Transparency, Corruption, and Democratic Institutions

This essay examines some of the institutional arrangements that underlie corruption in democracy. It begins with a discussion of institutions as such, elaborating and extending some of John Searle’s remarks on the topic. It then turns to an examination of specifically democratic institutions; it draws here on Joshua Cohen’s recent Rousseau: A Free Community of Equals. One of the central concerns of Cohen’s Rousseau is how to arrange civic institutions so that they are able to perform their public functions without being easily abused by their members for individual gain. The view that Cohen sketches on behalf of Rousseau offers a clear framework for articulating institutional corruption in democracy. With this account of democratic institutions in place, the essay turns the discussion to the role of transparency in deterring institutional corruption. The basic thought here is perhaps unsurprising: to ensure that a democratic institution is serving its public function and not being manipulated for self-interested gain, its activities must be subject to public scrutiny, and so these activities must be transparent to the public. Saying this makes the role of transparency in a well-functioning democracy clear, but it does not settle how transparency is to be realized. The essay argues that transparency can be realized in a democracy only by an extra-governmental institution that has several of the familiar features of the press. If this is correct, it follows that in its design and in many, though not all, of its activities, WikiLeaks provides a contemporary example of such an institution.

This essay examines some of the institutional arrangements that underlie corruption in democracy. It begins with a discussion of institutions as such, elaborating and extending some of John Searle's remarks on the topic. It then turns to an examination of specifically democratic institutions; it draws here on Joshua Cohen's recent Rousseau: A Free Community of Equals. One of the central concerns of Cohen's Rousseau is how to arrange civic institutions so that they are able to perform their public functions without being easily abused by their members for individual gain. The view that Cohen sketches on behalf of Rousseau offers a clear framework for articulating institutional corruption in democracy. With this account of democratic institutions in place, the essay turns the discussion to the role of transparency in deterring institutional corruption. The basic thought here is perhaps unsurprising: to ensure that a democratic institution is serving its public function and not being manipulated for self-interested gain, its activities must be subject to public scrutiny, and so these activities must be transparent to the public. Saying this makes the role of transparency in a well-functioning democracy clear, but it does not settle how transparency is to be realized. The essay argues that transparency can be realized in a democracy only by an extra-governmental institution that has several of the familiar features of the press. If this is correct, it follows that in its design and in many, though not all, of its activities, WikiLeaks provides a contemporary example of such an institution.

ABSTRACT:
This essay examines some of the institutional arrangements that underlie corruption in democracy. It begins with a discussion of institutions as such, elaborating and extending some of John Searle's remarks on the topic. It then turns to an examination of specifically democratic institutions; it draws here on Joshua Cohen's recent Rousseau: A Free Community of Equals. One of the central concerns of Cohen's Rousseau is how to arrange civic institutions so that they are able to perform their public functions without being easily abused by their members for individual gain. The view that Cohen sketches on behalf of Rousseau offers a clear framework for articulating institutional corruption in democracy. With this account of democratic institutions in place, the essay turns the discussion to the role of transparency in deterring institutional corruption. The basic thought here is perhaps unsurprising: to ensure that a democratic institution is serving its public function and not being manipulated for self-interested gain, its activities must be subject to public scrutiny, and so these activities must be transparent to the public. Saying this makes the role of transparency in a well-functioning democracy clear, but it does not settle how transparency is to be realized. The essay argues that transparency can be realized in a democracy only by an extra-governmental institution that has several of the familiar features of the press. If this is correct, it follows that in its design and in many, though not all, of its activities, WikiLeaks provides a contemporary example of such an institution.

INTRODUCTION
WikiLeaks, an organization that collects and disseminates individual, corporate, and governmental secrets, conceives of itself as part of the journalistic media. On the importance of the media and its own place in it, WikiLeaks says the following: Publishing improves transparency, and this transparency creates a better society for all people. Better scrutiny leads to reduced corruption and stronger democracies in all society's institutions, including government, corporations and other organisations. A healthy, vibrant and inquisitive journalistic media plays a vital role in achieving these goals. We are part of that media 2 . This self-ascription is contentious; U.S. Congressperson Peter King has sought, for example, to have WikiLeaks declared a foreign terrorist organization 3 . If we are to say whether WikiLeaks is part of the journalistic media, or a terrorist organization, or something else, we will need to know the details of WikiLeaks's activities. These are many and varied, and it is beyond the scope of this essay to treat them in much depth, although I shall make some remarks on them as the essay proceeds 4 . However, we will need to know more than just these details; to know whether to agree with WikiLeaks or King, we will need to know what it is to be part of the journalistic media in a democracy. My discussion here will be organized around this question.
If we understand the phrase 'journalistic media' liberally, then any individual or group that channels the day's events to an audience is part of the journalistic media. In this sense, when your child comes home and reports what has happened on the playground that day, she is part of the journalistic media. In this sense, even if WikiLeaks is properly understood as a terrorist organization, it is also part of the journalistic media, for its activities channel information about governments and corporations to anyone with an uncensored Internet connection. King does not appear to have this sense of 'journalistic media' in mind when he suggests that WikiLeaks is a terrorist organization; he seems to think that because it is a terrorist organization, it is not part of the journalistic media. The latter, but not the former, has a legitimate role to play in a democracy. WikiLeaks seems to have this narrower understanding in mind as well when it claims that the journalistic media reduces corruption and strengthens democracies.
In this essay I will defend a version of this last claim. I will argue that in a democracy one of the institutional roles of the journalistic media-or, as I shall speak of it, the press-is to reduce corruption. Once we understand this role and its place in a well-organized democracy, we have no choice, I will conclude, but to see WikiLeaks's more notorious activities as those of a well-functioning democratic press. My argument will proceed as follows. Because I will be focusing on the press as an institution, I will open with a discussion of what, in general, an institution is. My remarks here will be structured around John Searle's account of institutions. I will then proceed to an examination of specifically democratic institutions. My point of departure will be a discussion of Joshua Cohen's Rousseau: A Free Community of Equals 5 . Few historical figures have provoked more interpretive disagreement than Rousseau, so one might think that I am generating unnecessary confusion by discussing Rousseau here at all 6 . Let me defend this choice at the outset. My argument requires an account of democratic institutions that demonstrates both how they can be corrupted and how such corruption can be deterred. The fifth chapter of Cohen's book, titled "Democracy," offers just such an account. The key claims I will take from this chapter are, I think, true on any viable conception of democracy, so I suspect that the argument I will offer can be advanced without adopting Cohen's Rousseauean vocabulary. I nevertheless shall use that vocabulary, for I find the central thought it expresses-viz., that we can understand institutions as unified and guided by wills-to be of great use in examining institutional corruption.
With this account of democratic institutions in place, I turn the discussion to the role of transparency in deterring institutional corruption. The basic thought here is perhaps unsurprising: to ensure that a democratic institution is serving its public function and not being manipulated for self-interested gain, its activities must be subject to public scrutiny, and so these activities must be transparent to the public. Saying this makes the role of transparency in a well-functioning democracy clear, but it does not settle how transparency is to be realized. I argue that transparency can be realized in a democracy only by an extra-governmental institution that has several of the familiar features of the press. If I am correct on this point, it follows that in its design and in many, though not all, of its activities, WikiLeaks provides a contemporary example of such an institution 7 .

INSTITUTIONS
Let us open with a general discussion of institutions. The majority of contemporary academic work on institutions has been conducted by economists, and they tend to characterize institutions as "the rules of the game 8 ." Searle's view of institutions fits with this conception, for Searle defines institutions as bound by constitutive rules. The rules of chess provide a paradigm case of these sorts of rules. If one takes a chess set and moves the pieces around however one likes, one is not playing chess; one must move the pieces according to the rules of chess to count as playing chess. According to Searle, it is this count-as relation that is key to understanding institutions. The institution of marriage exists only because we count certain people as being married to one another; the institution of government exists only because we count what certain people say as the law; the institution of property exists only because we count certain physical boundaries as marking property rights. In each of these cases, Searle thinks, the institution exists because we collectively follow and enforce its defining rules; if not for these rules, the institution would not exist 9 .
By creating institutions, we not only structure our actions, but we also, Searle thinks, create deontic powers. Consider the institution of government, specifically in its legislative capacity. Governments have the power to make and to enforce laws. This power is not a brute physical power, such as gravity; rather, it is a power that arises from a society counting certain of its members as legislators. If we want to indicate that a given individual has or is constrained by a deontic power, we speak of that individual's rights, or duties, or obligations, etc. To speak of the deontic power of a citizenry to elect its own leaders, for example, we speak of the right to vote; to speak of the constraints instituted by property laws, we speak of the obligation not to trespass. The thought that institutions give rise to deontic powers can be, I think, quite helpful; I shall appeal to it on several occasions over the course of this essay 10 .
As for Searle's claim that all institutions are bound by constitutive norms: this may be correct, but there are some institutions that cannot be fully understood without an explication of their functional roles 11 . A national postal service, for example, might be bound by constitutive norms, but one will not understand this institution if one does not understand the function performed by a postal service. The function of many institutions is social maintenance, i.e., to help maintain the society to which it belongs. This is the broader function of a postal service. Its narrow function, of course, is to deliver mail and packages. Such things need to be delivered, however, to foster communication and commerce, which-in a commercial society, at least-are integral to the maintenance of that society.
This brief discussion of institutions gives us a framework within which to analyze both democracy as such and the organizations that may exist within a democracy. To be a democracy is to count as a democracy, so we can ask what is necessary for a multitude to count as a democratic organization. We can also ask if democracy has a function, and if so, what it is; if it has a function, we can ask what deontic powers must exist for it to execute its function. If some of these deontic powers belong only to some of the members of a democracy, then those powers will define democratic institutions. The function of these institutions will be to maintain the broader institution of democracy to which they belong. This way of thinking about democracy and democratic institutions lines up neatly with Cohen's interpretation of Rousseau, with one crucial omission: the framework just presented makes no use of Rousseau's central concept of the General Will. Adding this concept to the framework will enrich its explanatory power; let us turn, then, to a discussion of the General Will 12 .

DEMOCRATIC INSTITUTIONS
According to Cohen, the concept of the General Will is Rousseau's answer to the fundamental problem of collective self-rule 13 . The problem: "[t]o find a form of association that will defend and protect the person and goods of each associate with the full common force, and by means of which each, uniting with all, nevertheless obey only himself and remain as free as before 14 ." The function of the General Will, then, is to provide a solution to this problem. If the General Will is the solution to this problem, then either it just is the form of association described above or it produces that form of association. I will assume the latter and treat the General Will as a special kind of cause, one that can guide the action of individuals who belong to the association united by that Will. When the General Will causes a person to act, one of the act's effects is the partial production of the form of association described in the problem. The General Will cannot be an alien force determining action; rather, it must be a force produced and maintained by the individuals whose action it is to guide. An association that makes its own laws thus gives itself a General Will, and its activities follow this will to the extent that the members of the association abide by their own laws. When these members follow their own laws, their actions aim, in part, at the maintenance of a well-functioning social unity 15 .
Although the formation of a General Will makes it possible for the individuals bound by it to act because of it, its existence does not guarantee its causal efficacy. This point can be usefully put, I think, by noting two different ways that an action can be in one's self-interest. In an association that has a General Will, acting according to the law is in one's self-interest, with 'self' properly explicated by the first-person plural ("what is good for us"). Sometimes, however, acting against the law and so against the General Will is in one's self-interest, conceived of by the first-person singular ("what is good for me"). Establishing a General Will does not remove the capacity of an individual to self-conceive by the first-person singular, nor does it eliminate individual interests that are properly articulated by the first-person singular. Dropping the grammatical language, we can bluntly say that even in an association guided by the General Will, individual members will maintain their individual wills, which aim at satisfying individual desires. This fact of human psychology need not threaten the integrity of an association that has given itself a General Will: so long as every individual follows the General Will when it conflicts with her individual will, the association remains intact. Unfortunately, as Rousseau himself reminds us, "[m]en are wicked; a sad and constant experience makes proof unnecessary 16 ." Experience teaches us to expect, then, that an association guided by a General Will will contain individuals who, on occasion, knowingly act in accordance with their individual wills against the General Will.
This basic fact of human psychology has consequences for the way that deontic powers are to be divvied up in any association that can count as a democracy. The constitutive aim of any governmental agency in a democracy is to help maintain democratic unity, and it can only do this if its members act in accordance with the General Will. In order to pursue its specific aim, a governmental agency will be given a specific deontic power; for example, a regulatory agency, in order to perform its specific regulatory function, may be specially authorized to set and to collect fines for regulatory non-compliance. Now an institution acts only when the individuals who belong to that institution act in their capacity as members of that institution. In the case of governmental institutions, an institution itself can be understood as acting in accordance with the General Will if all of its members' governmental actions accord with the General Will. Should an individual, in his capacity as a member of a governmental institution, act on his individual will against the General Will, his action will thereby diminish the extent to which that institution acts in accordance with the General Will. This is what corruption amounts to on the Rousseauean model: exploiting institutional power to pursue one's individual will instead of the General Will.
The corruption here is the rupture of the unity produced by the General Will. One way to avoid this corruption, obviously, is to populate governmental institutions only with political saints, whose tendency to follow the General Will always wins out in the face of competing individual desires. If we cannot hope to populate our institutions exclusively with such individuals-and here, experience recommends against such hope-then we can alternatively deter corruption by structuring governmental institutions in a way that minimizes the temptation to exploit their deontic powers to satisfy competing individual desires. The less an institution tempts one to exploit its power for individually selfserving use, the more likely that institution's activities will accord with the General Will 17 .
Cohen discusses these institutional issues at length in his concluding chapter on democracy. A primary aim of this discussion is to describe how " […] to structure institutions so that individuals and groups are not typically in a position to design and implement policies which can reasonably be expected to yield benefits for themselves alone and to impose the bulk of them on others 18 ." Cohen's Rousseau makes a number of recommendations for achieving this structural goal. In any association where laws are made by representatives and not directly by all members, the temptation to exploit legislative office for individual gain will be diminished if representatives are elected, subject to frequent review, given specific and binding directions, and required to report all of their activities to their constituents. The thinking here is straightforward: to deter a representative from using her governmental powers to serve her individual interest, it must be possible to vote her out of office, and her actions as a representative should be reviewable by the public, lest she secretly exploit her office to pursue her individual self-interest. Similar restrictions must be placed on the executive in order to diminish the temptation to exploit its office for individual gain. It too is subject to public review, and its members may also be removed if they fail, in their capacity as executives, to abide by the General Will.
In a democratic organization, then, it must normally be possible for the people to know what its governmental officials do when they are acting in their capacity as these officials. Rousseau himself is explicitly clear about this: discussing the people as the supreme legislator of an association bound by a General Will, he remarks that "the Legislator, existing always, sees the effect of the abuse of its laws: it sees whether they are followed or transgressed, interpreted in good or bad faith; it watches over it; it ought to watch over it; that is its right, its duty, even its sworn oath 19 ." To call this a duty, even a "sworn oath," is to claim that this oversight is a deontic power that must be exercised for the association to count as a democracy. In order for the people to be capable of exercising this deontic power, members of state institutions must normally not be allowed to hide activities that involve the use of institutional power 20 .
This last point can be reformulated as a claim about transparency. As I shall be using the term, 'transparency' picks out an epistemic relation between a potential knower of some agential activity and the agent or agents who perform said activity. In the present context, the potential knowers are the members of the public, and the activity to be known is governmental activity 21 . A governmental activity is transparent to the extent that it is easy for a member of the public to gain knowledge about it. Transparency thus admits of degrees: the easier it is to know about a governmental activity, the more transparent it is. An activity is also more or less transparent as a function of the amount of information that is readily knowable about it. An activity can become more transparent as more information about it is made readily available. For example, reading The Guardian on 29 April 2009, one could have easily learned that the Obama administration was "urging" nations in Europe to accept released detainees from the U.S.'s prison at Guantánamo Bay 22 . At the time, this activity of urging was thus somewhat transparent. It was made more transparent, however, by a WikiLeaks cable release seven months later, which showed a willingness of the U.S. to pander to "small, meek Belgium" by promising them "prominence in Europe" for taking released detainees 23 . An activity can also go from being completely nontransparent to being transparent. Continuing to read the Guardian article just mentioned, one would have found then-U.S. Attorney General Eric Holder open to the possibility of former Bush officials being tried in Spain for acts of torture in Guantánamo. With the WikiLeaks cable release seven months later, one could later learn that U.S. officials were in fact working to deter this possibility by secretly pushing members of the Spanish government to drop the case 24 .

AN EQUILIBRIUM MODEL OF THE FREEDOM OF THE PRESS
If it is the "sworn oath" of democratic citizens to watch over their government, then democratic institutions must be arranged so that governmental activity is transparent. This gives rise to a pair of questions: how are the relevant institutions to be arranged? and what limits, if any, should there be on government transparency? Geoffrey Stone's work on the rights of the democratic press can be understood as addressing these questions. Stone argues that a democratic press has the right to publish classified material, although this right is not total. For example, he denies that the press has the right to publish state secrets that it obtains through bribery. Any information given to a media outlet without solicitation, however, ought in Stone's view to be publishable. Stone claims that this is a necessary counterbalance to a government's prerogative to keep its secrets, which it can best fulfill by imposing stringent regulations on what governmental employees can disclose to the public. Here is his argument for this position: If we grant the government too much power to punish the press, we risk too great a sacrifice of public deliberation; if we give the government too little power to control confidentiality "at the source," we risk too great a sacrifice of secrecy. The solution, which has stood us in good stead for more than two centuries, is to reconcile the irreconcilable values of secrecy and accountability by guaranteeing both a strong authority of the government to prohibit leaks and an expansive right of the press to publish them. This solution may seem awkward in theory and unruly in practice, but it makes perfect sense and has stood the test of time 25 .
Stone quotes Alexander Bickel, an early advocate of this equilibrium model of the freedom of the press, as describing this arrangement as a "disorderly situation 26 ." Stone thinks that in spite of, or perhaps because of, this disorder, the arrangement succeeds in balancing a government's need for secrecy against a public's right to know its activities.
Let us explicate this brief argument in terms of the Rousseauean framework developed above. Begin with Rousseau's claim that the people have not only a right but a duty, a "sworn oath," to watch over their government's activities. This duty follows from the duty of every individual to act in accordance with the General Will. A failure to monitor the government creates a context in which state institutions can tempt their members to exploit their institutional power to satisfy their individual wills, which corrupts the General Will. Now if we admit that at least some of what a government does-e.g., the plans it makes to capture a dangerous enemy in hiding-must be kept secret in order to preserve the democratic association, then the institutions that perform these specific activities must be granted the deontic power to keep the relevant information secret. This power, however, must not be so great that it is easy for the members of these institutions to exploit it to satisfy their individual wills. If the power of any such institution is easily exploitable, we can expect it to tempt individuals to use their official capacity in the institution to satisfy their individual wills, not the General Will.
Stone's solution to this problem is to create a pair of institutions that have apparently conflicting rights with respect to state secrets. One of these is a governmental institution, and its task is to keep secret all information whose secrecy is necessary for the maintenance of the social unity generated by the General Will. The institution may be populated by either some or all of the government's members; for lack of a better term, call this institution "the classifier." In a democracy, the classifier cannot take the form of a state censor whose unrestricted judgment determines what information should be classified. Any individual or board empowered to classify material in this manner would be too powerful, for the power of such an institution would be too tempting to the individual wills of its members; a society that included this sort of censor would not count as a democracy. In a democracy, then, the classifier must be governed by general laws. Given its task, this institution must be internal to the government: its function is to keep certain information within the government, and it cannot perform this function if it itself is outside of the government. This function also cannot be performed unless the laws that specify what information is to be classified are sufficiently general to keep all needed information secret. On Stone's view, in order to make sure that all needed information is in fact kept secret, these laws will need to be overly general, and in being so, they will end up legitimating the classification of information whose secrecy is not necessary for maintaining the democracy to which the classifier belongs. Without instituting a corrective, governmental activity will not be sufficiently transparent.
To correct for this over-classification of information, a second institution-call it "the press"-is established with the right to publish any secrets it obtains by legitimate means. To bring the relevant sense of 'legitimate' into view, at present we need only to contrast what Stone takes to be the limiting cases. At one end there is bribery, which is illegitimate; at the other end is any case in which the press is the passive recipient of unsolicited information, which Stone claims is legitimate. Whenever the press comes to have information in the latter manner, it has the right to publish that information. Its rights appear to conflict with those of the classifier: whereas the classifier has the right to keep secrets, the press has the right to disclose them. On Stone's view, the conflict here is merely apparent, however; if information is conveyed legitimately, the classifier has thereby lost its right to keep it secret. It may retain the right to punish any of its members who convey information to the press, but it cannot legitimately demand that the press not publish the information in its possession. If the information is conveyed by proper means from the classifier to the press, then so too is the right to publish that information conveyed; if the information is not conveyed properly, then neither is the right.
In order for the press to function in the way Stone describes, it, in contrast with the classifier, must be external to the government. To be clear, this is not to say that the press is external to the laws of the state: it is bound, as all institutions in the state are, to the laws that define the General Will. It is external to the government in that it is not bound by the rules that regulate and guide the government itself. There are at least two reasons that the press must be so external. First, if it were not, then its information would be subject to classification, and it thus could not play its de-classifying role. Second, and related, its function is to inform the sovereign, i.e., the people-by rendering governmental activity transparent, it allows them to fulfill their duty of watching over the making and implementing of laws. It is, in a sense, the eyes of democracy. Because its service to the people is watching over the making and implementing of laws, its deontic power must not be limited by the powers that constitute the government 27 .
A case can be made that, according to the equilibrium model, many of Wiki-Leaks's activities count as legitimate press activities. WikiLeaks is designed to accept unsolicited materials 28 . If a document is submitted according to this design, it will count as having been legitimately transferred. The organization is clearly external to any government. Indeed, it is arguable that it is external to any state; I will return to this point in the essay's conclusion. If we accept the equilibrium model, then it seems we should accept that WikiLeaks is protected by the expansive right that the model assigns to the press to publish the unsolicited secrets it receives 29 .

INSTITUTIONAL IRRATIONALITY
The equilibrium model's characterization of the press as necessarily external to the government is, I think, correct. The model goes wrong, however, in portraying the antagonism between the government and the press as making "perfect sense." This conflict institutionalizes a sort of practical irrationality, so the model itself may be considered as a sort of institutionalized irrationality. We can make the irrationality in question clear if we continue explicating the equilibrium model within our Rousseauean framework. For this, we will need to expand the framework by introducing the Rousseauean notion of a corporate will 30 . This will specifies the collective goal of an institution that is a part of the broader institution of a state; when it causes individuals within the unit to act towards its end, it unifies the activity of the unit into that of a single body. In a well-functioning democracy, the corporate wills of all of its institutions are instruments of the General Will. The goals of these corporate wills are instrumental goals that have as their final end the well-functioning of the social association to which they belong. Just as individual wills can run counter to the General Will, however, so too can corporate wills. When this happens, the activity of the body united by a corporate will pursues its own narrow collective goal or goals over and against the broader collective interest of the democracy to which it belongs.
The classifier and the press can be distinguished from one another as having different corporate wills. The corporate will of the classifier directs its members to maintain the secrecy of information that it has classified; the corporate will of the press directs its members to reveal information that helps the public govern itself. As already noted, in a well-functioning democracy each of these wills must be an instrumental will of the General Will, so these corporate wills function well only if a person who acts in accordance with one of them thereby acts in accordance with the General Will. On the equilibrium model, this is sometimes impossible, for sometimes acting on the corporate will of the classifier runs counter to the General Will. According to the model, some of the information that is classified by the classifier does not need to be classified and can be of benefit to the public for self-governance. When a member of the classifier classifies this information, she acts in accordance with the classifier's corporate will but against the General Will. This is a sort of practical irrationality, akin to setting an end for oneself and then adopting a means that runs counter to that end. On the equilibrium model, this sort of irrationality will sometimes occur, not due to an accident or incompetence, but due to institutional design.
For evidence that this design does not, as Stone asserts, make "perfect sense," consider the various reactions to Edward Snowden's revelations to The Guardian of some of the United States' National Security Administration's (NSA's) spying programs. Some believe that these programs infringe on the civil liberties of citizens of the United States; to them, it does not make "perfect sense" that the U.S. government has conducted these programs in secret. Others believe that the revelation of these secrets threatens the security of the United States; to them, it does not make "perfect sense" that these secrets can be published. Neither of these two camps sees the process of the programs' first being secret and then being revealed as a well-functioning democratic process; one camp thinks there never should have been such secret programs in the first place, while the other thinks that they never should have been revealed. According to the equilibrium model, however, this process just is democracy in action. Neither the pro-transparency nor the pro-secrecy camps, it seems, would agree.
If we apply the Rousseauean framework to Snowden's case, we can begin to see what is needed for this "disorderly situation" to make a little more sense. Prior to deciding to reveal the NSA's programs to The Guardian, Snowden was bound by a pair of conflicting wills. As a citizen of the United States, he was bound by its General Will; as a contractor for the NSA, he was bound by its corporate will. These wills pulled him in opposite directions. This fact alone is evidence that something had gone wrong. It is possible, of course, that what went wrong was Snowden's perception of what each will demanded; were this the case, it would be analogous correctly judging that a given action is incompatible with a given goal but then incorrectly judging that one was bound anyway to perform that action. If Snowden made no comparable failure of judgment, then the problem is that serving the General Will of the United States required violating the corporate will of the NSA, and vice versa. If a person finds herself in such a situation as a result of institutional design, it constitutes an institutional defect 31 . Now it may be that this sort of defect is an inevitable feature of a democracy. If we accept the basic elements of the Rousseauean framework, it is not clear how the possibility of this defect can be avoided. The people have a sworn duty to watch over their government, but on occasion the government will have to prevent some of its activities from becoming public knowledge. The problem lies in specifying the deontic powers of the institution that will serve as the classifier. As we have already noted, the classifier cannot be a state censor; the power of such an institution would be too tempting and thus would risk corruption. The remedy for this, it seems, must be the one Stone recommends, viz., a general set of rules that the classifier will follow. If the rules are to be sufficiently general, following them will inevitably result in the classification of information that the public would benefit from knowing; so, inevitably, there will be instances where abiding by the corporate will of the classifier will require one to act counter to the General Will. There may be no way to organize the relevant institutions within the Rousseauean framework that is immune to this possibility 32 .
Even if this is so, the framework does recommend structuring these institutions so as to minimize this possibility. It agrees with U. S. Supreme Court Justice Potter Stewart that the duty of maintaining state secrets requires " […] judgment and wisdom of the highest order […] [and] that moral, political, and practical considerations would dictate that a very first principle of that wisdom would be an insistence upon avoiding secrecy for its own sake 33 ." The equilibrium model is blind to this wisdom, for it places no internal restrictions on the extent to which the government can classify information. It is legitimate, according to the model, for the government to classify information aggressively, as long as there also exists a press that is sufficiently empowered to balance this aggressive classification. But as Cass Sunstein notes, "the notion that the government may control information at its source is at odds with the idea that the purpose of a system of free expression is to control the conduct of representatives 34 ." The Rousseauean framework agrees. In response to cases such as Snowden's, it recommends rectifying the disorder by focusing squarely on the lack of governmental transparency in an effort to bring the corporate will of the classifier in line with the General Will. It may not on its own make specific recommendations on how to align these wills, but that is no shortcoming of the framework; its role as a framework is to set standards and terms for democratic deliberation, not to resolve every deliberative matter. In cases such as Snowden's, this is not trivial, for it means that the democratic course of action-the course of action that accords with the General Will-is first to decide how to reorganize the relevant institutions to promote greater transparency and only later to worry about what punishment, if any, Snowden himself deserves.

THE PRESS
We have just used the Rousseauean framework to criticize the equilibrium model's conception of the classifier. When we apply the framework to the model's conception of the press, the result is not a critique, but rather an elaboration. As ever, we are only concerned here with the role the press may play in combating democratic corruption, which we have defined as the use of institutional power to serve a will other than the General Will. The institutionalization of a press that has the deontic power to disclose any governmental information it passively receives, even if that information is secret, aims at just this end. A press that is free to publish any governmental secret it passively receives institutionalizes a corrective to the temptation of individually self-serving secrecy.
Several notes are in order here on this freedom of the democratic press. First, as I intend the notion, it is a deontic power that constitutes a specific sort of activity, which in turn defines the institution of the press. The power is the power to serve as the eyes of democracy, that sworn oath, according to Rousseau, of every citizen. Anyone who publishes anything that is of use to the public in governing itself thereby exercises this power and engages in press activity. 'Publish' here has its etymological meaning, "to make public." On this definition, press activity does not require constructing a narrative based on primary sources, as journalistic writing does. The presentation of those materials itself counts as press activity, for such a presentation serves to render governmental activity transparent. To exercise this power and thereby perform a press activity, then, a person need not be employed by a traditional media outlet, such as a newspaper organization or a television network 35 . Should a group-e.g., Wiki-Leaks-be a non-traditional, online group that publishes primary sources, it is not thereby precluded from counting as a press organization 36 .
This freedom is instrumental to, not partially constitutive of, self-rule. A democracy governed by political saints would not require an institution with this free-dom. Given human nature, however-specifically, given the temptation to exploit the power of public office to satisfy one's individual will-the freedom is necessary. In order for humans to govern themselves, they must structure their governmental institutions to minimize this temptation, which threatens to corrupt their self-governance. The freedom of the press to publish state secrets is a necessary means to this end.
Next, the arguments here only concern the freedom to publish state secrets. These arguments do not immediately apply to private secrets. They will apply to private secrets that hide information of relevance to members of a democracy; the secret composition of toxic fluids injected into the ground during hydraulic fracturing might be such a case. They will not apply, however, to private secrets that are irrelevant for self-governance, e.g., my secret plan to utter 'Rosebud' on my 78 th birthday. I take no stand here on the press's right to publish such private secrets, should it come by them legitimately. If it has such a right, though, it is not one that is necessary for performing the function that is presently under discussion.
In speaking here of legitimate means of obtaining state secrets, I follow Stone, who follows the standard established in the U.S. Supreme Court case Bartnicki v. Vopper 37 . A secret is legitimately received if it is received passively, which in this context means without coercion, including bribery. At issue here is the quality of the received information. If someone is willing to pay someone else for a bit of information, the latter has an incentive to say what he thinks the former wants to hear, whether or not what is said is true. The press is of little use if it spreads misinformation, so it must be structured so that it can trust that the putative information it receives is accurate. Where passivity tapers off into coercion may not be clear-cut, but sorting out these details is beyond the present discussion.
The arguments here are compatible with the possibility of the press being liable for individual harm. Consider the following view, put forth as outlining sufficient conditions for such liability: if the press publishes information that leads to an individual's harm, and if the information does not concern any wrongdoing (imagine, e.g., it concerns a person's sexual orientation), and if the press had strong evidence that publishing the information would lead to the harm, it is liable, at least in part, for the harm. The arguments here are compatible with this view. To see this, imagine a bit of information that the government has made secret, that the press has discovered, that concerns a person engaged in permissible activity, and whose publication can be expected to lead to that person's harm. If the press releases the information in such a way that the expected harm results, it may be held liable for doing so. This liability poses no restriction on the press's freedom to reveal the secret, however, for it can publish the secret while concealing the identity of the relevant person, e.g., by redacting her name. If the press does this, it can perform the function described above without harming the individual mentioned in the information.
The key to the last point is a firm separation between the press's responsibility to individuals qua individuals and its right to publish governmental secrets. This leads me to my next point regarding the freedom of press articulated by these arguments, which concerns the absolute nature of this deontic power. If the role of the press is to prevent the government from consolidating power in such a way that risks corrupting self-rule, then the government cannot mark off a class of information as so secret that it is illegal, under any circumstances, for the press to publish it. The line of reasoning that leads to this conclusion parallels that of traditional liberal arguments against the criminalization of seditious libel 38 . The power to classify some information as absolutely unpublishable parallels the power to classify some political views as absolutely inarticulable: once the power exists, it institutionalizes the temptation to control the flow of information and ideas to satisfy the individual wills of the government's officers. If the role of the press is to check this power, then its capacity to publish secrets it legitimately receives must be absolute.
To some, surely, this will seem too strong. The test case here is whether during a time of war the press has the right to publish "the sailing dates of transports or the number and location of troops," which the U.S. Supreme Court has said is not a press activity worthy of legal protection in a democracy 39 . Set aside the harms that might come to the individual troops from publishing the dates; again, I do not deny that the press might be liable for such harms. The issue here, as it has been throughout this essay, concerns institutional design. What I have called the classifier has the power to make governmental activities and their results, such as a battle plan, secret. This power, if left unchecked, can be abused. In a state full of political saints, this power will not be abused, but most humans are not saints. It is the role of the press to make governmental officers act as if they were political saints. If the government has the power to designate information as beyond the right of the press to publish, then that power cannot be checked, and so it can be abused for individual gain. In a democracy, then, the government cannot have this power. To say this is not to deny that some information, such as the sailing dates of ships, ought to be kept from the public eye. Rather, as was suggested at the end of the previous section, it is to place the full responsibility of keeping this information secret on the government itself.

CONCLUSION: WIKILEAKS
I have sought here to explain one role of the press in a democracy; as it has seemed fitting to do so, I have tried to illuminate the explanation by discussing contemporary examples. Let me close by returning to the essay's central contemporary case and consider in a bit more detail whether WikiLeaks should count as a press institution, specifically in relation to the United States. Begin by focusing on the releases that made WikiLeaks notorious by the end of 2010, "The Afghan War Diary," "The Iraq War Logs," and "The Secret US Embassy Cables (Cablegate)." As noted in the introduction, Congressperson King believes that the release of these materials constitutes a sort of terrorism, presumably perpetrated against the United States and its citizens. If the arguments above are correct, then to the contrary, these releases accord with the General Will of the United States. It does not matter that WikiLeaks is an international organization, which itself is not a part of the United States. As noted above, the role of the press in democracy is instrumental, so an organization can perform this role in a state to which it does not belong. It performs this role just in case its activities function as the work of the press for that state. I have described this work as serving as the eyes of democracy: it keeps the sovereign, the people, informed of its government's activities, including secret activities. I have not argued here that a democratic people has a right to this information, though that may be true; rather, I have argued that publishing this information reveals a failure of the government to serve its people. Either the information does not need to be kept secret, in which case the government has misused its power to classify, or it should be kept secret, in which case the government has failed its responsibility to keep it so. Either way, a body that releases information that notifies a people of this governmental failure does the proper work of a democratic press for that people. Many of WikiLeaks's activities have done just this for the citizens of the United States. This is not to say that all that WikiLeaks has done, on the present account, is the proper work of the press. For example, WikiLeaks has published a 1975 manual describing the secret rituals of the Alpha Sigma Tau sorority 40 . Steven Aftergood has argued that "[t]his is not whistleblowing and it is not journalism. It is a kind of information vandalism 41 ." Whether or not this last description is apt, publishing an old sorority manual does not readily fit the account of press activity elaborated in the previous section. It is also arguable that its release of "The Afghan War Diary" fell short of the standards of proper press work. A number of international human rights groups, including The Afghan Independent Human Rights Commission and Amnesty International, criticized WikiLeaks for failing to redact the names of Afghan civilians from the diary 42 . The concern was that the inclusion of these individuals' names could threaten their safety. If part of the press's work is to avoid creating these sorts of threats, then WikiLeaks's failure to redact the names constitutes a failure to satisfy the norms that govern press activity. On the present account, however, this failure is narrowly and specifically related to these potential individual harms. As noted above, this should be kept separate from the role the press plays in monitoring and reporting on governmental activities. Though it may be at fault for failing to redact the names, WikiLeaks did not, on the present account, perform any similar wrongdoing by releasing the other contents of the diary.
Not everything that WikiLeaks has done, then, clearly counts as press activity, and it has not performed all of its press activity well. Nevertheless, much of what it has done counts on the Rousseauean model I have developed here as press activity, precisely because it serves to render governmental activities transparent. A full analysis of the legitimacy of all of WikiLeaks's activities would require a separate (and substantially longer) monograph. My hope is that the present essay has provided an adequate framework for such an analysis.