S. government classifies too
much information and at too high a level. This criticism has been made not
only by congressional committees and nongovernmental critics of the mte
ligence agencies or a given administration, but by insiders as well, r
example, a 1985 DoD commission on security practices noted that
 is warranted, current poney specifics mai me Mguci vi a
^^ument is responsible for the classification assigned but frequently, out
^oC orance or expedience, little scrutiny is given such determinations.
0 \ i r l v , while challenges to improper classifications are permitted, few
take the time to raise questionable classifications with the originator.12

s0me extent, as already noted, secrecy is a natural characteristic of any

vernment. Given the great and, to some extent, unique openness of the
g0 rican political system, the practice seems particularly anomalous and
An raised greater objection in the United States than elsewhere. In recent
haS*s' the same issue has surfaced in other countries, particularly Great
Britain, where the laws governing the release of governmental information
are much stricter than in the United States.
Aside from the primary harm— the unnecessary denial of information to
the public, reducing its ability to understand and debate national security
policy issues— overclassification also could have the secondary drawback of
reducing the credibility of the classification system, leading to a tendency to
disregard its rules. If many innocuous documents are classified, the inhibi­
tions against the unauthorized disclosure of classified information are bound
to be weakened.
Whether such pervasive overclassification exists in the U.S. government
is a much more difficult question to answer. While any system that deals
with millions of documents each year is bound to produce some dubious
results, the major sources of disagreement arise from different views of what
standards should be applied to the classification decision.
A common claim of those who believe that overclassification is rampant
is that documents are often classified to save government officials from
embarrassment or censure. These people claim, in other words, that clas­
sification serves to shield officials from public accountability for their ac­
tions. While this is an unacceptable motive for classification, the situation is
oiten more complicated; the embarrassment is often not simply to an official
who may have done something improper vis-à-vis the public, but to the
government itself vis-à-vis foreign powers.
As was discussed in the treatment of covert action, secrecy is sometimes
esired more for diplomatic reasons than because it is operationally neces-
Sar>- The same is true with regard to foreign policy altogether. As in the
Case covert action, however, governments have generally tended to be
d °fe °Pen about many facets of foreign policy and less concerned about
0rnatic niceties. Governments are more likely now than previously to
admit that they conduct espionage and other intelligence activities. Thus
there is less room for invoking the excuse of governmental embarrassment*
Nevertheless, valid diplomatic reasons may exist. For example, in 1977? the
Washington Post asserted that King Hussein of Jordan had received subsi
dies from the CIA, a claim that could only be harmful to him politically 13
Keeping such a relationship secret to avoid embarrassing and weakening the
political position of a friendly head of state is a legitimate motive for secrecy
that is not, strictly speaking, operationally necessary.

U nderclassification ?

Although it is often claimed that governments overclassify, it is worth

considering whether the opposite is more likely in certain areas. This might
occur with information developed outside the government that is significant
enough to national security to justify its being classified and controlled. The
most obvious example of this is information relating to nuclear weapons
technology. We would expect that information that could facilitate nuclear
proliferation would be carefully controlled.
In the United States, this issue was addressed in the Atomic Energy Act
of 1954, which defined restricted data as “ all data concerning (1) design,
manufacture, or utilization of atomic weapons; (2) the production of special
nuclear material; or (3) the use of special nuclear material in the production
of energy, . . ,’’14 Under the act, this type of information is subject to
controls regardless of whether it was developed inside or outside the gov­
ernment. (The same act also prohibits the dissemination of unclassified
information relating to the design of, and security measures for, various
facilities for the production or utilization of nuclear materials.15)
Nevertheless, considerable amounts of information about nuclear mate­
rials and weapons exist in the public domain in the United States. Thus
comes the occasional report that a bright physics graduate student, working
at a public library with unclassified sources, has drawn up plans for a nuclear
bomb that unspecified experts have acknowledged as viable.
Other information that cannot be classified, but that might, under certain
circumstances, deserve some protection include:
• financial information that would enable one to earn illegitimate (insider)
profits on financial or commodity markets,
• technological information a nation may wish to prevent adversaries
from obtaining, and
• personal information that might enable a hostile intelligence service
recruit government officials either by offering money to those wi
financial problems or by blackmailing them.
 ting this sort of information has involved communications security,
^ ssed below. However, it can, on occasion, involve other issues. For
d*scU j efforts have been made in the United States to apply the regula-
eXa concerning the export of technology to the dissemination of scientific
rï°nS , dealing with these technologies; however, no feasible scheme for
PaP -jpüshing this has been developed. In 1982 the deputy director of
a°ntral intelligence, Admiral Bobby Ray Inman, noted the problem in a
CC ch to the American Association for the Advancement of Science. Speak-
of theoretical and applied cryptologic research, he expressed concern
‘°that indiscriminate publication of the results of that research will come to
the attention of foreign governments and entities and, thereby, could cause
irreversible and unnecessary harm to U.S. national security interests.”
Among the areas where publication of technical information could harm
national security, Inman included “ computer hardware and software, other
electronic gear and techniques, lasers, crop projections, and manufacturing
procedures.” 16
Other than the rarely invoked Patent Secrecy Act of 1952, which allows
the government, on national security grounds, to order that an invention be
kept secret and to that end to withhold any patent for it, no effective mech­
anism exists for dealing with these issues. Attempts to invoke general export
controls to prevent, for example, the dissemination of technical papers at
conferences attended by foreigners have, by and large, been abandoned in
the face of public controversy.17

Security
Security measures are steps taken to obstruct a hostile intelligence service’s
ability to collect intelligence. Such measures are designed to prevent a
hostile intelligence service from either gaining access or exploiting any
access it may have to personnel, documents, communications, or operations
t0 gain important information; they constitute the wall surrounding classified
form ation. The more traditional aspects of security— which deal with
Protection against the adversary’s human intelligence collection efforts—
are discussed in this section; those that deal w'ith his technical collection
Capabilities are discussed later under the heading “ multidisciplinary
counterintelligence. ”

Personnel Security
pe
before ?e .secur^y involves procedures for screening potential employees
Inng them for a job that gives them access to information a hostile
 pioyees commue 10 meet me summitu?> tut iu autu mi^miauon \
screening procedure’s primary function is to judge the potential employee’
willingness and ability to keep classified information secret. The key ele
ments of this judgment involve the potential employee’s character and
loyalty.18 Judgments about character must consider both the individuals’
mental stability and whether, for any reason, they would be vulnerable to
blackmail by a hostile intelligence service.
In the United States, the screening investigation determines whether an
individual is granted a security clearance, that is, authorized access to clas­
sified information. The investigation relies on information supplied on a
security questionnaire by the individual, supplemented and verified by a
national agency check (interrogation of the data banks of various law en­
forcement and other government agencies), and depending on the sensitivity
of the information to which access is to be granted, interviews with friends
and acquaintances, present and former neighbors, work- and schoolmates,
and so forth. In the case of particularly sensitive information, a periodic
reinvestigation is required to maintain access.
Background investigations of this sort are probably not very effective in
assuring the loyalty and character of personnel to be granted access to
classified information. Various legal prohibitions, discussed in the next
chapter, prevent the FBI or other government agencies from collecting and
maintaining membership lists of what used to be called subversive organi­
zations such as the Communist party of the United States, even if the
organization openly advocates the violent overthrow of the U.S.
government.19 A member of such an organization, who was prudent about
discussing his membership with acquaintances and who did not disclose the
membership on the security clearance forms, could reasonably hope to
remain undetected.
In addition, societal changes in the United States have made it consider­
ably harder to get candid responses from acquaintances, colleagues, neigh­
bors, and others. First of all, there is a typically American resentment at
government snooping, which means that respondents are disinclined to pass
on negative information. In addition, the increased geographic mobility of
U.S. society implies that past acquaintances and neighbors are harder to
locate and may have only a superficial knowledge of the candidate. The
changed political climate means that respondents are less likely to take
questions of loyalty and subversion seriously. The same is true with respect
to questions about character and life-style. For example, the vast increase m
narcotics usage among the middle class (and, in particular, its young) during
the 1970s makes it harder to distinguish the cases in which past drug usage
11IV in Wincn 11
from tl*>se there is the question of whether informants can be assured that
•can d id remarks will remain confidential. Under the Privacy Act of
tlieIf individuals have extensive rights to access government files about
197 selves; while an exemption exists that allows an investigative agency to
ttien» Id the name of an informant who requests anonymity, this may not be
wlt. . , reassuring to potential informants who would be quite embarrassed,
en worse, if their identities were revealed.
orfirst such a potential informant may worry about the possibility that his
lame or some identifying fact (such as a phone number) will be released
A d v e r t e n t l y . Second, he may be concerned that, even though his identity
. not explicitly revealed, the subject of the investigation will be able to
deduce it from the substance of the information the government file con­
tains. (In other words, the release of a seemingly insignificant detail, which,
however, only the informant knew, could give away his identity. Since the
officer reviewing the file before its release would not likely be aware that this
detail provided a solid clue to the identity of the informant, he might well
release it, unaware of the damage he was causing.) Finally, as a Department
of Justice official told the Senate Judiciary Committee in 1978, “ In theory,
the [Privacy and Freedom of Information] acts provide an adequate basis for
protecting our sources, but whether they in fact do so is largely irrelevant as
long as our sources think they do not.“ 20
One method of augmenting the background investigation as a protection
against unsuitable personnel is the use of the polygraph machine, commonly
known as the lie detector. This technique has been used primarily by U.S.
intelligence agencies. Neither other Western intelligence services nor non-
Western ones place as much faith in it. Within the United States, the CIA
places the greatest emphasis on the polygraph. The agency requires that any
candidate for employment take a polygraph test, and that all personnel be
subject to periodic retesting as a condition of continued employment. How-
ever, extension of polygraph use to the rest of the U.S. government has been
strongly resisted, most notably by former Secretary of State George Shultz.
On the other hand, following the arrest for espionage of Geoffrey Prime, an
employee of the British Government Communications Headquarters
HQ, the British communications intelligence agency), U.S. officials
tolf^ ^ at l^C P°*y2raPh adopted by GCHQ as well. (Prime is believed
y ave compromised various comint capabilities that involved close U.K.-
les cooperation.) The implicit threat was that the United States might be
s willing to cooperate on communications intelligence matters if this were
n°l done.
 While use of the polygraph appears to have a strong deterrent effect, and
in many cases induces the revelation of information that was otherwise
concealed, its overall accuracy remains controversial.21 Aside from the
question of the frequency with which innocent subjects fail the test and are
unfairly rejected for employment or forced to resign as a result, foreign
intelligence services apparently have developed methods for beating i t 22

The Changing Nature o f the Threat

Since the mid-1970s, a succession of espionage cases has involved employ­

ees of U.S. government agencies or contractors.23 In these cases, the mo­
tivation has been primarily financial, although occasionally compounded by
emotional instability. Thus, these cases differ from the famous American
and English espionage cases of the 1940s and 1950s, in which the motiva­
tion was primarily ideological. This pattern suggests additional steps that
could be taken to improve personnel security, such as developing detailed
psychological profiles and instituting a system for alerting security officials
when individuals with access to classified information either run into finan­
cial difficulties or appear to be living beyond their means of support.

Physical Security
Physical security refers to the steps taken to prevent foreign intelligence
agents from gaining physical access to classified information. It deals with
such matters as the strength of safes in which classified information is kept
and alarm systems to detect any unauthorized intrusion into the areas in
which officials deal with classified information. For the most part, the re­
quirements of physical security are not arcane and differ only in degree, if
at all, from those a commercial establishment might take to prevent thefts of
merchandise or equipment.
One major difference, however, is that physical security seeks to safe­
guard not only the material objects, such as documents, that contain infor­
mation but also the information itself. This requires much stricter controls
on access to the relevant areas, since an intruder can quickly and unobtru­
sively implant a bugging device that would give a hostile intelligence service
access to what was being said within the classified work area. Therefore, d
is important to control what is taken into the area as well as what is removed.
Similarly, it is important to have some means of “ sweeping” an area to
detect any bugging devices so they can be removed. This in turn leads to the
development of less-detectable systems for monitoring conversations. The
high level of technical sophistication to which this spiral can lead has come
 ublic attention, most recently, with the problems concerning the new
t0 the United States is constructing to serve as the chancery (office
holding) for its Moscow embassy.
bUTlie building is constructed with precast concrete columns and beams the
viets manufactured without any U.S. surveillance. The Soviets took ad­
vantage of this to design a bugging system that could use the entire structure
as an antenna for picking up signals and relaying them. The system appears
to be quite sophisticated, and the United States has had difficulty under­
standing how it was intended to work. As former Defense Secretary James
Schlesinger testified in 1987,
In past years, the Soviets were sufficiently behind us that we were able to
detect penetrations, and neutralize them. . . . We now face a rising curve
of Soviet technology, with no gap between what the Soviets can do and
what we can do; indeed, in some areas they have been ahead of us. . . .

With respect to both embassy construction and operations, we have a lot

to learn from the Soviets. The Soviets have thought long and hard about
how to design embassies for security, and they have thought long and
hard about the construction process, . . .24

Counterespionage
The security measures discussed above are passive, since they do not go
after the hostile intelligence threat directly but seek to deny it access to
information. More active measures that try to understand how a hostile
intelligence service works to frustrate or disrupt its activities and ultimately
to turn those activities to one’s own advantage are usually referred to as
4Counterespionage. ’’

Surveillance Operations
An obvious way to learn about the activities of a hostile intelligence service
ls to mount surveillance operations (keep under constant observation)
against its officers wherever they operate. Such surveillance tries to deter-
minc where the officers go and with whom they communicate or are in
c°ntact; these contacts can then be used as leads for further investigation,
hile simple in concept, this is in practice a complex task. The officer’s
ec**aft, which was discussed earlier in connection with human intelli-
N0 ^ CO**ect*on’ ls devoted primarily to frustrating this sort of operation.
0n,y will the officer be trained in evading surveillance, but techniques
as brush passes and dead drops may be used to conceal the identity of
 lance is observed, it may prompt the officer to cancel any planned meetin
to avoid endangering his contact. Thus, it is important to hide the surveil
lance. To do this is difficult and requires a great deal of manpower, since an
officer trained in countersurveillance techniques will notice if the same
person is trailing him for any length of time.
Because this sort of surveillance is cumbersome and expensive, it js
important to target it on actual intelligence officers. One place to look, of
course, is at the kinds of official cover positions available to the hostile
intelligence service: diplomats, consular officials, trade representatives
journalists (for government-owned media), and employees of international
organizations such as the U .N ., when the employees are selected by their
own governments. The problem is to determine which individuals holding
these sorts of positions are really intelligence officers and which are what
they appear to be.
This problem can be attacked in several ways. In general, the more
knowledge an agent has about the operation of, for example, a given foreign
embassy, the more likely he can tell which officials seem engaged in actual
diplomatic or consular activity. If an official does not seem to be engaged in
such activity, he may be involved in intelligence wrork instead.
Similarly, it may be possible, by observing patterns of rotation and re­
placement of personnel, to determine who is replacing whom; if X has been
identified as an intelligence officer, X’s replacement is likely to be one as
well. Also, the same techniques used to collect intelligence in general may
be targeted on the hostile intelligence service and its presence in the agent’s
own country. The more an agent learns about that service and the way it
operates, the more effectively surveillance can be targeted. Finally, using
double agents, as discussed later in this chapter, may enable him to deter­
mine the identities of hostile intelligence officers.

Intelligence Collection
The most direct way to achieve counterespionage’s goals is to collect intel­
ligence directly from the hostile service, either by human or technical means.
(To distinguish this type of intelligence collection from the methods peculiar
to counterintelligence, it may be referred to as “ positive” foreign intelli­
gence.) Thus, the recruitment of a KGB official who was involved in all
espionage operations against the United States could solve the major U.S.
counterintelligence problem immediately. In this respect, counterintelli­
gence would not differ much from intelligence collection in general; its
^governm ental leadership, arraeu 1U1L&dj V/l Vmvi ----

pefectors
As with other types of humint collection targeted on a closed society like the
Soviet Union, recruiting and running an agent within the KGB or GRU
(military intelligence) is extremely difficult. Thus, the United States and
other Western countries have relied heavily on defectors from Soviet intel­
ligence services for counterespionage information. The most prominent re­
cent example was Vitaliy Yurchenko, the deputy chief of the KGB
department responsible for espionage against the United States and Canada.
After defecting in the summer of 1985, he provided information about
Soviet espionage successes against the United States. For example, although
he did not know their real names, the personal and operational details he
provided led to the arrest of Ronald Pelton, a former employee of the
National Security Agency, and the surveillance of Edward Lee Howard, a
former CIA employee.25

Double-Agent Operations
The other major method of conducting counterespionage is through double
agents. They are agents who, while pretending to spy for a hostile service,
are actually under the control of the country on which they are supposed to
be spying. Such agents may have originally been real spies who, upon being
detected, were 4‘turned,” or converted, into agents of the country on which
they were spying. Or they may be agents who pretended to volunteer to spy
tor the hostile intelligence service but who in fact remained loyal to their
country ( “ dangles” ). In between would be those who, having been ap­
proached by a hostile intelligence service, report the recruitment attempt to
their own country’s authorities and are encouraged by them to play along.
These three double-agent operations all serve the same counterintelli­
gence purposes. At the simplest level, they allow the counterintelligence
organization to penetrate the adversary’s cover mechanisms and identify
°fficers of the hostile intelligence service engaged in running agents. As a
rej>ult surveillance can be concentrated on the actual intelligence officers,
ho/ C ,*ess atten*i°n need be paid to the other officials who, although they
have the same cover status, really are diplomats, trade officials, and so on.
- — ~ .V. UVM kW U, »V W W J « » V *■* ’ * * * '" "

In addition to identifying the hostile intelligence officers, these operations

also
enable counterintelligence agents to learn their adversaries operational
 methods. Agents can observe how they pass instructions to their agents and
receive information from them, when and where they prefer to meet them
what precautions they take against being detected, and so forth. In short, by
knowing how and when the adversary communicates with his agents, coun­
terintelligence learns about its adversary’s tradecraft and can better counter
it. In addition, knowledge about the adversary’s tradecraft, and the general
pattern of his intelligence officers’ activities, contributes to counterintelli­
gence’s ability to identify them. Finally, if a double agent is provided with
some special piece of equipment, such as a radio transmitter specifically
designed for use by agents, his contact gets the chance to examine it. This
might lead to, for example, the interception of radio traffic between the
intelligence officers of the hostile service and their real agents.
From the instructions handlers give the double agents, his contact mieht
be able to learn about the hostile service’s collection priorities. This could
provide valuable clues to the adversary’s thinking about major issues.26
Alternatively, lack of interest in an area that would seem very important
might indicate that the adversary already had good sources of information
about it; this could provide an important clue for counterintelligence
investigation.
In addition to obtaining information about the hostile intelligence service,
its modus operandi, and its collection priorities, double agents can exert
some control over the service’s actions. The mere existence of a double
agent achieves this purpose to some extent: if the hostile service believes it
has an agent with access to specific information, it may not bother to recruit
another one. By deflecting the service’s activities, the double agent can
protect an important area of information. In any case, the handling of the
double agent absorbs the hostile intelligence officer’s time and effort, thereby
reducing the resources for recruiting and running real agents.
Furthermore, the window one double agent provides into the operations of
the hostile service may allow counterintelligence to dangle successfully
another double agent, perhaps one who has or appears to have access to
information the hostile service is particularly anxious to obtain. The first
double agent might also be able to support the bona fides of the second; for
example, the first agent could confirm the supposed fact that the second had
just received a promotion that gave him expanded access to sensitive data or
that he was having financial difficulties and desperately needed an additional
source of income. Finally, counterintelligence can use double agents to
deceive an adversary, leading him to misunderstand the situation and draw
the wrong conclusions, because it controls the information the double agents
pass to their supposed employers.
Double agents obviously must provide some information to their handlers
 remain credible. Frequently, this problem is solved with “ chicken
t0 __ostensibly classified, sensitive information that is, in fact, not very
te ortant. Alternatively, the double agent can provide true, important in­
f o r m a t i o n that the adversary is thought to have already obtained through

orne other channel. In these cases, counterintelligence must balance the

advantage of keeping the double agent credible against the damage done by
releasing the information, however trivial or duplicative, by which this
credibility is maintained. The goal is to provide as little useful information
as possible without raising the suspicions of the hostile intelligence service.
A more ambitious use of double agents involves having them provide a
judicious blend of information and misinformation designed to mislead the
adversary. This to some extent controls not only the adversary’s intelligence
collection but also his analytic capabilities. The payoff from doing this can
be enormous, although the difficulties involved are huge as well.
Deception is discussed more fully below. In this section, I look at two
examples of large-scale, very successful, double-agent operations, one con­
ducted during a war and one in time of peace.

Wartime Double Agents: The Double-Cross System

One of the best-known examples of a large-scale double-agent operation is
the Double-Cross System, by which the British, during almost all of World
War II, “actively ran and controlled the German espionage system in [their]
c o u n try "21 Starting with a single German agent who was detained on his
return to Britain from Germany at the beginning of the war, the British were
able to build up, under their control, a large network of supposed German
agents and keep the Germans believing in them until the end of the war.28
Among other things, the original agent sent back information the Ger­
mans used to produce false identification papers for additional agents to be
sent by parachute or boat, thus facilitating their capture by the British. In
°ther cases, new agents were told to contact ones already in place (and
a ready controlled by the British) for money or other aid. For the rest of the
Qar’ ^elped by some lucky breaks, the British kept control of the entire
nnan espionage network by intercepting new spies on their arrival.
s ° n Masterman, the MI5 (British security service) officer who ran the
ern, ^sted seven objectives for it:
to Contr°l the enemy espionage system,
lo catch fresh spies being infiltrated into the country,
gain knowledge of the personalities and methods of the German
Secret service.
 service,
• to get evidence of enemy plans and intentions from the questions aske \
by them,
• to influence enemy plans by the answers sent to the enemy, and
• to deceive the enemy about Britain’s plans and intentions.
The advantages for the British were enormous. For example, in 1940, the
original agent was given a code for communicating with Germany. Since
this code turned out to be “ the basis of a number of codes used by the
Abwehr [German military intelligence],’’ possession of it led to Britain’s
“ early and complete mastery of the [Abwehr code] system.” 29 Later in the
war, when the British became more confident of the Double-Cross System
they achieved the last two objectives, which involved deceiving the Ger­
mans. The climax came in 1944, when double agents were used to bolster
the deception campaign that induced the Germans to expect the main D-Day
landing at Calais rather than in Normandy.
Because of its cryptologic success (Ultra), British intelligence was in a
position to observe the German reaction to the information and misinfor­
mation reported by the double agents. Among other things, it could tell
which reports were forwarded to the top leadership in Berlin and whether
they were believed. This feedback was crucial for achieving the deception
objectives, because it allowed the British to emphasize points the Germans
had missed or modify the message to relieve any German doubts.
Controlling the enemy’s espionage network had its price: the network had
to produce enough intelligence to keep the Germans satisfied with it. Fur­
thermore, the information it produced could not be easily contradicted by
information available to the Germans through other intelligence channels,
including any uncaptured spies. (It turned out there were not any, but, of
course, the British could only become aware of that over time; at first, they
had to assume there might be other agents of whom they were unaware.)
However, this seemed a reasonable price to pay; since the Germans were
likely to have an espionage network in Britain in any case, it made sense tor
the British to run it.
In retrospect, the manager of the Double-Cross System believed that the
British running it were probably too cautious in the sense that the British
probably surrendered more true information than they had to and did not
make full use of the deception potential. Not only did the Germans possess
fewer other intelligence channels against w'hich the double-cross inform^
tion could be checked (as noted above, they had no loyal agents), but they
also proved to be remarkably trusting of the agents they thought they ha
 one occasion an agent was deliberately run in order to show the
rrm an s that he was under control, the object being to give them a false
■J a 0f our methods of running such an agent and thus to convince them
'hat the other agents were genuine. The theory was sound and the gaffes
•ornmitted were crass and blatant, but the object was not achieved . . .
. Germans continued to think of the agent as being genuine and
reliable!30

Cuban Double-Agent Operations Against the United States

On August 12, 1987, the Washington Post reported that a defector from the
Cuban foreign intelligence service, the DGI, had told CIA debriefers that
•‘an undetermined number of Cuban government officials, once believed by
the United States to be secretly working for the CIA, were feeding the
agency misleading or useless information prepared by the Cuban
DGI. • • -” 31 Following this revelation. Cuban media published detailed
accounts of the operations of many of these double agents, including pho­
tographs of them engaged in supposedly clandestine activities. That their
activities were known to the Cubans in such great detail tends to confirm
they were double agents for part, if not all, of the time they were purportedly
spying on Cuba. According to Cuban sources, some of the agents, appar­
ently aware that the United States now knew their true allegiance, sent
farewell messages to their U.S. handlers.
Without a careful analysis of the information provided by these double
agents, it cannot be determined whether Cuba was attempting systematically
to mislead the United States to take steps contrary to its interests. Even
without such an ambitious plan, the operation would have paid its way from
the Cuban point of view by absorbing large amounts of CIA time and effort
that otherwise might have obtained useful information; by revealing CIA
operational methods, thereby helping to counter its real agents, if any; and
y feeding U.S. intelligence analysis random pieces of information and
misinformation that confused it and prevented it from forming any coherent
Picture of events in Cuba.

Multidisciplinary Counterintelligence
(MDCI)
What Is MDCI?
Just
Ver as t^e hostile intelligence threat posed by a technically up-to-date ad-
™ Is not limited to human intelligence collection, so active counterin-
lemgence cannoi oe umiiea to counterespionage. is.amci, u must taice int0
account the full range of the adversary’s technical intelligence collecti0n
capabilities, including overhead photographic reconnaissance and coirtrnu
nications and signals intelligence.
Thus, the first task of MDCI is to assess the effectiveness of the adver
sary’s technical intelligence collection capabilities. This knowledge, in turn
indicates where one’s own information, communications, or activities are
vulnerable and how best to protect them. The actual measures that can be
taken may be labeled “ security” or “ technical countermeasures” and vary
from technology to technology. The multidisciplinary perspective, however
is important because it encourages one to look at the problem from the
adversary’s point of view.

Communications Security
Just as the interception of messages, both those in written and, more re­
cently, electronic form, is a major method of collecting intelligence, so
protecting the contents of messages has long been a major counterintelli­
gence task. In the twentieth century, this has focused on protecting elec­
tronic telecommunications transmitted either by wire or radio.
Messages transmitted by wire can be safeguarded by preventing the line
from being tapped (reading the message either by drawing a small amount
of current from the line or by sensing the fluctuations of the magnetic field
around the line caused by the current flowing through it). Radio signals, on
the other hand, cannot be protected; any receiver within range can pick them
up. Some signals, such as microwave transmissions (which are used to carry
long-distance telephone calls, among other things), are highly directional
(not dispersed in all directions but concentrated in a beam pointed at the
intended receiver), but even they can be intercepted if a receiver is properly
placed in line-of-sight contact with the emitter.
The contents of radio messages have been protected primarily through
encryption. Written messages can be encrypted in the ordinary fashion
before transmission. For voice messages, “ secure phones,” or “ scram­
blers,” are used; they distort the voice signals in a complex manner. Anyone
intercepting such a signal receives a meaningless jumble of sounds analo­
gous to the jumble of letters of a ciphertext. Restoring the original voice
message is comparably difficult.
Very secure scrambling devices are expensive. Few are available within
the U.S. government and the defense industry, and they are used only to
discuss classified information. Outside the government, they arc practica >
nonexistent. Thus, most long-distance telephone communication, vvhic
I l'-'* '
able to interception.
In the mid-1980s, in the United States, this situation came to the public’s
attention because of the controversy surrounding the location of the new
Soviet Embassy in Washington. Built in the Mount Alto neighborhood
several hundred feet above downtown Washington, the Soviet buildings
look out over the downtown and northern Virginia areas, including the
White House, State Department, Pentagon, and most other government
buildings. From this and other official Soviet buildings in the United States
(consulates and trade representatives in New York, Washington, and San
Francisco; the U.N. mission in New York; and residential and recreational
facilities in New York City, on Long Island, New York, and on Maryland’s
Eastern Shore), Soviet intelligence can intercept large volumes of long­
distance telephone calls.
In addition to government officials using ordinary nonsecure phones,
targets probably include high-technology corporations, commodity traders
dealing in products the Soviets buy or sell in large amounts (grain, oil, and
gold), and financial institutions. The Soviets also may target individuals
with access to classified information to try to learn compromising personal
details that can be used to blackmail them.
Various countermeasures have been suggested to reduce this vulnerabil­
ity. For example, one could redesign the microwave network to eliminate
those links passing near the Soviet intercept facilities and replace them with
underground cables. Alternatively, one could encrypt all signals transmitted
by microwave by scrambling the signal when it first goes on the air and
decrypting it when it has finished the microwave portion of its journey. Both
solutions have some technical weaknesses and would be expensive to im­
plement. In the future, the situation may be eased by the widespread use of
fiber-optic transmission, which is being adopted by communications com­
panies for economic reasons. Fiber-optic cables are virtually untappable
and, because of their great capacity (ability to transmit many signals simul­
taneously), tend to be cheaper than conventional copper cables or terrestrial
microwave.

Emanations Security
piece ot electrical equipment radiates electromagnetic waves; by inter-
lng these w av« it u- nnrriKia the characteristics of
waves, called “ em-
someone could reconstruct the text of a document being typed on
can be guarded against, for example, by shielding the electrical equipment
to reduce the intensity of the emanations, making them harder to intercept

Other Technical Countermeasures

Countermeasures can be devised to guard against other technical intelligence
collection means as well. With respect to satellite photographic reconnais­
sance, for instance, it is possible, assuming we can identify which of the
adversary’s satellites are for photographic reconnaissance, to predict their
orbits and to warn military or other sensitive installations to stop activities
and move sensitive equipment under cover when the satellites are able to
photograph them. Of course, the more we know about the satellites’ capa­
bilities, the easier it is to prescribe the necessary countermeasures. For
example, after we have determined the points on the earth’s surface directly
beneath the satellite’s orbit, wc still must know at what distance to the side
objects are still within the cameras’ range. Similarly, electronic or telemetry
intelligence collectors may be beatable by encryption, by jamming, by
shifting to frequency ranges the collector cannot pick up, or by ceasing
emissions when collectors are within range.

Deception and Counterdeception

So far, the discussion of counterintelligence has examined ways in which an
adversary’s intelligence collection capabilities may be countered. If done
successfully, the adversary will presumably lack the information needed to
analyze the situation in which he finds himself, and his actions, being blind,
will be less likely to serve his purposes. As the treatment of double-agent
operations has indicated, however, we can try to counter the adversary’s
intelligence operations more ambitiously by targeting his analysis capabil­
ity, that is, by taking steps to mislead him.

What Is Deception?
Deception is the attempt to mislead an adversary’s intelligence analysis
concerning the political, military, or economic situation he faces and to
induce him, on the basis of those errors, to act in a way that advances one s
own interests rather than his. It is considered a form of counterintelligence
because it attempts to thwart a major purpose of the adversary’s intelligence
operations; in addition, it often involves counterintelligence methods, sue
as double-agent operations.
 ' ful deception implies me orner siue s lnieingence ianure. m e re-
SÜCCe of course, need not be true; a side may make important mistakes in
verse’ Agence analysis even without any deception by the other side. Nev-
ilS 1 1 ss it is often possible, in cases of intelligence failure, to identify
eftl effort at deception by the other side. To what extent that effort is
s0rn^nsible for the failure is a more complicated question.
^Deception can be attempted in wartime or in time of peace, although one
would expect deception to be much more common in wartime. Deception
rin^es from tactical to strategic. Any battle that begins with a feint in one
seemr while the main weight of the attack falls on another, exemplifies
tactical deception. Examples of strategic deception are less common but
often very important, such as the Allies’ World War II deception operation
that misled the Germans about the location of the Normandy D-Day
landings.
Peacetime deception operations are not common, tend to be less well
known, and are sometimes hard to identify as such. Among the more spec­
tacular was one known as the “ Trust,” a Soviet organization that pretended
to be hostile to the new Communist regime but which was in fact established
and run from 1921 to 1927 by the forerunner of the current Soviet KGB, the
Cheka.32 Using this fake opposition group, the Soviets made contact with
anti-Communist émigré organizations and Western intelligence services,
thereby channeling and neutralizing any hostile activities they might under­
take; induced potential opponents within the Soviet Union to make contact
with it, thereby allowing the Cheka to learn their identities; and dissemi­
nated abroad false information about the internal state of the Soviet Union.33
During the 1950s and early 1960s, the Soviet Union engaged in deception
operations to convince the United States that the USSR possessed larger of­
fensive strategic nuclear forces than it in fact did. At a ceremonial flyover in
July 1955, for example, the Soviets “ displayed” twenty-eight Bison (Mya-4)
bombers— probably more than they had in their entire inventory— by having
the first group of planes circle around out of view of the spectators (including
the U.S. air attaché) and return for a second pass.34 Similarly, in the late
1950s, Soviet leaders, drawing on the prestige derived from the initial suc-
cesses ot their space program, made a series of exaggerated claims concerning
e Soviet ICBM program.35 In this way, the Soviets contributed to the
°mber gap and missile gap fears in the United States (that held that the So-
'jlets were about to surpass the United States in these measures of strategic
Yy ear caPability). The Soviet deception appeared aimed at inducing the
Be^r t0 ma^e P°Iitical concessions— such as concerning the status of
ln —that it would otherwise be insufficiently motivated to make.36
 * - * •■'“Vo
adversary to react to it. In wartime, for example, one might wish to launch
a surprise attack on the enemy, in which case the deception would be
devoted to convincing him that no attack is on the way. Sometimes, as in the
case of the Normandy D-Day landings, the enemy fully expects to be at
tacked, and it is unlikely one could convince him otherwise. In such a case
the deception is to convince him that the attack is coming at a time, in a
place, and/or in a manner other than what the actual plans call for.
In time of peace, on the other hand, it is less obvious what goal the
deception should pursue. One might wish to convince an adversary that one
is stronger than one really is to induce him to make political concessions that
he would not otherwise feel compelled to make. Alternatively, one might
wish to conceal one’s actual military strength to lull one’s adversary into
complacency and not provoke him into increasing his own military forces.
If one’s forces are limited by an arms control treaty, one could have the goal
of concealing a treaty violation, thereby leading the other party to continue
to limit his own forces.

The Prerequisites of Successful Deception

Blocking True Signals and M anufacturing False Ones

If we visualize the intelligence process as the reception and interpretation of

signals emitted by the activities of the side under observation, then imple­
menting a deception operation involves blocking, to the extent possible, the
true signals (those that reflect the actual activities) and substituting mislead­
ing signals. To use a simple example: the actual tank at point X is camou­
flaged, while a dummy plywood tank is placed out in the open at point Y.
If an enemy reconnaissance plane flies over the area, it may miss the tank at
X (the visual signal having been blocked by the camouflage) and report one
at Y (assuming that the false signal emanating from the plywood dummy is
similar enough to a true one). In the case of double-agent operations, the
real documents, for example, remain in government safes, while false ones,
produced as part of the deception operation, arc passed on by the double
agents to their handlers.
The first half of the task is the problem of security. If too many true
signals get through, the adversary is unlikely to be deceived, although he
may be so confused by the mixture of true and false signals that he cannot
form a coherent picture of the actual situation. Thus, the first prerequisite o
successful deception is the ability to block most, if not all, of the channe
 in p-11'
jties. **- " ra|) reason successful deception is more likely to occur in the
formeAhan in the latter.
R lo ck in s intelligence-gathering channels requires, among other things,
hensive knowledge of the intelligence channels by which the adver-
C° receives signals. A good counterespionage capability is necessary,
one well-placed human source could reveal the actual situation or, for
sl matter, the deception plan itself. Beyond that, one must know about the
adversary’s technical intelligence collection capabilities in order to thwart
them Thus, ships and planes can adopt radio silence (more properly “ emis­
sion control,” or “ emcon,” since radar and other electronic emissions
could give away the vehicle’s position as quickly as would radio transmis­
sions) to avoid detection by enemy comint or elint. If ground forces can rely
on landlines (communication via telephone or telegraph wires) rather than
radio, as did the German forces before the December 1944 Ardennes of­
fensive, the same effect may be achieved. Similarly, if the orbits of the
adversary’s photographic reconnaissance satellites are known, then activi­
ties can be halted and equipment moved indoors or camouflaged when the
satellites are overhead and able to photograph them.
The second half of the task, manufacturing false signals, is also planned
with the adversary’s human and technical intelligence collection capabilities
in mind. Double agents can pass whatever fake document or report one
wishes, as long as they can produce a plausible explanation for their access
to it. Deceiving technical collection systems is more complicated but pos­
sible. To achieve surprise at Pearl Harbor, the Japanese did not merely
impose radio silence on their attack fleet on its trip across the Pacific. In
addition, the fleet’s radio operators were kept busy passing bogus messages
to each other using transmitters based in Japan. Thus, if the radio signals had
been intercepted (it does not appear that they were) and the transmitters’
locations determined, the United States would have concluded that the Jap-
anese aircraft carriers were engaged in exercises in home waters.37 Simi-
lai|y, photographic reconnaissance can be fooled by dummy weapons or
vehicles; to know how accurate the dummies must be to be indistinguishable
rom real thing, one needs to know how good the adversary’s photo-
£raphic reconnaissance capabilities are.38

Feedback

n conducting a deception operation, one faces major uncertainties:

^ ere all the real signals blocked?
 • Did the manufactured signals reach the adversary .'
• Did he draw the desired conclusions from them?

To answer these questions, successful deception typically employs some

method of finding out how the adversary is assessing the situation. If he ‘
not alert enough to have noticed the false signals, or if he has not interpreted
them as the deceiver wished him to, then more can be manufactured to get
his attention and lead him to the desired interpretation. If he begins to sense
anomalies in the (false) picture of the situation the deceiver has planted in
his mind, new signals can be created to explain them away. If enough true
signals have reached the adversary to enable him to understand the situation
correctly, one may wish to abandon the deception and change plans.
Feedback can be obtained in many forms. In some cases (of wartime
deception or of deception in support of a surprise attack), the adversary’s
actions (or lack of them) may be sufficient indication of whether he has been
deceived. Thus, the absence of any signs of heightened military activity
(which would have been easily observable) by the Hawaiian-based U.S.
fleet in the first week of December 1941 told the Japanese what they needed
to know. In other cases, adequate feedback may require good intelligence
about the other side’s views of the situation; in the case of the British World
War II Double-Cross System, this feedback was provided by the Ultra in­
tercepts, which gave the British good access to the internal deliberations of
the German high command.
The more long-term and strategic the deception, the more important good
intelligence feedback becomes. The deceived party’s responses to such de­
ception take longer to become manifest; thus, one needs some other way of
knowing whether the bait has been taken. In general, peacetime deception
operations require better intelligence feedback than those conducted in war­
time. In wartime, the adversary is more likely to have to act quickly based
on his understanding of the situation, thus perhaps revealing whether he has
been deceived. During peace, the adversary is under less pressure to take
actions that indicate his understanding of the general situation.

D eception and Self-Deception

The false view of the situation one wishes an adversary to adopt must be
determined by the action one wishes the adversary to take. Nevertheless, the
view must be plausible to the adversary; in fact, success is more likely if the
deception scenario is based on what the adversary thinks is the case anyway-
For example, the D-Day deception worked so brilliantly in part because
Hitler was already convinced that the Allied landing would take place
faiai»* "'
nals only had to reinforce this view and prevent it from Deing unueniimcu
slgany true signals that managed to get through: this is an easier task than
^ tQ induce the adversary to adopt the deception scenario in the first
v e~Thus, it is not accidental that most of the impressive deception suc­
cesses involve a large element of self-deception.

Counterdeception
Experience shows that defeating every attempt an adversary might make at
deception is very difficult. Even as they were running Double-Cross, the
British were being similarly tricked by the Germans with respect to their
intelligence and guerrilla-type operations in the Netherlands. Starting in the
late summer of 1941, with the arrest of a Dutch intelligence agent working
for MI6 (the British foreign intelligence agency), the Germans managed to
catch all subsequent British agents sent to the Netherlands by the Special
Operations Executive (SOE, whose mission included sabotage and support
for anti-German resistance groups).
Having broken the cipher system used by these agents, the Germans
continued to send messages to Britain that purported to come from the
agents, who had in fact been arrested. Some arrested agents cooperated in
preparing and transmitting the messages; in other cases, German radio op­
erators impersonated them. Among other things, these messages made the
arrangements for additional airdrops of agents and supplies, which were in
turn captured immediately. The operation, which the Germans called Nord-
pol (North Pole) and Englandspiel (the match against England), continued
until two of the captured agents escaped from the camp in which they were
being held. At this point, concluding that the game would be given away in
any case, the German military counterintelligence officer in charge of the
operation chose to end it with a plaintext message to London on, appropri­
ately enough, April 1 (1944).39
One incident from this operation shows how strong the psychological
resistance is to the idea that one is being deceived. When British agents in
Holland enciphered their messages for transmission back to London, they
Were supposed to include a security check— a specified deviation (which
w°uld appear as a simple, random error) from the cipher system that would
reate to those receiving the message that they really were who they were
supposed to be and that they were operating freely and not under German
oinpulsion. An agent might be instructed to make a “ mistake” in the
in C ^ctter ° f each outgoing message. If the Germans obtained the cipher-
s>stem but did not understand the principle of the security check, any
 o- — j ---- -- ------- imtw *v r»uuiu nui humain the
required error.
In this particular incident, the Germans used their own radio operator to
transmit messages in the cipher system they obtained from a captured British
agent. Since a radio operator can often be identified by the manner in which
he hammers out the dots and dashes of Morse code (his “ handwriting” ) the
first message explained that a new operator had been recruited in Holland
because the original agent had sprained his wrist. This message contained no
security check at all and should have been recognized as bogus by SOE
headquarters in London; this was, after all, the very situation for which the
security checks had been devised. Instead, the staff officer in London
disturbed by the failure to follow proper procedure and seemingly oblivious
of the reason the procedure had been instituted, replied with an order to
“ instruct [the new radio operator] in the use of his security check.” 40
In any case, understanding deception is the first step toward figuring out
how to avoid being deceived; by understanding the factors that facilitate
deception, one can at least be alert to the possibility of deception and
recognize some warning signs. One is particularly vulnerable to deception
when one is dependent on a small number of channels of information and
when the adversary is aware, at least in general terms, of the nature of these
channels and their mode of operation.
For example, a heavy dependence on photographic reconnaissance satel­
lites, whose identities and orbits most likely will become known to those
whom they are photographing, may make one vulnerable to being deceived.
The adversary, knowing when his facilities can be photographed, may make
sure certain things are not seen: the items may be moved into garages or
sheds or covered with camouflage or tarpaulins whenever a photographic
reconnaissance satellite passes overhead. Similarly, the adversary knows
when to display dummy equipment to increase the chance of its being
noticed and mistaken for the real thing.
The situation is obviously worse (from the point of the potential deception
victim) if the adversary can find out ahead of time not only which sites can
be photographed but which will be photographed. Similarly, it is also worse
if the adversary knows the satellite’s precise capabilities. If he knows the
satellite’s ground resolution distance (roughly speaking, the smallest object
the satellite can detect), for instance, he knows how similar to the real thing
the fake pieces of equipment have to be to appear identical to the satellite.
For this reason, it was an important intelligence coup for the Soviets to
obtain the manual for the U.S. KH-11 photographic reconnaissance satellite.
They bought it for a mere $3,000 from William Kampiles, after he resigne
his position as a junior CIA officer in November 1977.41 It is also for this
 t h a t S & t £ l l U C p i i u i u a VW H IC H i i i i g i n m u ic a ic tiic wx u u a u u ic
reason
,ellite detects) are classified.
Sa'n ce one understands the risk of being deceived that comes from heavy
on a single known channel, one decides what to do about it. The
rcl corrective is to maximize what might be called “ unexpected
E lection ’’__taking photographs at times and places the adversary does not
C ticipate. Given the predictability of orbits, this is hard to do in the case of
Stellite reconnaissance; along the borders or coasts of the adversary’s ter-
litory where aerial photography is possible, this is easier to achieve.
Nevertheless, certain steps are possible even with respect to satellites.
Although the adversary may be able to determine (by studying its orbit, for
example) which satellites conduct photographic reconnaissance, it takes him
some time to do this and to warn his military facilities accordingly. During
the satellite’s first few hours of flight, it is likely that the pictures it takes are
unexpected. Hence, it makes sense to consider carefully which facilities
should be photographed on a satellite’s first few orbits. Similarly, thought
could be given to ways to prolong the period of unexpected collection,
perhaps by varying the launch and orbit characteristics of the satellites, so
that it takes the adversary more time to determine their function.42 Finally,
if it were feasible to have several photographic satellites in orbit at the same
time (including, if possible, dummies the adversary could not easily distin­
guish from operating satellites), the adversary might find it impossible to
protect sensitive military equipment and operations against all of them with­
out impeding his activities to an unacceptable degree.
Conversely, many proponents of signals intelligence (and, in particular,
communications intelligence) argue that they prefer it to other types of
intelligence collection because several factors ensure the reliability of the
information it collects. Comint may be collected in vast quantities, which
implies that the true signals are more likely to get through. By the same token,
it would take a larger effort to deceive comint by means of a fake radio
network (which would still require the use of real resources and personnel
10 create the signals) than that involved in displaying dummies or turning an
agent and running him as a double.43
Even more important is the fact that the adversary will ordinarily be
unable to tell which of his many communications channels others may be
reading; he may well transmit fake messages that are never intercepted,
1e some of the real ones are. However, even this is not foolproof: via
q ° naSe or some technical means an adversary may learn on which fre-
co CleS ° r cornmunicahons lines others are eavesdropping. If so, these
^ unication channels might be used in a deception effort.
im plicated situation of this sort, which illustrates the difficulty of
 - - 4 ----- A11C united
States and Britain dug a tunnel from the American into the Soviet гопе
Berlin to tap a set of telephone and telegraph cables that linked the Sov‘°
Air Force headquarters at Karlshorst with the city. The tunnel was b n*
jointly by MI6 and the CIA. Unfortunately, George Blake, a senior Ml6
officer in Berlin, was a Soviet spy. (He was finally tracked down arrect^
and convicted in 1961.)
Blake presumably told the Soviets about the tunnel as soon as he learned
about it. Even if he were not directly involved in the tunnel project, the
operation was so large and complicated that it is hard to believe that any alert
British or American intelligence officer in Berlin would not have quickly
picked up some idea of what was being done. Thus, one can assume the
Soviets learned promptly of the tunnel’s existence and purpose.
However, the Soviet telegraphic communications being transmitted on the
tapped lines were encrypted. According to John Ranelagh, whose history of
the CIA seems to have benefited from a considerable amount of information
from its former officers, the real secret of the affair was the existence of a
technique that enabled the CIA to recover the clear text of the encrypted
messages.44 This technique was not shared with the British; hence Blake
should not have been able to betray it. Thus, even knowing of the tap, the
Soviets might have continued to use the telegraph line for sensitive messages
in the belief that their encryption rendered them secure.
On the other hand, the British shared in the intelligence collected from the
Berlin tunnel operation. According to Ranelagh, “ The British were to carry
out the analysis of half the material” produced by the intercept.45 If so, then
it should have become evident that the United States had some method of
decrypting the intercepted messages; if Blake were aware of this, the Soviets
would have learned that the tunnel posed a real threat to them.
The tunnel operation ended April 21, 1956, about a year after the first
message w'as intercepted, when it was ostensibly discovered by Soviets in
the course of repairing one of the tapped cables that had been damaged by
heavy rainfall. At the time, the CIA, not knowing of the existence of a
Soviet spy in the MI6 Berlin office, regarded the Soviet discovery as acci­
dental. In retrospect, of course, one would have to consider the possibility
that the Soviets contrived to make the discovery appear so to protect Blake.
The interesting question is to what extent the “ take” from the operation
should be regarded as potentially deceptive. On the one hand, it seems like }
that the Soviets, via Blake, would have learned about the operation quickly«
that they let it go on so long suggests that they were using it to deceive
United States and Britain. On the other hand, one would have to look at t
actual information the operation produced; if it were very valuable then
t>eC° e iagj1 claims that the first indications that the Soviets had an agent in
in Berlin were acquired by the tunnel operation46). The intelligence
would have to be studied as a function of time to determine if the
ta*ce hprame aware of the operation before their ostensibly accidental
Sovi et s O , 4 7
discovery of it and, if so, when.
G any case, even without knowing that the adversary is listening to a
rticuiar communications link, a deception operation could create an entire
0f radio transmitters whose sole purpose is to send messages the adver­
sary might intercept; if the net is big enough, the adversary is bound to
tumble across it. This would be an expensive and elaborate measure, and
it is perhaps feasible only in wartime. In wartime, however, it is clearly
ossible: before the D-Day landings, the United States and Britain created
such a radio net to help deceive the Germans into believing that there existed
a large First U.S. Army Group, based in southeastern England, that was
poised to conduct the invasion against the Calais area. In fact, the Army
Group contained very few real soldiers; most of its divisions and other
components existed only on paper and in the radio chatter among the station
operators of the false network.
This was a key part of the D-Day deception operation; it not only sup­
ported the prelanding deception, but it also helped prevent the Germans
from moving reserves into Normandy in the crucial days following the
landings there. It allowed the Germans to interpret the Normandy action as
a diversion, a means of distracting their attention from the invasion that was
to come. According to a study of deception in World War II,

The networks which were set up to carry the fake radio traffic were
extensive and complicated. At the peak there were 22 fake formations
li e., the supposed headquarters of the army group and its constituent
armies and divisions]. . . .

fhere is little doubt that the simulated radio traffic, most of which was
carried on within a very short distance of the enemy and was therefore
easy to monitor, was— after the contribution of the double agents— the
most important factor in the overall deception.48

i he same factors help one determine the risk of being deceived by humint
cction capabilities. In particular, if one’s human intelligence collection
capabilities depend heavily on defectors, it increases one’s vulnerability to
the111^ ^ece*vec*- First, unlike the case in which one has an agent in place,
e adversary knows the defector’s identity. He can determine, more or less,
a form ation he had access to. Although he cannot prevent the defector
deception plan he develops. In this respect, information that comes frorri ^
defector is similar to “ expected collection.”
Second, the adversary can relatively easily plant a defector with fa]s
(deceptive) information designed to mislead one. If one relies heavily 0C
defectors, the adversary may provide several of them, whose reports al
though false, will nevertheless support each other. Suspicions that the Soviet
Union may have done just that to the United States, starting in the earl
1960s, led to some of the most severe internal conflicts in the history of the
CIA between those who took this possibility seriously and those who found
it overblown. The latter group finally triumphed when the CIA’s counter­
intelligence chief, James Angleton, the major proponent of the former view
was fired by DCI William Colby in December 1974.49

Counterintelligence Analysis
As the discussion of double agents and deception indicates, protecting the
integrity of one’s own intelligence operations can become very complicated,
involving much more than catching the occasional spy. In fact, cases may
be linked to each other in various ways; the overall task of guarding one’s
own intelligence apparatus against penetration and deception requires a
special office dedicated to counterintelligence analysis to serve as an insti­
tutional memory and to analyze these connections.
The best way to determine whether one’s own intelligence apparatus has
been penetrated is to acquire a high-level source in the hostile service (either
an agent in place or a defector). Even in this best case, however, it would
be a rare stroke of luck if the source could identify by name the spies in
one’s ranks. This would be the case only if the source were directly involved
in handling them or were a very high-ranking officer of the hostile service.
In the much more likely scenario, one’s source will be able to provide only
clues to the spies’ identities. Analytic work will be necessary to make those
clues yield results.
For example, one may discover that the adversary has had access to
several classified documents on a subject. Assuming they all came from the
same agent, one could review the distribution lists for the compromised
documents and note which officials had access to all of them. Alternatively,
the source may know that the spy held a meeting with his handler in a certain
foreign city on a specific date. Reviewing travel records would indicate
which officials with access to relevant information were in that city them
In addition, other clues may be available that suggest penetrations o
some sort. Thus, if one’s operation fails (an agent is discovered or
teeing—
was aware of the operation before it took place, it is necessary to investigate
how he may have learned of it. There will no doubt be many avenues
(individuals with access to the information as well as security vulnerabilities
or lapses) by which the information could have reached the adversary If
there is a series of such failures, however, it may be possible to narrow the
options; for example, maybe only one particular individual had authorized
access to all of them. In any event, such a series should alert an intelligence
service to the possibility that its adversary has effective human or technical
intelligence collection capabilities that must be discovered and neutralized.
 U

GUARDING THE
GUARDIANS
THE MANAGEMENT OF
INTELLIGENCE

T he preceding four chapters have discussed types of intelligence activities

and some issues that arise with respect to them. I now turn from the intel­
ligence activities themselves to the relationship between intelligence and the
government of which it is a part and to questions concerning the manage­
ment of intelligence by its nonintelligence superiors, or “ political masters.”
The management of intelligence presents two major sets of issues. The
first set arises from the secrecy in which intelligence activities are neces­
sarily conducted; it centers on the special difficulties that secrecy creates for
the political superiors whose job it is to oversee and control intelligence
activities. The second set arises from the uneasy relationship between ex­
pertise and policy-making. It deals with the problems of determining the
appropriate weight the views of the experts (who claim special knowledge
or expertise) should be given in governing the actions of the policymakers
(who have the actual authority to set policies and make decisions) and of
ensuring that the experts’ views receive the attention they deserve.

Secrecy and Control

a modem government, with its many thousands of employees arranged in
implicated bureaucratic structures, the problem of controlling their myriad
ancj ensuring that they are in accordance with the law and the
1Cles of their superiors, is bound to be difficult. With respect to intelli­
the CC a^enc*es' this hasic problem is compounded by secrecy, even though
secrecy springs from the legitimate need to keep knowledge of

131
CC IT din i n K / i l l ^ C I l C L üU U IC C 5 f IIJU U IU U J, ci ii vj u c u v m v /o u v /m m c DLlKI*
and restricted to the smallest possible number of officials within the goyC
emment. This secrecy hinders the management and control mechanisms th
are common elsewhere in the government.
In principle, of course, control of government activities is organized in
hierarchical fashion, so that every official is ultimately responsible, throu h
a chain of command, to the head of the government. Thus, the minimum
condition for control is that each superior have the right to know all the
information to which his subordinates have access. In general, this condition
is probably met, although the notion of plausible denial raises some impor­
tant questions in this regard.
Having theoretical access to information about an activity is not, however
the same as knowing enough to control it effectively. In other areas of gov­
ernmental activity, the direct top-down control exercised by superiors is sup­
plemented by other mechanisms, both formal and informal: examinations by
auditors or inspectors general, challenges and complaints by competing parts
of the bureaucracy, investigations by law enforcement agencies, legislative
oversight, press coverage, and complaints or other feedback from the public.
Applying these mechanisms to intelligence activities would require
spreading knowledge about them beyond the narrowest possible circle of
officials whose need to know derives from their actual involvement in them.
How far the knowledge is spread depends on the mechanisms, since one
may involve informing only a few additional executive branch officials;
another, members of the legislature and their staffs; and a third, the public
at large. In each case, a balance must be struck between the danger of
widening the circle of those with access to the information and the benefit
derived from the increased capability of the intelligence agencies’ political
superiors to control their work.

Plausible Denial
Because these control mechanisms all involve more dissemination of infor­
mation, one could say, in general, that a tension exists between secrecy and
effective control. With intelligence activities, however, achieving effective
control may be complicated by the doctrine of plausible denial, which
suggests that even fundamental top-down control by superiors may be placed
in doubt. According to this doctrine, intelligence activities that might cause
embarrassment (because they violate international law or for some other
reason) should be planned and executed in a way that allows the head o
government plausibly to deny that he had anything to do with them or even
^ ^ c o v e r t action, although it is also applicable to otner intelligence auiv-
° uch as espionage or aerial photographic reconnaissance violating the
itics •
ggt’s airspace.
This doctrine of plausible denial can complicate the control of intelligence
• 'ties to which it is applied. To be effective, it requires not only that
aC ledae of the activity be restricted to the smallest possible number of
Officials but also that there be no formal procedure by which it is approved
° d no paperwork in which the approval is recorded. The activity itself
a Id be conducted with a minimal amount of record keeping, and any tiles
created in the course of carrying it out would probably be destroyed once the
activity is completed. Obviously, this creates the conditions for misunder­
standings and uncertainty as to whether a specific action was authorized. In
particular, without a written record (or paper trail), it could easily prove
impossible to determine whether an activity had been approved by the head
of government or other senior nonintelligence officials.
All this does not mean that intelligence agencies in general, or any one in
particular, is, in the phrase attributed to Senator Church, a “ rogue ele­
phant.“ It suggests, however, that, other things being equal, an intelligence
agency may more easily “jump the rails” than other governmental organi­
zations. Against this inference, on the other hand, one should note that
intelligence agencies investigate the character of a potential employee more
carefully than other government offices and that the opportunity for illicit
personal enrichment in an intelligence agency is less than in many other
government offices that deal with economic statistics or regulation or that
conduct business on behalf of the government by, for example, awarding
contracts or leasing government lands for private exploitation.
Although Senator Church, chairman of the Senate committee formed in
1975 to investigate the intelligence agencies, used the term “ rogue ele­
phant ’ early in the investigation to describe the CIA, that characterization
was generally abandoned by the time the investigation was completed. In the
^lew Loch Johnson, a Church Committee staffer sympathetic to Senator
Urch, the chairman’s use of the term “ derived from a sense that the
evidence needed to be dramatized to have an effect upon the public,” and
,e ^na* report of the committee “ carefully steered clear of the ‘rogue
elePhant’ theory.” 1
sinai ma^0r actions the committee criticized, the attempts to assas-
ate Fidel Castro raised the most important questions about the agency
vol ra^s> According to the committee’s report on the CIA’s in-
Vement in assassination plots,
 ........... .............. — ^Uiu Con
elude that Presidents Eisenhower, Kennedy, or Johnson, their close ad
visors, or the Special Group [the interagency group that reviewed cov
action proposals] authorized the assassination of Castro.2

Instead, the record the committee found was bewilderingly vague ab

authorization. The key CIA officials said they felt they were fully authorized
to do what they did, but that this authorization had not been conveyed in
many words, either orally or in writing. In short, the committee confronted
precisely the sort of record, or rather lack of record, that one would expect
to find in a situation ruled by the doctrine of plausible denial. President John
F. Kennedy and his brother, Attorney General Robert F. Kennedy (who
played a major role in formulating policy toward Cuba), could have plau­
sibly denied any involvement or contemporaneous knowledge of the assas­
sination attempts.3
Thus, despite the a priori argument that control of intelligence agencies by
their political superiors may be problematic, the Church Committee record
does not actually show that this was the case regarding attempts to assas­
sinate Castro. Rather, it demonstrates the plausible denial doctrine working
successfully to shield the president from blame. This, of necessity, leaves
open the possibility that the CIA was out of control, but it seems unlikely,
given the lack of outrage from administration figures.4

For Whom Does Intelligence Work?

In any case, it is a peculiar control by a head of government that is con­
sistent with the plausible denial doctrine. It is unlike the ordinary control the
head of government exerts over departments or ministries through an official
chain of command and written orders, cabinet decisions, policy statements,
memorandums, and so forth. Its basis is a direct, or personal, loyalty of the
intelligence service to the head of government, rather than bureaucratic
subordination.
In past ages, intelligence services commonly worked for a monarch, chief
minister, or commander in chief in a personal capacity outside the ordinary
governmental structures. For example, Sir Francis Walsingham, the powerful
secretary of state to Queen Elizabeth I of England, supported a very effective
intelligence service largely out of his own funds and nearly went bankrupt in
the process.5 Similarly, the Duke of Marlborough, the British commander in
chief during the War of the Spanish Succession, used a traditional 2? perce
commission on the British-paid salaries of his army’s foreign troops, as w
as other perquisites (or tradition-sanctioned kickbacks, to use less eleg
lan^ ecret service. He regarded these funds as properly expended by him
CCS “ procuring intelligence and other secret service,” but he was later pros-
^°uted by his political opponents for embezzlement for his pains.6 In both
eC nces, the distinction between public and private funds or activities was
" Clearly, this sort of arrangement cannot easily accommodate itself to
C m odern bureaucratic state, in which the separation between officials’ public
duties and their private interests is supposed to be absolute.
But the issue goes beyond ensuring accountability for public funds. There
is the more fundamental question of who can authorize anyone to engage in
such activities at all. This is particularly important in a constitutional gov­
ernment like the United States in which the powers of government are
divided among the branches of government and political officials are re­
sponsible to the electorate. Under the U.S. Constitution, the president claims
the power to engage in intelligence operations on the basis of his foreign
policy and military responsibilities. It is not clear how the president’s sub­
ordinates can claim the right to exercise these powers without his authority.
At the same time, as discussed in chapter 4 on covert action, Congress has
insisted on a complete paper trail leading back from intelligence activities to
those who authorized them. Thus an important effect of congressional in­
volvement in this area has been to subject intelligence activities to formal
bureaucratic control within the executive branch.
But congressional involvement went beyond the question of how intelli­
gence activities should be authorized and controlled within the executive
branch. The House and Senate intelligence committees successfully asserted
a right to be kept “ fully and currently informed of all intelligence activities”
and to receive “ any information . . . concerning intelligence activities
which is in the possession” of any intelligence agency.7 Taken together with
a reinvigorated power of the purse (the power of Congress to appropriate or
withhold funds for intelligence, as for all other governmental, activities),
Congress could use and, to some extent, has used these levers to establish
a joint authority with the president over intelligence.
The result, in the words of former Deputy Director of Central Intelligence
Robert M. Gates,

ls that the CIA today finds itself in a remarkable position, involuntarily

Poised nearly equidistant between the executive and legislative branches.
e administration knows that the CIA is in no position to withhold much
information from Congress and is extremely sensitive to congressional
emands; the Congress has enormous influence and information yet re-
ma*ns suspicious and mistrustful.8
to explain a further problem of control: however adequate informal direction
can be for a head of government to control an intelligence service, it will not
let a legislative body feel it is exercising its authority over the service
effectively. For that, more formal means of control are necessary, such as
those already noted (formal right of access to information, notification of
certain types of activities, and the requirement that funds for intelligence
activities be appropriated by law). Thus, in the United States, the result of
congressional involvement has been to subject the CIA, as well as other
intelligence organizations, to full executive branch control and to the same
system of legislative oversight as any other part of the government.

Expertise and Policy

The previous section dealt with issues concerning how to manage intelli­
gence services that must conduct many activities secretly. Another set of
problems involves the difficult relationship between expertise and policy or
between those who possess specialized knowledge about an issue and those
who are authorized to determine and implement government policy con­
cerning it. This relationship is a problem in all parts of the government,
regardless of the policy areas with which they deal, and is thus a recurrent
theme in public administration theory.
Many issues and conflicts typically arise between expertise and policy. At
the center is the question of what role expertise ought to play in the policy­
making process. How should one reasonably draw the line between the
functions and responsibilities of the experts, on the one hand, and the
policymakers, on the other? Clearly, we w'ant policy to be guided by the best
information available; we would be very critical of a policymaker who
ignored the available facts and based his actions on his unsupported views
of what the world was like. At the same time, we want policy to be made
by those to whom the political system (via election or appointment) gives
the leadership authority; in any case, they must take ultimate responsibility
for their policies, regardless of the information on which the policies were
based.
Theoretically, contemporary social science resolves this problem by
means of the “ fact-value” distinction: social science can provide the facts,
(including contingent predictions such as “ if the government follows policy
X ,, it will obtain result Y / ’), but policymakers have a monopoly on choos­
ing the values to be pursued (whether to prefer result Y, to result Y2>or vice
versa).
This theoretical solution provides, however, very little useful guidance in
PraCUT ent predictions with any confidence in most areas; certainly with
C°n ct to the national security issues with which intelligence deals, this type
feS^redictive ability is not available. Second, most if not all of the debates
° f the national security arena center much more on differing assessments of
in afactual consequences of policies than on disagreements concerning which
values should be sought. Everyone can easily agree on peace, liberty, and
V sneritv as goals, but the means to achieve them are debated endlessly.
Conversely, if it could be stated with any certainty that a given policy would
attain these goals, the policy-making role would be diminished to the point
of extinction.
Since intelligence can provide only the roughest approximations of what
will happen, the theoretical fact-value distinction does little to illuminate the
appropriate roles of intelligence and policy. In practice, policymakers’
choices depend more on their views about the consequences of policies than
on their choices of goals. (Indeed, the choice of ultimate goals is considered
so self-evident that the issue is rarely addressed at all.) Thus, what policy­
makers do is not so clearly separated from what experts do as it first seemed;
hence, it is not surprising that the intelligence-policy relationship holds a
certain amount of tension. Each side often views the other side’s actions as
infringing on its territory.

“Killing the Messenger”

From the intelligence officer’s perspective, the difficulties between intelli­
gence and policy look like the latter’s fault. The difficulties stem, in their
view, from the policymakers’ tendency to disregard intelligence reports that
do not support the policies they wish to adopt or have already adopted and
to which they are committed. Intelligence officers think policymakers are
interested in the intelligence product only insofar as it can be used to support
their policies, which they have adopted for completely different reasons.
When the intelligence product does not support their policies, they either
tgnore it or try to “ cook” it by using political pressure to change its
intents.
This perspective has led to the view, which tends to predominate in both
academic and political discussions of intelligence, that the most important
c aracteristic of any intelligence-producing organization is its independence
r° m Policymakers. Only an independent intelligence agency, in this view,
f an res*st the political pressures that policymakers would otherwise bring to
ar on it to conform its analyses to their policy preferences.
n ^ e American context, this has provided the most important argument
its own collection and analysis capability. Such an organization, the ar
ment runs, will be free of the biases that would distort the output of
intelligence service that was part of the Department of State or Defense and
forced to take account of its parent organization’s policy preferences
budgetary interests.
This independence permits the intelligence agency to be guided entirel
by the data, without having to worry about how its product will affect the
policy process. Independence helps negate what may be called the tlkillin
the messenger” syndrome, the tendency to blame the messenger__the in­
telligence analyst— for bringing unwelcome news— analyses of the situation
that do not support the favored policy.
The common understanding of this syndrome is that it reflects some
irrationality on the part of the policymaker; it may perhaps best be illustrated
by a passage from Shakespeare’s Antony and Cleopatra. After threatening
the messenger who has informed her that Antony has married Octavia
Cleopatra explains her behavior as follows:

Though it be honest, it is never good

To bring bad news: give to a gracious tiding
A host of tongues; but let ill tidings tell
Themselves when they be felt.9

Not to be forewarned of unpleasant developments until they come up and

bite is an idea that obviously cannot guide an intelligence agency. Any
policymaker who adopted such a position would indeed be acting
irrationally.
However, a policymaker need not be influenced by this irrational motive
to wish to prevent an intelligence agency from reporting unwelcome news.
In a modem bureaucratic state, no official, even the head of state or gov­
ernment, receives intelligence information or assessments bearing on a given
policy that other high officials do not receive as well. Some of these other
officials may have been enemies of the policy and can be expected to use any
negative intelligence information or assessment to question and try to over­
turn it.
This difficulty is only exacerbated when intelligence is routinely shared
not only with the official’s colleagues and bureaucratic rivals, but also with
his political opponents (for example, the opposition party in Congress).
From this perspective, then, it is not at all irrational for a policymaker to
wish to ensure that intelligence provides the “ right” answer.
Of course, what is in the interest of any one policymaker need not be in
the nation’s interest. From the national point of view, we should see
^fortunately, these two goals may have conflictii1' ltc 1 t
ofthe best Policy reciuires that al1 of the releva,1| Inul
tl0nnable to everyone engaged in forming the polk V 1 \V x
ava and analyses will most likely support arg u m ^ 0
U<oposed policy; all these arguments should be we»j,
decision- Suppressing pieces of information or lines ^ ^,
hem from certain participants in the p o l i c y - m a k i n g l;V
1i bias the process and reduce the likelihood of it.yen; \ :\ h
P implementation, however, poses a different pnYler
chosen, the arguments against it must be ignored. , nK' 1 ,
basic policy question whenever a tactical decisim
reorient the policy in response to each new piece ‘v IIK:\ y
lead to a weak, confused, vacillating, and ultimatev- № y
The wide dissemination of the intelligence prod A - y y
bound to contradict the assumptions on which the ^ e 1'
reflect poorly on its progress, can look like a need iy11^ 1 . ‘
charge of implementing the policy. In the earthy : '
ident Lyndon Johnson, the problem is as follows: ^ .
milking a fat cow. You see the milk coming out, '
milk bubbles and flows, and just as the bucket is f y c
whips the bucket and all is spilled. That’s what 1
making.” 10

“Imperial Intelligence” p^ir

While the intelligence analyst may often feel that tlV1; ‘
ment to a policy leads him to ignore or to try to d is ^ •
information, policymakers may feel that although ^ ^
°nly tenuous judgments, policymakers are expected
as gospel and to make decisions accordingly. In ni^aIt‘,;J',
not clear why the intelligence judgment should be ^ 1
tative or objective than anyone else’s, including t h ^ J
For example, often an intelligence judgment d ea ly r'
1Ssues is based not on a sensitive nugget of hard data t!!
0 an agent inside the adversary’s cabinet or general y;;\
available to the policymaker (or to a journalist, for; ^
• *s Case» the policymaker may feel justified in d^J^ ii^i
gment rather than that of an anonymous (to the p y
and critlcs are ^kely to assert that he is ji*
^legitimately as Cleopatra. What may increase *
which it is based. (Except for the highest officials, policymakers may
learn very much about the sources for the written materials they receive
The policymaker may come to believe that this lack of information about
sources masks the fact that the main intelligence judgments are just specu
lative judgments that are not based on hard data.
In any case, the possibility that the policymaker is right— that his jud?
ment will be superior to that of the intelligence analyst— is not so small It
is true that, in the United States at least, the policy-making community is
often characterized by a high turnover rate among top officials, giving rise
to a certain amateurism. It may well be that, as an article by a U.S. intel­
ligence official put it, “ after a few months on the job [new intelligence
analysts] are among the most knowledgeable people in the government on a
particular issue, . . . [F]or the first time in their lives, they are writing for
an audience that knows less than they d o.“ 11
However, this is not necessarily the case, and policymakers who do have
some expertise in the area would likely dismiss an intelligence report that
reflects this view of their abilities. Furthermore, the “ policymaker may have
direct access to information unavailable to the analyst (confidential conver­
sations with foreign officials, for example) and, in any case, knows U.S.
policy— often a key piece of the puzzle— far better than the analyst.“ 12
Thus, in the absence of particularly secret information, or of a specialized
method of analysis, the intelligence analyst’s judgment often does not have
any special character that entitles it to be accepted over the judgment of
anyone else. As Charles Fairbanks has pointed out with respect to a con­
troversy over Soviet intentions toward Iran in 1985 (whether there was a
Soviet threat to Iran was a major question affecting U.S. policy toward that
country):
As for the means of analysis the Intelligence Community uses, these are
not anything arcane: the means is thinking, of the same kind that an
official outside the Intelligence Community or a reporter or a citizen
would use in trying to interpret similar facts. The composition of a Special
National Intelligence Estimate on [such] an issue . . . is essentially the
same exercise as the one the members of the [National Security Council]
staff engaged in, well or badly, when they decided to sell arms to Iran. It
is an exercise in policy reasoning.13
It is only when the result of such a reasoning process receives the label of
“ intelligence“ (and a classification stamp) that it becomes possible to Prc<
tend that what is involved “ is some hard nugget of fact that transcends the
UUa neutral test to evaluate them.” 14
aJ>This same situation may arise even when there is a hard fact available to
intelligence analyst, because very rarely will this fact speak for itself. It
' st be interpreted in some context, and this context is supplied, first of all,
hv the intelligence analyst. Yet understanding the context requires making
the speculative judgments I have been discussing.
These arguments go against the mainstream of current opinion on this
topic, which emphasizes the importance of maintaining the independence of
intelligence from policy and of forcing policy to listen to it. The consonance
of a policy with the intelligence view is often taken as a measure of its
reasonableness. What is lost is any sense of the solidity of the intelligence
view itself— is it based on incontrovertible fact or quite controvertible
speculation?
This point may be seen with respect to the circumstances surrounding the
Fairbanks article excerpted above. The conclusion of the SN1E to which he
refers— that the likelihood of a Soviet invasion of Iran was relatively small—
had been leaked to the press evidently to criticize “ the secret sale of U.S.
arms to Iran, which President Reagan ordered in January 1986 partly to
assist Iran against ‘intervention by the Soviet Union.’ ” 15
This motivation may also be seen in the article in which the leak appeared
that claims “ Casey’s amended analysis appears to have called into question
a primary White House rationale.” 16 In other words, the intelligence judg­
ment is used as the touchstone of the reasonableness of a policy, regardless
of the “ hardness” of the evidence on which it is based.

The Independence of Intelligence

Given these potential sources of friction in the intelligence-policy relation,
the question of how independent of policy intelligence should be is bound to
be a complicated one. In a military command, for instance, the chief intel­
ligence officer is as much a part of the commander’s staff (and hence as
much under his command) as any other aide; there is no question of any
independence. Indeed, the very meaning of the demand that an intelligence
agency be “ independent” is not clear. In principle, nobody would advocate
an intelligence service be independent from the head of government. Nev­
ertheless, as the above discussion indicates, people often want to use intel-
'fn nCe aS a neutra' arbiter of policy fights— not only among government
ctals but between government officials and their legislative and extragov-
menta! critics. This implies independence from all government officials,
 legislative body (as a representative of the public) or the public itself
As previously discussed, the main argument for independence derives
from the “ killing the messenger” syndrome: intelligence must be indepen
dent to enable analysts to tell the truth as they see it. In this context
independence means the ability to shield individual analysts from any pres'
sures or threats that might induce them to make their conclusions more
palatable to the product's consumer. To some extent, this sort of indepen­
dence can be achieved by the creation of a central intelligence agency, one
that is subordinate to the head of government, but not to the major intelli­
gence consumers such as the foreign affairs or defense ministries or depart­
ments. However, heads of government are themselves intelligence
consumers who may have invested a great deal of personal prestige in a
policy. In this case, the problem reappears at a higher level. Ultimately, as
long as the intelligence service is a part of the government, the only safe­
guards are the backbones of the chiefs of the intelligence services and their
willingness and ability to protect analysts from outside pressure.
Independence is particularly important when it is necessary to abandon
policies that are not working or that changed circumstances have rendered
obsolete or counterproductive. For this to take place, some mechanism must
ensure that all relevant information, positive and negative, is available for
reviewing the policy. As the Johnson quote suggests, policymakers must
devote most of their energy to implementing policies with some consistency
in the face of constant crosscurrents, doubts, political attacks— not neces­
sarily motivated by a sincere belief in the superiority of some other policy—
and so forth. Under these conditions, it is particularly difficult for such
officials to change gears and conduct a fully open-minded policy review.
In this situation, intelligence takes on an adversarial posture, since its
most important function will be to point out areas in which the policy is not
working or has become inappropriate. It must counteract, in other words,
the political pressure on bureaucrats to be good team players and cheerlead­
ers for current policies. Therefore, the intelligence service presenting infor­
mation that suggests the policy is obsolete or wrong-headed must have
enough independence to withstand the pressures such a situation generates.
One such situation that has attracted public attention is the intelligence role
in verifying compliance with arms control agreements. An accusation that a
party to an arms control agreement has violated it calls into question the wis
dom of having made the agreement in the first place or, at any rate, of con­
tinuing to adhere to it. The desire to maintain the agreement (or to maintain
or improve relations generally with the party in question) provides a strong
incentive to ignore or hide the disturbing evidence. In any case, the difficulty
 f f o r c i n g U 1C P « * V w ---------r v ..................... - - t o * — ----------------------------- « — ■
V
01 j0|ation is a strong inducement for denying that one exists.
3 VIn addition, an intelligence service must be sufficiently independent to
the initiative in looking at issues or areas of the world without policy
mmunity support. An intelligence service should scan the horizon, as it
C° e for potential issues and problems in areas that are not attracting the
Volicyrnakers’ attention. This warning function inevitably involves giving
the policymakers information that they have not requested and do not par­
ticularly want because they do not know its value.
Independence carries some disadvantages as well. Perhaps the greatest is
the risk that the intelligence work will become, or will be perceived as,
irrelevant to the policy process. This can occur because the intelligence
analyst does not know what problems appear most important to the policy­
makers or what options are being considered for dealing with them.
More than mere organizational or physical distance from the policymak­
ers leads to this problem. Intelligence analysts occasionally refuse to coop­
erate in implementing a policy they do not support. Former Deputy Director
of Central Intelligence Robert Gates provides an example of this problem:
When Secretary of State Alexander Haig asserted that the Soviets were
behind international terrorism, intelligence analysts initially set out, not to
address the issue in all its aspects, but rather to prove the secretary
wrong— to prove simply that the Soviets do not orchestrate all interna­
tional terrorism. But in so doing they went too far themselves and failed
in early drafts to describe extensive and well-documented indirect Soviet
support for terrorist groups and their sponsors.17
In other words, a policy to address the problem of international terrorism
that would have pressured the Soviet Union to reduce its support (whether
direct or indirect) of terrorist groups was thwarted by the unwillingness of
the intelligence analysts to cooperate, based on their view that such Soviet
involvement was a relatively small part of terrorism. Their view may have
been true, but that does not imply that the policy was incorrect: it might be,
tor sam p le, that the major psychological, sociological, or even political
causes of terrorism simply cannot be addressed by U.S. policy, or that
addressing them would conflict with other foreign policy goals.
if intelligence concentrates too much on its (admittedly necessary) ad-
Versarial role (such as would be involved in reviewing ongoing policies), it
makes it ap the more difficult for it to support the actual implementation of
£ lcy* In a supportive role, intelligence must concentrate its efforts on
ln2 and analyzing information relevant to implementing the policy. In
e above example, it might have meant discovering the mechanisms by
which an international terrorist group receives financial anu logistic supp0rt
from its patrons. In supporting policy implementation, the agency should be
willing not to insist on its view— that patron state support is not a major
cause of international terrorism— of why the policy is misguided or wrong

Intelligence and Democracy

Democracy and Secrecy
Secrecy, as already noted, raises a potential problem for the control of
intelligence agencies by their nonintelligence superiors. The legitimate need
to keep many details of their operations within the smallest possible circle
may facilitate a cover-up of unauthorized actions, thereby preventing higher
nonintelligence authorities from finding out about them.
In a democracy, however, secrecy may pose an additional problem: it has
the potential to call into question the political legitimacy, as opposed to the
actual control, of an intelligence service. If democracy is government not
only for the people, but of and by them as well, it is not surprising that
institutions that rely so heavily on secrecy can easily become the objects of
popular mistrust. This suspicion is not limited to intelligence services: Pres­
ident Woodrow Wilson’s promise that the international issues of the post-
World War I period would be resolved by “ open covenants, openly arrived
at” reflected public suspicion of prewar secret diplomacy as well as of
secrecy generally.

Legislative and Public Oversight

This skepticism concerning governmental secrecy, which is probably en­
demic in democratic societies, has led, in the United States and to a much
lesser extent in other democratic countries, to calls for oversight of intelli­
gence agencies by a part of the government that enjoys some independence
from the head of state. In the United States, a system of congressional
oversight by means of two committees created for that purpose (the House
of Representatives Permanent Select Committee on Intelligence and the
Senate Select Committee on Intelligence) was established in the mid-1970 s.
The cornerstone of this oversight system is the committees’ statutory right
to be “ fully and currently informed of all intelligence activities" by the
director of central intelligence and the heads of other intelligence entities.
This right to be informed is particularly significant in light of the congres­
sional power of the purse— its power to appropriate, or refuse to appropri­
ate, funds for any governmental activity. The result, according to former
Deputy Director of Central Intelligence Robert Gates, has been that
 overeigui — ••• ~—o— --r----- ^
. nce committees— far greater knowledge of and influence over the way
the CIA and other intelligence agencies spend their money than anyone in
the executive branch would dream of exercising, from expenditures in the
billions of dollars to line items in the thousands. . . . Congress may
actually have more influence today over the CIA’s priorities and its al­
location of resources than the executive branch.19

The congressional oversight system may be viewed as a compromise

between the requirement for secrecy and the desire to bring public opinion
to bear on the intelligence agencies to make sure that their secret activities
neither use means nor seek ends public opinion would condemn. Congres­
sional oversight allows the intelligence committees to serve as a sounding
board, a surrogate for the much wider debate (intra-executive branch, con­
gressional, and public) that might otherwise accompany a governmental
policy or initiative.
The ambiguities created by using the intelligence committees as a surro­
gate for public opinion are most readily seen regarding covert action, con­
cerns about which led to the introduction of congressional oversight in the
first place. The legal basis of the committees’ role is the requirement that
they be notified each time the president orders a covert action. While in
principle the notification requirement does not imply that the committees
possess a veto power over covert action programs, strong objections from
committee members can cause revision or even cancellation of a proposal,
if the political cost looks like it will be greater than the program’s expected
results.20 Thus, one purpose of prior notification might be described as to
test a proposal against a (somewhat restricted) cross section of political
opinion, represented by individuals independent of the president, to see if it
is out of line with fundamental beliefs or values.
Beyond this sounding board function, spreading the knowledge of covert
action to members of Congress creates the opportunity for bringing into play
the congressional power of the purse. This power can be exercised by the
intelligence committees themselves, or it can involve full-scale debate on
the floor of the Senate and House of Representatives, as in the case of the
various Boland amendments restricting or prohibiting aid to the Nicaraguan
resistance in the 1980s.
The fundamental claim on behalf of congressional oversight of covert
action is that it succeeds in combining the advantages of checks and balances
° n the one hand with those of secrecy on the other. Committee consideration
Serves as a surrogate for the full-scale public debate and democratic decision-
t^aking process that is of course incompatible with sccrccy.
retical terms. Practically, the issue is whether congressional oversight does
not inevitably lead to public revelation of the information, regardless of
what the rules say. In fact, there is little hard information about the source
of the numerous leaks that occur with respect to covert action, and inves
tigations of particular leaks almost never produce results. Given the impor­
tance journalists place on not revealing their sources, this is not surprising 21
Furthermore, even the information available may be misleading. A ref­
erence in a news report to a ‘‘congressional source” or an “ administration
official” may not indicate who first tipped the journalist off; the original
source may have spoken on “ deep background,” precluding any reference
to him at all.22 Thus, the reference in the article could refer to a source who
confirmed the story, not to the one who originally leaked it.23
In any case, there is no doubt that congressional oversight broadens the
number of people who are given access to information about covert action.
In general, the risk of a leak varies with the number of people with access
to the information. Furthermore, the congressmen and their staff aides typ­
ically are less used to dealing with classified information and work in a more
political setting than the average intelligence official. The requirement to
brief Congress probably leads to the involvement of additional people within
the executive branch itself, since some of those who handle congressional
relations have to be informed as well. In most cases, enough executive
branch employees have access that the expansion due to congressional over­
sight probably does not significantly increase the risk of a leak; in special
cases, where access is strictly limited within the executive branch, the
increase would be relatively greater, as would the risk of a leak.
While much of the public debate about congressional oversight concerns
leaks, the more theoretical challenge to congressional oversight of covert
action deals with the propriety and effectiveness of Congress adopting secret
procedures as a routine matter.
Thus, before the Church Committee, Morton Halperin argued against the
concept as follows:

[Better forms of control] cannot succeed in curing the evils inherent in

having a covert capability. The only weapon that opponents of a Presi­
dential policy, inside or outside the executive branch, have is public
debate. If a policy can be debated openly, then Congress may be per'
suaded to constrain the President and public pressure may force a change
in policy. But if secrecy is accepted as the norm and as legitimate, then
the checks put on covert operations can easily be ignored.24
 V ‘ | committees can acnieve in quiei neginiauuu wuu me cAauuvc uianut
sl° t^reat to “ go public” remaining in the background), it does contain a
U rnel of truth: in major disagreements, such as “ covert” aid to the Nica-
resistance, the norm of secrecy was abandoned, and the issue re-
!% ed full-scale and public congressional debate.
C Thus, the notion of congressional oversight conducted secretly contains a
^„contradiction: one wants to obtain the benefits of legislative deliberation
S intelligence matters, but one rules out from the start the major method of
sUch deliberation— full and open public debate.
These issues of intelligence secrecy and oversight often become confused
with what is properly a separate issue: to what extent does the fact that a
nation is a democracy affect the kind of foreign policy it ought to adopt?
This is particularly true with respect to covert action, which is the element
of intelligence that in the United States has attracted the most criticism as
beina antidemocratic. However, much of the criticism directed against the
CIA for its covert action programs would have been more properly directed
against the foreign policy those programs served.
In the United States, the main complaint against covert action has been
that it interferes in another country’s internal affairs. It seems to be a sec­
ondary question whether that interference is secret or overt. The argument
is sometimes made that such activity is undemocratic in the sense that it is
carried on covertly because domestic public opposition to it would be over­
whelming if it were known. This situation, however, is not so clearly the
case as those who make this argument claim: for example, the various covert
actions directed against Fidel Castro, including the assassination attempts,
might have received public support at the time they were under way. When
the Church Committee complained that these actions went counter to Amer­
ican public opinion, it was judging the CIA’s past covert actions not by
contemporaneous public opinion, but by the post-Vietnam public opinion of
the mid-1970s.
Since the mid-1970s, when the congressional intelligence oversight com­
mittees were created in the United States, other democratic countries have
considered establishing some form of independent oversight of intelligence
activities. For example, in 1984, the Canadian Parliament created the Se-
cunty Intelligence Review Committee (SIRC) to exercise oversight of the
country’s domestic intelligence apparatus, which was at the same time sep-
arated trom the Royal Canadian Mounted Police and made into a new
Civilian agency, the Canadian Security Intelligence Service.
Unlike the American oversight committees, the SIRC is not a committee
of Parliament but is composed of privy councillors appointed by the prime
minister after consultation with the opposition party leaders. (The PriVy
Council is a group whose members, usually former cabinet members, have
been officially recognized by the Canadian government as senior advisers )
The S1RC reports to Parliament annually, but itself has no legislative or
budgetary powers. The differences between the Canadian and U S. over­
sight mechanisms reflect the differences in the two political systems: the
main tension within the U.S. system is between the executive and legislative
branches of the government, which is not the case in Canada’s parliamentary
system with strong party discipline. Instead, in Canada, the necessary in­
dependence from the executive branch is obtained by relying on eminent
senior statesmen.25

Democracy, Counterintelligence, and Domestic

Intelligence
Rather than covert action, counterintelligence is the element of intelligence
most fundamentally affected by the democratic nature of the regime it serves;
the greatest number of difficulties in the relationship between democracy
and intelligence concern counterintelligence. In particular, these difficulties
involve defining the circumstances under which a government agency may
legitimately conduct surveillance of a citizen and determining the limits, if
any, on the amount and kind of such surveillance.
In the United States, this discussion tends to be conducted in terms of
constitutional law. For example, questions arise as to how the prohibition,
contained in the Fourth Amendment to the U.S. Constitution, against un­
reasonable searches and seizures should be interpreted with respect to na­
tional security cases.26 One such question involved the legitimacy of
wiretaps (which are considered a form of search) for such cases. It was
resolved by the Foreign Intelligence Surveillance Act of 1978, which cre­
ated a special court to grant warrants permitting wiretaps in national security
cases.27 Other proposals, which did not become law, would have estab­
lished similar procedures for other forms of surveillance.28
This focus on the admissibility or inadmissibility of means (which will
not be discussed in detail here) has been accompanied by a neglect of the
more fundamental questions concerning the proper ends of counterintelli­
gence or domestic intelligence in a democracy.29 What kind of information,
in other words, is needed? What purpose is such information supposed to
serve? What kinds of threats give rise to domestic intelligence requirements.
These questions are very difficult to answer with respect to a democratic
state, as opposed to a totalitarian or authoritarian one. In the latter
states, any opposition to the current leadership of the state is, in principle,
of intelligence interest. By constantly monitoring dissent, the government
strives to be in a position to take whatever steps necessary, either through
the ordinary law enforcement system or outside of it, to maintain its power.
In a democratic state, on the other hand, mere opposition to the govern­
ment of the day is not, and should not be treated as, a threat to the country.
The country’s national security interests (as opposed to the government’s
political interests) are not threatened by opposition as such, and domestic
intelligence collection about it is not needed. The question is, then, What
forms of activities are potentially threatening to national security such that
intelligence should be collected about them?
In particular, where is the line to be drawn between activities that are
legitimately of intelligence interest and those that are not? Should intelli­
gence ever be interested in someone’s opinions, in and of themselves, or
only when they are accompanied by activities? Should the criterion be the
legality or illegality of the activity, or are there activities that, although
legal, the government nevertheless has a legitimate need to know about and
hence may keep under some form of surveillance?

The Criminal Standard

Because of the focus on domestic intelligence means (especially in terms of

their constitutionality or lack of it) rather than ends, these questions have not
been addressed explicitly in public debates on intelligence. However, an
answer of sorts was suggested in the mid-1970s in the United States (at the
time of the Church Committee’s intelligence investigations and those of its
counterpart in the House of Representatives, chaired by Representative Otis
Pike), when an attempt was made to apply what was called the “ criminal
standard” to domestic intelligence. In other words, domestic intelligence
investigations would be strictly limited to situations where a violation of the
law has occurred or is about to occur. The implication is that the proper
scope ot domestic intelligence is delimited by the law; an illegal action may
a proper subject of domestic intelligence, but a legal one cannot be.30
While this principle was never promulgated in so many words, it set the
*°ne for the guidelines in which the attorney general in 1976 first set down
^ the conditions under which the FBI could conduct investigations and
e techniques it could use in carrying them out. Its impact also can be seen
In Foreign Intelligence Surveillance Act of 1978, the sole piece of
c°ngressional legislation in this area in that period.31
Sidelines were established by then-Attorney General Edward Levi
specify the circumstances under which the FBI could use various surveil-
domestic intelligence was divided into two parts— domestic security on the
one hand and foreign intelligence and counterintelligence on the other—and
separate guidelines were established for investigations in each category
The guidelines governing foreign intelligence and counterintelligence in­
vestigations were issued on a classified basis and have never been made
public in their totality.32 It is therefore difficult to know to what extent they
embody the criminal standard with respect to the investigations they govern
The vast majority of the investigations conducted under these guidelines
would deal with detecting and preventing espionage. Thus, they focus on
potential criminal activity. The same holds for the detection and prevention
of terrorist activities carried out under foreign direction, which is probably
the FBI’s second most important task to be accomplished in accordance with
these guidelines.
It is not clear to what extent the guidelines envisage the investigation of
legal political activities carried out on behalf of a foreign government, that
is, cooperation with the covert action or active measures conducted by a
foreign government and targeted against the United States. It is true that,
according to the guidelines,
the FBI is, under standards and procedures authorized in these guide­
lines, authorized to detect and prevent espionage, sabotage and other
clandestine intelligence activities, by or pursuant to the direction of for­
eign powers through such lawful foreign counterintelligence operations
within the United States and its territories, including electronic surveil­
lance, as are necessary or useful for such purposes.3'
This implies that generally legal activities such as political propaganda,
organizational work, and fund-raising are subject to surveillance when car­
ried out pursuant to the direction of a foreign power. Furthermore, theoret­
ically at least, anyone engaged in such activity on behalf of a foreign
government is obliged to register as a foreign agent under the Foreign Agent
Registration Act (FARA), and failure to register would itself constitute a
crime. 3 4
On the other hand, failure to register is rarely used as the sole basis for
a foreign counterintelligence investigation or for prosecution.35 One such
investigation that has come to light involved the Committee in Solidarity
with the People of El Salvador (CISPES), which was the subject of a
three-month FBI investigation in 1981. The investigation was conducted at
the request of the Criminal Division of the Department of Justice, t0
determine whether CISPES is required to register under” FARA as an agent
of the Salvadoran Frente Democratica Revolucionario, the political arm
 what extent the standards and procedures authorized in these guidelines
^ould permit any effective investigation of clandestine intelligence activities
that did not involve violence or the violation of any law (other than FAR A).
As opposed to the foreign counterintelligence guidelines, the domestic
security ones (called the Levi guidelines) have been made public.37 They
deal primarily with investigations of groups once called subversive or un-
American. While these are vague terms, they generally include groups (1)
that are hostile not merely to the government of the day and its policies but
to the constitutional structure and its fundamental principles, (2) that seek to
deprive some class of persons (such as a racial, ethnic, or religious group)
of their civil rights, or (3) more generally, that seek to bring about political
change by violent means.
The key point in the Levi guidelines is that investigation of such groups
is permitted only when a group (or an individual) is or may be engaged in
activities “ which involve or will involve the use of force or violence and
which involve or will involve the violation of federal law, . . . ” 38 This is
the essence of the criminal standard that effectively defines the govern­
ment’s interest in the domestic security area.
The most important motive for adopting this standard has been to ensure
that no surveillance takes place just because an individual or group is ex­
ercising its rights of freedom of speech, of the press, or of association, as
protected by the First Amendment to the U.S. Constitution.39 Two general
lines of argument support this goal. First, there is the view that government
investigation of an individual or a group is itself a kind of punishment from
which those who have done nothing illegal should be exempt. The second
argument holds that this type of investigation, either because of its intrinsic
unpleasantness for the target or because it seems to threaten future punish­
ment, inevitably “ chills” the activity (the exercise of the right) that forms
the basis of the surveillance. These reasons have some plausibility but are
not as compelling as is sometimes claimed. In constitutional terms, the
questions would be whether surveillance under these conditions would
amount to an abridgement of the freedoms of the First Amendment.
This issue cannot be dealt with here at length. A few points, however,
illustrate the cases for and against the criminal standard. On the one hand,
le violation of privacy involved in the notion that the government is keep-
*n£ tabs on an individual is indeed distasteful. This is true of all law en­
forcement investigations that target individuals who have not been convicted
any crime and who, indeed, may well not be charged with any. An
lnc°me tax audit, a customs inspection, and any number of other law
°rcement actions are invasions of privacy, which the government may
 * --------------- — fc“~ has
committed a crime.
Similarly, the possibility of government surveillance might scare people
away from controversial political activity, because they fear the inclusion of
their names in a file might lead to unpleasant consequences or for some other
reason. Nevertheless, the chilling effect of a secret investigation that leads to
no government activity can only be minimal, if anything. The effect seems
more equivalent to the social obloquy one might earn by espousing unpop­
ular opinions than to an abridgement of First Amendment freedoms. As a
matter of constitutional law, the Supreme Court has rejected the contention
that an individual suffered any harm from “ the mere existence, without
more, of a governmental investigative and data-gathering activity. . .
ruled that, as far as achieving the legal status ( “ standing“ ) to contest a
governmental investigation in court is concerned, “ allegations of a subjec­
tive ‘chill’ [resulting from the secret government investigation] are not an
adequate substitute for a claim of specific present objective harm or a threat
of specific future harm.” 40
As has been noted by Kenneth Robertson, this issue is typically discussed
in terms of constitutional law prohibitions against certain types of govern­
mental activity. The question also must be faced of whether these standards
allow the government to meet its legitimate needs for information about
what is going on in its own society.41 In at least two areas of interest an
argument can be made that, theoretically at least, these standards do not
allow the government enough leeway.

The Crim inal Standard and Personnel Security

The first area has to do with the government’s personnel security program
for screening applicants for government positions and current employees
before giving them access to classified information. For such purposes, it is
legitimate for the government to want to know if the applicants have any
affiliations or loyalties that might lead them to disclose classified informa­
tion to representatives of foreign powers or otherwise act against the coun­
try’s interests. In general, it is difficult to determine this information during
the screening, for all the reasons noted in chapter 5. The only way the
government can have some confidence that the applicant is disclosing any
such affiliation is if the government possesses other sources of information,
derived from surveillance of the organizations in question.
In this regard, it is important to understand how far the First Amendment s
protection extends. According to the Supreme Court’s decision in the case o
 n d e H O l U K v. i / / M u , w i m i m a ) ut; c d i i c u a u M i a n a u v u c a c y Ul U1C u v c i UUUW

f the U.S. government is protected. Advocacy can be forbidden only when

^‘such advocacy is directed to inciting or producing imminent lawless action
d is likely to incite or produce such action.”42 Thus, under the criminal
standard, an organization that openly called for the violent overthrow of the
U S government could not be subject to surveillance unless and until its ad­
vocacy passed from the abstract to the concrete, as defined above.
Furthermore, membership in such an organization can be considered crim­
inal only if it is active rather than nominal membership and involves “ knowl­
edge of the [organization’s] illegal advocacy and a specific intent to bring
about violent overthrow [of the U.S. government] ‘as speedily as circum­
stances would permit.’ ” 43 This implies that, under a criminal standard, it
would be impossible to conduct surveillance of such an organization for the
purpose of compiling a complete list of its members, since there would not
be evidence that each member met the standard for criminal membership,
even if the organization as a whole met the Brandenburg test.
To some extent, this was an intended result of the adoption of the criminal
standard. As John Elliff has noted, in describing the initial policy decisions
taken by Attorney General Edward Levi:
Levi rejected the federal employee security program as a basis for FBI
intelligence investigations. As Levi saw it, the Executive Order |on the
personnel security program] authorized the Bureau only to investigate
Executive Branch employees and applicants. It could not supply a basis
for the FBI’s authority to conduct domestic intelligence investigations in
general.44

In other words, investigations cannot start with a group’s public words or

actions. They have to start from the individual applicant for government
employment, whose affiliation with the group might be well hidden.
Without surveillance of such groups for the purpose of compiling a list of
their members, the federal government cannot, in principle, guard against
granting access to confidential information to, or placing in sensitive posi­
tions, individuals who are members of a political party loyal to a hostile
foreign power, of various cult groups fanatically devoted to a leader whose
ultimate aims may not be compatible with constitutional government, or of
ale groups that want to expel part of the American people from the body
Politic. This may or may not pose a problem at any given time, depending
on whether such groups exist and whether they are sophisticated enough to
Jelop a strategy of placing their members in sensitive positions in the
e eral government (penetrating it.)
The Criminal Standard and Counterterrorism

A somewhat different issue arises in the case of terrorist groups. While an

active terrorist group would meet any conceivable criminal standard for
surveillance, it can be extremely difficult to establish such surveillance if the
group is as sophisticated as the major terrorist groups of the past decades in
the Middle East and Western Europe. Given the rarity of known fixed
facilities and means of communication, which might be covered by technical
collection means, the most fruitful investigative technique, and often the
only useful one, is to penetrate the group with one’s own agent.
However, a sophisticated group will be suspicious of outsiders and will
take into its confidence only individuals with whom current members have
long personal associations or who have proved their loyalty to the group by
committing serious crimes. Absent a lucky break, it is difficult or even
impossible for an intelligence service, starting from scratch, to penetrate a
terrorist group once it is organized and in operation.
Intelligence agencies have to fall back on other strategies to deal with this
problem. First, they must be able to collect information about support
groups, which are groups of sympathizers who provide various kinds of
support— financial, logistic, political, and legal support— but do not take
part in the actual terrorist activities. They, of course, present themselves to
the public as solely political organizations engaged in legitimate political or
charitable activities.
Penetrating a support group, while much easier to accomplish, is not as
valuable as penetrating the terrorist group itself. Nevertheless, it can be very
useful: the kind of aid provided may make it possible to deduce the timing and
nature of future terrorist acts. Furthermore, a support group member, through
personal ties or actual recruitment by the terrorist group, may work his way
into a position to provide much more precise information. Finally, an agent
within a support group might be able to foil a terrorist act, either by providing
the police with a key piece of information (such as the description of the rental
car to be used in a terrorist act) or by directly sabotaging a bomb (for example,
by replacing the gunpowder with sawdust) or other equipment.
As explained by Shlomo Gazit, former director of Israeli Military
Intelligence,
Very few [terrorist] organizations can operate in a complete or full com
partmentalization and do not depend on networks of local supporters.
Such supporters help the terrorist organization, either because of ideo
logical motivation or through fear and blackmail, without being direct y
involved in terrorist operations. The importance of penetrating the sym
pathizers’ or supporters’ system lies in the fact that it is easier to penetra
 it than the more highly closed terrorist organizations. By penetrating this
supportive system it may be possible to penetrate the organization itself or
obtain indirect information about it.45

The criminal standard inhibits the penetration of support groups since they
typically present themselves to the public as political or charitable groups.
One group may describe itself as being engaged in political propaganda in
favor of various causes, which just happen to be the same as those espoused
by the terrorist group. Similarly, another will characterize its activity as the
charitable support of the wives and children of alleged terrorists who have
been imprisoned or killed. A third group may claim to limit its activities to
providing lawyers and funds for the accused terrorists’ legal defense.
Without getting an informer into the support group, it may not be possible
to establish the exact nature of the group’s additional (and not so innocent)
efforts on behalf of the terrorist group. But the use of this or other intrusive
techniques is permitted only if evidence is already available that the group
is involved in violent, criminal activities.46 In this respect, the criminal
standard is something of a catch-22: one cannot know about the support
group’s additional activities because one may not look, and one may not
look as long as one does not know.
A second way of overcoming the difficulty of conducting surveillance
against terrorist groups relies on the fact that terrorist groups are typically
offshoots of politically extremist, nationalist, or separatist groups; they are
usually composed of members of such groups who become frustrated with
the failure of less violent means to accomplish the groups’ goals. By having
informers in the groups from which terrorist groups are likely to form, one
would know about a terrorist group as soon as it formed. Having agents in
the original group would increase the possibility that a penetration agent
could be placed in the terrorist group from its beginning. A criminal stan­
dard, on the other hand, makes it impossible to maintain the initial surveil­
lance that could provide the first indication of a new terrorist group.
A particular circumstance that comes under the rubric of counterterrorism
involves the responsibility of the U.S. Secret Service to protect the presi­
dent, other senior government officials, and foreign dignitaries. As with
counterterrorism in general, the goal here is to prevent the criminal act from
taking place, rather than to react to it by punishing the perpetrator to deter
Ure attacks. The Secret Service’s responsibility requires advance intelli-
£ nce on assassination attempts against its protectees. A criminal standard
investigating groups that might engage in assassination may mean the
Cret Service does not receive warnings about assassination attempts until
er the group has engaged in illegal violence, which might be too late.
 Ill 1 ^ 0 Z., H I C 1 IC ilM iiy U C JJdl llU U ii u i u t m i it^ p u iio iu iv iv i u iv o w tici

vice explained the problem this way:

Today, the [Secret S]ervice has to be . . . concerned about the terror
ist. . . . The Secret Service needs to know . . . about the intentions and
activities of terrorists or other extremist groups. Further, we need to know
about this kind of threat before it materializes, not after.47
This official also noted that a reasonable allocation of the Secret Service’s
resources would require that it “ know, for example, that an organization
using extreme rhetoric is either disinclined to put words into action or
incapable of launching dangerous activities.” In his view, the current
problem is that the FBI is n o t. . . able to collect enough intelligence. This
partly because [o f]. . . Attorney General Levi’s Domestic Security Guide­
lines. . . . Prior to the Levi Guidelines, the Secret Service received from the
FBI a vast amount of intelligence information on individuals and poten­
tially violent groups. . . . Although it has been claimed that this resulted
in the Secret Service being overinformed— the service obviously had to sift
through an awful lot of information and sort out the wheat from the chaff
. . . the situation was ideal as far as the Secret Service was concerned.
The problem now is that investigations in important areas are simply not
being pursued. The result is a lack of information about the location,
structure, plans, and activities of groups that may not be labelled terrorist
as such, but that certainly are radical, . . . that support terrorist causes
and are potentially terroristic or violent.
Of course, this general point would apply to all law enforcement; in
general, very little law enforcement activity is devoted to preventing crime
(except by deterrence) as opposed to detecting it afterward and apprehend­
ing the perpetrators. Certainly, it would be impossible for ordinary law
enforcement organizations to set themselves the goal of preventing crime by
establishing an intelligence network that would warn them of planned crim­
inal activity. In counterterrorism and executive protection, however, be­
cause of the potential seriousness of a single incident, there is a strong desire
to prevent crime and not merely to punish it; this implies a requirement to
collect intelligence about individuals or groups who, although they have not
yet done so, are likely either to engage in such activities or to maintain close
ties with those that do.

The Case fo r the Criminal Standard

Taking note of the theoretical argument against the criminal standard for
domestic security investigations is much easier than determining what could
replace u- ---------
^ umented many cases in which the FBI had conducted investigations that
0 esented both a waste of resources and unwarranted intrusion into polit-
. 1 aCtivities that posed no domestic security threat. This “ overbreadth/’ as
Church Committee termed it, argued for a relatively clear-cut standard
t w0Uld limit investigations to those undeniably necessary. Necessary
investigations were defined as being targeted against specific acts of criminal
violence that had already occurred or were about to occur.
The examples of this overbreadth in domestic security investigations fall
under various headings. Perhaps the most glaring were those cases in which,
at White House direction, the FBI undertook investigations against the
president’s partisan opponents, apparently for purely political purposes.
Judge and former Deputy Attorney General Laurence Silberman testified in
1978 that
of all the abuses of various agencies of Government over the last 20 years,
1 think the single most egregious abuse was President Johnson’s direction
to the FBI to see if they could find any dirt on [Senator Goldwater’s] staff
2 weeks before the [1964 presidential] election.48

There are other similar examples of the use of the FBI against political
opponents in Congress, the media, or elsewhere.
In many other cases, investigations were directed against activist political
groups targeted under vague labels, such as

• “ rightist” or “ extremist” groups in the “ anticommunist field,”

• “ anarchistic or revolutionary beliefs,”
• “ black nationalists” and “ extremists,”
• “ white supremacists,” and
• “ agitators.” 49

In these cases, while there might have been some nexus between the
organization being investigated and potential violence or links to foreign
intelligence services, clearly the investigations were indeed overbroad in
forms of the individuals and groups involved, the extraneous information
collected about them, and the time unpromising investigations were allowed
t0 continue. The extraneous information that was collected was not only
retained in investigatory files, but freely disseminated within the govem-
^ Us’ *nf°rmation about legitimate political activities was made avail-
a o for the administration’s partisan political use.50
rom the history' the Church Committee compiled, it seems clear that
excesses were facilitated by the absence of clear standards for domestic
intelligence investigations. Thus, it is easy to see how a criminal standard
would be an attractive, clear-cut option. If that limitation is too narrow, then
one must search for a broader one that still would protect against the sorts
of abuses that have occurred.
This task cannot be accomplished quickly. To approach it in a responsible
manner, one would have to start by considering what domestic political or
other activities a democratic government legitimately needs to know about
to protect against threats to the country. In the United States, the legislation
executive orders, and guidelines dealing with domestic intelligence have
been cast in terms of the conditions that must be met before various inves­
tigative techniques may be employed. There is, on the other hand, no formal
statement of what the purpose of domestic intelligence is, or of what type of
information it is supposed to provide. Such a statement would have to be
elaborated before one could develop guidelines that focus on the goal (what
information is necessary) rather than on the means (which investigative
technique may be used).
By contrast, Canadian law defines the threats with which the domestic
intelligence agency, the Canadian Security Intelligence Service (CSIS),
should be concerned. The CSIS is authorized to collect information regard­
ing activities “ that may on reasonable grounds be suspected of constituting
threats to the security of Canada.” These are defined as:

• espionage or sabotage, or activities directed toward or in support of

espionage or sabotage,
• foreign-influenced activities that are detrimental to Canadian interests
and are clandestine or deceptive or involve a threat to any person,
• activities directed toward or in support of the threat or use of serious
violence to achieve a political purpose, and
• activities directed toward or intended ultimately to lead to the destruc­
tion or violent overthrow of the Canadian government.51
Even with an adequate definition of the scope of domestic intelligence,
questions will arise as to whether a given group constitutes a threat or is
simply engaging in hyperbolic rhetoric, and as to the amount and type o
evidence required for a group to be considered a threat that should be
investigated. But without an understanding of what domestic security infor
mation is required, it is impossible to judge whether specific restrictions on
investigative means are reasonable.
 7
TWO VIEWS OF
INTELLIGENCE

In the five preceding chapters I have discussed the four elements of

intelligence— the four general headings under which intelligence activities
may be categorized— as well as some of the issues concerning the relation­
ship between intelligence and the rest of the government and society of
which it is a part. I now return to the more general issue of the nature of
intelligence as a whole. I first compare what may be called the “ traditional”
view of intelligence, which emphasizes obtaining, protecting, and exploit­
ing secret information relevant to the struggle among nations, with a newer,
characteristically American view of intelligence that has evolved since
World War II.
do some extent, the “ traditional” view as developed here is my
invention— hence the quotation marks. While it reflects traditional concepts
°f intelligence, those concepts were not typically articulated in this manner.
(It should be remembered that public writings about intelligence were rare
in the past.) I use it here primarily as a foil to highlight the distinctive
f sPects of the American view of intelligence that has arisen since World
War II.
The “ traditional” view can be traced back to one of the oldest known
^ cmatic discussions of intelligence, which is found in a Chinese classic
entitled 7 he Art o f War, commonly attributed to a sixth century b .c . general
f nci Hulitary thinker named Sun Tzu. He explains the importance of intel-
118ence as follows:

159
 enemy whenever they move and their achievements surpass those of
ordinary men is foreknowledge.

What is called “ foreknowledge” cannot be elicited from spirits, nor from

the gods, nor by analogy with past events, nor from calculations. It must
be obtained from men who know the enemy situation [directly, i.e., men
with access to the enemy camp].1

Espionage is at the center of this understanding of intelligence. Given the

close connection between learning the enemy’s secrets and achieving victory
over him, it would follow that the protection of one’s own secrets (coun­
terintelligence) must be as important as the obtaining of the adversary’s. Sun
Tzu mentions two methods of achieving this:

• doubling enemy agents, presumably by making them a better offer, and

• sending into the enemy camp spies who have been deceived as to one’s
own situation, so that, once captured by the enemy and forced to talk,
they will unwittingly deceive him.2

The ultimate goal of this espionage is to elucidate the adversary’s strategy

so well that one can devise means of circumventing and defeating it at the
lowest possible cost. Indeed, from the point of view of Sun Tzu, for whom
defeating an enemy’s strategy is more important than defeating his army, the
intelligence component of international struggle is as vital as the armed
component:

Thus, what is of supreme importance in war is to attack the enemy’s

strategy;
Next best is to disrupt his alliances;
The next best is to attack his army.3

Intelligence is as much a struggle with an enemy as is armed combat; the

difference lies in the means employed.
In an article that calls for the “ growth of a truly American intelligence
system,” former Director of Central Intelligence William Colby distin­
guishes such a system from traditional intelligence systems, which he de­
scribes and dismisses, as follows:

For centuries, intelligence was the small, private preserve of monarchs

and generals. Governmental and military espionage ferreted out the se
crets of other powers in order to provide its sponsors with advantage
their dealings. Secret agents intrigued and subverted in order to discre
 an —rr —- io<cj auvcib<mc^» wiuiin ms own camp. The
spy was the prototype of this traditional “ intelligence” discipline.4
glsewhere, Colby defines the “ traditional concept of intelligence” as a
“ secret service which ferrets out an enemy’s secret plan and shares it with
monarch so that he can win a battle.” 5 He claims that the American
experience during World War II saw the rise of a new view of intelligence
that challenged the “ traditional” one in major respects. In the postwar
oeriod, this view had continued to develop both within U.S. intelligence and
amonc academic students of the subject. This chapter discusses that new
view and compares it with the “ traditional” one.

Historical Development of the

American View
The unsettled conditions brought about by the beginning of World War II,
and America’s subsequent entry into it, led to massive increases in resources
the United States devoted to intelligence. It also led to a new bureaucratic
development, the creation of a central intelligence service (known first as
the Office of the Coordinator of Information and then as the Office of
Strategic Services, or OSS) whose purposes included the collection, anal­
ysis, and correlation of national security information and data.6 Despite
President Franklin Roosevelt’s decision in July 1941 to create a central
capability to correlate information, none was in place by December 1941,
when the failure to foresee the Japanese attack on Pearl Harbor highlighted
the lack of intelligence correlation and analysis, since many relevant bits and
pieces of information were available.

The Centrality of Analysis

The OSS set up a research and analysis branch that was to analyze “ all the
relevant information, that was overtly available as well as that secretly
obtained.” 7 Although this goal was not fulfilled during World War II, it
remained the ideal situation, which was essentially realized with the creation
°f the Central Intelligence Agency in 1947.8
This emphasis on intelligence analysis, the intellectual work of piecing
together disparate bits of information to develop an accurate picture, sug-
gests a view of intelligence different from the “ traditional” one. This newer
v,ew pays less attention to secrecy and the means of overcoming it and more
Mention to the analytic function, an activity that comes to resemble social
Science research more than it does traditional espionage.
 - - - •' v/i ivj
War II by Sherman Kent, former OSS officer and future head of the OffiCe
of National Estimates, the Central Intelligency Agency office then in charge
of producing overall assessments of the world situation. His well-known
book, Strategic Intelligence fo r American World Policy, argued the case for
understanding intelligence as the scientific method (in its social science
variant) applied to strategic matters:

Research is the only process which we of the liberal tradition are willing
to admit is capable of giving us the truth, or a closer approximation to
truth, than we now enjoy. . . . we insist, and have insisted for genera­
tions, that truth is to be approached, if not attained, through research
guided by a systematic method. In the social sciences which very largely
constitute the subject matter of strategic intelligence, there is such a
method. It is much like the method of physical sciences. It is not the same
method but it is a method none the less.9

Of course, this social science method cannot be the whole of intelligence,

even if it is its heart. There must, after all, be data upon which it can work.
While recognizing that some of the data must be gathered by clandestine
means, Kent insists that the popular view of intelligence as dealing primarily
with secrets is a myth. Instead, in his view, intelligence analysis can rely on
nonsecret, openly available information, with only a small number of clan­
destinely gathered secrets added as seasoning. Referring to the knowledge
necessary to conduct foreign policy— the content of the strategic intelligence
about which he is writing— he says: “ Some of this knowledge may be
acquired through clandestine means, but the bulk o f it must be had through
unromantic open-and-above-board observation and research. “ 10
In its most extreme form, this new view equates intelligence with a sort
of universal, predictive social science. This social science was described by
former DCI Colby in 1981, as follows:
A new discipline specifically designed for intelligence analysis must be
refined, and the process of research and development has already begun.
It will step beyond academic analysis through new techniques to project
future probabilities rather than explain the past. Experiments in this new
discipline are by no means limited to the official intelligence community»
as they also take place in information science research centers, among
political risk analysts, and in the projections of the Club ot Rome, the
Global 2000 study, and others.11
The central importance of analysis is a major distinguishing characteristic
of the new view of intelligence. To understand this view better, I examine
 iiiit m uov/ i v w iiv iu c > c a
li auinuiidiiM U 1 lib
tne
truth- I’ll start by observing that, whatever one’s view of intelligence, one
must recognize an important role for at least some kinds of analysis.
First of all, collection activity immediately generates the need for some
analysis» either technical (such as decoding coded messages) or more gen­
eral (such as comparing reports from two different sources to determine
whether they support or contradict one another). Some evaluation of the
information received is needed to determine, for example, whether it should
be accepted as genuine or whether it should be regarded as a deliberate
attempt by the adversary to mislead.12
More significantly, it often may be too difficult to ferret out the relevant
secret information directly. As a result, the only alternative is to try to
deduce it from whatever data is available. This could include information
that the adversary does not consider, in itself, to be sensitive and so does not
keep secret.
For example, the most direct and reliable warning of imminent attack
would probably come from an agent within the enemy’s General Staff.
Failing this, one would have to attempt to deduce the enemy’s intention
from various pieces of information concerning the deployment of his armed
forces, their readiness, the status of his reserves, and so forth. In addition,
public information might be useful as well; before Pearl Harbor, for exam­
ple, the United States paid attention to the locations and planned itineraries
of Japanese merchant ships on the grounds that, were Japan planning to go
to war, it would make sure that its merchant fleet had been recalled to
relatively safe waters.
Thus far, those holding the 4‘traditional” view would not argue with the
necessity of such activities, despite their view that intelligence deals pri­
marily with secrets. The questions come as the sphere of intelligence con­
cerns expands to the dimensions of Colby’s “ new discipline . . . [of]
intelligence analysis,” a universal, predictive social science covering all
aspects of society.
Obviously, to manage military affairs and to conduct foreign policy, a
country’s governmental officials must know more about its potential adver­
saries than merely their military or diplomatic secrets. Other factors, such as
the potential adversary’s economic activity and potential, its demographic
trends, and its internal political forces and concerns, also must be consid­
ered. With respect to these factors, the difficulty is likely to lie not so much
ln gathering the raw data (although in the case of a closed society with an
lchosyncratic political system, this may present many challenges) as in eval­
e tin g the information correctly.
F°r example, despite the openness of the American political process, an
observer, whether American or not, has difficulty understanding, let alone
predicting, the course of a major political event such as a congressional vote
or a presidential campaign. Yet, clearly, to conduct its policy toward the
United States, another country would find that a good understanding of such
matters could be as important as obtaining U.S. national security secrets. As
Eliot Cohen has pointed out, the openness of American political life in the
1930s did not by itself enable its future opponents to understand it:

Although the United States . . . was nothing if not accessible to foreign

agents, both legal and covert, it was in another sense impenetrable to the
Axis powers because of their own failure to comprehend the workings of
democratic states. Both the Germans and the Japanese repeatedly under­
estimated the American polity— its tenacity and ingenuity, as well as its
ability to organize, improvise, and produce. And these colossal failures of
intelligence helped doom both states to ruinous defeat.13

In other cases, the bit of information most important for a nation to have
cannot be considered a secret because, when the information is most needed,
it does not even yet exist. Sherman Kent gives as an example of this the
intelligence problem the Soviet Union faced when it sought to determine
whether the United States would resist the 1950 North Korean invasion of
the South. Had Soviet agents been looking in U.S. files for documents
containing the U.S. decision whether to fight, they could not have found
any, since the decision was made by President Truman only after the inva­
sion began: “ Thus, if knowledge of the other man’s intentions is to be
divined through the reading of his intimate papers and one’s own policy is
to be set on the basis of what one discovers, here is a case where policy was
on the rocks almost by definition.” 14 Instead, Kent argues that such infor­
mation can only be had through research and analysis:
I have urged that if you have knowledge of Great Frusina’s [Kent’s
hypothetical great power] strategic stature [Kent’s term for the totality of
a nation’s capabilities— military, political and economic— to act on the
international scene], knowledge of her specific vulnerabilities, and how
she may view these, and knowledge of the stature and vulnerabilities of
other states party to the situation, you are in a fair way to be able to
predict her probable courses o f action.
To strengthen the reliability of your prediction you should possess two
additional packages of knowledge. First, you should know about the
courses of action which Great Frusina has followed in the past. •
Second: you should know, as closely as such things may be known, how
Great Frusinans are estimating their own stature in the situation.
 There can be no doubt that such information, and the thought processes
that lead to it, is a vital part of strategic thinking in general and high-level
foreign policy-making in particular. No policies can be adopted or imple­
mented without some view of the potential actions of other countries. Nev­
ertheless, it is not clear why it should be considered “ intelligence” and why
it should be located in the intelligence agencies. This issue, which involves
important questions concerning the relationship between intelligence and the
policymakers it serves, has already been discussed in the previous chapter.
For present purposes, it is sufficient to observe two connected conse­
quences of this view of intelligence, which emphasizes the centrality of
analysis. The first consequence is that a good deal of the thinking that goes
into policy analysis can be, and ought to be, carried out within the intelli­
gence agencies themselves. In intelligence agencies this work can then be
done according to some kind of scientific method. In other words, the social
sciences can, or should be able to, provide methods of approaching these
problems that deserve some of the respect and authority routinely granted to
science. As a result, in this view, a superior wisdom may be attributed to
intelligence results, even when these conclusions reflect not covertly ob­
tained and jealously guarded hard data, but speculation based on facts gen­
erally available to the policy community. In this view, it becomes possible
to use intelligence to “ grade,” or judge the correctness of, policy.
A second consequence is that intelligence analysis can be divorced from
the policy process and, indeed, be apolitical in nature. Intelligence is less a
specific aid to policy-making than it is a kind of living encyclopedia or
reference service whose information should be made available not only to
the policymakers’ political opponents in the legislature, but also to the
domestic public and the world at large.16 For example, Colby foresees an
era of free trade in intelligence, in which nations recognize the
mutual benefit from the free flow and exchange of information, in the
fashion that the SALT agreements recognize that both sides can benefit
from pledges against concealment and interference with the other’s na­
tional technical means of verification.17
However fantastic this may sound, it illustrates the power of the view that
mtelligence is, at bottom, an endeavor similar to social science, if not
efluivalent to it.

Espionage vs. Technical Intelligence

Collection
j^° matter how central to intelligence the analytic process is considered to
e ’ the raw data to be analyzed must come from somewhere. While it is true
that some of it can be found in open sources (publications, radio and tele
vision broadcasts, and so forth), the fact remains that some nations__jn
particular, the Soviet Union (even in the era of glasnost)— do not make
public even basic facts about their military forces and the economic re­
sources that support them. Thus, even if Kent’s view of the relative impor­
tance of open and secret sources were true for more ordinary states, it is
clearly not true of a state that conceals almost all the details of the compo­
sition of its military forces.
The Soviet Union’s particularly strong security measures made it very
difficult, in the post-World War II period, for the United States and other
Western countries to use espionage to obtain the missing basic data about
Soviet military forces. The problem was solved primarily by the develop­
ment of technical intelligence collection platforms that could pass above
Soviet territory— at first, the U-2 plane and then reconnaissance satellites.
These technical means not only leapfrogged over Soviet security precau­
tions, but appeared to be free of espionage’s other problems as well. The
whole problem of determining whether a human source was trustworthy
seemed to have been avoided by relying instead on machines, which could
not lie or be suborned. At the same time, access to a particular location
could be assured (although the ability to photograph it required the absence
of cloud cover). In this way, too, technical intelligence seemed more reli­
able than humint.
The result was an exaggeration of the importance of techint at the expense
of humint, a tendency that has been characteristic of the post-World War II
American view of intelligence. This made intelligence seem a more scien­
tific undertaking, not only because it required the latest and most sophisti­
cated technology, but also because it appeared that the uncertainty associated
with human agents could be overcome by technical collectors that provided
broad coverage of the adversary’s territory and enabled one to see what was
there for oneself.

The Depreciation o f Counterintelligence

Another characteristic of the new view of intelligence is its tendency to
depreciate the importance of counterintelligence in general, and deception
and counterdeception in particular. This is closely related to the notion that
intelligence is a variety of social science dealing with foreign countries.
Even though intelligence work obviously differs in some respects trom
social science, the two endeavors, according to this view, remain s i m i l a r -
For example, Kent notes that strategic intelligence is, in his terminology»
an “ extension,” in several senses, of the search for knowledge in genera •
One of these extensions is that “ difficult barriers,” which “ are put there on
by other nations,” often stand in the way. Nevertheless,

important as they are, these extensions, as I have called them, are external
to the heart of the matter: intelligence work remains the simple, natural
endeavor to get the sort of knowledge upon which a successful course of
action can be rested. And strategic intelligence, we might call the knowl­
edge upon which our nation’s foreign relations, in war and peace, must
rest. 18
For the "traditional” view, on the other hand, the fact that an adversary is
trying to keep vital information secret is the very essence of the matter; if an
adversary were not trying to hide his intentions, there would be no need for
complicated analyses of the situation in the first place.
These different stances toward the importance of secrecy reflect basic
differences with respect to what intelligence is. If, according to the “ tradi­
tional” view, intelligence is part of the real struggle with human adversar­
ies, we might say that in the new view intelligence, like science in general,
is a process of discovering truths about the world (or nature) that can be only
metaphorically called a struggle. In other words, while there are secrets of
nature, they are not pieces of information being jealously guarded from our
view; they are simply truths we have not yet discovered. The paradigmatic
intelligence problem is not so much ferreting out the adversary’s secret
intention (as the Korean War example shows, the adversary himself may not
know how he will react to future events) as it is of predicting his behavior
through social science methodology. This is particularly true the more the
emphasis in intelligence analysis shifts to research on long-term trends,
often societal and economic in nature. With respect to future social or
economic conditions, no real secret can be obtained, because the adversary
himself is uncertain what will happen.
The same tendency to say counterintelligence occupies a marginal place
ln intelligence also affects the importance accorded to deception and coun­
terdeception. By categorizing intelligence as social science endeavor, the
new view ignores the possibility of deception. Nature, while it may hide its
secrets from scientific investigators, does not actively try to deceive them.

The Audience fo r Intelligence

The two views of intelligence differ also regarding who the recipients of
tntelligence information, the consumers, are. In Colby’s formulation, the
traditional” view sees the head of state as the prime, and perhaps the only,
recipient of intelligence. This view was also reflected in the National Se­
curity Act of 1947, which established the Central Intelligence Agency and
placed it under the authority of the National Security Council and the pres­
ident, rather than under the Department of State or Defense. The implication
of this bureaucratic arrangement seems to be that the CIA would remain
close to high-level policymakers and would monitor and coordinate the work
of the other, departmental, intelligence agencies on their behalf.
Since the new view, as advocated by former DCIs Colby and Turner
tends to depreciate secrecy, it is not surprising that it also seeks to widen the
audience for intelligence as far as possible. For example, Turner envisages
the creation of an international satellite agency to conduct technical intelli­
gence collection on behalf of the United Nations, with the information being
made available to the whole world.19 Similarly, Colby argues that we are,
or should be, entering a period of free trade with respect to intelligence, in
which the large volume of information available because of modem tech­
nology “ can be seen to provide mutual rather than one-sided benefits.“ 20 A
philosophy to guide this new era is needed; such a philosophy “ must insist
on the recognition of mutual benefit from the free flow and exchange of
information.”

Intelligence and Moral Issues

It may seem strange, when writing on a topic in the usually hardheaded field
of national security studies, to discuss moral issues explicitly. Certainly, one
risks being accused of naïveté for doing so. Yet, those who reflect on the
subject of intelligence have often had to deal with the question of its mo­
rality or, more precisely, the difficult relationship between, for example, the
seeming immorality of inducing people to commit treason, on the one hand,
and its great usefulness and even necessity, on the other.21 In Sun Tzu’s
view, the key to success in intelligence lies in one’s ability to suborn offi­
cials in the enemy’s camp. Sun Tzu realizes that this is a questionable
business, especially for someone who regards loyalty as an important virtue.
He regards “ moral influence”— by which he means “ that which causes the
people to be in harmony with their leaders”— as the first fundamental factor
determining success or failure in war.22 He certainly seems aware that the
necessity of corrupting officials and subjects of other states can hardly be
squared with the ideal of harmony between subjects and rulers.
Even so, in a manner we might anachronistically refer to as Machiavellian,
he exhorts his reader to employ espionage. He argues that the opposite
course— which he expects his readers would regard as more “ honorable’ IS
in fact inhumane. After noting the great burden that war places on the people
and the disruption it causes to their lives, he admonishes in strident tones.
 One who confronts his enemy for many years in order to struggle for
victory in a decisive battle yet who, because he begrudges rank, honors
and a few hundred pieces of gold, remains ignorant of his enemy’s situ­
ation, is completely devoid o f humanity. Such a man is no general; no
support to his sovereign; no master of victory.23
The fact that Sun Tzu makes a moral, and not only a practical, argument for
espionage indicates that he feels it necessary to overcome his readers’ moral
qualms against it.
Despite the antiquity of this illustration, the moral question should not be
regarded as a mere curiosity. It has practical effects in modern times. For
example, in 1929, U.S. Secretary of State Henry L. Stimson, newly ap­
pointed by President Hoover, closed down the Black Chamber, the State
Department’s cryptographic bureau that had been reading foreign diplomatic
and other coded messages for more than a decade. He is said to have acted
on the grounds that “ gentlemen do not read each other’s mail.” 24 From the
point of view of U.S. success in World War II, it was fortunate that a
cryptographic capability was maintained in the Army and Navy, paving the
way for the major successes of the war in breaking the Japanese diplomatic
and naval codes.
The moral issue concerning intelligence resides in the fact that it is a
devious or underhanded form of struggle between nations, as compared with
armed combat, traditionally seen as more open (and hence more honorable).
In its very essence, it involves deception and depends heavily on inducing
the adversary’s citizens to commit treason. As the Stimson example illus­
trates, this type of activity was hardly congenial to the typical (and recur­
ring) American optimism toward placing international relations on a higher
moral plane. (That communications intelligence is now often regarded as
“ clean” compared to the “ messy” business of espionage may only dem­
onstrate the extent to which our standards of international morality have
been diluted.25)
Needless to say, Stimson’s view did not survive the unsettled interna­
tional conditions of the 1930s, let alone World War II itself. However, the
new view of intelligence, as sketched out above, has many characteristics
that make it seem more moral than the “ traditional” one in the eyes of its
Proponents:
the centrality of analysis (understood as a variant of science, and thus
partaking of its prestige) as opposed to espionage,
the primacy of “ clean” technical intelligence over “ messy” espionage,
the depreciation of counterintelligence and deception, diminishing the
awareness that intelligence involves a struggle between nations,
 • the expansion of intelligence to include areas (societal changes, demog­
raphy, narcotics trafficking, and so on) more remote from the political
and military core of the struggle among nations, and
• the widening of the intelligence audience, tending to convert a hand­
maiden of policy (which is used to enhance national interests, often at
the expense of others) into a morally neutral provider of information for
the government, its opponents, and eventually the domestic and world
public.

Thus, this new view of intelligence can be understood as a reassertion of

America’s optimistic outlook after the harsh realities of international rela­
tions during and after World War II.
 8
TOWARD A THEORY OF
INTELLIGENCE

In this concluding chapter, I return to the original question, What is intel­

ligence? On the basis of discussion of the elements of intelligence, we are
now better able to evaluate the various extant views of intelligence, partic­
ularly that presented in the preceding chapter as the “ new American view ,”
and to attempt a more general definition.
The word “ intelligence” is used to refer to a certain kind of knowledge,
to the activity of obtaining knowledge of this kind (and thwarting the similar
activity of others), and to the organizations whose function is to obtain (or
deny) it.1 Of these three categories (knowledge, activity, and organization),
the first is the most basic, since the other two are defined in terms of it.
Thus, the first problem in defining intelligence is to determine the scope of
the knowledge with which it is concerned.
Kent provides a definition of what he calls “ high-level foreign positive
intelligence” : the adjectives are meant to exclude intelligence that is oper­
ational or tactical intelligence (not “ high-level” ), domestic intelligence
(not "foreign” ), or counterintelligence (not “ positive” ). According to
Kent:

What is left is the knowledge indispensable to our welfare and security.

K is both the constructive knowledge with which we can work toward
peace and freedom throughout the world, and the knowledge necessary to
toe defense of our country and its ideals.2
^hus, the knowledge involved is that necessary for conducting foreign pol-
lcy and making the major decisions concerning the development and de­

171
 ployment of military forces in peacetime. It is the knowledge, according to
Kent, on which national security policy may be based.
Its very title (Strategic Intelligence fo r American World Policy) makes
clear that Kent’s book is a discussion of U.S. intelligence rather than intel­
ligence generally. Accordingly, his definition of intelligence is couched in
terms of the information the U.S. government requires. Nevertheless, there
does not seem to be anything about the definition or about the book as a
whole that would limit the applicability of the propositions to the United
States alone. The basic principles are all general, and, indeed, in consider­
ing what one wishes to know about a foreign country, Kent does not look at
America’s potential adversaries, but rather invents the country of “ Great
Frusina’’ to serve as the generic intelligence target. In this discussion, I
therefore take Kent’s statements as applying to intelligence generally.
As Kent implies (by his use of the qualifiers “ high-level,” “ foreign,”
and “ positive” ), his description is too narrow to define intelligence alto­
gether. One would have to take account, somehow, of what he calls
“ operational/tactical intelligence,” “ domestic intelligence,” and “ counter­
intelligence.” But the most important point to note is that Kent’s description
of “ high-level foreign positive intelligence,” as given above, implies that
these other kinds of intelligence are not necessary for “ our welfare and
security.”
It seems fair to assume that Kent’s definition is focused only on peacetime
intelligence requirements. In time of war, intelligence would have to support
not only strategic military decisions, but also the operations of military
forces in the field. Thus, the operational/tactical level of intelligence ought
to be included in any overall discussion of intelligence, and it is hard to
believe that Kent would disagree.
The situation may be different with respect to the other two limitations.
First, limiting intelligence to foreign subjects seems arbitrary, at least in
theory. Threats to the country and its ideals may be domestic as well as
foreign, or they may involve domestic groups with foreign ties. As we have
seen, defining the scope of domestic intelligence is a very difficult task.
Although we typically speak of national security, a nation acts in these
matters through its government, which generally regards the survival of its
form of government as a vital interest. Note that, in the definition above,
Kent refers to “ the defense of [the United States] and its ideals. ” This
recognizes that American democracy could be threatened by something that
did not threaten the country in a physical sense. Be this as it may, domestic
intelligence, in the sense of the information necessary for the regime to
protect itself against violent or revolutionary change, also seems to be a
necessary part of intelligence. Of course, what the proper subjects of that
information are, and how much of it is required, depends heavily on the
nature of the government (the regime).
In addition, the defense of the country may depend on thwarting an
adversary’s intelligence activities, as well as his military operations; coun­
terintelligence as well as positive intelligence must be regarded as necessary
for the “ defense of [the] country.” As compared with the previous example
(operational or tactical intelligence), it is less clear that Kent would agree to
these additions.
In any case, the scope of intelligence must be broadened to include these
areas as well. If intelligence is to provide the knowledge needed to conduct
national security policy, it must include the knowledge to support the actual
use of military forces to pursue national goals and the knowledge that
enables one to frustrate other countries’ intelligence activities. Thus, Kent’s
definition of intelligence must be significantly broadened if it is truly to be
“ the knowledge indispensable to [a country’s] welfare and security.”
At the same time, however, that we have to broaden Kent’s definition to
include knowledge relevant to conducting national security policy, in other
respects, the definition is already too broad. A great deal of information (for
example, technical knowledge in physics or engineering) may be needed to
make informed decisions on some military development questions, but such
information would not ordinarily be considered intelligence. Similarly, a
general understanding of meteorological phenomena is necessary to plan
military actions, but again this type of knowledge would not be considered
intelligence.3 In other words, even when some branch of natural science
makes a major contribution to a nation’s pursuit of its security interests, it
does not automatically become a part of intelligence.
The situation is less clear when it comes to the social sciences. The
subject matter of an intelligence analysis (for example, the political situation
° f a foreign country and how it is likely to evolve) may be similar to the
work done in the social sciences. Even so, the two kinds of analysis exhibit
important differences that suggest that different approaches may be neces­
sary even if the substantive content is similar.
To be useful, an intelligence analysis ought, in discussing the determi­
nants of the political situation in a foreign country, to emphasize those
factors that can be manipulated or changed; the consumer of the analysis is,
after all, typically interested in affecting that political situation and not just
knowing about it. An academic analysis, on the other hand, will be most
lnterested in discovering the most fundamental causes of a given situation,
even if, or especially if, they are immune to change. To the extent that social
science can predict the future course of events (which is, according to Kent
why it is useful for intelligence), it must regard the future as already
determined.4
Thus, the relationship between intelligence and social science is complex
The nature of this relationship can be better addressed in the context of
another major difference between intelligence and science— the close con­
nection between intelligence and secrecy, on the one hand, and science and
the free exchange of ideas, on the other.
From the point of view of the traditional understanding of intelligence (to
say nothing of the popular understanding), what seems to be most glaringly
missing from the definition of intelligence as a kind of knowledge is the
element of secrecy. For Kent, indeed, a concern with secrets is not an
inherent part of intelligence; while noting that one’s adversary may attempt
to deny one access to the information one requires, he regards this as an
incidental problem, akin to the organizational problems that derive from the
large size of a modern state’s intelligence establishment:

Important as they are, these extensions, as I have called them, [i.e.,

subtlety, expertise, clandestinity and size] are external to the heart of the
matter: intelligence work remains the simple, natural endeavor to get the
sort of knowledge upon which a successful course of action can be rested.5

However, the connection between intelligence and secrecy is central to

most of what distinguishes intelligence from other intellectual activities. By
depreciating its importance, Kent can maintain the position that there is no
fundamental difference between intelligence and social science; the Colby
article discussed earlier only takes this view to its logical conclusion when
it puts forward futurology of the “ Club of Rome’’ variety as the model for
intelligence to imitate.6
The connection between intelligence and secrecy in turn reflects the fun­
damental issue of the relationship between intelligence and science. What­
ever similarities might emerge, there is a fundamental difference in the
ultimate goals of the two enterprises. The goal of science is knowledge,
either for its own sake or to further the conquest of nature— the ability to
manipulate natural forces according to man’s will in the interests of his
comfort, health, longevity, and so forth. However, the concept of a struggle
with nature is only a metaphor. In fact, nature, although sometimes com­
plicated and difficult to understand, is indifferent to human efforts to con­
quer it and is not purposefully acting to obstruct them.
Intelligence, on the other hand, involves a real struggle with human
opponents, carried on to gain some advantage over them. It is not surprising?
therefore, that these opponents often are trying not only to obstruct one s
efforts to learn about them, but also to mislead and deceive one. One side’s
intelligence failure is likely to be another side’s counterintelligence success.
Conversely, an intelligence coup by one country implies a counterintelli­
gence or security failure on the part of its opponent.
Once we understand that intelligence is part of a struggle between two
countries, we see why counterintelligence is not an afterthought but is rather
a part of it. Not only is it important to limit or distort what one’s adversary
can learn about one, but one cannot even be sure of what one knows about
an adversary without counterintelligence capability to detect any deception
effort he might have undertaken.
One objection often made to this approach is that it ignores the important
role that open-source information can play in the intelligence process. This
objection is, however, based on a misunderstanding. Open-source informa­
tion is vital for both intelligence and social science; the important distinction
is that, in its intelligence role, it is primarily a means to get around the
barriers that obstruct direct access to the information being sought.7
Fundamentally, intelligence seeks access to information some other party
is trying to deny. Obtaining that information directly means breaching the
security barriers that the other party has placed around the information, by
intercepting communications, stealing documents, overflying restricted
areas and taking photographs, suborning officials with authorized access to
the information, or some other means. But in the absence of direct access to
the information, it may be possible to deduce it from other data— open-
source as well as whatever secret data is available— that can be analyzed.
In the context of this discussion it may appear that the difference between
the view of intelligence emphasizing the importance of secrecy and Kent’s
is mainly a matter of emphasis. The former view stresses the fact that the
adversary is keeping the desired information secret, while Kent stresses the
usefulness of open-source information in finding out what one needs to
know. To some extent, Kent’s emphasis on open sources reflects his view
that secret sources are often unreliable. Perhaps one’s agent on the enemy’s
general staff is really a double agent who is providing deceptive information;
perhaps he is an imaginative swindler (the operator of a paper mill) who is
elever enough to forge plausible war plans to sell at a high price; perhaps the
agent is genuine but his involvement in espionage is already known to the
enemy, who will arrest him just as he is about to relay a critical message.
While this is true, open-source data also can be manipulated to deceive,
fake for example the Arab deception measures designed to convince the
Israelis that no attack was imminent in the fall of 1973. They “ ranged from
welcoming Dr. Kissinger’s peace initiatives in September 1973 to planting
news items in a Lebanese newspaper about the neglect and deterioration of
the Soviet equipment in the Suez Canal zone.” 8 The point is that any type
of intelligence data is subject to distortion, and analysis of it must take this
possibility into account. Indeed, this is a fundamental difference between
intelligence analysis and social science: no one falsifies election returns for
the purpose of confusing the social scientist who is analyzing them (some­
one might, of course, do so for other reasons); the intelligence analyst is
not so fortunate. In the intelligence context, even open-source information
is not as innocent as it appears.
However, Kent’s depreciation of secret sources has a deeper motive.
Espionage is obviously limited to obtaining information the adversary al­
ready has. It is incapable of predicting the enemy’s behavior when he has
not yet made a decision and hence does not know what he will do. Kent, on
the other hand, believes that such prediction is a feasible intelligence task,
provided that intelligence learns to use the methods being developed in the
social sciences.
It is not only that this method is useful for statesmen. Their use of it is,
Kent implies, almost mandatory— anyone rejecting it can be accused of
relying on a crystal ball or, to put it more politely, intuition:

When the findings of the intelligence arm are regularly ignored by the
consumer, and this because of consumer intuition, he should recognize
that he is turning his back on the two instruments by which western man
has, since Aristotle, steadily enlarged his horizon of knowledge— the
instruments of reason and scientific method.9
Thus, whatever insight into the political situation a statesman may possess
is treated as intuition, as opposed to the “ reason and scientific method” of
the “ intelligence arm.”
If intelligence could reliably make the predictions implied in Kent’s dis­
cussion, the policymakers would indeed be foolish to ignore them, and their
jobs would be made much easier. Kent’s optimism on this point reflects the
general optimism of the social sciences in the 1940s and 1950s. This opti­
mism held that adopting a scientific method (such as quantitative methods)
and a scientific outlook (such as behaviorism) would enable the social
sciences to understand social and political phenomena in much the same way
(and ultimately with the same success) as physics understands the atom. In
particular, this understanding would be precise enough to support the pre­
dictive capability Kent attributes to intelligence.
In general, these prospects for the social sciences have not been realized
in recent decades, and they have little chance of realization. The predictive
abilities of intelligence are likely to remain much less than Kent envisaged.
This raises a much more difficult question about the proper relationship
between intelligence and policy.
Everyone would agree that intelligence is subordinate to policy, in the
sense that intelligence activities are directed toward serving the policymaker
(although, as already noted, this subordination would be trivial if the pre­
dictive capability of social science-based intelligence could be perfected).
The range of this service can be quite broad. At one extreme, intelligence is
sometimes used as a reference service, a source of answers to very specific
questions.10 At the other, it prepares extensive analyses, complete with
predictions, of major issues. In either case, however, it supports policy­
makers who must make the actual decision.
The belief in the availability (if not in the present, then in the near future)
of a social science method that would be as rigorous and fruitful as the
scientific method is with respect to the physical world naturally suggests that
those who arc expert in it deserve to be heeded regarding subjects with
which the method deals. Thus, policy should not only accept the facts
provided by intelligence (if relevant facts are available, it would be madness
to ignore them), but its assessments as well.
Unfortunately, such a social science method does not exist. In that case,
intelligence assessments that attempt to make predictions (especially con­
tingent predictions) do not differ from the conclusions that policymakers
might draw about the same situations. Other factors would determine the
relative status to be granted the two estimates.
The intelligence analyst probably has a greater background in the area and
almost certainly can devote greater resources (such as time and access to
data) to the effort. The policymaker is likely to know more about his gov­
ernment’s own policies in the area and other countries’ diplomatic reactions
to them, is more likely to focus on the possible actions to influence the
situation, and is, in any case, responsible for the outcome. The most com­
prehensive and policy-relevant assessment of the situation is likely to result
from a joint effort between the two groups.

The Dual Nature of Intelligence

Intelligence is concerned with that component of the struggle among nations
that deals with information. As such, it has a dual nature, one part governed
by the fact that it deals with information, the other part by the fact that it is
part of the conflict among nations. The first part, taken to its logical ex­
treme, gives rise to the notion of intelligence as a universal, predictive social
science that completely meets the needs of policymakers for information
about other countries, uiciuuuig C lU U U l UAVU lu v u tv u . ----------*_____________________________

entific information, intelligence of this sort would be intrinsically capable of

being shared; it would not lose the characteristic of being intelligence by
being disseminated widely.
The second part, that concerned with the struggle among nations, leads in
another direction. Because intelligence is part of a struggle, the obstacles to
understanding do not simply arise from the difficulties of the subject matter
the more important of them, and those that are potentially the most danger­
ous, are put there by one’s adversary, either in the form of information
denial or deliberate deception. Thus, whatever else one wishes to know, one
has to pay attention to the adversary’s intelligence services as well; indeed,
those sendees become a prime intelligence target since the reward for pen­
etrating and, at best, being able to manipulate them is so high. In this
respect, intelligence has an internal dynamic that tends to transform it into
a counterintelligence duel, in which each nation’s intelligence service seeks,
most of all, to penetrate and manipulate the intelligence service of its
adversary.
In his memoir, former DCI Colby makes the following complaint about
the way that James J. Angleton, the long-time head of the CIA’s counter­
intelligence staff, conducted the operations of his division: “ Indeed, we
seemed to be putting more emphasis on the KGB as the CIA’s adversary
than on the Soviet Union as the United States’ adversary.’’11 Obviously, an
intelligence service exists to serve the interests of its nation with respect to
the nation’s adversaries. But to do this, it often must focus particularly on
the adversary’s intelligence service.
When an intelligence service focuses primarily on the activities of the
adversary service, it runs the risk of descending into a “ wilderness of
mirrors,’’ in which nothing and nobody can be trusted and in which every­
thing may be the opposite of how it appears. If it constantly doubts the
validity of all available evidence, it cannot make much progress in under­
standing the outside world. Intelligence would be analogous to a science of
physics that concentrated on questions of epistemology to the exclusion of
experimentation.
If an intelligence service ignores its adversaries, however, it runs the risk
of being completely deceived and of completely misinterpreting the world it
is trying to understand. In intelligence matters, analysts can rarely be com­
pletely confident of the solidity of the foundations on which they are build­
ing; they must remain open to the possibility that their evidence is
misleading.
Intelligence is thus caught in a dilemma that reflects its dual nature.
Intelligence seeks to learn all it can about the world and its goal may be
er’s director of central intelligence, adopted as the CIA’s motto: “ And ye
shall know the truth and the truth shall make you free.’’12 But intelligence
can never forget that the attainment of the truth involves a struggle with a
human enemy who is fighting back and that truth is not the goal but rather
only a means toward victory.
 NOTES

In tro d u ctio n : W ritin g A b o u t In te llig e n c e

1. Sherman Kent, a former intelligence officer, is the best-known and most im­
portant representative of this school of thought. The argument was first expressed
in his book. S tra te g ic In te llig e n c e f o r A m e ric a n W o rld P o lic y (Princeton, N.J.:
Princeton University Press, 1949; reprint, 1966) and is discussed in detail in chapter
7.
2. Victor Marchetti and John D. Marks, T h e C IA a n d th e C u lt o f In te llig e n c e
(New York: Alfred A. Knopf, 1974), pp. 4, 12.
3. Jeffrey Richelson, T h e U .S . In te llig e n c e C o m m u n ity (Cambridge, Mass.: Ball­
inger, 1985) is a good example of this genre. In the footnotes, “ private informa­
tion“ is occasionally cited as a source; one should probably assume this refers to
leaked information, that is, unauthorized disclosures of information by government
officials.
4. It should be noted for the record that, prior to its publication, this book was
submitted to the Central Intelligence Agency and the Select Committee on Intelli­
gence of the U.S. Senate for security review. In agreeing that S ile n t W a rfa re does
not contain classified information, those organizations are not, of course, endorsing
its contents in any way.
5. Kenneth G. Robertson has noted that intelligence studies in the United King­
dom tend to be more historical in nature than in the United States. See his “ Editorial
Gomment: An Agenda for Intelligence Research,“ D e fe n se A n a ly sis 3, no. 2 (June
1987):95-101. This is due in part to the much stricter control the British govern­
ment exercises over information concerning intelligence in the present and the

181
recent past. In the United States the focus has tended to be more on current public
policy issues involving intelligence. Many issues of this sort were publicly debated
at the time of the mid-1970s congressional investigations of intelligence.

7. W h a t I s In te llig e n c e ?

1. Following Sherman Kent, whose book, S tra te g ic In te llig e n c e f o r A m eric a n

(Princeton, N.J.: Princeton University Press, 1949; reprint, 1966), is
W o rld P o lic y
organized according to this three-part description of intelligence.
2. The term “ adversary” is used here in a broad sense. A friendly government,
with which one is negotiating a treaty, is in this sense an adversary in the context
of the negotiation; each side is presumably seeking to maximize the benefits the
agreement affords it, at least partially at the expense of the other.
3. Herbert E. Meyer, R e a l-W o rld In te llig e n c e (New York: Weidenfeld & Nicol-
son, 1987), p. 6. Meyer is former vice-chairman of the National Intelligence Coun­
cil, the body responsible for producing the most comprehensive and authoritative
U.S. intelligence analyses.
4. A more detailed discussion of this point may be found in Roy Godson, “ In­
telligence: an American View,” in B ritish a n d A m e ric a n A p p ro a c h e s to In te lli­
g e n c e , ed. K. G. Robertson (London: Macmillan Press Ltd., RUSI Defence Studies
Series, 1987), pp. 5-16.
5. The Attorney General’s Guidelines on General Crimes, Racketeering Enter­
prise and Domestic Security/Terrorism Investigations, 1983, part III. The text of the
guidelines may be found in Roy Godson, ed., In te llig e n c e R e q u ire m e n ts f o r the
1 9 8 0 's: D o m e stic In te llig e n c e (Lexington, Mass.: Lexington Books, 1986), pp.
245-64; the discussion of criminal intelligence investigations appears on page 254.
6. In addition, an intelligence service has many opportunities to make money to
fund its own operations or to remit to its government: the clandestine operational
capability of a service could be used to conduct such profitable activities as drug
trafficking and other kinds of smuggling, and the clandestine collection capability
could obtain valuable insider-type information to support speculation on financial
and commodity markets. It is not clear to what extent nations have used their
intelligence services like this.
7. U.S. Commission on the Organization of the Government for the Conduct of
Foreign Policy (Murphy Commission), R e p o rt (Washington, D.C.: Government
Printing Office, 1975), p. 100.
8. See John Ranelagh, T h e A g e n c y : T he R ise a n d D e c lin e o f th e C IA (New York.
Simon & Schuster, 1986), pp. 199-200, for a discussion of the conflicts between
the Office of Policy Coordination (the U.S. government’s covert action agency) an
the Office of Special Operations (the CIA’s espionage arm.) See Christopher An­
drew, H e r M a je s ty 's S e c re t S e r v ic e : T h e M a k in g o f th e B ritish In te llig e n c e C o m -
(New York: Viking, 1986), pp. 476-77, for a discussion of the rivalry
m u n ity
between the British Secret Intelligence Service and its Special Operations Execu­
tive, which was established to carry out sabotage activities and support partisan
forces in Nazi-occupied Europe during World War II.

2. S p ie s, M a ch in e s, a n d L ib ra rie s: C o lle c tin g th e D a ta

1 . This categorization lists the most commonly used collection methods but is not
meant to be complete. For instance, intelligence information can also be collected
by stealing documents or codes from an adversary’s embassy. Such theft may be
considered humint, but it differs greatly from espionage as described in the text.
2. The term “ agent” is ambiguous: it usually refers to the source, although it
sometimes, as in the lexicon of the U.S. FBI, refers to the intelligence officer
instead.
3. For the sake of simplicity, the discussion in the text is phrased in terms of
intelligence officers tasked with spying on the country to which they are posted.
This is not necessarily the case: officers may be posted to Fredonia to recruit
Ruritanians (such as foreign service, intelligence, or military officers) who are also
stationed there. In this case, the officers may be declared (their intelligence con­
nection revealed) to the Fredonian authorities; the purpose of their cover would be
to avoid arousing Ruritanian suspicions. Also, intelligence officers may work in
their own country to recruit foreign diplomats stationed there.
4. Despite the risks of having Soviet nationals working in the U.S. Embassy in
Moscow, reportedly as, among other things, telephone receptionists and operators,
the State Department for many years resisted efforts by other parts of the U.S.
government to force it to end this practice. Eventually, a decision was reached to
reduce dependence on foreign nationals almost to zero. Before it was implemented,
the situation was resolved when, in response to the expulsion of a large number of
Soviet “ diplomats” from Washington in October 1986, the Soviets withdrew the
Soviet work force from the embassy. Presumably they believed that their nationals
would soon be expelled in any case; withdrawing them suddenly disrupted the
embassy’s functioning and let the Soviets respond to the American step without
escalating the conflict and inviting a further response.
5. Given that this claim was made by opponents of a proposed law, since enacted,
to prohibit the disclosure of the identities of intelligence officers, one must take it
with a grain of salt. Nevertheless, there seems to be a good deal of truth behind it.
Clearly certain practices, such as the extensive use of the (otherwise rare) Foreign
Service Reserve (FSR) status for intelligence officers, did help identify them.
6. Cover may also be provided by a business established, owned, and run by an
mtelligcncc service; such an organization is known as a proprietary. A proprietary
other than the collection of intelligence and hence can be useful for covert action
purposes.
7. This account is based on John Barron, K G B T o d a y: T h e H id d e n H a n d (New
York: Reader’s Digest Press, 1983; New York: Berkley Publishing Group, 1985),
pp. 247-314.
8. Frank Gibney, comp., T h e P e n k o v sk iy P a p e rs (New York: Doubleday & Com­
pany, 1965; New York: Ballantine Books, 1982), pp. 57, 87-88.
9. Burgess and Maclean both served at the Foreign Office, while Philby joined the
British Secret Service (MI6) and rose to be head of its counterintelligence section
and its Washington-based liaison officer with the CIA and FBI. See Andrew Boyle,
T h e F o u rth M a n (New York: Dial Press, 1979) for an exhaustive account of the
Soviet spy ring that had its roots in 1930s’ Cambridge.
10. See U.S. Senate, Select Committee on Intelligence, M e e tin g th e E sp io n a g e
C h a llen g e : A R e v ie w o f U n ite d S ta te s C o u n te rin te llig e n c e a n d S e c u rity P ro g ra m s,
99th Cong., 2d sess., 1986, S. Rept. 99-522, pp. 13-15, for a survey of the major
cases of espionage against the United States that came to light in 1981-86.
11. See David Wise, T h e S p y W h o G o t A w a y (New York: Random House, 1988)
for a detailed account of the Howard case.
12. This account is based on Thierry Wolton, L e K G B en F ra n c e (Paris: Bernard
Grasset, 1986), pp. 71-73.
13. Allan Dulles, T h e C ra ft o f In te llig e n c e (New York: Harper & Row, 1963;
Boulder, Colo.: Westview, 1985), p. 216.
14. Christopher Andrew, H e r M a je s ty 's S e c re t S e rv ic e : T h e M a k in g o f th e B ritish
(New York: Viking, 1986), p. 488.
In te llig e n c e C o m m u n ity

15. U .S . N e w s & W o rld R e p o rt, March 19, 1954, p. 62, as cited in Harry Howe
Ransom, C e n tra l In te llig e n c e a n d N a tio n a l S e c u rity (Cambridge, Mass.: Harvard
University Press, 1959), p. 23.
16. The story is told in David C. Martin, W ild e rn ess o f M irro rs (New York: Harper
& Row, 1980). As the title suggests, Martin’s view is that the conflicting stories can
never be resolved, and he recommends abandoning the effort to determine which
defectors were genuine and which were k‘plants.”
17. Even earlier, balloons were used as aerial observation posts by the first French
Republic in 1794 and by the Union Army in the American Civil War. In neither case
was great success achieved; both armies later disbanded their fledgling “ airforces.
See William E. Burrows, D e e p B la c k : S p a c e E sp io n a g e a n d N a tio n a l Security
(New York: Random House, 1987; New York: Berkley Publishing Group, 1 9 8 8 ),
pp. 26-28.
19. William Mitchell, M e m o irs o f W o rld W a r I: “ F ro m S ta rt to F in ish o f O u r
Q rea test W a r" (New York: Random House, 1960), p. 59, as cited in Russell F.
Weigley, T h e A m e ric a n W a y o f W a r (Bloomington, Ind.: Indiana University Press,
1977), p. 224.
20. Andrew, H e r M a je s ty 's S e c re t S e rv ic e , p. 136. At the same time (1915), the
appearance of the German Fokker monoplane fighter meant that aerial reconnais­
sance could be a dangerous business.
21. For a masterful discussion of British scientific intelligence in World War II by
one of its greatest practitioners, see R. V. Jones, T he W iza rd W ar: B ritish S cien tific
In tellig en ce, 1 9 3 9 -1 9 4 5 (New York: Coward, McCann & Geoghcgan, 1978). The
book was first published in Great Britain as M o st S e c re t W a r (London: Hamish
Hamilton, 1978).
22. John Prados, T h e S o v ie t E s tim a te . U .S . In te llig e n c e A n a ly sis a n d R u ssia n
M ilita ry S tre n g th (New York: Dial Press, 1982), pp. 29-30.
23. Stephen E. Ambrose, E ise n h o w e r: T he P resid en t, vol. 2 of E ise n h o w e r (New
York: Simon & Schuster, 1984), pp. 264-65.
24. Ibid., p. 228.
25. Testimony of Major General B. L. Schriever, In q u iry into S a te llite a n d M issile
Hearings before the Preparedness Investigating Subcommittee of the
P ro g ra m s,
Committee on Armed Services, U.S. Senate, 85th Cong., 2d sess., January 6-22,
1958, part 2, pp. 1633-35, as cited in Amrom Katz, “ Technical Collection in the
1980s,” in In te llig e n c e R e q u ire m e n ts f o r th e 1 9 8 0 's: C la n d e stin e C o lle c tio n , ed.
Roy Godson (Washington, D.C.: National Strategy Information Center, 1982), p.
116.
26. G. Zhukov, “ Space Espionage Plans and International Law,” In te rn a tio n a l
(Moscow) (October 1960), as reprinted in U.S. Senate, Committee on
A ffa irs
Aeronautical and Space Sciences, L e g a l P ro b le m s o f S p a c e E xp lo ra tio n , 87th
Gong., 1st sess., 1961, S. Doc. 26, p. 1100.
27. Burrows, D e e p B la c k , p. 107.
28. Katz, “ Technical Collection,” p. 102.
29. The detection and analysis of electromagnetic radiations emanating from ra­
dioactive sources or nuclear detonations (n u c in t) are not considered a part of elint.
See, for example, the definition of electronics intelligence in U.S. Joint Chiefs of
Staff, D ic tio n a ry o f M ilita ry a n d A s s o c ia te d T erm s, JCS Publication 1 (Washing­
ton, D.C.: Joint Chiefs of Staff, 1986), p. 126.
20. Thus, electromagnetic waves given off by electric typewriters, word proces­
sors, or computers can be intercepted and analyzed; in this manner, the texts or data
prepared on these machines might be recovered. To avoid this danger, the United
States has established a set of standards called “ Tempest.” These standards require
that machines of this sort be used to process secret information only if they are
surrounded by shielding that reduces the emanations and impedes any attempt to
intercept them.
31. The practice of intercepting messages predates the use of radio as a means of
communication: couriers were captured and letters opened long before radio ex­
isted. In addition, telegraph cables may be tapped if physical access to them is
available. Before and during World War I, the British gained a special advantage
from the fact that their companies owned and operated the major international cable
lines.
32. Precisely because radio messages might be intercepted, important messages are
likely to be sent in encrypted form. Thus comint and cryptanalysis, the breaking of
codes and ciphers, are intimately related, although cryptanalysis is probably more
reasonably seen as a branch of intelligence analysis than of intelligence collection.
33. For a full discussion of British naval comint during World War I, see Patrick
Beesley, R o o m 4 0 : B ritish N a v a l In te llig e n c e 1 9 1 4 -1 8 (New York: Harcourt Brace
Jovanovich, 1982).
34. An account of this operation may be found in CIA Clandestine Services His­
torical Paper, “ The Berlin Tunnel Operation,” June 1968. (This paper was released
by the CIA to me in sanitized form in accordance with the Freedom of Information
Act.) The possibility that the Soviets knew about the Berlin operation and used it to
pass misleading information is discussed in the section on deception in chapter 5.
35. In his memoir, Herbert O. Yardley, the head of the Black Chamber, is un­
characteristically coy on the subject of how he obtained the encrypted cables. He
does note, however, that the censorship imposed during World War I had been lifted
and that “ supervision of messages [had been] restored to the private cable compa­
nies.” T h e A m e ric a n B la c k C h a m b e r (Indianapolis: The Bobbs-Merrill Co., 1931;
reprint, New York: Ballantine Books, 1981), pp. 156-57.
36. Jones, T h e W iza rd W a r, p. 198.
37. For a fictional account of intelligence collection through sonar, see Tom Clan­
cy, T h e H u n t f o r R e d O c to b e r (Annapolis, Md.: Naval Institute Press, 1984).
Nonfictional accounts are somewhat harder to come by in the open literature.
38. Minor but unpredictable changes in orbit can be accomplished by firing small
rockets on the satellite; the number and magnitude of such maneuvers is limited by
the fuel capacity of the satellite, which is typically small.
39. ABM Treaty, art. XII, para. 1; Interim Agreement on the Limitation of Stra­
tegic Offensive Arms (SALT I), art. V, para. 1; and SALT II, art. XV, para. 1-
40. Since the provision recognizes the right to do only those things allowed by
international law, one might wonder why it is included at all. The answer appears
to be that the Soviets wished to make sure that the United States could not claim,
in order to provide itself with “ assurance of [Soviet] compliance,” a right to
conduct inspections on Soviet territory. In fact, the provision, as written, does not
confer a rig h t to do anything: rather, it imposes a d u ty to use NTM in accordance
with international law. The next paragraph, which likewise appears in all three
agreements, prohibits interference with NTM when they are “ operating in accor­
dance with [international law].”
41. See, for example, A.S. Piradov, ed.. In te rn a tio n a l S p a c e L a w , Boris Belitsky
trans. (Moscow: Progress Publishers, 1976), pp. 192, 218. This work notes the
fear, which it considers well grounded, that “ the use of satellites for the study of
natural resources in other countries might become a form of legalized economic
espionage . . .” (p. 216). The same views are expressed more recently in Gennady
Zhukov and Yuri Kolosov, In te rn a tio n a l S p a c e L a w , Boris Belitsky trans. (New
York: Praeger, 1984), pp. 133, 143.
42. Piradov, In te rn a tio n a l S p a c e L a w , p. 136.

43. Stansfield Turner, S e c re c y a n d D em o c ra c y : T he C IA in T ra n sitio n (Boston:

Houghton Mifflin, 1985; New York: Harper & Row, 1986), p. 92.

44. Jozef Garlinski, T h e E n ig m a W a r (New York: Charles Scribner’s Sons, 1979),

p. 16.
45. Sherman Kent, S tra te g ic In te llig e n c e f o r A m e ric a n W o rld P o lic y (Princeton,
N.J.: Princeton University Press, 1949; reprint, 1966), pp. 3-4. Kent—a veteran of
the Office of Strategic Services (OSS), the U.S. intelligence service of the World
War II era—provided a peculiarly American theory of intelligence right after World
War II.

46. V. Zaykin, “ ‘Secret’ Classification Removed,” Izv e stiy a , Sept. 3, 1988, p. 2,

reprinted and trans. in Federal Broadcast Information Service (FBIS), S o v ie t R e ­
p o rt, September 8, 1988 (FBIS-SOV-88-174), pp. 53-54.

47. One might wonder why such an operation should be conducted by an intelli­
gence agency at all, given its total reliance on open-source materials and on meth­
odologies developed in the academic world that are familiar to the Soviets.
Obviously, it need not be, but whether placing it in a nonclassified research center
would be a good idea depends on a balance of several factors. On the one hand, it
might be easier to attract personnel to such an open center, and its product would
be more easily shared with the academic community. On the other hand, special,
potentially cumbersome procedures would be needed to integrate this open-source
material with secret information or to use this information to support secret intel­
ligence activities. For example, a request for information needed to recruit a Soviet
official to do espionage would have to be camouflaged to conceal the reason for the
request.
48. Donald E. Queller, T h e O ffice o f A m b a ssa d o r in th e M id d le A g e s (Princeton
N.J.: Princeton University Press, 1967), p. 90.
49. What is described in the text is the ordinary activity of a military attaché, which
is essentially overt, although there may be cases where he attempts to travel in or
near restricted areas to observe objects the host country would prefer he did not. At
the same time, however, the position of military attaché (like any diplomatic po­
sition) could be used by an intelligence officer as cover.
50. For a discussion of the possibilities opened up by the relaxation in Soviet travel
regulations, see R. Keeler and E. Miriam Steiner, “ Collection,” in In tellig en c e
R e q u ire m e n ts f o r th e 1 9 9 0 's : C o lle c tio n . A n a ly sis, C o u n te rin te llig e n c e a n d C o v ert
A ctio n , Roy Godson ed. (Lexington, Mass.: D.C. Heath & Co., Lexington Books,
1989), pp. 42-57.

3. W h a t D o e s It A ll M e a n ? In te llig e n c e A n a ly sis a n d P ro d u ctio n

1. See David Kahn, T he C o d e b re a k e r s: T h e S to ry o f S e c re t W ritin g (New York:

Macmillan, 1967) for the most complete public treatment of the history of cryp­
tography. Since his book was published, however, revelations about the British
success in breaking German ciphers in World War II have added a new and most
important chapter to that history.
2. See Patrick Beesly, R o o m 4 0 : B ritish N a v a l In te llig e n c e 1 9 1 4 -1 8 (New York:
Harcourt Brace Jovanovich, 1982), pp. 3-7, 22-33.
3. See Barbara Tuchman, T h e Z im m e rm a n n T ele g ra m (New York: Macmillan,
1958, 1966) for a full account of this incident.
4. See Herbert O. Yardley, T h e A m e ric a n B la c k C h a m b e r (Indianapolis: The
Bobbs-Merrill Co., 1931; reprint, New York: Ballantine Books, 1981), p. 4.
5. Kahn, T h e C o d e b re a k e r s , p. 348.
6. The official Japanese name for the machine American cryptanalysts called
“ Purple” was Alphabetical Typewriter 97.
7. Such a mathematical theory was developed at the beginning of the 1930s by the
Polish mathematician-cryptographers who laid the groundwork for the early Polish
successes and the later Allied wartime successes against the German Enigma ma­
chine. See, for example, Jozef Garlinski, T h e E n ig m a W a r (New York: Charles
Scribner’s Sons, 1979), pp. 25, 196-204.
8. For example, there is now an academic-style journal called Cryptologia de­
voted to these topics. In addition, the U.S. government’s National Bureau of Stan­
dards has developed a data encryption standard (in the form of an enciphering
algorithm) for the electronic transmission and computer storage of unclassified
government data; it would also be available for businesses, such as banks, that
might want to encrypt their electronic transmissions.
9. More speculatively, it might be noted that on the basis of British World War II
experience skill in chess seems to be positively correlated with cryptologic ability.
If so, the Soviet Union would not suffer from a lack of potential cryptanalysts.
Similarly, the British cryptanalysts tended to be musical, as well.
10. Many possible mistakes in using a cipher can help the cryptanalyst. For ex­
ample, the German World War II procedure of repeating, at the beginning of each
message, a three-letter key was particularly damaging: simply knowing that the first
and fourth letters of the plaintext were the same (as were the second and fifth, and
the third and sixth) was a tremendous help to British cryptanalysts. Similarly,
routine message formats, standard salutations, or other stock phrases may provide
important clues for cryptanalysts.
11. Kahn, T h e C o d e b re a k e rs, p. 603. Ronald Lewin, T he A m e ric a n M a g ic : C o d es,
C ip h ers a n d the D e fe a t o f J a p a n (New York: Farrar Straus Giroux, 1982), p. 113.
12. Kahn. T h e C o d e b re a k e r s . p. 604.
13. Lewin, pp. 116-17. See also Edwin T. Layton, “ A n d 1
T he A m e ric a n M a g ic ,
W as T h e re '':
Pearl H a rb o r a n d M id w a y —B re a k in g th e S e c re ts (New York: William
Morrow and Company, Inc., 1985), pp. 453-56, for a discussion of this incident.
14. Lewin, T h e A m e ric a n M a g ic, p. 117.
15. Summary of Magic intelligence (derived from intercepted Japanese diplomatic
cable traffic) dated September 11, 1942, cited in B. Nelson Macpherson, “ The
Compromise of U.S. Navy Cryptanalysis After the Battle of Midway,” In te llig e n c e
a n d N a tio n a l S e c u rity 2, no. 2 (April 1987):321. The summary does not indicate the
date of the message from Japan to Lisbon.
16. See Christopher Andrew, H e r M a je s ty 's S e c re t S e rv ic e : T h e M a k in g o f the
B ritish In te llig e n c e C o m m u n ity (New York: Viking, 1986), pp. 331-32, for a
description of this incident.
17. Bob Woodward and Patrick E. Tyler, “ Libyan Cables Intercepted and De­
coded,” W a sh in g to n P o st, April 15, 1986, p. Al.
18. Stephen Engleberg, “ U.S. Aides Worried over Libya Cables,” N e w Y ork
April 17, 1986, p. A24.
T im es,

19. See Ronald Lewin, U ltra G o e s to W ar (London: Hutchinson & Co., 1978; New
York: McGraw Hill Book Co., 1987), pp. 99-103, for an account of Ultra intel­
ligence with respect to the raid.
20. For a rough analogy, one might imagine being confronted with a set of graphs,
lacking both labels and scales for the x and y axes, and being told only the general
nature of the phenomenon of which the graphs describe various aspects. One would
then be asked to figure out what each graph means and to describe the phenomenon
in detail.
21. The Second Common Understanding to article XV, paragraph 3, of the SALT
II Treaty states:
Each Party is free to use various methods of transmitting tclemetric information during
testing, including its encryption, except that, in accordance with the provisions of para­
graph 3 or article XV of the Treaty, neither Party shall engage in deliberate denial of
telemctric information, such as through the use of telemetry encryption, whenever such
denial impedes verification of compliance with the provisions of the Treaty.
22. Some have argued that U.S. insistence on including an antiencryption clause in
the SALT II Treaty backfired by alerting the Soviets to the fact that the United States
was successfully intercepting and using Soviet missile telemetry. On the other hand.
Senator Gordon J. Humphrey, basing his discussion on news reports, has suggested
that Soviet espionage success against a U.S. sigint-collecting satellite ‘‘may have
contributed to the Soviet decision to encrypt telemetry . . . ” (“ Analysis and Com­
pliance Enforcement,” in V erifica tio n a n d S A L T : T h e C h a lle n g e o f S tra te g ic D e ­
cep tio n , ed. William C. Potter [Boulder, Colo.: Westview Press, 1980], p. 112.

23. The President’s Unclassified Report to the Congress on Soviet Noncompliance

with Amis Control Agreements, February 1, 1985. This report reaffirms the finding
of the president’s earlier Report to the Congress of January 23, 1984.
24. Amrom Katz, a key figure in the development of aerial photography and photo
interpretation in the United States during and after World War II, has noted that,
despite the fact that Pis were able to identify the Cuban sites for Soviet intermediate-
range ballistic missiles on photographs taken from high altitudes, the U.S. Air Force
and Navy conducted low-level reconnaissance flights during the missile crisis to get
detailed photography that would be more convincing to the policymakers. “ Tech­
nical Collection,” in In te llig e n c e R e q u ire m e n ts f o r th e 1 9 8 0 's : C la n d estin e C o l­
lectio n , ed. Roy Godson (Washington, D.C.: National Strategy Information Center.
1982), p. 107.
25. “ Ground resolution” refers to the ability of a photograph to render barely
distinguishable a standard pattern consisting of parallel black-and-white lines (i.e.,
lines and spaces) of equal width. A ground resolution distance of, say, one foot
would describe a photograph in which the pattern whose line-plus-space width is
one foot would be rendered barely recognizable as such. In a finer pattern—one with
a smaller line-plus-space width—the lines Would not be distinguishable at all. (This
definition, and the citation in the text, are from Amrom Katz, “ Observation Sat­
ellites: Problems and Prospects,” Astronautics [April 1960]:5-6; “ essentially iden-
tical” to Amrom Katz, O b se rv a tio n S a te llites: P ro b le m s, P o ssib ilitie s a n d
P ro sp e cts, RAND Paper P-1707, May 25. 1959.)
26. For a discussion of crateology, see Robert S. Greenberger, “ Can CIA Cratol-
ogy Ultimately Outsmart Kremlin’s Shellology?” W a ll S tre e t J o u rn a l, January 10,
1985, p. Al.
27. See William E. Burrows, D e e p B la c k : S p a c e E sp io n a g e a n d N a tio n a l S ec u rity
(New York: Random House, 1987; New York: Berkley Publishing Group, 1988),
pp. 108-11, for a discussion of this point. Burrows claims that an intelligence
analyst at the Defense Intelligence Agency made this connection, although it is not
clear w'hether the inference was accepted elsewhere in the intelligence community
before the discovery' of the IRBM launch sites. DCI John McCone did expect the
Soviet missile deployments, but it appears that his prediction was based more on
political judgments about Nikita Khrushchev than on evidence of this sort. See John
Ranelagh, T h e A g e n c y : T h e R ise a n d D e c lin e o f th e CIA (New York: Simon &
Schuster, 1986), pp. 394-96.
28. The reason for the emplacement pattern in the first place was presumably that
it maximized the antiaircraft missiles’ effectiveness. Nevertheless, given the high
premium the Soviets placed on secrecy as an ingredient of the deployment plan, a
degradation in the SA-2s’ effectiveness would have been a small price to pay for not
letting the cat out of the bag.
29. Sherman Kent, S tra te g ic In te llig e n c e f o r A m e ric a n W o rld P o lic y (Princeton,
N.J.: Princeton University Press, 1949; reprint, 1966), pp. 12-13, gives the table
of contents of one such book about a country’ of interest prepared in Germany for
use during World War II:
I. G e n e ra l B a c k g ro u n d . Location. Frontiers. Area. History. Governmental and
Administrative Structure.
II. C h a ra cte r o f th e C o u n try. Surface Forms. Soils. Ground Cover. Climate.
Water Supply.
III. P eo p le. Nationalities, Language, Attitudes. Population Distribution. Settle­
ment. Health. Structure of Society.
IV. E co n o m ic . Agriculture. Industry. Trade and Commerce. Mining. Fisheries.
V. T ra n sp o rta tio n . Railroads. Roads. Ports. Airfields. Inland Waterways.
VI. M ilita ry G e o g ra p h y . [Detailed regional breakdown].
VII. M ilita ry E sta b lish m e n t in B ein g . Army: Order of Battle, Fixed Defenses,
Military Installations, Supply. Navy: Order of Battle, the Fleet, Naval Shore In­
stallations, Naval Air, Supply. Air: Order of Battle, Military Aircraft, Air Instal­
lations, Lighter than Air, Supply.
VIII. S p e c ia l A p p e n d ix e s . Biographical data on key figures of government. Local
geographical terminology. Description of rivers, lakes, canals. List and specifica­
tions of electric power plants. Description of roads. List of airdromes and most
important landing grounds. List of main telephone and telegraph lines. Money,
weights, and measures. Beaches [as for amphibious military operations].
30. House of Representatives, Committee on the Armed Services, L e sso n s L e a rn e d
a s a R e su lt o f th e U .S . M ilita ry O p e ra tio n s in G ren a d a : H e a rin g , 98th Cong., 2d
sess., 1984, H. Rept. 98-43, p. 24. Given the Reagan administration’s concern with
the Marxist-Leninist regime of Grenadian leader Maurice Bishop (whose arrest and
execution by his colleagues set in motion the events that led to the invasion), this
lack of maps and other basic data provides an illustration of a failure to anticipate
U.S. policy.
31. For a full discussion of this task, described as the avoidance of technological
surprise, see Michael I. Handel, “ Technological Surprise in War,” In te llig e n c e a n d
N a tio n a l S e c u rity 2, no. 1 (January 1987): 1-53. Handel characterizes the impor­
tance of scientific and technical intelligence as follows:
More than any other type of war in the past, modem warfare, based on the continuous
development of new weapons systems at an ever-accelerating pace, depends on intelli­
gence. . . . Victory or defeat (in the Battle of Britain] was often decided by the “battle of
intelligence’’ before combat had even begun. The role played by scientific technological
intelligence in war will increase in proportion to the technological advancement of the
adversaries and the use they make of state-of-the-art weaponry (p. 40).
32. R. V. Jones, T h e W iza rd W ar: B ritish S c ie n tific In te llig e n c e , 1 9 3 9 -4 5 (New
York: Coward, McCann & Geoghegan, 1978), p. 104. The Knickbein system
involved two unidirectional radio beams transmitted from different antennae that
intersected over the bombers’ target; the bombers would fly along one beam and
would drop their bombs when they crossed the other.
33. Eliot Cohen, “ Analysis,” in In te llig e n c e R e q u ire m e n ts f o r th e 1 9 90s: C o lle c ­
ed. Roy Godson (Lexington,
tio n , A n a ly s is , C o u n te rin te llig e n c e a n d C o v e rt A c tio n ,
Mass.: D. C. Heath & Co., Lexington Books, 1989), p. 82. Emphasis supplied.
34. Even in the case of Soviet economics, where the intelligence analyst’s access
to data the Soviets do not publicly release might give the intelligence analyst an
advantage over other experts, this has not necessarily been the case. In 1977, for
example, the CIA produced several studies of the Soviet oil industry that predicted
that production would have fallen to 8 million to 10 million barrels per day (mb/d)
by 1985 and that the Soviet Union would import (for its own use and for reexport
to Eastern Europe) between 3.5 and 4.5 mb/d. Academic and business exports were
skeptical of these results, and indeed they have turned out to be wide of the mark.
Soviet oil production stayed at or near the 12 mb/d level through the 1980s, and the
Soviet Union has continued to be an exporter, rather than an importer, of oil. For
a review of this issue generally sympathetic to the CIA, see U.S. Senate, Select
Committee on Intelligence, T h e S o v ie t O il S itu a tio n : A n E va lu a tio n o f C IA A n a ly se s
o f S o v ie t O il P ro d u c tio n : S ta ff R e p o r t , 95th Cong., 2d sess., 1978, Committee
Print.
35. Kent, S tra te g ic In te llig e n c e , pp. 7-8.
36. Intelligence agencies are sometimes regarded as competitors of the news me­
dia, often to the agencies’ disadvantage. For example, news reports that U.S.
officials followed the failed Panamanian coup d’etat of October 1989 on the Cable
News Network insinuated that this was evidence of a failure of American intelli­
gence. But it is an unfair comparison, since the CIA is not designed to provide
up-to-the-minute reporting.
37. Kent, S tra te g ic In te llig e n c e , p. 38.
38. Cord Meyer, F a cin g R e a lity : F ro m W o rld F e d e ra lism to th e C IA (New York:
Harper & Row, 1980), p. 352.
39. U.S. Senate, Select Committee to Study Governmental Operations with Re­
spect to Intelligence Activities ( “ Church Committee” ), F in a l R e p o rt, B o o k I:
F oreign a n d M ilita ry In te llig e n c e , 94th Cong., 2d sess., 1976, S. Rept. 94-755, pp.
272-73.
40. Consider the following reported remark of a high official of the U.S. State
Department: “ I had a friend of mine from the CIA come down yesterday to give me
a briefing on East Germany. He told me that he revised his paper three times on the
drive from the CIA to the State Department just because of things he was hearing
on the radio” (Thomas L. Friedman, “ In Quest of a Post-Cold War Plan,” N e w
Y ork T im es, November 17, 1989, p. A22). The remark was meant to illustrate how
quickly Eastern Europe was changing in November 1989; it also indicates that the
analyst’s briefing was concerned more with the latest events than with a deeper
understanding of the political dynamics of East Germany.
41. See Timothy M. Laur, “ Principles of Warning Intelligence,” in T h e M ilita ry
ed. Gerald W. Hopple and Bruce W. Watson (Boulder,
In tellig en c e C o m m u n ity ,
Colo.: Westview Press, 1986), pp. 149-68, for a fuller description of warning
analysis.

42. Gordon Brook-Shepherd, T h e S to rm B ird s: S o v ie t P o stw a r D e fe c to rs (New

York: Weidenfeld & Nicolson, 1989), pp. 330-31.

43. According to the most recent Executive Order on United States Intelligence
Activities (EO 12333, December 4, 1981), the intelligence community is defined as
the CIA, the National Security Agency, the Defense Intelligence Agency, “ offices
within the Department of Defense for the collection of specialized national foreign
intelligence through reconnaissance programs,” the State Department’s 1NR, and
the intelligence elements of the armed services, the FBI, and the Departments of the
Treasury and of Energy. Para. 3.4(f). The text of the executive order appears as
Appendix II in In te llig e n c e R e q u ire m e n ts f o r th e 1 9 8 0 ’s: E le m e n ts o f In te llig e n c e ,
rev. ed., ed. Roy Godson (Washington, D.C.: National Strategy Information Cen­
ter, 1983), pp. 117-37.
44. Avi Shlaim, “ Failures in National Estimates: The Case of the Yom Kippur
War,” W o rld P o litics 28, no. 3 (April 1976):368-69.
45. Alexander Orlov, H a n d b o o k o f In te llig e n c e a n d G u errilla W a rfa re (Ann Ar­
bor: University of Michigan Press, 1965), p. 10.
46. Barton Whaley, C o d e w o rd B a rb a ro ssa (Cambridge: MIT Press, 1973);
Shlaim, “ Failures in National Intelligence Estimates,” pp. 348-80; Michael I.
Handel, P erc e p tio n , D e c e p tio n a n d S u rp rise : T h e C a se o f th e Y om K ip p u r W ar
(Jerusalem: Leonard Davis Institute of International Relations, Jerusalem Paper No.
19, 1976); Roberta Wohlstetter, P e a rl H a rb o r: W a rn in g a n d D e c isio n (Stanford,
Calif.: Stanford University Press, 1962); Edwin T. Layton, ilA n d I W as T here:
P e a rl H a r b o r a n d M id w a y — B re a k in g th e S e c re ts (New York: William Morrow and
Company, Inc., 1985); Gordon W. Prange et al., A t D a w n W e S le p t: T he U n to ld
S to ry o f P e a rl H a rb o r (New York: McGraw-Hill Book Co., 1981); Harvey de
Weerd, “ Strategic Surprise in the Korean War,” O rb is 6, no. 3 (Fall 1962):435-
52.
47. See Cohen, “ Analysis,” in In te llig e n c e R e q u ire m e n ts f o r th e 1990s, pp. 84-
87, for a discussion of the latter event. Cohen argues that although the political
leadership in Washington may have been surprised by the Chinese attack, General
Mac Arthur, the U.S. and U.N. commander, was well aware of the massive infil­
tration of Chinese troops into Korea; the disaster was caused not so much by
surprise as by the use of inappropriate tactics against the Chinese Army. This will
be discussed in note 52 to this chapter.
48. Lew in, U ltra G o e s to W a r , pp. 317-18.
49. Stansfield Turner, S e c re c y a n d D e m o c ra c y : T h e C IA in T ra n sitio n (Boston:
Houghton Mifflin, 1985; New York: Harper & Row, 1986), p. 113. Turner received
the note on November 11, 1978.
50. A report by the staff of the U.S. Senate Intelligence Committee contends that
“ the question of oil price levels was analyzed in the context of a narrow supply-
and-demand framework, which tended to overlook both political influences and
such economic factors as elasticities of supply and demand.” It also claims that
“ political aspects of relationships among OPEC nations and the internal dynamics
of the Saudi Government. . . were not consistently integrated into the community’s
economic analysis” (Select Committee on Intelligence, Subcommittee on Collec­
tion, Production, and Quality, U .S . In te llig e n c e A n a ly sis a n d th e O il Issu e , 1 9 7 3 -
1974: S ta ff R e p o rt, 95th Cong., 1st sess., 1977, Committee Print, pp. 4, 5). This
suggests that the intelligence analysts involved practiced economic analysis that was
too theoretical and hence apolitical.
51. There were elements of tactical surprise, for example, with respect to using
artificial harbors to supply the Allied forces until major harbors such as Cher­
bourg could be captured. See Handel, “ Technological Surprise in War,” pp. 14-
15, for a discussion of the problem these artificial harbors posed for German
intelligence.
52. See Cohen, “ Analysis,” in In te llig e n c e R e q u ire m e n ts f o r th e 1 9 9 0 s, pp. 84-
87. General Mac Arthur’s command believed that U.S. supremacy in the air could
be used as effectively against the Chinese People’s Liberation Army (PLA) as it had
been against the North Korean People’s Army (NKPA). Ultimately, this mistaken
judgment reflected a failure to assess the significance of the differences between the
heavily mechanized and hence road-bound NKPA (formed on the Soviet model)
and the unmechanized and nonroad-bound PLA, whose more dispersed infantry
hordes presented less lucrative targets for aerial attack.
53. Lewin, U ltra G o e s to W a r, pp. 347-51.
54. Albert Wohlstetter, “ Is There a Strategic Anns Race?” F o reig n P o lic y , no. 15
(Summer 1974):3-20, and “ Rivals, But No Race,” F o re ig n P o lic y , no. 16 (Fall
1974):48-81.
55. This is similar to the question of “ paradigm shift” discussed in the philosophy
of science—the way in which one organizing framework, such as Newtonian phys­
ics, gives way to another, for example, relativity. However, two important differ­
ences, both of which suggest that paradigm shift can occur more easily with respect
to scientific theory than to intelligence analysis, should be noted: (1) the pioneers
of the new framework do not need the tolerance, let alone the agreement, of the
upholders of the old framework to publish their views, and (2) it is possible to
conceive and carry out experiments specifically designed to decide between the two
frameworks. Even so, it has been claimed that shifts are due not so much to the
conversion of existing scientists to the new framework as to its adoption by new
entrants into the discipline.
56. This is sometimes referred to by the psychological term “ projection,” which
means, more technically, the psychological mechanism by which one attributes to
others the feelings or characteristics (such as hostility or dishonesty) that one
dislikes about oneself and wishes to disown. While this mechanism may be a
psychological basis of mirror-imaging in some cases, the two concepts, which
operate at different levels, are not identical; for example, projection could not be the
cause of mistakenly attributing to an adversary a characteristic (such as peacefulness
or satisfaction with the status quo) that one possesses and with which one is com­
fortable.
57. The failure of Israeli intelligence in this case is discussed in Shlaim, “ Failures
in National Intelligence Estimates.” A key cause of the failure was the continued
adherence by Israeli intelligence, in the face of accumulating evidence of a possible
attack, to the “ conception,” or the view, that Egypt would not attack until it was
able to stage deep air strikes to destroy the Israeli Air Force and that Syria would
not attack without Egypt. The first and crucial part of the conception reflects, on the
military level, the same mirror-imaging that existed on the political level—the belief
that Egypt would not start a war it did not stand a good chance of winning.
58. A l-H a m is h m a r , September 14, 1975, as cited in Shlaim, “ Failures in National
Intelligence Estimates,” p. 362.
59. This passage was deleted from the final version of the NIE but survived as a
dissenting footnote expressing the separate views of the director of the INR of the
D e p a rtm e n t o f S ta te ( U .S . S e n a te , C h u rch C o m m itte e , F in a l R e p o rt, B o o k I: F o r ­
eig n a n d M ilita r y In te llig e n c e , 94th Cong., 2d sess., 1976, S. Rept. 94-755, p. 78).

60. Jones, T h e W iza rd W ar, p. 457.

61. Ibid., p. 458.
62. In principle, one could speak of failures of intelligence collection; however,
one rarely does so. Part of the reason may be that the most interesting cases of
intelligence failure are those for which relevant information was in fact available.
63. It may be noted that, to the extent that intelligence shifts its focus from political
and military matters to economic and social ones (as suggested by, for example,
former Director of Central Intelligence William Colby, and discussed in chapter 7),
this assertion becomes less convincing. An intelligence service concerned primarily
with forecasting socioeconomic trends might find that its adversaries were uncon­
cerned about its progress and not interested in impeding it. (But not necessarily: a
major Soviet deception effort of the 1920s tried to convince Western intelligence
services that communism in the Soviet Union was fading and that its leaders were
moving toward an unthreatening nationalism. John J. Dziak, C h e k isty : A H isto r y o f
the K G B [Lexington, Mass.: D.C. Heath and Co., Lexington Books, 1988], p. 48.)
Similarly, sometimes one wants one’s adversary to understand the situation cor­
rectly. For example, a strategy of nuclear deterrence (assuming one possesses
adequate forces to carry it out and is not bluffing) depends on the other govern­
ment’s correctly assessing that one’s forces, even after an initial attack, would still
be capable of wreaking unacceptable damage on its country.
64. See U.S. Senate, Select Committee on Intelligence, Subcommittee on Collec­
tion, Production and Quality, T he N a tio n a l In te llig e n c e E stim a te s A -B T ea m E p i­
so d e C o n c ern in g S o v ie t S tra te g ic C a p a b ility a n d O b jec tiv es, 95th Cong., 2d sess.,
1978, Committee Print.
65. Richard K. Betts, “ Analysis, War, and Decision: Why Intelligence Failures
Are Inevitable,” W o rld P o litic s 31, no. 1 (October 1978):61-89.
66. In the early 1980s, the CIA’s Intelligence Directorate was reorganized along
geographical lines, that is, the political, economic, and (for the most part) military
analysis with respect to a given country or region were combined in the same
division. (Previously, it had been organized functionally; for example, all the eco­
nomic analysis was in one office.) Although undertaken mainly to improve inter­
action with the policy community, which itself tends to be organized geographically,
this was a helpful step from the point of view of the problem discussed in the text.

4. W o rkin g B e h in d th e S c e n e s: C o v e rt A ctio n

1. Paragraph 3.4(h) of Executive Order 12333 of December 4, 1981, United

States Intelligence Activities [48 Fed. Reg. 58847 (1981)], reprinted in In te llig e n c e
 t h e 1 9 8 0 ' s : E l e m e n t s o f I n t e l l i g e n c e , rev. ed., ed. Roy Godson
R e q u ir e m e n ts f o r

(Washington, D.C.: National Strategy Information Center, 1983), p. 137.

2. Section 662 of the Foreign Assistance Act of 1961, as amended (22, U.S.C.
2242), known as the Hughes-Ryan Amendment. The amendment itself docs not
name the set of activities defined in this way, but they are commonly called covert
action.
The two definitions substantially overlap but do not coincide. Secret foreign
influence activities undertaken by a government agency other than the CIA would
be considered covert action under the executive order but not under Hughes-Ryan.
Secret CIA operations (other than intelligence gathering) in support of a hostage
rescue mission would be covert action under Hughes-Ryan but arguably not under
the executive order, if the purpose were solely humanitarian and not to advance any
foreign policy objectives. (The latter point is admittedly unclear, since it could be
said that obtaining the release of hostages is itself a foreign policy objective.) See
Americo R. Cinquegrana, “ Dancing in the Dark: Accepting the Invitation to Strug­
gle in the Context of ‘Covert Action,’ The Iran-Contra Affair and the Intelligence
Oversight Process,” H o u sto n J o u r n a l o f In te rn a tio n a l L a w 11, no. 1 (Fall
1988): 177—209, for a complete discussion of the various definitions of covert
action.
3. Donald E. Qucllcr, T h e O ffice o f A m b a ssa d o r in th e M id d le A g e s (Princeton,
N.J.: Princeton University Press, 1967), p. 93. Queller notes that “ [t]he line be­
tween legitimate gathering of information and espionage is difficult to draw, . . . ”
(p. 93). He also gives examples of ambassadorial involvement in bribery, subver­
sion, and assassination (pp. 93-95). We are somewhat relieved to read that, despite
their bad reputation, “ it does not seem that [Renaissance] ambassadors [, even those
from Venice,] commonly participated in assassinations” (p. 93).
4. See U.S. Congress, R e p o rt o f th e C o n g re ssio n a l C o m m itte e s In ve stig a tin g the
Ira n -C o n tra A ffa ir, 100th Cong., 1st sess., 1987, H. Rept. 433, S. Rept. 216, pp.
37-45, 67-69, 504, and passim for a discussion of this activity.
5. John Bruce Lockhart, a former officer of Britain’s Foreign Office with practical
experience in intelligence, uses the term “ special political action” as a synonym for
covert action; this presumably reflects British usage ( “ Intelligence: A British View”
in K. G. Robertson, ed., B ritish a n d A m e ric a n A p p ro a c h e s to In te llig e n c e [London:
Macmillan Press, 1987], pp. 37, 46).
6. This definition of active measures is taken from Richard H. Shultz and Roy
Godson, D e z in fo r m a ts ia : A c tiv e M e a su re s in S o v ie t S tra te g y (McLean, Va.:
Pergamon-Brassey’s, 1984), p. 193.
7. The United States has invented a term, “ public diplomacy,” that would cover
some of the overt active measures techniques but has not created the organizational
structures to implement it. National Security Decision Directive 77 (NSDD-77),
January 14, 1983, defines public diplomacy very broadly as “ those actions of the
U.S. Government designed to generate support for our national security
objectives.”
8. In their expose of the CIA, Victor Marchetti and John D. Marks mention
critically a police-training program in South Vietnam; however, their main com­
plaint is directed against the involvement of an academic institution (Michigan State
University in this case) in a covert program ( T h e C IA a n d th e C u lt o f In te llig e n c e
[New York: Alfred A. Knopf, 1974], p. 234).
9. Section 660 of the Foreign Assistance Act of 1961, as amended in 1974 by P.L.
93-559 [22 U.S.C. 2420]. This provision was watered down in 1985 by the enact­
ment of exemptions for Costa Rica, El Salvador, and Honduras.
10. See Shultz and Godson, D ezin fo rm a tsia , pp. 133-49, for a discussion of the
case and a detailed analysis of the contents of the newsletter.
11. Shultz and Godson, D e zin fo rm a tsia , p. 38.
12. The quoted material is from Bob Woodward, “ CIA Curried Favor With
Khomeini, Exiles,” W a sh in g to n P o st, November 19, 1986, pp. A l, 28. The de­
fection was reported in “ Soviet Diplomat in Iran Defects and Flees to Britain,”
N e w Y ork T im es, October 24, 1982, p. A 14.

13. Churchill relates the story of this attempt in his history of World War II. He
tried to get around Stalin’s suspicions by sending a “ short and cryptic” message
that did not warn of the German attack but merely discussed some German troop
movements that pointed in that direction; he hoped that such a message “ would
arrest [Stalin’s] attention and make him ponder,” thereby leading him to draw the
desired conclusion on his own. Unfortunately, the implementation of Churchill’s
tactic was somewhat botched by the British ambassador in Moscow (T h e G ra n d
A llia n c e , vol. 3 of T h e S e c o n d W o rld W a r [Boston: Houghton Mifflin Company,
1951], pp. 357-61).
14. Ladislav Bittman, T h e K G B a n d S o v ie t D isin fo rm a tio n : A n In sid e r s V iew
(McLean, Va.: Pcrgamon-Brassey’s, 1985), pp. 112-13.
15. Sec John Barron, K G B T o d a y: T h e H id d e n H a n d (New York: Reader’s Digest
Press, 1983; New York: Berkley Publishing Group, 1985), pp. 32-159, for
Levchenko’s account, as told to Barron, of his intelligence career and eventual
defection.
16. Ibid., pp. 76-81,85-90, and 93-94. The article, by suggesting that the double­
agent operation was an American provocation, was aimed at reducing public sup­
port for a tough stand by the Japanese government against Soviet diplomatic
pressure; the GRU agent was in fact released. The purpose of the letter—to create
public pressure on the Japanese government to accede to the Soviet demand that the
pilot be returned—was not achieved.
17. Total concealment may not be possible, depending on the case. Thus, a foreign
intelligence service can determine the location from which a “ black” radio station
is broadcasting by using direction-finding equipment. This information, however,
will not typically be available to the average listener.

18. The National Voice of Iran (NVOI) and its inflammatory anti-American rhet­
oric are discussed in a CIA study, “ Soviet Covert Action and Propaganda,” pre­
sented to the Oversight Subcommittee, Permanent Select Committee on
Intelligence, House of Representatives, February 6, 1980, by the deputy director for
operations, Central Intelligence Agency. The study is reproduced in U.S. House of
Representatives, Permanent Select Committee on Intelligence, Subcommittee on
Oversight, S o v ie t C o v e r t A c tio n (T h e F o rg e ry O ffe n siv e ): H ea rin g s, 96th Cong., 2d
sess., 1980. The discussion of the NVOI appears on pages 78-79.

19. Ray S. Cline, T h e C IA : R e a lity vs. M y th , rev. ed. (Washington, D.C.: Acrop­
olis Books, 1982), p. 151.

20. Cord Meyer, F a c in g R e a lity : F ro m W o rld F ed e ra lism to th e C l A (New York:

Harper & Row, 1980), pp. 110-13.

21. Novosti is not an acknowledged voice of the Soviet government or the Soviet
Communist party; instead, it is supposed to be an organ of public opinion as
represented by such associations as the Union of Journalists and the Union of
Writers (Shultz and Godson, D e zin fo rm a tsia , p. 28).

22. The version provided to foreign media included thirty-four paragraphs on So­
viet foreign policy that were absent from the N e w Y o rk T im e s version. In 1977, the
T im es claimed that these paragraphs had been forged by the CIA (John M. Crews-
don, “ The CIA’s 3-Decade Effort to Mold the World’s Views,” N e w Y ork T im e s ,
December 25, 1977, p. Al).
Crewsdon reported that the CIA had obtained “ an expurgated version” of
Khrushchev’s original speech, a version that had been “ prepared for delivery to the
nations of Eastern Europe, from which some 34 paragraphs of material concerning
future Soviet foreign policy had been deleted” ; the extra thirty-four paragraphs
made available to foreign media were written by CIA counterintelligence experts in
Khrushchev’s style and presumably designed to cause confusion in the Soviet bloc
countries. Crewsdon does not explain why the additional paragraphs were omitted
from the domestic version—the one plausible explanation, that the CIA withheld the
forged material from the T im e s to reduce the possibility that U.S. public opinion
would be misled, would have been at odds with the article’s general tenor.
John Ranelagh concludes, from the fact that the Soviets never disputed their
authenticity, that the additional paragraphs were genuine (T h e A g e n c y : T h e R ise a n d
D e c lin e o f th e C IA [New York: Simon & Schuster, 1986], p. 287n). According to
him, the CIA received two copies of the text, a complete version (presumably from
a source in the Soviet Union) as well as the expurgated (Eastern European) version.
However, he does not explain why the CIA gave only the expurgated version to the
T im es. He would, it seems, have to argue that domestic publication would have
placed the Soviet source in greater jeopardy than foreign publication.
23. U.S. Senate, Select Committee to Study Governmental Operations with Re­
spect to Intelligence Activities (“ Church Committee” ), F in a l R e p o r t, B o o k I:
F o reig n a n d M ilita ry In te llig e n c e , 94th Cong., 2d. sess., 1976, S. Rept. 94-755, p.
194.
24. See House Permanent Select Committee on Intelligence, S o v ie t C o v e rt A ctio n
(T h e F o rg ery O ffe n siv e ), for a discussion of this phenomenon in the 1960s and
1970s. Soviet use of this technique has continued into the 1980s; for recent exam­
ples, see U.S. Department of State, S o v ie t In flu e n c e A c tiv itie s : A R e p o rt on A c tiv e
M e a su re s a n d P ro p a g a n d a , 1 9 8 6 -8 7 , August 1987, pp. 29-32 and 79-80. A
question relating to a possible U.S. use of this technique has arisen with respect to
the CIA’s dissemination of Khrushchev’s “ secret speech” on Stalin. See note 22 to
this chapter.
25. See Shultz and Godson, D e zin fo rm a tsia , pp. 194-95.
26. The additional paragraphs—the authenticity of which has been disputed—dealt
only with foreign policy issues. See note 22 to this chapter.
27. Cline, T h e C IA , pp. 123-24.
28. NSC directive 10/2 (June 1948), as cited in ibid., p. 126.
29. Church Committee, F in a l R e p o r t , B o o k I, p. 145.
30. Cline, T h e C IA , pp. 150-51.
31. Church Committee, S ta ff R e p o r t: C o v e rt A ctio n in C h ile: 1 9 6 3 -1 9 7 3 , 94th
Cong., 1st sess., 1975, Committee Print, pp. 29, 45.
32. Ibid., p. 49. Emphasis in original.
33. Gregory F. Treverton, C o v e rt A c tio n : T h e L im its o f In te rv e n tio n in th e P o stw a r
W o rld (New York: Basic Books, 1987), p. 142.

34. Ibid., p. 143.

35. See Church Committee, A lle g e d A ssa ssin a tio n P lo ts In v o lv in g F o reig n L e a d ­
94th Cong., 1st sess., 1975, S. Rept. 94-465, for the com­
ers: In te rim R e p o rt,
mittee’s investigation of these issues.
36. Executive Order 11905, “ United States Foreign Intelligence Activities,” Feb­
ruary 18, 1976, sect. 5(g) [41 Fed. Reg. 7733 (1976)].
37. Executive Order 12036, January 24, 1978, sect. 2-305 [43 Fed. Reg. 3687
(1978)] and Executive Order 12333, December 4, 1981, sect. 2.11 [48 Fed. Reg.
59947 (1981)]. Both provisions drop the adjective “ political,” but the meaning is
presumably unchanged.
38. Turner’s testimony may be found in U.S. House of Representatives, Permanent
Select Committee on Intelligence, Subcommittee on Legislation, H .R . 1013, H R .
1 3 7 1 , a n d O th e r P ro p o sa ls W h ich A d d re s s th e Issu e o f A ffo rd in g P rio r N o tic e o f
 100th Cong., 1st sess., 1987, pp.
C o v e rt A c tio n s to th e C o n g re ss: H e a rin g s,
44-49. One might argue that such a rescue mission was humanitarian and not
undertaken to influence foreign conditions, events, or behavior and hence not a
covert action at all. However, Admiral Turner had no choice but to describe it as
such. First, it met the relevant legal tests: it was an activity engaged in by the CIA
for a purpose other than the collection of intelligence (the Hughes-Ryan standard),
and given that a great deal of the U.S. government’s foreign policy activity was
devoted to freeing the hostages, it may be said to have been an activity “ conducted
abroad in support of national foreign policy objectives . . . planned and executed so
that the role of the U.S. Government is not apparent or acknowledged publicly”
(the definition contained in President Carter’s Executive Order 12036 on United
States Intelligence Activities, which was then in effect). Second, whatever its
motive, the rescue attempt had a significant effect on international relations, since
it resulted in the closing of the Canadian Embassy in Teheran.
39. Sefton Delmer, B la c k B o o m e ra n g (New York: Viking Press, 1962), p. 120.
40. Note that the Executive Order 12333 definition of covert action cited above
refers to activities “ which are planned and executed so that the role of the United
States Government is not apparent or a c k n o w le d g e d p u b lic ly , . .” (emphasis sup­
plied). This would appear to envisage cases where the government’s role, though
apparent to an observer of international affairs, or even reported in the press, is still
not publicly acknowledged.
41. Nikita S. Khrushchev, K h ru sh c h e v R e m e m b e rs: T he L a st T e sta m e n t (Boston:
Little, Brown and Company, 1974), pp. 447-48.
42. See John Dyson, S in k th e R a in b o w ! A n E n q u iry in to th e ' G re e n p e a c e A ffa ir ’
(London: Victor Gollancz Ltd., 1986), pp. 157-86, for an account of the French
government’s handling of the affair.
43. Section 662(a) of the Foreign Assistance Act of 1961, as amended. This text
was itself amended in 1980 by the intelligence oversight provisions of the Intelli­
gence Authorization Act for Fiscal Year 1981 (P.L. 96-450).
44. In its F in a l R e p o rt, the Church Committee ignored the issue of plausible denial
in discussing the amendment’s effects. Instead, it describes the two results of the
amendments as (1) a statutory responsibility of the executive branch to inform
Congress about covert actions and (2) inclusion of the Senate Foreign Relations and
House Foreign Affairs among the committees to be informed {B o o k I: F o re ig n a n d
M ilita ry In te llig e n c e , p. 151).

45. R e p o r t o f th e C o n g re ssio n a l C o m m itte e s In v e stig a tin g th e Ira n -C o n tra A ffa ir,

100th Cong.* 1st sess., 1987, H. Rept. 100-433, S. Rept. 100-216, p. 271.
46. See Christopher Andrew, H e r M a je s ty 's S e c re t S e rv ic e : T he M a k in g o f th e
(New York: Viking, 1986), pp. 476-77. It has been
B ritish In te llig e n c e C o m m u n ity
suggested that SOE’s existence as an agency separate from the MI6 was due in part
to domestic political considerations: by placing SOE under the minister of economic
warfare, a member of the Labour party, Churchill met the demands of his coalition
partners for a share of control of the intelligence services. See M. R. D. Foot, S O E ,
A n O u tlin e H isto r y o f th e S p e c ia l O p e ra tio n s E x e c u tiv e , 1 9 4 0 -4 6 , rev. ed. (n.p.:
University Publications of America, 1986), pp. 19-20.
47. Church Committee, F in a l R e p o rt, B o o k I , p. 106.
48. The name was later changed to the Directorate of Operations (DDO).
49. Church Committee, F in a l R e p o rt, B o o k I, pp. 107-8.
50. For a critique of the concept of the elements of intelligence, see Kenneth G.
Robertson, “ The Study of Intelligence in the United States.” in C o m p a rin g F o r­
eig n In te llig e n c e : T he U .S ., th e U SSR , th e U .K . a n d th e T h ird W o rld , ed. Roy
Godson (McLean, Va.: Pergamon-Brassey’s, 1988), pp. 26-28.

5. S p y vs. S p y: C o u n te rin te llig e n c e

1. The term “ counterintelligence” is sometimes defined to include protection

against “ sabotage, or assassinations conducted for or on behalf of foreign powers,
organizations or persons, or international terrorist activities, . . . ” (The quoted
words are from the definition of counterintelligence in President Reagan’s Executive
Order 12333 of December 4, 1981, on United States Intelligence Activities, para.
3.4[a].) This is a reasonable development, since these activities are identical with
or similar to the kinds of covert action a hostile intelligence service might carry out;
protecting against them, therefore, is likely to involve many of the same skills and
methods as protecting against hostile intelligence activities.
2. The executive order definition of counterintelligence referred to in note 1,
above, specifically excludes “ personnel, physical, document or communications
security programs.” This reflects the bureaucratic fact that, in general, U.S. de­
partments and agencies that deal with secret information are responsible for the
security programs to protect that information, while the FBI and CIA have primary
responsibility for counterintelligence (as defined in the executive order) at home and
abroad, respectively. For our (theoretical) purposes, however, security should be
considered as a part of counterintelligence since it serves the same function.
3. Harold C. Relyea, “ The Presidency and the People’s Right to Know,” in
Harold C. Relyea, ed., T h e P re sid e n c y a n d In fo rm a tio n P o lic y (New York: Center
for the Study of the Presidency, 1981), pp. 11-19.
4. Executive Order 10290, September 24, 1951.
5. Executive Order 12356 on National Security Information, April 1, 1982.
6. Ibid., sect. 1.1(a).
7. National Security Act of 1947, sect. 102(d) (3) [50 U.S.C. 403 (d) (3)].
 8. K e ep in g the N a tio n 's S e c re ts, A Report to the Secretary of Defense by the
Commission to Review DoD Security Policies and Practices, November 1985, p.
23, cited in U.S. Senate, Select Committee on Intelligence, M ee tin g th e E sp io n a g e
C h a llen g e : A R e v ie w o f U n ite d S ta te s C o u n te rin tellig e n c e a n d S e c u rity P ro g ra m s,
99th Cong., 2d sess., 1986, S. Rept. 99-522, p. 66.
9. The current Soviet policy of g la sn o st might be viewed in this light, as a policy
established by the top leader (Gorbachev) to advance his political goals and weaken
his potential opponents in the party and state bureaucracies by allowing the news
media to publicize their faults and mistakes. This is, of course, not to say that
g la sn o s t cannot develop, and has not already developed, into something more
deeply rooted in Soviet society that the top leadership might find difficult to do away
with in the future.
10. Until 1989, the Soviet government published only a single number that it
described as its defense budget. However, the number was patently too small to
include all Soviet military expenditures. In 1989, along with the disclosure of a
much higher total defense expenditure, the Soviets explained that the previously
published budget figure included only the salaries and upkeep of the members of the
armed forces.
11. Similarly, the Soviet decision in 1989 to publish a more (but not completely)
plausible figure for its defense expenditures may have resulted from an internal
political factor—a desire to increase the new Supreme Soviet’s control over military
matters. Alternatively, the Soviets may have recognized that the secrecy surround­
ing the defense budget increased Western distrust and, for that reason, was an
obstacle to achieving more important foreign policy goals.
12. K eep in g the N a tio n ’s S e c re ts, p. 49.
13. Bob Woodward, “ CIA Paid Millions to Jordan’s King Hussein,” W a sh in g to n
February 18, 1977, p. Al.
P o st,

14. Atomic Energy Act of 1954 Sec. 11y [42 U.S.C. 2014y].
15. Sec. 148 [42 U.S.C. 2168].
16. Adm. B. R. Inman, “ National Security and Technical Information,” paper
presented to the annual meeting of the American Association for the Advancement
of Science, January 1982, mimeo (Washington, D.C.: The American Association
for the Advancement of Science, 1982).
17. The general export controls are contained in the International Trade in Arms
Regulations 22 CFR chapter I, subchapter M, part 125. These regulations explicitly
address the issue of the export of unclassified technical data; in general, an export
license is required except when the data are in the public domain, that is, readily
accessible to the public.
18. In the United States, at least theoretically, lack of loyalty to the government
disqualified a person from any federal civil service position; however, this criterion
has been in effect dropped with respect to jobs that do not require access to classified
information. See Guenter Lewy, “ The Federal Loyalty-Security Program,” in I n ­
te llig e n ce R e q u ire m e n ts f o r th e 1 9 8 0 ’s: D o m e stic In te llig e n c e , ed. Roy Godson
(Lexington, Mass.: Lexington Books, 1986), p. 147 and passim, for a discussion of
this development. The loyalty requirement is contained in President Truman’s Ex­
ecutive Order 10450, Security Requirements for Government Employment, April
27, 1953.
19. Provided that no violent, criminal action had been taken based on this belief.
Even then, it would not be permissible to compile a complete membership list of the
organization unless it could be shown that the membership at large was involved in
the organization’s violent activities. These issues are discussed further in the next
chapter.
20. U.S. Senate, Subcommittee on Criminal Laws and Procedures, Committee on
the Judiciary, H e a rin g s on th e E ro sio n o f L a w E n fo rc e m e n t In te llig e n c e a n d Its
Im p a c t o n the P u b lic S e c u r ity , part 8, 95th Cong., 2d sess., 1978, as cited in
Guenter Lewy, “ The Federal Loyalty-Security Program,” p. 152.
21. In 1983, the Office of Technology Assessment (OTA) of the U.S. Congress
produced a study generally critical of the claims made on behalf of the polygraph.
S cien tific V a lid ity o f P o ly g ra p h T estin g : A R ese a rc h R e v ie w a n d E v a lu a tio n —
T e c h n ic a l M e m o ra n d u m , OTA-TM-H-15 (Washington, D.C.: Government Printing
Office, 1983). The following year, the Defense Department, whose National Se­
curity Agency makes widespread use of the polygraph, published a rebuttal. U.S.
Department of Defense, T he A c c u ra c y a n d U tility o f P o ly g ra p h T estin g (Washing­
ton, D.C.: Government Printing Office, 1984).
22. See the discussion, later in this chapter, of the Cuban double-cross operation
run against the United States. Presumably, at least some of these double agents,
whom the CIA accepted as genuine, had been subjected to polygraph tests.
23. See U.S. Senate, Select Committee on Intelligence, M e e tin g th e E sp io n a g e
C h a llen g e : A R e v ie w o f U n ite d S ta te s C o u n te rin te llig e n c e a n d S e c u rity P ro g ra m s ,
99th Cong., 2d sess., 1986, S. Rept. 99-522, pp. 12-15, for a summary of the
major cases between 1980 and 1986.
24. Testimony of James Schlesingcr before the Senate Budget Committee, as cited
in “ For the Record,” W a sh in g to n P o st, July 1, 1987, p. A18.
25. Howard evaded FBI surveillance and turned up several months later in Mos­
cow. Yurchenko redefected to the Soviet Union, which suggests that his original
defection might have been bogus and that he might have been dispatched by the
Soviets to mislead U.S. intelligence or for some other reason. If so, giving up
Pelton and Howard—no longer employed by the United States, they were presum­
ably of no further use to the Soviets in any case—was intended to bolster Yur­
chenko’s authenticity or bona fides. See David Wise, T he S p y W ho G o t A w a y (New
York: Random House, 1988), pp. 17-21 and passim.
26. A particularly dramatic case (not, however, involving a double agent) is the
famous Japanese bomb-plot message of September 1941. In this message, the
Japanese consulate in Honolulu was instructed to provide the precise berths of the
U.S. Navy ships at Pearl Harbor. The United States decoded the message but did
not figure out why Japan wanted such detailed information. See Gordon W. Prange,
A t D a w n W e S le p t: T h e U n to ld S to ry o f P e a rl H a rb o r (New York: McGraw-Hill,
1981), pp. 248-52.
27. John C. Masterman, T he D o u b le -C ro ss S y ste m in th e W a r o f 1 9 3 9 to 1945
(New Haven: Yale University Press, 1972), p. 3 (emphasis in the original). This
work, by the man who managed the system, is the source of the account in the text;
further citations are from pp. 30-31, 38^40, 41, 49, and 58.
28. The original agent had worked for the German Abwehr (military intelligence)
in the late 1930s; the British, aware of his existence, allowed him to continue
operating in part because he had also had some dealings with British MI6 that would
have complicated any attempt to prosecute him.
29. Masterman, T h e D o u b le -C ro ss S ystem .

30. Ibid.
31. Michael Wines and Ronald J. Ostrow, “ Cuban Defector Claims Double Agents
Duped U.S.,” W a sh in g to n P o st, August 12, 1987, p. A8.
32. Formally, the All-Russian Extraordinary Commission to Combat Counterrev­
olution and Sabotage; “ Cheka” is an acronym based on the second and third words
of the formal name.
33. For a discussion of the Trust, see John J. Dziak, C h e k isty : A H isto r y o f the
KGB (Lexington, Mass.: D.C. Heath and Co., Lexington Books, 1988), pp. 47-50.
34. John Prados, T h e S o v ie t E stim a te : U .S . In te llig e n c e A n a ly sis a n d R u ssia n
M ilita ry S tre n g th (New York: Dial Press, 1982), pp. 42-43.
35. See Arnold L. Horelick and Myron Rush, S tra te g ic P o w e r a n d S o v ie t F o reig n
(Chicago: University of Chicago Press, 1966), pp. 42-116, for a full dis­
P o lic y
cussion of these exaggerated claims and the purposes they served.
36. See Horelick and Rush, S tra te g ic P o w e r , pp. 117-40.
37. Prange, A t D a w n W e S le p t, pp. 338, 353.
38. One would also want to know how the adversary processes the information. If,
for instance, because of the time pressure of combat, he will not have the oppor­
tunity to analyze the photographs fully before having to take action, it may be
possible to fool him with an inferior imitation. By the time detailed analysis reveals
to him that the object was a dummy, he has already acted on the basis of his
mistaken first impression. This is another reason successful deception operations are
more to be expected in wartime than in peacetime.
39. See M. R. D. Foot, S O E , A n O u tlin e o f th e S p e c ia l O p e ra tio n s E xe cu tive ,
1 9 4 0 -4 6 , rev. ed. (n.p.: University Publications of America, 1986), pp. 130-34.
For a more complete account, see the memoir of the German counterintelligence
officer who ran the operation, H. J. Giskes, L o n d o n C a llin g N o rth P o le (London:
William Kimber and Co., 1953). It does not appear that the Germans used this
channel to deceive the British on matters other than the operations of their supposed
SOE agents in the Netherlands.
40. This incident is recounted in H. M. G. Lauwers, “ Epilogue,” in Giskes,
L o n d o n C a llin g N o rth P o le , p. 194. Lauwers, a Dutch SOE agent captured by the
Germans, tried unsuccessfully to alert London to the deception operation by trans­
mitting messages without the proper security checks. Unfortunately, headquarters
ignored their absence and treated the messages as genuine.
41. Stansficld Turner, S e c re c y a n d D em o c ra c y : T he C IA in T ra n sitio n (Boston:
Houghton Mifflin, 1985; Harper & Row, 1986), p. 65.
42. As with everything else, a trade-off is involved here. Varying the orbit means
that some satellites follow one that is less than optimal in its coverage; what is
required is a balance between the ordinary measures of cost effectiveness and the
more subtle (and less measurable) requirement to counter the adversary’s conceal­
ment and deception operations.
43. David Kahn, an expert on cryptology, makes these points in “ Discussion,” in
ed. Roy Godson
In te llig e n c e R e q u ire m e n ts f o r th e 1 9 8 0 ’s: C la n d e stin e C o llec tio n ,
(Washington, D.C.: National Strategy Information Center, 1982), p. 119.
44. The CIA had noticed that one type of U.S. electric cipher machine transmitted
an electrical signal representing the plaintext character (the machine’s input) along
with the enciphered character (its intended output); this electrical signal, although
faint compared to the enciphered character, could be picked up from the transmis­
sion wire far away from the cipher machine. It turned out that Soviet cipher ma­
chines suffered from the same defect, enabling the United States to recover the
plaintext messages along with their encrypted versions (John Ranelagh, T h e A g e n c y :
T h e R ise a n d D e c lin e o f th e C IA [New York: Simon & Schuster, 1986], p. 140).

45. Ibid., p. 289.

46. Ibid., p. 295.
47. Why would the Soviets would want to close down the operation if they were
using it to pass deceptive information? They might have done so to harm CIA and
MI6 morale, or because they believed that the Western powers would soon discover
that the Soviets knew about the operation anyway, or because the cost in chicken
feed (legitimate information deliberately given away to bolster confidence in the
operation) was higher than they wished to pay. The discovery may have been
accidental in the sense that the repair party was not in the know and stumbled across
the tunnel because no one who was had thought to prevent it from working in the
tunnel area (if the Soviets were running a deception operation, knowledge of it
would have been kept within a small circle). Once the United States and Britain
became aware of the discovery, they would have become suspicious had the Soviet
continued to use the cable without removing the taps.
48. Charles Cruickshank, D e c e p tio n in W o rld W a r II (Oxford: Oxford University
Press, 1979), p. 182.
49. This entire controversy has been called the “ wilderness of mirrors” and is the
subject of a book with that title by David Martin (New York: Harper & Row, 1980).

6. G u a rd in g th e G u a rd ia n s: T h e M a n a g e m e n t o f In te llig e n c e

1. Loch Johnson, A S e a s o n o f In q u iry : T h e S e n a te In te llig e n c e In ve stig a tio n

(Lexington, Ky.: University Press of Kentucky, 1985), pp. 224, 268.
2. U.S. Senate, Select Committee to Study Governmental Operations with Re­
spect to Intelligence Activities (“ Church Committee” ), A lle g e d A ssa ssin a tio n P lo ts
In v o lv in g F o reig n L e a d e r s: In te rim R e p o r t , 94th Cong., 1st sess., 1975, S. Rept
94-495, p. 263.
3. On May 7, 1962, Attorney General Robert F. Kennedy was officially informed
of actions taken during 1960 and 1961 in connection with attempts to assassinate
Castro, specifically the CIA’s involvement in this connection with two Mafia fig­
ures, John Roselli and Sam Giancana. The formal necessity for the briefing arose
from the fact that the FBI wanted to prosecute Robert Maheu, the CIA’s go-between
with the Mafia figures, for installing an illegal wiretap, and the CIA wanted to
forestall this prosecution to prevent the revelation of the entire story. Kennedy is
reported to have been very angry about the Mafia involvement, which would have
complicated prosecution of Roselli or Giancana for any involvement they may have
had in organized crime; he is not reported as being angry about the assassination
attempt itself (ibid., pp. 131-34).
4. The best evidence for this proposition is of the “ dog that didn’t bark” variety.
One would have expected major Kennedy administration figures, such as former
Defense Secretary Robert McNamara, to express some outrage had they believed
that the CIA had undertaken such an action without approval by higher authority;
instead, McNamara emphasized, in testifying before the Church Committee, his
belief that “ the CIA was a highly disciplined organization, fully under the control
of senior officials of the government” (ibid., p. 158, citing McNamara’s testimony
of July 11, 1975). Robert Kennedy’s reaction in 1962 is reported to have been
similar; see the preceding note.
5. Christopher Andrew, H e r M a je s ty ’s S e c re t S e rv ic e : T he M a k in g o f th e B ritish
(New York: Viking, 1986), p. 1.
In te llig e n c e C o m m u n ity

6. See Winston S. Churchill, M a rlb o ro u g h : H is L ife a n d T im e s , vol. 6 (New

York: Charles Scribner’s Sons, 1938), pp. 482-84 and 526-29. Marlborough’s
opponents claimed that these monies were public funds and should have been
accounted for as such; since he spent the funds to pay his secret agents for infor­
mation, this would have been impossible. Churchill makes a strong case that his
illustrious ancestor behaved properly in using traditional sources of money to fi­
nance intelligence operations, which were not otherwise funded by the government.
7. These provisions originally appeared in the resolutions that established the
Senate and House intelligence committees and were enacted into law in 1980. The
National Security Act of 1947, as amended, sect. 501.
8. Robert M. Gates, “ The CIA and American Foreign Policy,” F o re ig n A ffa irs
66, no. 2 (Winter 1987/88):225.
9. A n to n y a n d C le o p a tra , Act II, scene 5, lines 85-88.
10. Henry Brandon, T h e R e tr e a t o f A m e ric a n P o w e r (New York; Doubleday &
Co., 1973), p. 103. President Johnson is said to have made this remark at a White
House dinner in the presence of Director of Central Intelligence Richard Helms.
11. “ Managing/Tcaching New Analysts,” S tu d ie s in In te llig e n c e 30, no. 3 (Fall
1986):3— 4. The name of the author (apparently a manager in the Directorate of
Intelligence, CIA) was deleted when the article was released in response to a
Freedom of Information Act request. I am grateful to Eliot Cohen for calling my
attention to this article. He discusses it in “ Analysis,” in In te llig e n c e R e q u ire m e n ts
f o r the 1 9 9 0 s: C o lle c tio n , A n a ly sis, C o u n te rin te llig e n c e a n d C o v e rt A c tio n , ed. Roy
Godson (Lexington, Mass.: D.C. Heath & Co., Lexington Books, 1989), pp.
71-96.
12. Cohen, “ Analysis,” p. 76.
13. Charles H. Fairbanks, Jr., “ Where Is the Secret?” W a sh in g to n P o st, February
25, 1987, p. A23. Emphasis in original.
14. Ibid.
15. Bob Woodward and Dan Morgan, “ Soviet Threat Toward Iran Overstated,
Casey Concluded,” W a sh in g to n P o st, January 13, 1987, p. Al. It was this news­
paper article that contained the original leak of the SNIE.
16. Ibid.
17. Robert M. Gates, “ The CIA and Foreign Policy,” F o reig n A ffa irs 66, no. 2
(Winter 1987/88):221.
18. The National Security Act of 1947, sect. 501, as amended [50 U.S.C. 413].
19. Gates, “ The CIA and Foreign Policy,” p. 225.
20. The 1980 intelligence oversight legislation explicitly provided that “ the fore­
going [notification] provision shall not require approval of the intelligence commit­
tees as a condition precedent to the initiation of any such anticipated intelligence
activity [i.e., covert action].” The National Security Act of 1947, as amended, sect
501(a)(1)(A) [50 U.S.C. 413(a)(1)(A)].
21. Jim Lehrer, of the Public Broadcasting System’s “ McNeil/Lehrer News
Hour,” was quoted by Radio TV Reports in March 1987, as follows: “ I think the
House and Senate Intelligence Committees are colanders of leaks, and it comes
from the staff. It doesn’t come from the principals . . . let me tell you, when you’re
30 years old or 34 and you’re carrying all that wisdom and heavy stuff in your head,
they [sic] are going to go tell it.” Coming from someone presumably in a position
to know, such a comment about sources, even in generalized form, is very rare.
22. Leon V. Sigal, R e p o rte rs a n d O fficia ls: T h e O rg a n iza tio n a n d P o litic s o f
(Lexington, Mass.: D.C. Heath & Company, 1973), pp. 113-14,
N e w sm a k in g
discusses the use of deep background to help an official “ establish an alibi.”
23. Once a reporter has the original lead, he can often force other officials to
comment by threatening to run the story anyway; thus, the official faces the choice
of having the story appear in print or on the air in a version that reflects the view
of the original leaker (who may be a bureaucratic rival) or of providing additional
information or another interpretation (confirming the basic facts in the process) in
an attempt to make sure the published story is more balanced from his point of view.
24. U.S. Senate, Select Committee to Study Governmental Operations with Re­
spect to Intelligence Activities ( “ Church Committee” ), F in a l R e p o rt: B o o k I:
F o re ig n a n d M ilita r y In te llig e n c e , 94th Cong. 2d sess., 1976, S. Rept. 94-755, p.
522. Halperin, an official of the American Civil Liberties Union, was arguing at this
point for the abolition of the capacity for carrying out covert operations.
25. This account is taken from a paper by the chairman of the SIRC, Ronald G.
Atkey, “ Security Intelligence Review Committee: Legislative Oversight and Gov­
ernment Policy in Canada,” prepared for the Conference on Intelligence and Policy,
sponsored by the Defense Intelligence College, Washington, D.C., August 26-28,
1986.
26. “ The right of the people to be secure . . . against unreasonable searches and
seizures, shall not be violated, and no warrants shall issue, but upon probable
cause, . . .” U.S. Constitution, Fourth Amendment.
27. P.L. 95-511 [50 U.S.C. 1801-11].
28. See, for example, the proposal to require warrants for “ unconsented physical
searches” in national security cases in the proposed National Intelligence Reorga­
nization and Reform Act of 1978 (S. 2525, 95th Cong., 2d sess., informally known
as the intelligence charter), sect. 341.
29. In a thorough survey entitled “ The Study of Intelligence in the United States,”
Kenneth G. Robinson notes: “ So far there has been little work on the nature of
internal threats and how these can be linked to intelligence requirements for do­
mestic intelligence.” C o m p a rin g F o re ig n In te llig e n c e : T h e U .S ., th e U SSR , the
U .K . a n d th e T h ird W o rld , ed. Roy Godson (McLean, Va.: Pergamon-Brassey’s,
1988), p. 19.
30. Of course, not all illegal actions are of interest from the point of view of
domestic intelligence; the criminal standard does not indicate what domestic intel­
ligence should do, only what it should not do.
31. Surveillance of U.S. persons (citizens and legal aliens) is permitted only if
there is probable cause to believe that the individual knowingly engages, for or on
behalf of a foreign power, in sabotage, international terrorism, or other clandestine
intelligence activities that involve or are about to involve a violation of criminal
law.
32. A sanitized version of these guidelines was released in connection with a
Freedom of Information Act request and appears in U.S. Senate, Select Committee
on Intelligence, N a tio n a l In te llig e n c e R eo rg a n iza ta io n a n d R e fo rm A c t o f 1978:
H e a r in g s , 95th Cong., 2d sess., 1978, Committee Print, pp. 774-90. According to
the guidelines, foreign intelligence (as opposed to counterintelligence) may “ be
collected only with the express approval of the Attorney General or his designee”
(para. VI.B.3, ibid., p. 782). This requirement would seem to imply that the
collection of foreign intelligence information was viewed as an exceptional activity.
33. Para. II.A, ibid., p. 775. Emphasis supplied.
34. The Foreign Agents Registration Act of 1938, as amended, requires “ public
disclosure by persons engaging in propaganda activities and other activities for or
on behalf of . . . foreign principals so that the government and people of the United
States may be informed of the identity of such persons and may appraise their
statements and actions in the light of their associations and activities.” (The citation
is from the act’s statement of “ Policy and Purpose.” The act itself, less this
statement, is codified at 22 U.S.C. 611-21.)
35. On February 7, 1989, Alan Thomson, director of the National Council of
American-Soviet Friendship, was arrested in connection with a bank deposit of
$17,000 in cash. The money was allegedly provided by the Soviet Society for
Friendship and Cultural Relations with Foreign Countries, which, according to an
FBI report, is directed by the International Department of the Central Committee of
the Communist party of the Soviet Union. While this may be the first arrest in the
United States connected with a Soviet active measures or covert action operation,
the actual charge centered on illegal avoidance of the requirement to report large
cash transactions rather than failure to register as a foreign agent. (Associated Press
story. February 8, 1989; W a sh in g to n P o st. February 8, 1989, p. A4.)
36. U.S. Senate, Select Committee on Intelligence, T he F B I a n d C IS P E S : R e p o rt,
101st Cong., 1st sess., 1989, S. Rept. 101-46, p. 21. A longer investigation of
CISPES was subsequently undertaken by the FBI on other grounds.
37. In 1983, the Levi domestic security guidelines were superseded by new
“ Guidelines on General Crimes, Racketeering Enterprise and Domestic Security/
Terrorism Investigations” promulgated by Attorney General William French Smith.
The text of both guidelines may be found in Roy Godson, ed.. In te llig e n c e R e ­
q u ire m en ts f o r th e 1 9 8 0 's: D o m e stic In te llig e n c e (Lexington, Mass.: D.C. Heath
and Co., 1986), pp. 225-64.
38. Levi, Domestic security guidelines, para. I.A., ibid., p. 225.
39. The Levi domestic security guidelines do not say precisely this; rather, they
state that “ all investigations undertaken through these guidelines shall be designed
and conducted so as not to limit the full exercise of rights protected by the Con­
stitution and laws of the United States” (para. II.B). This leaves open whether the
mere act of investigation is regarded as limiting free speech by chilling it; if so, then
an investigation based only on the individual’s or group’s protected speech would
violate this provision. A proposed charter for the intelligence agencies, introduced
in 1978, but not enacted, would have spelled this out specifically: “ No intelligence
activity may be directed against any United States person solely on the basis of such
person’s exercise of any right protected by the Constitution or laws of the United
States” (S. 2525, 95th Cong., 2d sess., the National Intelligence Reorganization
and Reform Act of 1978, sect. 241. The context, however, is the regulation of the
collection of foreign intelligence and counterintelligence information, rather than
domestic security information, with which this legislation did not deal).
40. L a ir d v. T a tu m , 408 U.S. 1 at 10, 13-14.
41. Robertson, “ The Study of Intelligence in the United States,” p. 18.
42. v. O h io , 395 U.S. 444 at 447 (1969). Emphasis supplied. B r a n ­
B ra n d e n b u rg
d en b u rg may be viewed as the culmination of a series of cases, D e n n is v. U n ite d
S ta te s , 341 U.S. 494 (1951), Y a tes v. U n ite d S ta te s, 354 U.S. 298 (1957), S c a le s
v. U n ite d S ta te s, 367 U.S. 203 (1961), and N o to v. U n ite d S ta te s, 367 U.S. 290
(1961), in which the Supreme Court evolved this rule and endowed it with consti­
tutional authority.
43. S c a le s v. U n ite d S ta te s, 367 U.S. 203 at 220 (1961). The Supreme Court is
paraphrasing and (with respect to the material inside the quotation marks) citing the
trial judge’s instructions to the jury; the Supreme Court later goes on to uphold these
instructions and state that the “ statute was correctly interpreted by the two lower
courts” (ibid, at 224).
44. John T. Elliff, T h e R e fo rm o f F B I In te llig e n c e O p e ra tio n s (Princeton, N.J.:
Princeton University Press, 1979), p. 57.
45. Shlomo Gazit and Michael Handel, “ Insurgency, Terrorism and Intelligence”
in In te llig e n c e R e q u ire m e n ts f o r th e 1 9 8 0 's: D o m e stic In te llig e n c e , p. 134.
46. The precise standard (the “ threshold” ) varies in the different guidelines. Under
the Levi domestic security guidelines, the FBI could not use an informer unless it
already possessed “ specific articulable facts giving reasons to believe that . . . a
group is or may be engaged in activities which involve the use of force or violence
and which involve or will involve the violation of federal law” (sec. II[I], pp.
227-28).
47. This, and the two subsequent, citations are from John M. Walker, Jr., “ Dis­
cussion,” in In te llig e n c e R e q u ire m e n ts f o r th e I 9 8 0 ’s: D o m e stic In te llig e n c e , pp.
194-96.
48. Testimony of Laurence H. Silberman, July 18, 1978, in U.S. Senate, Select
Committee on Intelligence, N a tio n a l In te llig e n c e R e o rg a n iza tio n a n d R e fo rm A c t o f
1978: H ea rin g s, 95th Cong., 2d sess., 1978, Committee Print, p. 616.

49. U.S. Senate, Select Committee to Study Governmental Operations with Re­
spect to Intelligence Activities (“ Church Committee” ), F in a l R ep o rt, B o o k II:
In te llig e n c e A c tiv itie s a n d th e R ig h ts o f A m e ric a n s, 94th Cong., 2d sess., 1976, S.
Rept. 94-755, p. 166.
50. See ibid., pp. 225-40, for cases in which purely political information was
collected and disseminated to administration political operatives.
51. C a n a d ia n S e c u rity In te llig e n c e A c t. Assented to June 28, 1984, Ottawa: Acts
of the Parliament of Canada, 2d session, 32d Parliament, 32-33, Elizabeth II,
1983-84, sects. 2, 12, as cited in Atkey, “ Security Intelligence Review Commit­
tee,” pp. 10-12.

7 T w o V iew s o f In te llig e n c e

1. Sun Tzu, T h e A r t o f W ar, trans. Samuel B. Griffith (Oxford: Oxford University

Press, 1963), pp. 144-45 (chap. 8, paras 3-4).
2. Ibid., p. 146 (chap. 8, paras. 9-10).
3. Ibid., pp. 77-78 (chap. 3, paras. 4-6). This view may be contrasted with that
of Carl von Clausewitz, for whom the destruction of the enemy’s fighting forces is
crucial. O n W ar, ed. and trans. by Michael Howard and Peter Paret (Princeton,
N.J.: Princeton University Press, 1976), p. 90 (bk. 1, chap. ii).
4. William E. Colby, “ Intelligence in the 1980s,” T h e In fo rm a tio n S o c ie ty 1, no.
1 ( 1981 ):53—54.
5. Ibid., p. 65.
6. On July 11, 1941, President Roosevelt ordered the establishment of “ the
position of Coordinator of Intelligence, with authority to collect and analyze all
information and data, which may bear upon national security; to correlate such
information and data, and to make [it] available to the President and to such
departments and officials . . . as the President may determine, . .” (Presidential
Order, “ Designating a Coordinator of Information,” para. 1). The order is re­
printed in Thomas F. Troy, D o n o v a n a n d th e CIA (Washington, D.C.: CIA Center
for the Study of Intelligence, 1981), p. 423.
 7. Colby, “ Intelligence in the 1980s,” p. 54.
8. See Ray S. Cline, T h e C IA : R e a lity vs. M y th , rev. ed. (Washington, D.C •
Acropolis Books, 1982), pp. 78, 80-81, for a discussion of the inability of the
research and analysis branch of the OSS to fulfill the analytic function of a central
intelligence operation. Cline states that “ [no] component of the OSS ever used
signals intelligence in its reporting” (p. 78). Thus, it lacked access to the major
intelligence sources (Ultra and Magic) available to the U.S. government.
9. Sherman Kent, S tra te g ic In te llig e n c e f o r A m e ric a n W o rld P o lic y (Princeton,
N.J.: Princeton University Press, 1949; reprint with a new preface, 1966), p. 155.
10. Ibid., pp. 3-4. Emphasis supplied.
11. Colby, “ Intelligence in the 1980s,” p. 59. The “ Club of Rome” produced a
series of studies in the late 1970s that projected demographic, economic, and
environmental trends into the twenty-first century.
12. According to Alexander Orlov, a defector from the Soviet NKVD (a prede­
cessor organization of the current KGB), analysis, or evaluation, in the Soviet
context “ concerns itself more with establishing the authenticity of the stolen doc­
uments rather than with the significance of the information. The political signifi­
cance of the information is evaluated principally by the policymaking members of
the government and the Party Presidium [Politburo].” Alexander Orlov, H a n d b o o k
o f In te llig e n c e a n d G u errilla W a rfa re (Ann Arbor; University of Michigan Press,
1965), p. 187.
13. Eliot A. Cohen, “ Analysis,” in In te llig e n c e R e q u ire m e n ts f o r th e 1 9 9 0 s: C o l­
ed. Roy Godson (Lex­
lectio n , A n a ly sis, C o u n te rin te llig e n c e a n d C o v e rt A c tio n ,
ington, Mass.: D.C. Heath & Co., Lexington Books, 1989), p. 83.
14. Kent, “ Preface to the 1966 Edition,” S tra te g ic In te llig e n c e , p. xxiv.
15. Ibid., pp. 58-59. Emphasis in original.
16. See, for example, Stansfield Turner’s suggestions in his intelligence memoir,
(Boston: Houghton Mifflin, 1985;
S e c re c y a n d D e m o c ra c y : T he C IA in T ra n sitio n
New York: Haiper & Row. 1986). After proposing the creation of an Open Skies
Agency that would release satellite reconnaissance information to the world at large,
he concludes:
Our intelligence capabilities arc suited to this special role on behalf of our own security
and the welfare of all mankind, because we have reconciled the necessary secrecy of
intelligence to the democratic processes on which our government is founded. . . . we
have opened vast new opportunities to demonstrate the superiority of our democratic
system through the employment of our intelligence capabilities to serve not only our
nation, but the rest of the world and all mankind (p. 285).
17. Colby, “ Intelligence in the 1980s,” p. 69.
18. Kent, S tra te g ic In te llig e n c e , p. viii. (Emphasis supplied.)
19. Turner, S e c re c y a n d D e m o c ra c y , pp. 279-85. Turner does not state clearly
how far he would go in making U.S. technical collection capabilities, or the infor­
mation obtained by means of them, available to the entire world. While he notes that
we " m ig h t " (emphasis supplied) begin by providing only “ intelligence collected
by systems so old that they were no longer a mystery to the Soviets” (p. 282), he
also believes that “ it will not be long before we reach a point where all satellite
photography will be so good that the differences between various models of satel­
lites will be insignificant” (p. 280).
20. This, and the subsequent, citation are from Colby, “ Intelligence in the 1980s,”
p. 69.
21. Of course, committing treason against a tyrannical or aggressive regime is
defensible in moral terms. In the case of an Oleg Penkovskiy, the motivation was
of such an ideological character, and Western intelligence officers did not have to
manipulate or suborn him in any manner. In other cases, however, where the
motivation is money or excitement, the intelligence officer often must play on these
weaknesses to help the potential spy succumb more fully to them. For a fuller
discussion of these issues, see E. Drcxcl Godfrey, “ Ethics and Intelligence,”
F o re ig n A ffa irs 56, no. 3 (April 1978):624^I2, and a response to it by Arthur L.
Jacobs in “ Comments and Correspondence,” F o re ig n A ffa irs 56, no. 4 (July
1978):867—75.
22. Sun Tzu, T he A r t o f W ar, pp. 63-64 (chap. 1, paras. 2-4).
23. Ibid., p. 144 (chap. 8, para. 2).
24. Henry L. Stimson and McGeorge Bundy, O n A c tiv e S e rv ic e in P e a c e a n d W ar
(New York: Harper Bros., 1948), p. 188. Stimson probably did not make this
famous and oft-quoted remark at the time; he expressed the thought later on, in the
course of defending his decision both as appropriate to 1929, when “ the world was
striving with good will for lasting peace, and in this effort all the nations were
parties,” and as not inconsistent with his later support (as secretary of war) for
cryptanalytic efforts (ibid). However, it is unlikely that the World War II successes
could have been achieved without an ongoing effort. For example, the successful
U.S. attack on the Japanese Purple enciphering machine, first used in 1939, de­
pended critically on the fact that the previous Red machine, dating from 1931, had
already been mastered. See Ronald Lewin, A m e ric a n M a g ic : C o d e s, C ip h e rs a n d
the D e fe a t o f J a p a n (New York: Farrar Straus Giroux, 1982), pp. 42-43.

25. Consider, for example, David Kahn’s discussion of this issue:

Immanuel Kant, in his book. P e r p e t u a l P e a c e , stated that spying is a kind of crime against
the international order because if discovered, it causes international difficulties. But this
doesn’t seem to happen with SIGINT (I n t e l l ig e n c e R e q u i r e m e n ts f o r t h e 1 9 8 0 ’s : C l a n ­
d e s t in e C o lle c tio n , ed. Roy Godson (Washington, D.C.: National Strategy Information
Center, 1982J, p. 120).
In fact, however, when, in 1931, Herbert Yardley, the head of the Black Chamber,
revealed that the United States had read the encrypted messages between Tokyo and
its delegation to the Washington Naval Arms Limitation talks (1921-22), it caused
quite a sensation in Japan (T h e A m e ric a n B la c k C h a m b e r [Indianapolis: The Bobbs-
Merrill Co., 1931; reprint, New York: Ballantine Books, 1981], pp. 187-211).
Similarly, E. Drexel Godfrey, Jr., a former CIA officer, argues that “ photo­
graphic and audio satellites and other interception devices are immensely expensive,
but they have the advantage of doing only minimal damage to the ethical standards
of the operators and processors” ( “ Ethics and Intelligence,” p. 637).

8. T o w a rd a T h eo ry o f In te llig e n c e

1. As noted in chapter 1, this threefold description of intelligence as knowledge,

activity, and organization is taken from Sherman Kent’s S tra te g ic In te llig e n c e f o r
A m e ric a n W o rld P o lic y (Princeton, N.J.: Princeton University Press. 1949; reprint
1966). Kent, however, does not regard the denial of information to others as a major
component of intelligence.
2. Kent, S tra te g ic In te llig e n c e , pp. 3-4.
3. If during time of war it is necessary to use clandestine or technical means (such
as agents’ reports or communications intercepts) to learn about weather conditions
over enemy territory, then the term “ meteorological intelligence” might be used to
describe the resulting information. However, the “ intelligence” part of the term
clearly refers to the methods by which the raw data are obtained, not the meteoro­
logical knowledge that allows it to be evaluated or that permits forecasts to be made
on the basis of it.
4. As Willmoore Kendall wrote in 1949, in an important review of Kent’s book:
“ The course of events is conceived [by the ‘state of mind’ reflected in Kent’s book]
not as something you try to influence but as a tape all printed up inside a machine;
and the job of intelligence is to tell the planners how it reads” ( “ The Function of
Intelligence,” W o rld P o litic s 1, no. 6 [July 1949]:549).
5. Kent, S tra te g ic In te llig e n c e , p. viii. Emphasis supplied.
6. William E. Colby, “ Intelligence for the 1980s,” T h e In fo rm a tio n S o c ie ty 1,
no. 1 ( 1981 ):59.
7. For this formulation of the relationship between intelligence and open-source
information, 1 am indebted to Michael Herman of Nuffield College, Oxford, w'ho
proposed it at a panel of the 1989 Convention of the International Studies
Association.
8. Avi Shlaim, “ Failures in National Intelligence Estimates: The Case of the Yom
Kippur War,” W o rld P o litic s 28, no. 3 (April 1976):355.
9. Kent, S tra te g ic In te llig e n c e , pp. 206-7.
10. According to Kent, this is known as “ spot intelligence,” or, less respectfully,
“ Information Please” (ibid., pp. 28-29). An intelligence service like the CIA thus
serves as a reference service, similar to the function performed for the U.S. Con­
gress by the Congressional Research Service of the Library of Congress.
11. William E. Colby, H o n o ra b le M e n : M y L ife in th e C IA (New York: Simon &
Schuster, 1978), p. 245.
12. Ray S. Cline, T h e C IA : R e a lity vs. M y th , rev. ed. (Washington, D.C.: Acrop­
olis Books Ltd., 1982), p. 175.
 INDEX

Allende, Salvador, 87-88 Castro, Fidel, 89, 133-34, 147. 207 n3

American Association for the Advance­ Church Committee (1975)
ment of Science, 105 and CIA, 207 n4
Analysis of intelligence information, 2, and covert action, 146—47
7, 8, 37-71, 77. S e e a l s o Photo in­ and domestic intelligence investiga­
terpretation tions, 133-34, 155-57
cryptanalysis, 38-45, 186 n32 National Intelligence Estimate on So­
forecasting, 8, 37, 196 n63 viet offensive nuclear forces (1969),
National Intelligence Council, 182 n3 65, 166
technical analysis, 37 and plausible denial, 201 n44
technical assistance, 76 Churchill, Winston, 81, 95, 198 nl3
telemetry analysis, 45-47 CIA (Central Intelligence Agency), 36,
U.S. access to Soviet telemetric in­ 55, 95, 128, 179
formation, 46 analysis of information and covert ac­
Anglcton, James J., 178 tion, 8, 54, 74, 86, 161-63
Assassination. S e e Castro, Fidel; CIA and assassinations, 89-90, 133-34
assistance to Italy, 75
and Church Committee, 207 n4
Bittman, Ladislav, 82 and Congress, 135-36
Board for International Broadcasting, 84, covert funding, 84, 86
93 and Cuban double-agent operations,
Brzezinski, Zbigniew, 60 115
“current events syndrome.” 55, 193
n40
Canada, 111 failures of, xv, 36, 56-57, 59-60, 63,
Canadian Security Intelligence Review 133-34, 205 n26
Committee (SIRC), 147-48 Intelligence Directorate of, 196 n66
Canadian Security Intelligence Service and Soviet forgeries, 82
(CS1S), 147, 158 submission of this book to, 181 n4
and domestic intelligence activities. 158 Classification of information, 100-5
Privy Council. 147-48 Atomic Energy Act (1954), 104
Royal Canadian Mounted Police, 147 communications security, 105, 116-17
Carter, Jimmy, 24, 60, 88-90 and defense budgets, 101—2

217
Classification of information ( c o n t . ) correlation with skill at chess, 189 n9
National Security Act (1947), 100, 168 history of, 39-42, 169
overclassification, 102-3
Patent Secrecy Act (1952), 105
public accountability for actions, 103-4 Deception operations, 2, 8, 82, 118-23.
technological information, 104-5 S e e a l s o Propaganda
Cline, Ray, 86-87 blocking intelligence-gathering chan­
Colby, William, 160-63, 165, 167-68, nels, 121
174, 178 false signals, 120-22
Collection of intelligence information feedback, 121-22
activities, 7-8, 11, 92, 125 disinformation, 81-82, 85
counterespionage, 8, 109-15, 121 and intelligence failures, 119-29
covert action, 7, 73 and self-deception, 122-23
denying certain information to adver­ true signals, 120-22
saries, 8, 182 n2 Domestic intelligence, 4, 148, 172
implementation of policy, 8 and constitutional law, 148-52, 209
motivations to commit espionage, 16- n26
17, 108 criminal standard. 149-58
nation’s foreign policy, 73, 171 and counterterrorism, 154-56
and nature of target, 18 and personal security, 152-53
“positive” foreign intelligence, 110— support groups, 153-56
11 Foreign Agent Registration Act
Counterdeception, 123-28 (FARA), 150-51, 210 n34, 210 n35
German use of against British in Neth­ Foreign Intelligence Surveillance Act
erlands, 123-24 (1978), 148-49
U.S.-U.K. tunnel operation (Berlin), Levi guidelines, 149-51, 153, 156,
126-27, 206-7 n47 210-11 n37, 211 n39, 211-12 n46
Counterespionage, 109-15 and national security cases, 148,
Counterintelligence, 7-9, 99-129, 166, 150, 209 n28
173, 202 nl, 202 n2 violation of privacy, 151-52
Coup d’état. S e e Covert action Double agents (of U.S.S.R.)
Covert action, 3, 8, 73-97 Blake, George, 126
coup d’état, 60. 73, 76, 87, 88 Burgess, Guy, 17, 184 n9
Foreign Assistance Act (1961), 197 and defectors, 20
n2, 198 n9 Hermann, Rudolf, 15
forgery, 85, 89, 91 Howard, Edward Lee, 17, 111, 204
Hughes-Ryan Amendment, 74, 93, 197 n25
n2, 200-1 n38 MacLean, Donald, 17, 184 n9
information and disinformation, 81-83, Pclton, Ronald, 111, 204 n25
85 Penkovskiy, Oleg, 16. 85, 214 n21
and plausible denial, 92-94, 132-34 Philby, Harold (“Kim”), 17, 184 n9
propaganda, 8, 83-85, 91 Prime, Geoffrey, 107
“special political action,” 74, 197 n5 Double-agent operations. 2, 18, 111-15,
Cryptanalysis, 25-27, 37, 38-42. S e e 120, 198 n16
a l s o Double-Cross System; Ger­ Yurchenko, Vitaliy, 111, 204 n25
many; Japan; United Kingdom Double-Cross System (U.K.), 18, 113-
British skills in WWII, 25, 43-45, 15, 119, 120, 122-23, 127. S e e a l s o
114, 122 Cryptanalysis; United Kingdom
fragility of, 42-45
history of, 39-42
mistakes in using cipher, 189 nlO Eisenhower, Dwight David, 22-23, 92-
U.S. breaking Libyan codes. 44, 102 93
Cryptology/cryptography and Allen Dulles. 18
Elliff, John, 153 11-20, 166. S e e a ls o Intelligence
Encryption, 2, 38. S e e a l s o Germany; officers
Japan; United Kingdom compared to technit, 30-33
cipher, 38, 40, 116 “dead drop.“ 20, 109
codebook, 38, 40 and deception, 127-28
C r y p t o lo g i a , 188-89 n8 defectors, 20. 83, 111, 127-28
data encryption standard. 188-89 n8 nonofficial cover officials (NOCs), 12-
fiber-optic transmission, 32, 117 14
microwave transmission, 116-17 official cover, 12-13, 110
scrambling devices, 116 personality reports, 15, 35
in World War 11, 41-43 proprietary cover, 183-4 n6
and Zimmerman telegram (1917), 40 quality control. 17-18
“source,” 11-12
Soviet use of against U.S. and U.K.,
Fairbanks, Charles, 140-41 16-17
FBI (Federal Bureau of Investigation), surveillance, 109-10
5-6, 149-50, 156-57 “tradccraft,” 19-20
Ford, Gerald, 89-90
France
DGSE (Direction Générale de la Secu­ Influence actions of foreign governments,
rité Extérieure), 93 77-80, 82
and German Enigma coding machine, Inman, Admiral Bobby Ray, 105
32 Intelligence
and R a i n b o w W a r r io r incident, 89, 90, consumers, 165, 173, 176
93 definition of, 1, 171
Secret Service (SDECE), 17 and economics, 6-7, 53
Soviet, 192 n34, 203 nl 1
development and deployment, 172
Gates, Robert M., 135, 143, 144-45 and law enforcement, 4-6
Gazit, Shlomo, 154-55 informers, use of, 5, 33
Germany (West), 15, 80, 81 military, 51-52
Abwehr, 114 political, 52
counterdeception, 123-24 recruited source of, 15-17, 183 n3
Enigma coding machine, 32, 41,42, relationship between intelligence and
45 the policymakers, 165, 176
intelligence failures in world wars, 18, terrorism, 5, 19, 33, 150, 154-56
25, 45, 59, 61, 113-15, 119-20, walk-in source of, 15-17
122, 127 Intelligence activities, 1, 74, 126, 168.
and Soviet targeting, 15, 86 S e e a l s o Cryptanalysis; Cryptology/
technical developments, 21, 27, 50 cryptography; Encryption
U.S.-U.K. tunnel (Berlin), 26, 127 collection and analysis of information,
Zimmerman telegram (1917), 40 2. 28-29, 177
Grenada, 34, 49 counterintelligence analysis, 2, 82,
GRU (U.S.S.R. Committee for Military 128-29, 166
Intelligence) decryption of coded messages, 2, 32,
double agent in, 83, 111 38, 39
use of moles, 15-17 electromechanical cipher machines, 32,
41,42, 45. 206 n44, 214 n24
espionage, 2. 80, 176
Haig, Alexander, 143 indications and warnings (I&W), 55-
Halperin, Morton. 146 57
Helms. Richard, 65 interception of communications, 2,
Human intelligence collection (humint), 24-27, 116-17
Intelligence activities ( c o n t.) “national technical means” of verifica­
radio traffic, 27, 41, 74, 116-17, 127 tion, 29-30, 187 n40
Knickbein system. 50, 192 n32 Outer Space Treaty, 29
research and analysis, 2, 164 restrictions on overflying another coun­
struggle between adversaries, 2, 161, try without consent, 22, 29
175, 177 Iran, 81, 83, 84, 90, 93
Intelligence failures seizure of U.S. Embassy in Tehran, 90
causes of failure, 62-67, 191-92 n30, Soviet intentions toward, 140^41
195 n57 U.S. “listening posts” in, 46
improper security checks, 206 n40 “Iron Curtain,” 18
mirror-imaging, 64-67, 71, 195 n56 Israel, 58, 59, 64-65, 154, 175, 195 n57
and deception operations, 119-29
solutions to intelligence failure, 67-71
types of failure, 59-61 Japan, 83, 169
Intelligence information attack on Pearl Harbor, 59, 63, 121
analysis and assessment of, 1, 2, 161, cryptanalysis, 42-43
178 “Purple” coding machine, 41
centrality of, 165, 169
free flow and exchange, 165
verification, 165 Kennedy, John, 24, 134
Intelligence officers, 11-12, HO. 183 n2 Kent, Sherman, 54, 162, 164, 181 nl
Foreign Service Reserve (FSR) sta­ S t r a t e g i c I n t e l l ig e n c e f o r A m e r i c a n
tus for, 183 n5 W o r ld P o li c y , 172, 191 n29, 215 nl
handler, 12 KGB (U.S.S.R. Committee for State Se­
illegal officers, 14 curity), 111. S e e a l s o Double agents
“legal” officers, 14 (U.S.S.R.)
“mailbox,” 12 acquisition of Western technology, 6,
“mole,” use of, 15 16
“station,” 13 Cheka, 119
Intelligence organizations, 1. S e e a ls o and Iran, 81, 82
CIA; FBI; KGB and moles, 15
British, 95, 113, 123, 126 NKVD, 59
characteristics of, 2-3 Khrushchev, Nikita, 24, 85, 93
clandestine collection capabilities of, Korean War (1950), 22, 59. 61
182 n6 Krcmlinology, 35
French, 17, 93
rivalries between, 183-84 n8
undercover agents, use of, 3, 15 Levchenko, Stanislav, 83
variations among, 58-59 Levi, Edward, 149-50, 153, 156. See
Intelligence “product,” 2, 37, 49, 53-71 a l s o Domestic intelligence
N a t i o n a l I n t e l l ig e n c e D a i l y (NID), 54 Libya, 44, 92, 102
national intelligence estimate (NIE),
58, 140
P r e s i d e n t 's D a i l y B r i e f , 54-55 Management of intelligence
special national intelligence estimate authorization issues, 134-35
(SNIE), 58, 141 democracy and secrecy, 144-48
International law. S e e a l s o SALT expertise and policy, 136-44
ABM Treaty, 29 funding. 134-35, 144
and diplomatic immunity, 12 history of, 134-35
Intermediate Range Nuclear Forces and independence of intelligence, 136.
(INF) Treaty, 36 140. 141-44
International Trade in Arms Regula­ paper trail, 133, 135
tions, 203 nl7 plausible denial doctrine, 132-34
 and policymakers, 134-35, 137-45 “Open Skies” plan, Soviet rejection
secrecy and control, 131-34 of, 22-23, 213 nl6
and intelligence, 174 Photo interpretation (PI), 21-24, 37 4 7 _
Masterman, John, 113 48, 124-25, 205 n38
McDonald, Admiral Wesley, 49 ground resolution distance, 47, 124
“Météorologie intelligence,” 215 n3 190 n25
Meyer, Cord, 54-55 and “Open Skies” plan, 22
Mirror-imaging. S e e Intelligence failures Propaganda, 83-85, 91
Multidisciplinary counterintelligence
(MDCI), 115-18
communications security, 116-17 Radio Free Europe, 83, 93
emanations security, 117-18 Radio Liberty, 83, 93
Radio Moscow, 76, 93
Ranelaugh, John. 126
National security, and the type of govern­ Reagan, Ronald, 44, 46, 89-90. 94
ment, 3-4 Iran-Contra affair, 93
National Security Agency (NSA), 54.
S e e a l s o Classification of informa­
tion SALT (Strategic Arms Limitations Talks)
National Security Council (NSC). 75 MI, 24, 29-30, 46. 57
National Security Decision Directive. 77, antiencryption clause in, 190 n22
197-98 n7 and national technical means (NTM),
National technical means (NTM). S e e 29-30
SALT Second Common Understanding to,
Nicaragua, 75, 89, 94. 145, 147 190 n21
Nixon, Richard, 65 violation of, 93
Noriega, Manuel, 90 Security, 99, 105-9, 120. S e e a l s o Coun­
terintelligence
personnel security, 105-8
Office of Strategic Services (OSS), 161 polygraphs, 107-8, 204 n21, 204
OPEC (Organization of Petroleum Ex­ n22
porting Countries), 34 Privacy Act (1974), 107
Open sources physical security, 108-9
collecting and cataloging information Shultz, George, 107
from. 34, 48-49, 175 Signals intelligence (sigint), 24-27, 213
data banks, 48-49 n8, 214 n25
diplomatic and attache reporting, 35- communications intelligence (comint),
36 25-27, 125
and intelligence agencies. 187-88 n47 compromise of, 107
predictions from, 34, 57-58, 192 n34 and cryptanalysis, 25-27. 186 n32
publications and broadcasts, 7, 33-35, electronic intelligence (elint), 25, 27
166 nucint, 185 n29
Organization of East Caribbean States, 49 telemetry intelligence (telint), 25, 27,
37, 45-46
“Tempest,” 185-86 n30
“Paradigm shift,” 195 n55 Silberman. Laurence, 157
Pathd, Pierre-Charles, 80, 82 Soviet Union. S e e U.S.S.R.
Photographic/imagery intelligence (phot- Surveillance, 19-20, 109-10, 210 n31
int or imint), 21-24, 47-48. 205 “brush pass,” 19-20, 109
n38 Foreign Intelligence Surveillance Act
aerial photorcconnaissance, 22-24, (1978), 148
28-29, 118, 120-21, 124-25, 133, Freedom of Information Act, 210 n32
190 n24 satellite, 28-29. 47-48, 206 n42
Technical intelligence collection (tech- use of double-agents against U.S.S.R.,
int), 20-36, 107, 121, 165, 169, 111, 204 n25
214 nl9. S e e a l s o Photographic/ U.S. Congress, 78
imagery intelligence; Signals intelli­ Boland Amendments, 145
gence House of Representatives Permanent
compared to humint, 30-33 Select Committee on Intelligence,
and deception operations, 121, 178 135, 136, 141, 144-45
emission control (emcon), 121 Hughes-Ryan Amendment, 74, 93, 197
fiber-optic/laser-beam communications, n2, 200-1 n38
32, 117 “leaks,” 146, 209 n2l, 209 n23
satellite reconnaissance, 23-24, 28-29, National Intelligence Reorganization
118, 120-21, 124-25 and Reform Act (1978), 211 n39
sensors, 28 Senate Select Committee on Intelli­
“signature,” 33, 47-48 gence, 144. 181 n4, 194 n50
Terrorism, 5, 75, 89 U.S. Department of State, 75, 85
Truman. Harry (S), 100, 164 biographic materials published by, 13
Turner, Admiral Stansfield, 30-31, 60, Black Chamber, 27, 41, 169, 186 n35,
90 214-15 n25
Tzu, Sun, 159-60, 168-69 Bureau of Intelligence and Research
(INR), 54, 68
U.S.S.R., 35, 56, 86. S e e a ls o KGB;
U-2 reconnaissance plane, 23-24, 92, GRU
166 in Afghanistan, 89, 91, 94
United Kingdom, 8, 16, 107, 134. S e e agents of influence, types of, 80
a ls o Cryptanalysis; Double-Cross and “black” propaganda, 83-84
System disinformation, 81-82, 85
British Government Communications embassy of, in Washington, D.C., 117
Headquarters (GCHQ), 107 glasnost, 31, 33, 53, 203 n9
communications intelligence break­ and “gray” propaganda, 84
throughs, 25, 27, 40, 50 Khrushchev, Nikita, secret speech of,
and D-Day landings, 18, 59. 61, 114. 85, 93. 199 n22
119, 122-23, 127 missiles in Cuba, 48, 191 n27, 191
and development of infrared and radar, n28
24 and use of front groups, 85
Government Code and Cypher School,
44
MI5, 113 Vance, Cyrus, 60
MI6, 95, 123
and not invented here syndrome. 66-
67 Wilson, Woodrow, 40. 41, 144
Special Operations Executive (SOE), World Peace Council, 76, 85
123 World War II. S e e a l s o Cryptanalysis;
use of aerial photography, 21 Encryption; Germany; Japan; United
United States Kingdom
Defense Intelligence Agency (DIA), British cryptanalysis skills, 25, 43-45,
54, 68. 191 n27 114, 122
Department of Defense (DoD), 101-2 C h ic a g o T r ib u n e article on Japanese
embassy in Moscow, 90, 183 n4 codes, 42-43
intelligence community. 193 n43 Double-Cross System, 18, 113-15,
Iran-Contra affair, 93-94 119, 120, 122-23, 127
Secret Service, 155-56
support for Nicaraguan resistance. 75,
89 Yardley, Herbert, 40
 ABOUT THE AUTHOR

Abram N. Shulsky, senior fellow at the National Strategy Information

Center (NSIC) in Washington, has served as the director of Strategic Arms
Control Policy in the Pentagon and as minority director (Democratic) of the
Senate Select Committee on Intelligence staff. He has been a consultant to
the President’s Foreign Intelligence Advisory Board and the acting repre­
sentative of the Secretary of Defense at the U.S.-USSR Nuclear and Space
Talks in Geneva. Dr. Shulsky is also the author of several articles on
intelligence and related national security matters.