VII. THE INFORMATION EXPLOSION

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP85G00105R000100130005-5
• Release Decision: RIPPUB
• Original Classification: T
• Document Page Count: 21
• Document Creation Date: December 12, 2016
• Document Release Date: July 30, 1998
• Sequence Number: 5
• Content Type: REPORT

V1111111111,0ue, ~~~ /. l ~. ~, l ~. l ~~ /~VIF~15e
CONTROL NO.
REFERRED TO
RECEIVED
RELEASED
SEEN BY
OFFICE
SIGNATURE
DATE
TIME
DATE
TIME
NAME & OFFICE SYMBOL
DATE
Handle Via Indicated Controls
TALENT-KEYHOLE
WARNING
This document contains information affecting the national security of the United States within the meaning
of the espionage laws U. S. Code Title 18, Sections 793 and 794. The law prohibits its transmission or
the revelation of its contents in any manner to an unauthorized person, as well as its use in any manner
prejudicial to the safety or interest of the United States or for the benefit of any foreign government to the
detriment of the United States. It is to be seen only by personnel especially indoctrinated and authorized
to receive information in the designated control channels. Its security must be maintained in accordance
with regulations pertaining toTALENT-KEYHOLE Control System.
ill~iil~/1~~1111~Z~l?l
 Approved For Release 0 DP85G00105R000100130005-5
HANDLE
VII. THE INFORMATION EXPLOSION
1. The Long Range Plan of 1965 is based on the unquestioned
assumption that CIA "must be allowed to grow to meet ever increasing
demands from the Government for intelligence on a world that becomes
constantly more complex." This proposition is repeated in a dozen ways
which all add up to endorsement of the belief that the more facts we
collect the better we do our job of safeguarding American security. It
could be distilled into an Orwellian slogan: MORE IS BETTER.
2. We have first to recognize that a bias in favor of amassing
indefinitely expanding quantities of information characterizes American
society far more than any other that ever existed. We are hypnotized by
statistics of every conceivable degree of trustworthiness, relevance, and
importance. Nowhere outside America is there so great a tyranny of
information over all the other factors affecting judgment--the tyranny
of the public opinion poll, market research, television ratings. Never
mind aberrations like the election of 1948 and the Edsel automobile;
these merely prove that what we needed was more facts.
3. Given its brief, intense history, it was inevitable that the intelli-
gence community would embody this American bias in its most extreme
form. In particular, two important factors have intensified our search
for more and more facts. One is the course of events during the decade
after World War II, in which our ignorance of many fundamental foreign
problems encouraged a kind of national paranoia that persists into the
present. The other is the way in which the two dominant segments of the
intelligence community--the military and the academic- -reinforce each
other's lifelong practice of treating the amassing of information as more
important than theory, speculation, or hard-thought analysis. Thus MORE
IS BETTER has been an article of faith with us from the beginning.
as well as analysis; the people who produce finished intelligence
TCS 18112-66
 Approved For Release 2002/06/18 : CIA-RDP85G00105R000100130005-5
TOP SECRET
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
make up less than 2%. The disparity is much less pronounced within CIA,
but even so its Directorate of Intelligence allocates fewer than 30% of its
professional and clerical slots to those functions which can be described
as production of finished intelligence; exactly the same percentage goes to
,_25X1A1a
and the remainder to collection, processing, and executive direction.
5. In our view the assumption that MORE IS BETTER is now a
dangerous anachronism, and ought to receive the community's most
rigorous and skeptical analysis. The reason is the Information Explosion.
The Long Range Plan of 1965 correctly says: "It is abundantly clear at
this time that our ability to process and analyze raw information has not
kept pace with our collection capability.'' The Plan alludes to needs for
increases in CIA personnel which add up too slots over the next five
25X1A1ayears, and include alone. It points out that these figures
require further study, but the implication that an increase of this magnitude
would enable CIA to use properly what is now being collected impresses us
as unduly optimistic.
6. At the same time the community goes on making plans for ever
larger explosions. A current study of one single aspect of the intelligence
function--early warning--holds out the hope, or threat, of the following
expansions of our ability to acquire raw data:
These figures are drawn from the calculations of and
describe the size and shape of the community in 1965. They can be only
rough approximations because of widely varying terminology and other
obstacles to statistical exactitude.
TOP SECRET
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
Approved For Release 2002/06/18 : CIA-RDP85GO0105R000100130005-5
 Approved For ReleaTOP 2SEG01-
HANDLE RDP85GO0105R000100130005-5
VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
7. It is clear that these and other developments can produce better
information on some specific problems than we now have, and the need for
this is not questioned. The bone of contention is quantity, and the problem
whether we can control it in such a way as to recognize and isolate and
correctly interpret (particularly with respect to enemy intentions) the
crucial items these systems would provide. The community's record with
respect to the exploitation of current collection inspires anxiety that we
are creating worse problems than we are solving; not merely of finding
the manpower to read out the new information, in itself a great problem
no matter how much we automate, but of developing the judgment and
discrimination to sift out the crucial facts and assess them correctly.
Automation and other technical improvements will bring the handling of
data under progressively better control (as to quantity) at the working
levels of the intelligence process, but the total product of these and other
quantum-jump systems can only increase the work load of policy-makers
from the President down who are already overworked. We must move
quickly to bring that product under control.
8. After several pages of describing future prospects for quantum
jumps in collection, the study just cited has only this to say about their
effects on analysis, estimative judgments, and determination of policy on
the basis of the information collected:
"One clear conclusion is inescapable: the investments
in sensor and vehicle development will require comparable
investments in means of exploitation and analysis. The
technical revolution in information collection, epitomized
Citations and quotations are from the third draft of the study on early
warning.
~1@@~OR QIAII3QIIA&5t
 Approved For Release 2002/06/18 : CIA-RDP85G00105R000100130005-5
TOP SECRET
HANDLE VIA.TALENT KEYHOLE CONTROL SYSTEM ONLY
by the reconnaissance field, is only just beginning
to be felt. Its full impact promises to be enormous
and may be expensive, but the great volumes of
information we are learning to collect must be dealt
with or wasted. Nowhere will this problem be sharper
Here is the Sorcerer's Apprentice commanding more and more broom-
sticks to haul more and more water--in real time.
10. The passage quoted above says that the technical revolution in
information collection is only just beginning to be felt. It is our conviction,
and the source of our anxiety, that the Information Explosion has been with
us for years and that our patchwork efforts to cope with it have concealed
the true disparity between collection and end use. The community already
holds warehouses full of unexploited SIGINT tapes and miles of photographic
film only superficially examined, though both are known to contain useful
TOP SECRET
 Approved For Release 2002/06/18 : CIA-RDP85G00105R000100130005-5
TOP
HANDLE VIA TALENT K Y1 OLE-CONTROL SYSTEM ONLY
information on our two top targets. But the explosion is by no means
confined to overhead reconnaissance and accumulation of signals. The
explosion in written communications was under way long before we were
born; it is tempting to date bureaucracy's affinity for it from the Franco-
Prussian War, when during the siege of Paris a Frenchman escaped to
Tours by balloon and
". . . set up the first microphotography unit ever to be
employed in war. Government dispatches in Tours
were reduced to a minute size, printed on feathery
collodion membranes, then rolled into a pellicle; so
that one pigeon could carry up to 40, 000 dispatches....
On reaching Paris, the dispatches were projected by
magic lantern and transcribed by a battery of clerks.
Sometimes one pigeon-load alone would require a whole
week to decipher and distribute. "*
11. Even during World War II Winston Churchill complained
repeatedly, and in vain, that the Information Explosion--i. e. , the quantity
of cable traffic--had become an intolerable burden and sometimes even a
positive hindrance to prosecution of the war. ** Later we ourselves used
to try to keep down the volume to save money, but Churchill's objection
* Alistair Horne, The Fall of Paris, New York, 1965, P. 128.
*x* For example, a memorandum to the Foreign Secretary, "I eel t a
this is an evil which ought to be checked. inis ers and Am assa ors
CPYRGHT abroad seem to think that the bigger the volume of their reports home,
the better is their task discharged. I try to rea all these telegrams,
and I think the volume grows from day to day.'' (The Second World
War, III, 723.) To General Ismay and others: "I see no need for these
long and pointless telegrams, and it is 'becoming quite impossible to
conduct military operations when everything has to be spread about the
Departments and around the world like this." (Ibid. , p. 724.) And
again to the Foreign Secretary: "The telegrams seem to be growing
longer and longer....I quite understand they all want to help the war
by increasing their output. In fact, they clog and hamper." (Ibid.,
I IV, 864. )
HANDLE VIA TALENT K YOLE CINTROL SYSTEM ONLY
Approved For Release 2002/06/18 : CIA-RDP85G00105R000100130005-5
 Approved For Relea(2RDP85G00105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
was more to the point: The greatest cost of our cable traffic is the time it
takes thousands of people to keep abreast of it. Yet in response to perpetual
demand for more and quicker communication, impressive technical advances
have brought this increase in the cable traffic processed by CIA Head-
quarters Signal Center over the years:
1950
1955
1960
1965
12. In this context the phrase "Information Explosion" has little
to do with exploitable raw intelligence data, because the number of
published CS reports has changed remarkably little over the years and
,__25X1A1a in 1965 amounted to some But it has everything to do with the
way intelligence officers at all levels spend their time, hours which cannot
be expanded with the expansion of reading matter. Some of this increase
in Commo capabilities must have improved the quality and speed of our
reporting and that of other agencies, and in particular improved the respon-
siveness of our operations; much of the time spent working on cables 25XlAla
at both ends must formerly have been spent on pouched dispatches. But
we doubt that these changes explain the I expansion of the system.
25X1A1a 25X1A1a
13. The greatest jump in our cable traffic, almost 0 from
1960 to 1965, illustrates how a system, developed to meet some important
need, acquires a life of its own whether the urgency persists or not. For
):c "Work expands so as to fill the time available for its completion."
Parkinson illustrated his discovery by some statistics drawn from the
British Navy and Colonial Office. Between 1914 and 1928 the Admiralty
(continued next page)
TOP SECRET
HAN, Tp, ov16 cGTr44 kt-'1JJ
1 ~$~ C $ 0( 1`b~l i M io '05-5
 Approved For Rele q0 WET-RDP85GO0105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
14. And now our Cable Secretariat, the National Military Command
Center, and others are looking for ways to cope with what the early-warning
study calls
". . . the sizeable increases expected in both message
volumes and numbers of electrical transmissions
over the next few years--perhaps as much as 8 to
10% a year. They are also experimenting with remote
print-out of the messages at the analysts' locations
and with electronic storage for extended periods and
retrieval by cathode ray tube display as well as hard
copy. As communications centers expand in capacity,
message handling down the line to the analyst will
come under heavy pressure to go automatic too."
But the capacity of each recipient of all those cables will remain exactly
what it was before: what he can read and absorb and act on in a day.
Unless, of course, he has to spend more time looking at pictures. The
total capacity of all recipients together has theoretically increased by
some of the few percentage points which reflect CIA personnel increases
(outside NPIC) since 1950, but any effort to prove this would have to take
into account the facts that cables go to many more readers than the
pouches used to, that we must read increasing numbers of State and
Defense cables in addition to our own, and that cables must compete
increasingly with the products of other compilations of material.
*(continued from Page VII - 6)
bureaucracy grew 78. 45% while the number of capital ships in commission
declined by 67. 74% and the number of officers and men declined by 31. 5%.
Between 1935 and 1954, while Great Britain was losing most of the colonies
it had accumulated over the past two centuries, the Colonial Office grew
from 372 bodies to 1661. It should be emphasized that the parallel
suggested here reflects not upon Commo but upon the work habits of its
users. Something of the same point could be made with respect to the way
we sought out things for the U-2 to do after May 1960, when it was no
longer usable to meet the urgent need for which it had been created.
O uEC RET
HAI J vYc~` o~ ~ i ~~11~6R F-cFA OdS W & 0V 3b&5-5
 Approved For Releaaq929/Rc RDP85G00105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
15. The effect of overhead reconnaissance on the collective work
load is impossible to measure, so that more attention is naturally paid
to backlogs and shortages of manpower in NPIC than to the ability of the
substantive producers to use the NPIC product adequately. We can only
concur in a recent general description of the problem which points out
that the explosion affects the processors, producers, and consumers of
photographic intelligence and cites these reasons:
"... the increased use of our ever-improving
reconnaissance capabilities, and increased
dependency on photography as other sources of
intelligence become relatively less productive,
and a general widening of interests into areas
previously neglected or ignored by intelligence.
(Emphasis added. )
In consequence,
"Volumes currently processed are many times
the 1960 volumes. For example, during fiscal
year 1965 an excess of two million feet of film
was exposed over Cuba and Vietnam alone."
17. We are now receiving open literature at the rate of 1, 600, 000'
items a year (114, 500 copies of books, 270, 600 of journals, and the
remainder of newspapers). "This a threefold increase over 1950. By
1970 the figure will be two million." In the fields of science and
technology and "sociology, " the figure rose from 16, 000 individual
titles in 1950 to 46, 000 in 1966. According to the Foreign Documents
Quotations and most statistics in paragraphs 15-24 are from Paul
Borel's Controlling Intelligence Information, written for presentation
to the in September 1966.
The opinions based on them here are far more pessimistic than Borel's.
TOP SECRET
HAWprovyc+0Ti?#&sW221~>CI~A~O~ ~50~OX05-5
 Approved For Releal
nn2g J& RDP85G00105R000100130005-5
edly
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
25X1Ala
Or in other words, we are looking for ways to speed up the flood we
must be shielded from. Meanwhile the USIB Committee on Documentation
is working on a scheme to acquire a quantity of open literature from the
Soviet Union, in return for three sets of U. S. patents. This would include
not only a number of serial publications and monographs, but copies of
all Soviet candidate doctoral dissertations--in triplicate.
19. Even so, CIA has been working on schemes to increase the
quantity of such material many times. One project, which still has bugs
TOP SECRET
HAN- oyJeAa a ITe YM 8E ?&RDPMG09J1W 0QIQQJW05-5
 Approved For ReleafIW pnk.el]--RDP85G00105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
in it, will some day make it possible for the processing of overt materials
from publications and broadcasts to be limited only by the speed with which
translators can dictate their translations onto tapes, because other stages
ending with computer printout will be automated. If automatic machine
translation ever becomes practicable, our ability to process overt foreign
information--i. e. , put it on paper in English--will jump from millions of
words per day to hundreds of millions. If we apply analogous techniques
to "exploiting" all foreign television, a medium we shall have to take
steadily more into account, we shall create the opportunity for another
quantum jump. By that time, for any evidence we can see to the contrary,
new projects for enormous expansion of technical collection of information
of all kinds will have required more and more warehouses for the storage
of more and more miles of tapes and films.
20. Meanwhile, Borel says "our total receipts of raw and finished
intelligence documents have shown little variation over the years. The
high water mark was reached in 1963 with a receipt figure of 409, 400
individual documents." Two factors make this high figure weigh especially
heavily in the collective work load. One is the special way these documents
compete with each other and with all unclassified material for attention.
Analysts must not only read as much raw information as possible but must
also read one another's finished products. It can be argued indeed that
they are one another's best customers, rather than the policy-makers who
cannot conceivably keep up with the flow. And the other factor, obviously
related to the first, is the way the number of copies of our classified
documents keeps growing. Cables are often published in from 35 to 50
copies, sometimes in more than 100, and many of these copies are circu-
lated to dozens or scores of readers. In addition, the prevalence of office
copying -machines "has encouraged secondary reproduction by recipient
offices to absurd proportions" --not just of cables, and no matter what
security restraints we put on this practice.
21. The most vivid illustration of all this proliferation is afforded
by the Office of Current Intelligence. With almost the same number of
authorized personnel as in 1955, OCI in 1965 succeeded in handling more
than twice as many incoming items, and in producing 2-1/2 times as many
of its own. But the total effect on all its customers' IN-boxes together
was much greater: Measured in "impressions" (number of pages x number
of copies per page), OCI annual production jumped in those ten years more
than 400%, from 7, 000, 000 to 36, 000, 000. Much of this flow is accounted
for by publishing the OCI Weekly Summary, which averages 28 pages in a
press run of 1570 copies of the Secret version and of 545 copies of the
code-word version called Weekly Review. OCI Special Reports go to 1646
TOP SECRET
HANAoy~_e 4014`#"1?1a405Y95W1@SM05-5
 Approved For Reid ee 0$p,CIA-RDP85G00105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
of the recipients of these two weeklies; thus, on 28 October 1966, a
productivity of 16, 460 impressions was achieved by an essay, ten pages
long counting the cover and a full-page map, which was entitled "Trends
in the Lesser Antilles. "*
22. Whatever the eventual benefits of automation, it is bound to
increase the need for discipline. So far, in some ways, it has made the
problem worse:." We see instances where basic input data is replicated,
reformatted, resorted, to the point where the output volume exceeds input
by perhaps 100 times." This is an excess which experience will no doubt
correct. Yet in all that we have heard of the prospects for automating
the intelligence community, there is still overwhelming emphasis on
quantity and speed and far too little on the more important factors of
quality and relevance.**
* After pages of travelogue ("St. Lucia... in 1962 opened its first luxury
beach hotel"), this report delivers the message that the mayor of a town in
Martinique was a Communist until the late 1940s, the mayor of a town in
Guadeloupe still is, and in certain unlikely circumstances the Communists
might reverse the steady decline they have been undergoing for nearly 20
years. We are informed that this and two other recent Special Reports were
published because the DCI had told OCI to keep an eye on the West Indies.
But we would argue that there is an important difference between keeping
track of a subject and publishing a pointless essay in many hundreds of
copies.
OCI publications are not usually so pointless. But managers throughout
CIA justify a good many such pieces of research and publication on the
ground that they keep up the analyst's morale, especially if his area is
quiet and dull. The same excuse is given by the Clandestine Services for
publishing raw reports known to be of marginal value or none. We consider
this excuse unworthy of a serious effort to produce good national intelli-
gence, and an important cause of the Information Explosion.
coninue next page)
HAND TOP SECRET
A ~ry o 1 FA I aTs g 9ii: Ff@$Gb(Yf:0'9 0 1005-5
 Approved For Releap 0~fMsk: --RDP85G00105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
23. This defect cannot be remedied by the automators; they are in
any case so devoted to hardware that they have invented a subtle deni-
gration, probably unintentional, of the programming (and' by extension
the thought that precedes programming) which determines the use to which
the hardware will be put. In their jargon this is software. The remedy
will have to be supplied by management: a workable definition of the
function of U. S. intelligence. *
24. Otherwise automation will accomplish the instant retrieval of
everything, no matter what, and infinite permutations of googols of binary
bits, no matter how trivial. One of the many functions of will be 25X1A2g
to "search the literature" and produce all the books, documents, reports,
etc. , on a given subject. What does the analyst do when the truckload is
delivered at the door ? Or when it is automated into another truckload of
electronic printout? That is his worry. Some electronics people obscure
this crucial limitation upon the system with another jargon phrase,
"individual channel capacity, " and change the subject. Yet there is no
* Our views on this point are influenced by the development of
for the Clandestine Services years ago. It became clear that great techno-
logical advances in storage and retrieval would have done more harm than
good if we had not first insisted upon drastically higher standards for what
was to be stored and retrieved. Automation therefore began with a new and
more rigorous definition of the purposes the information was to serve, a
severe purge of accumulated irrelevancies, and a program of education and
regulation throughout the Clandestine Services designed to acquaint all
users of information both with the advantages of automation and with the
necessity for self-discipline which these advantages imposed. The problem
the whole community now faces of what to automate is many times more
difficult than the one attacked by I, because it means redefining
the function of U. S. intelligence altogether, and the need for rigorous
standards of quality and relevance is correspondingly greater.
):():c (continued from Pate VII - 11)
VII - 12
TOY SEc.P VT
25X1A2g
25X1A2g
HANDkAr&36Ad ~1- ads 0~ :qW
I $ Wf6$P1640W0LW005-5
 Approved For ReIem9Q00f 1,$ e-RDP85G00105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
doubt that if its function is properly defined, automation can accomplish
many more marvels of the kind which caused a Soviet planning official to
say that CIA understood the Soviet economy better than the Soviet govern-
ment did. The biggest problem is managerial "software."
25. Given all these grisly statistics we would go far beyond Borel's
comment: "All in all a significant part of the information problem is of
our own creation. " He draws this conclusion only with respect to our
manufacture and distribution of classified documents, themselves only
a fraction of the "20 thousand individual series, in ten million issues,
published in 150 million copies, " which a survey of information inven-
tories and flows estimated that the community produces or handles every
year. In our view this problem is almost entirely of the community's
own creation, not imposed from outside, and is a product of the way it
has allowed its own "requirements, " both for collection and for production,
to proliferate unchecked. There would be a rough justice in blaming the
analysts for their own plight, because collectively they have stated require-
ments for everything under the sun. But it is more to the point to emphasize
that management has allowed this to happen.
26. We believe that the sum total of all our requirements, and the
Information Explosion they have caused to be created, are severely
detrimental to the American intelligence effort. In the long run it is',aot
the crude question of work load which matters most, nor even the point
that each item uses up customers' time and attention which cannot be given
to any other item, so that each of our products must receive steadily less.
What matters most is the question whether this quantity of information is
degrading the quality of all our work. It is the earnest conviction of those
of us who have studied CIA's requirements systems as a whole, and thought
of their effect on our work, that this is already happening, and can only
grow worse with each large new accretion. It is impossible to prove or
disprove this thesis, least of all by our customary reliance on statistics.
It has probably applied least where in the past it has mattered most--for
example, the community could hardly have given more intense attention
to Soviet military capabilities. But in many other important matters we
believe that the community's attention is becoming steadily more
superficial: that we cannot apply to the available information the depth of
analysis it requires for accurate judgments. In addition to those we have
suggested elsewhere, we offer these further indications, of course debata-
ble but worth considering:
a. The quantity of information on the Sino-Soviet dispute,
especially of official statements from both sides published by
HAND1 pVdAd1'I&R d6F2 0-M CObR RMg5G ( IBQO?'VQ,V0005-5
 Approved For Re1ease/8fr RCRDP85G00105R000100130005-5
HANDLE VIA TALENT'KEYHOLE CONTROL SYSTEM ONLY
25X1A8a
reinforces our geographically discrete, highly compart-
mented approach to understanding it. The Sinologues have no
time to master to their own satisfaction the wealth of information
from the Chinese side, including all that comes out in the 46, 000-
25X1A8a word volume and the Sovietologists are
equally swamped in the flood of broadcast and other information
on the USSR, and it is far beyond the capacity of either to study
in depth the behavior of the two governments together. One
solution is hiring more people, since we are especially short of
Chinese experts. Another is transferring people away from
attention to problems like "any indications, however indirect,
of Somali involvement in Eritrean dissidence" (an IPC target),
and onto important matters. But surely the quickest practical
change would be even greater selectivity as to what we read and
publish. The Sino-Soviet dispute has developed so slowly, with
such infinite repetition of arguments, that we would be better
occupied studying its underlying causes than in looking for
Byzantine subtleties in the latest several thousand words of
diatribe. Our compartmented, current-events approach kept
many people from even acknowledging the seriousness of the
dispute until it had been going on for several years. Of course
somebody must skim and screen as much as possible, but our
current handling of available reading matter makes skimmers
of us all.
b. As for China itself, we are informed that curious dis-
crepancies in the Chinese press, in the sort of items broad-
cast by the Chinese and monitored by II foreshadowed the 25X1A8a
current internal upheaval in China months before events drew
our attention to it. Perhaps greater selectivity would have
passed over these indications, perhaps not; at any rate turning
analysts into skimmers produced the same result as if the
indications had not been published. When we get down to
discussing the meaning of this upheaval, our preoccupation
with current events can only bolster the view that, since
Communists and their governments are pretty much alike,
this must be a Communist phenomenon to which some past
power struggle in Russia probably offers a reliable basis
for judgment. The extent to which it must also be a peculiatly
Chinese phenomenon gets lost in the welter of headlines and
daily bulletins.
TOP SECRET
HAN b Fi 'r s~ 9 I~ Q~ $6 0O T LSIgG l 41i 30 05-5
 Approved For Rele0(1?J4>-RDP85G00105R000100130005-5
0 121 HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
c. Our preoccupation with great quantities of information
about current events, at the expense of deeper study of what
causes the events to take place, is illustrated by our work on
Indonesia. On 1 July 1965 the U. S. intelligence community
ratified without reservation the view that:
"The principal development in Indonesia over the past
year has been the sharply accelerated growth of the
Communist Party (PKI) role in government. This
trend is likely to continue as long as Sukarno is in
control. Opponents of this trend are discouraged
and intimidated; even the military has all but lost
the will to resist.'
On 10 September a special supplementary estimate added that:
"Sukarno is the unchallenged leader of Indonesia
and will almost certainly remain so until death or
infirmity removes him from the scene.... Communist
fortunes in Indonesia will probably continue to prosper
so long as Sukarno stays in power.... If Sukarno lives,
it is probably that in two or three years the Indonesian
state will be sufficiently controlled by the Communists
to be termed a Communist state, even though Sukarno
remains the acknowledged leader. It will probably not
be possible, however, to detect any precise moment at
which the Communists 'take over,' unless Sukarno
chooses to proclaim it.... Conceivably, the PKI leaders
could become powerful enough to threaten Sukarno's own
dominance, but since his policies are likely to remain
along lines generally favorable to then, they are unlikely
to take risks in order to seize power. "**
These unusually confident judgments were based on an enormous
accumulation of surface facts which even in hindsight appear to
have been accurate; they were made by experts on Indonesia whose
qualifications cannot be impugned, and concurred in quickly by the
NIE 54/55-65, "Prospects for Indonesia and Malaysia."
NIE 55-65, "Prospects for and Strategic Implications of a Communist
Takeover in Indonesia."
CRET
HANDLE VIA TALENTOKEYHOL
EE CONTROL SYSTEM ONLY
Approved For Release 2002/06/18 : CIA-RDP85G00105R000100130005-5
 Approved For Releng?~QO MA:-EC [T
HANDLE VIA TALENT. KEYHOLE CONTROL SYSTEM ONLY
Board of National Estimates and USIB. Yet when the Communist
roof caved in three weeks later, events uncovered counterforces
so vast and violent that their nature ought to have been suspected.
The massacre of hundreds of thousands of Communists must have
had a motivation, a basis in a peculiarly Indonesian mixture of
politics and economics and religion and social institutions, which
our preoccupation with surface facts and current events kept us
from even imagining.
d. So with our approach to the Communist threat throughout
the underdeveloped world. Each group of experts is so busy
keeping up with quantities of current information on its own field
that there is no time for deeper study and comparison. When we
consider the Communist threat to, say, black Africa we can only
bring together experts on Africa, experts on international Commu-
ism, and experts on Russia and China, none of whom can be
deeply enough versed in the fields of the others, and add up their
anxieties instead of discriminating among them. When we do this
in turn to each potential Communist threat around the world, we
end up with a worst-case view of total Communist capabilities
which is greatly at variance with observable developments over
the past ten years.
e. For lack of deeper study we deceive outselves by applying
the narrowly economic concept of "underdeveloped countries" to
forms of society which have been highly developed for centuries,
but along lines too alien to our own for us to understand. We
therefore apply much the same standards of intelligence interest
to most "underdeveloped countries" as if they were all pretty
much alike, at least in their susceptibility to the attractions of
Communism; as if where the Communists try hardest they are
most certain to succeed. This neglects the emotional impact of
concepts like colonialism, nationalism, various forms of
xenophobia, racial and tribal animosities, and the search for a
national identity; some of these concepts we ignore as empty
slogans in the Cold War, and some we fail to see as obstacles
to any foreign ideology. In particular this approach neglects
the social force of ancient religions. We usually mention in
passing the impact of Islam or Buddhism, which varies widely
from country to country, as no doubt interesting but not crucial.
We are alarmed by the intense effort the Communists have been
putting for years into subverting the Muslim world, but we never
seem to ask ourselves why they have so little to show for it.
HAND, rr4&'p6rk S% 9&'U8 V FD MBjMs YM5]R .@OWO LVO05-5
 Approved For Rel1 e, 6VOg f l j fjA-RDP85G00105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
And for a long time we treated the Buddhist leaders of South Vietnam
as merely a set of particularly devious politicians. Yet when Muslims
run amok in Java and Hindus in Bali, and Buddhist monks and nuns
immolate themselves in South Vietnam, these oriental religions
must have more political and social force than our Cold-War simplifi-
cations take into account. The history and nature of religion in
China might help explain why China is now Communist and southern
Asia is not--just as in Russia the Communists subverted to their
own purposes the xenophobia, messianic zeal, and autocracy
inculcated for centuries by Russian Orthodoxy, but have had no
such advantage elsewhere in Europe. At any rate our black-and-
white simplicities have been much too simple.
f. A more specific example of damaging superficiality concerns
one aspect of the Vietnamese War. Faced month after month with
a lack of hard information from the scene, our experts had to
develop alternative bases for judgment, and these came inevitably
to include more and more statistics and extrapolations, unreliable
as these were known to be. One consequence was that our estimate
of the daily logistic requirements of the North Vietnamese Army and
the Viet Gong was extrapolated from American logistic requirements
with adjustments according to various untestable assumptions. A
U. S. Army study of some captured documents suggested that both
the CIA and the DIA figures might be much too high. Then it de-
veloped that some 7, 000 captured documents, of which a spot check
showed that 6, 000 bore on this problem, had been gradually accumu-
lating in our files for more than a year without being exploited for
this important purpose. So far have our work load and work habits,
reinforcing each other, led us away from using what the collectors
are in business to collect.
g. Meanwhile, in our effort to treat intelligence and policy
as two quite separate entities, we pay too little attention to the way
in which our own intelligence views affect the events themselves:
the question, for example, whether American policy based on our
view of the Communist threat does not increase the appeal of
Communism and thereby intensify the threat. In the cases of
TOP SEClRE
HAND, r-d,lddTdrlE Ma-sK)21(MLL'IEB C RDJU]k $R G0"A_V005-5
 Approved For Releas?6p/g6 1 RCS RDP85G00105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
Indonesia and Ghana we were fortunate, because the peoples rose
up and checked the threat for reasons of their own, not ours. But
what about Cuba? Did American policy, based on a worst-case
view expressed repeatedly by the intelligence community from
1959 on, cause Castro, with little effective help from Cuban
Communists or even (until later) from the Soviet Union, to convert
Cuba to a communist state?
27. To some readers these observations will seem remote from
the subject of collection requirements, but we believe that they are closely
connected. What we characterize here as superficialities are an amalgam
of preconceptions, simplifications, and work load which are dominated by
the belief that we must try to cover the whole world. The Long Range
Plan does not question its own explicit assumption that the "security
interests of the United States have expanded to include virtually every
inhabited spot on earth." The whole range of our stated requirements
from the PNIOs down to the most trivial item has justified such an
assumption and made our superficiality inevitable. But it is absolutely
essential that we do question it from now on. In the meantime we must
recognize that we do not understand the Asians and Africans and Latin
Americans; so long as our study is given largely to the surface events
of the moment, understanding them will be an unduly long, slow, expensive
process. This is an urgent reason for bringing the Information Explosion
under control and freeing time for deeper study of fewer subjects.
28. None of this highly negative recital should be read as ridiculing
our spectacular technical achievements, or denying that they have provided
crucial answers to some crucial questions, or arguing that we do not need
better information. But it does point to the necessity for the most earnest
consideration of the following propositions:
a. We are already collecting far more information than we can
satisfactorily use. What makes this alarming is not the quantity(cf
photographic film that receives only superficial scanning or of
SIGINT tapes that are not exploited at all; this kind of waste might
be tolerable, even unavoidable, if only we could be confident that
what we do use we use to the best of our collectively great abilities.
The real cause for concern is the danger to the quality of our
finished intelligence.
b. It is not the proper function of this Agency to know every-
thing about everything, even about all locally important develop-
ments in politics, economics, military affairs, and technology
TOP SECRET
HA~e~(~ir'A~~e'~Ol$~~Y ~~IAG?QD'1~'8~t0~~OC~$OD05-5
 Approved For Relea eo 028)F/OR RDP85G00105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
around the world. The same problem which faces us in converting
mountains of data into finished intelligence also faces the Executive
Branch in putting our finished intelligence to proper use--time to
study it, competition from other claims to attention, varied habits
of thought and work. But a further inhibition operates against full
use of our finished intelligence however perfectly we analyze and
interpret the data. We devote much attention to problems on which
the Executive Branch knows it is either unable or unwilling to take
action. Both Secretary McNamara and Secretary Rusk, thca forrrncr
with particularly persuasive force, have recently emphasized that
the U. S. cannot reform and police the whole world; this means to
us that much of our intelligence might as well not have been produced
for any practical difference it made. Yet we tell ourselves to act
as if every collectible scrap of information and finished intelligence
on all foreign developments were essential to the national interest.
This is to multiply the Pearl Harbor syndrome by the jigsaw theory
(that little scrap might be the missing piece) and get mediocrity.
It will take changing a great many attitudes inside the intelligence
community to bring it about, but we would like to hope for a time
when we can be more sure of ourselves, and right, about a few
important matters of which Indonesia is one example, and less
anxious about beating the newspapers to one more coup in Syria or
some other non-country.
c. As the only alternative to indefinite expansion, the Long
Range Plan says that "if the Government reaches a conscious decision
that the Agency should not expand to the degree that we propose, then
it must relieve the Agency of some of these responsibilities." The
implication that everything we do is the result of some specific re-
sponsibility laid on us by higher authority is dubious. We would
argue rather that the size and multiplicity of our work are to a
considerable degree a result of our own interpretation of broad and
vague guidelines, and that much of what we consider our specific
responsibilities is either self-imposed or responsive to no higher
authority than the management of the intelligence community itself--
the U. S. Intelligence Board. Witness the PNIOs, which originally
were to receive executive scrutiny by the National Security Council
but did not to any effective degree, and long ago came to represent
the community talking to itself. In this way traditions arising out
of our own past reactions to events take on the deceptive appearance
of fundamental imperatives.
d. Thus one of our most important functions is educating the
policy-makers to ask the right questions and to know what useful
TOP SECRET
HAND~pproved ejas ~0 M&F8 CuqN- 5( '(AiF0M]66-1I0005-5
 Approved For Releasy" 2/ RDP85G00105R000100130005-5
HANDLE VIA TALENT KEYHOLE CONTROL SYSTEM ONLY
answers they can rightfully expect from intelligence. Not that our
judgment of what is important is superior to theirs, but that it is
an important ingredient of their own capacity to make judgments,
and that our view of what it is practical for us to do to help them
is better based in operational and analytical experience.
e. It follows that it is up to us to redefine our own jobs. Keenly
aware that the community and CIA have habituated the Executive
Branch to certain high expectations, and that it would take diplomatic
skill and assiduity of a high order to revise them, we nevertheless
believe that the effort is worth making. Its first aim would be to
educate the policy-makers to an understanding that we would serve
their own interest better if we could concentrate our effort on the
crucial problems rather than try to cover the whole world compre-
hensively--and superficially- -in the way that has been assumed to be
essential ever since the Bogotazo of 1948. There could be two
improvements: better judgments, and less competition of secondary
matters for attention. Its second aim would be to educate manage-
ment within the intelligence community to the necessity for concen-
trating on the fundamentals and letting the incidentals go. Without
such education no possible combination of regulations, USIB reso-
lutions, systems analyses, and deliberations of boards, panels,
and committees will bring the Information Explosion under adequate
control.
HANDLE VIA A TOP SECRET
Approved _Tor he K1 i' : ~C~ 2 6 i~51 D !a&g3V005-5