As I have noted before, there are a number of paragraphs in the May 6, 2004 Goldsmith memo authorizing warrantless wiretapping that appear to be badly overclassified. Not only were many of the same paragraphs printed, almost verbatim, in unclassified fashion, in the White Paper released in January 2006. But many of those paragraphs contain nothing more than discussions of published statute.

Now, I hope to do a follow-up to this post on whether I’m right about this overclassification. But thus far, in asking around, no one outside of government has been able to see the logic behind the classification markings on some of these paragraphs, and the people who should know were unable to explain it.

The Overclassification of the March 13, 2003 Torture Memo

Now, I’m not just talking outtamyarse about the possibility that this is overclassified; the Bush Administration has a history of improper classification. It was a particular issue with the March 14, 2003 Yoo DOD Torture Memo. Here’s how former head of Information Security Oversight Office Bill Leonard described the classification of the memo at Russ Feingold’s 2008 secret law hearing:

The March 14, 2003, memorandum on interrogation of enemy combatants was written by DoJ’s Office of Legal Counsel (OLC) to the General Counsel of the DoD. By virtue of the memorandum’s classification markings, the American people were initially denied access to it. Only after the document was declassified were my fellow citizens and I able to review it for the first time. Upon doing so, I was profoundly disappointed because this memorandum represents one of the worst abuses of the classification process that I had seen during my career, including the past five years when I had the authority to access more classified information than almost any other person in the Executive branch. The memorandum is purely a legal analysis – it is not operational in nature. Its author was quoted as describing it as “near boilerplate.”! To learn that such a document was classified had the same effect on me as waking up one morning and learning that after all these years, there is a “secret” Article to the Constitution that the American people do not even know about.

Here are Leonard’s specific complaints about the memo:

In this instance, the OLC memo did not contain the identity of the official who designated this information as classified in the first instance, even though this is a fundamental requirement of the President’s classification system. In addition, the memo contained neither declassification instructions nor a concise reason for classification, likewise basic requirements. Equally disturbing, the official who designated this memo as classified did not fulfill the clear requirement to indicate which portions are classified and which portions are unclassified, leading the reader to question whether this official truly believes a discussion of patently unclassified issues such as the President’s Commander-in-Chief authorities or a discussion of the applicability to enemy combatants of the Fifth or Eighth Amendment would cause identifiable harm to our national security. Furthermore, it is exceedingly irregular that this memorandum was declassified by DoD even though it was written, and presumably classified, by DoJ.

Mind you, the Goldsmith memo is not as bad as the March 2003 memo. As we’ll see, every single paragraph includes a classification mark (though I believe some–if not many–of those are specious). But like the March 2003 memo, this one does not describe who classified it, when it could be declassified, nor a reason for declassification. And as I explained, the people who should be able to offer an explanation (like DOJ) are unable to.

When Feingold asked about the improper classification of the March 2003 memo (see PDF 53-54), DOJ explained,

Because none of the attorneys who participated in preparing the March 2003 memorandum remains at the Department of Justice, no current DOJ employees have first-hand knowledge of the circumstances surrounding the classification of that memorandum. We have consulted the Acting General Counsel of the Department of Defense and understand from him that the memorandum was classified under the authority of DoD using that agency’s classification authority because the memorandum related to the guidance of a DoD working group charged with developing recommendations of the Secretary of Defense concerning a range of possible interrogation techniques for use with alien unlawful combatant detained at Guantanamo Bay.

In other words, they claimed the memo was classified under derivative classification of the DOD Detainee Working Group.

Derivative of the Original White House Authorizations

That background helps us at least surmise what is claimed to have happened with this memo.  It says on its front page it was,

Derived from: “Presidential Authorization for Specified Electronic Surveillance Activities During a Limited Period to Detect and Prevent Acts of Terrorism Within the United States,” dated Oct. 4, 2001, and subsequent related Presidential authorizations [at least one line redacted]

In other words, this memo tells us it was derived from the original White House authorization of October 4, 2001 (note, not the John Yoo memo authorizing the program from the same date). From that, it’s a safe bet that, given that OLC is not a classification originator (that is, Jack Goldsmith couldn’t have classified this memo without violating the sometimes-pixie-dusted EO on classification), the White House (you know, someone like Dick Cheney?) must have classified this document as the originator of the documents of which it was a derivative.

Which brings us to what I believe to be either arbitrary, or badly manipulative, determinations of which paragraphs are classified.

Let’s start with one of my favorite examples. Page 20, footnote 17 reads, in its entirety,

17 See also 50 U.S.C. 1810 (providing for civil liability as well). (TS//SI[redacted]//NF)

Someone in government–almost certainly someone in the White House–claimed in 2004 that the mere citation of one clause of the FISA legislation and the admission that its plain language meant violation of FISA called for (in addition to the criminal penalties described in the body of the text, also classified Top Secret) civil penalties was Top Secret and compartmented.

This is not even what Leonard describes as legal analysis. Rather, it is the direct citation of the public language of a statute and the interpretation even non-lawyers like me would take from that statute.

Other similar examples can be found in footnotes 15 and 16, among others.

What the White House appears to have declared Top Secret is their Administration’s acknowledgment that they knew what the plain language of FISA meant.

Gaming the System of Secrecy

Now, as I said, in my non-expert opinion, much of this memo was misclassified. For example, I’ve cataloged below something like 55 examples of paragraphs that, in this release, have no redaction in their body. That says nothing in them is now considered even classified. None of these even include stray references that there is “a program.” (Indeed, the most common reason paragraphs were classified Top Secret is because they contained a reference to the name of the program. But none of these do.)

While some of these paragraphs are simple restatement of statute or historical narrative (always excluding reference to the Vietnam war and the wiretapping during that war subsequently prohibited by Keith and FISA), most of them are what Leonard calls “analysis:” Goldsmith’s interpretation of what a statute means.

But then, so are many of the paragraphs considered unclassified. See, for example, page 22, paragraph 3.

In my rough impression, what seems to distinguish legal analysis paragraphs marked unclassified from those marked Top Secret is the potential controversy asserted with particular analysis, particularly with regards to a congressional interpretation of the same statute.

If I had to guess, I’d say they might be hiding behind a Top Secret classification anything to which Congress might readily–and with generally accepted interpretation–object. If you just hide the controversial legal analysis, you see, it makes it harder for Congress to lay out how wrong you are!

On the other hand, there are a number (by my count, at least four) references to OLC opinions that were–at the time Goldsmith wrote this memo–completely secret (including one that has never before this memo been even mentioned). To a degree, it’s odd that whoever at the White House was so wildly overclassifying the rest of this memo left these references unclassified. Though, at the same time, it would give the memo the illusion of authority (multiple OLC opinions in support) without any access to the underlying thought of the memo.

Similarly, a number of the other unclassified analytical paragraphs are those that espouse a maximal interpretation of authority.

Declaring FISA’s Exclusivity Provision Top Secret

All of which brings me to my favorite example of what appears to be overclassification: Goldsmith’s admission that FISA includes a provision that says FISA is the exclusive means to electronically surveil. Here’s the language of the paragraph that someone in the 2004 White House (presumably) considered Top Secret.

Generally speaking, FISA provides what purports to be, according to the terms of the statute, the exclusive means for intercepting the content of communications in the United States for foreign intelligence purposes. Specifically, FISA sets out a definition of “electronic surveillance”–a definition that includes any interception in the United States of the contents of a “wire communication” to or from a person in the United States–and provides specific procedures that must be followed for the government to engage in “electronic surveillance” as thus defined for foreign intelligence purposes. As a general matter, for electronic surveillance to be conducted, FISA requires that the Attorney General or Deputy Attorney General approve an application for an order that must be submitted to a special Article III court created by FISA–the Foreign Intelligence Surveillance Court (FISC). See 50 USC 1804 (2000 & Supp I 2001). The application for an order must demonstrate, among other things, that there is probable cause to believe that the target is a foreign power or an agent of a foreign power. See id 1805(a)(3)(A). It must also contain a certification from the Assistant to the President or an officer of the United States appointed by the President with the advice and consent of the Senate and having responsibilities in the area of national security or defense that the information sought is foreign intelligence information (as defined by FISA), that cannot reasonably be obtained by normal investigative means. See id 1804(a)(7). FISA further requires details about the methods that will be used to obtain the information and the particular facilities that will be the subject of the interception. See id. 1804(a)(4), (a)(8).

Now, except for Goldsmith’s use of the word “purports,” the rest of the paragraph is a plain reading of the statute.

It is, in fact, a perfect statement of what the Bush Administration recognized the law on its face to mean, before all the wrangling to try to make the exclusivity paragraph disappear behind a Top Secret designation.

And that paragraph–not so much analysis as basic reading comprehension of what FISA required–the White House claimed could not be released because doing so would result in “exceptionally grave damage to the national security.”

Remember, the entire point of this memo was to find some way to get the Executive Branch out of complying with the plain language of the FISA statute.

And so, someone in the White House (presumably) declared the evidence that they knew exactly what the statute said to be so secret that it would cause grave damage to the national security.


The following inventory tracks paragraphs that appear completely unredacted in the recently released memo (suggesting none of the information in the paragraph is currently classified). For paragraphs that span two pages, I treat them on their ending page.

Top Secret paragraphs that probably should be unclassified

Page 19, footnote 15: a verbatim definition of FISA’s definition of “electronic surveillance.”

Page 20, paragraph 1: discussion of exclusivity provision and the usual manner for a FISA application.

Page 20, paragraph 2: description of criminal penalties for violating FISA as well as interlocking penalties in Title III.

Page 20, footnote 16: acknowledgment that DAG also has authority to approve FISA orders.

Page 20, footnote 17: notice that 50 USC 1810 provides for civil liability.

Page 24, paragraph 1: The admission that Egan held that constitutional avoidance should not be used to “rewrite language enacted by legislature,” continuing, “If Congress has made it clear that it intends FISA to provide a comprehensive restraint on the Executive’s ability to conduct foreign intelligence surveillance, then the question whether FISA’s constraints are unconstitutional cannot be avoided.”

Page 30, paragraph 1: a history of the use of SIGINT referring to published histories (but not referring to any example from Vietnam).

Page 30, footnote 23: an explanation that President Wilson’s EO on wartime censorship was extended to points near the Mexican border through which censorship might be avoided.

Page 32, footnote 24: an admission that Ruckelshaus admits repeals are disfavored, but insisting that “the ordinary restrictions in FISA cannot continue to apply if the Congressional Authorization is appropriately construed to have its full effect.”

Page 33, paragraph 1: Discussion of the “deter acts of international terrorism against the US” language of AUMF.

Page 33, paragraph 2: Admission that “if the Congressional Authorization actually had applied so broadly, the specific amendments to FISA that Congress passed a few weeks later in the PATRIOT Act would have been superfluous.”

Page 33, paragraph 3: Claim that PATRIOT changes were not superfluous because they had broader impact.

Page 34, paragraph 2: Claim that the 72-hour grace period for FISA orders was necessary because “there was bound to be a substantial increase in the volume of surveillance conducted under FISA, which would strain existing resources.”

Page 35, paragraph 1: Reassertion that AUMF “can thus be read” as specific authority that overrode FISA.

Page 35, paragraph 2: Claim that AUMF creates enough of an ambiguity to apply constitutional avoidance.

Page 38, paragraph 1: Assertion that even in peacetime POTUS has inherent authority to conduct warrantless searches for foreign intelligence.

Page 38, paragraph 2: Concession that when collecting foreign intelligence within the US, POTUS must comply with Fourth Amendment.

Page 39, paragraph 3: Assertion that foreign intelligence collection is another case of “special needs” because of greater flexibility and secrecy required.

Page 39, paragraph 4: Claim that “every federal court” has concluded that even in peacetime POTUS has inherent authority to conduct searches for foreign intelligence purposes w/o a warrant.

Page 40, paragraph 1: “To be sure, the Supreme Court has left this precise question open.”

Page 40, paragraph 2: Note that the cited federal cases two paragraphs above were decided after Keith.

Page 41, paragraph 2: Assertion that in 1970s, SCOTUS had barely begun to develop “special needs.”

Page 41, footnote 32: Dismissal of any tie between Truong and Vietnam War.

Page 41, footnote 33: Attribution of term “special needs” to Blackmun in TLO.

Page 42, paragraph 1: Recitation of executive practice treating wiretapping as an inherent authority.

Page 43, paragraph 1: Reiteration that these examples are all peacetime, with another reference to wartime surveillance ignoring Vietnam.

Page 44, paragraph 1: Assertion there are few precedents for this kind of conflict between Congress and executive authority.

Page 46, paragraph 1: Whether POTUS’ power to conduct foreign searches within the US is inherent presents difficult question.

Page 46, footnote 36, paragraph 1: Claim that power to make rules for regulation of naval and land forces and necessary and proper clause less likely source of Congressional authority to regulate foreign surveillance in the US.

Page 47, paragraph 1: Evidence that legislative history of FISA allows for doubt on constitutionality, including references to Laurence Silberman.

Page 47, footnote 37: Assertion that in spite of In Re Sealed Case’s [Goldsmith doesn't refer to it as such] dismissal of Silberman’s complaints about non-adversary process, it did not address inherent power.

Page 48, paragraph 1: Claim that Teddy Kennedy and Roman Hruska’s concerns about getting future Presidents to comply with FISA reflect “trepidation” about whether future Presidents would comply.

Page 48, paragraph 2: Discussion of objection from four members of HPSCI in 1978 and dissenters to the conference report about giving Article III Courts power over foreign affairs.

Page 49, paragraph 1: Citation of FISCR opinion accepting pre-FISA case law finding POTUS has inherent authority to conduct warrantless searches to obtain foreign intelligence.

Page 52, paragraph 2: Claims OLC has “long concluded” that the Commander-in-Chief is a substantive grant to POTUS (though citing only Rehnquist’s 1970 Cambodian Sanctuaries memo); then listing citing 19th century SCOTUS cases (including Milligan) to support his case.

Page 53, paragraph 1: Argument that President’s authority is at its height when responding to attack on US, citing Swift Justice opinion (interpreting a Leahy attempt to put limits on military commissions as illegal; the memo was among those withdrawn on January 15, 2009 and released on March 2, 2009), and the Rehnquist Cambodia memo.

Page 56, paragraph 1: This is, in my NAL opinion, weird. It admits there are times when the Executive says Congress could control its actions. Again, NAL, but I think In Re Sealed Case is as much an admission as anything cited here–and it basically accepted that Congress’ FISA regime was legitimate.

Page 57, paragraph 1: Discussion of times when, after the fact, Congress has sanctioned the President’s actions.

Page 57, paragraph 1: Discussion of times when POTUS has asked for sanction after the fact.

Page 58, paragraph 1: Discussion of Barreme; I discuss this briefly here.

Page 59, paragraph 1: Discussion of Barreme as restriction on commerce, not war.

Page 59, paragraph 2: Claim that contemporary cases suggest SCOTUS simply found Barreme to be too limited a war to grant inherent authority.

Page 59, paragraph 3: Elaboration of claim that Justice Marshall was just not sure the Quasi War was war enough to grant POTUS inherent powers.

Page 60, paragraph 1: More discussion of what contemporary decisions on the Quasi War mean about Barreme.

Page 60, paragraph 2: Conclusion that SCOTUS therefore did not considere the Quasi Wars a sufficiently full-scale war to trigger CinC powers.

Page 61, paragraph 1: Claim that this distinguishes Barreme from GWOT.

Page 61, paragraph 2: Assertion that Youngstown parallel is inapt.

Page 61, paragraph 3: General description of Youngstown, including note that Congress had chosen not to give POTUS power to seize industry during an emergency (as Congress had chosen not to give the President power to operate in the US in the Afghan AUMF).

Page 62, paragraph 1: Assertion that SCOTUS found the tie between POTUS’ CinC powers and his actions “was simply too attenuated.”

Page 62, paragraph 2: Assertion that Youngstown was extension of POTUS power into Commerce Clause.

Page 63, paragraph 2: Claim that Youngstown was POTUS bootstrapping power over commerce.

Page 64, paragraph 1: Claim that GWOT involved infiltration more than Korean war did (again, no mention of Vietnam).

Page 74, paragraph 1: This paragraph, appearing in the midst of completely redacted last section on why FISA is unconstitutional as applied, is worth citing in full:

To summarize, we conclude only that when the Nation has been thrust into an armed conflict by a foreign attack on the United States and the President determines in his role as Commander in Chief and sole organ for the Nation in foreign affairs that it is essential for defense against a further attack to use the signals intelligence capabilities of the Department of Defense within the United States, he has inherent constitutional authority to direct electronic surveillance without a warrant to intercept the suspected communications of the enemy–an authority that Congress cannot curtail. We need not, and do not, express any view on whether the restrictions impose in FISA are a constitutional exercise of congressional power in circumstances of more routine foreign intelligence gather that do not implicate an armed conflict and direct efforts to safeguard the Nation from a credible danger of foreign attack.

Page 101, last paragraph: As counterclaim to Katz’ privacy interests, an assertion that national security interests are the most important aspects of government.

Page 105, footnote 86: In a discussion of reasonableness, discussion of SCOTUS’ use of efficacy as a measure against “random and standardless searches.”

Unclassified paragraphs

Page 22, paragraph 3: Discussion of constitutional avoidance.

Page 23, paragraph 1: further discussion of constitutional avoidance (including information that “This Office has always adhered to the rule of construction described above…”

Page 23, footnote 19: Examples of OLC’s interpretation of constitutional avoidance, including a citation of July 22, 2002 Jay Bybee memo that has never been revealed, much less released.

Page 29, paragraph 2: A description of the AUMF authorizing “all necessary and appropriate force.”

Page 31, footnote 34: A description off how in the context of detention, OLC has also taken an expansive view of the AUMF, citing a June 27, 2002 John Yoo memo that was first released on March 2, 2009.

Page 34, paragraph 1: Recitation of earlier discussions of changing “the purpose” to “a significant purpose” (though not mentioning the opinion to David Kris written during the PATRIOT Act debate).

Page 38, paragraph 3: Review of Fourth Amendment, including note that probable cause not universal.

Page 38, footnote 29: Note that Fourth Amendment does not protect aliens outside the US.

Page 39, paragraph 1: Discussion of “special needs.”

Page 39, paragraph 2: Discussion of other permissible warrantless searches.

Page 45, paragraph 2: Assertion, “It is settled beyond dispute that … certain presidential authorities in [foreign affairs] are wholly beyond the power of Congress to interfere with by legislation.”

Page 46, footnote 36, paragraph 2: Discussion of previous OLC opinions interpreting Congress’ authority to regulate land and naval forces, including reference to March 13, 2002 Jay Bybee opinion first released on March 2, 2009.

Page 46, footnote 36, paragraph 3: Discussion providing narrow interpretation of Necessary and Proper.

Page 53, footnote 40: A discussion of a Union Army official’s declaration that troop movement decisions were exclusive to the commander in chief.

Page 56, footnote 42: A discussion of Thomas Jefferson’s orders to the naval forces fighting the Barbary Pirates.

Page 58, footnote 43: A discussion of statute in question under Barreme.

Page 58, footnote 44: Discussion of Talbot v. Seeman claiming that SCOTUS would have ruled that POTUS had authority in the Quasi War in absence of laws authorizing it.

Page 62, footnote 45: As part of discussion why GWOT different, details on when Ali al-Marri came to the US.

Page 63, footnote 46: Admission that circuit decision in Padilla held POTUS lacked inherent CinC authority to hold citizens on American soil.

Page 100, footnote 84: Reference to John Yoo’s Fourth Amendment Eviscerating opinion, first released on March 2, 2009.

Page 108, last paragraph: “Please let me know if we can be of further assistance.”