. 36
( 53 .)


This will exacerbate the anxiety that has already provoked govern-
ments to tighten their control over the out¬‚ow of government infor-
mation. (See Chapters 3 and 4.) Elected of¬cials and career bureau-
crats will argue that this degree of transparency goes too far, and
that the right to con¬dentiality for internal deliberations should
include a right to keep secret the topic of conversation as much as the
content of conversation. However, it is not clear how the bulk release
of metadata can be avoided under many FOI laws. Under U.S. law,
agencies might argue that metadata can be withheld because it is (in
the words of the federal FOI law) related to relatively unimportant
“internal practices” of an agency107 “ but this is unlikely to be a tenable
defense, because it will be widely recognized that the information is
being withheld precisely because it is critically important. Other laws
are crafted to deal with the content of speci¬c documents; it would
likely require a signi¬cant and controversial interpretation of exist-
ing law to allow the withholding of bulk data on the grounds that it
reveals compromising information about the pattern of conversation
within an agency.
There is an alternative response: Of¬cials may begin to evade the
system by generating documents that are not logged within their
EDRM system. This might easily be done if of¬cials are not pressed
to comply with ERDMS requirements. Indeed, our small sample of
data from TBS suggests that compliance with its EDRMS require-
ments could be spotty. Technological enthusiasts working for the
government™s Chief Information Of¬cer appear much more likely to
enter documents into RDIMS than employees in other TBS branches.
Routines for recording signi¬cant e-mail messages also seem erratic
across TBS™s branches, with a low proportion of total traf¬c being
entered into the system.
There is already a notorious precedent for the tactic of evasion. In
his early years as director of the FBI, Herbert Hoover revolutionized
its capabilities by introducing an advanced method of documents
management “ the Central Records System, which provided a stan-
dardized method of classifying information about FBI investigations
and simpli¬ed the task of searching for and retrieving information
by FBI employees. In retrospect we can see this as a precursor of
EDRMS. However, Hoover soon realized that there were records he
did not want logged in the widely accessible Central Records Sys-
tem, and within a few years he began to develop separate and secret

Liquid Paper

record-keeping procedures. Memoranda on sensitive (and often ille-
gal) activities were captioned “Do Not File” to ensure that they
would not be entered into the Central Records System. The sep-
arate procedures helped to thwart oversight. As Senator Richard
Schweiker observed during 1975 hearings on FBI misconduct, the
system allowed the bureau to say that its central records system
revealed no evidence of illegal conduct.108
Whether a comparable strategy of evasion could work in a contem-
porary bureaucracy is unclear. It is dif¬cult for large organizations to
get work done while relying on informal and covert procedures. The
willingness of employees to quietly accept practices that clearly con-
travene formal policy has also declined. This implies that the more
probable response to the threat posed by EDRM systems is likely to
be a reinterpretation or amendment of existing law.

Reacting to digitization
The digitization of government-held information appears, at ¬rst
glance, to provide extraordinary opportunities for improving trans-
parency. It seems to provide new tools for monitoring the work of
government itself “ either by accessing bulk data about routine gov-
ernment transactions or by pro¬ling the pool of unstructured data
held within government agencies. It also seems to provide new oppor-
tunities for monitoring business, by allowing access to data that gov-
ernment itself collects from the private sector. Finally, and most prob-
lematically, digitization creates new opportunities for monitoring
citizen behavior, by allowing businesses and other actors “ including,
in sometimes circuitous ways, government agencies “ easier access
to the mass of personal information that is aggregated within public
The extent to which these opportunities will remain available is
unknown and is, in fact, likely to be one of the major points of con-
tention in coming years. As the power of these new monitoring tech-
niques becomes apparent, we can expect the actors that are subject
to monitoring (businesses, citizens, governments) to react by seek-
ing to restrict the in¬‚ow of information to government databases, or
the out¬‚ow of that information from public agencies. In fact, we can
already see these battles over the in¬‚ow and out¬‚ow of digitized data

Blacked Out

In some respects, reaction against the impact of digitization is
not only predictable but justi¬ed. Closer monitoring of governmental
or business practices is one thing; tighter surveillance of individual
behavior is another. Indeed, it would be ironic if transparency laws
that are justi¬ed in the name of human rights had the effect, in prac-
tice, of compromising civil liberties. And it is far from clear that a
radically heightened capacity for monitoring governmental activity
is necessarily in the public interest. Although the prospect is now
entirely hypothetical, we could envisage a moment when businesses,
exploiting the opportunities posed by access to ERDMS metadata, are
capable of closely tracking the work of government agencies, inter-
vening to nip unfavorable policies before they are even partly formed.
Public agencies, subject to a degree of openness that is far greater than
that imposed on the businesses and other organizations with which
public of¬cials must work, might be incapacitated by transparency.
The result of these struggles may be the imposition of new legal
restrictions on access to digitized data, including rules that restrict
access to certain kinds of individuals or impose conditions on the use
of information. For many, the idea of differentiating access rights in
this way is anathema: It seems to undermine the long-standing prin-
ciple that rights of access should be equal and universal “ a notion
sometimes caught in the phrase “public is public.” However, the real-
ity has never been so simple. The act of inspecting records at the town
of¬ce or courthouse imposed its own costs: the knowledge needed to
determine the location of the records; the time needed to visit the
of¬ce; the nerve often needed to request even information that is, as
a settled matter of law, publicly accessible. These practical barriers
assured a rough form of access control, discouraging access by indi-
viduals who were not from the community, familiar with the law and
processes of government, and known by local of¬cials. Technological
change may be causing us to articulate access restrictions that were
already embedded in informal practice.
While new technologies destroy these old practical barriers to
access, new practical barriers could also be created. The growing pool
of digitized information will be accessible only to nongovernmental
organizations that have the technical expertise to understand how
information is structured within government agencies and under-
stand as well the agencies™ capacity to extract that information from
their databases in usable form. Manipulating that data, or extracting

Liquid Paper

meaning from a large mass of data, will also impose substantial bur-
dens on nongovernmental organizations. It is one thing to thumb
through a few thousand pages of memoranda; it is another to manip-
ulate tens of thousands of digitized records, each containing dozens
of ¬elds of data, released in tab-delimited ASCII format; and another
to make sense of the several hundred thousand lines of ERDMS meta-
data that might be generated in a few months by a reasonably sized
These new barriers to access can be substantial. Technical and
¬nancial constraints already limit the capacity of American journal-
ists to undertake computer-assisted reporting projects. The cost to the
nongovernmental organization Environmental Defense of its estab-
lishing a website to exploit the mass of Toxic Release Inventory data
collected by the EPA was initially $1.5 million; the work got done with
the support of private philanthropies.109 These are the constraints
imposed in the United States, which is af¬‚uent and advantaged by a
thriving media and nongovernmental sector. As a matter of practice,
the impediments to access to digitized government information in
other countries could be much more substantial. This has two impli-
cations. The ¬rst is that the opportunities for heightened transparency
could be undercut, not only by the reaction of other stakeholders, but
by the practical dif¬culties associated with extracting and manipu-
lating digitized data. The second is that problems in equity of access
to government information might be even more severe than they are


Many students of contemporary journalism argue that news produc-
tion consists largely of the reproduction of stock stories “ the retelling
of archetypal narratives already familiar to both journalists and their
audiences. The facts may change from year to year; the plot does not.
Relying on an archetypal story line, journalists achieve certain ef¬-
ciencies. They know what facts are needed to make the story work,
and they do not need to explain to readers or viewers what the story
is “about.” The story helps to organize reality. Moreover, the story line
imposes a moral order: When we begin to tell a story, or read a story,
we are led to a certain view of how it ought to end.1
This sounds very abstract. But at least one of these archetypal
narratives will be instantly familiar. Imagine the following story line:
Powerful of¬cials abuse their authority and injure innocent people.
They attempt to hide their abuses. Tenacious outsiders struggle to
reveal the facts, but are thwarted by of¬cial indifference and out-
right obstruction. Ultimately, however, the truth comes out. The citi-
zenry is outraged, and of¬cials are brought to account. Reforms are
introduced to prevent future abuses. The story comes to a satisfying
It is not dif¬cult to ¬nd this story line at work in the American
media. A famous example is the Watergate scandal, in which two
determined reporters for the Washington Post, Bob Woodward and
Carl Bernstein, exposed the role of the Nixon White House in orches-
trating a break-in at the Democratic Party national headquarters
and “dirty tricks” during the 1972 presidential campaign. Here poli-
tics merged with popular culture. The ¬lm version of the scandal,

Blacked Out

All the President™s Men, was pitched by Warner Brothers Studio as “The
most devastating detective story of the century!”2 The d´ nouement
was equal in scale to the con¬‚ict that preceded it: Not only did
the President resign, but Congress adopted a series of legislative
checks it hoped would prevent comparable scandals in the future (see
Chapter 3).
Even in 1972, this story line was not new. On the contrary, it was an
old mainstay of progressive reformers. Seventy years earlier, in 1902,
Ida Tarbell tallied the misdeeds of the Standard Oil trust in a series of
articles for McClure™s Magazine, fueling a controversy that led even-
tually to the breakup of the trust. Another journalist, Upton Sinclair,
exposed horri¬c conditions in the meat-packing industry in his 1906
book The Jungle, provoking Congress to pass new laws regulating the
industry. A third muckraker, Lincoln Steffens, detailed the corruption
of major American cities, prompting prosecutions and the turn-out
of legislators who had tolerated graft. The sequence was always the
same: abuse, expos´ , outrage, reform.
Revelation of misconduct was the crucial second step. “Sunlight
is said to be the best of disinfectants,” said Supreme Court Justice
Louis Brandeis in Harper™s Weekly in 1913, “electric light the most
ef¬cient policeman.”3 This faith in the catalytic power of disclosure
bolstered reformers in subsequent years: If only the facts were made
public (many thought), justice would follow.
In the spring of 2004 the controversy over abuse of detainees
held by U.S. forces in Iraq and at Guantanamo Bay seemed to be
evolving to ¬t this old story line. A recalcitrant executive branch
appeared determined to hide the facts from the American public.
(According to polls taken in early 2004, many Americans believed
that the Bush administration was “mainly trying to cover up” reports
of prisoner abuse.4 ) Nevertheless the facts came out. Documents
and photographs were leaked; investigations were undertaken; entire
books were written. It was clear that senior administration of¬cials
had attempted to contort traditional understandings of national and
international law to justify more aggressive methods of interroga-
tion. Poorly trained and inadequately supervised troops were seen to
have engaged in horri¬c abuses. Many of the military™s detainees were
acknowledged to have little or no intelligence value; they were minor
players caught by events.

The End of the Story?

Despite the administration™s secretiveness, these facts appeared to
take hold in the popular consciousness. According to polls, a majority
of Americans believed that their government was using torture in its
campaign against terrorism “as a matter of policy”; a larger majority
believed that their government was using “physical abuse that falls
short of torture.”5 (A post-election poll showed that two-thirds of the
American public believed that torture of prisoners by Americans had
taken place in Iraq and Afghanistan over the preceding two years; 40
percent believed that torture was still happening, in spite of the Abu
Ghraib controversy.6 ) Most Americans said that Justice Department
memos interpreting the legal constraints on torture contributed to
the Abu Ghraib abuses,7 and over one-third believed that leaders in
Washington were involved in a “decision” to abuse and torture the
Abu Ghraib detainees.8 Moreover, Americans said that this sort of
behavior violated norms of acceptable conduct; a substantial major-
ity said that torture was never acceptable. A majority was even pre-
pared to say that “physical abuse that falls short of torture” was never
By the early fall of 2004 the abuse story appeared to be nearing
its natural conclusion. All that was needed was a ¬tting conclusion “
the rendering of justice. Of¬cials had to be held accountable for the
outrage. There was already dissatisfaction among editorialists over


. 36
( 53 .)