Michael Alan Crooker v. Bureau of Alcohol, Tobacco & Firearms

WILKEY, Circuit Judge,

dissenting:

Today this court has determined en banc to repudiate the majority opinion in this court’s 1978 en banc decision in Jordan v. United States Department of Justice. 1 The rationale in Jordan — endorsed without reservation by five judges, a majority of the court — was one of strict construction of the exemptions to the Freedom of Information Act (FOIA), consonant with the avowed overall intent of Congress and a long line of our own decisions. The result compelled disclosure of documents whose publication may arguably handicap the Government’s law enforcement efforts, a fact which was openly recognized by both majority and dissent.

Now this court has decreed a remarkable reversal in both rationale and result. Yet the words of the statute remain the same; and the actions of the Congress in enacting and amending the FOIA are undeniable historical facts, as certain in 1978 as in 1981. Indeed, not one additional fact in the legislative history is relied on in Judge Edwards’ opinion, although there is a great deal of new interpretation. The Jordan majority opinion itself has been virtually ignored, as if it were natural and proper for this court to conduct an entirely de novo determination of the issue resolved decisively by a majority in the Jordan case. And, of course, if this court can switch its interpretation of Exemption 2 so completely in three years, it would not be difficult to switch back again in, say, ten years.

What this court is playing is a legislative game. The role it has assumed, years after the event, is that of a conference committee to reconcile the contradictory Senate and House interpretations of Exemption 2. Because of the legislative situation in 1966, such a reconciliation could not be achieved. Only one of these interpretations is correct; to my mind all evidence shows it must be the Senate version, a conclusion reached by virtually all commentators who have addressed the issue, and one we reached in Jordan three years ago. Clearly this court should not attempt to perform the omitted legislative act and work out a compromise which it thinks years after the event might have been acceptable to both houses.

It is not only improper for this court to work out a legislative compromise between the two houses as of 1966, it is also superfluous, because in practical effect that legislative compromise was worked out in 1976 with the enactment of the Government in the Sunshine Act containing the identical language in its Exemption 2 as in Exemption 2 of the Freedom of Information Act. What happened in 1976 was that both houses of Congress adopted the Senate interpretation of the identical phraseology used in both the FOIA and the Sunshine Act. In effect, the House determined that the meaning to be attributed to these words was that specified in the original Senate Report.

If the House is now in accord with the Senate interpretation of the English lan *1094 guage used in Exemption 2 of both the FOIA and the Sunshine Act, then it is the height of presumption for this court to go back to the House position of 1966 and assert that position today over the position of the Senate maintained in both 1966 and 1976. Such an assertion of legislative power, combined with the disregard shown for the decision issued by a majority of this en banc court a mere three years ago, can serve only to undermine the prestige of this court and diminish the respect which will be paid its decisions.

I. ANALYSIS OF EXEMPTION 2

A. The Language of Exemption 2

Exemption 2 provides that mandatory disclosure under the FOIA does not apply to matters that are “related solely to the internal personnel rules and practices of an agency.” 2 In Jordan the majority conducted an in-depth analysis of this language and concluded that its meaning was quite clear:

It is almost impossible to look at this short, simple exemption on its face, “related solely to the internal personnel rules and practices of an agency,” and say that this description was intended to cover the Manual and Guidelines here. The word “personnel” would normally connote matters relating to pay, pensions, vacations, hours of work, lunch hours, parking, etc. — precisely the kind of trivia that was indeed described by the Senate’s comment on the coverage of this particular exemption. 3

Today the majority asserts, without any explanation whatsoever, that “at least at first blush ... the BATF manual appears to be encompassed by the literal language of Exemption 2.” 4 From where comes this new interpretation, directly counter to that of Jordan? Judge Edwards cites two circuit court decisions adopting a similar reading, apparently implying that he is simply following the weight of federal authority. He makes no mention of the other courts which have read Exemption 2 as Jordan did. The Sixth Circuit, for example, held that “the plain import of the (b)(2) language” supported a narrow interpretation and that the “emphasis on personnel rules and practices suggests concern with conditions of employment and the like.” 5

What is worse is that the majority makes no attempt to refute the analysis of Jordan. Given that Jordan found that “personnel” was the real problem for the Government agency seeking to avoid disclosure by virtue of Exemption 2, one might have thought that the majority would confront Jordan’s reading of the word and explain why it rejects that reading. But the majority opinion says nothing about what the specific word “personnel” means in Exemption 2 or why Congress used it, and it fails even to indicate a recognition that the word was central to Jordan’s analysis. The majority merely holds that “the words ‘personnel rules and practices’ encompass not merely minor employment matters, but may cover other rules and practices governing agency personnel, including significant matters like job training for law enforcement personnel.” 6 This detailed definition, it turns out, derives entirely from two passages written *1095 by Judge Leventhal in Vaughn v. Rosen (Vaughn II) 7 and Jordan.

The first passage is used obviously to alter the meaning of the words used by Congress. “Related solely to,” under this novel construction, actually means “related predominantly to.” 8 In other words, the actual language of the exemption turns out to indicate a meaning contrary to that preferred by the majority; so the unambiguous word “solely” is changed to “predominantly,” a much more malleable word that enables a court to support whatever result it finds desirable.

That this construction is wrong was made quite clear by the Supreme Court in Department of Air Force v. Rose. 9 The Court had before it the Vaughn II decision, and it chose to adopt the reasoning of the majority rather than that of Judge Leventhal’s concurring opinion. In determining that the Senate Report better comported with the intent underlying Exemption 2, the Court cited with approval the Vaughn II majority’s finding that the distinction between exempt “ ‘minor or trivial matters’ ” and nonexempt “ ‘substantial matters’ ” was “ ‘a standard, a guide, which an agency and then a court, if need be, can apply with some certainty, consistency and clarity’.... ” 10 This unequivocal Supreme Court endorsement is precisely why Jordan rejected the vague and uncertain standard of “predominant internality” suggested by Judge Leventhal. And not only did Rose adopt this clear interpretation of Exemption 2, it specifically stated that the materials at issue could not be withheld because “[tjhey are not matter with merely internal significance” and “do not concern only routine matters.” 11 Had the Court meant to adopt Judge Leventhal’s standard it obviously would have used the word “predominantly” rather than “merely.” That the Court chose “merely,” which is a close approximation of the actual word “solely,” shows that the Court preferred a literal interpretation. Thus the Court followed the view of the Vaughn II majority. By resurrecting Judge Leventhal’s standard, the majority today has adopted what the Supreme Court in Rose demonstrated is an erroneous interpretation of Exemption 2.

The second passage quoted by the majority is from Judge Leventhal’s concurrence in Jordan :

Exemption 2 is applicable where the document consists of internal instructions to such government officials as investigators and bank examiners. In such a case disclosure would permit circumvention of the law, and there is no substantial, valid external interest of the community at large in revelation. 12

Apparently content that this passage is sufficient explanation, Judge Edwards simply concludes by saying that “[w]e agree with this interpretation of the literal meaning of Exemption 2.” 13

I submit that the latter passage is manifestly not an interpretation of the literal *1096 meaning of Exemption 2. The first sentence simply asserts that Exemption 2 covers internal instructions to government officials, and the second sentence is purely a statement of one view of what the appropriate policy should be, i.e., not to permit disclosure where it would allow circumvention of the law and where there is no important external interest in revelation. There is no hint in Exemption 2 that the goal is to keep instructions secret to prevent circumvention of the law. “Law enforcement” or a similar phrase would convey such an intent, but “personnel” does not. 14 Such concerns can be derived only from certain aspects of the legislative history, not from an analysis of the specific language of Exemption 2. The result is that the majority has side-stepped Jordan’s analysis of the language of Exemption 2. In its place the court has simply inserted, without further elaboration or justification, a different definition which was rejected in Jordan and in Rose.'

B. The House and Senate Reports

1. The conflict between the two reports

The heart of the controversy over the meaning of Exemption 2 is the conflict between the House and Senate Reports on the FOIA. The Senate Report states:

Exemption No. 2 relates only to the internal personnel rules and practices of an agency. Examples of these may be rules as to personnel’s use of parking facilities or regulation of lunch hours, statements of policy as to sick leave, and the like. 15

The House Report’s statement on Exemption 2 is completely different:

Operating rules, guidelines, and manuals of procedure for Government investigators or examiners would be exempt from disclosure, but this exemption would not cover all “matters of internal management” such as employee relations and working conditions and routine administrative procedures which are withheld under the present law. 16

The Jordan majority declared that “rarely can there be found two such contradictory explanations of a statute’s meaning than in the Senate and House Reports.” 17 The Sixth Circuit found that “[t]he disagreement between the houses is total.” 18 Many other commentators have called the two reports “diametrically opposed” 19 and have joined the Jordan majority in determining that the background of the two reports indicates that the Senate Report should be given effect. 20

*1097 In today’s opinion, however, the majority comes up with a startling new conclusion: the two reports are not really contradictory at all. Judge Edwards declares that “[t]he so-called contradiction between the House and Senate Reports . . . exists only with respect to the exemption of trivial employment matters. The House Report’s statement that Exemption 2 permits exemption of more substantive matters — such as ‘manuals of procedure for government investigators or examiners’ — is uncontroverted by the Senate Report.” 21 Because the Senate Report provides “only a non-exclusive list of examples of matters exempt from disclosure,” and because “it does not expressly limit the scope of Exemption 2 to [minor employment] matters,” 22 the majority determines that the two different statements can simply be added together to form one coherent explanation of the intent behind Exemption 2.

This is one of the most remarkable and novel principles of construction I have seen in a long time. Apparently the majority believes that if a report is silent on matter X, then that report fails to create even an inference that matter X is not within the legislative intent — even where the report lists examples of what it does cover which are quite different than matter X. I submit that the implausibility of this view in general is obvious. 23 And applied here it flies in the face of all other existing interpretations of these two reports. 24 The Senate Report states that Exemption 2 relates only to “rules as to personnel’s use of parking facilities or regulation of lunch hours, statements of policy as to sick leave, and the like.” 25 Under the majority’s interpretation it would be perfectly natural for the Report to have contained the phrase “instructions on conducting surveillance” in place of the phrase “and the like.” Perhaps my colleagues believe there is no difference between rules regarding parking facilities, lunch hours, and sick leave and those regarding covert surveillance by federal agents, but I think the difference between the two could not be greater.

Not only does the majority attempt to find no material conflict between the House and Senate Reports, it actually goes on to imply that the House Report should be given greater weight. Judge Edwards asserts that “the Senate gives us less guidance since the Senators made no useful comments during the Senate hearings, and held no debate at all on FOIA.” 26 The implication is that the Senate Report should be given little weight because of this lack of debate. Such a holding is directly contrary to the Supreme Court’s opinion in Rose. The Court noted that “[a]lmost all courts that have considered the difference between the Reports have concluded that the Senate Report more accurately reflects the *1098 congressional purpose,” 27 and it relied upon the Senate Report in reaching its decision. If, as the majority opinion implies, the Senate Report should be given little weight because of the lack of senatorial debate on the issue, surely it should have been disregarded in Rose as well.

Judge Edwards seeks to explain his reliance on the House Report on the ground that “[ujnder the Jordan interpretation, congressional intent expressed in the House is given no weight at all.” 28 To avoid this “extreme approach” he asserts that “the proper route is to try to reconcile legislative intent expressed in the House with that of the Senate.” 29 But under the majority opinion the central thrust of the Senate Report is utterly repudiated. Quite to the contrary of the majority’s interpretation, the clear meaning of the Senate Report is not that minor employment matters are among those exempted under Exemption 2, it is that only such minor matters are covered. No fair reading of the Senate Report can lead to the conclusion that it meant to include investigative manuals within Exemption 2. It is true that under Jordan the House Report is disregarded. But that is a result of the House’s own attempt to amend the meaning of the Senate version of Exemption 2 without actually changing the words of the statute. If the House had intended that Exemption 2 have a different meaning from that given to it by the Senate Report, it should have made changes in the legislation and then attempted to resolve the differences in conference.

2. The background on passage of the FOIA

Crucial to this court’s holding in Jordan was analysis of the background surrounding the issuance of the two committee reports, especially the fact that the House Report was prepared only after the Senate had unanimously adopted the FOIA, including Exemption 2, based on the Senate Report. As the Supreme Court found in Rose, quoting with approval from Vaughn II, “ ‘The second major consideration favoring reliance upon the Senate Report is the fact that it was the only committee report that was before both houses of Congress. The House unanimously passed the Senate Bill without amendment, therefore no conference committee was necessary to reconcile conflicting provisions’ . ...” 30 As Professor Davis has said, “no one will ever know whether the Senate Committee or the Senate would have concurred in the restrictions written into the House committee report.” 31 The Supreme Court in Rose adopted Vaughn IT’s statement of the significance of this fact:

“. . . [W]e as a court viewing the legislative history must be wary of relying upon the House Report, or even the statements of House sponsors, where their views differ from those expressed in the Senate. As Professor Davis said: ‘The basic principle is quite elementary: The content of the law must depend upon the intent of both Houses, not of just one.’ By unanimously passing the Senate Bill without amendment, the House denied both the Senate Committee and the entire Senate an opportunity to object (or concur) to the interpretation written into the House Report (or voiced in floor colloquy). This being the case, we choose to rely upon the Senate Report.” 32

How does the majority respond to this key argument in Jordan, which was specifically adopted by the Supreme Court in Rose ? It only cites Judge Leventhal’s concurring opinion in Vaughn II (five months before Rose), which found this “ ‘a novel *1099 and totally unpersuasive canon of statutory construction.’” 33 Though Judge Edwards does not specifically adopt this view, he appears to agree with it as he cites it directly following his criticism of the Vaughn II majority. And he makes no mention of the fact that Rose specifically quoted the Vaughn II majority with approval on this point, thus necessarily rejecting Judge Lev-enthal’s position. Thus the majority has either taken a position directly contrary to the Court in Rose or, recognizing that the Court’s view refutes its own interpretation of the legislative history, it has chosen to avoid commenting on the analysis entirely.

This analysis becomes even more powerful when it is recognized that “the expansive gloss placed on Exemption 2 and other sections of the Act by the House Report was the product of last minute chicanery by interested members of the House after the Senate had passed the bill and just as the full Committee in the House was about to report out the bill.” 34 Professor Davis has aptly termed this an “abuse of legislative history.” 35 Another commentator noted that the House Report “differed so substantially from its Senate counterpart . . . that it, in effect, amended the bill,” and cited the House Report’s discussion of Exemption 2 as the “most blatant example” of the House’s attempt to restrict the FOIA by committee report. 36 And the most damning evidence of all was provided by the testimony of Benny L. Kass, who was counsel to the Foreign Operations and Government Operations Committee from 1962 to 1965. In 1973 testimony at the Senate hearings on proposed amendments to the FOIA, Mr. Kass explained “why the House report is so different from the rest of the bill”:

The basic reason that the House bill is different was after the Senate passed the Freedom of Information Act and it was about to be reported out of the House Government Operations Committee, the Justice Department — Mr. Katzenbach, Mr. Wozencraft — came up and talked to Congressman Moss and said, look, we cannot support the bill. There are a number of changes that have to be made.

I kind of appeared as an emissary on behalf of the former chairman of this subcommittee to Congressman Moss and I said it is our reading from the Senate that the Senate has already passed this bill twice, that there should be no amendments. We wanted to move forward with it. We have played with it long enough.

So basically what was done under really almost an implied veto — I don’t think they ever specifically said they would veto it but there was an implied threat— we tried to compromise a number of the specific objections into the House report. I don’t think time permits going into these details. I have a very brief analysis which I was going to submit. I have to type it and I will submit it for the record, pointing out where the House kind of gave in to what the Justice Department wanted.

*1100 Mr. Kass then pointed out eight sections in the House Report in which the Justice Department was able to get the language it wanted. Not surprisingly, the seventh area was the Report’s description of Exemption 2. Mr. Kass concluded:

I don’t think it was a sellout but in any event it was really the price of getting the bill. It was my legal advice to both the chairman of this committee and the chairman, Congressman Moss, that the legislative history only interprets and does not vitiate in any way the legislation and that the legislation was strong and was there.

I think this is important just for the record to point out why the House report is different. Fortunately, as Mr. Dobrovir said, there have been a number of cases all of which have said that the House report is so different that we have to look to the statute and that the House report should not in any way undermine the basic statute that was passed by Congress in 1966. 37

How do my colleagues respond to this compelling evidence that the House Report was an effort to change the meaning of the bill the Senate had already passed, not by amending the language but by inserting a contradictory legislative intent? They ignore it. Mr. Kass’ testimony, which was so central to the opinion in. Jordan, is not quoted, cited, or rebutted in any point of Judge Edwards’ opinion. Professor Davis’ strong views condemning the House’s ae-tions as “an abuse of legislative history” are ignored as well. The majority merely asserts that since “the House hearings were completed before the Senate held hearings on and passed S. 1160,” “it is hardly sensible to disregard other consistent indications of House intent occurring after passage of the Senate bill on the speculative grounds that some members of the House engaged in ‘last minute chicanery’ to undermine the Senate’s action.” 38 This use of the word “speculative” highlights the majority’s unwillingness to join issue with the reasoning of the Jordan opinion. Mr. Kass’ testimony, which is not contradicted by the majority nor by the testimony of any other person involved in the congressional activities leading to the enactment of the FOIA, leaves no doubt that such an “abuse of legislative history” did in fact occur. Faced with this uncontroverted fact, so damaging to its position, the majority pretends it does not exist.

3. The significance of the House and Senate hearings

In addition to ignoring evidence supporting the Jordan majority’s rationale, the majority misconstrues the legislative history in attempting to demonstrate that the House actually expressed its intent on Exemption 2 prior to the Senate’s actions. The opinion asserts that “[t]he House hearings unequivocally reveal that Exemption 2 was intend *1101 ed to cover investigatory materials.” 39 When one checks closely the discussions cited by Judge Edwards, one finds that this “unequivocal” intention of “the House” actually derives from a single statement by Congressman Moss: “What [Exemption 2] was intended to cover was instances such as the manuals of procedure that are handed to an examiner — a bank examiner, or a savings and loan examiner, or the guidelines given to an FBI agent.” 40 The majority quotes the statement in text standing alone, as if it indicates beyond all doubt that this was the intent behind Exemption 2. It relegates to a footnote and fails even to discuss the exchange subsequent to Congressman Moss’ statement, an exchange which makes clear that while Congressman Moss may in fact have wanted Exemption 2 to cover such manuals, he recognized that its language would have to be changed to accomplish that purpose:

Mr. Moss. What this was intended to cover was instances such as the manuals of procedure that are handed to an examiner — a bank examiner, or a savings and loan examiner, or the guidelines given to an FBI agent.

Mr. Schlei. Ah! Then the word “personnel” should be stricken. Because “personnel” I think connoted certainly to us the employees relations, employee management rules and practices of an agency. What you meant was material related solely to the internal rules and practices of any agency for the guidance of its employees — something like that.

I do agree that there should be protection for the instructions given to FBI agents and bank examiners; people who, if they are going to operate in expectable ways, cannot do their jobs. Their instructions have to be withheld.

But I think that word “personnel” does not do the job well enough, Mr. Chairman. I am sure it can be done.

Mr. Moss. We will hope to seek a way of doing the job without exempting internal rules and practices.

Mr. Schlei. I suppose that could cover quite a lot of ground, Mr. Chairman.

Mr. Moss. Because I am afraid that we would there open the barn door to everything.

Mr. Schlei. Well, it is one of those things, Mr. Chairman, that just shows how hard it is to cover the whole Government with a few words. There are a number of problems.

Mr. Moss. Oh, we recognize the difficulty and the complexity, but we are perfectly willing to work at it. 41

The Jordan opinion makes clear the significance of this exchange:

[Congressman Moss] apparently wanted investigative manuals covered by the exemption, but he was told flatly that the word “personnel” precluded such interpretation. He acknowledged this, but stated his concern that excising “personnel” would “open the barn door” by leaving a broad exemption for all “internal rules and practices.” 42

The majority’s response is simply that “[s]ince in this ease we are seeking congressional intent, we cannot follow the suggestion [in Jordan] that Mr. Schlei’s interpretation of Exemption 2 is to be given predominant weight over that of a congressman and chief sponsor of the bill.” 43

This reflects an obvious misunderstanding of the significance of Mr. Schlei’s testimony. The end result of the discussion between Mr. Schlei and Congressman Moss was that the latter indicated that he understood that the language of Exemption 2 did *1102 not accord with his preferred interpretation of the purpose of the exemption. He most certainly did not in any way indicate that he thought the present language was satisfactory. Had he done so and thus disagreed with Mr. Schlei, his statement would be much. stronger evidence of the legislative intent underlying the specific words ultimately enacted into law. But all he said, and all that the majority concedes that he said, was that he wanted to protect investigatory records and that he would be “willing to work at it.” Thus this is not at all a question of deciding to give greater weight to the interpretation of a witness than to that of a congressman. It is rather one of relying on the fact that a witness’ interpretation of the exemption was not controverted by the congressman, who rather indicated that he would try to see if changes could be made to correct the deficiencies in the language.

This understanding is made even stronger by the fact that many other witnesses at the House hearings gave uncontroverted testimony that Exemption 2’s language did not extend far enough to protect sensitive internal information necessary to carry out agency functions. The Department of Defense submitted the following commentary on Exemption 2:

Although the second exception for “internal personnel rules and practices of any agency” is desirable as far as it goes, it makes no provision for the many other kinds of internal rules and practices equally deserving of protection and of no legitimate interest outside the agency. Moreover, it raises a question concerning the status of matters which cannot satisfy the requirement of relating “solely” to personnel rules and practices but involving other matters as well. It appears to be the intent of the provision to give no protection to those portions of records which relate to internal rules and practices of an agency when they are mixed with other information. An example of the kind of internal management rule that would receive no protection under section 1(c)(2) of H.R. 5012 is found in DOD Directive 4105.46 which prescribes the permissible price latitudes for DOD negotiators in cost-plus-fixed-fee contract negotiations. The undesirability of making such information generally available is obvious, but H.R. 5012 provides no basis for not doing so. 44

The Treasury Department shared the same basic concern:

Exemption (2). — This exemption reduces the existing exemption for matters “relating solely to the internal management of an agency” to matters related “solely to internal personnel rules and practices.” It reflects the view that all other internal management operations of the Government should be disclosed to any person at all. But internal operations include many matters which are of no public interest or which should not be made readily available, as a few examples will illustrate.

As we read the present exemption, it would not protect the Treasury Department if it refused to detail in advance the method it intended to employ in protecting the movement of currency from the Bureau of Engraving and Printing to its own cash room in the main building. Apparently, upon demand, the Treasury would have to supply the records of how it proposed to use its guard force. . . .

The Department recommends that this exemption be revised to exempt any matter relating solely to the internal management or procedure of an agency. 45

*1103 Both of these departments thus read Exemption 2 as applying only to minor personnel matters. Both gave specific examples of documents that might have to be disclosed even though they contained key instructions to agency personnel as to how to carry out their duties; and the Treasury Department example clearly involves the possibility of disclosure resulting in circumvention of the law. Combined with Mr. Schlei’s responses to Congressman Moss, these statements should have made clear to the House that the language of Exemption 2 did not appear to outside observers to encompass information such as instructions to agency personnel. If there had been anything close to what the majority calls an “unequivocal” indication that the members of the House wanted to exempt such materials, 46 the Representatives would have responded either by asserting that the proposed language did cover them or by subsequently amending the language to do so. They did neither, leading to the conclusion that they agreed with the witnesses’ interpretation of the statute as it then stood— and still stands.

As in the House, testimony in the Senate hearings showed that Exemption 2 was not read by any witness as applying to investigative manuals. Mr. Edwin Rains, the Assistant General Counsel of the Treasury Department, objected to Exemption 2 on the specific ground that it would not apply to these documents:

The second exemption protects matter “related solely to the internal personnel rules and practices of any agency.” This is good so far as it goes. But, we don’t believe that it goes far enough. There are many other internal agency documents which should not be released. The investigative manuals of the Secret Service and the Bureau of Narcotics, for example, contain information the release of which would only be of assistance to criminals in telling them how to plot their crimes so as best to escape detection. Surely such information should be protected. 47

And another independent observer, Mr. Robert Benjamin, Chairman of the ABA Committee on the Code of Federal Administrative Procedure, commented on Mr. Rains’ concern about protecting investigative manuals and stated that “we think that if anywhere besides in specific statutes it should be treated in [Exemption] 7, which has to do with investigative matters.” 48

This testimony is fully in accord with that before the House subcommittee. Mr. Rains specifically stated that Exemption 2 did not cover investigative materials, and Mr. Benjamin agreed that Exemption 2 did not cover them and that reference should instead be made to them in Exemption 7.

My colleagues are forced to recognize the existence of this testimony, but they then proceed to contend that it is irrelevant. Judge Edwards’ opinion dismisses the Plouse testimony on the ground that Congressman Moss’ initial statement on the *1104 purpose of Exemption 2 is controlling. And the Senate hearings are given equally short shrift:

The Senators themselves, however, made no comments on the witnesses’ objections or on the scope of Exemption 2. Consequently, the Senate hearings provide little enlightenment as to Congress’ intent concerning Exemption 2. 49

The majority further asserts:

Of course, a decision not to act on the objections to Exemption 2, especially on those raised by witnesses who opposed the entire bill, does not create an inference that the Exemption would have permitted disclosure of all law enforcement manuals. The Senators may have felt that the existing language of Exemption 2 adequately dealt with the expressed fears of these witnesses. 50

This argument is totally implausible. An inference certainly is created about the meaning of proposed statutory language when a house of Congress fails to change the language in the face of uncontradicted testimony by witnesses that such language would have a given result. On the majority’s view legislative hearings are irrelevant. What witnesses say about the proposed bill can be completely ignored, while a single congressman’s statement about what he would like the bill to accomplish can be seized upon as an unequivocal expression of the meaning to be given to specific statutory words. Judge Edwards has not cited a single person present at either the House or Senate hearings who contradicted the testimony of the many witnesses who declared that the wording of Exemption 2 did not protect investigative manuals. If congressmen had “felt that the existing language of Exemption 2 adequately dealt with the expressed fears of these witnesses,” surely one of .them would have said so. None did, and yet the language was not changed. Its literal meaning should be given effect.

4. The House debates

The majority also claims that the House debates are “highly instructive” in showing that the House intended Exemption 2 to cover investigative manuals. 51 Yet two of the three quotes it relies upon are irrelevant to this specific issue. The statements by Representatives Moss and Dole are general statements to the effect that the bill was attempting, as Representative Dole put it, to take “into consideration the right to know of every citizen while affording the safeguards necessary to the effective functioning of Government.” 52 Such quotes simply state the obvious and prove nothing about the intent underlying Exemption 2.

The third quotation relied upon by the majority does support its position. Representative Gallagher stated that “the bill also prevents the disclosure of other types of ‘sensitive’ Government information such as ... income tax auditors’ manual. . . . Income tax auditors’ manual would be protected under No. 2 — ‘related solely to internal personnel rules and practices.’ ” 53 The majority concludes:

From these brief, but uncontradicted, statements in the House debate on S. *1105 1160, we can see that the members of the House recognized that the unrestricted release of government documents might interfere with the “effective operation of the Government,” and that the nine exemptions were included in S. 1160 to prevent such interference. More specifically and important, supporters of S. 1160 were not challenged in their claim that government investigatory manuals were protected under Examination 2. 54

The majority’s claim that “these statements” by “supporters” of the bill were important is misleading given that there was only one such statement. Only Representative Gallagher, in an extremely brief remark, asserted that income tax auditors’ manuals would be exempt from disclosure under Exemption 2. His remark was not echoed by anyone else and it was not made during any specific debate over Exemption 2. Thus it is not surprising that no one “challenged” the claim. There is no obligation on the part of congressmen to challenge one representative’s isolated statement in order to keep it from becoming the definitive expression of the intent of the entire House. Judge Edwards’ opinion characterizes this single comment as constituting “the express feelings of one house,” 55 when actually the majority of persons who spoke on the subject before the House indicated their belief that Exemption 2 did not cover law enforcement manuals.

C. Section (a)(2)(G) of the FOIA.

Jordan held that section (a)(2)(C), which provides that agencies shall make available for public inspection “administrative staff manuals and instructions to staff that affect a member of the public,” 56 does not operate to exempt law enforcement manuals from disclosure under the FOIA. The majority explicitly affirms that part of the Jordan holding, yet goes on to argue that section (a)(2)(C)’s legislative history supports the finding that Exemption 2 covers such manuals. Judge Edwards notes that “several witnesses expressed their fears that [the phrase “staff manuals”] would mandate release of information that could damage or hinder law enforcement efforts,” 57 and that the Senate Report then recommended adding the word “administrative” to the expression “staff manuals” in an apparent effort to protect law enforcement information. The majority concludes that this indicates “Congress’ deep concern that manuals setting forth guidelines for auditing or inspection procedures should not be released to the public.” 58

In my view the legislative history of section (a)(2)(C) works against the majority’s conclusion. It is ironic that substantial reliance is placed on the testimony of witnesses at hearings on section (a)(2)(C) to the effect that its language would require release of information damaging to law enforcement efforts; the majority dismisses similar testimony about the language of Exemption 2 as indicative of nothing whatsoever. Moreover, the fact that the Senate subcommittee responded to these concerns by adding the word “administrative” and by including specific discussion of the need for protecting law enforcement information in the committee reports only goes to point out that there was an easily available way for Congress to implement its desire to protect law enforcement manuals. In sharp contrast to its actions in response to the testimony on section (a)(2)(C), neither house responded to the many witnesses who complained about the language of Exemption 2. Thus, the fact that “Congress acted specifically to prevent such a result” 59 with re *1106 gard to section (a)(2)(C) provides a strong indication that, in the last analysis, it did not provide similar protection in Exemption 2, as it made no effort to change the latter’s language.

D. The Overall Purpose of the FOIA

In Jordan the majority addressed the fact that the House Report on Exemption 2 cuts strongly against the overall purpose of the FOIA. It noted that “the sweeping interpretation of Exemption 2 favored by the House Report was incompatible with Congress’ expressed intent to cut back on the previous exemption for ‘internal management’ ” and that “the language of the House Report seemed less consonant with the overall scheme and general purpose of the Act than did the Senate Report.” 60 Quoting Vaughn II, the majority in Jordan concluded that “faced with a conflict in legislative history, the recognized principal purpose of the FOIA requires us to choose that interpretation most favoring disclosure.” 61 Other commentators have observed the same thing about the House Report. 62

The basic principle of construction of the FOIA has been enunciated by the Supreme Court, by this court, and by virtually all other courts on so many occasions that it is now quite familiar: disclosure is by far the overriding purpose of the FOIA. Indeed the Supreme Court in Rose quoted with approval the exact same passage relied on by Jordan :

“[W]e have repeatedly stated that ‘[t]he policy of the Act requires that the disclosure requirements be construed broadly, the exemptions narrowly.’ Thus, faced with a conflict in the legislative history, the recognized principal purpose of the FOIA requires us to choose that interpretation most favoring disclosure.” 63

Apparently undaunted by the Supreme Court’s adoption of this principle of construction from Vaughn II, the majority proceeds to criticize it as “too superficial” because “it fails to take into account all of the ‘cross-currents’ of concerns expressed by those members of Congress supporting the enactment of FOIA.” 64 Thus, in direct contrast to the expressed view of the Supreme Court itself, the majority has chosen to emphasize not disclosure but rather the “two crucial bút potentially conflicting interests: the right of the citizenry to know what the Government is doing, and a legitimate but limited need for secrecy to maintain the effective operation of Government.” 65 The phrase “cross-currents” is repeated again and again in Judge Edwards’ opinion, and the need for effective Government operation is held to be “fundamental” to the statutory scheme. 66 The majority finally concludes that “even common sense . . . would seem to belie any suggestion that Congress enacted a statute whose provisions undermined its criminal statutes in *1107 the effectiveness of law enforcement agencies." 67

If the approach taken by the majority — ■ with the emphasis on “common sense” and on the “cross-currents of concern” — had been followed consistently with regard to all nine exemptions, the FOIA would have had little impact on the Government of this nation. Simply put, “common sense” is not the standard for judicial construction of the FOIA. It is an extraordinary strong statute requiring disclosure of information even where balancing of the costs and benefits would indicate to most people that the Government should be permitted to withhold the information. It is the Congress which has made this balancing of costs and benefits, and this court has no higher duty than to carry out what Congress has enacted as law, our personal judgments to the contrary notwithstanding. For this court now to start emphasizing the FOIA’s dual purposes rather than its mandate for disclosure is nothing short of a dramatic revision of the intent underlying the Act as well as of fifteen years of judicial interpretation.

II. SIGNIFICANCE OF LATER CONGRESSIONAL ACTS

A. The Government in the Sunshine Act

The majority in Jordan supported its interpretation of the meaning of Exemption 2 by reference to the action of the House of Representatives in enacting a provision identical to the wording of Exemption 2 as part of the Government in the Sunshine Act of 1976. 68 Jordan stated the following:

The position of this Court in Vaughn II has recently been vindicated by the action of the House of Representatives itself in passing the “Government in the Sunshine Act of 1976.” Professor Davis has recently suggested the relevance of the Sunshine Act to interpretation of the Freedom of Information Act:

The Freedom of Information Act, Advisory Committee Act, Privacy Act, and Government in the Sunshine Act all deal with the subject matter of openness of records and of meetings. Each of the four statutes has its own function. Each provision of each statute may be interrelated to one or more provisions of the other statutes. Furthermore, the various statutes often use language that is identical with the language of another statute. The meaning of that language may depend not only on legislative history and interpretations with respect to the language of the one statute, but it may depend upon legislative history and interpretations with respect to the identical language that is used in one of the other statutes.

Of course, Professor Davis is correct, for it is a well established principle that courts may look to subsequent legislation as an aid in the interpretation of prior legislation dealing with the same or similar subject matter. Indeed, Chief Justice Marshall stated the principle that, if it can be gathered from a subsequent statute in pari materia what meaning the legislature attached to the words of a former statute, this will amount to a legislative declaration of its meaning, and will govern the construction of the first statute.

Applying this principle, it is highly significant that the Government in the Sunshine Act, enacted in 1976, carries over verbatim most of the exemptions in the Freedom of Information Act, including the specific language of Exemption 2. Thus, 5 U.S.C. § 552b(c)(2) exempts from the Act’s open meeting requirement portions of meetings likely to “relate solely to the internal personnel rules and practices of an agency.” The House Report to the Sunshine Act gives the same narrow interpretation to this exemption as the Senate did in 1965:

(2) This exemption includes meetings relating solely to an agency’s internal personnel rules and practices. It is intended to protect the privacy of staff *1108 members and to cover the handling of strictly internal matters. It does not include discussions or information dealing with agency policies governing employees’ dealings with the public, such as manuals or directives setting forth job functions or procedures. As is the case with all of the exemptions, a closing or withholding permitted by this paragraph should not be made if the public interest requires otherwise.

It thus appears that by 1976 the House of Representatives had repudiated the sweeping language concerning Exemption 2 contained in its 1966 report on the Freedom of Information Act. 69

The majority’s only response is to “note that we do not find it particularly significant that Congress enacted in 1976 a provision in the Government in Sunshine Act identically worded to Exemption 2 of FOIA. ...” 70 This is remarkable given *1109 that the Senate Report to the Sunshine Act expressly noted the relationship: “This wording parallels the Freedom of Information Act, 5 U.S.C. 552(b)(2). This exemption does not include directions to agency personnel concerning their responsibility vis a vis the public, such as manuals explaining job functions.” 71 If both the Senate and the House thought that FOIA Exemption 2 covered instruction manuals, is it conceivable that they would use exactly the same language in the Sunshine Act Exemption 2, which quite clearly does not extend to such manuals? Obviously the answer must be no. Yet under the majority’s interpretation either the two identically worded exemptions have precisely opposite meanings or else the Sunshine Act Exemption 2 would have to be interpreted in a manner contradictory to the unequivocal intent expressed in both committee reports. Either result is absurd.

Rather than attempt to contradict this aspect of the Jordan opinion, the majority misconstrues the argument: “We cannot follow the suggestion in Jordan . . . that the House, in an entirely different act, repudiated its understanding of a provision in FOIA. Had the House meant to invest Exemption 2 in FOIA with a different meaning, it would have done so at any of the several times Congress amended FOIA.” 72 This reflects a failure to consider a basic principle of statutory construction which, as Jordan indicated, dates back to the beginning of the American judicial system. The point is that a clear statement by Congress of the meaning of a statutory phrase sheds considerable light on what Congress intended when it used that phrase in an earlier statute. What we have here are four committee reports on one specific statutory phrase, three of which are in basic agreement. By opting for the interpretation contained in the first (and second) Senate Report, the House in 1976 indicated that it agreed that FOIA Exemption 2 did not extend to instruction manuals. Thus it is not a question of investing Exemption 2 with a new meaning, it is one of explaining what Exemption 2 meant in the first place. The majority does not respond to this argument because no response is possible. Congress’ act in 1976 is clear evidence that Jordan was correct and that the majority here is wrong.

B. Section (b)(7)(E) of the FOIA

Judge Edwards’ opinion also relies on the congressional intent he derives from the 1974 amendments to the FOIA, one of which revised Exemption 7 to provide that records need not be disclosed if they are “investigatory records compiled for law enforcement purposes, but only to the extent that the production of such records would . . . disclose investigative techniques and procedures.” 73 The majority points out that the purpose of this amendment was to “force the courts to follow the original in *1110 tent of Congress in enacting FOIA.” 74 It then draws the conclusion that the 1974 amendment showed that Congress wanted to protect law enforcement information, and that since the amendment was characterized as merely embodying Congress’ original intent, Congress necessarily must have desired Exemption 2 also to protect law enforcement manuals.

This argument has several grave defects. First, the majority has stated the history of the 1974 amendments so as to make it seem to have been motivated by a desire to protect law enforcement information. In fact, the direct opposite is true: the 1974 amendments were a response to broad judicial construction of the exemptions. The motivating pressures to amend were thus in favor of more disclosure. Congress’ purpose in 1974 was to limit the law enforcement exemption, not strengthen it, as witness what happened precisely to Exemption 7 itself.

Moreover, there is no necessary connection between a congressional intent to protect law enforcement information under Exemption 7 and a similar intent under Exemption 2. As one commentator has declared, “The fact that Congress decided to protect information on investigatory techniques used in specific investigations does not imply that Congress wished to protect all such information regardless of the context in which it arose.” 75 The majority tries to counter this by finding that Exemptions 2 and 7 necessarily go together: “It would be inconsistent to no small degree to hold that Exemption 2 would not bar the disclosure of investigatory techniques when contained in a manual restricted to internal use, but that Exemption 7(E) would exempt the release of such techniques if contained in an ‘investigatory record.’ ” 76

Far from being “inconsistent to no smáll degree,” such a construction is perfectly logical. Indeed, Senator Hart, the primary sponsor of the 1974 amendment to Exemption 7, introduced during the floor debates a report of a key bar association committee which explains why only techniques in investigative records should be exempt:

The fear that disclosure of investigative technique in general will hinder an agency’s operations appears to be illusory. The methods used for such investigations are widely known and relatively limited in type and scope. The realistic problems are those we have already met — the need to preserve the identity of sources of information in particular cases, the need to assure an impartial trial and to protect reasonable personal privacy. 77

Not only does this indicate that Exemption 7(E) is compatible with the Jordan interpretation of Exemption 2, it provides yet another independent opinion that Exemption 2 does not apply generally to investigatory techniques; if this had been the reasonable explanation, surely the bar committee or some other observer would have discussed the relationship between the two exemptions. 78 They did not for the simple reason *1111 that Exemption 2 was not considered a law enforcement exemption.

The relationship between Exemptions 2 and 7 ultimately proves precisely the opposite of the majority’s argument. I have shown why it is logical to exempt specific investigatory records while allowing disclosure of general investigatory techniques; the majority has not and, I submit, cannot show why Exemption 7(E) is necessary at all if Exemption 2 works to exempt all information regarding investigatory techniques. Under the court’s holding today, all such information can be withheld because it is necessarily internal and has the possibility of allowing circumvention of the law. Accordingly, Exemption 7(E) is redundant, and Congress need not have bothered with a long and heated debate on its amendment. That Congress did so bother suggests that it saw Exemption 7(E) as the chief source of protection for law enforcement information and Exemption 2 as unrelated.

III. CASE LAW ON EXEMPTION 2

A. Department of Air Force v. Rose

In Rose the Supreme Court quoted extensively from this court’s opinion in Vaughn II and decided that the Senate Report should prevail over the House Report in resolving the facts of that case. As Judge Edwards states, “Rose left open the question that faces us in the present case.” 79 He goes on to assert that “the Supreme Court considered as substantial the argument that Exemption 2 might be construed to cover internal agency materials where disclosure might risk circumvention of the law.” 80 It is true that the Supreme Court deliberately avoided deciding the issue, but the opinion in no way implied that the Court would come out as this court has today. On the contrary, as one commentator. has noted,

Rose, while stopping short of applying the Vaughn rationale to law enforcement cases, nonetheless endorsed that reasoning. Even if Rose did not mandate applying the Senate Report to cases involving law enforcement information, the logic of Rose supports such an application. 81

And, as noted earlier, Rose specifically adopted Vaughn II’s narrow reading of the words of Exemption 2 rather than Judge Leventhal’s construction, revived today by the majority, of the word “solely” as meaning “predominantly.” 82 Thus although Rose did not decide the issue before us today, its logic and reasoning are directly in accord with the Jordan majority opinion which has now been repudiated.

B. D.C. Circuit Case Law

There is not much that need be said here. This court’s decision in Vaughn II and its en banc decision in Jordan completely repudiated the House Report in favor of the Senate Report. The subsequent panel decision in Cox v. United States Department of Jus tice, 83 was an attempt by the two dissenters in Jordan to distinguish Jordan in a manner utterly inconsistent with the specific hold *1112 ing of the case. 84 And the decisions in Lesar v. United States Department of Just ice 85 and Alien v. CIA 86 both relied on Jordan and in no way indicated that its reasoning was faulty. The majority’s assertion that Lesar’s “logic and even result . . . seem at odds with that in Jordan ” 87 cannot survive a fair reading of Lesar. The symbols used in FBI records unquestionably were solely of internal significance for administrative convenience and were routine “minor or trivial matters” of no public interest. 88 Jordan, in contrast, involved matters of substantive agency policy that clearly were of substantial public interest.

C. Case Law In Other Circuits

From the majority’s presentation of the Exemption 2 decisions of other circuits one would think that its position had the overwhelming support of the existing case law. Judge Edwards does admit at the end of the discussion that the case law on Exemption 2 is “scattered,” 89 a rather unusual word to describe a situation in which the circuits are split four to two against his position on the issue. He goes on to state that “[i]t is important to note, however, that every circuit that has considered the issue, except the District of Columbia in Jordan, has held that FOIA does not mandate the release of investigatory manuals where disclosure may risk circumvention of agency regulation.” 90 This perhaps better than any other part of the opinion exemplifies the approach to this case taken by the majority — what matters to them is that their result be consistent with that in other circuits, even if many of those cases relied on a rationale that their own opinion repudiates!

That is precisely what has happened in this case. (The majority states that “[w]e are also concerned that every other circuit considering the issue has barred the mandatory release of these materials, whether through the use of Exemption 2 or section (a)(2)(C).” 91 It thus finds significance in the fact that the Sixth and Eighth Circuits have found that section (a)(2)(C) exempts law enforcement manuals from disclosure— even though these courts explicitly repudiated the analysis of Exemption 2 adopted *1113 today and agreed with the analysis set forth in Jordan and even though Judge Edwards’ opinion expressly rejects their interpretation of section (a)(2)(C). This truly is a new step in our jurisprudence. ■ Reliance is placed upon a decision merely because it reaches the same result, even though it is based on an absolutely contradictory analysis of the issues. 92

What should of course be clear is that the only relevance of these precedents is what they say about the meaning of Exemption 2. An analysis of this issue demonstrates not merely a split in the circuits, but that prior to today’s decision four circuits had found that Exemption 2 exempts only minor housekeeping matters, while only two circuits had held to the contrary.

The Fifth, Sixth, and Eighth Circuits have all ruled that Exemption 2 does not apply to instruction manuals. In Hawkes v. Internal Revenue Service 93 the Sixth Circuit found that the Senate and House Reports could not be reconciled and that the Senate Report should be given effect. In a holding similar to that of Jordan, the court ruled that Exemption 2 relates “only to the employee-employer type concerns upon which the Senate Report focused.” 94 This decision was relied upon by the Fifth Circuit in Stokes v. Brennan 95 which required disclosure of a training manual for safety and health officials. 96 And the Eighth Circuit relied on Vaughn II, Rose, and Jordan in two opinions adopting the interpretation contained in the Senate Report and holding that Exemption 2 does not exempt law enforeement manuals. 97

Prior to today’s decision, therefore, the Fifth, Sixth, Eighth, and District of Columbia Circuits had all ruled that the Senate Report should prevail over the House Report and that Exemption 2 relates only to routine employment matters. The opposite conclusion has been reached by two circuits, the Second and Ninth. The major decision was Caplan v. Bureau of Alcohol, Tobacco & Firearms, 98 which, as Judge Edwards notes, “relied heavily on language in the Supreme Court’s decision in Rose.” 99 Moreover, the Second Circuit’s opinion makes clear that its decision was based primarily upon the policy consideration that “individuals embarked upon clandestine and illicit operations” would be enabled through disclosure to frustrate government investigation.” 100 The court’s analysis of legislative history consisted solely of citing the *1114 Senate and House Reports, and its entire reasoning seems to be that the Supreme Court’s refusal to decide the issue in Rose left the Second Circuit “free to give weight to the House Report.” 101 The major independent commentary on law enforcement manuals and Exemption 2 concludes that

the Caplan opinion’s reasoning is suspect. Based upon the Supreme Court’s failure to proscribe use of the House Report in the interpretation of exemption 2 in all cases, the court apparently concluded that it need not justify its preference for the House interpretation of the exemption over that of the State. The opinion thus failed to deal with the arguments for preferring the Senate Report advanced in Vaughn . . . . 102

Thus while Caplan is precedent in support of the majority opinion today, its reasoning is not at all persuasive.

The same is true of the Ninth Circuit’s decision in Hardy v. Bureau of Alcohol, Tobacco & Firearms. 103 The Hardy court engaged in even less analysis of legislative history than Caplan. Like Caplan and the majority opinion today, Hardy’s approach was explicitly result-oriented. The court noted that five circuits had refused to release similar materials but that three different rationales were employed; its response was simply to pick which of the three it preferred, without engaging in any substantial analysis of the background of the FOIA. Indeed, the court did not feel that it needed even to acknowledge that the Caplan interpretation of Exemption 2 was a minority view, rejected by four other circuits. The reason was that, like the majority today, it felt that reaching the preferred result was all that mattered. 104

In sum, even after today’s decision, the circuits are completely split on the meaning of Exemption 2. Only if one accepts the majority’s result-oriented approach can one find it significant that no circuit has yet found that a law enforcement manual must be disclosed. The issue before us is Exemption 2. In rejecting the Jordan rationale, this court will not create a greater consistency among the circuits, but rather will further increase the division that currently exists. If this precedent has any meaning at all, it is that Jordan should have been followed.

IV. THE MAJORITY’S NEW TEST

The majority today holds that “if a document for which disclosure is sought meets the test of ‘predominant internality,’ and if disclosure significantly risks circumvention of agency regulations or statutes, then Exemption 2 exempts the material from mandatory disclosure.” 105 This standard— which is contrary to the standard specifically adopted in Rose 106 — is drawn primarily from Judge Leventhal’s concurring opinions in Vaughn II and Jordan. As Judge Leven-thal saw it, the distinction was between documents that were for predominantly internal purposes and documents that were of legitimate public interest. The majority adopts this test but then immediately asserts that “it is not for the courts to determine when disclosure is in the public inter *1115 est.” 107 This latter statement is certainly right, and it may be possible to avoid this kind of judicial policymaking if the Rose and Jordan approaches are followed and the inquiry is limited by a clear line as to whether the materials are related solely to internal personnel rules and practices. As Rose found, quoting with approval from Vaughn II, the line drawn “ ‘between minor or trivial matters and those more substantial matters which might be the subject of legitimate public interest ... is a standard, a guide, which an agency and then a court, if need be, can apply with some certainty, consistency and clarity’ . ...” 108

In contrast, there is simply no way to employ a fuzzy standard of “predominant internality” without introducing enormous discretion for judges to make their own assessments of the “public interest.” If something is predominantly internal, it therefore has only a limited impact on the public. But this can be determined only if the court makes a value judgment that the public has no important interest in disclosure — -a judgment which this court has repeatedly held has already been made by Congress.

Moreover, the two parts of the majority’s test — both of which must be satisfied if the documents are to be withheld — contradict one another. If certain materials satisfy the test of “predominant internality,” meaning they do not have a substantial external impact, how can their disclosure result in circumvention of agency regulations? It makes no sense to suggest that a member of the public could evade the law by acquiring documents which have no external impact on the public. If the information disclosed is important enough and of the type to permit persons to circumvent the law, it by definition must have a direct external impact on the public.

It is thus apparent that the majority’s standard actually can never be met: a document cannot both be predominantly internal and lead to circumvention of agency regulations, it can only be one or the other. Since the two standards contradict one another, the judge must choose between them. For the majority the choice is clear: the circumvention standard is the crucial one, while the “predominant internality” standard is simply manipulated to support the desired result.

This is made clear by the lengths to which the majority goes in attempting to find that the BATF manual at issue here has only a predominantly internal effect. Judge Edward’s opinion asserts that “[t]he instructions to BATF agents contained in the BATF manual are not written to regulate the public.” Instead, he claims, it “consists solely of instructions to agency personnel. There is no attempt to modify or regulate public behavior — only to observe it for illegal activity.” 109 To find the majority arguing fervently that instructions on covert surveillance techniques for federal agents to use on members of the public do not have an external impact is astonishing. Have my colleagues been asleep for the past ten years while controversy has raged over the investigative practices of the FBI and CIA? Have they failed to observe the controversy over the BATF itself which has been brewing in the past few months, a controversy centering specifically on the *1116 tactics. used by BATF agents? 110 This is not merely a case in which “internal personnel rules and practices have ... a tiny ripple effect outside their area of prime impact.” 111 On the contrary these instructions have an absolutely direct and substantial impact on members of the public.

The distinction drawn between regulating or modifying public behavior and observing it is ludicrous. Do my colleagues mean to say that if this BATF manual contained instructions on how to use illegal bugging techniques to spy on members of the public they would still find that it was of only predominantly internal concern, since all that would be going on is mere “observation”? If, on the other hand, this manual instructed agents on how to make arrests, would my colleagues thereby allow it to be disclosed because this constitutes a “modification” of public behavior, even though disclosure clearly could result in circumvention of the law?

This is not merely a hypothetical example. Both Caplan and Hardy, the two precedents on which the majority relies heavily, involved a request for a BATF manual entitled “Raids and Searches.” 112 The district judge in Caplan “found that the withheld portions of the pamphlet in issue included descriptions of the equipment used by agents in making raids, the methods of gaining entry to buildings used by law breakers, factors relating to the timing of raids, and the techniques used by suspects to conceal contraband.” 113 Surely the majority would not contend that such techniques are mere forms of “observation” rather than methods of “modifying” public behavior. Surely there is not a more direct and extreme form of government regulation of the public than breaking into a building and searching persons suspected of criminal activity. But the majority would have to deny this unless it would decide Caplan and Hardy differently, ruling that disclosure is required because the documents in question involve regulation and modification of public behavior and therefore is not predominantly internal. I cannot believe my colleagues intend to reach such a result, but to avoid it they would have to repudiate the very standard they create today.

This example makes clear that the distinction proposed by the majority is untenable. Presumably it has been used as a means of buttressing the decision to disclose this particular manual and of appearing to place some limits on the breadth of Exemption 2. But I suspect that it is sufficient for the majority that this can be called a “law enforcement manual” and that these are instructions to agency personnel. Pre *1117 sumably the BATF and similar agencies will in the future not make the mistake of disclosing a substantial part of their manuals, as the BATF did here, but will rather withhold everything remotely relating to how agency personnel conduct their jobs, at least where those personnel are responsible for enforcing some law or regulation — a large category indeed.

Of course this does not mean that Exemption 2 will always be applied by this court. This new standard leaves this court with enormous discretion in defining what constitutes a sufficient risk of circumvention of agency regulations and what is predominantly internal, thus enabling the court to pick and choose those documents it wants disclosed and those it wants withheld. Predictably, this judicially discretionary standard may become popular with judges who prefer these public policy choices to be made by courts and not the legislature.

V. THE MAJORITY’S ATTEMPT TO DISTINGUISH THE RESULT IN JORDAN

After repudiating the rationale of Jordan, Judge Edwards’ opinion goes on to assert that the result there would be “identical under the test we announce today.” 114 The majority attempts to support this assertion first by noting that no finding had been made in Jordan that disclosure would risk circumvention of the law. As Judge Ginsburg forthrightly acknowledges in her concurring opinion, however, there is no basis for the implication that release of the documents in Jordan would not help individuals evade detection by law enforcement agents. 115 On the contrary Jordan specifically noted that the Justice Department had warned of the pernicious consequences that would result from disclosure:

Public disclosure of these materials would alert members of the public to those situations, persons, and offenses for which prosecution is withheld, selectively applied, or disposed of by pre-trial diversion. Individuals could then successfully exploit these policies by committing crimes within these select categories, thereby escaping prosecution. For example, publication of a policy of non-prosecution (or prosecution at a lesser degree of seriousness) for possession of certain quantities of specific narcotic drugs would serve only to encourage dealers and users of narcotics to carry lesser quantities of the drug than those specified in our guidelines. 116

Not content to rely on the circumvention distinction, the majority goes on to assert that the materials in Jordan are distinguishable because they were a source of “secret law,” and thus were related to regulation of the members of the public. In contrast, it asserts, the manual in this case relates only to observing public behavior rather than modifying or regulating it. 117 I *1118 find this distinction preposterous. As Judge Ginsburg notes in her concurring opinion, the prosecutorial guidelines in Jordan also did not inform citizens as to how to conduct their activities in conformance with statutory requirements, they simply described how government prosecutors go about enforcing the law. 118 As one commentator put it:

The reasons behind requiring the disclosure of “secret law” — giving members of the public full knowledge of their rights and obligations under the substantive law of the federal government — are inapplicable to enforcement guidelines at issue in Jordan. The guidelines do not inform members of the public of their rights under the law, but rather, their chances of evading the consequences for breaking the law. 119

This attempt to reinterpret Jordan as hinging upon the existence of “secret law” is startling, given that only Judge Bazelon saw fit to raise the point in the Jordan decision. And the attempt to preserve Jordan on this basis is the final and clearest evidence of the majority’s willingness to ignore precedent in implementing its preferred result. Not willing to refute the Jordan rationale directly or to reverse the decision completely, this court has determined to hold that a three-year-old en banc decision was based on a ground that only one out of nine members of the court thought important enough even to mention at the time. So much for the binding weight of our en banc decisions, so much for the respect earned and deserved by this court.

VI. CONCLUSION

Judge Edwards’ opinion for the court expressly repudiates the rationale of the Jordan majority. In this dissent I have argued that the majority’s construction of the language and legislative history of Exemption 2 is erroneous and that the interpretation adopted by the five-judge majority in Jordan should prevail. In conclusion it is worth reiterating a most remarkable fact about Judge Edwards’ opinion: he repudiates the rationale of the Jordan opinion without refuting or even addressing its critical bases. Why doesn’t the majority refute or offer an alternative to the Jordan majority’s definition of “personnel” in Exemption 2? Why doesn’t it challenge or even cite the many commentators who argue, along with the Jordan majority, that the House Report constituted an abuse of legislative history and that its interpretation of Exemption 2 should be disregarded in favor of the interpretation contained within the Senate report? Why doesn’t the majority challenge or offer any. contrary evidence to Jordan’s finding that the House Report was concocted as an effort to placate the Justice Department without having to go through the appropriate process of amending the language of Exemption 2? And why doesn’t it explain why Congress used the same language in Exemption 2 of the Sunshine Act as in the FOIA? (When this court is presented with a Sunshine Act Exemption 2 case, will it follow today’s interpretation of the language, thereby rejecting the express intent of Congress declared in both committee reports? Or will it follow the two committee reports and read the exemption very narrowly, thus holding that Exemption 2 of the Sunshine Act and Exemption 2 of the FOIA — identically worded provisions in two statutes enacted for the same fundamental purpose — have directly contradictory meanings?)

Surely if the rationale of a three-year old majority opinion for an en banc court is to be discarded, there is an obligation on the part of the new majority to challenge directly and refute the analysis of the exist *1119 ing opinion. Yet the court essentially ignores Jordan’s majority opinion during its lengthy discussion of the legislative history, stating that “[t]he legislative history of Exemption 2 has been recounted by this court in [Jordan and Vaughn II],” but that “because the case before us was heard en banc, and because the outcome of the case turns on congressional intent as expressed in the legislative history, we feel that the history merits a fresh review.” 120 Quite obviously neither of these “reasons” explains at all why a “fresh review” is necessary. The fact that this case turns on an interpretation of the legislative history is not a sufficient reason for ignoring a dispositive en banc interpretation of that history; precisely to the contrary, the very purpose of determining issues en banc is to set forth binding precedents for the future. And the fact that this case was heard en banc certainly does not imply that the work of a previous en banc court can be readily discarded. This is especially true given that the majority has not based its reversal on any new evidence bearing on the meaning of Exemption 2, but has rather just rehashed the same evidence that was before the Jordan court in 1978.

The true explanation unfortunately is all too evident from the majority’s opinion. A “fresh review” is undertaken — no matter at what sacrifice to logic, our own precedent, or specific congressional action in 1966, 1974, and 1976 — because applying the Jordan rationale would lead to a result disfavored by the majority. This explains why some of the judges who sat in Jordan so casually have altered their positions in only three years. None of them saw fit three years ago to echo Judge Bazelon’s concern about “secret law”; now they agree it was central to the Jordan decision. The rest of Jordan lies in shambles: the dissenting opinion’s major argument there remains decisively rejected; the “public interest” test which was crucial to Judge Leventhal’s concurring opinion has been rejected, thus rendering his approach unworkable; and the majority opinion has been repudiated. Yet these judges are content to see these opinions they once joined rejected, because today’s decision promises to support the result they prefer as a matter of policy.

I also prefer that result as a matter of policy, but Congress did not agree in 1966 or in 1976, and it is not my place or that of my colleagues to amend Congress’ work. 121 My colleagues have made much of the fact that they find it hard to believe Congress would have intended to allow FOIA to impede law enforcement activities. The answer to that position was given by the Supreme Court in TVA v. Hill, 122 which was similar to this case in that it involved another extremely powerful statute, the Endangered Species Act, that mandated a foolish result, the enjoining of the operation of a virtually completed dam in order to save a small number of three-inch fish. Chief Justice Burger noted the strong argument that “the burden on, the public through the loss of millions of unrecoverable dollars would greatly outweigh the loss of the snail darter,” but he responded that “neither the Endangered Species Act nor Art. Ill of the Constitution provides federal courts with authority to make such fine utilitarian calculations.” 123

This principle applies with equal force here. Respect for the acts of Congress cannot be an on-again off-again affair. It is not for this court to correct the errors Congress has made, no matter how egregious they might be. 124 There is underway in the *1120 Congress an effort to rewrite substantially the Freedom of Information Act, including the problems under Exemption 2 which were recognized but not solved in 1966. As a policy matter, I hope the Congress does a complete job of revision, but as a judge I can never join in the rewriting of this or any other law by this court. 125

*1121 My colleagues’ concern, expressed both in the majority opinion and in Judge Mikva’s concurring opinion,, that “absurd” results would occur if the Jordan rationale were applied to this and similar cases bespeaks a lack of confidence in the ability of Congress to overcome its prior mistakes. The recent experience regarding a case similar to the one before us shows this view to be mistaken. In Long v. Bureau of Economic Analysis, 126 the Ninth Circuit ruled that plaintiffs were entitled under the FOIA to receive copies of computer tapes prepared by the Internal Revenue Service for use with its Taxpayer Compliance Measurement Program, a program designed to aid the IRS in enforcing the income tax laws. When the decision was announced, the Government argued forcefully that release of the data would enable income tax evaders to circumvent the law. While disclosure was stayed pending Supreme Court review, Congress acted swiftly. It attached to a major bill a provision stating that no provision of law “shall be construed to require the disclosure of standards used or to be used for the selection of returns for examination, or data used or to be used for determining such standards, if the Secretary [of the Treasury] determines that such disclosure will seriously impair assessment, collection, or enforcement under the internal revenue laws.” 127 The Supreme Court subsequently vacated and remanded Long for reconsideration in light of the new provision. 128 Surely this is the preferable way of correcting policy mistakes. Congress responded directly to the policy problem and cleared up the ambiguity that had previously existed. The courts were left to construing the law, not determining what it should be.

The test of a defender of free speech is his willingness to defend the right to speak of a person with whom he emphatically disagrees. The test of a judge who has taken an oath to uphold the law is the objectivity with which he interprets and enforces laws of which he thoroughly disapproves. By trying to implement their own vision of appropriate policy, in this case attempting to find a “compromise” between what they perceive as conflicting policy goals, my colleagues have adopted an approach which can only diminish public confidence in the work of our court. 129

. 591 F.2d 753 (D.C.Cir.1978) (en banc).

. 5 U.S.C. § 552 (b)(2) (1976).

. Jordan, 591 F.2d at 763 .

. Majority opinion (maj. op.) at 1056.

. Hawkes v. IRS, 567 F.2d 787, 797 (6th Cir. 1972) ; see Cox v. United States Dep’t of Justice, 576 F.2d 1302, 1309-10 (8th Cir. 1978); Stokes v. Brennan, 476 F.2d 699, 703 (5th Cir. 1973) ; Benson v. GSA, 289 F.Supp. 590, 594 (W.D.Wash.1968) (“plain words-of the statute . . . seem to provide an exemption for material dealing solely with the physical management of the agency’s work force”), aff’d on other grounds, 415 F.2d 878 (9th Cir. 1969) (Exemption 2 issue not argued on appeal).

. Maj. op. at 1056. Judge MacKinnon adheres to his Jordan view that the word “internal” modifies both “rules” and “practices,” but that “personnel” modifies only “rules.” Concurring opinion of Judge MacKinnon at 1078. Not even the majority opinion adopts this interpretation, which Jordan made clear “is violative of basic rules of English grammar, contrary to the legislative history of the exemption, and incompatible with the general purpose of the Act.” Jordan, 591 F.2d at 764 .

. 523 F.2d 1136 (D.C.Cir.1975).

. Maj. op. at 1056-1057 (quoting Vaughn II, 523 F.2d at 1150-51 (Leventhal, J., concurring)) (emphasis added).

. 425 U.S. 352 , 96 S.Ct. 1592 , 48 L.Ed.2d 11 (1976).

. Id. at 365 , 96 S.Ct. at 1601 (quoting Vaughn II, 523 F.2d at 1142 ).

. Id. at 370, 96 S.Ct. at 1603 (emphasis added).

. Maj. op. at 1057 (quoting Jordan, 591 F.2d at 783 (Leventhal, J., concurring)).

. Id. Although Judge Edwards thus relies almost entirely on Judge Leventhal’s interpretation of Exemption 2, he goes on to repudiate the essence of Judge Leventhal’s standard, namely that Exemption 2 is applicable where there is “no substantial, valid external interest of the community at large in revelation.” Jordan, 591 F.2d at 783 (Leventhal, J., concurring). The majority declares that “nor is it for this court to decide which disclosures are in the public interest” and that “we reject language [in Cox and in Judge Leventhal’s concurring opinion in Jordan ] suggesting that the courts are to decide when there is a legitimate public interest in disclosure.” Maj. op. at 1074 n.60. As discussed below, this leaves the majority with a test it cannot apply, except by mere assertion as a means of justifying its desired result. See pp. 1072-1073 infra.

. “[I]f the FOIA’s drafters had intended a complete exemption for law enforcement manuals, they could easily and unambiguously have provided such an exemption in subsection (b). The failure to do so indicates that this was not their intent.” Comment, The Status of Law Enforcement Manuals Under the Freedom of Information Act, 75 Nw.U.L.Rev. 734, 747 (1980) [hereinafter Northwestern Comment].

. S.Rep.No.813, 89th Cong., 1st Sess. 8 (1965).

. H.R.Rep.No.1497, 89th Cong., 2d Sess. 10 (1966).

. Jordan, 591 F.2d at 767 .

. Hawkes v. IRS, 467 F.2d 787, 796 (6th Cir. 1972); see Stokes v. Brennan, 476 F.2d 699, 702 (5th Cir. 1973) (“definite conflict”).

. See Note, The Freedom of Information Act: A Critical Review, 38 Geo.Wash.L.Rev. 150, 154 (1969) [hereinafter Geo.Wash.Note]; Northwestern Comment, supra note 14, at 748; Note, The Freedom of Information Act: Shredding the Paper Curtain, 47 St. John’s L.Rev. 694, 716 (1973) ([hereinafter St. John’s Note], See also Note, The Freedom of Information Act —The Parameters of the Exemptions, 62 Geo.L.J. 177, 183 (1973) (“conflicting” reports); Project, Government Information and the Rights of Citizens, 73 Mich.L.Rev. 971, 1051 (1975) (Senate Report “quite different” from House Report); Comment, The Freedom of Information Act: A Survey of Litigation Under the Exemptions, 48 Miss.L.J. 784, 792 (1977) (“conflicting congressional reports”) [hereinafter Miss.Note].

. See, e.g., J. O’Reilly, I Federal Information Disclosure § 12.03, at 12-6 (1980) (“courts should restrict their reading to the literal text and its initial Senate Report explanation”); Davis, The Information Act: A Preliminary Analysis, 34 U.Chi.L.Rev. 761, 785-86 (1967) (Senate Report “seems fully faithful to the words of the statute,” whereas House Report “tries to change the meaning of the legislative language”); Geo.Wash.Note, supra note 19, at 153-54; Miss.Note, supra 19, at 793; Northwestern Comment, supra note 14, at 758; St. John’s Note, supra note 19, at 717.

. Maj. op. at 1061.

. Id. at 1058, 1065. Judge Edwards also cites passages from the Senate Report which deal with the overall philosophy of the FOIA, and he concludes that “the Senate Report clearly recognized that the broad policy of full disclosure must be tempered in order to protect the ‘operation of our government.’ ” Id. at 1058. Yet this only goes to undermine the majority’s own position. If the Senate Report “clearly recognized” the need to protect Government operations, it surely would not have been silent about this concern in relation to Exemption 2 had Exemption 2 been intended to account for it.

. The majority’s argument runs directly counter to the well-known maxim of statutory interpretation “Expressio unius est exciusio alteri-us.” This maxim “is a product of ‘logic and common sense.’ It expresses the learning of common experience that generally when people say one thing they do not mean something else.” C. D. Sands, 2A Sutherland Statutory Construction § 47.24, at 127 (4th ed. 1973).

. Not even the Second and Ninth Circuits, which have agreed with the majority on the meaning of Exemption 2, felt capable of denying the direct conflict between the House and Senate Reports. See Caplan v. Bureau of Alcohol, Tobacco & Firearms, 587 F.2d 544, 546-47 (2d Cir. 1978); Hardy v. Bureau of Alcohol, Tobacco & Firearms, 631 F.2d 653, 655 (9th Cir. 1980).

. S.Rep.No.813, 89th Cong., 1st Sess. 8 (1965).

. Maj. op. at 1065.

. Rose, 425 U.S. at 363-64 , 96 S.Ct. at 1600-01 .

. Maj. op. at 1075 n.63.

. Id.

. Rose, 425 U.S. at 366 , 96 S.Ct. at 1601 (quoting Vaughn II, 523 F.2d at 1142 ).

. K. Davis, Administrative Law Treatise, § 3A.31 at 175-76 (1970 Supp.), quoted in Jordan, 591 F.2d at 769 .

. Rose, 425 U.S. at 366 , 96 S.Ct. at 1601 (quoting Vaughn II, 523 F.2d at 1142 ) (citation omitted).

. Maj. op. at 1067 (quoting Vaughn II, 523 F.2d at 1148 (Leventhal, J., concurring)). Judge Mikva also adopts Judge Leventhal’s position on this issue, stating that “Judge Leven-thal, whose expertise on issues of legislative history was second to none, thoroughly demonstrated the error in the argument Judge Wilkey tries to make from [the chronology of the reports].” Concurring opinion of Judge Mikva at 1088. Granting Judge Leventhal’s recognized expertise on legislative issues, it nonetheless is strange that Judge Mikva unabashedly contends that Judge Leventhal “thoroughly demonstrated” the error of a particular argument that was accepted by a majority of this court in Vaughn II , by a majority in Jordan , and by a majority of the Supreme Court in Rose.

The majority also states in a footnote that “the problem before us is that the intent of the House and that of the Senate, as expressed in the Committee Reports, are not identical. It is the job of this court to find the interpretation that best comports with the understanding of both houses.” Maj. op. at 1067 n.41. This does not respond to the issue whether the Senate Report should be given “predominant” weight; it simply begs the question.

. Jordan, 591 F.2d at 768 (emphasis in original).

. Id. at 769 (quoting K. Davis, supra note 31, § 3A.31, at 175-76.

. Geo.Wash.Note, supra note 19, at 153.

. Jordan, 591 F.2d at 769 (quoting Freedom of Information, Executive Privilege, Secrecy in Government: Hearings on S. 1142 et al. Before the Subcomms. on Administrative Practice and Procedure and the Subcomm. on Separation of Powers of the Senate Comm, on Judiciary and the Subcomm. on Intergovernmental Relations of the Senate Comm, on Government Operations (Volume 2), 93d Cong., 1st Sess. 122 26 (7, 8, 11, 26 June 1973) (testimony of Benny L. Kass)) (emphasis in original opinion).

. Maj. op. at 1065 n.39. Judge Mikva’s concurring opinion likewise notes the arguments about the House’s abuse of legislative history without saying more than that such arguments indicate “a lamentable tendency to scorn the legislative process.” Concurring opinion of Judge Mikva at 1087. I find this refusal to treat the argument seriously especially ironic coming from Judge Mikva, since the basic thrust of his opinion is that judges should look past the evidence contained in hearings and reports to gain a “reasonable perspective on the legislative process.” Id. at 1088. This is precisely what Jordan did in focusing on the testimony of Mr. Kass as to what actually went on in preparing the House Report. Judge Mikva, like the majority, does not deny that this “abuse of legislative history” occurred. He might have been expected, then, to discuss why this evidence can be ignored or why he thinks it has been misused. But he does neither. It is surprising that his desire to inquire into the realities of the legislative process stops where that inquiry produces evidence that one house has attempted unilaterally to change the meaning of what the other house has enacted.

. Maj. op. at 1059.

. Id. at 1059 (quoting Federal Public Records Law (Part I): Hearings Before a Subcomm. of the Comm, on Government Operations on H.R. 5012 et al., 89th Cong., 1st Sess. 29 (30 Mar.-5 Apr. 1965) [hereinafter House Hearings on H.R. 50121).

. Jordan, 591 F.2d at 766 (quoting House Hearings on H.R. 5012, supra note 40, at 29-30) (emphasis added).

. Jordan, 591 F.2d at 766 (emphasis in original).

. Maj. op. at 1059 n.24 (emphasis in original).

. House Hearings on H.R. 5012, supra note 40, at 219 (letter from L. Niederlehner, Defense Dep’t General Counsel, to Rep. William Dawson, Government Operations Comm. Chairman).

. Id. at 229 (Treasury Dep’t memorandum). The majority asserts that “[ojther witnesses before the House Subcommittee apparently felt that no problem existed under Exemption 2.” Maj. op. at 1059 n.24. These “other witnesses” turn out to be only one witness, Clark Mollen-hoff, who, like Congressman Moss, stated only that Exemption 2 should exempt instructions to law enforcement agents. He did not give his opinion as to the meaning of the proposed language of Exemption 2. See House Hearings *1103 on H.R. 5012, supra note 40, at 151 (statement of Clark R. Mollenhoff).

. Maj. op. at 1059, 1065.

. Administrative Procedure Act: Hearings Before the Subcomm. on Administrative Practice and Procedure of the Comm, on the Judiciary on S. 1160 et al., 89th Cong., 1st Sess. 34 (12-21 May 1965) (statement of Edwin F. Rains [hereinafter Senate Hearings on S. 11601. The Treasury Department also submitted the following criticism:

The exemptions would not protect the agency’s instructions to its staff on methods of law enforcement since the proposed section removes the present exemption for matters of “internal management of an agency” and provides only an exemption for “internal personnel rules and practices” (2). In fact, subsection (b) would require public inspection of staff manuals and instructions to staff “that affect any member of the public.” All Treasury’s law enforcement staff manuals affect members of the public, particularly those who would violate the law and those who would be victimized by such violation. The Narcotics Bureau’s instructions, for example, could hardly be effectively enforced if any narcotics dealer could examine the Bureau’s instructions to Treasury agents.

Id. at 31. See also id. at 439 (letter from Fred B. Smith, Treasury Dep’t General Counsel, to Sen. Eastland, Judiciary Comm. Chairman).

. Id. at 112.

. Maj. op. at 1058. The majority does note one comment by a senator at the Senate hearings on S. 1666, the predecessor to S. 1160:

During the Senate hearings on S. 1666 subcommittee members made only one comment in response to fears expressed by some witnesses that S. 1666 would impede law enforcement efforts. When Paul Dixon, Chairman of the Federal Trade Commission, spoke in opposition to the bill, Senator Long reassured him that: “The Committee, I am sure, and the Congress, wouldn’t want to do anything that would seriously hurt the law enforcement provision of any agency.”

Id. at 1057 (citation omitted). Once again the majority apparently believes that vague and general statements to the effect that Congress did not want to impede law enforcement constitute convincing evidence that the Congress could not have intended Exemption 2 not to apply to investigative manuals. That it is forced to rely on such statements demonstrates the weakness of its position.

. Maj. op. at 1058 n.18.

. Id. at 1061.

. 112 Cong.Rec. 13655 (remarks of Rep. Dole), quoted in maj. op. at 1061.

. Id. at 13659 (remarks of Rep. Gallagher), quoted in maj. op. at 1061.

. Maj. op. at 1061.

. Id.

. 5 U.S.C. § 522 (a)(2)(C) (1976).

. Maj. op. at 1062.

. Id. at 1063 .

. Id. Indeed, if Congress did “act specifically to prevent” disclosure of law enforcement manuals under section (a)(2)(C), it obviously did not then think that the existing Exemption 2 prevented the identical result.

Apparently not realizing that the two arguments analytically cancel each other, Judge MacKinnon, as in his dissent in Jordan , argues that both Exemption 2 and section (a)(2)(C) were intended specifically to exclude law enforcement matters from disclosure. He cites generously from the legislative history of sec *1106 tion (a)(2)(C) and concludes that Congress clearly intended the section to exempt law enforcement manuals and that “there was no necessity for excluding them by Exemption 2.” Concurring opinion of Judge MacKinnon at 10. After conceding this, however, he concludes that Exemption 2 was meant to cover this material as well. Id.

Obviously, if Exemption 2 already took care of the matter, there was no need to make the change in section (a)(2)(C) relied on by Judge MacKinnon’s opinion. Judge Edwards and the majority understand this, and repudiate (as did seven of the nine judges in Jordan) Judge MacKinnon’s argument. Maj. op. at 1055 n.ll. Both Judge MacKinnon and Judge Robb concur in Judge Edwards’ opinion without noting that the whole logic of its Exemption 2 basis is a refutation of the MacKinnon-Robb (a)(2)(C) argument in Jordan and now here.

. Jordan, 591 F.2d at 768 .

. Id. (quoting Vaughn II, 523 F.2d at 1142 ).

. See, e.g., J. O’Reilly, supra note 20, § 12.03, at 12-5 to 12-6; Project, supra note 19, at 1053.

. Rose, 425 U.S. at 366 , 96 S.Ct. at 1601 (quoting Vaughn II, 523 F.2d at 1142 ) (citations omitted).

. Maj. op. at 1067.

. Id. at 1062.

. Id. at 1065.

. Id. at 1074.

. Pub.L.No.94-409, § 3(a), 90 Stat. 1241 (codified at 5 U.S.C. § 552b (1976)).

. Jordan, 591 F.2d at 770-71 (footnotes omitted) (emphasis in original opinion).

. Maj. op. at 1062 n.30. Apparently unsatisfied with the majority opinion’s failure to respond directly to the Sunshine Act arguments, Judge Mikva makes three responses. First, he complains of “the debatable wisdom of construing legislation that is not now before the court.” Concurring opinion of Judge Mikva at 1087 n.1. The answer is that consideration of the Sunshine Act is important because it has a direct bearing on the meaning of the words used in Exemption 2 of the FOIA: “[I]f a subsequent act on the same subject affords complete demonstration of the legislative sense of its own language, the rule which has been stated, requiring that the subsequent should be incorporated into the foregoing act, is a direction to courts in expounding the provisions of the law.” Alexander v. Alexandria, 9 U.S. (5 Cranch) 1 , 7, 3 L.Ed. 19 (1809) (Marshall, C. J.). Judge Mikva’s response is that Exemption 2 of the Sunshine Act is unhelpful because its language “is invoked in entirely different contexts.” Concurring opinion of Judge Mikva at 1087 n.1. But the FOIA and the Sunshine Act are anything but “entirely different.” Both are disclosure statutes designed for the same basic purpose; the drafters of the Sunshine Act clearly recognized the close similarity when they modeled it after the FOIA. Judge Mikva’s third and final point evinces a willingness to conclude that the Sunshine Act Exemption 2 does cover job function manuals, even though both committee reports say precisely the contrary. See id. By asking whether Congress “truly” intended to open meetings discussing investigative manuals he apparently suggests an entirely subjective test under which committee reports can be safely ignored. Indeed, he suggests as much elsewhere in the opinion by asking, “May not judicial notice be taken of the fact that committee reports are often authored entirely by staff members, and that in the rush and flurry of events active congressmen may never have an opportunity to read these reports at all?” Id. at 1088. In a case of clear legislative mistake I might agree, but that is not the case here since the Senate Report is consistent with all the testimony before the Senate committee. In my view, Judge Mikva’s attempt to obtain a “reasonable perspective on the legislative process,” id., is dangerously unconstrained and invites judges to ignore the objective evidence in front of them about what Congress intended.

Judge MacKinnon also responds to the Sunshine Act arguments. He begins by asserting that “the statutory provisions of the Sunshine Act which precede Exemption 2 provide that agency meetings are open ‘Except in cases where the agency finds that the public interest requires otherwise’ ... 5 U.S.C. § 552b(c). This statutory provision provides an obvious basis for an agency to refuse to disclose law enforcement matter that would permit law violators to evade detection.” Concurring opinion of Judge MacKinnon at 1084: This reading of the statute is simply wrong. Section 552b(b) provides that meetings shall be open except as provided in subsection (c), which lists exemptions; subsection (c) contains the further exception, however, that even if an exemption applies the agency may open the meeting if it finds that the public interest so requires. “[Ejven if a matter falls within an exemption, the discussion must be open where the public interest so requires . . . . ” H.R.Rep.No.880, 94th Cong., 2d Sess. 3 (1976). The “public interest” provision thus enables the agency to open a meeting which could be kept closed; it does not, as Judge MacKinnon asserts, provide a blanket public interest exception to the open meeting requirement.

Judge MacKinnon then goes on to assert that my argument is undercut by the fact that Exemption 7 of the Sunshine Act would allow meetings to be closed if the BATF manual here were discussed. In particular he thinks that Sunshine Act Exemption 7(F), exempting law enforcement material that would “endanger the life or physical safety of law enforcement personnel,” would apply to the BATF manual. Consequently, he reasons, Exemption 2 of the Sunshine Act would not have to be read to cover the manual. Id. at 1085.

I am at a complete loss in trying to comprehend the conclusion Judge MacKinnon draws from this: “[Tjhe Sunshine Act can be said to be even more specifically protective of law enforcement matters than the Freedom of Infor *1109 mation Act. The dissent’s reliance on the Sunshine Act is accordingly groundless.” Id. at 1086 (emphasis in original). This conclusion is startling given that Exemption 7 of the FOIA provides exactly the same protection for law enforcement matters as does Sunshine Act Exemption 7: disclosure is not required where it might “interfere with enforcement proceedings,” 5 U.S.C. § 552 (b)(7)(A); id. § 552b(c)(7)(A), “disclose investigative techniques and procedures," id. § 552(b)(7)(E); id. § 552b(c)(7)(E), or “endanger the life or physical safety of law enforcement personnel,” id. § 552(b)(7)(F); id. § 552b(c)(7)(F).

The Sunshine Act provides no more protection for law enforcement information than does the FOIA. Thus, if Exemption 7 of the Sunshine Act demonstrates that Exemption 2 of the Sunshine Act does not cover law enforcement manuals, then FOIA Exemption 7 demonstrates equally well that FOIA Exemption 2 does not cover law enforcement manuals. Indeed, I argue below that reading FOIA Exemption 2 as broadly as does the majority makes FOIA Exemption 7(E) entirely unnecessary. See p. 1111 infra. Perhaps Judge MacKinnon is correct that the BATF manual would have been exempt under the language of Exemption 7(F), but that issue is not before us because the Government did not, for whatever reason, choose to defend its refusal to disclose on the basis of Exemption 7.

. S.Rep. 354, 94th Cong., 1st Sess. 21 (1975).

. Maj. op. at 1062 n.30.

. 5 U.S.C. § 552 (b)(7)(E) (1976).

. Maj. op. at 1064 (emphasis in original).

. Northwestern Comment, supra note 14, at 763.

. Maj. op. at 1065.

. 120 Cong.Rec. 17034 (1974) (quoting Report of the Committee on Federal Legislation of the Association of the Bar of the City of New York, Amendments to the Freedom of Information Act) (emphasis in original); see Northwestern Comment, supra note 14, at 763 n.147.

. See Reply Brief for Plaintiff-Appellant on Rehearing En Banc at 4 n.2 (“If the two exemptions are as closely related in purpose as the government suggests, one would expect some discussion of that interrelationship in the relevant legislative history. Significantly, however, there is not a single reference to exemption 2 in the legislative history of exemption 7(E).”).

Congress’ belief that Exemption 2 is unrelated to law enforcement information is further evidenced by the conference report on Exemption 7, which Judge Edwards quotes without discussion:

The conferees wish to make clear that the scope of this exception against disclosure of “investigative techniques and procedures” should not be interpreted to include routine techniques and procedures already well known to the public, such as ballistics test, fingerprinting, and other scientific tests or commonly known techniques. Nor is this exemption intended to include records falling *1111 within the scope of subsection 552(a)(2) of the Freedom of Information law, such as administrative staff manuals and instructions to staff that affect a member of the public.

Maj. op. at 1064 n.36 (quoting H.R.Rep.No. 1380, 93d Cong., 2d Sess. 12-13 (1974)). This demonstrates both a desire to limit the investigative law enforcement exemption and a recognition of the relationship with section (a)(2)(C); Exemption 2 is not mentioned because Congress considered it irrelevant to the subject matter.

. Maj. op. at 1066. Earlier in the opinion the majority contradicts itself, asserting that Rose did in fact decide this issue: “By noting that ‘the primary focus of the House Report was on exemption of disclosures that might enable the regulated to circumvent agency regulation,’ the Supreme Court in Rose recognized that the conflict between the committee reports exists only as regards minor employment matters.” Id. at 1061 n.27. This assertion is astonishing given that the Supreme Court quite clearly intended to avoid altogether the issue of which report should control in a case of circumvention of agency regulations.

. Id. at 1066.

. Northwestern Comment, supra note 14, at 753.

. See p. 1095 supra.

. 601 F.2d 1 (D.C.Cir.1979).

. Cox stated that:

The exemption covers those portions of law enforcement manuals that prescribe the methods and strategy to be followed by law enforcement agents in the performance of their duties. See Ginsburg, Feldman & Bress v. Federal Energy Administration, Civ.No. 76-0027 (D.D.C. June 18, 1976) aff’d per curiam by an equally divided court, 192 U.S.App.D.C. 143 , 591 F.2d 752 (1978) (en banc), cert. denied, 441 U.S. 906 , 99 S.Ct. 1994 , 60 L.Ed.2d 374 (1979). It does not apply to “secret law” contained in the rules and practices by which an agency regulates its own staff. Jordan v. United States Department of Justice, 192 U.S.App.D.C. 144, 154 , 591 F.2d 753, 763 (1978) (en banc).

Id. at 4. This was a total misstatement of the existing law. The rationale of the Jordan majority had nothing to do with “secret law”; and the Ginsburg decision was merely one of the district court, since the panel opinion by Judge MacKinnon had been vacated and the district court opinion had been affirmed by an equally divided court. Moreover, the Jordan majority had expressly repudiated the reasoning of the vacated panel opinion in Ginsburg, see Jordan, 591 F.2d at 763-64 & n.37. Cox therefore should not have relied on Ginsburg, especially since Jordan was an en banc and therefore controlling decision.

Nonetheless, Judge MacKinnon insists on citing the panel opinion in Ginsburg as valid authority. He asserts that Judge Leventhal would have voted with the majority in Ginsburg and that this makes the Ginsburg panel opinion “entitled to substantial consideration.” Concurring opinion of Judge MacKinnon at 1077. This is extrajudicial speculation of the worst kind. Judge Leventhal recused himself in Ginsburg. Judge MacKinnon ignores this and concludes that since he thinks he knows how Judge Leventhal would have voted, we can treat the Ginsburg panel opinion as having been affirmed. I am appalled at this attempt to construct binding precedent on the basis of the probable vote of a recused, and now late, judge.

. 636 F.2d 472 (D.C.Cir.1980).

. 636 F.2d 1287 (D.C.Cir.1980).

. Maj. op. at 1069-1070.

. Rose, 425 U.S. at 365 , 96 S.Ct. at 1601 (quoting Vaughn II, 523 F.2d at 1142 ).

. Maj. op. at 1072.

. Id.

. Id. at 1074.

. That the majority’s approach is explicitly result-oriented is made clear by its decision to organize the discussion of the case law in other ■ circuits into “two categories: those that have exempted law enforcement investigatory manuals under section (a)(2)(C); and those that have used Exemption 2 to exempt internal law enforcement manuals whose disclosure would risk circumvention of the law.” Maj. op. at 1070. Were the court really interested in how the other circuits interpreted Exemption 2 it would have divided the cases into those that have exempted law enforcement manuals under Exemption 2 and those that have not.

. 467 F.2d 787 (6th Cir. 1972).

. Id. at 797 .

. 476 F.2d 699 (5th Cir. 1973).

. Id. at 701-02 . In a later decision another panel of the Fifth Circuit found that disclosure of a certain manual would not impede law enforcement efforts and thus stated that “there is no need for us to choose between the Second Circuit’s and the D.C. Circuit’s interpretation of Exemption 2 when disclosure would risk circumvention of the law.” Sladek v. Bensinger, 605 F.2d 899, 902 (5th Cir. 1979). Under Stokes , however, the choice necessarily would be to follow Jordan.

. Cox v. United States Dep’t of Justice, 576 F.2d 1302, 1309-10 (8th Cir. 1978); Cox v. Levi, 592 F.2d 460, 462 (8th Cir. 1979). See also Kuehnert v. FBI, 620 F.2d 662, 667 (8th Cir. 1980) (expressing doubt that FBI “information for investigative lead purposes” falls under Exemption 2).

. 587 F.2d 544 (2d Cir. 1978).

. Maj. op. at 1071. Such heavy reliance on Rose, of course, is misplaced since the Supreme Court’s opinion leaves the issue open and, if anything, suggests that the Jordan interpretation is correct. See p. 1111 supra.

. Caplan, 587 F.2d at 547 . This, of course, is an impermissible concern since the FOIA permits every person, regardless of motive, to take advantage of disclosure. Unfortunately, Judge Mikva makes the same error in his concurring opinion. See note 124 infra.

. Caplan, 587 F.2d at 547 .

. Northwestern Comment, supra note 14, at 752 (emphasis added). The article went on to note that “[t]he court in Caplan did not justify its conclusion that the House Report, despite its relative vagueness and insulation from scrutiny by the Senate, was preferable to the Senate Report; it was no answer to point out that the Supreme Court had not conclusively ruled that the Senate Report must be preferred.” Id. at 753.

. 631 F.2d 653 (9th Cir. 1980).

. See id. at 656-57 . The court did acknowledge that courts in the District of Columbia Circuit and the Eighth Circuit had arrived at a contrary interpretation of Exemption 2, but it made no mention of the fact that courts in the Fifth and Sixth Circuits had come to a similar conclusion. See id. at 656 . The court also made no mention of the fact that a district court in its own circuit had determined ten years earlier that the Senate Report’s interpretation should be given effect over that of the House Report. Benson v. GSA, 289 F.Supp. 590, 594-95 (W.D.Wash.1968), aff’d on other grounds, 415 F.2d 878 (9th Cir. 1969).

. Maj. op. at 1074.

. See pp. 1095 supra.

. Maj. op. at 1056 n.12; see id. at 1074 n.60.

. Rose, 425 U.S. at 365 , 96 S.Ct. at 1601 (quoting Vaughn II, 523 F.2d at 1142 ). The Supreme Court has also held that the nine exemptions “are plainly intended to set up concrete, workable standards for determining whether particular material may be withheld or must be disclosed.” EPA v. Mink, 410 U.S. 73, 79 , 93 S.Ct. 827, 832 , 35 L.Ed.2d 119 (1973) (emphasis added). The majority’s standard is anything but concrete or workable.

. Maj. op. at 1073, 1075 (emphasis in original). At a different point Judge Edwards states that “it is uncontroverted that the disputed rules here have been developed predominantly for internal uses.” Id. at 1073. If this is the meaning of the “predominantly internal” standard, then there actually is no standard at all. Every document prepared within an agency to be given to employees of the agency would be exempt because it was “developed predominantly for internal uses” rather than meant for external publication. Such a definition would reduce the FOIA to uselessness.

. The National Rifle Association recently came out with “a half-hour television documentary that characterizes the agency as abusive government stormtroopers,” a documentary in which a Democratic congressman describes the Bureau “as a ‘jack-booted group of fascists who are ... a shame and a disgrace to our country.’ ” Earley, NRA Aims at U.S. Agency, Wash.Post, 23 July 1981, at Al, col. 1. The BATF calls these charges grossly misleading, see id. at A34, col. 1, but their accuracy is not the question here. What is important is that there is undeniably an important public controversy about the tactics of BATF agents. See also Reply Brief for Plaintiff-Appellant on Rehearing En Banc at 3 (citing studies criticizing methods of BATF agents).

Judge Edwards quotes the concurring opinion in Vaughn II , which argued that Exemption 2 was based in part on “ ‘a recognition that management of government needs some elbow room in developing and revising internal practices, so as to achieve efficiency, without becoming embroiled in continuous public discus- sión.’ ” Maj. op. at 1073 n.58 (quoting Vaughn II, 523 F.2d at 1150 (Leventhal, J., concurring) (emphasis added)). Even if this argument were correct, it obviously is inapplicable here. The BATF is already “embroiled in continuous public discussion” because of its controversial tactics, and exempting the manual in this case will not alter this situation.

. 523 F.2d at 1151 (Leventhal, J., concurring). See Ferguson v. Kelley, 448 F.Supp. 919, 922 (N.D.Ill.1977) (FBI leads and status information not exempt under Exemption 2 because they “may reveal how the FBI proceeds in investigating a member of a ‘subversive organization,’ the Socialist Workers Party, — a matter of genuine and significant interest to the public.”) :

. Hardy, 631 F.2d at 655 ; Caplan, 587 F.2d at 545 .

. Caplan, 587 F.2d at 545 .

. Maj. op. at 1074.

. See concurring opinion of Judge Ginsburg at 1091.

. Jordan, 591 F.2d at 758 .

. Maj. op. at 1075. Judge Mikva also reinterprets the basis of Jordan. He quotes at length from Judge Bazelon’s opinion in Jordan, and is now prepared to read that opinion’s brief discussion of “secret law” as the true holding of the case. Concurring opinion of Judge Mikva at 1086. Like the majority opinion, Judge Mikva’s opinion fails to answer Judge Ginsburg’s demonstration that the attempt to distinguish the results in Jordan and this case is a failure. And like the majority opinion, his opinion does not find it the least bit troublesome that an en banc majority opinion is repudiated and replaced by the views of a single judge in that case. This, of course, will serve only to encourage judges to express individual views even when joining those of another, since they can hope that a later court will reinterpret what the original decision meant.

Ironically, Judge MacKinnon also now finds “secret law” to have been important in Jordan, even though he dissented without even mentioning the point. Concurring opinion of Judge MacKinnon at 1076. He goes even further, asserting that because Judge Bazelon wrote a separate opinion, “there was never a majority of the court in full support of the entire rationale of Jordan." Id. This provides yet another example of how the judges in the majority are willing to distort the basic principles of jurisprudence under which we have operated for so long, solely for the purpose of saving themselves from acknowledging forthrightly that they are overruling a recent en banc deci *1118 sion. Judge Bazelon joined the opinion of the court without stating any reservations whatsoever about its rationale. His decision to make some additional comments did not diminish the meaning of his adherence to the majority opinion, unless Judge MacKinnon wishes now to rewrite the settled practices of our court.

. See concurring opinion of Judge Ginsburg at 1092.

. Northwestern Comment, supra note 14; at 758.

. Maj. op. at 1058 n.15.

. Congress intended “to allow only the most limited protection for such manuals. The reasons behind this intent are not entirely clear, but, its existence suggests that courts should not substitute their policy judgments for those properly made by Congress.” Northwestern Comment, supra note 14, at 737.

. 437 U.S. 153 , 98 S.Ct. 2279 , 57 L.Ed.2d 117 (1978).

. Id. at 187 , 98 S.Ct. at 2298 .

. Judge Mikva purports to agree with this proposition, but then says that “neither may we ascribe to Congress errors that it clearly did not intend.” Concurring opinion of Judge Mikva at 1089. Throughout his concurrence *1120 Judge Mikva repeatedly asserts, without elaboration, that Congress could not have intended to permit disclosure of investigative manuals. He finds it “patent,” id. at 1087, and “irrefutable,” id. at 1086, that Congress did not intend “such an absurd and unreasonable result,” id., and he urges that this result would make “scofflaws of the Congress, or at least some of its members,” id. No explanation is given other than that Crooker “is a federal prisoner,” id. at 1086, and “an adjudged felon” who seeks to use FOIA “for his own ends,” id. at 1090. Yet surely Judge Mikva is aware that the FOIA is available without exception to any person who seeks to invoke it; the fact that Crooker or any other information requester is a criminal is irrelevant to the analysis.

Apparently Judge Mikva feels qualified to decide when a statute’s policy is erroneous and therefore invalid; the majority of the Supreme Court expressed the opposite view in TVA v. Hill, as noted above. And he also feels that it is unnecessary even to respond to arguments showing that the policy has a logical and reasonable basis. During the 1974 debates on amending the FOIA, Senator Hart introduced the argument that information on general investigative techniques need not be kept secret even though information on specific investigations should be made exempt. See p. 1110 & note 77 supra. Judge Mikva has not attempted to counter this argument. Moreover, he declares that the BATF manual in question here falls in the category of trivial matters in which there is no public interest in disclosure, see concurring opinion of Judge Mikva at 1087, without commenting on the incontrovertible fact that the subject of investigative techniques of BATF agents is in fact one of heated public controversy and debate. See pp. 1115-1116 & note 110 supra. Far from being absurd, then, disclosure of this manual might well contribute to the resolution of this public issue.

More generally, Judge Mikva argues that the views of “many distinguished judges” demonstrate that my analysis must be wrong. Concurring opinion of Judge Mikva at 1090. Although perhaps Judge Mikva believes that a mere mortal judge like myself should fall silent in the face of allegedly contrary views of judges who are distinguished, I venture to suggest that the many passages he cites are inapplicable to the facts of this case. I am in hearty agreement that courts should not reach ridiculous results, but I strongly disagree with Judge Mikva’s apparent view that attempting to judge the reasonableness of a statute is the most basic principle of statutory construction. Judges should attempt to find the congressional intent by considering the objective evidence before them. Certainly they should use sound judgment in assessing that evidence, but only in rare and extreme circumstances should they override that evidence in order to implement a more reasonable, from their perspective, policy result. Judge Mikva appears to advocate frequent resort to his principle of judging what is reasonable policy and implementing the same. Indeed, in this case he has joined the majority in ignoring — not refuting, but simply ignoring — critical evidence concerning the actions of the House in preparing the House Report. See p. 1100 & note 38 supra. The only conclusion I can reach is that determining what the “reasonable” policy should be is, under the princi- ■ pies of the majority, by far the overriding concern. I am confident that this is far from what Justices Frankfurter, Reed, Field, Harlan, and Brewer had in mind. See concurring opinion of Judge Mikva at 1089-1090.

. Judge Mikva suggests that I am the one who seeks to legislate here and that by stating my personal views on the substantive policy matter at issue I have exceeded “the bounds of the judicial role.” Id. at p. 1090. By implying that I have reached my decision so as to cause Congress to reach a certain legislative result, Judge Mikva ignores the unequivocal record of my views on Exemption 2. In Vaughn II (1975), Jordan (1978), and again today (1981) I have adhered to a consistent interpretation of the meaning of the words and legislative history of Exemption 2. Those views lead inexorably to the result I would reach in this case. Had I been concerned about reaching a certain legislative result I never would have interpreted Exemption 2 as I did; for Judge Mikva now to imply that that is my goal is rather odd.

Judge Mikva also implies that it is improper for a judge to state publicly his views on a matter of legislative policy. If so, this is a strikingly new conception of the judicial role. American judges have long spoken out on public issues, not only in articles and speeches, but in their judicial opinions as well. See, e.g., Furman v. Georgia, 408 U.S. 238, 369 , 92 S.Ct. 2726, 2792 , 33 L.Ed.2d 346 (1972) (Marshall, J., concurring) (“Assuming knowledge of all the facts presently available regarding capital punishment, the average citizen would, in my opinion, find it shocking to his conscience and sense of justice.”); id. at 375 , 92 S.Ct. at 2796 (Burger, C. J., dissenting) (“If we were possessed of legislative power, I would either join with Mr. Justice BRENNAN and Mr. Justice MARSHALL or, at the very least, restrict the use of capital punishment to a small category of the most heinous crimes.”); id. at 405-06 *1121 (Blackmun, J., dissenting) (“I yield to no one in the depth of my distaste, antipathy, and, indeed, abhorrence, for the death penalty .... Were I a legislator, I would vote against the death penalty .... ”); United States v. DeCost er, 624 F.2d 196, 299-300 (D.C.Cir.1976) (Bazel-on, J., dissenting) (stating that “the problem of inadequate representation of the indigent cannot be solved by the courts alone” and recommending steps to be taken by the bar and the legislatures). Indeed, it is ironic that Judge Mikva makes this criticism, since his own opinion expresses the unequivocal view that it would be absurd for a legislature to vote to permit disclosure of investigative manuals, a view disputed by arguments introduced by the late Senator Hart, one of the principals in the FOIA debate. See p. 1110 & note 77 supra.

. 646 F.2d 1310 (9th Cir. 1981).

. Economic Recovery Tax Act of 1981, Pub. L.No.97-34, § 701(a), 95 Stat. 350 (codified at 26 U.S.C.A. § 6103 (b)(2) (West Supp.1981)).

. 450 U.S. 1301 , 101 S.Ct. 1073 , 67 L.Ed.2d 322 (20 Oct. 1981).

. A last minute deletion by Judge MacKin-non has made note 70, paragraph 2, supra, no longer relevant. Note 70, paragraph 4, supra, is also affected by a late revision, as he now apparently agrees that FOIA Exemption 7 has the same scope as Sunshine Act Exemption 7. See concurring opinion of Judge MacKinnon at 1085. Unfortunately, his new conclusion is as untenable as his old one. If Exemption 7 of the Sunshine Act would exempt the BATF manual at issue here, and if the FOIA and the Sunshine Act are both “specifically protective of law enforcement matters,” id. at 1085, then my point in dissent must be right: protection for law enforcement materials was not intended to arise from Exemption 2 in either statute, but from Exemption 7. Judge MacKinnon’s argument is now a complete non sequitur.

Two other final additions by Judge MacKin-non are equally erroneous. First, his contention that “[t]he Sunshine Bill Report, dated 1976, reflects the same point as Judge Leven-thal’s 1978 interpretation of ‘predominant inter-nality,’ ” see concurring opinion of Judge MacKinnon at 1081, is clearly wrong. He ignores the fact that Judge Leventhal first set forth his “predominant internality” standard not in 1978, but in 1975 — before the Sunshine Act was adopted. See Vaughn II, 523 F.2d at 1150-51 (Leventhal, J., concurring). This stan *1122 dard was available to the Supreme Court in Rose, but the Court rejected it by adopting the narrower construction of the Vaughn II majority. See p. 1095 supra. This Rose interpretation, based on the Vaughn II majority opinion, was specifically adopted by the Sunshine Act Report, which stated that Exemption 2 was adopted “with recognition of the Supreme Court’s interpretation of the analogous Freedom of Information Act exemption in Department of Air Force v. Rose.” S.Rep.No. 1178, 94th Cong., 2d Sess. 15 (1976) (conference report). It therefore is nonsense to argue that the Sunshine Act Report in any way reflects the reasoning advocated by Judge Leven-thal. Second, the argument that manuals of law enforcement instructions are not encompassed within “discussions or information dealing with agency policies governing employees’ dealings with the public, such as manuals or directives setting forth job functions or procedures,” see concurring opinion of Judge Mac-Kinnon at 1081, is obviously implausible.