STUDIES IN INTELLIGENCE [Vol. 23 No.3, Fall 1979]

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP80-00630A000100080001-1
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 59
• Document Creation Date: December 16, 2016
• Document Release Date: December 22, 2004
• Sequence Number: 1
• Publication Date: January 1, 1979
• Content Type: STUDY

Approved For Release 2905/01/06 : CIA-RIpP80-0063OA000100080001-1
cTUDIES
IN
'ELLI,GENCE
CENTRAL INTELLIGENCE AGENCY
VOL. 23 No. 3 FALL 1979
TR-SINT 79-003
0824
Approved For Release 20@5/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/q
Materials in the Studies are in general to be reserved to U.S.
personnel holding appropriate clearances. The existence of this journal
is to be treated as information privy to the U.S. official community.
All copies of each issue beginning Summer 1964 are numbered
serially to facilitate control, accountability and recall. Copies which
are no longer needed by recipients should be returned to the Editor's
office, 1036 C of C Bldg., or may be destroyed if a notice showing
Volume number, Issue number, and Copy number is sent to the
Editor,
25X1
25X1
All opinions expressed in the Studies are those of the authors.
They do not necessarily represent the official views of the
Central Intelligence Agency or any other component of the
intelligence community.
Warning Notice
Sensitive Intelligence Sources and Methods Involved
NATIONAL SECURITY INFORMATION
Unauthorized Disclosure Subject to Criminal Sanctions
25X1
25X1
Approved For Release 2005/01j106 : CIA-RDP80-og630A000100080001-1
 Approved For Release 200 P80-00630A000100080001-1
STUDIES IN INTELLIGENCE
Articles for the Studies in Intelligence may be
written on any theoretical, doctrinal, operational, or
historical aspect of intelligence.
The final responsibility for accepting or rejecting
an article rests with the Editorial Board.
The criterion for publication is whether or not, in
the opinion of the Board, the article makes a contribu-
tion to the literature of intelligence.
GEORGE ALLEN
HELENE L. BOATNER
MAURICE C. ERNST
DONALD P. GREGG
DONALD E. SMITH
JAMES H. TAYLOR
Additional members of the Board are
drawn from other CIA components.
25X1
Approved For Release 200q/01/06 : CIA-IRDP80-00630A000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Contributions to the Studies or communications to the editors may come from
any member of the intelligence community or, upon invitation, from persons outside.
Manuscripts should be submitted directly to the Editor, Studies in Intelligence, Room
1036 Chamber of Commerce Building and need not be coordinated
through channels. They should be typed in duplicate, double-spaced, the original on
bond paper. Articles may be classified through Secret. Supplements, separately
distributed, can accommodate articles of higher classification.
For inclusion on the regular Studies distribution list call your office dissemination
center or the Center for the Study of Intelligence, For back issues and on
other questions, call the Office of the Editor, erox-quality copies of back
articles can be obtained from Document Services Branch, Office of Central Reference,
I Iwhere a complete file of Studies is available on microfiche.
THE STUDIES IN INTELLIGENCE AND
SHERMAN KENT AWARDS
An annual award of $500 is offered for the most significant contribution to the
literature of intelligence submitted for publication in the Studies. The prize may be
divided if the two or more best articles submitted are judged to be of equal merit, or it
may be withheld if no article is deemed sufficiently outstanding. An additional $500 is
available for other prizes.
Except as may be otherwise announced from year to year, articles on any subject
within the range of the Studies' purview, as defined in its masthead, will be
considered for the awards. They will be judged primarily on substantive originality
and soundness, secondarily on literary qualities. Members of the Studies editorial
board and staff are of course excluded from the competition.
The editorial board will welcome readers' nominations for awards but reserves to
itself exclusive competence in the decision.
Approved For Release 2005/0106 : CIA-RDP8T-0063OA000100080001-1
25X1
25X1
25X1
25X1
25X1
 Approved For Release 20 -RDP80-0063OA000100080001-1
CONTENTS
Fall 1979
President Truman and the Congolese SAM .......................... Dino A. Brugioni
19
Captain Harry and military nomenclature (SECRET NoFORN)
The OSS Assessment Program ........................................ Donald W. MacKinnon
21
Situational testing of overseas operations candidates
Analysis, War, and Decision ...................................................... Richard K. Betts
35
Why intelligence failures are inevitable
Intelligence in Recent Public Literature ................................................................
55
Intelligence Vignette ................................................................................................
68
Approved For Release 20Q5/01/06 : CIA-FIDP80-00630A0001000800Q,1-1
 Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
CONTRIBUTORS TO THIS ISSUE
Richard K. Betts ("Analysis, War and Decision") served on the staff of the Senate
Select Committee on Intelligence and was a staff member of the National Security
Council. He is currently a Research Associate at the Brookings Institution.
I)r. Louis W. Tordella (Lewin's "Ultra Goes to War" and Van der Rhoer's "Deadly
Magic") was for many years Deputy Director of the National Security Agency.
Approved For Release 2005/01/p6 : CIA-RDP$0-00630A000100080001-1
iv I I
25X1
25X1
25X1
25X1
 Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
5 October 1979
MEMORANDUM FOR: Recipients of Studies in Intelligence,
Volume 23 No. 3 Fall 1979 (TR-SINT 79-003)
SUBJECT: Erratum
The Fall 1979 Edition of Studies in Intelligence contains
an error in pagination. In the article "The OSS Assessment
Program" by
commencing on page 21,
pages 32 and 33 are reversed. The text properly proceeds
from the foot of page 30 to the top of page 33, then to
page 32, and concludes on page 34.
Readers are requested to staple this notice to the
blank page following the Contents page in their copy of
Studies.
The editor regrets this oversight and offers his
apologies to the author and the readers of Studies in
Intelligence.
Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
 25X1 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Next 16 Page(s) In Document Exempt
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-00630A000RR901-1
NOFORN
Captain Harry and
military nomenclature
PRESIDENT TRUMAN AND THE CONGOLESE SAM
The U.S. Presidents who followed Harry Truman in office made it a practice to
provide the former President with periodic intelligence briefings. During the Johnson
administration this assignment was usually performed by the Deputy Director of
Central Intelligence, General Marshall S. Carter. General Carter had a penchant for
injecting humor into his briefings, and Truman appreciated it.
During the early days of the 1964 Congo crisis, USAF aircraft were dispatched on
low-level reconnaissance flights over missionary outposts in an effort to determine the
whereabouts of the missionaries and to assess the condition of their missions. At the
time, there were more than 3,000 missionaries scattered throughout the Congo. A U.S.-
Belgian airborne rescue operation, which would eventually save the lives of 250
European and American hostages held in Stanleyville, was in the planning stage in the
fall of 1964.
Flying reconnaissance at low levels is always hazardous because of the threat of
small-arms fire, but we at the National Photographic Interpretation Center were
surprised to see on a photo acquired on one of these missions a Congolese peasant in
the act of throwing a spear at the reconnaissance aircraft. General Carter saw the
photograph and decided to use it in his briefing for Mr. Truman.
He told the President the photo depicted "an unrevetted surface-to-air missile
system. The missile, a Mark I, Mod 1, has a manual inertial guidance system, a mobile
launcher with a one-sling power velocity on takeoff, and considerable pucker power
on impact. It is, however, of questionable accuracy and has an undetermined C.E.P.
The refire capabilities have not yet been determined."
Truman was delighted with the presentation, but memories of his own service as
an Army officer moved him to reflect, soberly: "I hope someone has advised the pilots
to fly a little higher. Just imagine the effect on a young Air Force pilot if his career
service record showed that he had been downed by a spear."
The President paused, reflected a moment, and then smiled: "I'm sure, in that
case, you military people would make certain that the record identified the weapon a
a `SAM, Mark I, Mod 1.' " General Carter agreed.
(This article, and the accompanying photograph overleaf, are classified Secret No
Foreign Dissem.)
MORI/HR
P from
pg. 19-20
S -pved For Release 2005/01/06: CIA-RDP80-00630A0001000800d'P 1
 A. r v d For Release 2005/01/06 : CIA-RDP80-0063OA0001 00080001 -1
Ar roved For Release 2005/01/06 : CIA-RDP80-00630A000100080('ybf-41 1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
How CIA's predecessor separated sheep from goats and
launched a modern management technique
During the first year of its operation, there were three channels of entry into OSS:
recruitment of military personnel by the Personnel Procurement Branch, recruitment
of civilians by the Civilian Personnel Branch, and recruitment of both military and
civilian personnel through the initiative of individual OSS members-all of this
without benefit of any professional or uniform screening process. Nobody knew who
would make a good spy or an effective guerrilla fighter. Consequently, large numbers
of misfits were recruited from the very beginning, and this might have continued had
it not been for several disastrous operations such as one in Italy for which, on the
assumption that it takes dirty men to do dirty works, some OSS men had been
recruited directly from the ranks of Murder, Inc., and the Philadelphia Purple Gang.
The need for professional assistance in selection was obvious, but resisted by many in
the organization.
In October 1943 an OSS official back from London suggested that a program of
psychological-psychiatric assessment similar to that in the English War Office
Selection Boards be set up in the OSS. This idea was picked up and pushed by Robert
C. Tryon, psychologist on leave from the University of California, who was Deputy
Chief, Planning Staff, OSS. He recommended that an assessment center be set up in
the Schools and Training Branch and in collaboration with three other California
Ph.D.'s, James A. Hamilton, John W. Gardner, and Joseph Gengerelli, he set about
planning the first assessment center in the United States.
By November a physical facility had been acquired: the Willard Estate, the
spacious residence and grounds of the owners of the Hotel Willard in Washington
which was to become Station S (for Schools and Training, though most preferred to
think the S stood for Secret) in Fairfax, Virginia, some 18 miles from Washington.
The first planning conference for the program of Station S was held in early
December with Henry A. Murray (Harvard) and Donald Adams (Duke) joining the
California psychologists. Shortly thereafter, the director of OSS authorized the
establishment of an assessment unit, and 15 days later the first assessment was held
with a skeleton staff.
Although the push for an assessment program in OSS came from the California
psychologists, the ultimate form and nature of the program was most importantly
shaped by Henry A. Murray.
In addition to the speed with which it was initiated, the assessment program
suffered from other handicaps. Although backed by General Donovan and some of the
bureau chiefs, it was opposed by others and especially by the military. We lacked
* Adapted from "How Assessment Centers Got Started in the United States: The OSS Assessment
Program," copyright 1974 Development Dimensions, Inc., Pittsburgh, Pa. Reprinted by permission.
MORI/HR
P from pg.
21-34
Approved For Release 2005/01/06 : CIA-RDP80-0063OA0001000800c I1.1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The OSS Assessment Program
knowledge about the assignments, most of them novel, to which our assessees would be
sent. Without job analyses, we did not know what specifically we were assessing for.
We needed experts to write job descriptions, but there were none in the field. At best,
job assignments were described by single terms: language expert, cartographer, news
analyst. Had we known what specific skills would be required there would have been
so many of them as to preclude a testing of them all.
Later we would learn more about what was required for successful execution of
OSS assignments from branch chiefs who had by then received more specific job
descriptions, from reports of returnees, from assessors who had received training in the
OSS schools, and from assessors who had traveled abroad for firsthand observations.
In the beginning it was the lack of specific knowledge that led us to conclude that
our assessments could not be of the specific skills of a given candidate for a specific job
but rather in each case an assessment of the "man as a whole," the general structure of
his being, and his strengths and weaknesses for rather generally described
environments and situations. As it turned out, there were some advantages to our
having taken this stance toward assessment, for we soon discovered that very often
assessees were never assigned to the job for which they were recruited. Typically, two
to eight months elapsed between assessment and job assignment overseas. The
candidate had first to be trained and by the time that had been accomplished, the war
had moved on and the job for which he had been recruited no longer needed to he
done.
Only those destined for overseas assignment were assessed; those who remained in
the States were exempt. At first our assessment reports were only for the information
of bureau chiefs. They were free to accept or reject our recommendations as they saw
fit. After two months all that changed. By order of General Donovan only those who
received a positive recommendation from Station S could be sent overseas. This was
personally flattering, but very frustrating to our scientific egos since it meant that
anything like pure validity studies of our assessment operation could not be made. This
order meant that Station S with its three-and-a-half-day program would not be able to
assess all those destined for overseas duty. Thus it was that in late winter 1944, a one-
day assessment center, Station W, was set up in Washington to assess a large number of
candidates, many of whom were to be assigned to headquarters and rear bases
overseas rather than to operations in the field. Two months later a center to assess
candidates recruited on the West Coast was established at Laguna Beach, California.
This was Station WS. Later assessment stations to screen native agents were set up in
Ceylon, Kunming (Yunnan Province), Calcutta, and Hsian. During the period of their
operation, Stations S and W assessed 5,391 recruits.
The program for Station S, set up hurriedly and with little knowledge of what
OSS assignments would entail, was bound to undergo many changes. There were seven
periods in the history of Station S, but there were no radical changes in the program
during the last six periods (June 1944 to V-J Day, September 1945) during which time
I served as Director of Station S, and this is the program I shall describe.
I have spoken of some handicaps under which the assessment program had to
operate. Let me mention two others. From its inception, the OSS had to guard against
infiltrations by foreign agents, and, of course, if its operations were to succeed they
had to be kept secret. But these concerns hardly justified the extreme secrecy which
was maintained and which merely added to the atmosphere of cloak-and-dagger
mystery that enshrouded the organization. This was nowhere more obvious than in the
recruitment of OSS personnel. Barred from mentioning the OSS by name but free to
talk about mysterious, exciting overseas assignment with a government agency, the
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The OSS Assessment Program
pitch made by OSS recruiters was especially attractive to the bored, to the
pathologically adventuresome, to those neurotically attracted to danger, and to
psychopaths in general. Thus, we had more than our share of misfits to weed out, and
of course it is psychopaths who have a special talent to make a good impression over
brief periods of time. Under the given circumstances, the branch representatives who
briefed candidates for their visit to Station S were wary about telling them much about
the work for which they had been recruited. Many, when seen in assessment, had no
idea of what their assignment would be, and some even thought they were in the State
Department!
The other handicap we faced, or so it seemed at first, was that we would not
know whom we were assessing. That is, they would not be allowed to reveal to us or to
the other members of their assessment class their true names. In December 1943 when
Station S was established, many of the recruits into OSS were refugees from Europe,
often with families and relatives in occupied territory. If their true identity were to
become known to others in the organization who had infiltrated it they might become
subject to blackmail through threats to relatives abroad.
Not knowing who our assessees were, was not, however, entirely disadvantageous.
Quite innocently we could be a check on the misuse of nepotism, turning down even a
Roosevelt or a Donovan if his performance at Station S left something to be desired.
In any case, we had to live with the anonymity of our assessees, so what we did
was to turn it into an assessment exercise!
Assessment at Station S ran in two-week cycles, during which three groups of 18
candidates would be assessed: the first group would arrive on a Sunday afternoon and
leave the following Thursday morning; that afternoon a second group would arrive
and, assessed over the weekend, would leave Monday morning; that afternoon a third
group would arrive and leave Friday morning. That weekend the staff would draw its
breath and prepare for the next two-week cycle which would begin again on Sunday
afternoon.
Each recruit for OSS, upon arrival in Washington, would report to the Branch
representative who would supervise him throughout training and who had authority
then and there to reject the candidate if he was obviously unsuitable. Otherwise and
usually, the recruit would be told something, but not much, about the type of work for
which he had volunteered, and he would fill out a Student Information Sheet.
He was then informed that he would spend the next three and a half days at an
assessment school in the country, where he would be asked many questions and given
many tests by a group of psychologists and psychiatrists. He was told there was no
reason to feel apprehensive, that most who had gone through the school had enjoyed it.
He was urged to do his best. He was also told he would go incognito for his own
protection. And finally he chose a name by which he would be known during
assessment.
During that day in Washington he was given a medical examination, and that
afternoon he would report to Schools and Training Headquarters, an old brick school
house that stood where Watergate now stands. He would report at 3:30 if he was an
enlisted man, at 4:00 if a civilian, and at 4:30 if an officer. Reporting was staggered in
this fashion so that each might know the status of some of the men in his own group,
the fewer the better, but not the others, e.g., no officer would know who were enlisted
men and who were civilians, et cetera.
Upon arrival at Headquarters, the candidate would be taken to a room, stripped
of his clothing and anything else that might identify him, issued army fatigues and
Approved For Release 2005/01/06 : CIA-RDP80-0063OA0001000800QI-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The OSS Assessment Program
boots, and ushered into a waiting area. At 5:00 the group would be put in a completely
closed army van and driven over the eighteen winding miles to Station S.
After a brief welcome and orientation to the program, the candidates were told
that during their stay at S each would have to build up and maintain as completely as
possible a cover story for himself, claiming to have been born where he wasn't, to have
been educated in institutions other than those he had attended, to have been engaged
in work or profession not his own, and to live now in a place that was not his true
residence. The cover story was to be maintained with staff and students alike at all
times except under "X conditions," at which time candidates could reveal anything
about themselves except their name and true identity. X conditions prevailed when
filling out a Personal History Form, in a Life History Interview, when a staff member
established X conditions, and when in case of need a student asked for X conditions
and a staff member granted them.
Thus, Station S was one grand simulation, each candidate having the task of
consistently simulating a person he was not. Since many of the candidates would
eventually be living abroad under cover, the relevancy of the cover story task seemed
clear enough.
The variables rated on the Station S report for each candidate offer a rationale for
the assessment exercises which were developed for the program. The rated dimensions
were these: Motivation, Practical Intelligence, Emotional Stability, Social Relations,
Leadership, Physical Ability, Observation and Reporting, Propaganda Skills, and
Maintaining Cover. Each variable was rated on a 6-point scale: Very Inferior, Inferior,
Low Average, High Average, Superior, Very Superior. Note that to force a spread of
ratings, using more of the whole scale, no rating of Average was given. If the
dimension was not measured, that fact was indicated in the final report.
Components of each of the dimensions were printed on the face sheet in the
following manner:
Motivation: energy, zest, effort, initiative, war morale, interest in
assignment.
Practical Intelligence: speed and accuracy of judgment, resourcefulness in
solving problems.
(motional Stability: emotional control and maturity, absence of neurotic
symptoms.
Social Relations: social awareness, goodwill, teamwork, tact, absence of
annoying traits.
Leadership: social initiative, organizing ability, ability to evoke cooperation.
Ihysical Ability: agility, daring, ruggedness, stamina.
Observation and Reporting: ability to search, question, observe and recall,
infer, report.
Propaganda Skills: ability to affect others through acts, words, or displays.
Maintaining Cover: caution, ability to remain inconspicuous, bluff, mislead,
keep a secret.
If one of the components of a variable was especially high, let us say tact, it would
be underlined: if especially low it would be crossed out.
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The OSS Assessment Program
The core of the assessment program was a rather detailed life-history psychiatric
Interview conducted by a senior staff member. In preparation for the interview, the
interviewer read a Personal History Form which the candidate had filled out as well a
Projective Questionnaire, a Sentence Completion Test, a Health Questionnaire, and a
Work Conditions Survey which described 43 conditions that the candidate rated on a
six-point scale, ranging from (1) "Highly desirable" to (6) "Would make job
impossible." In the light of all this material the interviewer had some impression of the
candidate and knew the areas in which he would want to get more information in the
interview.
Two conventional measures of intellectual ability were administered, the Otis
Self-Administering Test of Ability and a Vocabulary Test, plus two tests of special
aptitudes, the Bennett Mechanical Comprehension Test and the Signal Corps Code
Aptitude Test, but we were less interested in our candidates' scores on intelligence
tests than in the effectiveness with which they used whatever intelligence they had.
And this, as well as most of the dimensions of behavior which we rated, was more
clearly to be seen in the most novel aspect of the OSS program-the situational tests
which we developed, tests which today would be more often referred to as simulations
or simulation exercises. Our objective was to observe as much of our candidates'
behavior in situations which simulated as realistically as feasible the kinds of situations
they would be likely to encounter in OSS assignments.
For those destined to work in Secret Intelligence, the abilities to search, to
question, to observe and recall, to infer, and to report would be of crucial importance
and a variety of tests (some situational) to tap the dimension we called Observation
and Reporting were devised. Examples of such tests in which, of course, other
variables could also be observed and rated were:
Belongings Test. In this test the candidate would be taken to a bedroom in
which 26 items such as articles of clothing, written materials, a time table,
newspaper clippings, a ticket receipt, etc., were placed openly on the bed, chairs,
and tables. His task was to examine them, to size up the man who left them, to
learn all he could about the person, what he was like, etc. Objects could be picked
up and examined, but had to be replaced. After four minutes the candidate was
taken to another room to answer a 36-item questionnaire, a test of the candidate's
ability to observe and draw correct inferences.
Map Memory Test. Here the candidate is to assume that he is an agent
operating in the field and that he has just made a secret rendezvous with a courier
who has a map of the territory that the agent will be covering. After a few
minutes the courier must leave with the map and since it would be dangerous for
the agent to have it, he must memorize it. After eight minutes to examine the
map, it is taken away and the candidate answers a set of multiple-choice
statements about the terrain of the map.
Interrogation Test. In this situational test a junior staff member plays the
role (a standardized one) of an escaped prisoner of war who is interrogated by the
candidate. Background material for the interrogation is supplied in the form of a
map and some general information about the imprisonment and escape of the
prisoner, who had served as a tail gunner of a B-29 before its crash landing.
Another staff member sits by and rates the candidate's skill in Developing
Rapport, Asking Productive Questions, Following Up Leads and Hints by the
"Prisoner," etc.
For those destined to work in Morale Operations (MO), the ability to affect others
through acts, words, or displays would be crucial, and special tests were devised in
Approved For Release 2005/01/06: CIA-RDP80-0063OA000100080CO-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The OSS Assessment Program
which these skills might be displayed, thus providing the assessors with some estimate
of the Propaganda Skills of candidates. Two examples of this type of test of
propaganda skills would be the OWI Test and the Manchuria Test. In the former, a
test of one's sensitivity to cultural differences, the candidate was to assume that he was
working for OWI in Korea and knew nothing about the country. He had 20 minutes to
indicate the kind of information he would want to have in order to work up a
propaganda program designed to win Koreans to our side. In the Manchuria Test, the
candidate, given some background facts, was, over a day and a half, to prepare two
pieces of propaganda to lower the morale of Japanese railway workers and guards on
the South Manchurian Railway. In this task candidates were urged to use whatever
graphic skills they had.
One of the better known of our situational tests was the Brook Test (a Leaderless
Group Situation), in which a group of six candidates would be taken to a shallow, quiet
stream whose banks were eight feet apart. On one bank was a heavy rock, on the other
a log. There were trees on both sides and scattered on the side where the group stood
were a number of hoards, none long enough to reach from bank to bank, three lengths
of rope, a pulley, and a barrel with both ends knocked out.
The candidates were told: "In this problem you have to use your imagination.
Before you, you see a raging torrent so deep and so fast that it is quite impossible to
rest anything upon the bottom of the stream. The banks are sheer, so it will he
impossible to work except from the top of them.
"You are on a mission in the field, and having come to this brook you are faced
with the task of transporting this delicate range finder, skillfully camouflaged as a log,
to the far bank, and of bringing that box of percussion caps, camouflaged as a rock, to
this side. In carrying out this assignment, you may make use of any materials around
here. When the job is done, all of you, as well as any material you used, are to be back
on this side. *
The group was to work in an area about 15 feet along the bank marked by two
white stakes.
In this exercise one could easily note the difference between asserted leadership
and effective leadership, and ample opportunity was given to observe differences
among the candidates on such variables as Energy and Initiative, Effective
Intelligence, Social Relations, Leadership, and Physical Ability.
A second Leaderless Group Situation was The Wall. Here the candidates were led
to a wall 10 feet high by 15 feet long. Behind the wall which the candidates could see
was another wall of the same height and length, parallel to it, the two walls being eight
feet apart. On the candidates' side of the wall there were a heavy log, an old board a
few inches longer than the log, and a couple of 2 X 4s, two feet and three feet in
length.
The candidates were informed (fancy being mixed with fact as in all situational
tests) that although they could not see it, the barrier before them actually consisted of
two walls (fact) separated by a 200-foot canyon (fancy). They were escaping from
some Japanese soldiers. In order to save themselves they would have to get to the other
side of the far wall and to fulfill their mission they would have to take their king-size
bazooka (the log) with them. They might get across the wall in any way they wished,
but they must not walk around the ends of the wall, and, of course, whoever or
whatever fell into the canyon would be counted as lost.
Ap -oved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 A roved For Rel ase 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The Approved Assessment Program
The same variables observed and rated at The Brook were observed and rated at
The Wall.
The Brook and The Wall gave some advantage to men with marked initiative and
assertiveness. To discover real ability when leadership was assigned, five other field
tests were employed. The task of the leader in the Mined Road was to get his group
across a mined road in 10 minutes; in the Sentry to get his group, with one man
wounded and unable to walk, across a road guarded by a sentry (played by a junior
staff member in uniform) whose regular beat meant that for about 45 seconds of each
of his tours he could not see the section of the road where the group would try to cross.
In the Agent's Room, the leader has just received a radio message that a friendly agent
had in a hurried getaway left some important papers behind; the task would be to
recover the papers. The instructions were given at the entrance to a large, unused
carriage house on the second floor of which was a crude living room (the agent's
room). Members of the group were told that they had been passing as natives, had
forged credentials, were dressed in native costume, and could speak the native tongue.
Midway through this exercise a German guard armed with a rifle approached the
carriage house.
The other two assigned leadership exercises, the Blown Bridge and Killing the
Mayor, like the three just described, provided good opportunity to observe each
candidate's leadership skills as well as other variables: Practical Intelligence,
Emotional Stability, Social Relations, and Physical Ability.
An indoor, verbal counterpart of the outdoor, physical Brook and Wall exercises
was a Leaderless Group Discussion. The topic was "What are the major postwar
problems facing the United States and (if you have time) along what lines do you think
they should be solved?" The dimensions observed and rated in this situation were
Leadership, Effective Intelligence, Energy and Initiative, and Social Relations.
To measure resistance to stress and frustration tolerance, important facets of
Emotional Stability, two situational tests were devised. Construction, which came to
be known as Behind the Barn, for that was where this exercise was held, required the
candidate to direct two helpers in the task of building a five-foot cube structure with
seven-foot diagonals on the four sides, using a huge "tinker toy" set of materials. The
candidate had 10 minutes in which to accomplish the task. All the physical work was
to be done by the helpers, junior staff members who played the role of Kippy (passive,
sluggish, and something of a stumblebum) and Buster (aggressive, critical, constantly
making impractical suggestions). Both were insulting, faultfinding characters. In the
history of Station S, this job was never completed in the allotted time. Some candidates
gained insight into the problem, but more often they became so involved and so
frustrated that they had difficulty in handling their frustration and controlling their
anger. A few physically attacked their helpers, and some asked to be relieved from the
program after this exercise.
After the Construction test, the candidate was taken aside by a senior staff
member who in a Post-Construction Interview provided the candidate an
opportunity for catharsis. Some admitted their anger and confessed to having almost
hit their "helpers" while others, with quivering lips and trembling hands, denied they
had been in any way emotionally upset in the exercise. This interview was also used to
test the candidate's ability to maintain cover. Since a person is particularly inclined to
break cover in moments of relaxation following stress, the interviewer would inquire
about earlier similar experiences which the candidate had had and how he had acted
then. If the candidate was not on guard he would be led into talking about past
Approved For Release 2005/01/06 : CIA-RDP80-00630A0001000800a? 1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The OSS Assessment Program
experiences of a similar kind in which he had performed better and this, of course,
was a breaking of cover since X conditions had not been established.
In the Stress Interview proper, candidates, one at a time, were instructed that
they have just been found by a night watchman in a government building in
Washington going through papers in an office. The candidate is not an employee of
the agency, nor does he have any identification papers. The night watchman is
bringing him to the security office for questioning. The candidate was given 12
minutes in which to think up an innocent cover story that would explain his being in
the building, after which he was to knock on a designated basement door. Upon being
ordered to enter, he was seated in a dark room, the only light being a blinding
spotlight directed at his face. For ten minutes he would be questioned about his
presence in the building. At first quietly and in an understanding, sympathetic
manner, he was encouraged to give more and more detail, then suddenly the
questioner would turn sadistic, confronting him with inconsistencies and implausibil-
ities in his story with ever-increasing tempo and crescendo of voice. The staff member
who questioned the candidate was a New York trial lawyer skilled in techniques of
cross-examination. Suddenly the examiner would break in with, "We now have
abundant evidence you have not been telling the truth. That is all." Then after
whispered conversation among the staff who had been sitting behind the examiner
observing the candidate but not seen by him except as dim forms, the examiner would
announce, "It is our decision that you have failed this test." After five seconds, during
which his reactions were observed, he was told to go to a room on the third floor.
There he found a staff member, feet on the desk, reading a paper, who would inquire,
"Mow are things going?" and in this Post-Stress Interview many a candidate would
break cover when sympathetically asked if he had ever been in a comparable
situation. The Stress and Post-Stress Interviews provided ample opportunity to
observe and to rate Emotional Stability, Maintaining Cover, and Motivation for
assignment in OSS.
Late afternoon of the second day all candidates under 45 who had passed the
physical examination before coming to Station S ran an Obstacle Course especially
designed to test physical agility and daring.
That evening candidates were assigned in pairs to act in Improvisations. These
were role-playing dyadic situations dreamed up by the staff in sessions referred to as
"Brain storms" and tailor-made for each candidate to help resolve doubts which the
staff might still have about him. For example, there might be some question left as to
how Candidate A takes criticism or rejection, and how much tact Candidate B can
muster. The situations they might be called upon to improvise, the emphasis being put
upon what they would actually do if confronted with such a situation rather than
acting it out, could be the following:
"A. moved to a small city about three months ago and opened a business
there. He has been doing quite well and one month ago sent in an application for
membership in a club in the town. He has heard nothing in response to his
application and goes to the home of B., a prominent member of the club, with
whom he is pleasantly acquainted. (A. is then sent out of the room and B. is told
that A. has received several blackballs.) A. is then called back into the room" and
the improvisation begins. *
After each improvisation the participants' reactions were questioned, commented
upon, and criticized by the participants, and also by the other candidates and staff.
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The OSS Assessment Program
Each candidate was observed in two improvisations. At a halfway break, liquor was
brought in and continued to be available throughout the evening and as long into the
morning as anyone wanted.
After Improvisations, candidates were divided into two teams to debate the
question: "What Shall We Do with Germany After the War?" The Debate provided
opportunity to observe and rate the dimensions of Energy and Initiative, Effective
Intelligence, Social Relations, Leadership, and especially since liquor was still flowing,
the Ability to Maintain Cover.
The unmentioned and never named Liquor Test led to much behavior which
otherwise would not have been observed. During the last month of OSS we were
grateful for an OSS operator, an Army captain who had parachuted into France to
work with a resistance group, who, staying up as long as any of the candidates wished,
relieved the professional staff of this duty.
On the morning of the last full day of assessment the candidates' opinions of and
attitudes toward each other were collected. The Sociometric Questionnaire asked
candidates "With whom would you enjoy continuing your acquaintance?", "Which
men expressed the most realistic and convincing opinions in the debate last night?",
"If you were a member of a group on a dangerous mission, whom would you prefer to
have as your leader?", and other questions of the same sort. The Judgment of Others
test required the candidates to write five personality sketches for the five assessees
whom they felt they had come to know best. Data from these two exercises were
available to the staff in forming their final impressions of the candidates.
During the rest of the day, while the staff was preparing its S Report on each of
the candidates, the assessees were involved in solving the Murder Mystery. For this
exercise they were given copies of the mythical Fairfield Chronicle which reported
the finding of the dead body of a woman on a road near S. Working in competing
groups of six, their task was to wring from three junior staff members (standard role
players), who were designated as possible informants, the information necessary for
them to make proper inferences and thus to solve "the murder."
Athletic Events-broad jump, high jump, and shot-put-more in the spirit of a
game than as a serious test, and a Baseball Game between the candidates and junior
staff members not involved in the staff conference engaged the interest of the assessees
during the last afternoon.
Reports on the candidates' performance in these last day simulations and games
were available to the staff before their deliberations ended.
For each assessment class, usually consisting of 18 assessees, the staff was divided
into teams of two senior staff member (professionals with Ph.D. or M.D. degrees) and
one junior staff member (enlisted men who had had some training in psychology).
Each team was assigned to a group of five to seven candidates. The senior members
conducted the life history interviews; the junior member administered special
individual tests and interpreted the projective test protocols. Otherwise, the work of
both senior and junior members was the same, namely, to develop as a group as
complete a conception as possible of each candidate in the subgroup assigned to them.
During the various situational tests the behaviors of each participant were
carefully noted by both senior and junior members of the responsible team, each staff
member rating each assessee on the variables relevant to the particular exercise.
Usually immediately following each situational test, the staff team met to discuss their
impressions of the candidates and the ratings they had assigned to them. The purpose
Approved For Release 2005/01/06: CIA-RDP80-0063OA000100080W-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The OSS Assessment Program
of the meeting was to come to agreement upon the ratings to be assigned to each
assessee on each of the rated variables.
The staff meeting for the preparation of the Station S Report on the candidates
ran from after lunch on the last full day of the assessment until all reports had been
finished, which was often close to midnight.
Before the staff meeting each interviewer had written a first draft of his section
of the reports on the men he had interviewed and for whom he was the primary
assessor. This was a description of the candidate as a person and of the major events of
his past history that seemed to shed light on his present personality, with special
emphasis being given to picturing how the candidate might be expected to behave in a
variety of situations and circumstances in the future. Also before the staff meeting the
member of the staff who was responsible for reporting on the behavior of the
candidate at Station S as revealed in the various situational tests, the situationist, as he
was known, had written his section of the final report, which not only described the
behavior of the candidate at S but offered specific support for the statements and
predictions made in the first part of the report written by the interviewer.
A huge board which covered one wall of the staff room offered a graphic display
of how each candidate had been rated by the staff team assigned to him on each of the
variables in each of the relevant tests and situations. Ratings were represented by
thumb tacks, red if above average, blue if below. While the interviewer and the
situationist read their reports, staff members looked at the board and the thumbtacks
to see if the reporters' statements were supported by the ratings. If not, there was
discussion until general agreement was reached about the changes that would have to
he made in the overall rating of the variables on the Station S Report or in the
statements in the written report. Sometimes the reports could be revised in the staff
meeting. If not, they would be rewritten by the report writer after the meeting.
The final report consisted of a face sheet on which the candidate was rated on the
nine variables already described, plus other sheets which carried a character sketch of
the candidate based upon the interviewer's insights and the staff's obsevations of the
candidate at S, and recommendations concerning overseas assignment made on a
S-point scale: Not Recommended, Doubtful, Recommended with Qualifications,
Recommended, Highly Recommended. A similar scale (Not Recommended to Highly
Recommended) was used to indicate candidate's fitness for work at (1) a rear base,
(2) an advanced base, and (3) at or behind enemy lines; his fitness for (1) higher,
(2) middle, or (3) lower level of authority and responsibility; and finally his fitness for
different types of assignment, e.g., administrator, intelligence officer, operational
agent, etc.
And these Station S reports, sometimes worked over long past midnight of the last
day of assessment, were already at OSS Headquarters in Washington when the
candidates arrived back there the next morning.
Iow effective was the OSS assessment program? We cannot say with certainty,
but accepting our validity figures at face value we were forced to conclude that we
were not very successful in predicting performance overseas. But were errors mainly
in the assessment process, or in the appraisal process, or in both? Again we cannot say
with certainty, but we do know that the appraisal process as carried out left very much
to be described. Our appraisal data were of four types:
L. Overseas Staff Appraisal. These were appraisals made by OSS assessment
staff members on the basis of interviews with the immediate chief or
commanding officer and, if possible, associates of our "graduates" overseas.
Approved For Release 2005/01/06: CIA-RDP80-0063OA000100080001-1
 oved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
The Assessment rogram
Footnote from a Guinea Pig
I can't resist a brief personal footnote to Donald MacKinnon's fascinating account of how OSS
assessed its recruits. Having been one of those assessed in this imaginative experiment, I can provide a
glimpse or two of the process as seen from the other end of the microscope.
The assessment process at Station "S", described so well in MacKinnon's OSS Assessment
Program, was to many of us assessed a splendid lark-sort of an English country house party except
for the phalanx of psychologists scrutinizing our every move. And even they seemed more like
recreation directors on a cruise ship than inquisitors. On arrival at the old Willard place in Fairfax I had a
fleeting moment to admire a magnificent stand of English boxwood before I was swept into the swirl of
activity calculated to take my measure. We were sternly admonished to stick to our cover story and
keep our identities secret. We were warned that efforts would be made by the staff to trick us into giving
our true selves away so we must be eternally vigilant. This kind of enforced anonymity was in a way
exhilarating-like taking a vacation from one's old self.
We were split into teams. Only with considerable research much later did I discover that my
colleagues in real life were the president of a prominent bank in Paris, a ballet dancer, a disc jockey, a
college professor, and a giant who had played pro football before the army inducted him. They had all
adopted cover identities-most of which struck me as implausible. Hiding my real Foreign Service
background, I chose to masquerade as a lumberjack since I had once had a summer fling at that and
could more or less talk the jargon. But I, too, didn't fit my new role very well.
The battery of written aptitude tests was tedious, as always. But the live exercises were great fun
even though we knew we were being watched carefully and rigorously graded. The "Brook Test" was
my favorite. As described in MacKinnon's article, our team was indeed faced with a large boulder sitting
heavily beside a small stream. It was explained that we would be timed while we sought to organize
ourselves and make the best use of the boards and pulleys strewn about to move the rock to the other
side of the stream. But before the earnest mentor could finish his instructions, our football hero grabbed
the rock, hugged it to his chest as he would some sort of outsized, overweight football, and leaped
nimbly to the other side of the stream for a touchdown. How our team was graded on leadership,
effective intelligence, incipient bravery, or whatever, I had no way of knowing. But we had at least
broken a time record.
In the "Behind the Barn" test each of us, with the help of two assistants provided by management,
took turns at trying to erect a giant tinkertoy cube. It soon became clear that the two "helpers," known
as "Kippy" and "Buster," were intent on sabotaging the effort. Never one to fight a hopeless problem, I
used my allotted ten minutes to relax in a hayloft. As MacKinnon wrote, this test evoked a variety of
responses. Some well-adjusted persons took it in the spirit of high humor-even joined in the highjinks,
turning it into a sort of impromptu "Three Stooges" routine. Others, of course, let their frustrations
show. Rumors were rife at the time that the only student to complete the tinkertoy had been a strapping
Texan who flattened Kippy and Buster with two well-aimed blows. As a result, the Texan became a sort
of folk hero at the station-perhaps a symbol that we could win the war after all so long as we had
problem-solvers like that.
The "Liquor Test," rather gingerly touched on by MacKinnon, was presumably based on the old
saw, In Vino Veritas. Under the guise of a relaxing farewell party we were plied with liquor, presumably
so that our testers could catch us with our guards down. This is an interesting approach but I don't think
it worked very well with our team. The college professor announced that he never touched the stuff. The
linebacker poured more than his share into his generous stomach without its appearing to have any
effect. The disc jockey, exhausted by his weekend, fell asleep promptly, while the Paris bank president
shrewdly switched to French which went over the heads of our observers. Playing my woodsman cover
to the hilt, I recall talking expansively to my observers about buying the Willard's prize boxwood after
the War. So the evening passed. I doubt if it all proved much.
OSS's membership list was long. It read like a cross between Who's Who and the Social
Register. Many of the young and untried later rose to positions of talent and prominence. It would be
interesting to know how the assessment profiles of the like of Arthur Goldberg, David Bruce, Julia Child,
Ralph Bunche, Walt Rostow, Sterling Hayden, Stewart Alsop, John Gardner, Douglas Dillon, and Paul
Mellon read after they had paraded anonymously through Stations W or S.
Not that our team was a typical microcosm, but it acquitted itself well in real action as far as I could
later determine. The disc jockey with no apparent talents distinguished himself behind the lines in
France and Norway. The somewhat musty professor did magnificently in North Africa. The pro-ball star
never surpassed his accomplishment of getting the rock across the stream, but I'm told he did his best
in the Far East. The Paris banker served gloriously with the French undergound and re-entered liberated
Paris in triumph.
MacKinnon's conclusion that the OSS assessment program had not been very successful in
predicting overseas performance is discouraging. All these years I have cherished the illusion that by
surviving the Station S trial I had been permitted to join an elite corps, pre-determined by scientific
testing to win the war. Now I discover that the Station S assessment record was only 14% more
accurate than what could have been done by a roulette wheel. Still, the experiment was a brave effort
and perhaps contributed to the science of testing. I shall try to find at least a measure of satisfaction in
having been a cheerful guinea pig. John W. Waller
Approved For Release 2005/01/06 : CIA-RDP80-00630A0001000800M-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The OSS Assessment Program
Table 1
Correlations between S and W Assessment Job Ratings and Appraisal Ratings
(After OSS Assessment Staff, 1948, p. 423)
S Job Rating W Job Rating
Type of Appraisal (Classes S-45 on) (All Classes)
r N r N
Overseas
Staff Appraisal
.37 a
88
53 a
83
Returnee
Appraisal
.19 a
93
.21 a
173
Theater
Commanders Appraisal
.23
64
15
158
Reassignment Area Appraisal .08
53
.30 a
178
Cases in which correcting r for restricted sample made a significant difference; r given in each case
is the corrected one.
These recommendations are reproduced below with the thought that the readers
of this monograph may find it both interesting and informative as they review them to
ask themselves such questions as these: How many of these recommendations have
been carried out in the setting up of assessment centers in business, in industry, in
government, and in education? If adopted have they proved helpful? And, if they
have been ignored, were they rejected for good reason? And finally, what further
recommendations should now be made for the improvement of assessment centers?
The recommendations follow:
i. Select a staff of suitable size and competence, diversified in respect to age,
sex, social status, temperament, major sentiments, and specific skills, but
uniform in respect to a high degree of intellectual and emotional flexibility
(p. 473).
2. Before designing the program of assessment procedures, conduct a
preliminary study of the jobs and job holders of the organization (p. 475).
f. Make an adequate functional analysis of each of the roles for which
candidates are to be assessed as well as an analysis of the environments in
which each role must he fulfilled (p. 476).
2.2. Obtain from members of the organization a list of attributes of personality
which, in their opinion, contribute to success or failure in the performance of
each role (p. 476).
2.3. After a careful survey, analysis, and classification of the information
obtained by these obsevations and interviews (recommendations 2.1 and 2.2),
make a tentative list of the personality determinants of success or failure in
the performance of each role. These determinants will constitute the
variables which, if possible, will be measured by the assessment procedures
(p. 477).
2.4. Define, in words that are intelligible to members of the organization, a
tentative rating scale for each personality variable on the selected list as well
as for the overall variable, Job Fitness (p. 479).
2.5. Devise a satisfactory system for appraising the performance of members of
the organization both at this time and later (p. 481).
2.6. Obtain appraisals of a properly distributed sample of the present members of
the organization (p. 484).
Appr82ed For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 The AC7 J o~s ess or Rerlease 2005/01 /06 : CIA-RDP80-00630A000100080001-1
2. Theater Commander's Appraisal. From Spring 1944, each returnee was
reported on and his personality traits were rated by his immediate superior.
3. Reassignment Area Appraisal. From Fall of 1944 a center was set up
(Area F) for the re-assessment or re-evaluation of personnel who had completed a
tour of duty in Europe or the Middle East for reassignment in the Far East.
4. Returnee Appraisal. Beginning in the summer of 1945, each returnee was
asked to report on others known to him, rating their performance Low, Medium,
or High.
When assessment job ratings were correlated with appraisal job ratings for
Stations S and W, validity coefficients were all positive but disappointingly low,
ranging from .08 to .53, depending upon the source of the appraisal data, as shown in
Table 1. And, much to our surprise, the one-day assessments at W appeared to have
been more effective than the three-and-a-half-day assessments at S.
We of the OSS staff are indebted to Jerry S. Wiggins (1973) who, making certain
assumptions and using improved principles for estimating outcomes of predictions,
worked over our data and came up with a more favorable picture than the one we had
drawn.
His estimate is that at S, if we had used only random selection, our percent of
correct decisions would have been 63%, but actually 77% were correct. Corresponding
estimates for Station W are that by random selection 66% would have been correct,
but actually 84% were correct. This means that at S, assessment effected a 14%
increase in correct decisions over random selection and at Station W an 18%
increment. Considering the crucial nature of the assignments, increments of 14% and
18% of correct decisions are not unimportant.
It is interesting to speculate as to why the briefer (one day) assessments of Station
W were more accurate than the longer (three-and-a-half-day) assessments of Station S.
It may have been that the procedures used at W were more efficient than those
employed at S, although this seems unlikely. Perhaps the Staff at W was more
competent than the one at S: more psychiatrists served on the staff at W and less use
was made of junior assessors than at S. Differences in the populations assessed at the
two stations could also have been a determining factor. Those assessed at W were more
often high echelon executives in the organization, women secretaries, and office
workers many of whom had already spent some time in the Washington headquarters,
while those sent to S for assessment were either the more difficult cases who were
already presenting perplexing problems or were men destined for more hazardous
duty under more stress and danger than would be experienced by the W assessees.
Indeed, of those assessed at W, 74% received rear base assignments with only 15%
serving behind enemy lines. In contrast, only 29% of Station S graduates received rear
base assignments while 43% operated behind enemy lines. Among possible
explanations of the differential success rates of the two assessment centers the one that
seems least plausible is the notion that the staff at Station S suffered from a
superabundance of information about their assessees, while the staff at Station W, with
less information, had just what they needed and no more to make the kinds of
decisions called for in the OSS assessment program. But the fact remains that we
cannot say with certainty why the assessments at W surpassed those at S. Indeed, still
today the optimal length of assessment center programs remains an unanswered
question, one which should long since have been subjected to empirical investigation.
At the end of their report on the OSS program, Assessment of Men, the OSS staff
made a number of recommendations which it was hoped would remedy some of the
defects of assessment as practiced in the OSS.
Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001 1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
The OSS Assessment Program
2.7. Examine the defects of the appraisal system as revealed in practice
(recommendation 2.6), and correct these by revising, where necessary, the
lists of variables, the definitions, the rating scales, or the other elements.
2.8. Obtain the figures necessary for a brief numerical statement of the personnel
history of the organization over the last four or five years (p. 485).
3. Design a program of assessment procedures which will reveal the strength of
the selected variables; for assessing these variables set up scales which
conform to the rating scales that were defined for the purpose of appraisal
(p. 485).
4. Build a conceptual scheme in terms of which formulations of different
personalities can be made (p. 488).
5. Set up an efficient punch-card system which will permit periodic statistical
analyses of assessment findings (p. 490).
6. Assess candidates for a long trial period without reporting ratings or decisions
to the organizations (p. 491).
How far and in what directions the state of the art of assessment has moved
beyond that which obtained in the assessment program of the OSS is a fascinating
chronicle, but that is another story for another time.
OSS Assessment Staff. Assessment of men. Rinehart and Company, New
York, 1948.
Smith, R. H. OSS: The secret history of America's first central
intelligence agency. University of California Press, Berkeley,
California, 1972.
Wiggins, J. S. Personality and prediction: Principles of personality
assessment. Addison-Wesley Publishing Company, Reading, Massa-
chusetts, 1973.
(The foregoing article is Unclassified.)
fildproved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Reflections of an intelligence scholar
and Senate Select Committee investigator
ANALYSIS, WAR AND DECISION: WHY INTELLIGENCE
FAILURES ARE INEVITABLE
The Editors commend this article to all producers and consumers of
estimative and warning intelligence. It is rejrinted from World Politics,
Vol. XXXI No. 1 (October 1978), Copyright?1978 by Princeton University
Press, by permission of the copyright holder.
Military disasters befall some states, no matter how informed their leaders are,
because their capabilities are deficient. Weakness, not choice, is their primary
problem. Powerful nations are not immune to calamity either, because their leaders
may misperceive threats or miscalculate responses. Information, understanding, and
judgment are a larger part of the strategic challenge for countries such as the United
States. Optimal decisions in defense policy therefore depend on the use of strategic
intelligence: the acquisition, analysis, and appreciation of relevant data. In the best-
known cases of intelligence failure, the most crucial mistakes have seldom been made
by collectors of raw information, occasionally by professionals who produce finished
analyses, but most often by the decision makers who consume the products of
intelligence services. Policy premises constrict perception, and administrative
workloads constrain reflection. Intelligence failure is political and psychological more
often than organizational.
Observers who see notorious intelligence failures as egregious often infer that
disasters can be avoided by perfecting norms and procedures for analysis and
argumentation. This belief is illusory. Intelligence can be improved marginally, but
not radically, by altering the analytic system. The illusion is also dangerous if it abets
overconfidence that systemic reforms will increase the predictability of threats. The
use of intelligence depends less on the bureaucracy than on the intellects and
inclinations of the authorities above it. To clarify the tangled relationship of analysis
and policy, this essay explores conceptual approaches to intelligence failure,
differentiation of intelligence problems, insurmountable obstacles to accurate
assessment, and. limitations of solutions proposed by critics.
Case studies of intelligence failures abound, yet scholars lament the lack of a
theory of intelligence., It is more accurate to say that we lack a positive or normative
*For corrections or comments whose usefulness exceeded my ability to accommodate them within
space limitations, thanks are due to Bruce Blair, Thomas Blau, Michael Handel, Robert Jervis, Klaus Knorr,
H. R. Trevor-Roper, and members of the staff of the National Foreign Assessment Center.
'For example, Klaus Knorr, "Failures in National Intelligence Estimates: The Case of the Cuban
Missiles," World Politics, xvi (April 1964), 455, 465-66; Harry Howe Ransom, "Strategic Intelligence and
Foreign Policy," World Politics, xxvii (October .1974), 145. MORI/HRP
from pg.
35 35-54
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
theory of intelligence. Negative or descriptive theory-the empirical understanding of
how intelligence systems make mistakes-is well developed. The distinction is
significant because there is little evidence that either scholars or practitioners have
succeeded in translating such knowledge into reforms that measurably reduce failure.
Development of a normative theory of intelligence has been inhibited because the
lessons of hindsight do not guarantee improvement in foresight, and hypothetical
solutions to failure only occasionally produce improvement in practice. The problem
of intelligence failure can be conceptualized in three overlapping ways. The first is the
most reassuring; the second is the most common; and the third is the most important.
1. Failure in perspective. There is an axiom that a pessimist sees a glass of water
as half empty and an optimist sees it as half full. In this sense, the estimative system is
a glass half full. Mistakes can happen in any activity. Particular failures are accorded
disproportionate significance if they are considered in isolation rather than in terms of
the general ratio of failures to successes; the record of success is less striking because
observers tend not to notice disasters that do not happen. Any academician who used a
model that predicted outcomes correctly in four out of five cases would be happy;
intelligence analysts must use models of their own and should not be blamed for
missing occasionally. One problem with this benign view is that there are no clear
indicators of what the ratio of failure to success in intelligence is, or whether many
successes on minor issues should be reassuring in the face of a smaller number of
failures on more critical problems.2 In the thermonuclear age, just one mistake could
have apocalyptic consequences.
2. Pathologies of communication. The most frequently noted sources of
breakdowns in intelligence lie in the process of amassing timely data, communicating
them to decision makers, and impressing the latter with the validity or relevance of
the information. This view of the problem leaves room for optimism because it implies
that procedural curatives can eliminate the dynamics of error. For this reason, official
post mortems of intelligence blunders inevitably produce recommendations for
reorganization and changes in operating norms.
3. Paradoxes of perception. Most pessimistic is the view that the roots of failure
lie in unresolvable trade-offs and dilemmas. Curing some pathologies with
organizational reforms often creates new pathologies or resurrects old ones;, perfecting
intelligence production does not necessarily lead to perfecting intelligence consump-
tion; making warning systems more sensitive reduces the risk of surprise, but increases
the number of false alarms, which in turn reduces sensitivity; the principles of optimal
analytic procedure are in many respects incompatible with the imperatives of the
decision process; avoiding intelligence failure requires the elimination of strategic
preconceptions, but leaders cannot operate purposefully without some preconceptions.
In devising measures to improve the intelligence process, policy makers are damned if
they do and damned if they don't.
' "As that ancient retiree from the Research Department of the British Foreign Office reputedly said,
after serving from 1903-50: 'Year after year the worriers and fretters would come to me with awful
predictions of the outbreak of war. I denied it each time. I was only wrong twice.' " Thomas L. Hughes, The
Fate of Facts in a World of Men-Foreign Policy and Intelligence-Making (New York: Foreign Policy
Association, Headline Series No. 233, December 1976), 48. Paradoxically, "successes may be
indistinguishable from failures." If analysts predict war and the attacker cancels his plans because surprise
has been lost, "success of the intelligence services would have been expressed in the falsification of its
predictions," which would discredit the analysis. Avi Shlaim, "Failures in National Intelligence Estimates:
The Case of the Yom Kippur War," World Politics, xxviii (April 1976), 378.
? Compare the prescriptions in Peter Szantun and Graham Allison, "Intelligence: Seizing the
Opportunity," with George Carver's critique, both in Foreign Policy, No. 22 (Spring 1976).
AI3oroved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Ana~ylsuoa Fanc~ib~c~-gir~005/01/06 :CIA-RDP80-00630A000100080001-1
It is useful to disaggregate the problem of strategic intelligence failures in order to
elicit clues about which paradoxes and pathologies are pervasive and therefore most in
need of attention. The crucial problems of linkage between analysis and strategic
decision can be subsumed under the following categories:
1. Attack warning. The problem in this area is timely prediction of an enemy's
immediate intentions, and the "selling" of such predictions to responsible authorities.
Major insights into intelligence failure have emerged from catastrophic surprises: Pearl
Harbor, the Nazi invasion of the U.S.S.R., the North Korean attack and Chinese
intervention of 1950, and the 1973 war in the Middle East. Two salient phenomena
characterize these cases. First, evidence of impending attack was available, but did not
flow efficiently up the chain of command. Second, the fragmentary indicators of
alarm that did reach decision makers were dismissed because they contradicted
strategic estimates or assumptions. In several cases hesitancy in communication and
disbelief on the part of leaders were reinforced by deceptive enemy maneuvers that
cast doubt on the data.,
2. Operational evaluation. In wartime, the essential problem lies in judging the
results (and their significance) of interacting capabilities. Once hostilities are under
way, informed decision making requires assessments of tactical effectiveness-"how
we are doing"--in order to adapt strategy and options. In this dimension, the most
interesting insights have come from Vietnam-era memoirs of low-level officials and
from journalistic muckraking. Again there are two fundamental points. First, within
the context of a glut of ambiguous data, intelligence officials linked to operational
agencies (primarily military) tend to indulge a propensity for justifying service
performance by issuing optimistic assessments, while analysts in autonomous non-
operational units (primarily in the Central Intelligence Agency and the late Office of
National Estimates) tend to produce more pessimistic evaluations. Second, in contrast
to cases of attack warning, fragmentary tactical indicators of success tend to override
more general and cautious strategic estimates. Confronted by differing analyses, a
leader mortgaged to his policy tends to resent or dismiss the critical ones, even when
they represent the majority view of the intelligence community, and to cling to the
data that support continued commitment.' Lyndon Johnson railed at his Director of
? Roberta Wohlstetter, Pearl Harbor: Warning and Decision (Stanford: Stanford University Press
1962); Barton Whaley, Codeword Barbarossa (Cambridge: The M.I.T. Press 1973); Harvey De Weerd,
"Strategic: Surprise in the Korean War," Orbis, vi (Fall 1962); Alan Whiting, China Crosses the Yalu (New
York: Macmillan 1960); James F. Schnabel, Policy and Direction: The First Year (Washington, D.C.:
Department of the Army 1972), 61-65, 83-85, 274-78; Michael I. Handel, Perception, Deception, and
Surprise: The Case of the Yom Kippur War (Jerusalem: Leonard Davis Institute of International Relations,
Jerusalem Paper No. 19, 1976); Shlaim (fn. 2); Abraham Ben-Zvi, "Hindsight and Foresight: A Conceptual
Framework for the Analysis of Surprise Attacks," World Politics, xxvin (April 1976); Amos Perlmutter,
"Israel's Fourth War, October 1973: Political and Military Misperceptions," Orbis, xix (Summer 1975); U.S.,
Congress, House, Select Committee on Intelligence [hereafter cited as HSCI], Hearings, U.S. Intelligence
Agencies and Activities: The Performance of the Intelligence Community, 94th Cong., 1st sess., 1975;
Draft Report of the House Select Committee on Intelligence, published in The Village Voice, February 16,
1976, pp. 76-81.
' David Halberstam, The Best and the Brightest (New York: Random House 1972); Morris Blachman,
"The Stupidity of Intelligence," in Charles Peters and Timothy J. Adams, eds., Inside the System (New
York: Praeger 1970); Patrick J. McGarvey, "DIA: Intelligence to Please," in Morton Halperin and Arnold
Kanter, eds., Readings in American Foreign Policy: A Bureaucratic Perspective (Boston: Little, Brown
1973); Chester Cooper, "The CIA and Decision-Making," Foreign Affairs, Vol. 50 (January 1972); Sam
Adams, "Vietnam Cover-Up: Playing War With Numbers," Harper's, Vol. 251 (June 1975); Don
Oberdorfer, Tet! (Garden City, N.Y.: Doubleday 1971). For a more detailed review, see Richard K. Betts,
Soldiers, Statemen, and Cold War Crises (Cambridge: Harvard University Press 1977), chap. 10.
Approved For Release 2005/01/06 : CIA-RDP80-0063OA0001000800(A71
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
Central Intelligence (DCI) at a White House dinner: "Policy making is like milking a
fat cow. You see the milk coming out, you press more and the milk bubbles and flows,
and just as the bucket is full, the cow with its tail whips the bucket and all is spilled.
That's what CIA does to policy making." I From the consensus-seeking politician, this
was criticism; to a pure analysts, it would have been flattery. But it is the perspective
of the former, not the latter, that is central in decision making.
3. Defense planning. The basic task in using intelligence to develop doctrines
and forces for deterrence and defense is to estimate threats posed by adversaries, in
terms of both capabilities and intentions, over a period of several years. Here the
separability of intelligence and policy, analysis and advocacy, is least clear. In dealing
with the issue of "how much is enough" for security, debates over data merge murkily
into debates over options and programs. As in operational evaluation, the problem lies
more in data mongering than in data collecting. To the extent that stark
generalizations are possible, the basic points in this category are the reverse of those in
the previous one.
First, the justification of a mission (in this case, preparedness for future
contingencies as opposed to demonstration of current success on the battlefield)
prompts pessimistic estimates by operational military analysts; autonomous analysts
without budgetary axes to grind, but with biases similar to those prevalent in the
intellectual community, tend toward less alarmed predictions.' Military intelligence
inclines toward "worst-case" analysis in planning, and toward "best-case" analysis in
operational evaluation. (Military intelligence officials such as Lieutenant General
Daniel Graham were castigated by liberals for underestimating the Vietcong's strength
in the 1960's but for overestimating Soviet strength in the 1970's.) Air Force
intelligence overestimated Soviet air deployments in the "bomber gap" controversy of
the 1950's, and CIA-dominated National Intelligence Estimates (NIE's) under-
eslimated Soviet ICBM deployments throughout the 1960's (over-reacting, critics say,
to the mistaken prediction of a "missile gap" in 1960).1
Second, in the context of peacetime, with competing domestic claims on
resources, political leaders have a natural interest in at least partially rejecting military
estimates and embracing those of other analysts who justify limiting allocations to
defense programs. If the President had accepted pessimistic CIA operational
evaluations in the mid-1960's, he might have withdrawn from Vietnam; if he had
accepted pessimistic military analyses of the Soviet threat in the mid-1970's, he might
have added massive increases to the defense budget.
Betts (fn. 5), 160-61, 192-95. On bias within CIA, see James Schlesinger's comments in U.S.,
Congress, Senate, Select Committee to Study Governmental Operations with Respect to Intelligence
Activities [hereafter cited as SSCI], Final Report, Foreign and Military Intelligence, Book I, 94th Cong., 2d
sess., 1976, 76-77.
"Ibid., Book IV, 56-59; William T. Lee, Understanding the Soviet Military Threat: How CIA
Estimates Went Astray (New York: National Strategy Information Center, Agenda Paper No. 6, 1977), 24-
37; Albert Wohlstetter: "Is There a Strategic Arms Race?" Foreign Policy, No. 15 (Summer 1974);
Wohlstetter, "Rivals, But No Race," Foreign Policy, No. 16 (Fall 1974); Wohlstetter, "Optimal Ways to
Confuse Ourselves," Foreign Policy, No. 20 (Fall 1975). There are exceptions to this pattern of military and
civilian bias: see ibid., 185-88; Lieutnant General Daniel Graham, USA (Ret.), "The Intelligence Mythology
of Washington," Strategic Review, iv (Summer 1976), 61-62, 64; Victor Marchetti and John Marks, The CIA
and the Cult of Intelligence (New York: Knopf 1974), 309.
PjWroved For Release 2005/01/06: CIA-RDP8O-0063OA000100080001-1
 Ari4a ~sv 0 FV an or RTasro2005/01/06 : CIA-RDP80-0063OA0001 00080001 -1
Some chronic sources of error are unique to each of these three general categories
of intelligence problems, and thus do not clearly suggest reforms that would be
advisable across the board. To compensate for the danger in conventional attack
warning, reliance on worst-case analysis might seem the safest rule, but in making
estimates for defense planning, worst-case analysis would mandate severe and often
unnecessary economic sacrifices. Removing checks on the influence of CIA analysts
and "community" staffs ? might seem justified by the record of operational evaluation
in Vietnam, but would not be warranted by the record of estimates on Soviet ICBM
deployments. It would be risky to alter the balance of power systematically among
competing analytic components, giving the "better" analysts more status. Rather,
decision makers should be encouraged to be more and less skeptical of certain
agencies' estimates, depending on the category of analysis involved.
Some problems, however, cut across all three categories and offer a more general
basis for considering changes in the system. But these general problems are not very
susceptible to cure by formal changes in process, because it is usually impossible to
disentangle intelligence failures from policy failures. Separation of intelligence and
policy making has long been a normative concern of officials and theorists, who have
seen both costs and benefits in minimizing the intimacy between intelligence
professionals and operational authorities. But, although the personnel can be
segregated, the functions cannot, unless intelligence is defined narrowly as the
collection of data, and analytic responsibility is reserved to decision makers. Analysis
and decision are interactive rather than sequential processes. By the narrower
definition of intelligence, there have actually been few major failures. In most cases of
mistakes in predicting attacks or in assessing operations, the inadequacy of critical data
or their submergence in a viscous bureaucracy were at best the proximate causes of
failure. The ultimate causes of error in most cases have been wishful thinking, cavalier
disregard of professional analysts, and, above all, the premises and preconceptions of
policy makers. Fewer fiascoes have occurred in the stages of acquisition and
presentation of facts than in the stages of interpretation and response. Producers of
intelligence have been culprits less often than consumers. Policy perspectives tend to
constrain objectivity, and authorities often fail to use intelligence properly. As former
State Department intelligence director Ray Cline testified, defending his analysts'
performance in October 1973 and criticizing Secretary Kissinger for ignoring them:
Unless something is totally conclusive, you must make an inconclusive
report.... by the time you are sure it is always very close to the event. So I
don't think the analysts did such a lousy job. What I think was the lousy job
was in bosses not insisting on a new preparation at the end of that week
[before war broke out].... the reason the system wasn't working very well is
that people were not asking it to work and not listening when it did work.,,
II. BASIC BARRIERS TO ANALYTIC ACCURACY
Many constraints on the optimal processing of information lie in the structure of
authority and the allocation of time and resources. Harold Wilensky argues
persuasively that the intelligence function is hindered most by the structural
0 The U.S. intelligence community includes the CIA, Defense Intelligence Agency (DIA), National
Security Agency, the intelligence branches of each military service, the State Department Bureau of
Intelligence and Research, the intelligence units of the Treasury and Energy Departments, and the FBI.
Before 1973, coordination for national estimates was done through the Office of National Estimates, and
since then, through the National Intelligence Officers. The Intelligence Community Staff assists the Director
of Central Intelligence in managing allocation of resources and reviewing the agencies' performance.
10 HSCI, Hearings (fn. 4), 656-57.
Approved For Release 2005/01/06 : CIA-RDP80-0063OA0001000800a4-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
characteristics of hierarchy, centralization, and specialization." Yet it is precisely these
characteristics that are the essence of any government. A related problem is the
dominance of operational authorities over intelligence specialists, and the trade-off
between objectivity and influence. Operators have more influence in decision making
but are less capable of unbiased interpretation of evidence because they have a vested
interest in the success of their operations; autonomous analysts are more disinterested
and usually more objective, but lack influence. Senior generalists at the policy level
often distrust or discount the judgments of analytic professionals and place more
weight on reports from operational sources.,, In response to this phenomenon, the
suggestion has been made to legislate the requirement that decision makers consider
analyses by the CIA's Intelligence Directorate (now the National Foreign Assessment
Center) before establishing policy.13 Such a requirement would offer no more than
wishful formalism. Statutory fiat cannot force human beings to value one source above
another. "No power has yet been found," DCI Richard Helms has testified, "to force
Presidents of the United States to pay attention on a continuing basis to people and
papers when confidence has been lost in the originator." " Moreover, principals tend
to believe that they have a wider point of view than middle-level analysts and are
better able to draw conclusions from raw data. That point of view underlies their
fascination with current intelligence and their impatience with the reflective
interpretations in "finished" intelligence."
'[he dynamics of decision are also not conducive to analytic refinement. In a
crisis, both data and policy outpace analysis, the ideal process of staffing and
consultation falls behind the press of events, and careful estimates cannot be digested
in time. As Winston Churchill recalled of the hectic days of spring 1940, "The
Defence Committee of the War Cabinet sat almost every day to discuss the reports of
the Military Co-ordination Committee and those of the Chiefs of Staff; and their
conclusions or divergences were again referred to frequent Cabinets. All had to be
explained or reexplained; and by the time this process was completed, the whole scene
had often changed." 11 Where there is ample time for decision, on the other hand, the
previously mentioned bureaucratic impediments gain momentum." Just as informa-
tion processing is frustrated by constraints on the time that harried principals can
spend scrutinizing analytic papers, it is constrained by the funds that a government
can spend. To which priorities should scarce resources be allocated? The Schlesinger
Report of 1971, which led to President Nixon's reorganization of U.S. intelligence,
noted that criticisms of analytic products were often translated into demands for more
" Wilensky, Organizational Intelligence (New York: Basic Books 1967), 42-62, 126, 179.
'` Ibid., passim. The counterpoint of Cooper (fn. 5) and McGarvey (fn. 5) presents a perfect
illustration.
11 Graham Allison and Peter Szanton, Remaking Foreign Policy: The Organizational Connection
(New York: Basic Books 1976), 204.
Quoted in SSCI, Final Report (fn. 7), I, 82.
Ibid., 267, 276; SSCI, Staff Report, Covert Action in Chile 1963-1973, 94th Cong., 1st sess., 1975,
48-49. The Senate Committee deplored the tendency of decision makers to focus on the latest raw data
rather than on refined analyses, a practice that contributed to the intelligence failure in the 1974 Cyprus
crisis. SSCI, Final Report (fn. 7), 1, 443. But the failure in the October War was largely due to the reverse
phenomenon: disregarding warning indicators because they contradicted finished intelligence that
minimized the possibility of war. HSCI Draft Report (fn. 4), 78; Ben-Zvi (fn. 4), 386, 394; Perlmutter (fn. 4),
453.
', Churchill, The Gathering Storm (Boston: Houghton Mifflin 1948), 587-88.
"Where the end is knowledge, as in the scientific community, time serves intelligence; where the
end is something else-as in practically every organization but those devoted entirely to scholarship-time
subverts intelligence, since in the long run, the central institutionalized structures and aims (the
maintenance of authority, the accommodation of departmental rivalries, the service of established doctrine)
will prevail." Wilensky (fn. 11), 77.
Ai roved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
extensive collection of data, but "Seldom does anyone ask if a further reduction in
uncertainty, however small, is worth its cost." 18 Authorities do not always know,
however, which issues require the greatest attention and which uncertainties harbor
the fewest potential threats. Beyond the barriers that authority, organization, and
scarcity pose to intelligence lie more fundamental and less remediable intellectual
sources of error.
1. Ambiguity of evidence. Intelligence veterans have noted that "estimating is
what you do when you do not know," 10 but "it is inherent in a great many situations
that after reading the estimate, you will still not know." 20 These observations highlight
an obvious but most important obstacle to accuracy in analysis. It is the role of
intelligence to extract certainty from uncertainty and to facilitate coherent decision in
an incoherent environment. (In a certain and coherent environment there is less need
for intelligence.) To the degree they reduce uncertainty by extrapolating from
evidence riddled with ambiguities, analysts risk oversimplifying reality and
desensitizing the consumers of intelligence to the dangers that lurk within the
ambiguities; to the degree they do not resolve ambiguities, analysts risk being
dismissed by annoyed consumers who see them as not having done their job.
Uncertainty reflects inadequacy of data, which is usually assumed to mean lack of
information. But ambiguity can also be aggravated by an excess of data. In attack
warning, there is the problem of "noise" and deception; in operational evaluation
(particularly in a war such as Vietnam), there is the problem of overload from the high
volume of finished analyses, battlefield statistics, reports, bulletins, reconnaissance,
and communications intercepts flowing upward through multiple channels at a rate
exceeding the capacity of officials to absorb or scruitinize them judiciously. (From the
CIA alone, the White House received current intelligence dailies, Weekly Reports,
daily Intelligence Information Cables, occasional Special Reports and specific
memoranda, and analyses from the CIA Vietnam Working Group.) Similarly, in
estimates for defense planning, there is the problem of innumerable and endlessly
refined indices of the strategic balance, and the dependence of assessments of
capabilities on complex and variable assumptions about the doctrine, scenarios, and
intentions that would govern their use.
Because it is the job of decision makers to decide, they cannot react to ambiguity
by deferring judgment.21 When the problem is an environment that lacks clarity, an
overload of conflicting data, and lack of time for rigorous assessment of sources and
validity, ambiguity abets instinct and allows intuition to drive analysis. Intelligence
can fail because the data are too permissive for policy judgment rather than too
constraining. When a welter of fragmentary evidence offers support to various
interpretations, ambiguity is exploited by wishfulness. The greater the ambiguity, the
greater the impact of preconceptions.22 (This point should be distinguished from the
Quoted in SSCI, Final Report (fn. 7), I, 274.
Sherman Kent, "Estimates and Influence," Foreign Service Journal, XLVI (April 1969), 17.
Hughes (fn. 2), 43.
"The textbooks agree, of course, that we should only believe reliable intelligence, and should never
cease to be suspicious, but what is the use of such feeble maxims? They belong to that wisdom which for
want of anything better scribblers of systems and compendia resort to when they run out of ideas." Carl von
Clausewitz, On War, ed. and trans. by Michael Howard and Peter Paret (Princeton: Princeton University
Press 1976), 117.
22 Robert Jervis, The Logic of Images in International Relations (Princeton: Princeton University
Press 1970), 132; Jervis, Perception and Misperception in International Politics (Princeton: Princeton
University Press 1976), chap. 4; Floyd Allport, Theories of Perception and the Concept of Structure, cited
in Shlaim (fn. 2), 358. Cognitive theory suggests that uncertainty provokes decision makers to separate
rather than integrate their values, to deny that inconsistencies between values exist, and even to see
contradictory values as mutually supportive. John Steinbruner, The Cybernetic Theory of Decision
(Princeton: Princeton University Press 1974), 105-8.
Approved For Release 2005/01/06 : CIA-RDP8O-0063OA0001000800SI-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
theory of cognitive dissonance, which became popular with political scientists at the
time it was being rejected by psychologists.) 23 There is some inverse relation between
the importance of an assessment (when uncertainty is high) and the likelihood that it
will be accurate. Lyndon Johnson could reject pessimistic NIE's on Vietnam by
inferring more optimistic conclusions from the reports that came through command
channels on pacification, interdiction, enemy casualties, and defections. Observers
who assume Soviet malevolence focus on analyses of strategic forces that emphasize
missile throw-weight and gross megatonnage (Soviet advantages); those who assume
more benign Soviet intentions focus on analyses that emphasize missile accuracy and
numbers of warheads (U.S. advantages). In assessing the naval balance, Secretary of
L)efense Rumsfeld focused on numbers of ships (Soviet lead), and Congressman Les
Aspin, a critic of the Pentagon, focused on total tonnage (U.S. lead).
2. Ambivalence of judgment. Where there are ambiguous and conflicting
indicators (the context of most failures of intelligence), the imperatives of honesty and
accuracy leave a careful analyst no alternative but ambivalence. There is usually some
evidence to support any prediction. For instance, the CIA reported in June 1964 that a
Chinese instructor (deemed not "particularly qualified to make this remark") had told
troops in a course in guerrilla warfare, "We will have the atom bomb in a matter of
months."-' Several months later the Chinese did perform their first nuclear test. If the
report had been the only evidence, should analysts have predicted the event? If they
are not to make a leap of faith and ignore the data that do not mesh, analysts will issue
estimates that waffle. In trying to elicit nuances of probability from the various
possibilities not foreclosed by the data, cautious estimates may reduce ambivalence,
but they may become Delphic or generalized to the point that they are not useful
guides to decision. (A complaint I have heard in conversations with several U.S.
officials is that many past estimates of Soviet objectives could substitute the name of
any other great power in history-Imperial Rome, 16th-century Spain, Napoleonic
France-and sound equally valid.) Hedging is the legitimate intellectual response to
ambiguity, but it can be politically counterproductive, if the value of intelligence is to
shock consumers out of wishfulness and cognitive insensitivity. A wishful decision
maker can fasten onto that half of an ambivalent analysis that supports his
predisposition .25 A more objective official may escape this temptation, but may
consider the estimate useless because it does not provide "the answer."
3. Atrophy of reforms. Disasters always stimulate organizational changes
designed to avert the same failures in the future. In some cases these changes work. In
many instances, however, the changes persist formally but erode substantively.
Standard procedures are constant. Dramatic failures occur only intermittently. If the
reforms in procedure they have provoked do not fulfill day-to-day organizational
needs-or if, as often happens, they complicate operations and strain the
organization's resources-they fall into disuse or become token practices. After the
" Sec William J. McGuire, "Selective Exposure: A Summing Up," in R. P. Abelson and others, eds.,
Theories of Cognitive Consistency (Chicago: Rand McNally 1968), and Irving L. Janis and Leon Mann,
Decision Making: A Psychological Analysis of Conflict, Choice, and Commitment (New York: Free Press
1977), 213-14.
21 CIA Intelligence Information Cable, "Remarks of the Chief of the Nanking Military Academy and
Other Chinese Leaders on the Situation in South Vietnam," June 25, 1964, in Lyndon B. Johnson Library
National Security Files, Vietnam Country File [hereafter cited as LBJL/NSF-VNCF], Vol. XII, item 55.
's See for example, U.S., Department of Defense, The Senator Gravel Edition: The Pentagon Papers
(Boston: Beacon Press 1971) [hereafter cited as Pentagon Papers], Vol. 11, 99; Frances Fitzgerald, Fire in the
take (Boston: Atlantic-Little, Brown 1972), 364; Special National Intelligence Estimate 53-64, "Chances for
a Stable Government in South Vietnam," September 18, 1964, and McGeorge Bundy's covering letter to the
President, in LBJL/NSF-VNCF, Vol. XIII, item 48.
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 AnaaPiE Y*grF8SoRf 1eecqMn2005/01/06: CIA-RDP80-0063OA000100080001-1
postmortem of North Korea's downing of a U.S. EC-121 monitoring aircraft in 1969,
there was, for several months, a great emphasis on risk assessments for intelligence
collection missions. Generals and admirals personally oversaw the implementaton of
new procedures for making the assessments. Six months later, majors and captains
were doing the checking. "Within a year the paperwork was spot-checked by a major
and the entire community slid back to its old way of making a `quick and dirty'
rundown of the JCS criteria when sending in reconnaissance mission proposals." 26 The
downing of the U-2 over the Soviet Union in 1960 and the capture of the intelligence
ship Pueblo in 1968 had been due in part to the fact that the process of risk assessment
for specific collection missions, primarily the responsibility of overworked middle-
level officers, had become ponderous, sloppy, or ritualized.27 At a higher level, a
National Security Council Intelligence Committee was established in 1971 to improve
responsiveness of intelligence staff to the needs of policy makers. But since the
subcabinet-level consumers who made up the committee were pressed by other
responsibilities, it lapsed in importance and was eventually abolished.28 A comparable
NSC committee that did serve tangible day-to-day needs of consumers to integrate
intelligence and policy-the Verification Panel, which dealt with SALT-was more
effective, but it was issue-oriented rather than designed to oversee the intelligence
process itself. Organizational innovations will not improve the role of intelligence in
policy unless they flow from the decision makers' views of their own needs and unless
they provide frequent practical benefits.
None of these three barriers are accidents of structure or process. They are
inherent in the nature of intelligence and the dynamics of work. As such, they
constitute severe constraints on the efficacy of structural reform.
III. THE ELUSIVENESS OF SOLUTIONS
If they do not atrophy, most solutions proposed to obviate intelligence
dysfunctons have two edges: in reducing one vulnerability, they increase another.
After the seizure of the Pueblo, the Defense Intelligence Agency (DIA) was
reprimanded for misplacing a message that could have prevented the incident. The
colonel responsible developed a careful microfilming operation in the message center
to ensure a record of transmittal of cables to authorities in the Pentagon.
Implementing this check, however, created a three-to-four hour delay--another
potential source of failure-in getting cables to desk analysts whose job was to keep
reporting current.29 Thus, procedural solutions often constitute two steps forward and
one step back; organizational fixes cannot transcend the basic barriers. The lessons of
Pearl Harbor led to the establishment of a Watch Committee and National Indications
Center in Washington. Although this solution eliminated a barrier in the
communication system, it did not prevent the failure of timely alert to the Chinese
intervention in Korea or the 1973 October War, because it did not eliminate the
ambiguity barrier. (Since then, the Watch Committee has been replaced by the DCI's
"Patrick J. McGarvey, CIA: The Myth and the Madness (Baltimore: Penguin 1974), 16.
27 David Wise and Thomas B. Ross, The U-2 Affair (New York: Random House 1962), 56, 176, 180;
Trevor Armbrister, A Matter of Accountability (New York: Coward-McCann 1970), 116-18, 141-45, 159,
187-95; U.S., Congress, House, Committee on Armed Services, Report, Inquiry Into the U.S.S. Pueblo and
EC-121 Plane Incidents [hereafter cited as Pueblo and EC-121 Report], 91st Cong., 1st sess., 1969, 1622-24,
1650-51; U.S., Congress, House, Committee on Armed Services, Hearings, Inquiry Into the U.S.S. Pueblo
and EC-121 Plane Incidents [hereafter cited as Pueblo and EC-121 Hearings], 91st Cong., 1st Bess., 1969,
693-94, 699-700, 703-7, 714, 722, 734, 760, 773-78, 815-16.
21 SSCI, Final Report (fn. 7), I, 61-62; HSCI Draft Report (fn. 4), 82.
29 McGarvey (fn. 26), 16.
Approved For Release 2005/01/06 : CIA-RDP80-00630A0001000800 1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
Strategic Warning Staff.) DIA was reorganized four times within its first ten years; vet
it continued to leave most observers dissatisfied. The Agranat Commission's review of
Israel's 1973 intelligence failure produced proposals for institutional reform that are
striking because they amount to copying the American system of the same time-
which had failed in exactly the same way as the Israeli system.30 Reform is not
hopeless, but hopes placed in solutions most often proposed-such as the following-
should be circumscribed.
1. Assume the worst. A common reaction to traumatic surprise is the
recommendation to cope with ambiguity and ambivalence by acting on the most
threatening possible interpretations. If there is any evidence of threat, assume it is
valid, even if the apparent weight of contrary indicators is greater. In retrospect, when
the point of reference is an actual disaster attributable to a mistaken calculation of
probabilities, this response is always justifiable, but it is impractical as a guide to
standard procedure. Operationalizing worst-case analysis requires extraordinary
expense; it risks being counterproductive if it is effective (by provoking enemy
countermeasures or pre-emption), and it is likely to be ineffective because
routinization will discredit it. Many Israeli observers deduced from the 1973 surprise
that defense planning could only rest on the assumption that no attack warning will be
available, and that precautionary mobilization should always be undertaken even
when there is only dubious evidence of impending Arab action.33 Similarly, American
hawks argue that if the Soviets' intentions are uncertain, the only prudent course is to
assume they are seeking the capability to win a nuclear war.
In either case, the norm of assuming the worst poses high financial costs.
Frequent mobilizations strain the already taut Israeli economy. Moreover, countermo-
bilization can defeat itself. Between 1971 and 1973, the Egyptians three times
undertook exercises similar to those that led to the October attack; Israel mobilized in
response, and nothing happened. It was the paradox of self-negating prophecy.32 The
Israeli Chief of Staff was sharply criticized for the unnecessary Cost. 33 The danger of
hypersensitivity appeared in 1977, when General Gur believed Sadat's offer to come
to Jerusalem to be a camouflage for an Egyptian attack; he began Israeli maneuvers in
the Sinai, which led Egypt to begin maneuvers of its own, heightening the risk of
accidental war.34 To estimate the requirements for deterrence and defense, worst-case
assumptions present an open-ended criterion. The procurement of all the hedges
possible for nuclear war-fighting-large increments in offensive forces, alert status,
hardening of command-control-and-communications, active and passive defenses-
" Shlaim (fn. 2), 375-77. The proposals follow, with their U.S. analogues noted in parentheses: appoint
a special intelligence adviser to the Prime Minister (Director of Central Intelligence) to supplement the
military chief of intelligence; reinforce the Foreign Ministry's research department (Bureau of Intelligence
and Research); more autonomy for non-military intelligence (CIA); amend rules for transmitting raw
intelligence to research agencies, the Defense Minister, and the Prime Minister (routing of signals
intelligence from the National Security Agency); restructure military intelligence (creation of DIA in 1961);
establish a central evaluation unit (Office of National Estimates). On the U.S. intelligence failure in 1973,
see the HSCI Draft Report (fn. 4), 78-79.
Shlaim (fn. 2), 379; Handel (fn. 4), 62-63.
az Ibid., 55.
Shlaim (fn. 2), 358-59. The Israeli command estimated a higher probability of attack in May 1973
than it did in October. Having been proved wrong in May, Chief of Staff Elazar lost credibility in
challenging intelligence officers, complained that he could no longer argue effectively against them, and
consequently was unable to influence his colleagues when he was right. Personal communication from
Michael Handel, November 15, 1977.
.1' Washington Post, November 27, 1977, p. A17.
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Appr d For (ease 2005/01/06: CIA-RDP80-00630A000100080001-1
Analyses, ar and ecision
would add billions to the U.S. defense budget. Moreover, prudent hedging in policy
should be distinguished from net judgment of probabilities in estimates.35
Alternatively, precautionary escalation or procurement may act as self-fulfilling
prophecies, either through a catalytic spiral of mobilization (a la World War I) or an
arms race that heightens tension, or doctrinal hedges that make the prospect of nuclear
war more "thinkable." Since evidence for the "action-reaction" hypothesis of U.S. and
Soviet nuclear policies is meager, and arms races can sometimes be stabilizing rather
than dangerous, the last point is debatable. Still, a large unilateral increase in strategic
forces by either the United States or the Soviet Union would, at the least, destroy the
possibility of gains desired from SALT. A surprise attack or defeat make the costs of
underestimates obvious and dramatic; the unnecessary defense costs due to
overestimates can only be surmised, since the minimum needed for deterrence is
uncertain. Worst-case analysis as a standard norm would also exacerbate the "cry
wolf" syndrome. Unambiguous threat is not an intelligence problem; rather, the
challenge lies in the response to fragmentary, contradictory, and dubious indicators.
Most such indicators turn out to be false alarms. Analysts who reflexively warn of
disaster are soon derided as hysterical. General William Westmoreland recalled that
the warnings that had been issued before the 1968 Tet Offensive were ignored. U.S.
headquarters in Saigon had each year predicted a winter-spring offensive, "and every
year it had come off without any dire results.... Was not the new offensive to be
more of the same?" 36
Given the experience of intelligence professionals that most peacetime indicators
of suspicious enemy activity lead to nothing, what colonel who has the watch some
night will risk "lighting up the board" in the White House simply on the basis of weak
apprehension? How many staffers will risk waking a tired President, especially if they
have done so before and found the action to be needless? How many distracting false
alarms will an overworked President tolerate before he makes it clear that aides should
exercise discretion in bothering him? Even if worst-case analysis is promulgated in
principle, it will be compromised in practice. Routinization corrodes sensitivity. Every
day that an expected threat does not materialize dulls receptivity to the reality of
danger. As Roberta Wohlstetter wrote of pre-Pearl Harbor vigilance, "We are
constantly confronted by the paradox of pessimistic realism of phrase coupled with
loose optimism in practice." 31 Seeking to cover all contingencies, worst-case analysis
loses focus and salience; by providing a theoretical guide for everything, it provides a
practical guide for very little.
2. Multiple advocacy. Blunders are often attributed to decision makers'
inattention to unpopular viewpoints or to a lack of access to higher levels of authority
by dissident analysts. To reduce the chances of such mistakes, Alexander George
proposes institutionalizing a balanced, open, and managed process of debate, so that no
relevant assessments will be submerged by unchallenged premises or the bureaucratic
strength of opposing officials.36 The goal is unobjectionable, and formalized multiple
advocacy certainly would help, not hinder. But confidence that it will help
systematically and substantially should be tentative. In a loose sense, there has usually
15 Raymond Garthoff, "On Estimating and Imputing Intentions," International Security, a (Winter
1978), 22.
?f Westmoreland, A Soldier Reports (Garden City, N.Y.: Doubleday 1976), 316. See the postmortem
by the President's Foreign Intelligence Advisory Board, quoted in Herbert Y. Schandler, The Unmaking of
a President (Princeton: Princeton University Press 1977), 70, 76, 79-80.
a' Wohlstetter (fn. 4), 69.
51 George, "The Case for Multiple Advocacy in Making Foreign Policy," American Political Science
Review, Vol. 66 (September 1972). My usage of the term multiple advocacy is looser than George's.
Approved For Release 2005/01/06 : CIA-RDP80-00630A0001000800$ -1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
been multiple advocacy in the U.S. policy process, but it has not prevented mistakes in
deliberation or decision. Lyndon Johnson did not decide for limited bombing and
gradual troop commitment in Vietnam in 1965 because he was not presented with
extensive and vigorous counterarguments. He considered seriously (indeed solicited)
tinder Secretary of State George Ball's analysis, which drew on NIE's and lower-level
officials' pessimistic assessments that any escalation would be a mistake. Johnson was
also well aware of the arguments by DCI John McCone and the Air Force from the
other extreme-that massive escalation in the air war was necessary because
gradualism would be ineffective." The President simply chose to accept the views of
the middle-of-the-road opponents of both Ball and McCone.
To the extent that multiple advocacy works, and succeeds in maximizing the
number of views promulgated and in supporting the argumentive resources of all
contending analysts, it may simply highlight ambiguity rather than resolve it. In
George's ideal situation, the process would winnow out unsubstantiated premises and
assumptions about ends-means linkages. But in the context of data overload,
uncertainty, and time constraints, multiple advocacy may in effect give all of the
various viewpoints an aura of empirical respectability and allow a leader to choose
whichever accords with his predisposition. 40 The efficacy of multiple advocacy (which
is greatest under conditions of manageable data and low ambiguity) may vary
inversely with the potential for intelligence failure (which is greatest under conditions
Of confusing data and high uncertainty). The process could, of course, bring to the
surface ambiguities where false certainty had prevailed; in these cases, it would be as
valuable as George believes. But if multiple advocacy increases ambivalence and
leaders do not indulge their instincts, it risks promoting conservatism or paralysis.
Dean Acheson saw danger in presidential indecisiveness aggravated by debate: " `I
know your theory,' he grumbled to Neustadt. `You think Presidents should be warned.
You're wrong. Presidents should be given confidence.' " " Even Clausewitz argued
that deference to intelligence can frustrate bold initiative and squander crucial
opportunities. Critics charged Henry Kissinger with crippling U.S. intelligence by
refusing to keep analysts informed of his intimate conversations with foreign leaders.-
To do so, however, would have created the possibility of leaks and might thereby have
crippled his diplomatic maneuvers. It is doubtful that Nixon's initiative to China could
have survived prior debate, dissent, and analysis by the bureaucracy.
It is unclear that managed multiple advocacy would yield markedly greater
benefits than the redundancy and competitiveness that have long existed. (At best it
would perfect the "market" of ideas in the manner that John Stuart Mill believed
"v Henry F. Graff, The Tuesday Cabinet (Englewood Cliffs, N.J.: Prentice-Hall 1970), 68-71; Leslie
H. Gelb with Richard K. Betts, The Irony of Vietnam: The System Worked (Washington, D.C.: Brookings,
forthcoming), chap. 4; Ball memorandum of October 5, 1964, reprinted as "Top Secret: The Prophecy the
President Rejected," Atlantic Monthly, Vol. 230 (July 1972); McCone, memorandum of April 2, 1965, in
I.BJL/NSF-VNCF, Troop Decision folder, item 14b.
"' Betts (fn. 5), 199-202; Schandler (fn. 36), 177. George (fn. 38), 759, stipulates that multiple advocacy
requires "no major maidistribution" of power, influence, competence, information, analytic resources, and
bargaining skills. But, except for resources and the right to representation, the foregoing are subjective
factors that can rarely be equalized by design. If they are equalized, in the context of imperfect data and
time pressure, erroneous arguments as well as accurate ones will be reinforced. Non-expert principals have
difficulty arbitrating intellectually between experts who disagree.
Quoted in Steinbruner (fn. 22), 332.
Clausewitz (fn. 21), 117-18; HSCI Hearings (fn. 4), 634-36; William J. Barnds, "Intelligence and
Policymaking in an Institutional Context," in U.S., Commission on the Organization of the Government for
the Conduct of Foreign Policy [hereafter cited as Murphy Commission], Appendices (Washington, D.C.:
GP.O., June 1975), Vol. VII, 32.
Aplgoved For Release 2005/01/06 : CIA-RDP8O-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
made liberalism conducive to the emergence of truth.) The first major reorganization
of the American intelligence community in 1946-1947 emphasized centralization in
order to avert future Pearl Harbors caused by fragmentation of authority; the latest
reorganization (Carter's 1977 extension of authority of the Director of Central
Intelligence over military intelligence programs) emphasized centralization to
improve efficiency and coherence. Yet decentralization has always persisted in the
overlapping division of labor between several separate agencies. Recent theorists of
bureaucracy see such duplication as beneficial because competition exposes
disagreement and presents policy makers with a wider range of views. Redundancy
inhibits consensus, impedes the herd instinct in the decision process, and thus reduces
the likelihood of failure due to unchallenged premises or cognitive errors. To ensure
that redundancy works in this way, critics oppose a process that yields coordinated
estimates-negotiated to the least common denominator, and cleared by all agencies
before they are passed to the principals. George's "custodian" of multiple advocacy
could ensure that this does not happen. There are, of course, trade-off costs for
redundancy. Maximization of competition limits specialization. In explaining the
failure of intelligence to predict the 1974 coup in Portugal, William Hyland pointed
out, "if each of the major analytical components stretch their resources over the same
range, there is the risk that areas of less priority will be superficially covered." 43
The problem with arguing that the principals themselves should scrutinize
numerous contrasting estimates in their integrity is that they are constantly
overwhelmed by administrative responsibilities and "action items"; they lack the time
to read, ponder, and digest that large an amount of material. Most intelligence
products, even NIE's, are never read by policy makers; at best, they are used by
second-level staffers as background material for briefing their seniors.44 Consumers
want previously coordinated analyses in order to save time and effort. In this respect,
the practical imperatives of day-to-day decision contradict the theoretical logic of
ideal intelligence.
3. Consolidation. According to the logic of estimative redundancy, more analysis
is better than less. Along this line of reasoning, Senate investigators noted critically
that, as of fiscal year 1975, the U.S. intelligence community still allocated 72 percent
of its budget for collection of information, 19 percent for processing technical data,
and less than 9 percent for production of finished analyses. On the other hand
according to the logic of those who focus on the time constraints of leaders and the
confusion that results from innumerable publications, quantity counteracts quality.
The size of the CIA's intelligence directorate and the complexity of the production
process "precluded close association between policymakers and analysts, between the
intelligence product and policy informed by intelligence analysis." 45 For the sake of
clarity and acuity, the intelligence bureaucracy should be streamlined.
This view is consistent with the development of the Office of National Estimates
(ONE), which was established in 1950 and designed to coordinate the contributions of
the various organs in the intelligence community for the Director of Central
Intelligence. DCI Walter Bedell Smith envisioned an operation of about a thousand
people. But William L. Langer, the scholar Smith imported to organize ONE, wanted
45 HSCI, Hearings (fn. 4), 778.
44 SSCI, Final Report (fn. 7), IV, 57; Roger Hilsman, Strategic Intelligence and National Decisions
(Glencoe, Ill.: Free Press 1956), 40. During brief service as just a low-level staff member of the National
Security Council, even I never had time to read. all the intelligence analyses relevant to my work.
11 SSCI, Final Report (fn, 7), I, 344, and IV, 95 (emphasis deleted).
Approved For Release 2005/01/06: CIA-RDP80-00630A0001000800V-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
a tight group of excellent analysts and a personnel ceiling of fifty. Langer prevailed,
and though the number of staff members in ONE crept upwards, it probably never
exceeded a hundred in its two decades of existence. 16 Yet ONE could not eliminate the
complexity of the intelligence process; it could only coordinate and integrate it for the
production of National Intelligence Estimates. Other sources found conduits to
decision makers (to Cabinet members through their own agencies, or to the President
through the National Security Council). And some policy makers, though they might
dislike the cacophony of multiple intelligence agencies, were suspicious of the
consolidated NIE's, knowing that there was pressure to compromise views in order to
gain agreement. Over time, the dynamics of bureaucracy also blunted the original
objectives of ONE's founder. From a cosmopolitan elite corps, it evolved into an
insular unit of senior careerists from the CIA. The National Intelligence Officer
system that replaced ONE reduced the number of personnel responsible for
coordinating NIE's, but has been criticized on other grounds such as greater
vulnerability to departmental pressures. Bureaucratic realities have frustrated other
attempts to consolidate the intelligence structure. The Defense Intelligence Agency
was created in 1961 to unify Pentagon intelligence and reduce duplicative activities of
the three service intelligence agencies, but these agencies regenerated themselves; in
less than a decade they were larger than they had been before DIA's inception."
The numerous attempts to simplify the organization of the analytic process thus
have not solved the major problems. Either the streamlining exercises were short-lived,
and bureaucratization crept back, or the changes had to be moderated to avoid the
new dangers they entailed. Contraction is inconsistent with the desire to minimize
failure by "plugging holes" in intelligence, since compensating for an inadequacy
usually requires adding personnel and mechanisms; pruning the structure that
contributes to procedural sluggishness or complexity may create lacunae in substantive
coverage.
4. Devil's advocacy. Multiple advocacy ensures that all views held by individuals
within the analytic system will be granted serious attention. Some views that should
receive attention, however, may not be held by anyone within the system. Virtually no
analysts in Israel or the United States believed the Arabs would be "foolish" enough to
attack in 1973. Many observers have recommended institutionalizing dissent by
assigning to someone the job of articulating apparently ridiculous interpretations to
ensure that they are forced into consideration. Establishing an official devil's advocate
would probably do no harm (although some argue that it may preversely facilitate
consensus-building by domesticating true dissenters or providing the illusory comfort
that all views have been carefully examined; 41 worse, it might delude decision makers
into believing that uncertainties have been resolved). But in any case, the role is likely
to atrophy into a superfluous or artificial ritual. By the definition of the job, the devil's
advocate is likely to be dismissed by decision makers as a sophist who only makes an
argument because he is supposed to, not because of its real merits. Institutionalizing
devil's advocacy is likely to be perceived in practice as institutionalizing the "cry
wolf" problem; "There are limits to the utility of a `devil's advocate' who is not a true
devil." i6 He becomes someone to be indulged and disregarded. Given its rather sterile
Ray S. Cline, Secrets, Spies, and Scholars (Washington, D.C.: Acropolis 1976), 20.
Gilbert W. Fitzhugh and others, Report to the President and the Secretary of Defense on the
Department of Defense, By the Blue Ribbon Defense Panel (Washington, D.C.: G.P.O., July 1970), 45-46.
" Alexander George, "The Devil's Advocate: Uses and Limitations," Murphy Commission,
Appendices (fn. 42), 11, 84-85; Jervis, Perception and Misperception (fn. 22), 417.
- Ibid., 416.
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
definition, the role is not likely to be filled by a prestigious. official (who will prefer
more "genuine" responsibility); it will therefore be easier for policy makers to dismiss
the arguments. In order to avert intelligence failures, an analyst is needed who tells
decision makers what they don't want to hear, dampening the penchant for wishful
thinking. But since it is the job of the devil's advocate to do this habitually, and since
he is most often wrong (as would be inevitable, since otherwise the conventional
wisdom would eventually change), he digs his own grave. If the role is routinized and
thus ritualized, it loses impact; but if it is not routinized,. there can be no assurance that
it will be operating when it is needed.
Despite the last point, which is more important in attack warning than in
operational evaluation or defense planning, there is a compromise that offers more
realistic benefits: ad hoc utilization of "real devils." This selective or biased form of
multiple advocacy may be achieved by periodically giving a platform within the
intelligence process to minority views that can be argued more persuasively by
prestigious analysts outside the bureaucracy. This is what the President's Foreign
Intelligence Advisory Board and DCI George Bush did in 1976 by commissioning the
"Team B" critique of NIE's on Soviet strategic objectives and capabilities. Dissenters
within the intelligence community who were skeptical of Soviet intentions were
reinforced by a panel of sympathetic scholars, with a mandate to produce an analysis
of their own.50 This controversial exercise, even if it erred in many of its own ways (as
dovish critics contend), had a major impact in promoting the re-examination of
premises and methodology in U.S. strategic estimates. The problem with this option is
that it depends on the political biases of the authorities who commission it. If it were
balanced by a comparable "Team C" of analysts at the opposite extreme (more
optimistic about Soviet intentions than the intelligence community consensus), the
exercise would approach regular multiple advocacy, with the attendant limitations of
that solution. Another variant would be intermittent designation of devil's advocates in
periods of crisis, when the possibility of disaster is greater than usual. Since the role
would then be fresh each time, rather than ritualized, the advocate might receive a
more serious hearing. The problem here is that receptivity of decision makers to
information that contradicts preconceptions varies inversely with their personal
commitments, and commitments grow as crisis progresses.51
5. Sanctions and incentives. Some critics attribute intelligence failures to
dishonest reporting or the intellectual mediocrity of analysts. Suggested remedies
include threats of punishment for the former, and inducements to attract talent to
replace the latter. Other critics emphasize that, will or ability aside, analytic integrity
is often submerged by the policy makers' demands for intelligence that suits them;
"the NIEs ought to be responsive to the evidence, not the policymaker." 52 Holders of
this point of view would institutionalize the analysts' autonomy. Unobjectionable in
principle (though if analysts are totally unresponsive to the consumer, he will ignore
them), these implications cannot easily be operationalized without creating as many
problems as they solve.
Self-serving operational evaluations from military sources, such as optimistic
reports on progress in the field in Vietnam or pessimistic strategic estimates, might
SO U.S., Congress, Senate, Select Committee on Intelligence, Report, The National Intelligence
Estimates A-B Team Episode Concerning Soviet Capability and Objectives, 95th Cong. 2d sess., 1978;
New York Times, December 26, 1976, pp. 1, 14; Washington Post, January 2, 1977, pp. Al, A4.
" George H. Poteat, "The Intelligence Gap: Hypotheses on the Process of Surprise," International
Studies Notes, 111 (Fall 1976), 15.
sz Cline (fn. 46), 140.
Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001R1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
indeed be obviated if analysts in DIA, the service intelligence agencies, and command
staffs were credibly threatened with sanctions (firing, nonpromotion, reprimand, or
disgrace). Such threats theoretically could be a countervailing pressure to the career
incentives analysts have to promote the interests of their services. But, except in the
most egregious cases, it is difficult to apply such standards without arbitrariness and
bias, given the problem of ambiguity; it simply encourages an alternative bias or
greater ambivalence. Moreover, military professionals would be in an untenable
position, pulled in opposite directions by two sets of authorities. To apply the
sanctions, civil authorities would have to violate the most hallowed military canon by
having civilian intelligence officials interfere in the chain of command. In view of
these dilemmas, it is easier to rely on the limited effectiveness of redundancy or
multiple advocacy to counteract biased estimates.
Critics concerned with attracting better talent into the analytic bureaucracy
propose to raise salaries and to provide more high-ranking positions (supergrades) to
which analysts can aspire. Yet American government salaries are already very high by
academic standards. Those who attribute DIA's mediocrity (compared to CIA), to an
insufficient allocation of supergrades and a consequent inability to retain equivalent
personnel are also mistaken; as of 1975 the difference in the grade structures of DIA
and CIA had been negligible.53 And the fact that CIA analysts cannot rise to a
supergrade position (GS-16 to 18) without becoming administrators is not convincing
evidence that good analysts are underpaid; GS-15 salaries are higher than the
maximum for most tenured professors.
Non-military analysts, or high-ranking soldiers with no promotions to look
forward to, have fewer professional crosspressures to contend with than military
intelligence officers. But an analyst's autonomy varies inversely with his influence, and
hortatory injunctions to be steadfast and intellectually honest cannot ensure that he
will be; they cannot transcend political realities or the idiosyncrasies of leaders.
Richard Helms notes that "there is no way to insulate the DCI from unpopularity at
the hands of Presidents or policymakers if he is making assessments which run counter
to administrative policy. That is a built-in hazard of the job. Sensible Presidents
understand this. On the other hand they are human too." Integrity untinged by
political sensitivity courts professional suicide. If the analyst insists on perpetually
bearing bad news, he is likely to be beheaded. Helms himself succumbed to policy
makers' pressures in compromising estimates of the MIRV capabilities of the Soviet SS-9
missile in 1969, and the prospects for Cambodia in 1970.54 The same practical
psychological constraints are reflected in an incident in which Chief of Naval
Operations Elmo Zumwalt, who had already infuriated Nixon and Kissinger several
times with his strategic estimates, was determined to present yet another unwelcome
analysis; Secretary of Defense Schlesinger dissuaded him with the warning, "To give a
briefing like that in the White House these days would be just like shooting yourself in
the foot." 5J
6. Cognitive rehabilitation and methodological consciousness. The intertwining
of analysis and decision and the record of intelligence failures due to mistaken
"' SSCI, Final Report (fn. 7), I, 352. A valid criticism is that military personnel systems and promotion
standards penalized intelligence officers, thus encouraging competent officers to avoid intelligence
assignments. t'his situation was rectified in the service intelligence agencies by the early 1970s, but not
within DIA. Ibid.; Betts (fn. 5), 196-97.
" SSCI, Final Report (fn. 7), I, 77-82. See also U.S., Congress, Senate, Committee on Foreign
Relations, Hearings, National Security Act Amendment, 92d Cong., 2d sess., 1972, 14-24.
"Zumwalt, On Watch (New York: Quadrangle 1976), 459.
AWroved For Release 2005/01/06: CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
Analysis, War and Decision
preconceptions and unexamined assumptions suggest the need to reform the
intelligence consumers' attitudes, awareness, and modes of perception. If leaders were
made self-conscious and self-critical about their own psychologies, they might be less
vulnerable to cognitive pathologies. This approach to preventing intelligence failure is
the most basic and metaphysical. If policy makers focused on the methodologies of
competing intelligence producers, they would be more sensitive to the biases and leaps
of faith in the analyses passed to them. "In official fact-finding ... the problem is not
merely to open up a wide range of policy alternatives but to create incentives for
persistent criticism of evidentiary value." 51 Improvement would flow from
mechanisms that force decision makers to make explicit rather than unconscious
choices, to exercise judgment rather than engage in automatic perception, and to
enhance their awareness of their own preconceptions.57
Unlike organizational structure, however, cognition cannot be altered by
legislation. Intelligence consumers are political men who have risen by being more
decisive than reflective, more aggressive than introspective, and confident as much as
cautious, Few busy activists who have achieved success by thinking the way that they
do will change their way of thinking because some theorist tells them to. Even if they
could be forced to confront scholarly evidence of the dynamics of misperception, it is
uncertain that they could consistently internalize it. Preconception cannot be
abolished; it is in one sense just another word for "model" or "paradigm"-a construct
used to simplify reality, which any thinker needs in order to cope with complexity.
There is a grain of truth in the otherwise pernicious maxim that an open mind is an
empty mind. Moreover, the line between perception and judgment is very thin, and
consumers cannot carefully scrutinize, compare, and evaluate the methodologies of
competing analyses, for the same prosaic reason (the problem of expertise aside) that
impedes many proposed reforms: they do not have the time to do so. Solutions that
require principals to invest more attention than they already do are conceptually valid
but operationally weak. Ideally, perhaps, each principal should have a Special
Assistant for Rigor Enforcement.
Although most notable intelligence failures occur more often at the consuming
than the producing end, it is impractical to place the burden for correcting those faults
on the consumers. The most realistic strategy for improvement would be to have
intelligence professionals anticipate the cognitive barriers to decision makers'
utilization of their products. Ideally, the Director of Central intelligence should have a
theoretical temperament and personal skills in forcing unusual analyses to the
attention of principals; he might act as George's "custodian" of the argumentation
process. To fulfill this function, the DCI should be not only a professional analyst and
an intellectual (of the twelve DCI's since 1946, only James Schlesinger met those
criteria, and he served for only three months), but also a skilled bureaucratic
politician. These qualifications seldom coincide. The DCI's coordinating staff and
National Intelligence Officers should be adept at detecting, making explicit, and
exposing to consumers the idiosyncracies in the assessments of various agencies-the
reasons that the focus and conclusions of the State Department's Bureau of
Intelligence and Research differ from those of DIA, or of naval intelligence, or of the
CIA. For such a procedure to work, the consumers would have to favor it (as opposed
to negotiated consensual estimates that would save them more time). There is always a
latent, tension between what facilitates timely decision and what promotes
thoroughness and accuracy in assessment. The fact that there is no guaranteed
Wilensky (fn. 11), 164.
Jervis, Perception and Misperception (fn. 22), 181-87.
Approved For Release 2005/01/06 : CIA-RDP80-00630A00010008001
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Analysis, War and Decision
prophylaxis against intelligence failures, however, does not negate the value of
incremental improvements. The key is to see the problem of reform as one of modest
refinements rather than as a systematic breakthrough.
W. LIVING WITH FATALISM
Organizational solutions to intelligence failure are hampered by three basic
problems: most procedural reforms that address specific pathologies introduce or
accent other pathologies; changes in analytic processes can never fully transcend the
constraints of ambiguity and ambivalence; and more rationalized information systems
cannot fully compensate for the predispositions, perceptual idiosyncrasies, and time
constraints of political consumers. Solutions that address the psychology and analytic
style of decision makers are limited by the difficulty of changing human thought
processes and day-to-day habits of judgment by normative injunction. Most theorists
have thus resigned themselves to the hope of marginal progress, "to improve the
`batting average'-say from .275 to .301-rather than to do away altogether with
surprise.''
There is some convergence in the implications of all three ways of
conceptualizing intelligence failures. Mistakes should be expected because the
paradoxes are not resolvable; minor improvements are possible by reorganizing to
correct pathologies; and despair is unwarranted because, seen in perspective, the
record could he worse. Marginal improvements have, in fact, been steadily instituted
since World War Ii. Although many have indeed raised new problems, most have
yielded a net increase in the rationalization of the system. The diversification of
sources of estimates of adversaries' military power has grown consistently, obviating
the necessity to rely exclusively on military staffs. The resources and influence of
civilian analysts of military data (principally in the CIA's Office of Strategic Research
but also in its Directorate of Science and Technology) are unparalleled in any other
nation's intelligence system. At the same time, the DCI's mechanism for coordinating
the activities of all agencies-the Intelligence Community Staff-has grown and
become more diverse and representative, and less an extension of the CIA, as more
staffers have been added from the outside. In 1972, a separate Product Review
Division was established within the staff to appraise the "objectivity, balance, and
responsiveness" of intelligence studies on a regular basis. It has conducted
postmortems of intelligence failures since then (the Yom Kippur War, the Cyprus
crisis of 1974, the Indian nuclear test, and the seizure of the Mayaguez).- (Previously,
postmortems had been conducted by the analysts who had failed, a procedure that
hardly guaranteed objectivity.)
Within the Pentagon, capabilities for estimates relevant to planning were
enhanced with the establishment of an office for Net Assessment, which analyzes the
significance of foreign capabilities in comparison with U.S. forces. (CIA, DIA, and
NIE's only estimate foreign capabilities.) Civilian direction of military intelligence
was reinforced by an Assistant Secretary of Defense for Intelligence after the 1970
recommendation of the Fitzhugh Commission, and an Under Secretary for Policy in
1978. .Experiments in improving communication between producers and consumers
have been undertaken (such as, for example, the testing of a Defense Intelligence
"Knorr (fn. 1), 460.
," SSCI, Final Report (fn. 7), I, 276, and IV, 85; U.S., Congress, House, Committee on Appropriations,
Hearings, Supplemental Appropriations for Fiscal Year 1977, 95th Cong., 2d sess., 1977, 515-621;
Washington Post, February 15, 1977, p. A6; Paul W. Blackstock, "The Intelligence Community Under the
Nixon Administration," Armed Forces and Society, 1 (February 1975), 238.
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA0001 00080001 -1
Analysis, War and Decision
Board in late 1976). The dominance of operators within the intelligence community
has also waned-especially since the phasing out of paramilitary operations in
Southeast Asia and the severe reductions in size and status of CIA's covert action
branch that began in 1973. Dysfunctions in the military communications system,
which contributed to crises involving intelligence collection missions in the 1960s (the
Israeli attack on the U.S.S. Liberty and North Korea's seizure of the Pueblo) were
alleviated (though not cured) by new routing procedures and by instituting an
"optimal scanning system" in the Pentagon.?0 Statistical analyses of strategic power
have become progressively more rigorous and comprehensive; as staffs outside the
executive branch-such as the Congressional Budget Office-have become involved
in the process, they have also become more competitive.?1
Few of the changes in structure and process have generated more costs than
benefits. (Some critics believe, however, that the abolition of the Office and Board of
National Estimates and their replacement with National Intelligence Officers was a
net loss.) But it is difficult to prove that they have significantly reduced the incidence
of intelligence failure. In the area of warning, for instance, new sophisticated
coordination mechanisms have recently been introduced, and since the institution at
the time of the 1974 Cyprus crisis of DCI "alert memoranda"-"brief notices in a
form which cannot be overlooked" --no major warning failure has occurred. But the
period of testing is as yet too brief to demonstrate that these adaptations are more
effective than previous procedures. In the area of operational evaluation, it is clear
that there was greater consciousness of the limitations and cost-ineffectiveness of aerial
bombardment during the Vietnam War than there had been in Korea, due largely to
the assessments made by the offices of Systems Analysis and International Security
Affairs in the Pentagon and Secretary of Defense McNamara's utilization of CIA
estimates and contract studies by external analytic organizations." Yet this greater
consciousness did not prevail until late in the war because it was not a consensus; Air
Force and naval assessments of bombing effectiveness contradicted those of the critical
civilian analysts. Nor has the elaboration and diversification of analytic resources for
strategic estimates clearly reduced the potential for erroneous planning decisions.
Determination of the salience and proper weight of conflicting indicators of strategic
power and objectives or of the comparative significance of quantitative and
10 Joseph C. Goulden, Truth is the First Casualty (Chicago: Rand McNally 1.969), 101-4; Phil G.
Goulding, Confirm or Deny (New York: Harper & Row 1970), 130-33, 269; Pueblo and EC-121 Hearings
(fn. 27), 646-47, 665-73, 743-44, 780-82, 802-3, 865-67, 875, 880, 897-99; Pueblo and EC-121 Report (fn.
27), 1654-56, 1662-67; Armbrister (fn. 27), 196ff, 395; U.S., Congress, House, Committee on Armed
Services, Report, Review of Department of Defense Worldwide Communications: Phase 1, 92d Cong., 1st
sess., 1971, and Phase II, 2d sess., 1972.
f1 See, for example, James Blaker and Andrew Hamilton, Assessing the NATO/Warsaw Pact Military
Balance (Washington, D.C.: Congressional Budget Office, December 1977).
12 SSCI, Final Report (fn. 7), I, 61; Thomas G. Belden, "Indications, Warning, and Crisis Operations,"
International Studies Quarterly, XXI (March 1977), 192-92.
?a Pentagon Papers, IV, 111-12, 115-24, 217-32. CIA critiques of bombing results began even before
the Tonkin Gulf crisis. CIA/OCI, Current Intelligence Memorandum, "Effectiveness of T-28 Strikes in
Laos," June 26, 1964; CIA/DDI, Intelligence Memorandum, "Communist Reaction to Barrel Roll Missions,"
December 29, 1964. But ambivalence remained even within the CIA, which occasionally issued more
sanguine evaluations-e.g., CIA Memorandum for National Security Council, "The Situation in Vietnam,"
June 28, 1965 (which McGeorge Bundy called directly to the President's attention), and CIA/OCI,
Intelligence Memorandum, "Interdiction of Communist Infiltration Routes in Vietnam," June 24, 1965. (All
memoranda are in LBJL/NSF-VNCF, Vol. I, item 5, Vol. III, items 28, 28a, 28b, Vol. VI A, items 4, 5, 8.)
See also Pentagon Papers, IV, 71-74. See also the opposing assessments of the CIA, the civilian analysts in
the Pentagon, and the Joint Chiefs in NSSM-1 (the Nixon Administration's initial review of Vietnam policy),
reprinted in the Congressional Record, Vol. 118, part 13, 92d Cong., 2d sess., May 10, 1972, pp. 16749-836.
Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001Lfr1
 Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
Analysis, War and Decision
qualitative factors is inextricable from the political debate over foreign policy:
uncertainties always remain, leaving the individual's visceral fears or hopes as the
elements that tilt the balance of judgment.
Although marginal reforms may reduce the probability of error, the unresolvable
paradoxes and barriers to analytic and decisional accuracy will make some incidence
of failure inevitable. Concern with intelligence failure then coincides with concern
about how policy can hedge against the consequences of analytic inadequacy.
Covering every hypothetical vulnerability would lead to bankruptcy, and hedging
against one threat may aggravate a different one. The problem is thus one of priorities,
and hedging against uncertainty is hardly easier than resolving it. Any measures that
clarify the cost-benefit trade-offs in policy hedges are measures that mitigate the
danger of intelligence failure.
One reasonable rule in principle would be to survey the hypothetical outcomes
excluded by strategic premises as improbable but not impossible, identify those that
would be disastrous if they were to occur, and then pay the price to hedge against
them. This is no more practicable, however, than the pure form of worst-case analysis,
because it requires willingness to bear and inflict severe costs for dubious reasons.
Escalation in Vietnam, after all, was a hedge against allowing China to be tempted to
"devour" the rest of Southeast Asia. The interaction of analytic uncertainty and
decisional prudence is a vicious circle that makes the segregation of empirical
intelligence and normative policy an unattainable Platonic ideal.
In the simplest situation, the intelligence system can avert policy failure by
presenting relevant and undisputed facts to non-expert principals who might
otherwise make decisions in ignorance. But these simple situations are not those in
which major intelligence failures occur. Failures occur when ambiguity aggravates
ambivalence. In these more important situations-Acheson and Clausewitz to the
contrary-the intelligence office may perform most usefully by not offering the
answers sought by authorities, but by offering questions, acting as a Socratic agnostic,
nagging decision makers into awareness of the full range of uncertainty, and making
the authorities' calculations harder rather than easier. Sensitive leaders will reluctantly
accept and appreciate this function. Most leaders will not; they will make mistakes,
and will continue to bear the prime responsibility for "intelligence" failures. Two
general values (which sound wistful in the context of the preceding fatalism) remain to
guide the choice of marginal reforms: anything that facilitates dissent and access to
authorities by intelligence producers, and anything that facilitates skepticism and
scrutiny by consumers. The values are synergistically linked; one will not improve the
use of intelligence without the other. (A third value, but one nearly impossible to
achieve, would be anything that increases the time available to principals for reading
and reflection.)
Intelligence failures are not only inevitable, they are natural. Some are even
benign (if a success would not have changed policy). Scholars cannot legitimately view
intelligence mistakes as bizarre, because they are no more common and no less
excusable than academic errors. They are less forgivable only because they are more
consequential. Error in scholarship. is resolved dialectically, as deceptive data are
exposed and regnant theories are challenged, refined, and replaced by new research. If
decision makers had but world enough and time, they could rely on this process to
solve their intelligence problems. But the press of events precludes the luxury of
letting theories sort themselves out over a period of years, as in academia. My survey
of the intractability of the inadequacy of intelligence, and its inseparability from
mistakes in decision, suggests one final conclusion that is perhaps most outrageously
fatalistic of all: tolerance for disaster.
Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
ULTRA GOES TO WAR by Ronald Lewin. McGraw-Hill, New York, 1978.
DEADLY MAGIC by Edward Van Der Rhoer. Charles Scribner Sons, New York,
1978.
We have here two recent publications focused on the European and the Pacific
theaters of World War II and both worthy of your attention. Each is substantially
correct in broad outline and in its account of how the products of cryptanalysis
influenced the outcome of armed conflict in the two theaters. Each contains a large
number of technical errors when examined critically from the cryptanalyst's
viewpoint, but these errors do not detract from the bigger picture the books portrary.
Correction of them would require classification of this review whilst contributing very
little more to the broad outline of the presentation.
Lewin's book repeatedly highlights one of the more important aspects of national
strength, namely that good intelligence is neither a substitute for good generalship nor
can it take the place of effective weaponry well used by motivated personnel. Most of
us in the intelligence community well know that only soldiers, sailors, and aviators
actually win wars. Without substantially more than token force in the right place at
the right time one seldom can expect even a draw. But with the prolific and accurate
source of information provided by carefully selected Ultra decrypts it became possible
to use limited resources to hold in check or prevail over markedly superior forces.
These limited resources, however, had to exist in actuality and sufficiency in the right
spot at the right time-even Ultra could not compensate for the overwhelming
German superiority in Greece and Crete.
The almost daily problems of managing and "deploying" scarce cryptanalytic
personnel and machine resources are mentioned more than once. Lewin provides an
excellent account of the internal and informal working organization at Bletchley Park
and the personal relationships between the senior personnel that greatly facilitated the
frequent shifts in emphasis made to meet developing situations. The published
selections from the Enigma decrypts he had available make immediately obvious to
the reader that even a decrypted operations order is often not very illuminating unless
read in full context. Relevant collateral information, operational background, and a
thorough understanding of the scene as viewed by the German commanders were
always necessary to get more than superficialities from most decrypts. Of course, this
seems a trivial observation once the reader reflects that two correspondents send and
receive messages based not only on the actual operational scene but also on their
background and briefing and on their service experience. The translators and
intelligence analysts in Hut 3 at Bletchley usually managed to project themselves into
the middle of the German operational planning and execution.
The account of how Ultra was made available for use by field commanders while
at the same time protecting its source is excellently done. That not all commanders
made effective use of it, either because of vanity or mistrust or ineptitude, is set forth
clearly and dispassionately. It also becomes very obvious that some claims to intuitive
knowledge of German operations could well have been credited to Ultra.
MORI/HRP
from pg.
55-58
Approved For Release 2005/01/06 : CIA-RDP80-0063OA00010008000151
 Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
Books
For the first time in the present spate of somewhat repetitive books which
purport to pull aside the veil of secrecy over the US/UK Sigint operations in World
War II there is an intelligence appreciation of "Y" material and its essential
complementarity to Ultra. "Y" material comes from such Sigint field operations as
direction finding and position fixing of emitters, decryption of lower level codes and
ciphers, plain language interception and exploitation, traffic analysis, and similar
operations. The author has done a good job of showing how the two interworked as the
British used them. U.S. practices were similar but not identical to those of the U.K. in
the collection and melding of field-derived communications intelligence material with
available Ultra from American or British sources.
Another point clearly made is that of the need for overriding security lest the
source dry up and be unavailable when needed most. The myth that Churchill knew
in advance of the Coventry raid is again effectively demolished. This reviewer
believes, however, that even had the exact location of the raid been known but
attributable only to Ultra without a plausible cover story, the decision would have
been not to warn Coventry in any open manner. At that stage in the war preservation
of' the source was of utmost importance because Ultra was still essential for finishing
the battle of Britain and winning the battle of the Atlantic. Some steps might have
been taken to minimize casualties but an open warning and a panic evacuation would
have hopelessly compromised the source. Churchill and his cabinet were spared that
problem.
Lewin's book is flawed in that he has had no access to nor apparent knowledge of
the U.S. Enigma decrypts, a major effort of comparable results fully shared with our
British allies. He probably had little if any contact with the sizeable U.S. product in
this area or else he did not recognize it. Nor had he much if any contact with any
German naval Enigma decrypts* and, as judged from his book, no knowledge
whatever of the major U.S. contributions to this battle area. This is not to say that he
has not mentioned U.S. personnel who were stationed at Bletchley, but his
relationships with them were apparently those between users and intelligence analysts.
Finally, it might be best simply to ignore the author's very sketchy and limited
account of the Pacific war and the American use of material containing the same
word, Ultra, but derived from decrypts which involved neither the Enigma nor the
"Purple." His sketchy account of the war in that theater is woefully incomplete,
inaccurate, and contributes nothing to what is an otherwise excellent study of the uses
and influence of Ultra on the campaigns of the European theater.
Van Der Rhoer's Deadly Magic is a major contrast in style and scope to Lewin's
hook.
Deadly Magic is the author's intensely personal account of day-to-day actions of
one section of the U.S. Navy's Op-20-G, today the Naval Security Group, as seen
through the eyes of a linguist and intelligence analyst. His contact with the Japanese
naval material was direct and first hand but it began with the receipt of decrypts from
the analytic sections and ended with the forwarding of translations and relevant data
to senior watch officers. Because of these circumstances one cannot expect any
personal observations of high-level policy decisions and reactions, nor are there any.
The book, however, does contain an excellent and balanced account of the use of
Sigint in the naval war in the Pacific. The battles of Midway and the Coral Sea, the
shoot-down of Admiral Yamamoto, the Guadalcanal campaign, the war against the
AAroved For Release 2005/01/06: CIA-RDP80-00630A000100080001-1
 Books
Japanese "marus," and the final naval battles are among the areas interestingly
handled and highlighted. In these accounts, Van Der Rhoer manages to shift skillfully
from the first-person, I-was-there, type of account to a carefully presented historical
description of what occurred as it is now known.
In April 1942, the author arrived for duty in Op-20-GZ after participating for
some months in a U.S. Navy Japanese language course taught at Harvard. He mentions
that he had done several years of private study at a Japanese Buddhist church in New
York City, originally intended for his use in commercial pursuits. Accordingly, he was
not personally involved in any of the pre-war cryptanalytic effort against the Jap-
anese "Red" or "Purple" machines nor in any of aspects of the 14-part message of
6-7 December 1941. His account of those events and the reactions of official
Washington is, however, substantially correct but without any of the personal touches
so often found in the rest of his book. His reminiscences about his colleagues, many of
them known to this reviewer, add interest and serve to recapture the spirit of the
times.
The account of the planning for the interception and shoot-down of Yamamoto
gives an interesting replay of the intensive soul searching by certain high-ranking
Washington officials before the operation was approved. That aspect of the operation
is not usually related. What is omitted, however, is any mention of the intense de-
bate that occurred between those who were afraid that the coincidence of having 16
P-38s at the extreme radius of their range at the exact time Yamamoto's plane would
pass (which information was available only from high-level decrypts, i.e., Ultra), and
those who felt that the death of Yamamoto would more than compensate for possible
hazarding of the source. What finally influenced the decision to attempt the
interception was the realization that there was little, if anything, the Japanese
communications security personnel could do in either the short or the long run to
effect drastic changes in the major operational codes because of the insuperable
physical problems of making distribution of new codes and additive pad materials to
their far-flung outposts and ships. Several times during the war in the Pacific the
Japanese attempted changes and each time they were defeated by clever
cryptanalysis, facilitated to some extent by the major problems of production and
distribution of the voluminous materials needed for use at so many different and
widely separated points. The attrition of the Japanese merchant shipping, the
"marus," is shown to have been made possible by a steady flow of decrypted noon
positions to U.S. submarines for interception. One or two sample messages give a good
idea of the usual contents on which action was taken.
No doubt because of his work in Op-20-GZ, which was not directly concerned
with any cryptanalysis, Van Der Rhoer badly mixes up the "Purple," the German-
produced "Enigma," and miscellaneous codes. He is quite inaccurate when he says
that the Purple was an adaptation of. the Enigma-it simply was not, and its
cryptography was completely different from that of the Enigma. Nor did the Japanese
Navy use the Purple or regularly employ any other machine. Whether or not the
existence of the Enigma machine inspired the Japanese cryptographers to develop a
machine peculiarly their own is purely speculative, but the resulting device was not
even a remote second cousin of the Enigma. The principal Japanese naval operational
code was the so-called JN-25 with its various derivatives, enciphered with a
periodically changed additive of increasing complexity as the war went on. This
cryptanalysis was, however, completely separate from GZ, in which the author spent
his war years. This and a few other errors in fact do not seriously detract from Van
Der Rhoer's very readable account of the naval war in the Pacific.
Approved For Release 2005/01/06 : CIA-RDP80-00630A00010008006?-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Books
Both books are interesting and substantially accurate in broad outline and most
details. Lewin's book, however, will no doubt have more impact in historical circles,
for it cannot help but bring into bold belief the fact that brilliant intuitive generalship
was often given a tremendous assist by a peek or two at the opponent's hole card.
Many less-than-inspired performances could have been immeasurably improved had
the commanding general in charge but been able to put aside a personal objective or
bias in order to listen and possibly comprehend. Each book in its own style makes a
completely convincing case that, even when available from Churchill and/or
Roosevelt, slightly disguised Ultra had no influence on major Soviet actions which
Stalin took or refused to take. Barbarossa is only one case in point. Interesting reading!
Ap goved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Books
DOUBLE EAGLE: THE AUTOBIOGRAPHY OF A POLISH SPY WHO DEFECTED
TO THE WEST, by Mr. X with Bruce F. Henderson and C. C. Cyr. Bobbs-
Merrill, Indianapolis/New York, 1979.
Double Eagle has the unmistakable flavor of a committee report.
Mr. X is a journeyman intelligence officer who, prior to his exit from the Polish
Intelligence Service, performed a useful service to the West as an agent in place.
Following his break with the Poles, X dutifully pitched a number of colleagues in the
Urzad Bezpieczenstva, or UB, on behalf of his new sponsor.
His book consists of the standard chapters in which X reveals how he came to be
caught up in the UB, the view from the inside, disillusionment, escape and Poland
today. The book abounds in excessive discretion which effectively embalms the
purport and distracts the reader.
It is probable that the real Mr. X has an engaging story to tell; regrettably, it is not
told in Double Eagle.
The period during which Mr. X served the UB (1953-1967) was a time of some
excitement:
? Col. Jozef Swiatlo, former deputy director of Department 10 and confidant of
General Ronkowski, defected to Berlin in December 1953.
? Col. Pawel Monat, Polish Military Attache in Washington, defected in the
summer of 1959 and later, with John Dille, wrote Spy in the US. Both Swiatlo
and Monat testified before the congressional committees of the day and were
given much notice in the U.S. press.
? In the summer of '59, the controversial Col. Michal Goleniewski burst into the
public prints with undeserved credit for having provided the basis for
counterintelligence actions against Houghton, Gee, the Krogers (nee Cohens),
Scarbeck, Lonsdale (Molody), Vassal, George Blake, Col. Wennerstroem and
Israel Beer.
? Wladyslaw Tykocinski, chief of the Polish Military Mission in West Germany,
who held the rank of major general, defected on Sunday, 16 May 1965.
The reader's credulity is overtaxed by X's failure to note these events in terms of
their effect on the Polish service. Aside from his pervasive coyness, X quite justifiably
might have felt diminished in the company of the other, more important, Polish
defectors.
Mr. X concedes that he "was one of the highest ranking intelligence officers of a
communist country ever to flee to the West...... Among the publicized Polish
defectors, Mr. X ranks number five out of five.
Although he makes no mention of Swiatlo or Monat, he does provide a puzzling
paragraph on Goleniewski:
In 1960, the chief of the Scientific Department of the Polish Intelligence
Service, Col. Goleniewski, failed to return to Poland from one of his foreign trips.
MORI/HRP
from pg.
59-60
Approved For Release 2005/01/06 : CIA-RDP80-00630A0001000800(f1
 Approved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
Books
Goleniewski traveled abroad quite often, mainly to meet our agents working for
Scientific Intelligence. We suspected that Goleniewski had become a traitor, that
he had disappeared somewhere in West Germany and was in the hands of the
CIA. Goleniewski would surely have been assassinated by his former colleagues in
the UB if they could have got their hands on him, but they could not....
Mr. X then details "for the first time" the dispatch of Wladyslaw Mroz, chief of
the Polish illegal intelligence network in western Europe. Mroz was beaten, kicked and
shot to death in a city dump in Paris. He had resisted recall to Warsaw and had taken
to augmenting his income by working for a Western service. The ignominious
departure of X's mentor, Mroz, was "a key factor in what was to he my decision to
become a double agent for the West...." My differences with the regime were
irreconcilable.... After several more years (emphasis added-Editor) of soul
searching, I switched sides and went to work for the West...." We are left with the
impression that the bloody execution of Mroz, which seems to have been calculated as
a disincentive to treachery, had an opposite effect on X. Having given October 1960 as
the time of his decision, X tells us later "I am never going to say precisely when my
decision to work for the West was made, in what year. I want the Polish authorities to
he forever confused on this point, to make it more difficult for them to assess the
damage I did to them...."
While on assignment in Oslo, X received a recall cable and, despite the
irnportunings of his CIA case officer, decided to return to Warsaw. Shortly after his
arrival, he was discharged from his cover position in the Ministry of Foreign Affairs
and expelled from the Party.
On Friday, February 17, 1967, X confides, he was born again. His revivification
attended his successful escape by hoarding an SAS flight from Warsaw to Copenhagen.
There is no doubt that X performed useful services and earned his keep. The
passions of his youth have since cooled and his perspectives have changed. His original
motives for acting against the UB as a penetration and later as a defector are obscured
in retrospect. He confesses a sense of abandonment and regrets his now attenuated
contact with intelligence. No longer concerned with vengeance and retribution, on
approaching his fiftieth birthday, he talks about redeeming his son from the barren
prospects of life in Poland as the child of a traitor. The account closes on the
insubstantial hope that its publication will somehow aid in gaining the release of his
son.
Double Eagle is an inconsistent book in which the man rather than the act is
revealed. While it is not a source of insight into the UB, it does convey the important
rationalizations of a man who acted against his country. X's apologia should hold
considerable interest for the behaviorists.
Aproved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
 pproved For Release 2005/01/06 : CIA-RDP80-0063OA0001 00080001 -1
Boo S
AMERICAN ESPIONAGE: FROM SECRET SERVICE TO CIA, by Rhodri Jeffreys-
Jones. Free Press, New York, 1977, 276 pp.
THE ARMIES OF IGNORANCE: THE RISE OF THE AMERICAN INTELLI-
GENCE EMPIRE, by William R. Corson. Dial Press, New York, 1977, 640 pp.
Our first author does not know what he is writing about, and he has therefore
written a confused and confusing book. By contrast, our second man, both informed
and discerning, has written an intelligible and readable, albeit wordy and overly
ambitious work. Thus dissimilar in quality, they are similar in their approach, in what
is a new departure in the literature on CIA and American intelligence. They both seek
to develop a much lengthier pre-World War II historical dimension to the story of CIA
than has hitherto been offered, and for this reason they merit joint consideration here.
Rhodri Jeffreys-Jones-could we for convenience's sake occasionally call him
J-J?-holds a doctorate in history from Cambridge University and is currently
teaching at the University of Edinburgh. Using the skills of his art, Jeffreys-Jones has
dug deeply in several archives and brought forth much interesting material. The bulk
of this is drawn from the World War I period and centers on American
counterintelligence, with a strong emphasis on some British ramifications.
In writing up this material, however, J-J has chosen to tie it together with the
thread of the development of central intelligence in, for him, the formative years from
1898 to 1947. In doing so he has clearly included much material-Britain's Somerset
Maugham in revolutionary Russia and the literary-psychological last chapter on "The
Spy in the Mind"-which simply do not belong there. He has likewise omitted
much-most notably and unpardonably even a mention of the establishment in 1941
of the Coordinator of Information (COI), which has the best of all claims to being this
country's first central intelligence organization. Additionally, and perhaps because of
his fundamental shift from espionage to central intelligence, he has clearly
disregarded-to the reader's disadvantage-Winston Churchill's observation on the
writing of history, namely, that strict chronology is the secret of good narrative. In this
book one really never gets out of World War I and the 1920s.
More fundamentally defective, however, is J-J's identification of espionage with
intelligence: he clearly considers everybody in the intelligence business to be a spy.
Thus, those Research and Analysis personnel, when they moved from the defunct
Office of Strategic Services (OSS) in 1945 into the Department of State, became "the
department's spies" (p. 194) and, reorganized as State's Interim Research and
Intelligence Service, they "advised the State Department on intelligence matters,
formulated clandestine policy, and coordinated the work of other undercover
agencies ...." (p. 189) Rubbish! But what else can be expected of an academician who
can describe a "Human Espionage Activities" file as "chillingly labeled"? (p. 58) And
who can be so naive as to describe Assistant Secretary of State Berle-FDR's informal
link with his intelligence agencies-as going "through the elaborate procedure of
gaining entry to the office of the security-conscious J. Edgar Hoover."?
With these faults weighing him down, Jeffreys-Jones has tried to tell the story of
central intelligence in terms of a see-saw struggle between the military and the
civilians. Such a thesis is not indefensible, but when spelled out in J-J's confused, naive
MORI/HRP
from pg.
61-64
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Books
way, it beggars belief. It is, for instance, reading history backwards and blowing it up
beyond recognition to see as the first American central, civilian, peacetime
intelligence service the Secret Service of 1898 as run by Treasury and supported by
State, which the author asks us to view as "the intelligence giant of the future." (p. 37)
It is stunning to be told that State's "future," which began in World War I when,
under counsellors Lansing and Polk, "the sinister hand of the spy (was) regulated by
the dexterous hand of the diplomat," (p. 45) was unfortunately ended in 1927 when
U - I, the department's "pivotal intelligence unit," was abolished by Secretary Kellogg
because its bureaucrats had been overwhelmed by the flood of paper brought on by
the typewriter and carbon paper! (pp. 158-159) Finally, events are just messed up
incomprehensibly when we are told that the civilians lost out to the military in 1942,
when OSS was established, and again in 1947 with the coming of CIA, and that they
subsequently re-established some element of balance only because "State Department
lawyers" had cleverly and unobtrusively reworded the military's draft of the National
Security Act of 1947.
More useful is William R. Corson, who now lives in Potomac, Maryland. A retired
Marine, Corson says he was "one of the many who briefed (President) Nixon on
specific (intelligence) operations" (p. 37), and by other scattered autobiographical
notes he indicates he has had considerable exposure to the Washington end of the
intelligence community, some members of which, the reviewer has been advised,
regarded him as superficial and subjective. Corson has interviewed some former
officials, extensively used the OSS War Report and Congressional reports, used
unpublished secondary materials, such as the Bruce Bidwell G-2 history, and has
looked into some primary sources-though not as much as his "Source Notes" suggest.
He writes with self-assuredness of the complexities and subtleties of intelligence, the
bureaucracy, and national politics, though he is often glib and subjective.
For instance, there are those "intelligence community leaders." They show up in
the first sentence on the first page where they are seen in 1974 viewing "with a sense
of foreboding" Nixon's fall from power, and thereafter, and beginning back in World
War I, they show up regularly-faceless, generally nameless, changing, but always as a
tight, powerful little bureaucratic clique, almost "an invisible government," whose
thoughts, attitudes, ambitions, battles, successes and failures, joys and disappointments
Corson, sometimes knowledgeably, too often impressionistically, divines, records, and
evaluates. It is always interesting, often convincing, but one is often provoked to
explode with a "How in hell can you say this?!"
Just as troublesome, though not as frequently obtrusive, is the organization of his
material. For him there is a fundamental distinction between, roughly, 140 years of
uninstitutionalized intelligence activity-the years of what he calls "patriot spies" like
Nathan Hale and of short-lived wartime efforts and organizations-and 60 years of
institutionalized intelligence. For him the distinction gives rise to a triadic
periodization consisting of a prehistory (an indefensible use of the word), an historical
past, and a modern period. For him the prehistory is the first 140 years, the historical
past is World War I to 1945, and the subsequent years are the modern period. But he
has not written it that way; he has devoted 483 pages to the last two periods and then
followed them with 114 pages, an "Afterword" (surely one of the longest in publishing
history) on the prehistorical period.
Whatever the justification for this organizational oddity, it does injustice to much
of Corson's own material and to the history of the permanent intelligence agencies
about which he writes. For instance, for one who is writing about "the rise of the
American intelligence empire," he unjustifiably-in the reviewer's opinion-relegates
ANroved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 BoAoproved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
to prehistory the establishment and development of many components of today's
intelligence community. Thus, the establishment of naval intelligence in 1882 and of
military intelligence in 1885 and their subsequent growth down to World War I are
prehistorical. So also, the development of the early Secret Service, so active then in
counterintelligence, and of the role of the Bureau of Investigation (the early FBI) in
counterintelligence are left to prehistory. Corson's own material would seem to dictate
more forthright treatment of the history of permanently institutionalized depart-
mental intelligence as the origin of the problem of central intelligence.
Beginning then with World War I, and taking the investigative agencies as they
then were, Corson sees the story not as does Jeffreys-Jones, not as a civilian-military
struggle, but as a bureaucratic struggle among those agencies, especially Treasury's
Secret Service and Justice's Bureau of Investigation, for jurisdictional turf and
supremacy, particularly in the counterintelligence field in the United States. The story
of this struggle, especially Treasury Secretary McAdoo's abortive effort to set up a
central intelligence bureau for counterintelligence, is most interesting and worthy of
more elaboration. The counterintelligence struggle is picked up again by Corson on
the eve of World War II, but he sees little happening in the positive intelligence field
until the appointment of William J. Donovan as Coordinator of Information (COI) in
July 1941.
Corson's account of the events leading up to that appointment contains some
interesting new material, and, though occasionally unbalanced, is one of the best yet to
appear. His new material shows those "intelligence community leaders" apparently
playing a greater role in Donovan's famous 1940 trip to London-from whence sprang
his own rise as a wartime intelligence chief-than had been previously known.
Corson also strikes out anew in paying attention to a hitherto completely ignored
aspect of COI's history, namely, its transformation into the Office of Strategic
Services. Most writers, if not ignorant of COI, have dismissed the change largely as a
change of name. Considerably more was involved than that, and Corson rightly
elaborates, though not always correctly, on the process whereby COI, reorganized as
the OSS, became a supporting arm of the Joint Chiefs of Staff. Where Corson errs is in
attributing the process to the establishment of the JCS and Donovan's felt necessity to
join it. Rather, it was the change from peace to war which made Donovan realize that
his civilian agency, if it were to survive in wartime, needed a military affiliation; and
what Corson misses is Donovan's quick perception of the JCS as the elevated military
body which could ensure him against imminent conquest by his foes, especially G-2.
Disappointing, and somewhat surprising, is Corson's too brief account of CIA's
establishment in 1947. Whereas he rightly spent pages on the origin of both COI and
OSS, and whereas he gives reasonably adequate coverage to the establishment in 1946
of the National Intelligence Authority and the Central Intelligence Group, he
dismisses the drafting, debating, and passage of the National Security Act of 1947 and
the establishment of CIA in less than two pages! The Congressional Record is full of
material that could have been used, and one would think that the last few years of
discussion of the agency's charter and record would have necessitated some extensive
account of the 1947 act.
In covering what he called the modern period, Corson discusses not the
bureaucratic struggle but the relationship of each of the presidents, from Truman to
Carter, to the intelligence community and its activities and problems. His chapter on
"Changing the Intelligence Guard" interestingly highlights, as he sees it, the
community's problem of periodically indoctrinating a new president in the
Approved For Release 2005/01/06 : CIA-RDP80-00630A0001000800O 31
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Books
intelligence business. He sees Truman as having been deliberately kept in the dark,
Eisenhower keeping the community at arm's length, Kennedy being conned, Johnson
keeping the community off balance, and Nixon early taking charge in businesslike
fashion.
Subsequently Corson writes many pages on each of these presidents, but it is just
here that he is overly ambitious, albeit interesting. What he does is not write a
balanced chronological account of each presidency-even if the documents for doing
so were available to him-but discuss each presidency on the basis of a few significant,
publicized events therein, e.g., Truman and the Italian elections, Kennedy and the
Bay of Pigs of course, and Nixon and the Huston plan and the Schlesinger report on
the reorganization of the community. Readers with more knowledge than the reviewer
of these years and events will find it worthwhile to take their own measure of Corson.
His concluding chapter on the future of American intelligence, though written
before the issuance of E.O.12036 and the introduction of S.2525 in the Senate-both
deaing with the structure and powers of the community-will remain useful at least
until new legislation on the agency or community is passed. He thinks there should be
more objective evaluation of the community's tasks and operations, that the
community should be more genuinely centralized, and a strengthened DCI should be
relieved of control of CIA. It is not clear, however, that he thinks a new head of "the
American intelligence empire" should actually be an "emperor."
Speaking of the book's subtitle makes one return to the title itself. Since it is used
in the text four times (pp. 456, 479, 481, and 482) it must be Corson's and not a
publisher's imposition. What of any value it imports or signifies, however, eludes this
reader. It seemingly depreciates the "foot soldiers and company-grade officers" for
whose success Corson clearly has the best wishes.
Thomas F. Troy
THE CIA AT WORK, by Lauran Paine. Robert Hale, London, 1977, 192 pp.
From CIA's point of view, Lauran Paine means well, but his book is
disappointing.
Paine, whose books on Vietnam, Captain John Smith, Bolivar, Saladin, and on
espionage, assassins, and terrorists are unknown to this reviewer, says in this work that
CIA, for all its ineptness and bloatedness, its blunders and crimes, is-in this anarchic,
violent world, in this "the most dangerous time in history," (p.182) and in the face of
the extensive, unremitting machinations of the Soviet KGB-"the West's best first line
of defense" (p.185). As such it merits defense against the vilifiers and fools who would
"dismantle" or "neuter" it (p.185). Such is Paine's message which he iteratively and
vigorously propagates in this slim volume.
The work, however, is an undocumented, opinionated rehash of yesterday's and
today's headlines about CIA's involvement in such places as Chile, the Bay of Pigs,
Iran, Iraq, and Southeast Asia. It is distressingly marred by errors, misconceptions,
exaggerations, and fantasies. (The Office of Strategic Services was not established in
October 1941, and it was not modeled after the FBI. Who believes that 38 percent of
MORI/HRP
from pg.
64-65
A00roved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
Books
the Agency's population are bilingual, and what "knowledgeable people" think the
CIA budget is "closer to seventy billion dollars" than one billion dollars (p.27)? And
tell. us more about CIA's control of the National Security Council since 1969-"a
condition which still exists (p.121)." The writing in this book, while journalistically
short and punchy and generally readable, is often strangely capitalized and
syntactically confusing, and always unnervingly self-confident and assertive.
For the undiscerning reader who needs reassurance that the world is, in Richard
Helms' simple phrase, "an untidy place" in which intelligence has an important role to
play, this book may be a morale booster. Otherwise it is not of much value.
ON THE BORDER OF WAR AND PEACE: POLISH INTELLIGENCE AND
DIPLOMACY IN 1937-1939 AND THE ORIGINS OF THE ULTRA SECRET,
by Richard A. Woytak. Columbia University Press, New York, 1979, 141 pages.
This book is a valuable addition to the literature on the "Ultra Secret." Woytak
has meticulously detailed the work of the Polish Intelligence Service in the pre-war
years which led directly to the triumphs of Allied code-breakers. One could wish only
that the author had chosen to write a longer and more popular treatment. Copious
footnotes and an extensive bibliography are included.
ULTRA AND THE BATTLE OF THE ATLANTIC, by Patrick Beesly, et al.
Reprinted from Cryptologic Spectrum.
Includes the texts of three papers presented at the Naval Symposium at the U.S.
Naval Academy on 28 October 1977. The speakers were Patrick Beesly, Jurgen
Rohwer, and Kenneth Knowles taking the British, German, and American views,
respectively. Also included is an article by Harold C. Deutsch-"The Historical
Impact of Revealing the Ultra Secret."
EXPLODING STAR: A YOUNG AUSTRIAN AGAINST HITLER, by Fritz Molden.
Morrow, New York, 1979, 280 pages.
Well-told story by a man with a story to tell. Molden's account of his own
experiences in the Austrian resistance movement is both well-paced and beautifully
written. Of special interest is Molden's account of his work with OSS in the last two
years of the war.
MORI/HRP
from pg.
65-68
Approved For Release 2005/01/06 : CIA-RDP80-00630A00010008000451
 Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
OPERATION SUNRISE: THE SECRET SURRENDER, by Bradley F. Smith and
Flan Agarossi. Basic Books, New York, 1.979, 234 pages.
Operation Sunrise was the OSS and Allen Dulles' operation to arrange the in-
dependent surrender of Axis forces in Northern Italy, which was effected on
29 April 1945. This hook adds little, substantively, to Allen Dulles' own account in his
The Secret Surrender. But it does place the operation in a historical perspective which
will not endear the hook to fans of Allen Dulles.
SONJAS RAPPORT, by Ruth Werner. Verlag Neues Lehen, Berlin, 1977, 343 pages.
The personal memoir (in German) of the author's twenty years as a Soviet
intelligence agent (GRU) in China, Switzerland, Poland and England. She is best
recalled in a number of intelligence cases under her names of the time: Ruth
Kuczynski, Ursula Hamburger, and Ursula Burton.
OPERATION FISH, THE RACE TO SAVE EUROPE'S WEALTH, by Alfred
1)raper. Cassell, London, 1979, 377 pages.
The title refers to the secret British transfer of gold and negotiable securities to
North America late in 1939 and 1940. The author does not limit himself to this
operation but includes the efforts to secure the national wealth of Norway, Holland
and France. The story is told in a chatty, anecdotal manner using much newly
declassified material. A surprisingly interesting exposition of a footnote to the history
of World War II.
SII)ESHOW: KISSINGER, NIXON, AND THE DESTRUCTION OF CAMBODIA,
by William Shawcross. Simon and Shuster, New York, 1979, 467 pages.
This is an exhaustive and well-written account of the recent history of Cambodia-
Kampuchea. Mr. Shawcross obtained much material through the Freedom of
Information Act and more than 300 personal interviews. If there is any fault to be
found with Sideshow it is, perhaps, that it is a book to be studied rather than read.
'I'IIE REFORM OF FBI INTELLIGENCE OPERATIONS, by John 7'. Elliff.
Princeton University Press, Princeton, N.J., 1979, 248 pages.
Dr. Elliff, a staff member of the Senate Select Committee on Intelligence and
consultant to the Police Foundation, discusses the various problems involved in
drawing up guidelines for FBI intelligence operations. Well-researched and well-
written, it nevertheless leaves one with the impression that Dr. Elliff doesn't
understand the field of intelligence. The hook raises many important questions, but
unfortunately, answers few of them.
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 eon proved For Release 2005/01/06 : CIA-RDP80-00630A000100080001-1
THE PIIILADELPHIA EXPERIMENT, PROJECT INVISIBILITY: AN ACCOUNT
OF A SEARCH FOR A SECRET NAVY WARTIME PROJECT THAT MAY
HAVE SUCCEEDED-TOO WELL, by William L. Moore in consultation with
Charles Berlitz. Grosset & Dunlap, New York, 1979, 177 pages.
This book would seem to be an attempt to plumb the depths of the credulity of
the American public. Trading on the name of Charles Berlitz (of Bermuda Triangle
fame), the author has assembled a confused mass of half-truth and unsubstantiated
rumor on everything from UFO's to "invisibility." The Philadelphia Experiment
allegedly was an attempt by the Navy during World War II to render a ship and crew
invisible and to teleport it to and from Norfolk in a few minutes by using Einstein's
"unified field" theory. This is the sort of book which will probably be serialized by the
National Enquirer.
ADMINISTRATIVE SECRECY IN DEVELOPED COUNTRIES, by Donald C.
Rowat. Macmillan, London, 1979, 364 pages.
Twelve country reports dealing with the delicate balance of the government's
need for secrecy in certain areas and the public's need for access to information.
Freedom of Information laws in Denmark, Finland, Norway, Sweden and the United
States are evaluated. Countering reports deal with Belgium, Canada, France,
Hungary, the United Kingdom, West Germany and Yugoslavia.
QUANTITATIVE APPROACHES TO ]POLITICAL INTELLIGENCE: THE CIA
EXPERIENCE, edited by Richards J. Heuer, Jr. Westview Press, Boulder, Colo.,
1978, 181 pages.
A collection of articles on the application of various quantitative methods to
specific analytical problems in CIA. These include articles on regression analysis,
multidimensional scaling, and Bayesian statistics. Definitely not for the non-
professional.
"The Intelligence of Stupidity: Understanding Failures in Strategic Warning," by
Steve Chan. In American Political Science Review, Vol. 73 No. 1, March 1979.
Good article, somewhat broader than the title implies.
"Computer Security: The Achilles Heel of the Electronic Air Force?" by Roger R.
Schell. In Air University Review, Jan-Feb 1979.
An examination of Air Force "tiger team" penetration of a computer system, the
high incidence of vulnerability in program "fixes," and a plea for multi-level
computer controls based on the security kernel concept.
Approved For Release 2005/01/06 : CIA-RDP80-00630A0001000800-1
 Appro xd~For Release 2005/01/06 :CIA-RDP80-00630A000100080001-1
Dr. Albert Einstein, whose centenary birthday is being celebrated this year by
statesmen, physicists, mathematicians and ordinary citizens, was possibly the best-
known Nobel prizewinner of all time. As such he was the object of countless requests
for advice and assistance on all sorts of problems; he counseled the President on the
employment of the atomic bomb, and he answered schoolchildren's requests for help
with their math homework. On the evidence of the letter reproduced below, it
appears that he was consulted on intelligence and security problems as well. The
bright and ambitious Navy yeoman who sought the professor's advice apparently
theorized that the security of encrypted messages could be enhanced if the encrypted
version contained fewer characters than the original. The idea has obvious
implications for the secure and rapid transmission of intelligence, but as Einstein
pointed out, it can't be done. The original letter resides in the Walter Pforzheimer
Collection on Intelligence Services, Washington, D.C.
THE INSTITUTE FOR ADVANCED STUDY
SCHOOL OP MATHEMATICS
PRINCETON, NEW JERSEY
Ho':ard L.' cVitty,Cox?'ain "S'TR
Public I:elations Office
U.:S.'Iaval Amphibious Traininn Pase
Tort Pierce,Florila.
Accordind to my opinion it is euite impossible
to express a serie of d.ipAts through a lesser serie of
digets so that the former can be reconstructed through a
code. This holds,of course,only for the case that the given
serie do not have regularities -,hich are Icno?'n inde^endently.
the serie could,of course, be expressed by a smaller number of
,igns if instead of dip-A'ts a larger number of elementary signs
are used. F.i. if you have hundred di??e-ent elementary signs
you can get numbers in a centesimal system instead of in a
decimal system. But I do not think that such augmentation of
the number of elementary signs would be practical for any
purpo se.
Sincerel,, Yours,
Professor Albert Einstein.
Approved For Release 2005/01/06 : CIA-RDP80-0063OA000100080001-1
 Approved For Release 2005/01/06(t kl .rDP80-00630A000100080001-1
Approved For Release 2005/01~g ~'CP1RDP80-00630A000100080001-1