Democracy and Homeland Security: Strategies, Controversies, and Impact

Part I. Papers Presented in Panels 1–4

Security’s Province in a Democratic Society

Alice Ristroph, University of Utah, S. J. Quinney College of Law

Jameel Jaffer, American Civil Liberties Union

A central theme of post–September 11, 2001, national security policy has been the control and regulation of information by the federal government, especially the executive branch. In general, the federal government has increased both surveillance and secrecy: it has expanded its own power to collect information even as it has imposed new limits on the ability of private individuals to obtain information about government activities.1 Few would disagree with the proposition that national security sometimes demands that government activity be kept secret or that private activity be made public. But once the door is opened to such practices, it becomes extraordinarily difficult to close it. From secret detentions to secret immigration hearings to the surveillance of libraries and bookstores, new policies make clear that national security can and will be used to justify ever-increasing restrictions of democratic liberties. As a practice, “security” has become a tool used by a small government elite to access, control, and selectively disseminate information. As a concept, “security” is slippery and potentially limitless in the hands of a government that claims not just the power to ensure security but also the authority to determine what security is, and what it requires.

We are not the first to observe that new national security policies involve significant and troubling restrictions of democratic liberties. What has not been sufficiently emphasized, and what we hope to illuminate with our focus on security policies that regulate information, is that these restrictions curtail not only individual freedom but also the democratic political process and the benefits generated by that process. Beyond the respect for individual choice and the protection of basic political rights, democracy promises to produce better policies: policies that are subject to critical review and thus better planned, more innovative, and ultimately more effective. When we regulate information in ways that impede collective decision making and democratic accountability, we lose these benefits. If we renounce democracy in pursuit of security, the result is likely to be not only less democracy but also less security.

This chapter first surveys the changes in information law and policy affected since September 11. We then consider theoretical arguments about the role of information in the democratic political process. Of course, proponents of new security policies have made the familiar claim that national security requires exceptions to general democratic principles. Security is said to need its own province, a nondemocratic territory within a democratic society. We examine the ways in which this province, previously contained by the temporal limits of discrete international conflicts, claims more and more territory as the “war on terrorism” makes crisis the norm rather than the exception.

As national security policies colonize more and more democratic practices, many Americans have looked to the judicial system to protect democratic liberties from the expansion of executive power. Proponents of expanded national security authority respond with the claim that courts are ill suited to strike the appropriate balance between security and liberty. At the end of our chapter, we argue that while the courts can and should defend democratic liberties from government overreaching, we misconceive the role of the judiciary if we think its only task is to balance personal liberties and national security. A key role of the judiciary is to preserve the democratic political process by minimizing government secrecy and protecting individuals’ ability to scrutinize and review government practices.

The government now possesses far more authority to monitor the activities of people living in the United States than it did before September 11. New surveillance measures can be roughly divided into two categories. First, there are broad monitoring programs that might be termed “generalized surveillance.” Second, the government possesses “investigative” surveillance powers that focus on individual suspects. Generalized surveillance programs aim to gain early intelligence of potential threats to national security and, to some degree, to discourage disfavored activity through the threat of quick detection. Such surveillance programs also aim to compile information about broad categories of people and to store that information for possible future use in more targeted investigations. Generalized surveillance is not based on individualized suspicion and is not primarily an effort to gather information about a particular suspect, act, or event.

Examples of general surveillance programs are plentiful. In May 2002 the U.S. attorney general issued guidelines permitting the FBI to infiltrate religious and political groups, even if such groups are not known to be linked to any wrongful activity whatsoever.2 Beginning in November 2001 and continuing into 2003, the Justice Department administered a “voluntary” interview program in which thousands of noncitizens of Arab and South Asian descent were interviewed about their families, their religious and political views, and their ties to foreign countries.3 In June 2002 the Justice Department announced a “special registration” program under which male nationals of predominantly Muslim countries are required to register in person with the government and submit to ongoing monitoring by the immigration authorities.4 In January 2003 the Justice Department launched the Student and Exchange Visitor Information System (SEVIS), which requires colleges, universities, and language schools to provide the government with information about international students.5

Perhaps the most notorious generalized surveillance program is the Total Information Awareness (TIA) program, a program that did not (quite) happen. This initiative of the obscure Defense Advanced Research Projects Agency (DARPA) would have collected and organized electronic files on every resident of the United States, drawing from financial, telephone, e-mail, consumer, prescription, medical, academic, and travel records. After extensive public criticism of the program’s initial design led Congress to ban its deployment, the Pentagon changed the name of the program to Terrorism Information Awareness.6 The Senate imposed restrictions on this second incarnation of TIA and then shut down the Pentagon office that ran it.7 Even so, the government has continued to pursue similar initiatives. The proposed Computer Assisted Passenger Pre-screening System II (CAPPS II), for example, aims to collect essentially the same information in order to create profiles for use by airlines.8 The profiles would be used to rate the security risk posed by each prospective flier. Proposed more recently, the Multi-state Anti-Terrorism Information Exchange (Matrix) is a state-level data-mining program that would use government and private databases to create dossiers about individuals; the dossiers would then be available to federal and state law enforcement officers. Matrix has garnered extensive criticism, and several of the states that initially expressed interest later declined to participate.9

In addition to these generalized surveillance programs, the government possesses “investigative” surveillance powers that permit more intrusive inquiries relating to individual suspects. The principal source of these new investigative authorities is the USA PATRIOT Act, which Congress signed into law only weeks after the September 11 attacks.10 The PATRIOT Act made numerous changes to the laws that govern FBI surveillance. For our purposes, it will suffice to note only a few examples. Section 213 of the act authorized the FBI to conduct what are known as “sneak-and-peek” searches in criminal investigations. The provision essentially allows the FBI to delay notifying a person whose home or office is searched until days or weeks after the search is complete. Section 215 of the act authorizes the FBI to obtain orders from a secret intelligence court requiring any person or organization to disclose “any tangible thing.” The provision allows the FBI to obtain membership rosters from political organizations, circulation records from libraries, and medical records from hospitals without individualized suspicion. Section 218 of the act expands the class of investigations in which the FBI can rely on the Foreign Intelligence Surveillance Act (FISA). FISA allows the FBI to conduct physical searches or install wiretaps without meeting the usual requirements of the Fourth Amendment (probable cause, particularity, and notice). Section 505 of the act expands the FBI’s authority to issue National Security Letters (NSL) requiring certain kinds of organizations—financial institutions, credit reporting companies, Internet service providers—to turn over records relating to their customers or subscribers. Previously, the FBI could issue an NSL only if it had reason to believe that the subject of the letter was a foreign agent (a spy) or a terrorist. The PATRIOT Act removed the individualized suspicion requirement and empowered the FBI to use the NSL provision against anyone at all, provided the FBI believes that the information sought is relevant to an investigation. There is no judicial oversight.

Congress has enacted other legislation that expands the government’s investigative surveillance authority. The Intelligence Authorization Act of 2004 further expanded the NSL authority by broadening the class of “financial institutions” that can be served with NSLs to include casinos, telegraph companies, travel agencies, pawnbrokers, and all other businesses whose cash transactions, in the secretary of the treasury’s view, “have a high degree of usefulness in criminal, tax, or regulatory matters.”11 The president recently endorsed legislation—the Anti-Terrorism Tools Enhancement Act—that would provide the FBI with administrative subpoena authority in terrorism investigations.12 The authority would allow the FBI to issue subpoenas for documents without having to work through a grand jury. Still other new surveillance programs have been initiated less formally. For example, in January 2003 the New York Times reported that the FBI ordered each of its field offices to “establish a yardstick for the number of terrorism investigations and intelligence warrants” by counting the number of Muslims and mosques in its district.13

New surveillance authorities—authorities that allow the government to obtain enormous amounts of information about the activities of people living in the United States—are all the more striking when contrasted with the unprecedented scope of refusals to release information about government activity. The Bush administration’s penchant for secrecy was evident even before September 11, 2001, but the executive branch has guarded information even more jealously since then.14

The opening salvo in the administration’s stepped-up campaign against access to information was a memorandum issued by the attorney general in October 2001 concerning the Freedom of Information Act (FOIA).15 Between 1993 and 2001, federal agencies administered the FOIA with what was essentially a presumption of transparency: they released records in response to public request unless disclosure would result in foreseeable harm. The attorney general’s October 2001 memorandum directed federal agencies to withhold records if they could find any sound legal basis for doing so. The memorandum thus replaced a presumption of transparency with one of secrecy. The new policy sharply restricts public access to information under the FOIA. Records were withheld even under the old policy, of course. But many more are withheld under the new one.16

FOIA policy is only one aspect of what has become a multifaceted campaign against public access to information about government activity. The government refused to release the names of the hundreds of immigrants detained in the FBI and INS sweeps that took place in the weeks and months after the 9/11 attacks. When these detainees were brought before judges, the government insisted that the immigration hearings be held behind closed doors.17 In criminal proceedings, the government has increasingly relied on evidence gathered under the Foreign Intelligence Surveillance Act.18 Defendants who are prosecuted on such evidence are not afforded access to the underlying surveillance warrants; as a result, the public, too, is denied access.19 The government has also increased its reliance on secret evidence in other contexts, relying on such evidence not only in prosecutions of accused terrorists but also in the now-discredited espionage case against James Yee, a chaplain at the Guantanamo Bay Naval Base, and in the espionage prosecution of Senior Airman Ahmad I. Halabi, a translator at the same base.20

Some of the new secrecy policies are connected to the expanded surveillance measures detailed above. To some degree, investigative surveillance depends on secrecy.21 The government does not (and could not) notify a surveillance target that it is wiretapping his or her phone. After the PATRIOT Act, it does not need to notify a target (until days or weeks later, anyway) that it has searched his or her home. Those private citizens who are ordered to turn information over to the government are often prohibited from disclosing that the government sought information from them. Section 215 of the act, for example, includes a subsection that prohibits any organization served with a Section 215 order from disclosing to any person that the government sought or obtained information from them. The National Security Letter provisions include similar language. Notably, all of the gag provisions are permanent; they persist even after the government has arrested the surveillance target or determined that the target is not after all a terrorist or criminal. The secrecy, that is to say, outlasts any security justification that could conceivably be offered.

These new expansions of secrecy and surveillance are but a subset of the legal and policy changes that the Bush administration has implemented since September 11, 2001. It is worth noting, however, that information policy is a threshold issue that must be addressed before we can fully evaluate other areas of security policy. Former Supreme Court Justice Benjamin Cardozo once wrote that “freedom of expression is the matrix, the indispensable condition, of nearly every other form of freedom.”22 Access to information is similarly indispensable to the democratic political process; we obviously cannot have an informed discussion about anything if we are not informed. We turn now to an analysis of the role of information in a democracy, considering first the arguments for the free exchange of information and then the argument that national security matters require a different approach to information.

Central to the notion of modern democratic government are two principles about the qualifications necessary to make political decisions. The first principle is one of equality in at least some basic measure of political opportunity. Every individual is entitled to participate in the political process. Basic political rights are not contingent on natural intelligence, education level achieved, training completed, or certifications attained. The second principle is an endorsement of collective decision making. It professes that the collective is a better decision-making body than any single individual or group of individuals; the most important political decisions are to be made by the many rather than the one or the few. Jeremy Waldron has called this second principle “the doctrine of the wisdom of the multitude,” and he locates it in the following passage from Aristotle:

For the many, of whom each individual is not a good man, when they meet together may be better than the few good, if regarded not individually but collectively, just as a feast to which many contribute is better than a dinner provided out of a single purse. For each individual among the many has a share of excellence and practical wisdom, and when they meet together, just as they become in a manner one man, who has many feet, and hands, and senses, so too with regard to their character and thought. Hence the many are better judges than a single man of music and poetry; for some understand one part, and some another, and among them they understand the whole.23

Note that in order for this justification of collective decision making to work, each individual must be guaranteed basic rights of political participation. Even though some individuals are “good” and some are not, each has some measure of wisdom, and a collective decision-making process allows these diverse forms and measures of wisdom to be combined.

To clarify this democratic approach to qualifications to rule, consider an alternative approach. In the famous argument of Plato’s Republic, a just city is one ruled by philosopher-kings, an elite class of wise men who have unique understandings of what is good. Political wisdom is emphatically not universal. The varying capacities of human beings are described in the noble lie’s claim that some are born with, as Plato wrote, gold “mixed in at their birth,” others with silver, still others with bronze or iron.24 In other words, Plato argues that the wise should rule, and the many are not wise; variations in natural intelligence as well as in education and training produce individuals with radically different mental capabilities. The response of liberal democratic theory is not to deny differences in natural intelligence or in aptitudes for learning but to argue that at least some degree of political wisdom does belong to each individual and to the collective as a whole.

Of course, differences in natural intelligence, education, and training obviously do affect our views of who is qualified to make political decisions. This observation does not defeat the claim that the many should rule, but it makes evident the need to distinguish among different kinds of political decisions. Consider the following types of decisions: (1) the selection of an (elected) official, such as a president, legislator, or governor; (2) the selection of nonelected officials, such as cabinet members, judges, or advisors; (3) the determination of a broad strategy in a particular policy area, such as whether and to what extent the government should provide publicly funded health care; (4) decisions about how to implement the strategy.

Liberal democratic theory assumes that a collective composed of individuals—the many—will and should participate in all of these political decisions. But most individuals will participate only indirectly: I vote for a delegate to the Electoral College who has promised to vote for the presidential candidate that I favor. I favor this presidential candidate because she has made—or I think will make—political decisions (whether to go to war, whether to fund health care) in the way that I want these decisions to be made. Some of the decisions that she will make entail the selection of other decision makers, and again I vote for a delegate who will vote for this candidate because I think this candidate has chosen or will choose decision makers (judges, advisors, agency heads) who will in turn make other decisions in the way that I want them to be made.

This is a detailed (and perhaps cumbersome) way of saying that modern democracy is both bureaucratic and indirect. But the details should be articulated, because they make clear that the selection of specialized decision makers is in no way a renunciation of the principle that the many should rule. Some forms of knowledge or wisdom—knowledge of the functioning of the Medicare system, for example, or an understanding of economic principles that can be used to predict when and whether Medicare will go bankrupt—are specialized. There is little question that educated, trained specialists possess some kinds of knowledge that many do not, and there is little question that certain decisions should be made by persons with this kind of specialized knowledge. Nonetheless, the democratic quality of our political system depends upon the fact that those who make broad strategic decisions and those who make specific decisions about policy implementation are ultimately accountable to the many.

A complex, mass democracy such as the modern United States permits (and probably requires) different classifications of decision makers—ordinary citizens, elected leaders, appointed specialists. But this division of the decision-making process does not necessitate a hierarchical approach to access to information. Instead, elected and appointed officials are ultimately accountable to the general populace for their decisions on policy issues. Accountability requires that the information relied on by specialists is at least available to the people. Most private individuals may not care to peruse the minutiae of Medicare finances, but that information must be available to the interested citizen who does want to see for herself how the system is being run. Frequently, information used by policy specialists is examined by watchdog groups, political opponents, and academic scholars, and doubtless, such scrutiny affects the decisions made and care exercised by the policy specialists. In short, a system that guarantees certain minimal rights of political participation must also ensure access to information.

A further requirement of democracy vis-à-vis information is developed by Waldron as he examines the doctrine of the wisdom of the multitude: Democracy presumes that collective decision making is not just the aggregation of random choices, but a deliberative process that actually produces a better decision than any one person could make alone.25 This deliberative process requires (among other things) the free exchange of information. It is not simply the possession of information by particular individuals that improves the collective decision-making process, but the public sharing and evaluation of information. Information is to be distinguished from misinformation by subjecting all purported facts to scrutiny in the marketplace of (facts and) ideas.26 This argument follows, of course, John Stuart Mill’s classic defense of free speech and open political debate in On Liberty, although Mill described the deliberative process in less gentle terms than those used by many contemporary advocates of deliberative democracy. “Truth . . . is so much a question of the reconciling and combining of opposites that very few have minds sufficiently capacious and impartial to make the adjustment with an approach to correctness, and it has to be made in the rough process of a struggle between combatants fighting under hostile banners.”27

Thus individual rights of speech, access to information,28 political expression, and deliberation help ensure that decisions made by the collective will be wise. Collective political wisdom depends on both individual wisdom and individual freedom.29 This point is important because it means that requests for government information from particular individuals must be honored even if the majority seems uninterested in that information. It means, too, that individuals within a democracy are not mere passive recipients of information as the government doles it out. Democracy does demand a principle of general disclosure by the government, but it further requires that individuals actively pursue information about government activities—not only information about what the government is doing, but also the information that specialized decision makers rely upon to make policy or strategy choices. In short, the democratic informational relationship is one in which government accountability is made possible by individuals’ access to both information with which officials make decisions and information about the decisions and activities of officials.

The theoretical arguments articulated so far suggest that current practices of official secrecy threaten the democratic nature of our politics, but does democratic theory have anything to say about surveillance?30 First, it should be reiterated that surveillance depends on secrecy, so that to accept a regime of surveillance is to endorse at least some measure of government secrecy. But the violations of individual privacy entailed in surveillance have further implications for the democratic political process. For democratic leaders to be accountable to their constituents, those constituents must not only have information about the leaders’ activities but also be willing to voice opposition to said activities or leaders. Extensive surveillance has a chilling effect and makes individuals less likely to engage in critical review of official actions.31 The Supreme Court articulated these dangers in United States v. United States District Court: “National security cases . . . often reflect a convergence of First and Fourth Amendment values. . . . History abundantly documents the tendency of Government—however benevolent and benign its motives—to view with suspicion those who most fervently dispute its policies. Fourth Amendment protections become the more necessary when the targets of official surveillance may be those suspected of unorthodoxy in their political beliefs.”32

The threat to democratic politics posed by surveillance is particularly acute when the government has authority to conduct surveillance on the basis of political affiliations and to use information gained through such surveillance to prosecute on the basis of political affiliation.

This account of democracy’s unease with secrecy and surveillance is not particularly controversial. But defenders of more extensive information regulation—and other expanded government powers—argue that national security requires an exception to normal democratic practices. On this account, security policy occupies its own undemocratic province within an otherwise democratic state. National security policy cannot and should not be held to democratic standards, the argument goes; we must simply acknowledge and accept “a tension of ‘tragic dimensions’ between democratic values and responses to emergencies.”33

Several defenses are advanced in favor of the unique, extrademocratic nature of national security policy. One argument addresses directly the democratic argument about qualifications to rule, and holds that with respect to security, collective decision making is not the best approach. A collective decision-making process takes too long; threats to security often require swift, decisive action. Furthermore, decisions about security should be made by trained and experienced specialists. Of course, as discussed earlier, many decisions are delegated to specialists in other policy areas, so the argument here is more than the mere need for expertise. The claim is that this type of expertise should not be subject to democratic review—on certain matters, the many do not know best. On national security matters, the few should have the last word.

A second argument focuses on secrecy in particular and emphasizes that disclosure of certain facts is itself dangerous. The classic example is atomic secrets. To make such information publicly available would be to invite disaster, for it would surely reach our enemies and they would have additional means to destroy us. It is not only weapons technology that is kept secret under this justification; many other pieces of information (military strategies, for example) are similarly kept classified on the ground that their disclosure would make the country more vulnerable.

In practice, the view of security as an extrademocratic province endorses not only extensive secrecy but also limitations or outright abandonment of various individual constitutional rights. For example, it is argued that the Fourth Amendment ought not to apply with full force when national security is at stake. Thus, there is little attempt to justify the administration’s expansive new surveillance measures on democratic grounds, and these measures are not subjected to ordinary constitutional standards.34 The argument is again one of impending danger, although here it is somewhat less evident how the proposed democratic restraint (the Fourth Amendment’s requirements) will lead to doom. But the claim must be that aggressive and even unconstitutionally aggressive surveillance is, in some circumstances, necessary to identify and intercept would-be terrorists. The common refrain is that the Constitution is not a suicide pact.35

Along these lines, Oren Gross has proposed an “Extra-Legal Measures” model of emergency powers in which public officials would be authorized a priori to act “extra-legally when they believe that such action is necessary for protecting the nation and the public.”36 Gross relies on such tired old favorites as the ticking bomb/torture fable, in which the government is “absolutely certain” that a bomb is about to go off, that they have the person who placed the bomb in custody, and that if they torture said person, he or she will disclose the location of the bomb. In such a situation, Gross argues, it would be appropriate to use torture. “Terrorists . . . do not observe the rule of law; neither should we.”37

Such arguments claim for “national security” a distinct territory in the landscape of government policy. In this territory, decision-making power belongs to the expert few rather than the many. Important questions for proponents of democracy, then, are whether and how the undemocratic powers of these expert few can be limited. Typical means of limiting government power are electoral accountability and the enforcement of individual rights by the courts. The first requires an informed electorate; the second requires courts willing and able to enforce constitutional rights. The expansive theory of national security described above precludes the application of either of these limits to national security powers.38

We must look, then, to other means of delineating the boundaries of security’s province. Before September 11, many expansions of government power in the name of national security were subjected to a temporal limit. Wars did not—and could not—last forever. They were simply too costly in terms of lives, resources, and public approbation (for even if wars begin with great popular support, that support tends to wane with time). Accordingly, extrademocratic measures justified as wartime necessities were unlikely to become permanent fixtures on the legal or political landscape. So, for example, President Abraham Lincoln suspended the writ of habeas corpus during the Civil War, but the end of the war brought the return of the writ. And the U.S. military infamously imposed a curfew and then internment on Japanese Americans in the midst of World War II, but the internment ended with the war.39

The war on terrorism has no such expiration date—as officials in the Bush administration have acknowledged. Vice President Richard Cheney has said that the war on terror “may never end.”40 Indeed, it is hard to imagine how it could end. There is no single enemy or discrete group of enemies who could surrender, and no clear way to tell when the enemies have been defeated. Furthermore, the very secrecy that is allegedly justified by the war on terrorism provides the administration an incentive and a means to perpetuate the war. If a state of crisis or emergency gives the executive greater powers, the executive has every incentive to preserve those powers by using them to declare the crisis ongoing. The periodic announcements of changes in the color-coded threat levels serve precisely this function. If it seems that public fear and the concomitant acquiescence to undemocratic practices are abating, a well-timed elevation of the threat level (and perhaps an instruction to buy duct tape) can renew executive power. And how would private individuals challenge this tactic? They are denied the information with which to assess threats themselves.

There is no temporal limit to the war on terrorism; emergency has become “the new normalcy.”41 An alternative limiting principle is one based on subject matter. We might try to limit the undemocratic powers described earlier to certain substantive policy areas—foreign intelligence gathering, military strategy, or court proceedings involving individuals suspected of terrorist activity. Unfortunately, subject matter fares no better than time in establishing boundaries for security’s province. A central aspect of the war on terrorism is intelligence gathering, and intelligence can potentially be gathered in a vast range of policy areas. National security policy now reaches subjects as diverse as banking, electronic communications, higher education, religious exercise, aviation and transportation, and of course all manner of public (and formerly private) records from drivers’ licenses to credit reports to rental car receipts. Any policy area from which information about individuals may be collected is a potential target of intelligence efforts in the name of national security.42

Without a subject matter, temporal, electoral, or constitutional limit, security’s province can grow as big as its rulers choose. The very definition of security becomes the prerogative of the elite few who claim to provide it. An increasingly broader array of government powers is defended as necessary to preserve security, and private individuals have no effective means of challenging this claim since they have little if any access to the pertinent information. Note that wariness about the scope of security’s province need not be based on a conspiracy theory or an assumption that those who claim undemocratic powers do so in bad faith. Doubtless, many public officials who claim to act in the interests of the nation have the best of intentions. But the fact that despotism is based in good faith is no consolation to those who value democratic processes.

A few specific examples illustrate the ways in which the regulation of information has impeded the democratic review of national security policy. Consider the government’s response to public interest organizations’ requests for a list of immigrants who had been detained in the post–September 11 INS and FBI sweeps. In the weeks and months following the September 11 attacks, those agencies detained well over a thousand immigrants. Public interest organizations, concerned that the government had detained individuals based solely on their race, religion, or nationality, submitted an FOIA request for information about the detainees. The government denied the request, refusing to make public the number of people arrested, their names, their lawyers’ names, the reasons for their detention, or the location of their detention.43 The government relied upon the so-called mosaic theory to defend its nondisclosure. According to the mosaic theory, any fact related to counterterrorism or national security efforts, no matter how seemingly trivial, is part of a larger mosaic. Until the mosaic is complete, the actual significance of any one piece of information is uncertain. The government must strive to piece together the mosaic of facts to identify and intercept terrorist activity, and it must simultaneously prevent would-be terrorists from gaining enough information about the government’s activities to piece together their own mosaic and thwart the government’s efforts.

The government also enlisted the mosaic theory to justify its refusal to release statistical information about the FBI’s use of surveillance authorities created by the PATRIOT Act. Public interest organizations sought statistics indicating the number of times that the FBI had relied on new surveillance authorities—the number of times the FBI had, for example, used Section 215 of the act to demand records, or had issued National Security Letters under Section 505 of the act. The Justice Department refused to release the information. With respect to the request for information about Section 215, the Justice Department argued that the statistic sought by the plaintiffs, while innocuous in itself, might, in the context of all of the other information that is and might become publicly available, provide “a very significant and meaningful picture of the FBI’s investigative efforts in the post–September 11, 2001, war on terror.” In other words, the government withheld the information on the grounds that the information might be meaningful. Of course, that the information might be meaningful was precisely the reason that the plaintiffs asked for it in the first place. At least in the national security context, the government seems to have adopted the view that the public is entitled to information about government activity only if the information sought is demonstrably meaningless.

The mosaic theory allows the government to keep information secret even without showing that its disclosure would be dangerous, since the argument is that any and all information is potentially dangerous. Further, the government need not explain even the potential danger; it can simply assert that information might be “significant” in some as-yet-undetermined way. The mosaic theory dispenses with any need to justify national security measures with any specificity; it allows the vague threat of danger to provide a blanket justification for government secrecy. Notably, the mosaic argument is literally limitless; it would justify the withholding of any information. If information could be withheld merely because it might contribute in some unspecified way to a more complete picture of government activity, the government could never be compelled to release information. To continue the geographic metaphor, security’s province can grow to an empire as the pursuit of security colonizes other democratic practices.

The mosaic theory enables widespread government secrecy and dispenses with a need to justify the secrecy with reference to specific dangers. The mere invocation of a national security interest now permits the government to exclude the public from decision making. The consequence is that security policy—or whatever else is kept secret—is not subject to democratic review and is insulated from both criticism and suggestions for improvement. The public cannot criticize government policy if it has no idea what the government’s policies are. Nor can the public hold the government accountable for abuses and excesses of which it isn’t aware.

Ultimately, the secrecy permitted by the mosaic theory and other aspects of national security policy can become a policy in itself—an approach to ruling and a strategy for holding and expanding government power.44 In fact, the Attorney General’s continued insistence on secrecy may be driven just as much by a desire to obscure the limits of government surveillance as to conceal its abuses and excesses. As the master of information, the government protects and expands its own power; it also has an increased capacity to manipulate the public through the strategic release of isolated bits of information. Two examples of such strategic releases are the leak of CIA operative Valerie Plame’s name after her husband criticized President Bush and the attempt to declassify Richard Clarke’s 2002 testimony to Congress in order to undermine him after he criticized the Bush administration in 2004.45 A third recent strategic declassification is Jamie Gorelick’s 1995 memo addressing the need to keep counterintelligence efforts separate from criminal investigations, released the day before John Ashcroft’s testimony to the September 11th Commission so that Republicans could call for Gorelick’s resignation from the Commission.46

The danger of ceding to national security an extrademocratic province is clear. But to date, efforts to limit the expansion of security’s province have not had much success. Many attempts by individuals and civil liberties groups to use constitutional litigation to challenge new security policy have been met with the claim that the judiciary is ill suited to decide the appropriate balance between liberty and security. Law professor Oren Gross, whose Extra-Legal Measures model is discussed above, makes this claim—but as Gross also acknowledges, members of the executive are no better suited and may in fact be less suited to strike that balance.47 It is obviously risky to give those who exercise power the authority to decide the scope of their power.

But more importantly, the rhetoric of “balancing” can cause us to misunderstand the difficult decisions that national security requires us to make. What popular parlance calls balancing liberty with security is, in most cases, the restriction of a known liberty in the name of a hoped-for but uncertain benefit. And that benefit is not any affirmative act or recognizable event but only a negative: the nonoccurrence of attacks or other hostile action. Because security is the absence of events rather than the presence of anything in particular, it is extraordinarily difficult to determine what policy best ensures security.48 Devising national security policy involves not so much balancing as guessing. This uncertainty is good reason for a more democratic decision-making process—here we need the benefit of many heads, the checks and balances of multiple participants, and especially the continual critical review of current practices with an eye toward possible improvement. Even though it may sometimes be prudent to restrict specific liberties to increase the likelihood that planned attacks will be detected and thwarted, the decision to restrict those liberties should be made democratically.

It has long been argued that through the protection of civil rights and civil liberties, American courts protect the democratic political process. In the context of national security policy, the role of the judiciary is not simply to balance liberty with security, but to ensure that security policies are products of, rather than substitutes for or limitations on, democratic procedures. For courts to serve this role, they must not defer to the executive branch at its mere invocation of national security powers. We suggest three general principles that could guide courts faced by a challenge by individuals or groups to a particular element of security policy.

Specificity. As should be clear, extensive government secrecy is one of the most troubling aspects of post–September 11 security policy; a principle of specificity aims to ensure that secrecy is a last resort rather than a general policy. Under this principle, a governmental decision to withhold information from the public cannot be justified merely by invoking the government’s interest in protecting national security. Rather, the government must advance an argument that both identifies a more specific harm that would result from the disclosure of the information and distinguishes the information at issue from information that is already in the public domain. In the context of the Freedom of Information Act, the specificity principle would mean the end of the mosaic argument. The mosaic argument, recall, is that apparently innocuous information must be withheld because such information, when combined with other information, could be of use to our enemies. The specificity principle would require, first, that the government explain exactly what harm would result from the information at issue.49 The argument that the information would be of use to our enemies is insufficiently specific because it fails to explain how the information would be of use. The specificity principle would also require the government to distinguish the apparently innocuous information at issue from information that is already in the public domain. In other words, it would require the government to advance an argument that could not as easily be used to justify any withholding. The government must explain why this information in particular must be withheld.

This demand for greater specificity in justifications of secrecy does not imply that we should no longer defer to those who do have special experience or expertise in the area of national security. On the contrary, one of the aims of the specificity principle is to ensure that such experience or expertise is actually exercised. Requiring officials to give reasons for their judgments means that they are more likely to consider facts carefully and actually exercise judgment in a particular case (rather than, say, simply repeating historical practice). Another aim of the specificity principle is to ensure that we defer to security experts in their areas of expertise but not outside them. An argument for the withholding of information from the public is likely to depend in some part on specialized expertise but also, in some part, on nonspecialized expertise. The specificity principle is meant to ensure that we do not defer to executive expertise where such expertise does not exist.

Narrow Tailoring. Here we borrow a principle from the “strict scrutiny” doctrine of constitutional law—the principle that the government may not restrict fundamental individual rights unless the restriction is narrowly tailored to the government interest that authorizes the restriction in the first place. In the context of information policy, the narrow-tailoring principle would require that any restriction on the public’s access to information be as limited as possible. The principle would require, for example, that any withholding of information be limited in duration; the government would have to justify at the outset not only the withholding of the information but the duration of the withholding. The information would be released automatically at the expiration of the term unless the government presented a new justification for extending the term of secrecy. The narrow-tailoring principle would operate similarly to limit the term of gag provisions that apply to surveillance laws. In its current form, the gag provision that applies to those who receive orders issued under Section 215 of the PATRIOT Act (50 USC 1861) is both categorical (i.e., the gag applies in every case, without reference to the government’s need for secrecy) and indefinite. It states: “No person shall disclose to any other person (other than those persons necessary to produce the tangible things under this section) that the Federal Bureau of Investigation has sought or obtained tangible things under this section.” Gag provisions that apply to other surveillance provisions suffer from the same deficiencies.50 The specificity principle would require the government to justify the gag on a case-by-case basis, and the narrow-tailoring principle would require the gag to be of limited duration.

Accountability. The accountability principle requires that the government identify the government agent who is ultimately responsible for any decision undertaken under the rubric of national security. This principle requires careful documentation of national security policymaking, especially since some decisions will not be immediately disclosed. Records of deliberations, proposals, and ultimate decisions must be carefully maintained. Accountability requires that we adopt a principle of signed classification: no document should ever be simply “classified,” but rather must be “classified by order of [name].” Requiring officials to take responsibility for their secrecy decisions means that they will be more likely to exercise careful judgment in making those decisions, and that once the decision becomes public (as it eventually will, under the narrow-tailoring requirement), the public will have the ability to hold the right person or agency accountable for excessive secrecy.

As we finish writing this chapter, the September 11th Commission continues to examine witnesses and documents in its effort to evaluate national security policies and practices in the years and months prior to the attacks. The Commission has had its share of partisan wrangling, and it is not necessarily an ideal model for a public review of national security policy. Yet few would argue that the Commission does not serve an important purpose. Suppose that this sort of detailed review of government practices were not simply an election year response to a public angered and grieved by hostile attacks on the United States, but it was instead a standard element of democratic politics. A standard review of security policy would have even more benefits—and few of the shortcomings—of the September 11th Commission: It would be less partisan and sensationalist, it would allow for even greater participation, and it would be more likely to produce changes to bad policies before those policies lead to national disaster.


In citing works in the notes, several legal abbreviations are used and are identified below.

6th Cir.                                        U.S. Court of Appeals for the Sixth Circuit

Cong.                                         Congress

D.D.C.                                         U.S. District Court for the District of Columbia

F.3d                                            Federal Reporter, 3rd series

F. Supp.2d                                 Federal Supplement, 2nd series

For. Intel. Surv. Ct. Rev.            Foreign Intelligence Surveillance Court of Review

Id.                                                In legal writing, the equivalent of ibid. (Latin for “in the same place”); refers to a
single work (e.g., book, court case, law) cited in the note immediately proceeding.

sec.                                             section

sess.                                           session

U.S.                                             U.S. Supreme Court

1. These developments in information law and policy are, of course, only a subset of the wide array of political developments following September 11. They merit special attention, however, because they severely curtail the public’s ability to evaluate the entire panoply of actions that the government takes in the name of national security. Racial and ethnic profiling and changes in immigration policy are two areas of post–September 11 policy changes that this chapter does not discuss in detail. Note, however, that some of the surveillance tactics discussed herein target particular groups—Arabs, Muslims, or South Asians.

2. Susan Schmidt and Dan Eggen, “FBI Given More Latitude: New Surveillance Rules Remove Evidence Hurdle,” Washington Post, May 30, 2002.

3. Brooke A. Masters and Cheryl W. Thompson, “U.S. Plans to Query More New Arrivals: 3,000 Foreign Nationals Added to List,” Washington Post (March 21, 2002), sec. A,; and Dan Eggen, “Missing Iraqis Sought: FBI Hunts for Thousands Here Illegally,” Washington Post (January 27, 2003), sec. A,.

4. Eric Schmitt, “Ashcroft Proposes Rules for Foreign Visitors,” New York Times (June 6, 2002), sec. A. The immigration program is called the National Security Entry-Exit Registration System (NSEERS).

5. Diana Jean Schemo, “Electronic Tracking System Monitors Foreign Students,” New York Times (February 17, 2003), sec. A.

6. Ariana Eunjung Cha, “Pentagon Details New Surveillance System: Critics Fear Proposed Extensive Use of Computer Database Raises Privacy Issues,” Washington Post (May 21, 2003), sec. A.

7. Carl Hulse, “Congress Shuts Pentagon Unit over Privacy,” New York Times (September 26, 2003), sec. A.

8. Steve Johnson, “Critics Point to Snoop Factor in Airline Security: New System Includes Background Checks,” Detroit Free Press (June 10, 2003).

9. Ray Burton III, “States Flee Matrix, Conn. Stays: Most States Abandon Controversial Data Collection Program,” Connecticut Law Tribune (March 29, 2004). Another generalized surveillance program that has not fared well is the Terrorism Information and Prevention System (TIPS), which would have enlisted a million letter carriers, meter readers, cable technicians, and other workers with access to private homes as informants for the Justice Department. In response to public concern about the proposal, Congress included language in the 2002 Homeland Security Act prohibiting the creation of the program. For more information on the legislation, see Homeland Security Act of 2002, Public Law 107-296, sec. 880, U.S. Statutes at Large 116 (2002): 2135, 2245.

10. Uniting and Strengthening America by Providing Appropriate Tools Required to Intercept and Obstruct Terrorism (USA PATRIOT) Act of 2001, Public Law 107-56, U.S. Statutes at Large 115 (2001): 272.

11. See Intelligence Authorization Act for Fiscal Year 2004, Public Law 108-177, sec. 374, U.S. Statutes at Large 117 (2003): 2599, 2628.

12. See Antiterrorism Tools Enhancement Act of 2003, HR 3037, 108th Cong., 1st sess. (September 9, 2003).

13. Eric Lichtblau, “F.B.I. Tells Offices to Count Local Muslims and Mosques,” New York Times (January 28, 2003).

14. Editorial, “Dodging a Bullet: Official Secrets Act Derailed,” Bergen County (NJ) Record (September 10, 2001). In 2001, for example, Republican legislators proposed a bill to increase criminal penalties for the disclosure of classified information but were forced to abandon the bill after a public outcry. See Classified Information Protection Act of 2001, HR 2943, 107th Cong., 1st sess. (September 21, 2001). Vice President Dick Cheney refused, for example, to provide records of his energy task force in response to a request from the Sierra Club. See Judicial Watch, Inc. v. National Energy Policy Development Group, 219 F. Supp.2d 20 (D.D.C. 2002); Naftali Bendavid, “Lawmakers Battle Bush over Secrecy: Executive Privilege Debate Rages Anew,” Chicago Tribune (September 8, 2001). The White House similarly refused to provide information about senior adviser Karl Rove’s closed-door meeting with Intel executives, a meeting that took place while Rove held more than $100,000 of Intel stock. See Bendavid, “Lawmakers Battle Bush”; and Josh Chafetz, “The White House Hides History,” New Republic (September 3, 2001).

15. U.S. Department of Justice, Office of Information and Privacy, FOIA Post: New Attorney General FOIA Memorandum Issued, October 12, 2001,

16. Ironically, one of the leading figures of the current administration was one of the most eloquent advocates of the FOIA when it was enacted in 1966. Secretary of Defense Donald Rumsfeld, then a congressman from Illinois, argued: “No matter what party has held the political power of Government, there have been attempts to cover up mistakes and errors. Significantly, [the FOIA] provides for an appeal against arbitrary decisions by spelling out the ground rules for access to Government information, and, by providing for a court review of agency decisions under these ground rules, [the FOIA] assures public access to information which is basic to the effective operation of a democratic society. . . . [O]ur democratic society is not based upon the vested interests of Government employees. It is based upon the participation of the public who must have full access to the facts of Government to select intelligently their representatives to serve in Congress and in the White House. This legislation provides the machinery for access to government information necessary for an informed, intelligent electorate.” See Congressional Record 13654 (June 20, 1966).

17. See Detroit Free Press v. Ashcroft, 303 F.3d 681, 683 (6th Cir. 2002).

18. Ted Bridis, “Record Terror Warrants Issued in 2002,”, May 1, 2003,

19. See War and National Defense, U.S. Code 50 (1994), subchap. 1806(f), which provides for in camera and ex parte review of evidence obtained under the Foreign Intelligence Surveillance Act.

20. “Pentagon Drops Chaplain Charges,” BBC News, March 20, 2004,; Sam Stanton, “Air Force Criticized in Spy Case,” Sacramento Bee (March 18, 2004), sec. A.

21. Even generalized surveillance—surveillance aimed at altering behavior rather than detecting specific acts—requires some secrecy. In designing the Panopticon, a utopian prison in which surveillance was to supplant violence as the principal means of social control, Jeremy Bentham proposed that perfect surveillance must be both visible and unverifiable: visible because citizens must understand that they may be monitored at any time, and unverifiable because citizens cannot be allowed to know when they are not being monitored. See Jeremy Bentham, Panopticon Letters, ed. Miran Bozovic (London: Verso, 1995), 29–95. But if surveillance is to be unverifiable, its limits must be secret. As Foucault noted, the effectiveness of surveillance depends on a disparity in information between the state and its citizens. Secrecy mystifies surveillance; it creates and sustains the requisite asymmetry between state and citizen. Michel Foucault, Discipline and Punish: The Birth of the Prison, trans. Alan Sheridan (New York: Vintage Books, 1995), 195–228.

22. Palko v. Connecticut, 302 U.S. 319, 327 (1937).

23. Aristotle, Politics, bk. 3, chap. 11: 1281a43–b9, quoted in Jeremy Waldron, “The Wisdom of the Multitude: Some Reflections on Book 3, Chapter 11 of Aristotle’s Politics,” Political Theory 23, no. 4 (November 1995): 563-84, esp. 564. As Waldron notes, Aristotle did not conclude from this argument that democracy was the preferred regime. One could argue that a collective composed of an aristocratic elite is superior to one composed of the entire populace, the rabble, and hoi polloi. Of course, Aristotle is not widely viewed as a democratic theorist, but contemporary democratic theorists often draw upon elements of Aristotle’s works.

24. Plato, Republic, ed. and trans. Allan Bloom (New York: Basic Books, 1991), 415a-c.

25. Waldron, “Wisdom of the Multitude,” 569–71.

26. As Justice Oliver Wendell Holmes put it, “The ultimate good desired is better reached by free trade in ideas. . . . the best test of truth is the power of the thought to get itself accepted in the competition of the market.” Abrams v. United States, 250 U.S. 616 (1919) (Judge O. W. Holmes dissenting).

27. J. S. Mill, On Liberty, ed. Elizabeth Rapaport (Indianapolis: Hackett, 1978), 46.

28. It has been pointed out that there is no affirmative “right to know” in the Constitution—that is, no specified right to access government information. The absence of an affirmative right to know is, in my view, only an indication of how fundamental to democracy the principle of free information is. One might also note that there is no “right to think” or even a “right to disagree” in the Constitution. As the quote from James Madison at the beginning of this chapter indicates, the founding fathers viewed knowledge (including knowledge of government activities) as fundamental to democracy.

29. Frances H. Foster, “Information and the Problem of Democracy: The Russian Experience,” American Journal of Comparative Law 44 (1996): 276–77, 289–91. Foster has argued that information policy in post-Soviet Russia has failed to encourage or sustain democratic politics. A major reason for the failure, according to Foster, is that information rights were framed and justified in terms of the collective rather than in terms of the individual. Because “the collective” acts as a unit only through the government, collective information rights can be defined and restricted as government officials see fit. To ensure that information will serve its democratic function of enabling citizens to review government actions, information rights must be individual rights.

30. “[W]hile the relationship between democracy and publicity on one hand and aristocracy and secrecy on the other is evident, the status of privacy is not so clear.” Maure L. Goldschmidt, “Publicity, Privacy, and Secrecy,” Western Political Quarterly 7 (September 1954): 404.

31. Many of the most forceful defenses of privacy against government surveillance and other intrusions focus on respect for individual autonomy and individual dignity. Those arguments are beyond the scope of this chapter, which focuses on the ways in which secrecy and surveillance affect the democratic political process. But here our lack of emphasis on individual autonomy and dignity is certainly not meant to suggest that those principles lack merit. To the contrary, they may provide an even stronger argument against the kinds of surveillance practices established in the United States after September 11.

32. United States v. United States District Court, 407 U.S. 297, 313–14 (1972).

33. Oren Gross, “Chaos and Rules: Should Responses to Violent Crises Always Be Constitutional?” Yale Law Journal 112 (2003): 1011, 28. Gross borrows the phrase “tragic dimensions” from Pnina Lahav, “A Barrel without Hoops: The Impact of Counterterrorism on Israel’s Legal Culture,” Cardozo Law Review 10 (1988): 529, 531.

34. See, for example, In re Sealed Case, 310 F.3d 717, 746 (For. Intel. Surv. Ct. Rev. 2002). The court upheld amendments to the Foreign Intelligence Surveillance Act (FISA) on the grounds that “the procedures and government showings required under FISA, if they do not meet the minimum Fourth Amendment warrant standards, certainly come close.”

35. Terminello v. City of Chicago, 337 U.S. 1, 37 (Judge Robert Jackson dissenting, 1949). Jackson wrote: “There is danger that, if the Court does not temper its doctrinaire logic with a little practical wisdom, it will convert the constitutional Bill of Rights into a suicide pact.”

36. Gross, “Chaos and Rules,” 1011, 23.

37. Id., 1097.

38. Id., 1111–112; Gross does propose some kind of democratic accountability as a limitation on extralegal action taken in the name of national security, but it is difficult to see how his proposal would operate in practice. Gross does not say when the extralegal action would be disclosed or how the process of public ratification would take place. He simply asserts in rather general terms that we should not worry about constitutional violations—democracy will take care of itself. The likely result of his proposal seems no different from the status quo: government officials would simply keep extralegal or undemocratic measures secret until and unless they felt they could disclose them without adverse political consequences. When extralegal measures are disclosed long after the fact (and after administrations and circumstances have changed), the public has little capacity to refuse to “ratify” them or to hold particular officials accountable.

39. For a brief history of the internment of Japanese Americans and the Supreme Court’s decision to uphold the internment as constitutional, see David Cole, Enemy Aliens: Double Standards and Constitutional Freedoms in the War on Terrorism (New York: New Press, 2003), 88–100.

40. Bob Woodward, “CIA Told to Do ‘Whatever Necessary’ to Kill Bin Laden: Agency and Military Collaborating at ‘Unprecedented’ Level; Cheney Says War on Terror ‘May Never End,’” Washington Post (October 21, 2001), sec. A,.

41. Id.

42. Gross, “Chaos and Rules,” 1073–89. Gross also considers possible limitations on the scope of emergency powers and finds each proposed limitation unworkable. He considers temporal limitations akin to those we discuss above, including geographic distinctions (which would limit emergency power to certain geographic areas), a domestic affairs/foreign affairs distinction akin to what we call here a subject-matter limitation, and communal divisions (which would allow emergency powers to be used only against certain noncitizen targets). Gross argues that none of these distinctions can provide any meaningful limitation on government power—which makes his ultimate endorsement of the Extra-Legal Measures model all the more puzzling.

43. Center for National Security Studies v. DOJ, 215 F. Supp.2d 94, 96 (D.D.C. 2002).

44. Daniel Patrick Moynihan, “Secrecy as Government Regulation,” PS: Political Science and Politics 30 (June 1997): 160–65; Moynihan argues that secrecy is itself a form of regulation.

45. Richard B. Schmitt, “Investigation Launched into Alleged White House Leak,” Los Angeles Times (September 29, 2003) sec. A,; Philip Shenon and Douglas Jehl, “Sept. 11 Panel Scrutinizing Past Testimony,” New York Times (April 1, 2004), sec. A.

46. Neil A. Lewis, “Rule Created Legal ‘Wall’ to Sharing Information,” New York Times (April 14, 2004).

47. Gross, “Chaos and Rules,” 1034, 1038–39, 1052.

48. No hostile attacks on American soil took place between December 7, 1941, and September 11, 2001. Was the homeland “secure” in the intervening sixty years and, if so, was that security attained through specific policies or by chance?

49. To say this another way, “[W]hen the Government defends a regulation on speech as a means to redress past harms or prevent anticipated harms, it must do more than simply posit the existence of the disease sought to be cured.” Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622, 664 (1994) (internal quotation marks omitted).

50. See, for example, Crimes and Criminal Procedure, U.S. Code 18 (2001), sec. 2709. This section of the code deals with the FBI’s authority to issue what are known as national security letters.