. 8
( 53 .)


that someone will see a telltale pattern.”78
This approach to reform can be regarded as a combination of
an old problem with new technology. The problem of “connecting
the dots” was described forty years ago in a classic study of the

Secrecy and Security

intelligence failure before the attack on Pearl Harbor. The United
States had ample signs of an impending attack in 1941, Roberta
Wohlstetter argued, but the critical signals were lost in a “buzzing
and blooming confusion” of irrelevant information, or “noise.”79 The
task of distinguishing signals from noise constituted the intelligence
analysts™ key challenge. Then “ as now “ it was complicated by limited
organizational resources. (The FBI™s counterterrorism head told the
9/11 Commission that he wished he had had “500 analysts looking
at Bin Ladin threat information” in the summer of 2001 “ “instead
of two.”80 )
The most extreme example of an attempt to distinguish signals
from noise may be the University of California™s SETI project “ an
attempt to discern evidence of extraterrestrial intelligence from an
overwhelming amount of data collected by the world™s largest radio
telescope. The task of analysis requires enormous computing power,
which as a practical matter would be unavailable if investigators were
compelled to rely on a single computer. Instead, SETI researchers
developed an alternative approach, in which analytic tasks are under-
taken by a network of over three million personal computers that
receive data from SETI and process it using otherwise idle capacity.
Researchers estimated that the approach had yielded the computing
power equivalent to a $300 million supercomputer.81
This “distributed-computing” model has become an increasingly
popular approach for handling complex analytic problems. The post-
9/11 investigations recommended what is essentially the bureaucratic
equivalent, proposing a network in which data is widely shared and
that harnesses the analytic capacity of a much larger group of spe-
cialists. A prerequisite, according to the Commission, is the loosening
of Cold War-era rules that gave greater weight to the risks of inad-
vertent disclosure than it did to the bene¬ts of broad dissemination.
The approach to reform is liberal but still limited: It proposes better
information sharing principally inside the community of government
Nevertheless, a similar logic could be used to justify broader
information sharing to stakeholders outside government as well.
Like many of¬cials inside government, the general public remained
unaware of high-level concern about impending attacks in the sum-
mer of 2001. An informed public might have observed actions that
gave evidence of looming threats: It was an attentive citizen, after

Blacked Out

all, who alerted the FBI to the suspicious behavior of the alleged
“twentieth hijacker,” Zacarias Moussaoui.82 (A few weeks before the
September 11th attacks, actor James Woods also observed suspicious
behavior by four Middle Eastern men on a transcontinental ¬‚ight; he
later identi¬ed two as 9/11 hijackers.83 ) Representatives of the vic-
tims™ families later suggested that the public might have made sense
of events on the morning of September 11 more quickly “ perhaps tak-
ing steps to minimize the effect of the attacks “ if they had been told
about the intelligence community™s assessment of the risk of attack.84
Before September 11, the congressional inquiry concluded in 2003,
“the U.S. Intelligence Community was involved in ¬ghting a ˜war™
against Bin Ladin largely without the bene¬t of what some would
call its most potent weapon in that effort: an alert and committed
American public.”85
The Bush administration™s post-9/11 decision to invade Iraq pro-
vided further evidence of the ways in which excessive secrecy could
undermine, rather than enhance, national security. By fall 2004, it
had become clear that the war in Iraq had been justi¬ed on the basis
of intelligence that was badly ¬‚awed and twisted in its representation
to the public, and that planning for the post-combat occupation of
Iraq had also been inadequate. A policy of secrecy had aggravated the
weaknesses in analysis and planning, and created the possibility of
misrepresentation “ resulting in a prolonged and bloody campaign of
Public opinion polls conducted in early 2003 showed that Ameri-
can public opinion was overwhelmingly in favor of war with Iraq.86
Research found that support for the war was closely tied to popular
beliefs about Iraqi complicity in the September 11th attacks and the
threat that Iraq posed to American security. In a succession of polls,
a majority of Americans said that they believed there was clear evi-
dence of Iraqi involvement in the September 11th attacks; a near-
majority said that Saddam Hussein had been “personally involved”
in the attacks. Overwhelming majorities also believed that Iraq had
weapons of mass destruction.87 In a CBS poll conducted on the eve
of war in March 2003, 45 percent of Americans said that Iraq posed a
threat requiring immediate military action, and a similar proportion
expected that war would be “fairly quick and successful.”88
These assessments were later repudiated. In July 2004 the Sen-
ate Committee on Intelligence, reviewing the evidence available to

Secrecy and Security

the U.S. intelligence community, concluded that there was no evi-
dence that Saddam Hussein had tried to employ al Qaeda to conduct
terrorist attacks, and “no evidence proving Iraqi complicity or assis-
tance in an al-Qaida attack.”89 A CIA reappraisal in October 2004
also questioned earlier evidence linking the Hussein government to
an alleged al Qaeda leader, Abu Musab al-Zarqawi.90 In the same
month, the CIA™s Iraq Survey Group “ a team of experts given the
responsibility of searching for weapons of mass destruction in post-
occupation Iraq “ reported that the Iraqi government did not possess
such weapons at the time of the invasion, and was not actively seeking
to produce them.91 The head of the occupational authority also said
in October 2004 that the American government had underestimated
the number of troops that would be needed to preserve order in Iraq
after the invasion.92 A leaked CIA report concluded that Iraq faced a
signi¬cant risk of civil war.93
How had the American public come to hold opinions about Iraq
that proved to be so badly misguided? One reason was the Bush
administration™s public misrepresentation of the evidence available to
American intelligence agencies in the months before the war “ a tactic
that was feasible because of the secrecy that prevented a more com-
plete view of the available evidence. For example, a National Intelli-
gence Estimate produced by the CIA in October 2002 noted important
disagreements within the intelligence community about the threat
posed by Iraq, and its links to al Qaeda. However, the Estimate itself
remained classi¬ed and inaccessible to the general public. Legisla-
tors with access to the Estimate and other intelligence summaries
were compelled to sign agreements pledging not to release classi¬ed
information, and the CIA resisted declassi¬cation requests.94 After the
completion of the Estimate, however, senior administration spokes-
men repeatedly made public statements about the Iraqi threat and
its connection to al Qaeda that ignored internal disputes over the
interpretation of evidence that had been recorded in the classi¬ed
Some commentators suggested that public misperceptions were
also a re¬‚ection of misperceptions within the intelligence community
about the threat posed by Iraq before invasion. Borrowing a term
coined by psychologist Irving Janis,96 the Senate Intelligence Com-
mittee suggested in July 2004 that intelligence analysts had fallen into
a “collective groupthink” that led them to ignore evidence that was

Blacked Out

inconsistent with their preconceptions about the Iraqi threat.97 One
of the prerequisites for groupthink is the insulation of a group from
external forces that would challenge prevailing views.98 In fact, the
intelligence community has developed mechanisms “ such as inter-
nal “red teams” tasked with contesting the dominant interpretation
of evidence “ designed to mimic the role played by such external
forces. But the Senate Committee says that the intelligence com-
munity™s collective predisposition was so strong that these internal
mechanisms were not deployed. (The strain of events may have con-
tributed to the corrosion of critical analysis: According to another
report, at a critical moment in 2003, CIA head George Tenet dis-
missed a dissenting view about the reliability of intelligence on the
Iraqi threat “with words to the effect of ˜yeah, yeah,™ and that he was
˜exhausted.™”99 )
In August 2004, an internal CIA study concluded that its capacity
to analyze intelligence may have been compromised in other ways.
The leaked report suggested that the agency™s analytic branch had
“never been more junior and inexperienced” and that its ability to
assess intelligence was compromised by “tradecraft weaknesses.”100
If either view “ groupthink or limited analytic capacity “ is right,101
then it creates a powerful case for greater openness “ so that other
stakeholders can perform the essential function of weighing evidence
and challenging preconceptions about its meaning.
The British government became immersed in a similar contro-
versy over its handling of intelligence on the Iraqi threat. In 2003, a
parliamentary committee criticized the government for publicly exag-
gerating the evidence of an imminent threat, and for its unwilling-
ness to provide access to intelligence papers and personnel.102 In one
instance, Prime Minister Tony Blair unwittingly relied on material
that was found to have been plagiarized from a decade-old graduate
student thesis.103 Blair eventually conceded that the pre-war evidence
on Iraq had been largely wrong.104
The American government™s approach to post-war planning suf-
fered from the same weaknesses. In 2004, it was reported that classi-
¬ed pre-war assessments had warned of the possibility of prolonged
and intense internal con¬‚ict in Iraq.105 Before the invasion, however,
the Bush administration had publicly rebuffed pessimistic assess-
ments about the occupation; at the same time, it refused to release
details about its post-war planning. Pressed by reporters in February

Secrecy and Security

2003, Defense Secretary Donald Rumsfeld said that “it™s not useful”
to release details on post-war planning for public discussion.106 Two
weeks before the invasion, a Council on Foreign Relations report crit-
icized the Bush administration for its failure to “fully describe to
Congress and the American people the magnitude of the resources
that will be required to meet post-con¬‚ict needs . . . [or] their per-
spectives on the structure of post-con¬‚ict governance.”107 The com-
plaint was shared by humanitarian groups who hoped to participate
in reconstruction efforts.108 Journalist James Fallows later concluded
that key planners within the Bush administration had developed their
own “groupthink” on the question of occupation “ a blindness to evi-
dence of potential dif¬culties in reconstruction. “Everyone had that
˜Stalingrad stare™,” a senior administrator told Fallows. “People had
been doing stuff under pressure for too long and hadn™t had enough
The weaknesses in decision making that preceded the Iraqi war
have been seen before. In 1968, Professor James C. Thomson, Jr.,
wrote a widely acclaimed article in The Atlantic magazine that
attempted to explain the weaknesses in the United States govern-
ment™s policy toward Vietnam. Thomson, who had served in the
Kennedy and Johnson administrations, predicted that historians
would look back at the Vietnam years and wonder how “men of
superior ability, sound training, and high ideals” could have made
decisions that were “regularly and repeatedly wrong.” The answer,
thought Thomson, could be found largely in the process of decision
making itself. The concentration of responsibility at the top led to
executive fatigue and an inability to respond to new and dissonant
information. This was compounded by a lack of expertise within
key agencies and “closed politics” of policy making on sensitive
(Ironically, Defense Secretary Robert McNamara understood the
weaknesses of the process by which decisions on Vietnam were being
made. Unknown to Thomson, McNamara had taken the unusual step
a few months earlier of commissioning a large study of American
decision making on Vietnam. Unfortunately, the Pentagon Papers “ as
they were eventually known “ did little to improve the quality of gov-
ernment policy. Classi¬ed as TOP SECRET, the papers were largely
inaccessible inside government until they were leaked by Daniel
Ellsberg in 1971.)

Blacked Out

The problems observed by Thompson during the Vietnam con¬‚ict,
and experienced again in the months before the Iraq war, are likely
typical of large public bureaucracies. The concentration of authority
at the top of the bureaucratic pyramid means that leaders and their
advisors are overwhelmed with information, juggling problems that
are often outside their area of expertise. Fatigue, confusion, and igno-
rance about key facts are commonplace. In most circumstances, a pol-
icy of openness helps to check the damage that might arise because
of these bureaucratic pathologies. Transparency allows outside actors
to challenge evidence and present their own “ to assess the merits of
proposed policies and present alternatives. The analytic capacity
of a few bureaucracies is aided by the vastly larger analytic capacity
of the public sphere as a whole. It is, in a metaphorical sense, another
application of the distributed computing model used in the SETI
But this does not hold true in the security sector, where long
tradition “ and an instinctive reaction against disclosure of sensitive
information “ militate against transparency. This is justi¬ed in the
name of national security. Perversely, however, the security sector is
probably the one area where the consequences of poor analysis are
most severe, and where the more substantial analytic capacity of the
public sphere is most badly needed. In the long run, it may be a pol-
icy of openness, rather than secrecy, that best promotes security, by
avoiding the tremendous costs that can follow from poor bureaucratic
decision making.

Hardened targets
There is another powerful argument for greater openness on ques-
tions such as the threat posed by Iraq: the protection of basic political
rights. A decision to go to war is arguably one of the most important
choices that a nation can be expected to make, because it involves
an explicit gamble with human lives. Citizens are entitled to expect
that they will be given an opportunity to make an informed judg-
ment about the need for war. In the case of preventative rather than
defensive war, this expectation cannot be dismissed on the grounds
of urgency. Respect for the fundamental right to self-determination
demands greater openness.111 Secrecy, by contrast, compels the pub-
lic to defer to the judgment of a narrow elite.112



. 8
( 53 .)