dissenting.
Although professing to maintain the accommodation of interests struck in New York Times Co. v. Sullivan, 376 U. S. 254 (1964), the Court today is unresponsive to the constitutional considerations underlying that opinion. Because I believe that some constraints on pretrial discovery are essential to ensure the “uninhibited [and] robust” debate on public *203issues which Sullivan contemplated, id., at 270, I respectfully dissent.
I
At issue in this case are competing interests of familiar dimension. States undeniably have an interest in affording individuals some measure of protection from unwarranted defamatory attacks. Libel actions serve that end, not only by assuring a forum in which reputations can be publicly vindicated and dignitary injuries compensated, but also by creating incentives for the press to exercise considered judgment before publishing material that compromises personal integrity. See Gertz v. Robert Welch, Inc., 418 U. S. 323, 341-342 (1974); Rosenblatt v. Baer, 383 U. S. 75, 86 (1966).
Against these objectives must be balanced society’s interest in promoting unfettered debate on matters of public importance. As this Court recognized in Sullivan, error is inevitable in such debate, and, if forced to guarantee the truth of all assertions, potential critics might suppress statements believed to be accurate “because of doubt whether '[truthfulness] can be proved in court or fear of the expense of having to do so.” 376 U. S., at 279. Such self-censorship would be incompatible with the tenets on which the First Amendment and our democratic institutions are founded. Under a representative system of government, an informed electorate is a precondition of responsive decisionmaking. See Associated Press v. United States, 326 U. S. 1, 20 (1945); Grosjean v. American Press Co., 297 U. S. 233, 250 (1936); A. Meikle-john, Free Speech and its Relation to Self-Government 88-89 (1948). To secure public exposure to the widest possible range of information and insights, some margin of error must be tolerated. Thus, absent knowing falsity or reckless disregard for the truth, the press is shielded from liability for defamatory statements regarding public figures. Curtis Publishing Co. v. Butts, 388 U. S. 130 (1967); New York Times Co. v. Sullivan, supra.
*204Yet this standard of liability cannot of itself accomplish the ends for which it was conceived. Insulating the press from ultimate liability is unlikely to avert self-censorship so long as any plaintiff with a deep pocket and a facially sufficient complaint is afforded unconstrained discovery of the editorial process. If the substantive balance of interests struck in Sullivan is to remain viable, it must be reassessed in light of the procedural realities under which libel actions are conducted.
II
The potential for abuse of liberal discovery procedures is of particular concern in the defamation context. As members of the bench and bar have increasingly noted, rules designed to facilitate expeditious resolution of civil disputes have too often proved tools for harassment and delay.1 Capitalizing on this Court’s broad mandate in Hickman v. Taylor, 329 U. S. 495, 507 (1947), reaffirmed in Schlagenhauf v. Holder, 379 U. S. 104, 114-115 (1964), that discovery rules be accorded a “broad and liberal” scope, litigants have on occasion transformed Fed. Rule Civ. Proc. 26 devices into tactics of attrition. The possibility of such abuse is enhanced in libel litigation, for many self-perceived victims of defamation are animated by something more than a rational calculus of their chances of recovery.2 Given the circumstances under which *205libel actions arise, plaintiffs’ pretrial maneuvers may be fashioned more with an eye to deterrence or retaliation than to unearthing germane material.
Not only is the risk of in terror em discovery particularly pronounced in the defamation context, but the societal consequences attending such abuse are of special magnitude. Rather than submit to the intrusiveness and expense of protracted discovery, even editors confident of their ability.to prevail at trial or on a motion for summary judgment may find it prudent to “ ‘steer far wid[e] of the unlawful zone’ thereby keeping protected discussion from public cognizance.” Rosenbloom v. Metromedia, Inc., 403 U. S. 29, 53 (1971) (plurality opinion; citation omitted). Faced with the prospect of escalating attorney’s fees, diversion of time from journalistic endeavors, and exposure of potentially sensitive information, editors may well make publication judgments that reflect less the risk of liability than the expense of vindication.3
Although acknowledging a problem of discovery abuse, the Court suggests that the remedy lies elsewhere, in “major changes in the present Rules of Civil Procedure.” Ante, at 177. And somewhat inconsistently, the Court asserts further that district judges already have “in fact and in law . . . ample powers ... to prevent abuse.” Ibid. I cannot agree. Where First Amendment rights are critically implicated, it is incumbent on this Court to safeguard their effective exercise. By leaving the directives of Hickman and Schlagenhauf unqualified with respect to libel litigation, the Court has abdicated that responsibility.4
*206In my judgment, the same constitutional concerns that impelled us in Sullivan to confine the circumstances under which defamation liability could attach also mandate some constraints on roving discovery. I would hold that the broad discovery principles enunciated in Hickman and Schlagenhauf are inapposite in defamation cases. More specifically, I would require that district courts superintend pretrial disclosure in such litigation so as to protect the press from unnecessarily protracted or tangential inquiry. To that end, discovery requests should be measured against a strict standard of relevance. Further, because the threat of disclosure may intrude with special force on certain aspects of the editorial process, I believe some additional protection in the form of an eviden-tiary privilege is warranted.
Ill
The Court of Appeals extended a privilege subsuming essentially two kinds of discovery requests. The first included questions concerning the state of mind of an individual journalist, principally his conclusions and bases for conclusions as to the accuracy of information compiled during investigation. The second encompassed communications between journalists about matter to be included in the broadcast. 568 F. 2d 974, 978 (CA2 1977). Reasoning that discovery of both forms of material would be intrusive, that the intrusion would be inhibiting, and that such inhibition would be inconsistent with *207the editorial autonomy recognized in Miami Herald Publishing Co. v. Tornillo, 418 U. S. 241 (1974), and Columbia Broadcasting System, Inc. v. Democratic National Committee, 412 U. S. 94 (1973), the Court of Appeals concluded that a privilege from disclosure was essential. 568 F. 2d, at 975.
With respect to state-of-mind inquiry, that syllogism cannot withstand analysis. For although discovery may well be intrusive, it is unclear how journalists faced with the possibility of such questions can be “chilled in the very process of thought.” Id., at 984. Regardless of whether strictures are placed on discovery, reporters and editors must continue to think, and to form opinions and conclusions about the veracity of their sources and the accuracy of their information. At best, it can be argued only that failure to insulate the press from this form of disclosure will inhibit not the editing process but the final product — that the specter of questions concerning opinion and belief will induce journalists to refrain from publishing material thought to be accurate. But as my Brother Brennan notes, ante, at 192-193, this inhibition would emanate principally from Sullivan’s substantive standard, not from the incremental effect of such discovery. So long as Sullivan makes state of mind dispositive, some inquiry as to the manner in which editorial decisions are made is inevitable. And it is simply implausible to suppose that asking a reporter why certain material was or was not included in a given publication will be more likely to stifle incisive journalism than compelling disclosure of other objective evidence regarding that decision.5
*208I do not mean to suggest, as did the District Court here, that Tornillo and Columbia Broadcasting have “nothing to do” with this case. 73 F. R. D. 387, 396 (SDNY 1977). To the contrary, the values of editorial autonomy given recognition in those decisions should inform district courts as they monitor the discovery phase of defamation cases. But assuming that a trial judge has discharged his obligation to prevent unduly protracted or inessential disclosure, see supra, at 206, I am unpersuaded that the impact of state-of-mind inquiry will of itself threaten journalistic endeavor beyond the threshold contemplated by Sullivan.
External evidence of editorial decisionmaking, however, stands on a different footing. For here the concern is not simply that the ultimate product may be inhibited, but that the process itself will be chilled. Journalists cannot stop forming tentative hypotheses, but they can cease articulating them openly. If prepublication dialogue is freely discoverable, editors and reporters may well prove reluctant to air their *209reservations or to explore other means of presenting information and comment. The threat of unchecked discovery may well stifle the collegial discussion essential to sound editorial dynamics. As we recognized in United States v. Nixon, 418 U. S. 683, 705 (1974): “[T]hose who expect public dissemination of their remarks may well temper candor with a concern for appearances ... to the detriment of the decisionmaking process.” (Footnote omitted.) Cf. NLRB v. Sears, Roebuck & Co., 421 U. S. 132, 151 (1975). Society’s interest in enhancing the accuracy of coverage of public events is ill-served by procedures tending to muffle expression of uncertainty. To preserve a climate of free interchange among journalists, the confidentiality of their conversation must be guaranteed.
It is not enough, I believe, to accord a discovery privilege that would yield before any plaintiff who can make a prima facie showing of falsity. See ante, at 197-198 (opinion of Brennan, J.). Unless a journalist knows with some certitude that his misgivings will enjoy protection, they may remain unexpressed. See 568 F. 2d, at 994 (Oakes, J., concurring). If full disclosure is available whenever a plaintiff can establish that the press erred in some particular, editorial communication would not be demonstrably less inhibited than under the Court’s approach. And by hypothesis, it is precisely those instances in which the risk of error is significant that frank discussion is most valuable.
Accordingly, I would foreclose discovery in defamation cases as to the substance of editorial conversation.6 Shielding *210this limited category of evidence from disclosure would be unlikely to preclude recovery by plaintiffs with valid defamation claims. For there are a variety of other means to establish deliberate or reckless disregard for the truth, such as absence of verification, inherent implausibility, obvious reasons to doubt the veracity or accuracy of information, and concessions or inconsistent statements by the defendant. See St. Amant v. Thompson, 390 U. S. 727, 732 (1968). To the extent that such a limited privilege might deny recovery in some marginal cases, it is, in my view, an acceptable price to pay for preserving a climate conducive to considered editorial judgment.
I would therefore direct the Court of Appeals to remand this case to the District Court for determination first, whether the questions concerning Lando’s state of mind satisfy the criteria set forth in Part II of this opinion, and second, whether respondents waived the privilege defined in Part III for prepublication discussions.
See Bell, The Pound Conference Follow-up: A Response from the United States Department of Justice, 76 F. R. D. 320, 328-329 (1978) ; Erikson, The Pound Conference Recommendations: A Blueprint for the Justice System in the Twenty-First Century, 76 F. R. D. 277, 288-290 (1978); Lasker, The Court Crunch: A View from the Bench, 76 F. R. D. 245, 252 (1978); A. B. A. Litigation Section, Report of the Special Committee for the Study of Discovery Abuse (Oct. 1977); Stanley, President's Page, 62 A. B. A. J. 1375 (1976); Burger, Agenda for 2000 A. D. — A Need for Systematic Anticipation, 70 F. R. D. 83, 95-96 (1976); 4 J. Moore, Federal Practice ¶ 26.02 [3] (2d ed. 1976).
See Anderson, Libel and Press Self-Censorship, 53 Texas L. Rev. 422, 435 (1975).
As the facts of the instant case illustrate, that expense can be considerable. The deposition of Lando alone consumed 26 days and close to 3,000 pages of transcript. See 568 F. 2d 974, 982 (CA2 1977).
Although the separate opinions of my Brothers Powell and Stewart display greater solicitude for First Amendment values than does the opinion for the Court, I believe that they too elide the critical issue presented by this case. Under the “broad and liberal” standard of Hickman, surely *206disclosure of what was known to a journalist but “was not published,” ante, at 200 (opinion of Stewart, J.), will often be germane to whether that individual proceeded with deliberate or reckless disregard for the truth. And admonishing district courts to monitor discovery in the “interest of justice,” ante, at 180 (opinion of Powell, J.) or to prevent “undue burden or expense,” ibid., adds little to the guidance already afforded by Rule 26 and cannot adequately mitigate the burdens on the press so long as Hickman’s directive remains in force. Moreover, neither opinion is directly responsive to the effect of discovery on editorial discussion. See infra, at 208-209.
Respondents in this case produced a considerable amount of evidence regarding preparation of the broadcast:
"Lando answered innumerable questions about what he knew, or had seen; whom he interviewed; intimate details of his discussions with interviewees; and the form and frequency of his communications with sources. The exhibits produced included transcripts of his interviews; volumes of reporters notes; videotapes of interviews; and a series of drafts of the ‘60 *208Minutes’ telecast. Herbert also discovered the contents of pre-telecast conversations between Lando and Wallace as well as reactions to documents considered by both.” 568 F. 2d, at 982 (footnote omitted).
As an abstract proposition, it is not self-evident why disclosure of this material, for which no privilege was sought, would be less likely to inhibit the final publication than state-of-mind inquiries, which in most cases would presumably elicit self-serving responses. Indeed, as the Court acknowledges, plaintiffs may “rarely be successful in proving awareness of falsehood from the mouth of the defendant himself.” Ante, at 170.
Thus, I seriously doubt that state-of-mind questions will substantially “increase the likelihood of large damages judgments in libel actions.” Ante, at 191 (opinion of BrennaN, J.). But neither can it be disputed that such questions might on occasion generate answers useful to plaintiffs in defamation suits. See, e. g., Davis v. Schuchat, 166 U. S. App. D. C. 351, 355-356, 510 F. 2d 731, 735-736 (1975); Goldwater v. Ginzburg, 414 F. 2d 324, 334-335 (CA2 1969), cert. denied, 396 U. S. 1049 (1970); Varnish v. Best Medium Publishing Co., 405 F. 2d 608, 612 (CA2 1968), cert. denied, 394 U. S. 987 (1969).
Contrary to the Court’s intimation, ante, at 165,169-170, this would not be the first instance in which protection apart from the Sullivan malice standard has been extended to safeguard the constitutional interests implicated in libel suits. For example, lower courts have displayed sensitivity to First Amendment values in assessing motions to compel disclosure of confidential sources, see Cervantes v. Time, Inc., 464 F. 2d 986, 992-994 (CA8 1972), cert. denied, 409 U. S. 1125 (1973), and motions by defendants for summary judgment. See Washington Post Co. v. Keogh, 125 U. S. *210App. D. C. 32, 34-35, 365 F. 2d 965, 967-968 (1966), cert. denied, 385 U. S. 1011 (1967).
Different considerations would, of course, obtain if a privilege for editorial communications were sought in conjunction with criminal proceedings. Cf. New York Times Co. v. Jascalevich, 439 U. S. 1331 (1978) (Marshall, J., in chambers); United States v. Nixon, 418 U. S. 683, 712-713 (1974); Branzburg v. Hayes, 408 U. S. 665 (1972); id., at 741-743 (Stewart, J., dissenting).