CENTRAL INTELLIGENCE AGENCY AND WILLIAM J. CASEY, DIRECTOR OF CENTRAL INTELLIGENCE, PETITIONERS V. JOHN CARY SIMS AND SIDNEY M. WOLFE No. 83-1075 In the Supreme Court of the United States October Term, 1983 On Writ of Certiorari to the United States Court of Appeals for the District of Columbia Circuit Brief for the Petitioners TABLE OF CONTENTS Opinions below Jurisdiction Statutory provisions involved Statement Summary of argument Argument: The Freedom of Information Act does not require the Central Intelligence Agency to disclose the identities of sources of intelligence-related information A. The term "intelligence sources" should be given its plain meaning B. The court of appeals' definition of "intelligence sources" produces results that Congress could not have intended Conclusion Appendix OPINIONS BELOW The opinion of the court of appeals (Pet. App. 1a-16a) is reported at 709 F.2d 95. The opinions and order of the district court (Pet. App. 21a-34a) are unreported. An earlier opinion of the court of appeals (Pet. App. 35a-64a) is reported at 642 F.2d 562. One of the earlier opinions of the district court (Pet. App. 73a-93a) is reported at 479 F. Supp. 84; the other earlier opinions and orders of the district court (Pet. App. 66a-72a, 94a-97a) are unreported. JURISDICTION The judgment of the court of appeals was entered on June 10, 1983 (Pet. App. 19a-20a). A petition for rehearing was denied on August 17, 1983 (Pet. App. 17a). On November 9, 1983, the Chief Justice extended the time in which to file a petition for a writ of certiorari to December 15, 1983, and on December 5, 1983, the Chief Justice further extended the time in which to file a petition for a writ of certiorari to December 29, 1983. The petition was filed on that date and was granted on March 5, 1984. The jurisdiction of this Court is invoked under 28 U.S.C. 1254(1). STATUTORY PROVISIONS INVOLVED Portions of the Freedom of Information Act, 5 U.S.C. 552, the National Security Act of 1947, 50 U.S.C. 401 et seq., and the Central Intelligence Agency Act of 1949, 50 U.S.C. 403a et seq., are reproduced at App., infra, 1a-3a. QUESTION PRESENTED Whether the Central Intelligence Agency must disclose the identity of a source of intelligence information under the Freedom of Information Act whenever it cannot demonstrate that it had to guarantee confidentiality in order to obtain the kind of information that the source supplied, even though Section 102(d)(3) of the National Security Act of 1947, as incorporated in Exemption 3 of the FOIA, exempts "intelligence sources" from disclosure. STATEMENT 1. Respondents, invoking the Freedom of Information Act (FOIA), 5 U.S.C. 552, filed a request with the Central Intelligence Agency seeking certain information about a CIA project known as MKULTRA. MKULTRA was initiated in the 1950's in "response to possible use by the Soviets and the Chinese of chemical and biological agents as instruments of interrogation and brainwashing" (Pet. App. 37a (footnote omitted); see id. at 73a). The project involved "research into 'chemical, biological and radiological materials capable of employment in clandestine operations to control human behavior'" (id. at 21a (footnote and citation omitted)). See also Pet. App. 89a (affidavit of Director of Central Intelligence Turner). MKULTRA research was conducted by a large number of private scientists, in the United States and abroad, affiliated with universities, research foundations, and similar institutions (Pet. App. 66a, 89a). At least 80 institutions and 185 private researchers participated (id. at 36a). A total of 149 subprojects were funded by the CIA (I C.A. App. 14-61). /1/ Most of these subprojects involved legitimate research into a variety of chemical, biological, psychological, and sociological subjects. In a few of the subprojects, researchers surreptitiously administered drugs to unwitting subjects. This wholly improper conduct is now expressly forbidden by executive order. Exec. Order No. 12,333, Section 2.10, 46 Fed. Reg. 59941, 59952 (1981). See also Project MKULTRA, the CIA's Program of Research in Behavioral Modification: Joint Hearing Before the Select Comm. on Intelligence and the Subcomm. on Health and Scientific Research of the Senate Comm. on Human Resources, 95th Cong., 1st Sess. 16, 17, 35 (1977) (Testimony of Director of Central Intelligence Turner) (hereinafter cited as Project MKULTRA Hearing). The CIA has attempted to notify the persons who were unwittingly subjected to tests. See id. at 36. The MKULTRA project was examined in 1963 in a report from the Inspector General of the CIA to the Director of Central Intelligence (II C.A. App. 118-145). Subsequently, congressional committees and a Presidential commission thoroughly examined the project, taking extensive testimony from both the Director of Central Intelligence and the former CIA personnel who had supervised the project. These inquiries gave "broad publicity" (Pet. App. 37a) to MKULTRA and the allegations of abuses connected with it. See S. Rep. 94-755, 94th Cong., 2d Sess., Bk. I, at 389-392 (1976); Report to the President by the Commission on CIA Activities Within the United States 226 (June 1975); Human Drug Testing by the CIA, 1977: Hearings on S. 1893 Before the Subcomm. on Health and Scientific Research of the Senate Comm. on Human Resources, 95th Cong., 1st Sess. (1977); Project MKULTRA Hearing, supra. 2. Respondents sought the grant proposals and contracts awarded under the MKULTRA program and the names of the institutions and individuals that performed research. The CIA made available to respondents all of the grant proposals and contracts. The CIA also contacted the institutions that had performed research, and approximately two-thirds of them voluntarily disclosed their identities to the public. The Agency accordingly disclosed them to respondents. Pet. App. 39a, 73a-74a. The CIA cited Exemptions 3 and 6 of the FOIA, 5 U.S.C. 552(b)(3) and (6), as the bases for not releasing the names of the other institutions and the individual researchers. /2/ Only the Exemption 3 claim is now at issue. Exemption 3 of the FOIA provides that an agency need not disclose "matters that are * * * specifically exempted from disclosure by statute * * * provided that such statute * * * refers to particular types of matters to be withheld" (5 U.S.C. 552(b)(3)(B)). The statute on which the CIA relied is Section 102(d)(3) of the National Security Act of 1947, 50 U.S.C. 403(d)(3). Section 403(d)(3) provides in part: (T)he Director of Central Intelligence shall be responsible for protecting intelligence sources and methods from unauthorized disclosure * * *. 3. Respondents then brought this action under the FOIA in the United States District Court for the District of Columbia. See 5 U.S.C. 552(a)(4)(B). They sought to compel disclosure of the names withheld by the CIA. The district court ordered disclosure of these names, apparently reasoning that the term "intelligence sources" in 50 U.S.C. 403(d)(3) did not include the MKULTRA researchers, of if it did, Section 403(d)(3) was not specific enough to qualify as an Exemption 3 statute (see Pet. App. 77a-79a). The court of appeals vacated the district court's order and remanded for reconsideration (Pet. App. 35a-64a). The court observed that it had consistently held that Section 403(d)(3) "qualifies as a withholding statute under Exemption 3" (id. at 44a). The court also noted that in its numerous previous decisions dealing with Exemption 3 and Section 403(d)(3), it had assumed that the phrase "intelligence sources and methods" has "a plain meaning" (ibid.). But upon further consideration of the "relevant statutory enactments" (see id. at 47a) -- which the court identified as the FOIA, the National Security Act of 1947, and the Central Intelligence Agency Act of 1949, 50 U.S.C. 403a et seq. -- the court concluded that the phrase "intelligence sources and methods" is in fact "ambiguous" (Pet. App. 49a) and must be interpreted in a way that reflects "Congress's sensitivity to the need for discrimination in identifying particular types of matters exempted from disclosure" (id. at 47a). The court of appeals acknowledged that the CIA's proposed definition of "intelligence sources" -- essentially, individuals or entities that provide intelligence information (see Pet. App. 46a) -- was a supportable interpretation of the phrase. But instead of accepting this definition, the court decided that "(a)nalysis should * * * focus on the practical necessity of secrecy * * * (and should) avoid an overbroad discretionary standard" (id. at 50a). The court then concluded (ibid.): (A)n "intelligence source" is a person or institution that provides, has provided, or has been engaged to provide the CIA with information of a kind the Agency needs to perform its intelligence function effectively, yet could not reasonably expect to obtain without guaranteeing the confidentiality of those who provide it. Judge Markey, in an opinion concurring in part and dissenting in part, urged that the district court's rejection of the Agency's Exemption 3 claim be affirmed but remarked (Pet. App. 62a): "Whatever may be said of the wisdom or morality of the MKULTRA program and its operation, the Agency's need for the research data "to perform its intelligence function effectively' has not been challenged on this record." 4. On remand, the district court began by explicitly rejecting respondents' contention that the MKULTRA research was not "needed to perform the CIA's intelligence function" (Pet. App. 22a). The court explained (id. at 22a-23a): "In view of the agency's concern that potential foreign enemies could be engaged in similar research and the desire to take effective counter-measures, * * * (the Agency) could reasonably determine that this research was needed for its intelligence function." The district court also rejected respondents' contention that some of the researchers were not intelligence sources because they provided "materials or supplies," rather than "information," to the CIA (id. at 22a); the court reasoned that "it is irrelevant whether a source tells the CIA how to synthesize a substance or synthesizes the substance itself and sells it to the CIA (because) in either case, what is essentially being provided is information" (ibid.). The district court then turned to the question whether the CIA could show, as the court of appeals' definition requires, that it could not reasonably have expected to obtain the information supplied by the MKULTRA sources without guaranteeing confidentiality to them (Pet. App. 23a). The district court acknowledged that the Agency considered the relationships between it and the MKULTRA researchers to be confidential (ibid.). The district court further noted that "(f)or understandable reasons, the Agency wished its interest in this subject matter kept secret, and feared that disclosure would jeopardize its ability to continue its research" (id. at 24a). But the district court ruled that this was not sufficient to satisfy the court of appeals' test because "the chief desire for confidentiality was on the part of the CIA. * * * (C)onfidentiality was normally guaranteed * * * solely to protect the CIA" (ibid.). In addition, the court remarked, many MKULTRA projects involved research that "goes on constantly at many places" and therefore "could have been done without a guarantee of confidentiality" (id. at 26a). After reviewing the Agency's submissions about particular cases, the district court found that some of the researchers had sought, and received, express promises of confidentiality from the Agency. The court ruled that the identities of these researchers and the institutions with which they had been associated need not be disclosed (Pet. App. 26a). The district court also exempted other researchers from disclosure for various reasons (see id. at 26a-27a), 30a-31a). In total, the court ordered the disclosure of the names of 47 of the researchers and the institutions with which they had been affiliated (id. at 21a-34a). 5. Both sides appealed, and a divided panel of the court of appeals reversed the district court's "determination regarding which of the individual researchers satisfy the 'need-for-confidentiality' portion of the definition of 'intelligence source' promulgated in" the court of appeals' earlier opinion. The court affirmed the district court's ruling in other respects. Pet. App. 11a. The court of appeals peremptorily rejected the Agency's suggestion that it reconsider the portion of the definition requiring the Agency to show that it had to guarantee confidentiality in order to obtain the information supplied by a source (Pet. App. 4a). Instead, the court of appeals criticized the district court for not following this aspect of the definition closely enough. The court of appeals remarked that "the (district) court's attention to questions of this order was deflected by its interest in whether the agency had, in fact, promised confidentiality to individual researchers" (id. at 5a). The court of appeals held that the district court's decision automatically to exempt from disclosure those researchers to whom the CIA had promised confidentiality was erroneous: "Proof that the CIA did or did not make promises of secrecy (either express or tacit) to specific informants * * * (cannot) be dispositive of the question whether a given informant qualifies as an 'intelligence source'" (id. at 6a). Specifically, the court of appeals ruled that even a source of intelligence information who received an express promise of confidentiality would have to be revealed if the source requested such a promise only because he was "unreasonably and atypically leery of providing the agency with innocuous information" (Pet. App. 6a). The court reasoned that "if the agency readily and openly could have obtained, from other sources, data of the sort (such a source) provided, he would not constitute an 'intelligence source'" (ibid. (footnote omitted)). The court also remarked that allowing the Agency to refuse to disclose the identities of all sources of intelligence that requested confidentiality could permit "widespread evasion of the letter and spirit of the FOIA" (id. at 7a) because it "would (be) * * * easy for the agency" to suggest to intelligence sources "that they sign a form expressing their desire for secrecy" (id. at 6a n.7). Judge Bork wrote a separate opinion, concurring in part and dissenting in part (Pet. App. 12a-16a). He criticized several aspects of the court of appeals' definition of "intelligence sources," urging in particular that there is "no reason to think that section 403(d)(3) was meant to protect sources of information only if secrecy was needed in order to obtain the information." Specifically, Judge Bork explained, "(t)he mere fact that the CIA pursues certain inquiries tells our adversaries much that there is no reason to think Congress intended them to know." He reasoned that "(o)ne need not be an expert in intelligence work to know that it is often possible to deduce what a person is doing, thinking, or planning by knowing what question he is asking or what information he is gathering. That is true even when the answers and information are publicly available." Id. at 15a. Judge Bork also strongly criticized the majority's conclusion that the FOIA sometimes requires the CIA to break a promise of confidentiality it has given to an intelligence source. He stated (Pet. App. 13a-14a): Many persons who expect pledges of confidentiality to be honored will be shocked to learn, long after they give information in return for such a promise, that their identities will be disclosed. * * * (I)n this very case, retrospective application of (the majority's definition) * * * may be profoundly unjust. It will certainly be so if it results in the disclosure of the identities of * * * researchers who fully, and justifiably, expected the government to keep its commitment and to protect them from the wide range of dangers that may have concerned them when they insisted on confidentiality. This is not an honorable way for the government of the United States to behave, and the dishonor is in no way lessened because it is mandated by a court of the United States. Judge Bork urged that by authorizing courts to force the CIA to break its promises of confidentiality, the majority's approach "produces pernicious results. * * * Because of the ever-present possibility of a future breach of trust ordered by the judiciary under the vague standard laid down today, the CIA will probably lose many future sources of valuable intelligence" (Pet. App. 13a-14a). Judge Bork remarked that under the court of appeals' definition of "intelligence sources," "individuals who give information to the CIA on the understanding that their names will be kept secret cannot rely on the promise of confidentiality if the information turns out to be the sort the CIA can get elsewhere without promising secrecy, something the sources of the information will often not be in a position to know. There is, moreover, no guarantee that a judge, examining the situation years later and deciding on the basis of a restricted record, will come to an accurate conclusion" (id. at 13a). Judge Bork then concluded (ibid.): The CIA and those who cooperate with it need and are entitled to firm rules that can be known in advance rather than vague standards whose application to particular circumstances will always be subject to judicial second-guessing. Our national interest, which is expressed in the authority to keep intelligence sources and methods confidential, requires no less. The court of appeals denied the CIA's petition for rehearing and suggestion of rehearing en banc. Judges Wilkey, Bork, and Scalia voted in favor of rehearing en banc. Pet. App. 18a. SUMMARY OF ARGUMENT A. The court of appeals' decision is inconsistent with the plain meaning of 50 U.S.C. 403(d)(3), which protects "intelligence sources" from disclosure without limitation or qualification. The district court specifically ruled that the MKULTRA project was within the scope of the Central Intelligence Agency's intelligence function, and the court of appeals did not question that ruling. Thus, the MKULTRA researchers were, literally, sources of intelligence information. That should have been the end of the inquiry. Nothing in the legislative history of Section 403(d)(3) suggests that the term "intelligence sources" should be given anything other than its plain meaning. On the contrary, the legislative history shows that Congress was aware that the CIA would derive intelligence from a large number of sources, and that these sources would be very diverse in character. Nonetheless, Congress did not attempt to differentiate among sources of intelligence information; it simply protected "intelligence sources" from disclosure. Moreover, Congress was acutely aware of the importance of secrecy to the intelligence-gathering process. Indeed, some of the congressional hearings on Section 403(d)(3) were held in secret and were only recently declassified. The court of appeals appears to have arrived at its narrow construction of the term "intelligence sources" by interpreting Section 403(d)(3) in a way that would reflect what the court considered to be the pro-disclosure philosophy of the Freedom of Information Act. This is a fundamental error. Exemption 3 of the FOIA incorporates by reference the exemptions from disclosure contained in the statutes to which it refers, and the legislative history of Exemption 3 expressly identifies Section 403(d)(3) as one of the statutes that is incorporated by reference. The disclosure of intelligence sources is therefore governed not by the "spirit" of the FOIA but by Section 403(d)(3) -- a statute enacted shortly after World War II in a climate quite different from that which prevailed at the time of the FOIA. B. In addition, the court of appeals' definition of "intelligence sources" leads to results that Congress could not possibly have intended. In the intelligence area, because the stakes are so high, it is crucially important to the CIA that it be able to give its sensitive intelligence sources as absolute a guarantee of confidentiality as possible, and that it be perceived by potential sources as being able to keept its commitments. "The Government has a compelling interest in protecting both the secrecy of information important to our national security and the appearance of confidentiality so essential to the effective operation of our foreign intelligence service." Snepp v. United States, 444 U.S. 507, 509 n.3 (1980), quoted in Haig v. Agee, 453 U.S. 280, 307 (1981)(emphasis added). The court of appeals' approach -- by requiring the CIA to reveal the identity of a source of intelligence information whenever a court determines, after the fact, that the Agency could have obtained the same kind of information without guaranteeing confidentiality -- would necessarily undermine the CIA's efforts to assure potential sources that their identities will not be revealed under circumstances that could cause them great harm. Moreover, the court of appeals' definition would require the Agency to disclose intelligence sources whenever the information they provide also happens to be in the public domain. This, too, would damage the Agency in a number of ways: by revealing to hostile foreign powers the subjects in which the Agency is interested; by making it difficult for the Agency to obtain information that, while theoretically available to the public, is far more easily obtained from a source that insists on confidentiality; and perhaps by requiring the Agency to disclose the identity of even a very sensitive source, if that source happened also to provide information that the CIA could have obtained without promising confidentiality. The legislative history of Section 403(d)(3) shows that Congress understood that the Agency would rely heavily on intelligence sources of the kind that the court of appeals' definition would require the Agency to disclose. There is no indication that Congress intended to exclude such sources from the unqualified protection it afforded to "intelligence sources" in Section 403(d)(3). In sum, Congress was aware that intelligence information would be provided by numerous and diverse sources, and it chose to enact an unqualified measure protecting "intelligence sources" from disclosure. There is no reason to give that term anything other than its literal meaning. ARGUMENT THE FREEDOM OF INFORMATION ACT DOES NOT REQUIRE THE CENTRAL INTELLIGENCE AGENCY TO DISCLOSE THE IDENTITIES OF SOURCES OF INTELLIGENCE-RELATED INFORMATION A. The Term "Intelligence Sources" Should Be Given Its Plain Meaning In this case, as in another Freedom of Information Act case recently decided by the Court, "(t)he plain language of the statute * * * is sufficient to resolve the question presented" (United States v. Weber Aircraft Corp., No. 82-1616 (Mar. 20, 1984), slip op. 6). 1. The only issue in this case is the meaning of the term "intelligence sources" in Section 102(d)(3) of the National Security Act of 1947, 50 U.S.C. 403(d)(3). Exemption 3 of the FOIA provides that the FOIA does not require an agency to disclose "matters that are * * * specifically exempted from disclosure by statute * * * provided that such statute * * * refers to particular types of matters to be withheld" (5 U.S.C. 552(b)(3)(B)). It is beyond dispute that Section 403(d)(3) is one of the statutes referred to by Exemption 3; the court of appeals twice acknowledged this "well-established" point (Pet. App. 2a n.1; see id. at 44a), and respondents have not contended otherwise. Indeed, the legislative history of Exemption 3 explicitly identifies Section 403(d)(3) as a principal example of an Exemption 3 statute. See H.R. Rep. 94-880, 94th Cong., 2d Sess., Pt. 2, at 15 n.2 (1976). See also H.R. Rep. 93-1380, 93d Cong., 2d Sess. 12 (1974); S. Rep. 93-854, 93d Cong., 2d Sess. 16 (1974); S. Rep. 98-305, 98th Cong., 1st Sess. 7 n.4 (1983). And the courts of appeals have consistently held that Section 403(d)(3) is an Exemption 3 statute. See, e.g., Gardels v. CIA, 689 F.2d 1100, 1103 (D.C. Cir. 1982); Goland v. CIA, 607 F.2d 339, 350 (D.C. Cir. 1978), cert. denied, 445 U.S. 927 (1980); National Commission on Law En orcement and Social Justice v. CIA, 576 F.2d 1373, 1376 (9th Cir. 1978). Section 403(d)(3) specifically authorizes the Director of Central Intelligence to protect "intelligence sources and methods" from disclosure. It follows that the FOIA does not require the disclosure of the identity of any entity that is an "intelligence source" within the meaning of Section 403(d)(3). 2. a. The court of appeals did not appear to deny that the MKULTRA researchers were, literally, sources of intelligence. The district court specifically ruled that the CIA "could reasonably determine that (the MKULTRA) research was needed for its intelligence function" (Pet. App. 22a-23a), and the court of appeals did not question this ruling. Indeed, on the first appeal, Judge Markey, who would have ordered the researchers' identities disclosed without further proceedings, nonetheless acknowledged that "the Agency's need for the research data "to perform its intelligence function effectively' has not been challenged on this record" (id. at 62a). The MKULTRA researchers were, therefore, "intelligence sources" within the literal meaning of that term. That should have been the end of the inquiry. This Court has frequently emphasized that the plain language of a statute is the surest guide to Congress's intentions (see, e.g., United States v. Rodgers, No. 83-620 (Apr. 30, 1984), slip op. 4; Steadman v. SEC, 450 U.S. 91, 97 (1981)), and there is no reason to believe that Congress meant the words of Section 403(d)(3) to have something other than their plain meaning -- that all sources of intelligence are protected from disclosure. "Absent a clearly expressed legislative intention to the contrary," the "language of the statute itself" must "ordinarily be regarded as conclusive." Consumer Product Safety Commission v. GTE Sylvania, Inc., 447 U.S. 102, 108 (1980). In this case, there is nothing approaching "the kind of compelling evidence of congressional intent that would be necessary to (warrant) * * * look(ing) beyond the plain statutory language" (Weber Aircraft Corp., slip op. 9). Congress did not say -- as the court of appeals has held -- that the Director of Central Intelligence is authorized to protect intelligence sources only if such protection is needed to obtain information that otherwise could not be obtained. Nor did Congress say that only confidential or nonpublic intelligence sources are protected. In other provisions of the FOIA and in the Privacy Act, a related statute, Congress has protected "confidential source(s)," sources of "confidential information," and sources that provided information under an express promise of confidentiality. See 5 U.S.C. 552(b)(7)(D); 5 U.S.C. 552a(k)(2) and (5). But Section 403(d)(3) contains no such language; Congress simply protected all sources of intelligence. Indeed, as the court of appeals acknowledged in its first opinion in this case, that court's prior decisions dealing with Section 403(d)(3) and Exemption 3 "simply assumed the phrase ('intelligence sources') to have a plain meaning" (Pet. App. 44a). And in subsequent cases in which the definition it devised in this case had apparently not been brought to its attention, the District of Columbia Circuit has continued to give the term "intelligence sources" its plain meaning. See, e.g., Afshar v. Department of State, 702 F.2d 1125, 1130 (1983) ("The Freedom of Information Act bars the courts from prying loose from the government even the smallest bit of information that * * * would disclose intelligence sources or methods."); Gardels, 689 F.2d at 1104. b. While the legislative history of Section 403(d)(3) does not specifically address the meaning of the term "intelligence sources and methods" -- presumably because Congress did not see any ambiguity in the phrase -- it suggests no reason whatever to doubt that Congress intended to give the Director of Central Intelligence broad power to protect the secrecy of the intelligence process. The National Security Act of 1947 was enacted shortly after World War II. Section 403 established the CIA and empowered it, among other things, "to correlate and evaluate intelligence relating to the national security" (50 U.S.C. 403(d)(3)). The legislative history of Section 403 shows that Congress was concerned about reports of shortcomings in American intelligence before Pearl Harbor and during World War II and was determined to improve the nation's capacity to gather and analyze intelligence in peacetime as well as in war. See, e.g., S. Rep. 239, 80th Cong., 1st Sess. 2 (1947); H.R. Rep. 961, 80th Cong., 1st Sess. 3-4 (1947); 93 Cong. Rec. 9444 (1947). See also Commission on Organization of the Executive Branch of the Government, Intelligence Activities: A Report to the Congress 29-30 (1955). At least two aspects of the legislative history shed light on the scope of the protection Congress afforded to "intelligence sources and methods." First, Congress was well aware that the CIA would derive intelligence from a large number of diverse sources. Congress created the CIA because it envisioned that the government would have to collect and analyze a "mass of information" in order to survive in the postwar world. See S. Rep. 239, supra, at 2 ("(T)o meet the future with confidence, we must make certain * * * that a central intelligence agency collects and analyzes that mass of information without which the Government cannot either maintain peace or wage war successfully"). See also 93 Cong. Rec. 9397 (1947) (remarks of Rep. Wadsworth) ("The function of that agency is to constitute itself as a gathering point for information coming from all over the world through all kinds of channels."); National Defense Establishment: Hearings on S. 758 Before the Senate Comm. on Armed Services, 80th Cong., 1st Sess., Pt. 3, at 669 (1947) (statement of Charles S. Cheston, former military intelligence official) (The agency must have "authority to analyze and correlate information from all sources.") (hereinafter cited as Senate Hearings); National Security Act of 1947: Hearings on H.R. 2319 Before the House Comm. on Expenditures in the Executive Departments, 80th Cong., 1st Sess. 112 (1947) (remarks of Rep. Boggs) (the Director of Central Intelligence "is dealing with all the information and the evaluation of that information, from wherever we can get it") (hereinafter cited as House Hearings); Senate Hearings, supra, at 132 (statement of Fleet Admiral Nimitz) ("(T)he Central Intelligence Agency (is) charged with responsibility for collection of information from all available sources * * *. (I)ntelligence is a composite of authenticated and evaluated information covering not only the armed forces establishment of a possible enemy, but also his industrial capacity, racial traits, religious beliefs, and other related aspects."); id. at 497 (statement of General Vandenberg, Director of Central Intelligence Group) ("Collection in the field of foreign intelligence consists of securing all possible data pertaining to foreign governments or the national defense and security of the United States."). Congress was also advised of the extraordinary diversity of intelligence sources. The classic secret agent, Congress was told, is only one such source. Allen W. Dulles, an important figure in wartime military intelligence who subsequently became Director of Central Intelligence, explained that "American businessmen and American professors and Americans of all types and descriptions who travel around the world are one of the greatest repositories of intelligence that we have." National Security Act of 1947: Hearings on H.R. 2319 Before the House Comm. on Expenditures in the Executive Departments, 80th Cong., 1st Sess. 22 (June 27, 1947) (published 1982) (hereinafter cited as Secret House Hearings); /3/ see id. at 28. Another high-ranking intelligence official emphasized "the great open sources of information * * * such things as books, magazines, technical and scientific surveys, photographs, commercial analyses, newspapers, and radio broadcasts, and general information from people with a knowledge of affairs abroad" (Senate Hearings, supra, at 492 (statement of General Vandenberg)). Second, Congress was acutely aware of the importance of secrecy. See Snepp v. United States, 444 U.S. 507, 512 (1980) ("The continued availability of * * * (intelligence) sources depends upon the CIA's ability to guarantee the security of information that might compromise them and even endanger the(ir) personal safety."). Congress was advised in graphic terms, by high-ranking intelligence officials, of the deadly peril that faced intelligence sources whose identities were revealed. See Secret House Hearings, supra, at 10-11 (statement of General Vandenberg); id. at 20 (statement of Allen W. Dulles). And Congress was told that even American citizens who supply intelligence information "close up like a clam" unless they can hold the government "responsible to keep the complete security of the information they turn over" (Secret House Hearings, supra, at 22 (statement of Allen W. Dulles)). /4/ The committees of both Houses went into executive session to consider the proposed legislation; the Secret House Hearings, supra, were declassified only in 1982. See id. at v-viii; S. Rep. 239, supra, at 1. A member of the House committee stated on the floor (93 Cong. Rec. 9444 (1947) (statement of Rep. Manasco)): We were sworn to secrecy and I hesitate to even discuss this section because I am afraid I might say something, because the Congressional Record is a public record, and divulge some information here that would give aid and comfort to any potential enemy we have * * *. The things we say here today, the language we change, might endanger the lives of some American citizens in the future. We know of no suggestion in the legislative history that Congress thought the CIA might be too secretive. Against this background, Congress specified that the Director of Central Intelligence is responsible for "protecting intelligence sources and methods from unauthorized disclosure"; that language did not appear in the Administration draft. Compare H.R. 2319, 80th Cong., 1st Sess. Sec. 202 (1947), with H.R. 4214, 80th Cong., 1st Sess. Sec. 105(d)(3) (1947); see H.R. Rep. 961, supra, at 3-4. /5/ Nothing in this legislative history remotely indicates that Congress intended Section 403(d)(3) or its crucial language -- "intelligence sources and methods" -- to be construed narrowly or in a way that would promote the disclosure of intelligence sources to the public. Nor is there any basis for concluding that Congress was concerned to restrict the authority of the Director of Central Intelligence to withhold information. Congress was advised that the CIA would draw upon a large and diverse group of intelligence sources; Congress was clearly aware that secrecy was extremely important; and it granted the Director of Central Intelligence unqualified authority to protect the secrecy of sources. Congress plainly intended the Director's authority to have the broadest scope. 3. The court of appeals appeared to proceed from the premise that the term "intelligence sources" must be given a narrow meaning in order to avoid "broad agency discretion" (Pet. App. 45a; see id. at 50a) and thereby to serve what the court of appeals considered to be the pro-disclosure "spirit" of the FOIA (see, e.g., id. at 7a, 42a; see also id. at 41a-45a, 47a, 50a). This approach is fundamentally misconceived. The disclosure of intelligence sources is governed by Section 403(d)(3), not by the substantive standards or "spirit" of the FOIA. The purpose of Exemption 3 is to make it clear that the FOIA does not repeal by implication certain other statutes. See H.R. Rep. 1497, 89th Cong., 2d Sess. 10 (1966). As we noted, it is beyond dispute that Section 403(d)(3) is one of the statutes identified by Exemption 3. The legislative history is explicit that Exemption 3 "incorporat(es) by reference exemptions contained in (the) * * * statutes" it identifies. H.R. Conf. Rep. 94-1441, 94th Cong., 2d Sess. 14 (1976); see id. at 25. /6/ Thus, the only question in this case is the interpretation of the term "intelligence sources and methods" in Section 403(d)(3). The supposed pro-disclosure philosophy of the FOIA is quite irrelevant to that question; as we have shown, Section 403(d)(3) was enacted in a climate far different from that which prevailed at the time the FOIA was enacted. The court of appeals' anachronistic attempt to impute its own skepticism about CIA secrecy (a skepticism it also attributed to the Congress that enacted the FOIA) to the post-war Congress that establihsed the Agency appears to be the fundamental error that led to its wholly unjustifiable definition of "intelligence sources." The court of appeals offered other justifications for narrowing the explicit protection afforded to intelligence sources by Section 403(d)(3), but none of them is any more substantial. For example, the court relied on a provision of the Central Intelligence Agency Act of 1949, 50 U.S.C. 403g, which provides in part: In the interests of the security of the foreign intelligence activities of the United States and in order further to implement the proviso of section 403(d)(3) of this title that the Director of Central Intelligence shall be responsible for protecting intelligence sources and methods from unauthorized disclosure, the Agency shall be exempted from the provisions of (5 U.S.C. (1958 ed.) 654) /7/ and the provisions of any other law which require the publication or disclosure of the organization, functions, names, official titles, salaries, or numbers of personnel employed by the Agency. The court of appeals stated that "Section 403g provides specific protection for most of the CIA activities and contractual relationships about which the Agency has expressed greatest concern * * * (and) evinces a congressional awareness that Section 403(d)(3) * * * would require construction and interpretation limiting executive discretion to withhold; otherwise it would have felt no need to 'implement' the original proviso by listing the specific matters exempted from disclosure under Section 403g." Pet. App. 49a-50a. This reasoning is erroneous in many ways. Section 403g does not in fact protect "most of the CIA activities * * * about which the Agency has expressed greatest concern." Section 403g applies only to "personnel employed by" the CIA, and many important intelligence sources might not be regarded as personnel employed by the CIA; one example is Americans travelling abroad, who, Director Dulles advised Congress, "are one of the greatest repositories of intelligence that we have" (Secret House Hearings, supra, at 22). Other sources, as well, furnish information simply because they wish to aid the Agency or the United States. But the extent to which Section 403g protects intelligence sources is, in any event, irrelevant. Congress enacted both Section 403(d)(3) and Section 403g, and it cannot be seriously disputed that both Section 403g and Section 403(d)(3) are included within Exemption 3. See page 15, supra; National Commission, 576 F.2d at 1376. Section 403(d)(3) should not be given an artificially narrow interpretation that finds no support in its language or legislative history merely because Section 403g also exists to protect what Congress described as "the confidential nature of the Agency's functions" (H.R. Rep. 160, 81st Cong., 1st Sess. 6-7 (1949); see S. Rep. 106, 81st Cong., 1st Sess. 1 (1949)). See also Baker v. CIA, 580 F.2d 664, 667-669 (D.C. Cir. 1978). Cf. SEC v. National Securities, Inc., 393 U.S. 453, 468 (1969). The court of appeals' suggestion that the enactment of Section 403g reveals Congress's awareness that Section 403(d)(3) "require(s) construction and interpretation limiting executive discretion to withhold" (Pet. App. 49a-50a) is similarly a non sequitur. The court of appeals would read Section 403g as if it superseded, narrowed, or exhausted the content of Section 403(d)(3), but what Section 403g says is that it "further * * * implement(s)" Section 403(d)(3). The most likely explanation of Congress's decision to enact Section 403g is that in 1949, there were other statutes that might have been construed to require the disclosure of information about CIA employees -- 5 U.S.C. (1958 ed.) 654, which is mentioned in Section 403g but has since been repealed, was apparently the statute Congress had in mind (see page 24 note 7, supra) -- and Congress thought it advisable explicitly to exempt the Agency from such statutes. But before the FOIA was enacted, there was no statute that could have been thought to require the disclosure of intelligence sources generally. When Congress enacted the FOIA, it included Exemption 3 and, as we noted, specified that Section 403(d)(3) is an Exemption 3 statute. Nothing in this pattern of congressional activity suggests that the protection that Exemption 3 and Section 403(d)(3) afford to intelligence sources should be given less than its full, literal meaning. /8/ Finally, the court of appeals stated that its definition of "intelligence sources" was justified because when Congress enacted Section 403(d)(3), "(s)ecrecy seems to have been a concern only insofar as it was pertinent to protection of the national security. Analysis should therefore focus on the practical necessity of secrecy. * * * Section 403(d)(3) must be interpreted in functional terms" (Pet. App. 50a). As we will explain, the court of appeals' definition of intelligence sources reveals that the court was ill-informed about the ways in which secrecy is "functional" in the intelligence area and the reasons that secrecy can be a "practical necessity." See pages 29-41, infra. But the more fundamental point is that Congress has already determined the extent to which secrecy is a "practical necessity" and is "pertinent to protection of the national security"; Congress's judgment is relfected in its unqualified mandate to the Director to "protect() intelligence sources and methods from unauthorized disclosure." It was not open to the court of appeals to second-guess Congress by deciding that it is only sometimes necessary to protect intelligence sources from disclosure. 4. Contrary to some of the suggestions made by respondents and the court of appeals (see, e.g., Memo. in Opp. 5, 9, 10-12; Pet. App. 45a, 47a, 50a), interpreting Section 403(d)(3) according to its plain meaning will not give the CIA unlimited authority to withhold documents requested under the FOIA. The CIA may engage only in authorized intelligence activities. See 50 U.S.C. 403(d); Exec. Order No. 12,333, Section 1.51, 46 Fed. Reg. 59941, 59944 (1981). Moreover, the Agency's intelligence-gathering operations are subject to a number of statutory restrictions. See 50 U.S.C. 403(d)(3) ("(T)he Agency shall have no police, subp(o)ena, law-enforcement powers, or internal-security functions."). This case, however, does not now involve any dispute over the meaning of the term "intelligence" or the breadth of the Agency's intelligence function. /9/ As we have noted, the district court ruled that the Agency "could reasonably determine that (the MKULTRA) research was needed for its intelligence function" (Pet. App. 22a-23a), and the court of appeals did not disturb that ruling. The only question, therefore, is whether Section 403(d)(3) protects from disclosure a "source" of information that is acknowledged to be necessary for the Agency's "intelligence" function. That question is answered by the plain language of the statute. B. The Court of Appeals' Definition of "Intelligence Sources" Produces Results That Congress Could Not Have Intended The severe difficulties that would be created if the court of appeals' definition of "intelligence sources" were routinely applied in FOIA cases are a further reason for interpreting that term according to its plain meaning. The court of appeals' definition would require disclosures that could be extremely damaging to the CIA's ability to carry out its mission, and Congress could not have intended to permit such disclosures. Indeed, there is specific evidence in the legislative history of Section 403(d)(3) that Congress intended to preclude many of the kinds of disclosures that would be required by the court of appeals' decision. But the fact that the court of appeals' definition of "intelligence sources" leads to results that are obviously inconsistent with Congress's intentions is not just an indictment of that court's particular formulation; it is a further illustration of why Congress chose to give unqualified protection to "intelligence sources and methods." The court of appeals, notwithstanding its concern with the "practical necessity of secrecy" and its effort to devise a "functional" definition, did not recognize that its definition produces wholly unacceptable results when applied to many categories of intelligence sources. Nor did the court of appeals fully understand the manifold and complex ways in which secrecy is vital to intelligence gathering. Congress protected "intelligence sources and methods" without qualification, and the profound shortcomings of the court of appeals' definition suggest that a narrower protection can risk interfering with the intelligence-gathering mission of the CIA in important ways that are not always apparent. 1. The court of appeals grievously underestimated the importance of providing intelligence sources with an assurance of confidentiality that is as absolute as possible. Under the court of appeals' appraoch, the CIA will be forced to disclose an intelligence source whenever a court determines, after the fact, that the Agency could have obtained the kind of information supplied by the source without promising confidentiality. Indeed, the court of appeals carried this approach to the point of holding that the Agency will be required to betray an explicit promise of confidentiality if a court determines that the promise was not necessary -- or, in the court of appeals' words, if a court decides that the intelligence source to whom the promise was given was "unreasonably and atypically leery" of cooperating with the CIA (Pet. App. 6a). /10/ Few things will have as devastating an impact on the CIA's ability to carry out its mission as forced disclosure of the identities of its sources. "The Government has a compelling interest in protecting both the secrecy of information important to our national security and the appearance of confidentiality so essential to the effective operation of our foreign intelligence service." Snepp, 444 U.S. at 509 n.3, quoted in Haig v. Agee, 453 U.S. 280, 307 (1981)(emphasis added). If potential intelligence sources begin to perceive that the Agency will be unable to maintain the confidentiality of its relationship to them, many can be expected to refuse to supply information to the Agency in the first place. This will be true no matter how rational or well-founded their perception is. As this Court has recognized in both the intelligence area and other contexts, what is crucial is not just whether the government has in fact betrayed a confidence but "the appearance that confidentiality ha(s) been breached" or might be breached (Baldrige v. Shapiro, 455 U.S. 345, 361 n.17 (1982)(emphasis in original)). "An uncertain privilege, or one which purports to be certain but results in widely varying applications by the courts, is little better than no privilege at all" (Upjohn Co. v. United States, 449 U.S. 383, 393 (1981)). Certainty is especially vital in the intelligence area, where the risks are enormous: if an intelligence source's cooperation becomes known -- indeed, even if it is revealed that he cooperated with the CIA in some apparently innocuous way -- he may face not only great embarrassment but far more severe consequences. See Haig v. Agee, 453 U.S. at 285 & n.7; J.A. 44-46 (affidavit of Louis J. Dube). Consequently, even if there is only a relatively small probability that a court will order disclosure of a source's identity, that will be of little comfort to the source. In order to induce him to cooperate, the CIA will have to provide as absolute an assurance of confidentiality as it possibly can. "The continued availability of * * * (intelligence) sources depends upon the CIA's ability to guarantee the security of information that might compromise them and even danger (their) personal safety" (Snepp, 444 U.S. at 512). See also J.A. 20-21. Cf. Weber Aircraft Corp., slip op. 10-11 n.23; Williams v. FBI, 730 F.2d 882, 885-886 (2d Cir. 1984). Nor will it reassure a potential intelligence source to learn that a court will order his identity revealed only after examining the facts of the case and determining that he was "atypically or unreasonably leery," that the Agency could have obtained the same information from another source without guaranteeing confidentiality, /11/ or that some similar standard was met. An intelligence source will "not be concerned with the underlying rationale for disclosure of" his cooperation with the CIA if his cooperation was secured "under assurances of confidentiality" (Baldrige, 455 U.S. at 361). Moreover, a court's decision whether an intelligence source will be harmed if his identity is revealed will often require complex political, historical, and psychological judgments, sometimes about societies very different from our own. /12/ There is no reason for anyone -- especially a potential intelligence source whose life may be at stake -- to have great confidence in a court's ability to make those judgments correctly. Indeed, many potential intelligence sources, especially those who are not American citizens, are likely to perceive courts as unpredictable institutions that are influenced by concerns that have little in common with the world in which the potential source and his CIA contact must operate. A potential source is also likely to realize that a court's decision whether to reveal his cooperation with the agency may occur well in the future, at a time when the source's concern for confidentiality may seem less understandable to everyone except the source himself. See also Pet. App. 92a (affidavit of Director of Central Intelligence Turner) ("(A) unilateral breach of confidentiality and trust by the United States Government will be viewed as an arrogant disregard for the lives or safety or reputations of those who have contributed to our intelligence activities."). 2. The court of appeals also failed to recognize that when Congress protected "intelligence sources" from disclosure, it was not simply protecting sources of secret intelligence information; as we have noted, Congress was aware that secret agents are not the most typical intelligence sources. Many important sources provide the Agency with information that members of the public could, at least in theory, also obtain. But under the court of appeals' definition of "intelligence sources," the Agency cannot withhold the identity of a source of intelligence information if that information is also publicly available, because the Agency can obtain such information without guaranteeing confidentiality. This approach is divorced from the realities of intelligence work and is demonstrably inconsistent with Congress's understanding of the purposes of Section 403(d)(3). First, as Judge Bork explained, another government can learn a great deal about "what subjects (are) of interest to the CIA" by examining the public sources of information that the Agency is exploiting: "One need not be an expert in intelligence work to know that it is often possible to deduce what a person is doing, thinking, or planning by knowing what question he is asking or what information he is gathering. That is true even when the answers and information are publicly available. The mere fact that the CIA pursues certain inquiries tells our adversaries much that there is no reason to think Congress intended them to know." Pet. App. 15a. The facts of this case themselves suggest an example. When the CIA decided to investigate "brainwashing" and the countermeasures that might be taken, it might have turned to sources, such as journals and ongoing research projects, that are available to the public. But a foreign government that learned the sources that the Agency was consulting would have been able to infer both the general nature of the CIA's project and the directions that its inquiry was taking. See Pet. App. 89a-90a (affidavit of Director of Central Intelligence Turner) ("Throughout the course of the Project, CIA involvement or association with the research was concealed in order to avoid stimulating the interest of hostile countries in the same research areas."). /13/ Similarly, the court of appeals, in its first opinion, suggested that the excessive breadth of the CIA's proposed definition of "intelligence sources" was revealed by the Agency's acknowledgment that its definition would "apply even to periodicals -- including Pravda and the New York Times -- from which (the Agency) culls information that informs its view of foreign nations and their policy intentions" (Pet. App. 46a (footnote omitted)). But the disclosure that the CIA consults Pravda and the New York Times is innocuous not because those periodicals are publicly available but because it is the disclosure of a fact -- the fact that the Agency consults these newspapers -- that is already commonly assumed to be true. An obscure Eastern European technical journal might also be available to members of the public; but disclosure of the fact that the CIA subscribes to that journal could easily thwart the CIA's efforts to exploit its value as an intelligence source. In addition, much information that is publicly available in principle can be difficult to obtain from public sources, and far more easily obtained from sources whose identity both the source and the Agency might legitimately wish to protect. An example of such information might be details of the travel plans or financial transactions of an individual whom the CIA is observing. Such information is theoretically in the public domain, but as a practical matter the Agency may have to obtain it from a source who would cease to provide it if his identity were revealed. The court of appeals would apparently require the disclosure of such a source's identity -- since the information was available by means that did not involve a guarantee of confidentiality -- thereby impairing the Agency's ability to recruit such sources in the future. Finally, the court of appeals apparently failed to recognize that many intelligence sources will provide the CIA with both highly sensitive information -- of a kind that would never be supplied without a promise of confidentiality -- and public or "innocuous" information that the Agency could have obtained elsewhere without guaranteeing secrecy. For example, a source in a foreign nation might supply, in addition to much sensitive information, a report of a crop failure that American journalists also learned about and reported in newspapers. If an FOIA request were then filed for all sources of information about the crop failure, the court of appeals' definition might be interpreted to require the Agency to disclose its relationship with the sensitive foreign source -- even though that disclosure would be likely to damage both the Agency and the source. /14/ In these respects, as well, the court of appeals' definition is harmful not only because it will force certain disclosures but because of the prospective effect it will have on the CIA's operations. If the Agency knows that it will be required to reveal its sources of public information in response to a proper FOIA request, it can be expected to alter its techniques so as to reduce the damage that such a disclosure might cause. In this way, the court of appeals' failure simply to protect all "intelligence sources" would force the CIA to depart from what it considers the best use of intelligence sources out of concern that some of those sources might become public. The legislative history of Section 403(d)(3) confirms what is obvious in any event -- that Congress would never have countenanced the results that the court of appeals' approach produces. As we have noted, Congress was specifically made aware that many -- indeed most -- "intelligence sources" provide information that is, at least in principle, publicly available; nonetheless, Congress gave no indication that it intended to exclude such sources from the protection of Section 403(d)(3). For example, General Vandenberg, who had been Director of the Central Intelligence Group, the CIA's predecessor (see 11 Fed. Reg. 1337 (1946); pages 21-22 note 5, supra), explained to Congress that "roughly 80 percent of intelligence should normally be based" on "the great open sources of information * * * such things as books, magazines, technical and scientific surveys, photographs, commercial analyses, newspapers, and radio broadcasts, and general information from people with a knowledge of affairs abroad." Senate Hearings, supra, at 492. Indeed, General Vandenberg asserted that this kind of intelligence had been neglected before World War II and urged that it be a principal concern of the post-war agency (ibid.). Essentially the same points were made by Director Dulles. See id. at 526; Secret House Hearings, supra, at 22. Thus, when Congress gave the Director of Central Intelligence unequivocal authority to "protect() intelligence sources and methods from unauthorized disclosure," it had previously been informed by the Director's predecessor and the future Director that the great preponderance of intelligence sources would be public sources. In these circumstances, Congress must have intended the Director's authority to extend to public sources. Similarly, Admiral Inglis, then Chief of Naval Intelligence, testified that intelligence agents in foreign countries are often used to "confirm or not what we have deduced from * * * Russian propaganda broadcasts" (Secret House Hearings, supra, at 63). Congress, therefore, unlike the court of appeals, was aware that highly secret intelligence sources will sometimes also report information that the CIA can obtain, or has obtained, from open sources. It is inconceivable that Congress, after having received Admiral Inglis's testimony, would have decided to deny the Director of Central Intelligence the authority to protect sources who function in this way. /15/ 3. In sum, as Congress knew when it enacted Section 403(d)(3) in 1947, intelligence sources can take a variety of forms: the secret agency; a cooperative foreign citizen or an American travelling abroad; an unwitting source who does not realize he is conveying information to the CIA and would stop conveying it if he did; or a publicly available periodical or radio broadcast. And Congress was aware that these sources could provide many different kinds of information, ranging from the most sensitive secrets to apparently innocuous facts from which Agency experts might derive an important lead. Secrecy is important in each of these contexts. Not only is secrecy necessary to protect a source supplying sensitive information; the Agency must also protect its reputation for being able to guarantee confidentiality to potential sources. In addition, the CIA must ensure that it can continue to exploit unwitting sources; and in the case of public sources of information, secrecy is important to avoid revealing to observers the subjects in which the Agency is interested. Congress, acting in light of these complex realities of intelligence gathering, did not attempt to delineate the categories of intelligence sources or information that were to remain confidential. It simply gave the Director of Central Intelligence unqualified authority to "protect() intelligence sources and methods from unauthorized disclosure." There is no indication that Congress intended these terms to have anything other than their literal meaning. Accordingly, the court of appeals erred when it attempted to devise a narrower definition. CONCLUSION The judgment of the court of appeals should be reversed. Respectfully submitted. REX E. LEE Solicitor General RICHARD K. WILLARD Acting Assistant Attorney General KENNETH S. GELLER Deputy Solicitor General DAVID A. STRAUSS Assistant to the Solicitor General ROBERT E. KOPP LEONARD SCHAITMAN Attorneys STANLEY SPORKIN General Counsel EDWARD PAGE MOFFETT Attorney Central Intelligence Agency JUNE 1984 /1/ "I C.A. App." refers to the Joint Appendix filed in the court of appeals on the first appeal, court of appeals docket numbers 79-2203 and 79-2104. "II C.A. App." refers to the Joint Appendix filed on the second appeal, court of appeals docket numbers 82-1945 and 82-1961. /2/ The CIA did release these names to congressional committees investigating MKULTRA (see Project MKULTRA Hearing, supra, at 8, 13, 49), but it requested that the committees keep the names confidential, and the committees honored the request. /3/ The committees of both Houses conducted hearings in executive session. As we note below in text, the executive session hearings of the House committee were only recently declassified and published. /4/ Since the earliest days of the Republic, secrecy has been recognized as vital to the successful gathering of intelligence. In a letter of July 26, 1797, issuing orders for an intelligence mission, George Washington wrote to Colonel Elias Dayton: "The necessity of procuring good intelligence, is apparent and need not be further urged. All that remains for me to add is, that you keep the whole matter as secret as possible. For upon secrecy, success depends in most Enterprises of the kind, and for want of it they are generally defeated * * *." The Writings of George Washington 478-479 (J. Fitzpatrick ed. 1933). See also United States v. Curtiss-Wright Export Corp., 299 U.S. 304, 320 (1936). /5/ The phrase "protecting intelligence sources and methods" was derived from the Presidential Directive of Jan. 22, 1946 (11 Fed. Reg. 1337), which was incorporated by reference in the Administration bill (H.R. 2319, 80th Cong., 1st Sess. Sec. 202 (1947)). This derivation scarcely suggests a narrow construction; the directive was issued by President Truman shortly after the war to establish the National Intelligence Agency and the Central Intelligence Group and to charge these predecessors of the CIA with "assur(ing) the most effective accomplishment of the intelligence mission related to the national security" (11 Fed. Reg. 1337 (1946)). They were accordingly made "responsible for fully protecting intelligence sources and methods" (id. at 1339). /6/ That is also how this Court has interpreted Exemption 3. In Baldrige v. Shapiro, 455 U.S. 345 (1982), the Court first considered whether 13 U.S.C. 8(b) and 9(a) were Exemption 3 statutes (455 U.S. at 354-355); after determining that they were, the Court turned to the language and legislative history of those provisions and did not further consider the FOIA (id. at 355-359). Similarly, in Consumer Product Safety Commission v. GTE Sylvania, Inc., supra, the Court, upon determining that 15 U.S.C. 2055(b)(1) was an Exemption 3 statute, had no difficulty holding that the time limits governing disclosure under that statute -- not the more restrictive time limits of the FOIA -- would control whenever a request for information covered by Section 2055(b)(1) was made. See 447 U.S. at 121-122. As this holding reflects, once a statute is determined to be an Exemption 3 statute, the entire regime it establishes to govern disclosures -- its time limits, its substantive criteria, and its "spirit" -- supersedes the standards found in the FOIA. /7/ This statute required the Civil Service Commission to compile an Official Register of the United States, "which shall contain a full and complete list of all persons occupying administrative and supervisory positions in the legislative, executive, and judicial branches of the Government * * *. The register shall show the name; official title; salary, compensation, and emoluments; legal residence and place of employment for each person listed therein * * *." This provision was repealed by the Act of July 12, 1960, Pub. L. No. 86-626, Sec. 101, 74 Stat. 427. /8/ The court of appeals also stated (Pet. App. 49a; see also id. at 11a n.13) that the existence of Exemption 1 of the FOIA, 5 U.S.C. 552(b)(1), which applies to properly classified documents, warrants a narrowed interpretation of Section 403(d)(3). (The identities of the MKULTRA researchers are not classified.) This approach is inconsistent even with prior precedent in the District of Columbia Circuit. See Gardels, 689 F.2d at 1107. It is true that the identities of intelligence sources will frequently be classified information, but that does not affect the interpretation of the independent exemption provided by Exemption 3 and Section 403(d)(3); as this Court has recently ruled, the interpretation of an FOIA exemption should not be distorted because a different exemption may also apply to some of the same documents. See FBI v. Abramson, 456 U.S. 615, 629-630 (1982). See also pages 2-4 of the Reply Brief for the Petitioners filed in support of the petition for a writ of certiorari. /9/ Of course, the Agency's determination that a particular person or entity is a source of intelligence information should be accorded great deference by a reviewing court (see, e.g., Halperin v. CIA, 629 F.2d 144, 148 (D.C. Cir. 1980); Gardels, 689 F.2d at 1104-1105), and, as Judge Bork noted, the court should not be permitted to substitute its judgment for that of the Director in the exercise of his sound discretion (see Pet. App. 15a-16a). /10/ The court of appeals' ruling that the Agency will sometimes be required to betray explicit promises of confidentiality is extraordinary. "Great nations, like great men, should keep their word." FPC v. Tuscarora Indian Nation, 362 U.S. 99, 142 (1960) (Black, J., dissenting), quoted in Heckler v. Mathews, No. 82-1050 (Mar. 5, 1984), slip op. 19. In a sense, however, the court of appeals' requirement that the Agency must be prepared to breach explicit promises of confidentiality is only the tip of the iceberg; it reflects the deeper misconceptions of the court of appeals' approach. The assurance of confidentiality given to an intelligence source is seldom recorded in writing, and often it is not even explicit. It is ususally so obvious to all concerned that the source wants to remain confidential that no explicit understanding is needed; an assurance of confidentiality is implicit in the relationship between the Agency and the source. See J.A. 17, 25-26 (affidavit of M. Corley Wonus). Cf. Londrigan v. FBI, 722 F.2d 840, 844-845 (D.C. Cir. 1983). Indeed, an experienced CIA operations officer explained, in an affidavit submitted in this case, that many sources would refuse to memorialize an agreement of confidentiality because doing so would create an additional document, linking them with the CIA, that might fall into the wrong hands. See J.A. 43 (affidavit of Louis J. Dube). For these reasons, the practical damage caused by the court of appeals' definition would not be greately reduced even if it were amended so as to require courts to honor fully explicit promises of confidentiality. /11/ There is, of course, no reason to believe that courts generally have the expertise needed to determine whether the Agency would have been able to obtain certain information without promising confidentiality to a source. /12/ In Fitzgibbon v. CIA, 578 F. Supp. 704 (D.D.C. 1983), motion for reconsideration pending, for example, a district court applying the court of appeals' definition of "intelligence sources" has ordered the disclosure of CIA sources in the Dominican Republic on the basis of judgments such as these: only the Trujillo regime, which has been deposed, would have taken action against these sources (id. at 719 n.50); the current regime is "stable" and "has disavowed all ties with Trujillo's politics, attitudes and methods" (ibid.); since Trujillo and his police chief are dead, the sources need not fear retaliation from individuals sympathetic to Trujillo (id. at 720 n.60); and indeed "(m)any of the sources * * * far from being embarrassed by revelation, might well be thought to be popular, particularly in the Dominican Republic, for having helped, no matter how slightly, to work against a dictator now unpopular and scorned" (id. at 721 n.61). Leading experts on the politics and culture of a society are likely to differ on the extent to which generalizations like these are accurate. Plainly a court should not be in a position of having to make such judgments. /13/ For example, the district court discussed several MKULTRA subprojects involving various kinds of scientific research (Pet. App. 25a-26a) and concluded that the CIA was required to disclose the persons who conducted this research because "(i)t seems clear that such research, which goes on constantly at many places, could have been done without a guarantee of confidentiality" (id. at 26a). The court of appeals specifically approved this reasoning as a correct application of its definition of "intelligence sources." Id. at 5a. It appears that neither court considered the possibility that a foreign power might have been interested in learning that the CIA was conducting research in these areas. /14/ A district court applying the court of appeals' definition has recently interpreted it in precisely this way (Fitzgibbon v. CIA, 578 F. Supp. 704, 716 n.35 (D.D.C. 1983), motion for reconsideration pending): Under (the court of appeals' definition), a document reporting on a conversation between a CIA agent and a source on a wholly innocuous subject would not be protected even if the source is a highly placed official of a government hostile to the United States. Because of its innocuousness, the information presumably could have been obtained from any number of individuals without a promise of confidentiality * * *. We do not acquiesce in this interpretation of the court of appeals' opinion -- the district court in this case, for example, apparently did not interpret the court of appeals' opinion in this way (see Pet. App. 26a) -- but the opinion is susceptible of such a reading. Obviously, even a possibility that such disclosures might occur would seriously damage the government's ability to gather intelligence. /15/ Moreover, the CIA obtains much valuable intelligence from unwitting sources -- sources who reveal information to a person who is, unbeknownst to the source, a CIA contact. A CIA operational officer filed an affidavit in this case in which he used real examples to explain how such sources are used (J.A. 47-48): (a) (Some) (i)ntelligence sources * * * do not realize they are intelligence sources. Such a source might, for example, be a foreign official who regularly discusses official business problems and concerns in candid terms with an old and trusted confident who repeats such information to the CIA. If the original source in such circumstances (learned that he was a source as the result) * * * of an FOIA request, the relationship between the source and his confidant would obviously be be destroyed. A useful intelligence source would likely refuse to provide any additional information and the intermediary would probably suffer some form of retaliation for having betrayed the source's trust. (b) (Some) (i)ntelligence sources * * * do not know they are reporting to the CIA. In such a case, the intelligence source might be knowledgeable of the plans and activities of a terrorist group or a narcotics smuggling ring of which he was a member. He might, in fact, believe he was reporting to an individual who was intent upon negating the efforts of CIA to learn of the plans and activities of such organizations. Should an FOIA request (cause) * * * the source (to learn of his status) * * *, this source would obviously cease providing information. No doubt the intermediary who had passed the information along to CIA would suffer severe consequences. * * * * * Such circumstances described above * * * are neither imaginary nor uncommon * * *. It is unclear how the court of appeals' definition should apply to unwitting sources; the district court ruled that unwitting sources necessarily cannot be kept confidential by the CIA. See Pet. App. 24a ("(T)he fact that the CIA has maintained that even the names of the unwitting researchers must remain undisclosed impeaches its contentions with respect to those researchers who were witting of CIA involvement"). APPENDIX