National Security Response 127 Harv. L. Rev. F. 86

Whose Secrets?

Responding to David Pozen, The Leaky Leviathan: Why the Government Condemns and Condones Unlawful Disclosures of Information, 127 Harv. L. Rev. 512 (2013)

Response To:


Professor David Pozen has written an absolutely stellar account of the uses, meanings, and regulations of leaks of secret information by government officials. At the most general level, Pozen argues that the United States’s apparently chaotic system of leak regulation — draconian penalties that are almost never imposed; threats to punish leakers but not publishers — masks a deeper logic. This system, “although ritualistically condemned by those in power, has served a wide variety of governmental ends at the same time as it has efficiently kept most disclosures within tolerable bounds.”1 Pozen’s nuanced and subtle account of what those governmental ends are and how they are served by the existing regulatory regime is both wholly persuasive and deeply insightful.

The word “governmental,” however, covers quite a bit of ground and might usefully be deconstructed. As Pozen’s account makes clear, the broad ends that are served by the leak regime are most decidedly those of the President and other executive branch officials. Indeed, Pozen notes the pervasive sense among such officials that they “own” the secrets that they produce and control.2 Various executive branch officials collect and develop that information, classify that information as secret, and then, as Pozen so deftly demonstrates, selectively leak the “secret” information that furthers their agendas. This bare-bones summary of Pozen’s positive account suffices to make one key point pellucidly clear: “secret” is a political category, not a natural one. Facts in isolation do not cry out for secrecy; facts within a specific political context do.3

If this observation at first seems too obvious to justify the expenditure of pixels, consider that it is emphatically not the way that we tend to talk about government secrets. The fact that someone in the executive branch has chosen to mark some piece of information as “secret” sets the presumptive terms for how that information may properly be used, shared, and disclosed. As Pozen notes, “virtually any deliberate leak of classified information to an unauthorized recipient is likely to fall within the reach of one or more criminal statutes,”4 and “[n]o court has ever accepted a defense of improper classification.”5 In other words, executive branch officials determine what information is secret, a determination to which other political actors are expected to (and do) defer. Executive branch officials frequently leak “secret” information without adverse consequence, because a permissive regime for such leaks broadly serves executive branch purposes. But when information is leaked outside of the normal, executive-friendly parameters of this regime, the executive reserves the right to prosecute, with the possibility of severe punishment. Secrets are treated as belonging to the executive; little wonder, then, that “there are substantially more unattributed disclosures ‘for’ the President than ‘against’ him.”6

One can certainly understand why Presidents and other high-ranking executive branch officials find this regime generally appealing. But what’s in it for the rest of us? Unless we can think of some good reason to believe that, when it comes to secret-keeping and secret-spilling, what’s good for the President is good for America, we should be thinking about ways of reframing this secrecy regime. As Pozen suggests, it does produce some public goods,7 and we should be careful not to sacrifice them. At the same time, if it tends to produce public benefits largely when they coincide with benefits to the executive, then those benefits are destined to remain partial (in both senses of the word). By way of correction, I propose that we take more seriously the idea that these are our collective secrets as a political community, rather than the executive branch’s secrets.

Of course, the fact that they are, in some sense, “our” secrets does not mean that we can all know them. That would undermine the very concept of secrecy. But it turns out that there is another institution, chosen by the people and responsive to them, tasked with making regular decisions on their behalf. It is time that we began thinking seriously about the role of Congress in our secrecy regime, a role that is given constitutional sanction by the Speech or Debate Clause, which allows members to speak freely, without fear of punishment by the executive or the courts.8

The shining example of this use of congressional power, which I have discussed in some detail elsewhere,9 is Mike Gravel, the first-term Senator from Alaska who, in 1971, used his exalted perch as chair of the Buildings and Grounds Subcommittee of the Environment and Public Works Committee to hold a late-night hearing at which he released over four thousand pages of the Pentagon Papers into the public record. Senator Gravel’s act has been largely eclipsed by the Supreme Court’s ruling the next day in New York Times Co. v. United States10 (The Pentagon Papers Case), which lifted an injunction on publishing the Papers in the press. But had the timing been different, or the Court majority been more hospitable to the Nixon Administration’s claims (there were, after all, three dissenters in the Pentagon Papers Case11), we might well remember Senator Gravel as the hero in that episode. Indeed, the Court’s ruling left open the possibility of post-publication criminal prosecution of press outlets,12 and Senator Gravel, concluding that fear of such prosecution led to timidity on the part of the newspapers, arranged for a version of the Papers to be published in book form.13 We may remember the Pentagon Papers incident as a story about the heroism of the courts and the press, but Senator Gravel released more of the Papers, and he did it earlier.14

Nor is Senator Gravel the only member to have made use of his Speech or Debate Clause protections to release theretofore secret information. Professor Kathleen Clark has described a number of such disclosures in the national security context,15 including everything from Representative Michael Harrington’s 1974 leak of information about American efforts to overthrow Chilean President Salvador Allende,16 to releases by the Pike and Church Committees over the objections of the Ford Administration,17 to Representative Robert Torricelli’s 1995 release of information regarding the role of the CIA in murders in Guatemala.18 Indeed, in the case of the Harrington release, the House Ethics Committee chose not to impose any internal discipline, “finding that the information had not been properly classified”19 — a conclusion that, as noted above, no court has ever reached. Outside of the national security context, Representative Darrell Issa in 2012 used his Speech or Debate Clause immunity to enter into the Congressional Record a letter that disclosed the contents of a sealed wiretap application as part of the congressional investigation into the “Operation Fast and Furious” gun-smuggling operation.20

In the national security context, the most recent additions to this list are Senators Ron Wyden and Mark Udall, both of whom are members of the Intelligence Committee. During the 2011 debates over reauthorizing portions of the PATRIOT Act, they announced on the Senate floor that the Obama Administration had adopted a secret, implausible interpretation of portions of the Act dealing with domestic surveillance, although they did not release the details of this secret interpretation.21 In doing so, they took what Pozen has termed “deep secrets” — that is, secrets that we did not even know existed — and made them shallower.22 One might think, however, that they should have gone further, that they should have made public the content of those decisions as well.23 And when it came to light that the NSA was engaged in a massive program of domestic surveillance, Professor Bruce Ackerman urged the Senators to reveal information they had learned in closed briefings, in the interest of having an informed debate on legislative responses to the surveillance program.24

As the examples adduced so far demonstrate, members of Congress have a range of options available to them when dealing with “secret” information. They can simply accede to executive branch demands for secrecy; they can make the secret shallower by disclosing its existence but not its content; and they can release varying amounts of the actual content. Which of these is appropriate will, of course, be highly situation dependent. But given the extent to which executive branch secrecy determinations are made to advance executive branch interests, there is no reason for Congress to offer automatic deference to those determinations.

Of course, there are rules about such things — the Speech or Debate Clause prevents punishment of members in any other place, but the Constitution explicitly reserves to each house the right to discipline its members.25 The rules of the Senate Intelligence Committee create an elaborate procedure26 for disclosing classified information, requiring a committee vote, notification of the Majority and Minority Leaders, and then notification of the President. If the President objects, then disclosure requires a vote of the full Senate in closed session. Any committee member can set the procedure in motion by requesting a vote, and the rules require the committee to hold a vote within five days when such a request has been made. If the committee votes to disclose and the President objects, either the Majority and Minority Leaders jointly or a majority of the Committee suffices to bring the matter to the Senate floor in closed session; once the matter is brought to the floor, debate is governed by a tight timeline, and the issue cannot be filibustered. Release of information outside of these procedures must be investigated by the Ethics Committee; if it determines that there has been a significant breach of confidentiality, then it must recommend a punishment to the full Senate. The House has similar procedures for its Intelligence Committee.27

It is not clear whether either Senator Wyden or Senator Udall has attempted to invoke these procedures — as Michael Stern has noted, it is possible that the Committee interprets its own rules so as to prevent disclosure of an attempt to invoke the disclosure procedures.28 If that is the case, it is deeply unfortunate — the secrecy should, at the very least, be made shallower by disclosing to the public that there is ongoing debate about releasing information. This would alert the public that there is something of consequence being withheld, and members would have to face questions about how they voted on the disclosure issue and why. And if Senators Wyden and Udall have not attempted to invoke the disclosure procedures, then an explanation should be demanded of them. They obviously believe that disturbing information is being withheld, and they obviously are frustrated — a frustration that appears to be shared by members of both parties and both houses — by what they see as a pattern of lies from the executive branch.29 Where is the Gravelian spirit?

Of course, there will be the usual litany of objections. At some point, truly damaging information will be disclosed, either by a vindictive chamber or by a member releasing information on his own. Perhaps. But it is not at all clear why we would think that Congress is especially prone to this. Senator Gravel, after all, released only about 4100 of the Pentagon Papers’ 7800 total pages, and he excised various names from the pages he did release.30 By contrast, the two major indiscriminate disclosures in recent history — those by Chelsea Manning and Edward Snowden — have come from people with executive branch ties. Indeed, it is at least possible that, if members of Congress were more willing to evaluate and consider releasing classified information, future Mannings and Snowdens might be willing to leak to them rather than engaging in indiscriminate public releases.31 In that case, at least the final decision on releasing information to the public would be made by a democratically accountable official, who would be likely to exercise at least some measure of care — as Senator Gravel did — to avoid releasing especially damaging information.

More importantly, perhaps, the question of what disclosures count as damaging depends, as noted above, on a politically constructed understanding of our national security interests. A regime that, as Pozen documents, tolerates and even encourages widespread leaking by executive branch officials to advance their agendas, but tells terrifying tales about the consequences of disclosures by members of Congress pursuing their agendas, will inevitably give the executive branch an outsized role in constructing those national security interests. I, for one, can think of no a priori reason to prefer the national security visions of the Scooter Libbys and James Clappers of the world to those of the Ron Wydens and Mark Udalls.

The skeptics will, however, suggest that the real threat is executive backlash. If Congress discloses information that the executive does not want disclosed, then the executive will simply stop giving Congress information. Again, perhaps. But Congress has ways of making the executive talk, ranging from contempt citations32 to funding cut-offs33 to refusing to pass legislation that the executive branch wants.34 The point is not that Congress should release everything (it shouldn’t!), nor that it should always get its way over the President. The point is simply that secrecy determinations are matters for interbranch politics like any other. We do ourselves a serious disservice when we allow the executive branch to claim exclusive ownership over this part of the political commons.

* Professor of Law, Cornell Law School. Thanks to Dave Pozen, Aziz Rana, Catherine Roach, and Mariah Zeisberg for helpful and thought-provoking comments on earlier drafts. Any remaining errors or infelicities are, of course, my own.

1. David E. Pozen, The Leaky Leviathan: Why the Government Condemns and Condones Unlawful Disclosures of Information, 127 Harv. L. Rev. 512, 517 (2013).

2. Id. at 603 & n.410.

3. In this regard, it is helpful to think in terms of what Professor Mariah Zeisberg has called “security orders” — political constructs that “specif[y] and enact[] the content of a national security interest; what will be construed as threats to that interest; and a supportive construction of constitutional war authority.” Mariah Zeisberg, War Powers: The Politics of Constitutional Authority 95 (2013). To the extent that a certain piece of information “must be” secret for national security reasons, that imperative arises only in the context of, and with respect to the terms of, a particular security order. See also Aziz Rana, Who Decides on Security?, 44 Conn. L. Rev. 1417, 1451—90 (2012) (tracing the twentieth-century rise of a security order emphasizing executive expertise).

4. Pozen, supra note 1, at 524—25.

5. Id. at 523.

6. Id. at 570.

7. These are discussed primarily in Part IV of the piece. See id. at 605—33.

8. U.S. Const. art. I, § 6, cl. 1 (“[F]or any Speech or Debate in either House, [Senators and Representatives] shall not be questioned in any other Place.”). For more on the clause, including its history, see generally Josh Chafetz, Democracy’s Privileged Few: Legislative Privilege and Democratic Norms in the British and American Constitutions 87—110 (2007).

9. See Josh Chafetz, Congress’s Constitution, 160 U. Pa. L. Rev. 715, 745—50 (2012).

10. 403 U.S. 713 (1971) (per curiam).

11. Id. at 748 (Burger, C.J., dissenting); id. at 752 (Harlan, J., dissenting); id. at 759 (Blackmun, J., dissenting).

12. Justices Stewart and White explicitly held open the possibility of post-publication criminal punishment. See id. at 730 (Stewart, J., concurring); id. at 733, 740 (White, J., concurring). Combined with the three dissenters, they made a majority that was at least open to punishing the leak.

13. See Chafetz, supra note 9, at 747—49. For the comparative scope of the different published versions of the Papers, see id. at 745 n.160.

14. This makes the Court’s subsequent holding that Gravel’s decision to arrange for private publication of the Papers was outside the scope of the Speech or Debate Clause all the more unfortunate. Gravel v. United States, 408 U.S. 606, 622—27 (1972). I’ve criticized that decision elsewhere. See Chafetz, supra note 8, at 99—100; Chafetz, supra note 9, at 748—49; see also Sam J. Ervin, Jr., The Gravel and Brewster Cases: An Assault on Congressional Independence, 59 Va. L. Rev. 175, 184—88 (1973).

15. Kathleen Clark, Congress’s Right to Counsel in Intelligence Oversight, 2011 U. Ill. L. Rev. 915, 940—46.

16. Id. at 942.

17. Id. at 943.

18. Id. at 945—46.

19. Id. at 943.

20. See Jonathan Strong & Amanda Becker, Moving on From Contempt Vote, or Not, Roll Call, July 12, 2012, available at 2012 WLNR 14553858.

21. See Chafetz, supra note 9, at 750—51.

22. See generally David E. Pozen, Deep Secrecy, 62 Stan. L. Rev. 257 (2010).

23. See Trevor W. Morrison, Constitutional Alarmism, 124 Harv. L. Rev. 1688, 1724—30 (2011) (book review) (arguing that the work of the Office of Legal Counsel should be disclosed); Sudha Setty, No More Secret Laws: How Transparency of Executive Branch Legal Policy Doesn’t Let the Terrorists Win, 57 U. Kan. L. Rev. 579, 601—29 (2009) (same).

24. Bruce Ackerman, Breach or Debate, Foreign Pol’y (Aug. 1, 2013),

25. U.S. Const. art. I, § 5, cl. 2. On the houses’ disciplinary power generally, see Chafetz, supra note 8, at 207—35. See also Chafetz, supra note 9, at 754 (noting that the Speech or Debate Clause “must be read in pari materia” with each house’s disciplinary authority).

26. This procedure is laid out in S. Res. 400, 94th Cong. § 8 (1976) (last amended 2007), reprinted as amended in S. Select Comm. on Intelligence, 112th Cong., Rules of Procedure, 10, 14—16 (Comm. Print 2011); see also S. Select Comm. on Intelligence, 112th Cong., supra, r. 9.7—.8, at 6 (incorporating the procedures of S. Res. 400, as amended, into the standing rules of the Committee).

27. See Rules of the House of Representatives, r. X, cl. 11(f)—(g), reprinted in H.R. Doc. No. 112-161, at 535-40 (2013).

28. See Michael Stern, Congressional Release of Classified Information and the Speech or Debate Clause, Point of Order (Aug. 6, 2013, 9:27 AM),

29. See Jeremy Herb, Intelligence Chief Clapper Apologizes for ‘Erroneous’ Statement to Congress, The Hill, July 3, 2013, at 3; Ron Wyden, Mark Udall & Martin Heinrich, End the N.S.A. Dragnet, Now, N.Y. Times, Nov. 26, 2013, at A25.

30. See Chafetz, supra note 9, at 751.

31. Indeed, Daniel Ellsberg initially approached Senators Fulbright, McGovern, and Mathias with the Pentagon Papers, hoping that they would hold hearings on them. Although those Senators were initially supportive, their eventual “cold feet” led Ellsberg to approach the press. He subsequently learned that Senator Gravel shared his anti-war leanings, and approached him. Ellsberg insists that, from the beginning, he preferred legislative hearings to release to the newspapers. See Daniel Ellsberg, Foreword to Mike Gravel & Joe Lauria, A Political Odyssey 9, 9—10 (2008).

32. See generally Josh Chafetz, Executive Branch Contempt of Congress, 76 U. Chi. L. Rev. 1083 (2009).

33. See generally Chafetz, supra note 9, at 725—35.

34. See Ackerman, supra note 24 (“If administration officials refuse to testify at secret sessions in the future, they will most likely alienate the fence-sitters [on a vote to shut down the NSA wiretapping program]. A boycott would also doom the renewal of the Patriot Act when it comes up for reconsideration in 2015.”).