(concurring):
I concur with much of Chief Judge Kaufman’s opinion, his broad answer to the certified question and the overall judgment. Because this case breaks new ground in an area of utmost importance, it warrants setting forth the somewhat different First Amendment analysis I use to reach the ultimate result, even at the risk of some repetition. In the process I will also set forth my own slightly more detailed views on the approach that should be taken by the district court on remand, for whatever guidance they may supply.
I
As we know, in New York Times Co. v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964), the Supreme Court “constitutionalized” the law of defamation by subjecting it to requirements deriving from and implicit in the First Amendment.1 *985In cases involving publication of matter pertaining to public affairs and involving public officials — as this case does2 — to be liable a defendant must have acted with “actual malice”3 as constitutionally defined. Sullivan contemplates not only that the alleged defamatory statements are false but that the libel defendant knew that they were false or made them with reckless disregard of their truth or falsity. Gertz v. Robert Welch, Inc., 418 U.S. 323, 342, 94 S.Ct. 2997, 41 L.Ed.2d 789 (1974); New York Times Co. v. Sullivan, supra, 376 U.S. at 279-80, 84 S.Ct. 710. The appropriate standard is whether “the defendant in fact entertained serious doubts as to the truth of his publication.” St. Amant v. Thompson, 390 U.S. 727, 731, 88 S.Ct. 1323, 1325, 20 L.Ed.2d 262 (1968). In this respect, ill will toward the plaintiff, bad motive, hatred, spite or even desire to injure — malice in the traditional as opposed to the constitutional sense — is not involved. See Old Dominion Branch No. 496, National Association of Letter Carriers v. Austin, 418 U.S. 264, 281, 94 S.Ct. 2770, 41 L.Ed.2d 745 (1974); Buckley v. Littell, 539 F.2d 882, 889 (2d Cir. 1976), cert. denied, 429 U.S. 1062, 97 S.Ct. 785, 50 L.Ed.2d 777 (1977).
This case presents the broad question whether in a case involving allegations of actual malice discovery should be “liberal” as provided generally by the Federal Rules of Civil Procedure4 and as held by the trial judge, or should be restricted in certain ways. The restriction urged on us specifically is thaf matters of “editorial process”5 should not be discoverable at all or only under certain limitations. The exact relief requested remains somewhat unclear since appellant’s brief would have us remand to the district court “with instructions for that Court to redetermine the issues raised by the motions [pursuant to Rule 37 of the Federal Rules of Civil Procedure], giving due regard to First Amendment considerations as set forth by this Court.” Brief for Appellant at 8, 32-33. Within these broad parameters we are invited in this extraordinary though not unique interlocutory appeal on a discovery order,6 to set some limits in Sullivan cases on the untrammeled, roving discovery that has become so prevalent in other types of litigation in today’s legal world. Not without the doubt that any venture on untrod paths may bring, I am willing to join in accepting the invitation.
*986At the outset, there are familiar landmarks. There is, for example, no necessary internal inconsistency between a First Amendment limitation on compelled discovery in Sullivan cases on the one hand and liberal rules of civil procedure on the other.7 The rules of civil procedure expressly contemplate limitations in at least two areas. First, where discovery would result in “oppression” of or “undue burden” on a person whose deposition is being taken, a court may limit or even forbid discovery.8 Certainly, therefore, a plaintiff’s attempt to prove actual malice may be restrained where the discovery he seeks satisfies the oppression standard of Rule 26(c). And second, Rule 26(b)(1) excepts privileged matters from compelled discovery.9 Not surprisingly, a privilege insulating journalists’ confidential sources from compelled discovery in civil litigation has been recognized by this court and others. Baker v. F & F Investment, 470 F.2d 778 (2d Cir. 1972) (civil rights litigation), cert. denied, 411 U.S. 966, 93 S.Ct. 2147, 36 L.Ed.2d 686 (1973); Cervantes v. Time, Inc., 464 F.2d 986 (8th Cir. 1972) (public figure libel litigation), cert. denied, 409 U.S. 1125, 93 S.Ct. 939, 35 L.Ed.2d 257 (1973); see Apicella v. McNeil Laboratories, Inc., 66 F.R.D. 78 (E.D.N.Y.1975); Loadholtz v. Fields, 389 F.Supp. 1299 (M.D.Fla.1975). But see Dow Jones & Co. v. Superior Court, 364 Mass. 317, 303 N.E.2d 847 (1973). The limitation on discovery of journalists’ sources suggests by analogy that other First Amendment limitations on discovery in Sullivan cases may be similarly appropriate. The focus of inquiry thus becomes whether the First Amendment protects matters which constitute the editorial process from compelled discovery and, if so, the extent of that protection.
II
I agree with the Chief Judge that compelled discovery of the editorial selection process implicates the First Amendment. • I arrive at this position not on First Amendment grounds generally but in light of what seems to be the Supreme Court’s evolving recognition of the special status of the press10 in our governmental system and the concomitant special recognition of the Free Press clause of the First Amendment. Mr. Justice Stewart, in a seminal speech at the Yale Law School, has characterized this trend as a structural, institutional differentiation between freedom of speech and freedom of press.11 The trend has found expression that is both developmental and fundamental in Miami Herald Publishing Co. v. Tornillo, 418 U.S. 241, 94 S.Ct. 2831, 41 L.Ed.2d 730 (1974). There the Court established an area of protection of the press against “intrusion into the function of editors.” Id. at 258, 94 S.Ct. at 2839. The Court held unconstitutional a Florida statute requiring newspapers to print the replies of political candidates who had been editorially attacked. Chief Justice Burger’s opinion for a unanimous Court explained that “governmental regulation” of the “crucial process” of “editorial control and judgment” cannot be exercised consistently with evolving First Amendment guarantees of a free press. Id.12 Tornillo expressly adopt*987ed some of the premises and many of the implications of Columbia Broadcasting System, Inc. v. Democratic National Committee, 412 U.S. 94, 93 S.Ct. 2080, 36 L.Ed.2d 772 (1973). Columbia Broadcasting had emphasized the First Amendment right of broadcasters to make independent editorial programming decisions. In holding that neither the Communications Act nor the First Amendment requires broadcasters to accept paid editorial advertisements, Chief Justice Burger stated:
For better or worse, editing is what editors are for; and editing is selection and choice of material. That editors — newspaper or broadcast — can and do abuse this power is beyond doubt, but . [calculated risks of abuse are taken in order to preserve higher values.
Id. at 124-25, 93 S.Ct. at 2097.
“Governmental regulation” surely includes judicial as well as legislative regulation; the First Amendment binds the courts just as it binds the other branches of government.13 Tornillo and Columbia Broadcasting thus suggest and support, if they do not compel, the proposition that the First Amendment will not tolerate intrusion into the decision-making function of editors,14 be it legislative or judicial action.
The district court declined to apply the editorial process concept simply on the basis that Tornillo and Columbia Broadcasting had “nothing to do” with the proper bound*988aries of discovery in a libel case. Herbert v. Lando, 73 F.R.D. 387, 396 (SDNY 1977). While this notion may be appealing initially, the principle enunciated in these cases— that the editorial process of the press is entitled to special protection — has, I think, not just pertinent, but altogether controlling ramifications.15 As Mr. Justice Stewart has pointed out, “the Free Press guarantee is, in essence, a structural provision of the Constitution.”16 He continues:
*987[Njothing in our holding allows government at any level to forbid Pittsburgh Press to publish and distribute advertisements commenting on the Ordinance, the enforcement practices of the Commission, or the propriety of sex preference in employment. Nor, a fortiori, does our decision authorize any restriction whatever, whether of content or layout, on stories or commentary originated by Pittsburgh Press, its columnists, or its contributors. On the contrary, we reaffirm unequivocally the protection afforded to editorial judgment and to the free expression of views on these and other issues, however controversial.
*988Most of the other provisions in the Bill of Rights protect specific liberties or specific rights of individuals: freedom of speech, freedom of worship, the right to counsel, the privilege against compulsory self-incrimination, to name a few. In contrast, the Free Press Clause extends protection to an institution. The publishing business is, in short, the only organized private business that is given explicit constitutional protection.
This basic understanding is essential, I think, to avoid an elementary error of constitutional law. It is tempting to suggest that freedom of the press means only that newspaper publishers are guaranteed freedom of expression. They are guaranteed that freedom, to be sure, but so are we all, because of the Free Speech Clause. If the Free Press guarantee meant no more than freedom of expression, it would be a constitutional redundancy .
It is also a mistake to suppose that the only purpose of the constitutional guarantee of a free press is to insure that a newspaper will serve as a neutral forum for debate, a “marketplace for ideas,” a kind of Hyde Park corner for the community. A related theory sees the press as a neutral conduit of information between the people and their elected leaders. These theories, in my view, again give insufficient weight to the institutional autonomy of the press that it was the purpose of the Constitution to guarantee
The primary purpose of the constitutional guarantee of a free press was . to create a fourth institution outside the Government as an additional check on the three official branches.
Stewart, “Or of the Press,” 26 Hastings L.J. 631, 633-34 (1975) (emphasis in original).
The structural or institutional aspect of the Free Press guarantee is not, as Justice Stewart points out, some filigree added at the final stages of design by the architects of the Constitution. Rather it is at the core of the construct, vital to the tensile integrity of our government. See New York Times Co. v. United States, 403 U.S. 713, 717, 91 S.Ct. 2140, 29 L.Ed.2d 822 (1971) (Black, J., concurring); United States v. National Committee for Impeachment, 469 F.2d 1135, 1142 (2d Cir. 1972). To the extent that the independent exercise of editorial functions is threatened by governmental action, the very foundations of the architectural masterpiece that is our form of government are shaken, the supporting columns weakened.17
Tornillo and Columbia Broadcasting recognize the inviolability of the editorial function. As such they reflect a keen judicial *989recognition of the role of the press in American society and its need for protection, a trend that has been evidenced by practically solid judicial response in favor of protection.
The doctrine of prior restraint, prohibiting government from censoring publications in advance, is of the highest constitutional magnitude.18 This presumption19 against the constitutional validity of prior restraints is particularly strong when the intrusion affects the communication of news or commentary on current events. See Nebraska Press Association v. Stuart, 427 U.S. 539, 559, 96 S.Ct. 2791, 49 L.Ed.2d 683 (1976); Near v. Minnesota ex rel. Olson, 283 U.S. 697, 717, 51 S.Ct. 625, 75 L.Ed. 1357 (1931).
Further solidifying judicial recognition of the compelling institutional need for an independent press, again in a test that has been nothing less than momentous, is the decision that notwithstanding the Sixth Amendment guarantee to a fair trial by an impartial jury, judicial restraints on the publication of information concerning criminal trials will be tolerated only when less drastic methods of avoiding the effects of pretrial publicity are useless.20 Nebraska *990Press Association v. Stuart, supra, 427 U.S. at 562, 96 S.Ct. 2791. The Freedom of Press guarantee against governmental intrusions on the editorial process is surely at least as strong when, as here, other constitutional rights are not at stake.
In short, the principles underlying the access and prior restraint cases apply in this situation even though the government is neither ordering the content of a publication nor directly restraining publication. The critical question is whether government is impermissibly impeding the editorial function of the press; the time21 at which this intrusion occurs should not — it cannot —matter.22 Because broad discovery orders compelling disclosure of the editorial selection process can result in a chilling of “the free interchange of ideas within the news room,” ante at 980, the “crucial process” of “editorial control and judgment” protected by the Freedom of Press clause. Miami Herald Publishing Co. v. Tornillo, supra, 418 U.S. at 258, 94 S.Ct. 2831, is in as much jeopardy as if the court had restrained publication ab initio. For self-censorship in the future resulting from the prior judicial order is a foreseeable, perhaps a likely, result. “The special vice of a prior restraint is that communication will be suppressed, either directly or by inducing excessive caution in the speaker, before an adequate determination that it is unprotected by the First Amendment.” Pittsburgh Press Co. v. Pittsburgh Commission on Human Relations, 413 U.S. 376, 390, 93 S.Ct. 2553, 2561, 37 L.Ed.2d 669 (1973). This is a concern directed at the “institutional viability” of the press. Id. at 382, 93 S.Ct. 2553. Uninhibited discovery into the motivations of the editor in a libel action poses precisely the danger sought to be avoided by the landmark cases which have established the prior restraint doctrine as hornbook constitutional law. Accordingly, the principles underlying the doctrine necessitate the application of the Free Press guarantee to protect the independence of the press against such discovery.
This is by no means the first instance in which First Amendment considerations *991have dictated special procedural rules.23 It is not even the first situation in which a court has held that the First Amendment extends protection in a libel case beyond the standards for liability established in Sullivan in order to prevent undue chilling from the litigation process itself.24
By recognizing that Tornillo and Columbia Broadcasting require a constriction of the normal discovery rules to protect the editorial selection process of the press from compelled scrutiny, we simply add an additional procedural rule in the interest of ensuring an independent, institutional freedom of the press.25 Only the level of protection remains to be determined.26
*992III
In selecting the appropriate level of protection for the editorial process, we are faced with three theoretical possibilities. First, we might conclude as did the lower court that Sullivan has struck the ultimate appropriate balance so that the libel plaintiff must be permitted a level of discovery coterminous with the substantive law of constitutional libel. If so, then the plaintiff would be permitted to inquire into every aspect of the defendant’s state of mind at the discovery stage with little or no inhibition. Second, we might decide that while Sullivan has generally struck the substantive balance, it does not preclude restraint on compelled discovery, specifically where First Amendment values are unnecessarily threatened by the nonconstitutional interest in liberal discovery. We could adapt the test developed in the disclosure of confidential source cases: evidence of the editorial process is discoverable only when it is direct evidence of a highly relevant matter which cannot otherwise be obtained.27 And finally, we might opt for the conclusion that the editorial process, is subject to constitutional privilege and that actual malice must be proved by evidence other than that obtained through compelled disclosure of matters at the heart of the editorial process.
The answer is not free of doubt. Strict logic leading to the selection of option one has surface appeal. Approach number two seems like a reasonable compromise at first blush. But in the delicate area of precious First Amendment liberty, see Baker v. F & F Investment, supra, 470 F.2d at 785, a subtly discerning eye is necessary. Hard cases make for hard choices; vision must not only be acute, it must also be peripheral.
The argument of strict logic — that the Sullivan test of knowing-or-reckless-falsity assumes open-ended discovery for the purpose of proving actual malice — is deficient in several respects. First, the Sullivan Court in no way indicates that it is doing anything more than setting forth substantive rules. It does not deal with the method of proving actual malice.28 Actual malice can be proved in a number of ways. Logical inferences from the inconsistency, say, between a television program’s content and contrary facts which a plaintiff might independently establish would provide an obvious starting point for such proof.29 *993Moreover, a plaintiff might adduce circumstantial evidence from participants or interviewees on the television program. In this case, for example, documents furnished under the Freedom of Information Act indicate that Lando’s state of mind may be provable without directly impinging on the editorial process.30 While I offer no opinion on the admissibility or adequacy of this evidence to prove actual malice, it is clear that an editor’s state of mind can be examined without discovering facts at the heart of the editorial process. Limiting discovery to those matters and persons not at the heart of the editorial process does not transform the Sullivan rule into a nullity for putatively libeled public figures. They can prove actual malice without endangering the editorial process which Tornillo held to be protected by the First Amendment.31
Second, the Sullivan balance, by permitting plaintiffs the opportunity to prove actual malice, deems a certain level of chilling-effect fallout to be consistent with the First Amendment.32 However, permitting compelled discovery of the editorial process would indubitably increase the level of chilling effect in a way ostensibly not contemplated by Sullivan. Thus, it is one thing to tell the press that its end product is subject to the actual malice standard and that a plaintiff is entitled to prove actual malice; it is quite another to say that the editorial process which produced the end product in question is itself discoverable. Such an inquiry chills not simply the material published but the relationship among editors. Ideas expressed in conversations, memoranda, handwritten notes and the like, if discoverable, would in the future “likely”33 lead to a more muted, less vigorous *994and creative give-and-take in the editorial room. This incremental chilling effect exceeds the level of chilling effect contemplated by the Sullivan balance.
Finally, the fatal flaw of the strict logic position is its failure even to consider Tornillo’s and Columbia Broadcasting’s ramifications. It ignores the special status which the Free Press guarantee accords to the editorial process.34
The compromise position is similarly defective. While taking account of Tornillo’s mandates, it falls short of the protection required by Sullivan and Tornillo. Admittedly, the compromise test accounts for the argument advanced above that evidence of actual malice is obtainable from sources other than compelled discovery of editors’ state of mind because the compromise test itself requires that the evidence be not otherwise obtainable.35 However, the incremental chilling effect engendered by this test, while not as great as in the case of uninhibited discovery, is still significant. First, the editorial relationship may be chilled if its dynamics are subject to forced scrutiny. The knowledge that in a certain number of cases the editorial process will be discoverable is itself likely to chill that process, because no editor can visualize when a court will consider relevancy to be “high” or evidence to be “direct” or “otherwise unobtainable.” Beyond this, the compromise test is vague, difficult of application, and hence likely to be the subject of constant litigation.36 In effect, the discovery process itself, and the resulting litigation over the “directly-related,” “highly-relevant” and “otherwise-unobtainable” standards, are not merely likely to make editors more cautious, but inevitably will require them to be. The chilling effect of the compromise test is, therefore, of a greater degree than that tolerated by Sullivan with the gloss of Tornillo and Columbia Broadcasting.
There is an additional reason for rejecting the compromise position: it developed in a very different context from that at issue here. Baker v. F & F Investment, supra, 470 F.2d at 783-84, in its discussion of Garland v. Torre, 259 F.2d 545 (2d Cir.), cert. denied, 358 U.S. 910, 79 S.Ct. 237, 3 L.Ed.2d 231 (1958), indicates that the appropriate standard for compelled disclosure of journalists’ sources in civil litigation is that the evidence is closely related to the very essence of the plaintiff’s case and that the information is not obtainable from other sources.37 However, in Baker and Garland very different First Amendment interests were at stake from those at issue here. In those cases, the information sought to be disclosed, whether or not vital to the plaintiff’s case, was far removed from the editorial process. In this case, the plaintiffs do *995not seek discovery on the periphery of the editorial process. That they have already done, as Chief Judge Kaufman notes, to the tune of 2,903 pages in deposition testimony and of 240 exhibits. Rather, plaintiffs now seek to discover the very heart of the editorial process. This they may not do consistently with Tornillo’s and Columbia Broadcasting’s solicitude for the editorial process. I, therefore, would conclude that Tornillo, Columbia Broadcasting and Sullivan mandate full protection of the editorial process from compelled disclosure. This is true because as soon as facts are set in their context there is editorial selection; as soon as that process is subject to scrutiny, there is a suppression effect; and as soon as there is such an effect, the freedom of the press has evaporated.
IV
I pass then to when editorial process immunity from compelled disclosure is properly invoked. The parameters of the editorial process concept will become more definite in the context of future cases. The obvious starting point, however, is the Chief Justice’s delineation in Tornillo : “[t]he choice of material” to go into the broadcast, “the decisions made” on the duration and “content” of the broadcast, and “treatment of public issues and public officials — whether fair or unfair — constitute the exercise of editorial control and judgment.” 418 U.S. at 258, 94 S.Ct. at 2840. Thus, Tornillo mandates that the mental processes of the press regarding “choice of material,” duration, and “content” of the broadcast are to be protected from scrutiny.38
At this stage of the proceedings we are not capable of determining which discovery demands fall within the editorial process privilege. I agree with the Chief Judge that at least five of the broad categories into which the district court grouped plaintiff’s questions seemingly fall within the privilege. Each area of dispute relates to Lando’s conclusions, opinions, intentions, or conversations concerning people or leads to be pursued, the veracity of persons interviewed, and Lando’s reasons for the inclusion or exclusion of certain material. By permitting interrogation into these areas of editorial selection, the judiciary improperly intrudes upon freedom of the press just as the Florida legislature did in Tornillo. Whether the intrusion is judicial or legislative, the result is an unconstitutional suppression effect. There may, however, be individual questions inappropriately grouped within the protected categories. I think it is open to the district judge to determine in specific instances that Lando was not engaged in the process of editorial selection. With these caveats I concur in the general answer to the certified question and in the remand to the district court for a determination of the nature of each question in dispute.
. See Restatement (Second) of Torts, Special Note at 3 (Tent. Draft No. 21, 1975); Eaton, The American Law of Defamation Through Gertz v. Robert Welch, Inc. and Beyond: An Analytical Primer, 61 Va.L.Rev. 1349, 1364-1408 (1975); Robertson, Defamation and the First Amendment: In Praise of Gertz v. Robert Welch, Inc., 54 Tex.L.Rev. 199, 199 (1976). For an early explication of the change of the tort law of defamation into a subject of constitu*985tional dimension, see Kalven, The New York Times Case: A Note on “the Central Meaning of the First Amendment," 1964 Sup.Ct.Rev. 191.
. It involves the conduct of the Vietnam war, a public issue, New York Times Co. v. United States, 403 U.S. 713, 724, 91 S.Ct. 2140, 29 L.Ed.2d 822 (1971) (Douglas, J., concurring), and a United States Army officer who was a public official and employee, who by his charges against the military establishment unquestionably made himself a public figure, thereby inviting “attention and comments.” See Gertz v. Robert Welch, Inc., 418 U.S. 323, 345, 94 S.Ct. 2997, 41 L.Ed.2d 789 (1974); Buckley v. Littell, 539 F.2d 882, 885-86 (2d Cir. 1976), cert. denied, 429 U.S. 1062, 97 S.Ct. 785, 50 L.Ed.2d 777 (1977).
. The phrase “actual malice” is a “term of art” that evidently is now “studiously avoided” by the Supreme Court. See Eaton, supra note 1, at 1370 n.87. However, it is a shorthand phrase for Sullivan’s “knowing falsity or reckless disregard of truth” test. It is used here in its accurate sense.
. Fed.R.Civ.P. 26(b)(1).
. While the area for which appellant seeks protection, “editorial process,” may initially seem vague, guidance is provided by Chief Justice Burger’s statements in Miami Herald Publishing Co. v. Tornillo, 418 U.S. 241, 258, 94 S.Ct. 2831, 41 L.Ed.2d 730 (1974), and in Columbia Broadcasting Sys., Inc. v. Democratic Nat’l Comm., 412 U.S. 94, 124-25, 93 S.Ct. 2080, 36 L.Ed.2d 772 (1973). See text accompanying note 37 infra. The specific areas of inquiry sought by the plaintiff are set out in Chief Judge Kaufman’s opinion, ante at 983.
. See Socialist Workers Party v. Attorney General, 565 F.2d 19 (2d Cir. 1977) (no interlocutory review of discovery order absent certification, “manifest abuse of discretion,” or legal question of “extraordinary significance”); cf. Time, Inc. v. McLaney, 406 F.2d 565, 566 (5th Cir.) (interlocutory review of a denial of summary judgment appealable because of critically important First Amendment issue), cert. denied, 395 U.S. 922, 89 S.Ct. 1776, 23 L.Ed.2d 239 (1969).
. I do not mean to suggest that Judge Haight in his scholarly opinion below relied simply on the discovery rules.
. Fed.R.Civ.P. 26(c); cf. Branzburg v. Hayes, 408 U.S. 665, 707-08, 92 S.Ct. 2646, 33 L.Ed.2d 626 (1972) (grand jury harassment of press impermissible).
. Fed.R.Civ.P. 26(b)(1).
. That the broadcast media are “press” is reasonably well established. E. g., Cox Broadcasting Corp. v. Cohn, 420 U.S. 469, 496-97, 95 S.Ct. 1029, 43 L.Ed.2d 328 (1975); Red Lion Broadcasting Co. v. FCC, 395 U.S. 367, 394, 89 S.Ct. 1794, 23 L.Ed.2d 371 (1969); United States v. Paramount Pictures, Inc., 334 U.S. 131, 166, 68 S.Ct. 915, 92 L.Ed. 1260 (1948).
. Stewart, “Or of the Press,” 26 Hastings L.J. 631, 633 (1975); see text accompanying notes 16-17 infra.
. Concurring, Mr. Justice White noted:
Regardless of how beneficient-sounding the purposes of controlling the press might be, we prefer “the power of reason as applied through public discussion” and remain intensely skeptical about those measures that would allow government to insinuate itself into the editorial rooms of this Nation’s press.
*987Miami Herald Publishing Co. v. Tornillo, supra, 418 U.S. at 259, 94 S.Ct. at 2840 (White, J., concurring) (footnote omitted). But see Red Lion Broadcasting Co. v. FCC, supra, 395 U.S. at 385-86, 89 S.Ct. 1794, holding that political editorializing by radio broadcasters is subject to the “fairness doctrine” under the Federal Communications Act. As to the right of access guaranteed by Red Lion, see F. Friendly, The Good Guys, The Bad Guys and the First Amendment: Free Speech vs. Fairness in Broadcasting (1976). Compare B. Schmidt, Freedom of the Press vs. Public Access (1976), with Abrams, In Defense of Tornillo, 86 Yale L.J. 361 (1976) (book review).
The Red Lion opinion, as well as the Act of Congress on which it is based, have been increasingly criticized for failing to consider the institutional aspects of the press. E. g., Note, Press Protections for Broadcasters: The Radio Format Change Cases Revisited, 52 N.Y.U.L. Rev. 324, 339 (1977) [hereinafter Note, Press Protections]. The criticism suggests that the opinion and Act fail to take into account the distinction between the Speech and Press clauses. As Chief Judge Bazelon has noted:
If one group has a right of access or a right to have the licensee present that group’s point of view, there is no independent press; there is only a multitude of speakers. That might be permissible if the First Amendment protected only free speech. However, it also protects the press.
Bazelon, FCC Regulation of the Telecommunications Press, 1975 Duke L.J. 213, 235 (footnote omitted); see Note, Press Protections, supra at 339 n. 102.
It has been asserted that historically there was no differentiation between Speech and Press guarantees. L. Levy, Legacy of Suppression: Freedom of Speech and Press in Early American History 173-74 (1960); Lange, The Speech and Press Clauses, 23 U.C.L.A.L.Rev. 77, 88-99 (1975); Note, Press Protections, supra at 342. But the High Court arguably has established the differentiation by now, whether or not it was the “original position.” Compare Stewart, supra note 11, at 631-37, and Nimmer, Introduction — Is Freedom of the Press a Redundancy: What Does It Add to Freedom of Speech?, 26 Hastings L.J. 639, 645-50 (1975), with Lange, supra at 77 Passim.
. Sullivan itself recognizes limitations on the courts imposed by the First Amendment. So, too, does Nebraska Press Ass’n v. Stuart, 427 U.S. 539, 96 S.Ct. 2791, 49 L.Ed.2d 683 (1976) (judicially imposed gag order violates First Amendment rights of the press).
. See Pittsburgh Press Co. v. Pittsburgh Comm'n on Human Relations, 413 U.S. 376, 391, 93 S.Ct. 2553, 2561, 37 L.Ed.2d 669 (1973). There the Court, in upholding the constitutional validity of an ordinance prohibiting sex-based help-wanted advertisements, emphasized:
. Government intrusion into editorial functions may result from compelled disclosure of editorial conclusions, opinions and intentions. To be sure, Tornillo and Columbia Broadcasting do not specifically deal with First Amendment limitations on discovery. But they give concrete form to the structural concept of press freedom in the editorial selection process which would otherwise be subjected to stress, if not internally weakened to the point of non-repair, by the probing drill of unrestrained discovery.
. Stewart, supra note 11, at 633-34.
. But see Carey v. Hume, 160 U.S.App.D.C. 365, 373-74, 492 F.2d 631, 639-40 (1974) (MacKinnon, J., concurring) (immense power of modern media requires that reporter divulge sources in civil libel suit), cert. dismissed, 417 U.S. 938, 94 S.Ct. 2654, 41 L.Ed.2d 661 (1974). The power that Judge MacKinnon (and an advocate of the Florida access statute in Tornillo) fears may, however, be dealt with in other ways. The FCC has, for example, won court approval to bar newspaper ownership of broadcast/television media in a single locale in the future. See National Citizens Comm, for Broadcasting v. FCC, 181 U.S.App.D.C. 1, 555 F.2d 938 (1977), cert. granted, - U.S. -, 98 S.Ct. 52, 54 L.Ed.2d 70 (1977).
. Prior restraints have generally been condemned as the most egregious violations of press freedom. E. g., Nebraska Press Ass’n v. Stuart, supra, 427 U.S. at 559, 96 S.Ct. at 2803 (“A prior restraint . . . has an immediate and irreversible sanction. If it can be said that a threat of criminal or civil sanctions after publication ‘chills’ speech, prior restraint ‘freezes’ it at least for the time.”) (footnote omitted); See Times Film Corp. v. City of Chicago, 365 U.S. 43, 53, 81 S.Ct. 391, 5 L.Ed.2d 403 (1961) (Warren, C. J., dissenting). The precise parameters of the prior restraint doctrine have never been delineated. Near v. Minnesota ex rel. Olson, 283 U.S. 697, 51 S.Ct. 625, 75 L.Ed. 1357 (1931), expanded the classic English definition of prohibiting publication without advance approval of the government to preclude an injunction against publication of malicious and scandalous matter enforced by a subsequent contempt order. The later threat of an injunction was viewed as a prior restraint; while the metaphor of “chilling effect” had not yet been devised it was already operational. The broad scope of the doctrine is apparent from the holding in Grosjean v. American Press Co., 297 U.S. 233, 56 S.Ct. 444, 80 L.Ed. 660 (1936), where a tax imposed on newspapers publishing advertisements and measured by the amount of circulation was found to be an unconstitutional prior restraint because the tax lowered advertising revenues and restricted circulation. Id. at 244-45, 56 S.Ct. 444. Justice Sutherland explained the scope of the prior restraint doctrine as enunciated in Near:
The conclusion there stated is that the object of the constitutional provisions was to prevent previous restraints on publication; and the court was careful not to limit the protection of the right to any particular way of abridging it. .
Judge Cooley has laid dotvn the test to be applied — “The evils to be prevented were not the censorship of the press merely, but any action of the government by means of which it might prevent such free and general discussion of public matters as seems absolutely essential to prepare the people for an intelligent exercise of their rights as citizens.” 2 Cooley’s Constitutional Limitations, 8th ed., p. 886.
Id. at 249-50, 56 S.Ct. at 449.
The broad discovery order in this case operates after publication to deter free editorial choice concerning subsequent publications. Because the order operates as a prior restraint, as did the tax in Grosjean, it should be presumed invalid. See T. Emerson, The System of Freedom of Expression 503-12 (1970). Furthermore, the order impinges on the most sensitive functions of the press, the motivations behind and thought processes involved in editorial decisions; thus, the presumption should be given all the more force. The prior restraint cases, as well as Tornillo and Columbia Broadcasting, appear to be directed against the danger of self-censorship by the press arising from concern with subsequent executive, legislative or judicial scrutiny. It is the protection these cases afford that truly gives rise to the concept of the press as a Fourth Estate, coequal in our democratic republic in constitutional respect, even though not incorporated formally into our governmental system as a structuralized entity.
. New York Times Co. v. United States, supra, 403 U.S. at 714, 91 S.Ct. 2140; Organization for a Better Austin v. Keefe, 402 U.S. 415, 419, 91 S.Ct. 1575, 29 L.Ed.2d 1 (1971); Bantam Books, Inc. v. Sullivan, 372 U.S. 58, 70, 83 S.Ct. 631, 9 L.Ed.2d 584 (1963).
. Another example of increasing judicial recognition of the fundamentality of the Press guarantee is the newly perceived status of commercial speech. See Bates v. State Bar of Ariz., 433 U.S. 350, 97 S.Ct. 2691, 53 L.Ed.2d 810 (1977); Linmark Assocs., Inc. v. Township of Willingboro, 431 U.S. 85, 97 S.Ct. 1614, 52 L.Ed.2d 155 (1977); Virginia State Bd. of Pharmacy v. Virginia Citizens Consumer Council, Inc., 425 U.S. 748, 96 S.Ct. 1817, 48 L.Ed.2d 346 *990(1976); Bigelow v. Virginia, 421 U.S. 809, 95 S.Ct. 2222, 44 L.Ed.2d 600 (1975). The Sullivan doctrine itself, subjecting the common law of libel to First Amendment limitations, while addressed to expression generally, New York Times Co. v. Sullivan, 376 U.S. 254, 269, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964), also reflects a pervasive Supreme Court awareness of the arterial flow to our system of democracy given through the channels of a free press. The Sullivan doctrine was, after all, enunciated in a suit against a newspaper. The opinion of the Court drew extensively on the history of the discredited Alien and Sedition Acts, id. at 273-77, emphasizing Madison’s Report to the effect that “the press has exerted a freedom in canvassing the merits and measures of public men, of every description, which has not been confined to the strict limits of the common law.” Id. at 275 (quoting 4 Elliot’s Debates on the Federal Constitution 570 (1876)). Mr. Justice Goldberg, concurring, advocated unconditional First Amendment protection to criticize official conduct. Id. at 298. He, too, relied heavily on the need for robust debate to ensure a stable, responsive democratic government. See id. at 300. See generally 1 N. Dorsen, P. Bender & B. Neuborne, Emerson, Haber & Dorsen’s Political and Civil Rights in the United States 20-51 (4th ed. 1976) [hereinafter 1 N. Dorsen], for an historical analysis of the First Amendment.
. The phrase “prior restraint” obviously no longer connotes a strict temporal meaning. Supreme Court cases indicate, see note 18 supra, that meddling with press freedoms after the printing or programming has occurred operates as a “prior restraint” on future editorial decisions by virtue if the chilling effect created by that interference. See T. Emerson, supra note 18, at 511.
. It has often been acknowledged that trivial distinctions between types of governmental intrusion will not be relied upon when the effects impinge the Free Press guarantee.
In rejecting the argument that there is a meaningful difference between government restricting the content of press communications and compelling the press to publish what “reason tells them should not be published,” the Court in Tornillo noted that “[g]overnmental restraint on publishing need not fall into familiar or traditional patterns to be subject to constitutional limitations on governmental powers.” 418 U.S. at 256, 94 S.Ct. at 2839. See Baker v. F & F Investment, 470 F.2d 778, 785 (2d Cir. 1972) (Kaufman J.), cert. denied, 411 U.S. 966, 93 S.Ct. 2147, 36 L.Ed.2d 686 (1973).
. See, e. g., Broadrick v. Oklahoma, 413 U.S. 601, 611-14, 93 S.Ct. 2908, 37 L.Ed.2d 830 (1973) (relaxed standing rules to challenge statutes allegedly violative of the First Amendment); Baker v. F & F Investment, supra, 470 F.2d at 783 (rejecting disclosure of journalists’ sources in civil rights case except under limited circumstances).
. Indeed, in Sullivan the Court qualified those standards by the self-imposed procedural limitation that it would make an independent examination of the entire record. New York Times Co. v. Sullivan, supra, 376 U.S. at 285 & n.26, 84 S.Ct. 710, see Time, Inc. v. Pape, 401 U.S. 279, 284, 91 S.Ct. 633, 28 L.Ed.2d 45 (1971); Buckley v. Litteli, supra, 539 F.2d at 888. A number of courts taking a normally cautious attitude toward summary judgment have been somewhat more relaxed in constitutional libel cases. Washington Post Co. v. Keogh, 125 U.S.App.D.C. 32, 34-35, 365 F.2d 965, 967-68 (1966), cert. denied, 385 U.S. 1011, 87 S.Ct. 708, 17 L.Ed.2d 548 (1967); see 1 N. Dorsen, supra note 21, at 693-95. To prevail, moreover, again as established by Sullivan itself, a plaintiff’s evidence must be of “convincing clarity,” rather than a mere preponderance. New York Times Co. v. Sullivan, supra, 376 U.S. at 285-86, 84 S.Ct. 710, see Wasserman v. Time, Inc., 138 U.S.App.D.C. 7, 9, 424 F.2d 920, 922 (Wright, J., concurring), cert. denied, 398 U.S. 940, 90 S.Ct. 1844, 26 L.Ed.2d 273 (1970); Freund, William J. Brennan, Jr., 86 Yale L.J. 1015, 1016 (1977). First Amendment considerations have been held to lead more readily to a finding of forum non conveniens, Buckley v. New York Post Corp., 373 F.2d 175, 183-84 (2d Cir. 1967), and lack of jurisdiction, New York Times Co. v. Connor, 365 F.2d 567, 570-73 (5th Cir. 1966).
. To the extent that this conclusion creates an inconsistency between judicial treatment of the press on the one hand and non-press defendants on the other, the inconsistency may be said to rest upon the non-redundant nature of the Freedom of Press guarantee. See note 12 supra. But see Restatement (Second) of Torts § 580B, Comment d at 29-30 (Tent. Draft No. 21, 1975), discussing the scope of Gertz v. Robert Welch, Inc., supra:
The defendant in the Gertz case was the publisher of a magazine. The Court speaks frequently of “news media” and “communications media” and states the rule in terms of a “publisher or broadcaster.” The precise holding of the case therfore, [sic] does not extend beyond a statement published by a member of the communications media; and the constitutional requirement of fault on the part of the defendant may turn out to be limited to this holding, though this seems unlikely.
Of course, there would be no inconsistency if the Sullivan rule were abandoned in favor of non-liability to public figure plaintiffs.
. From its inception the Sullivan rule, as a matter of substantive law, has not been spared of criticism. Justices Black, Douglas and Goldberg thought the actual malice test constitutionally deficient for inquiring in the first place into the editor’s state of mind. As Mr. Justice Black explained:
“Malice,” even as defined by the Court, is an elusive, abstract concept, hard to prove and hard to disprove. The requirement that malice be proved provides at best an evanescent protection for the right critically to discuss public affairs .
376 U.S. at 293, 84 S.Ct. at 733 (Black, J., concurring). Mr. Justice Goldberg elaborated further by noting that the right to criticize official misconduct, “ ‘to speak one’s mind’ . about public officials and affairs . should not depend upon a probing by the jury of the motivation of the citizen or press.” Id. at 298, 84 S.Ct. at 735 (Goldberg, J., concurring) (citations and footnote omitted). And in his footnote two, Mr. Justice Goldberg quoted from Mr. Justice Jackson’s dissent in United States v. Ballard, 322 U.S. 78, 92-93, 64 S.Ct. 882, 88 L.Ed. 1148 (1944), that it is difficult to separate, practically or philosophically, “what is believed” from “what is believable.” 376 U.S. at 298-99 n.2, 84 S.Ct. 710. (Goldberg, J., concurring). Professor Emerson has written that the actual malice rule is
subject to the very same defects that led the majority of the [Sullivan ] Court to reject broader tests of liability[,] . leaves the speaker with roughly the same degree of risk as the earlier rules of negligence and engenders approximately the same amount of self-censorship . . . [and] imposes an impossible problem of judicial administration.
T. Emerson, supra note 18, at 535-36.
*992Professor Emerson also attacks the Sullivan majority’s rationale, set forth in Garrison v. Louisiana, 379 U.S. 64, 85 S.Ct. 209, 13 L.Ed.2d 125 (1964), that calculated falsehood is no essential part of any exposition of ideas as inconsistent with the concept of Sullivan itself. See T. Emerson, supra, at 536. But even inferior federal courts know that upon occasion judicial opinions may involve compromises in expression to reach a result in fact; here there was at least a procedural accommodation, note 24 supra.
However persuasive these arguments might be if we were writing on a clean slate, as an inferior court we are of course bound to follow the rule of Sullivan. Thus the editor’s state of mind, not vis-á-vis the plaintiff, but vis-á-vis the truth or falsity of what is being published about the plaintiff, is a proper focus of inquiry.
. See note 36 and accompanying text infra.
. Appellee argues that Tornillo and Columbia Broadcasting have not affected media liability under Sullivan and its progeny, relying on the Court’s reaffirmation of the Sullivan liability standard in Gertz v. Robert Welch, Inc., supra, 418 U.S. at 342, 94 S.Ct. 2997, decided on the same day as Tornillo. Brief for Appellee at 29 & n.*. This argument, however, overlooks the distinction between standards of liability and the means of proving liability under the appropriate standard. Tornillo and Columbia Broadcasting have not altered the substantive law of libel established in Sullivan. They articulate broad First Amendment protection of editorial process decisions. The issue, then, is simply whether and to what extent this protection encompasses compulsory disclosure of the mental processes of editors. The free press principles of these cases are being applied to limit a procedural rule, not to alter the substantive law of libel.
. A case in our court goes further perhaps than any other in permitting proof of bad motive directed toward the plaintiff to show the reckless disregard of truth that the actual malice test of Sullivan requires. See Goldwater v. Ginzburg, 414 F.2d 324, 342 (2d Cir. 1969), cert. denied, 396 U.S. 1049, 90 S.Ct. 701, 24 L.Ed.2d 695 (1970) (“evidence of negligence, of motive and of intent may be adduced for the purpose of establishing, by cumulation and by appropriate inferences, the fact of a defendant’s recklessness or of his knowledge of falsity”). Goldwater, however, was not concerned with the *993boundaries of plaintiff’s discovery into the editorial process but rather dealt principally with the sufficiency of the evidence adduced and the correctness of the judges, charge to the jury on the appropriate standard of substantive law.
. The documents, prepared by Lieutenant Colonel F. B. Reed, Jr., indicate that in anticipation of the broadcast Lando told Colonel Reed that “Lando’s stated premise is that Herbert is a liar and he has stated that if he can’t develop a sufficient number of incidents in which Herbert’s account can not [sic] be debunked, then there will be no story.” Brief for Appellee at 56 (Exhibit I). On December 4, 1972, Lando interviewed Colonel J. Ross Franklin. During the interview, at which Reed was present, Lando “persisted] in [his] contention that he is interested in debunking Herbert. . After the interview he informed me that Mike Wallace has agreed to do the narration and is equally convinced that the story is in debunking Herbert. Lando asserts that he has [the] final decision on the segment and it will not go unless he can convincingly portray Herbert as the bad guy.” Id. at 57 (Exhibit II). The third document reveals that on December 20, 1972, Lando in a meeting with Major General Sidle “indicated that his peice [sic] is aimed at debunking Herbert in his long fight against the Army. Further Lando indicated that he would focus some attention on the failure of the media to check out Herbert’s story prior to ‘puffing him up’. [Sic.] He plans to focus on four or five events whcih [sic] are contained in Herberts [sic] book and factually destroy Herbert’s credibility.” Id. at 58 (Exhibit III).
. Limiting plaintiff’s discovery concededly may deprive him of adducing the best proof of malice in the common law sense of ill will toward the plaintiff. But Sullivan itself distinguishes common law malice from actual malice. Limiting proof of actual malice as defined in Sullivan resembles other rules of evidence which limit the “search for truth” in the interests of a higher social policy. See, e. g., Fed.R. Evid. 407, precluding introduction of subsequent remedial measures to prove negligence in order to encourage the promotion of safety.
. Hotchner v. Castillo-Puche, 551 F.2d 910 (2d Cir. 1977), cert. denied, - U.S. -, 98 S.Ct. 120, 54 L.Ed.2d 95 (1977), provides an excellent example of the inhibiting effect that Sullivan may exercise on First Amendment freedoms. In Hotchner, the defendant had written a biography of Ernest Hemingway with uncomplimentary references to the plaintiff, a public figure. Prior to publication the editor, pursuant to recommendations of the legal department of Doubleday & Co., Inc., suggested that a number of passages “be eliminated or toned down.” Id. at 912. Even though the author “vouched for the statements” in his manuscript, he “accepted the suggested modification.” Id. The self-censorship was nonetheless imposed despite the fact that it may have been unnecessary in view of the court’s subsequent conclusion that “[w]here a passage is incapable of independent verification, and where there are no convincing indicia of unreliability, publication of the passage cannot constitute reckless disregard for truth.” Id. at 914.
. See Lamont v. Postmaster General, 381 U.S. 301, 307, 85 S.Ct. 1493, 1496, 14 L.Ed.2d 398 (1965) (postal regulation that those wishing to *994receive “communist political propaganda” had to request it of the Post Office held unconstitutional because “any addressee is likely to feel some inhibition” at doing so (emphasis added)); Talley v. California, 362 U.S. 60, 64, 80 S.Ct. 536, 4 L.Ed.2d 559 (1960) (requirement that the names and addresses of those who wrote and/or sponsored the distribution of handbills held unconstitutional because it “would tend to restrict freedom to distribute information and thereby freedom of expression” (emphasis added)).
. I do not make the distinction between the institutional press and the individual pamphleteer which Judge Meskill suggests in dissent, post at 997. Rather, the distinction I draw is between communicative functions properly protected under the Free Press clause and expression protected by the Free Speech guarantee.
. Of course, the number of instances that the evidence is in fact not otherwise obtainable in some form should be few. This does not mean that the chilling effect in the editorial room would be concomitantly reduced. The fear of such discovery and of the full scale utilization of the litigation process as permitted by the compromise test is likely to stifle the flow of ideas in the editorial room.
. See New York Times Co. v. Sullivan, supra, 376 U.S. at 277-78, 84 S.Ct. 710, 11 L.Ed.2d 686.
. Cf. Branzburg v. Hayes, supra (rejecting argument of journalists who had witnessed crimes that they were not obliged to testify before grand juries with respect thereto). In Baker v. F & F Investment, supra, 470 F.2d at 783, the Second Circuit distinguished Branzburg and thereby protected journalists from compelled disclosure of sources in civil litigation.
. To the extent that Buckley v. Vidal, 50 F.R.D. 271 (S.D.N.Y.1970), is inconsistent with this opinion, this court obviously now declines to follow it.
While the question need not be answered here, I would assume that the very same protection would be afforded the press in a trial on the merits. Compelled testimony at trial on matters privileged from compelled pretrial discovery should similarly be privileged, because this case concerns privilege, not undue oppression. See text accompanying notes 7-9 supra.