delivered the opinion of the Court.
By virtue of the First and Fourteenth Amendments, neither the Federal nor a State Government may make any law “abridging the freedom of speech, or of the press . . . .” The question here is whether those Amendments should be construed to provide further protection for the press when sued for defamation than has hitherto been recognized. More specifically, we are urged to hold for the first time that when a member of the press is alleged to have circulated damaging falsehoods and is sued for injury to the plaintiff's reputation, the plaintiff is barred from inquiring into the editorial processes of those responsible for the publication, even though the inquiry would produce evidence material to the proof of a critical element of his cause of action.
I
Petitioner, Anthony Herbert, is a retired Army officer who had extended wartime service in Vietnam and who received *156widespread media attention in 1969-1970 when he accused his superior officers of covering up reports of atrocities and other war crimes. Three years later, on February 4, 1973, respondent Columbia Broadcasting System, Inc. (CBS), broadcast a report on petitioner and his accusations. The program was produced and edited by respondent Barry Lando and was narrated by respondent Mike Wallace. Lando later published a related article in Atlantic Monthly magazine. Herbert then sued Lando, Wallace, CBS, and Atlantic Monthly for defamation in Federal District Court, basing jurisdiction on diversity of citizenship. In his complaint, Herbert alleged that the program and article falsely and maliciously portrayed him as a liar and a person who had made war-crimes charges to explain his relief from command, and he requested substantial damages for injury to his reputation and to the literary value of a book he had just published recounting his experiences.
Although his cause of action arose under New York State defamation law, Herbert conceded that because he was a “public figure” the First and Fourteenth Amendments precluded recovery absent proof that respondents had published a damaging falsehood “with 'actual malice’ — that is, with knowledge that it was false or with reckless disregard of whether it was false or not.” This was the holding of New York Times Co. v. Sullivan, 376 U. S. 254, 280 (1964), with respect to alleged libels of public officials, and extended to “public figures” by Curtis Publishing Co. v. Butts, 388 U. S. 130 (1967).1 Under this rule, absent knowing falsehood, liability requires proof of reckless disregard for truth, that is, that the defendant “in fact entertained serious doubts as to the truth of his publication.” St. Amant v. Thompson, 390 U. S. 727, 731 (1968). Such “subjective awareness of probable falsity,” Gertz v. Robert Welch, Inc., 418 U. S. 323, 335 n. 6 (1974), may be found if “there are obvious reasons to doubt *157the veracity of the informant or the accuracy of his reports.” St. Amant v. Thompson, supra, at 732.
In preparing to prove his case in light of these requirements, Herbert deposed Lando at length and sought an order to compel answers to a variety of questions to which response was refused on the ground that the First Amendment protected against inquiry into the state of mind of those who edit, produce, or publish, and into the editorial process.2 Applying the standard of Fed. Rule Civ. Proc. 26 (b), which permits discovery of any matter “relevant to the subject matter involved in the pending action” if it would either be admissible in evidence or “appears reasonably calculated to lead to the discovery of admissible evidence,” the District Court ruled that because the defendant’s state of mind was of “central importance” to the issue of malice in the case, it was obvious that the questions were relevant and “entirely appropriate to Herbert’s efforts to discover whether Lando had any reason to doubt the veracity of certain of his sources, or, equally significant, to prefer the veracity of one source over another.” 73 F. R. D. 387, 395, 396 (SDNY 1977). The District Court rejected the claim of constitutional privilege because it found nothing in the First Amendment or the relevant cases to permit or require it to increase the weight of the injured plaintiff’s *158already heavy burden of proof by in effect creating barriers “behind which malicious publication may go undetected and unpunished.” Id., at 394. The case was then certified for an interlocutory appeal under 28 U. S. C. § 1292 (b), and the Court of Appeals agreed to hear the case.3
A divided panel reversed the District Court. 568 F. 2d 974 (CA2 1977). Two judges, writing separate but overlapping opinions, concluded that the First Amendment lent sufficient protection to the editorial processes to protect Lando from inquiry about his thoughts, opinions, and conclusions with respect to the material gathered by him and about his conversations with his editorial colleagues. The privilege not to answer was held to be absolute. We granted certiorari because of the importance of the issue involved. 435 U. S. 922 (1978). We have concluded that the Court of Appeals misconstrued the First and Fourteenth Amendments and accordingly reverse its judgment.
II
Civil and criminal liability for defamation was well established in the common law when the First Amendment was adopted, and there is no indication that the Framers intended to abolish such liability. Until New York Times, the prevailing jurisprudence was that “[l]ibelous utterances [are not] within the area of constitutionally protected speech . . . .” Beauharnais v. Illinois, 343 U. S. 250, 266 (1952); see also Roth v. United States, 354 U. S. 476, 482-483 (1957); Chaplinsky v. New Hampshire, 315 U. S. 568, 571-572 (1942) ; Near v. Minnesota ex rel. Olson, 283 U. S. 697, 707-708 (1931). The accepted view was that neither civil nor crimi*159nal liability for defamatory publications abridges freedom of speech or freedom of the press, and a majority of jurisdictions made publishers liable civilly for their defamatory publications regardless of their intent.4 New York Times and Butts effected major changes in the standards applicable to civil libel actions. Under these cases public officials and public figures who sue for defamation must prove knowing or reckless falsehood in order to establish liability. Later, in Gertz v. Robert Welch, Inc., 418 U. S. 323 (1974), the Court held that nonpublic figures must demonstrate some fault on the defendant’s part and, at least where knowing or reckless untruth is not shown, some proof of actual injury to the plaintiff before liability may be imposed and damages awarded.
These cases rested primarily on the conviction that the common law of libel gave insufficient protection to the First Amendment guarantees of freedom of speech and freedom of press and that to avoid self-censorship it was essential that liability for damages be conditioned on the specified showing of culpable conduct by those who publish damaging falsehood. *160Given the required proof, however, damages liability for defamation abridges neither freedom of speech nor freedom of the press.
Nor did these cases suggest any First Amendment restriction on the sources from which the plaintiff could obtain the necessary evidence to prove the critical elements of his cause of action. On the contrary, New York Times and its progeny made it essential to proving liability that the plaintiff focus on the conduct and state of mind of the defendant. To be liable, the alleged defamer of public officials or of public figures must know or have reason to suspect that his publication is false. In other cases proof of some kind of fault, negligence perhaps,5 is essential to recovery. Inevitably, unless liability is to be completely foreclosed, the thoughts and editorial processes of the alleged defamer would be open to examination.
It is also untenable to conclude from our cases that, although proof of the necessary state of mind could be in the form of objective circumstances from which the ultimate fact could be inferred, plaintiffs may not inquire directly from the defendants whether they knew or had reason to suspect that their damaging publication was in error. In Butts, for example, it is evident from the record that the editorial process had been subjected to close examination and that direct as well as indirect evidence was relied on to prove that the defendant magazine had acted with actual malice. The damages verdict was sustained without any suggestion that plaintiff's proof had trenched upon forbidden areas.6
*161Reliance upon such state-of-mind evidence is by no means a recent development arising from New York Times and similar cases. Rather, it is deeply rooted in the common-law rule, predating the First Amendment, that a showing of malice on the part of the defendant permitted plaintiffs to *162recover punitive or enhanced damages.7 In Butts, the Court affirmed the substantial award of punitive damages which in Georgia were conditioned upon a showing of “wanton or reckless indifference or culpable negligence” or “ ‘ill will, spite, hatred and an intent to injure 388 U. S., at 165-166. Neither Mr. Justice Harlan, id., at 156-162,8 nor Mr. Chief Justice Warren, concurring, id,., at 165-168, raised any question as to the propriety of having the award turn on such a showing or as to the propriety of the underlying evidence, *163which plainly included direct evidence going to the state of mind of the publisher and its responsible agents.9
Furthermore, long before New York Times was decided, certain qualified privileges had developed to protect a publisher from liability for libel unless the publication was made with malice.10 Malice was defined in numerous ways, but in gen*164eral depended upon a showing that the defendant acted with improper motive.11 This showing in turn hinged upon the intent or purpose with which the publication was made, the belief of the defendant in the truth of his statement, or upon the ill will which the defendant might have borne toward the plaintiff.12
*165Courts have traditionally admitted any direct or indirect evidence relevant to. the state of mind of the defendant and necessary to defeat a conditional privilege or enhance damages.13 The rules are applicable to the press and to other defendants alike/14 and it is evident that the courts across the country have long been accepting evidence going to the editorial processes of the media without encountering constitutional objections.15
*166In the face of this history, old and new, the Court of Appeals nevertheless declared that two of this Court’s cases had announced unequivocal protection for the editorial proc*167ess. In each of these cases, 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), we invalidated governmental efforts to pre-empt editorial decision by requiring the publication of specified material. In Columbia Broadcasting System, it was the requirement that a television network air paid political advertisements and in Tornillo, a newspaper’s obligation to print a political candidate’s reply to press criticism. Insofar as the laws at issue in Tornillo and Columbia Broadcasting System sought to control in advance the content of the publication, they were deemed as invalid as were prior efforts to enjoin *168publication of specified materials.16 But holdings that neither a State nor the Federal Government may dictate what must or must not be printed neither expressly nor impliedly suggest that the editorial process is immune from any inquiry whatsoever.
It is incredible to believe that the Court in Columbia Broadcasting System or in Tornillo silently effected a substantial contraction of the rights preserved to defamation plaintiffs in Sullivan, Butts, and like cases. Tornillo and Gertz v. Robert Welch, Inc., were announced on the same day; and although the Court’s opinion in Gertz contained an overview of recent developments in the relationship between the First Amendment and the law of libel, there was no hint that a companion case had narrowed the evidence available to a defamation plaintiff. Quite the opposite inference is to be drawn from the Gertz opinion, since it, like prior First Amendment libel cases, recited without criticism the facts of record indicating that the state of mind of the editor had been placed at issue. Nor did the Gertz opinion, in requiring proof of some degree of fault on the part of the defendant editor and in forbidding punitive damages absent at least reckless disregard of truth or falsity, suggest that the First Amendment also foreclosed direct inquiry into these critical elements.17
*169In sum, contrary to the views of the Court of Appeals, according an absolute privilege to the editorial process of a media defendant in a libel case is not required, authorized, or presaged by our prior cases, and would substantially enhance the burden of proving actual malice, contrary to the expectations of New York Times, Butts, and similar cases.
Ill
It is nevertheless urged by respondents that the balance struck in New York Times should now be modified to provide further protections for the press when sued for circulating erroneous information damaging to individual reputation. It is not uncommon or improper, of course, to suggest the abandonment, modification, or refinement of existing constitutional interpretation, and notable developments in First Amendment jurisprudence have evolved from just such submissions. But in the 15 years since New York Times, the doctrine announced by that case, which represented a major development and which was widely perceived as essentially protective of press freedoms, has been repeatedly affirmed as the appropriate First Amendment standard applicable in libel actions brought by public officials and public figures. Curtis Publishing Co. v. Butts, 388 U. S. 130 (1967); St. Amant v. Thompson, 390 U. S. 727 (1968); Gertz v. Robert Welch, Inc., 418 U. S. 323 (1974); Time, Inc. v. Firestone, 424 U. S. 448 (1976). At the same time, however, the Court has reiterated its conviction — reflected in the laws of defamation of all of the States— that the individual’s interest in his reputation is also a basic concern. Id., at 455-457; Gertz v. Robert Welch, Inc., supra, at 348-349.
We are thus being asked to modify firmly established constitutional doctrine by placing beyond the plaintiff’s reach a range of direct evidence relevant to proving knowing or reckless falsehood by the publisher of an alleged libel, elements that are critical to plaintiffs such as Herbert. The case for *170making this modification is by no means clear and convincing, and we decline to accept it.
In the first place, it is plain enough that the suggested privilege for the editorial process would constitute a substantial interference with the ability of a defamation plaintiff to establish the ingredients of malice as required by New York Times. As respondents would have it, the defendant’s reckless disregard of the truth, a critical element, could not be shown by direct evidence through inquiry into the thoughts, opinions, and conclusions of the publisher, but could be proved only by objective evidence from which the ultimate fact could be inferred. It may be that plaintiffs will rarely be successful in proving awareness of falsehood from the mouth of the defendant himself, but the relevance of answers to such inquiries, which the District Court recognized and the Court of Appeals did not deny, can hardly be doubted. To erect an impenetrable barrier to the plaintiff’s use of such evidence on his side of the case is a matter of some substance, particularly when defendants themselves are prone to assert their good-faith belief in the truth of their publications,18 and libel plaintiffs are required to prove knowing or reckless falsehood with “convincing clarity.” New York Times Co. v. Sullivan, 376 U. S., at 285-286.
Furthermore, the outer boundaries of the editorial privilege now urged are difficult to perceive. The opinions below did not state, and respondents do not explain, precisely when the editorial process begins and when it ends. Moreover, although we are told that respondent Lando was willing to testify as to what he “knew” and what he had “learned” from his interviews, as opposed to what he “believed,” it is not at all clear why the suggested editorial privilege would not cover knowledge as well as belief about the veracity of published *171reports.19 It is worth noting here that the privilege as asserted by respondents would also immunize from inquiry the internal communications occurring during the editorial process and thus place beyond reach what the defendant participants learned or knew as the result of such collegiate conversations or exchanges. If damaging admissions to colleagues are to be barred from evidence, would a reporter’s admissions made to third parties not participating in the editorial process also be immune from inquiry? We thus have little doubt that Herbert and other defamation plaintiffs have important interests at stake in opposing the creation of the asserted privilege.
Nevertheless, we are urged by respondents to override these important interests because requiring disclosure of editorial conversations and of a reporter’s conclusions about the veracity of the material he has gathered will have an intolerable chilling effect on the editorial process and editorial decision-making. But if the claimed inhibition flows from the fear of damages liability for publishing knowing or reckless falsehoods, those effects are precisely what New York Times and other cases have held to be consistent with the First Amendment. Spreading false information in and of itself carries no First Amendment credentials. “ [T]here is no constitutional value in false statements of fact.” Gertz v. Robert Welch, Inc., supra, at 340.
Realistically, however, some error is inevitable; and the difficulties of separating fact from fiction convinced the Court in New York Times, Butts, Gertz, and similar cases to limit *172liability to instances where some degree of culpability is present in order to eliminate the risk of undue self-censorship and the suppression of truthful material. Those who publish defamatory falsehoods with the requisite culpability, however, are subject to liability, the aim being not only to compensate for injury but also to deter publication of unprotected material threatening injury to individual reputation. Permitting plaintiffs such as Herbert to prove their cases by direct as well as indirect evidence is consistent with the balance struck by our prior decisions. If such proof results in liability for damages which in turn discourages the publication of erroneous information known to be false or probably false, this is no more than what our cases contemplate and does not abridge either freedom of speech or of the press.
Of course, if inquiry into editorial conclusions threatens the suppression not only of information known or strongly suspected to be unreliable but also of truthful information, the issue would be quite different. But as we have said, our cases necessarily contemplate examination of the editorial process to prove the necessary awareness of probable falsehood, and if indirect proof of this element does not stifle truthful publication and is consistent with the First Amendment, as respondents seem to concede, we do not understand how direct inquiry with respect to the ultimate issue would be substantially more suspect.20 Perhaps such examination will lead to liability that would not have been found without it, but this does not suggest that the determinations in these instances will be inaccurate and will lead to the suppression of protected information. On the contrary, direct inquiry from the actors, which affords the opportunity to refute inferences that might otherwise be drawn from circumstantial evidence, suggests *173that more accurate results will be obtained by placing all, rather than part, of the evidence before the decisionmaker. Suppose, for example, that a reporter has two contradictory reports about the plaintiff, one of which is false and damaging, and only the false one is published. In resolving the issue whether the publication was known or suspected to be false, it is only common sense to believe that inquiry from the author, with an opportunity to explain, will contribute to accuracy. If the publication is false but there is an exonerating explanation, the defendant will surely testify to this effect.21 Why should not the plaintiff be permitted to inquire before trial? On the other hand, if the publisher in fact had serious doubts about accuracy, but published nevertheless, no undue self-censorship will result from permitting the relevant inquiry. Only knowing or reckless error will be discouraged; and unless there is to be an absolute First Amendment privilege to inflict injury by knowing or reckless conduct, which respondents do not suggest, constitutional values will not be threatened.
It is also urged that frank discussion among reporters and editors will be dampened and sound editorial judgment endangered if such exchanges, oral or written, are subject to inquiry by defamation plaintiffs.22 We do not doubt the direct relationship between consultation and discussion on the one hand and sound decisions on the other; but whether or not there is liability for the injury, the press has an obvious interest in avoiding the infliction of harm by the publication *174of false information, and it is not unreasonable to expect the media to invoke whatever procedures may be practicable and useful to that end. Moreover, given exposure to liability when there is knowing or reckless error, there is even more reason to resort to prepublication precautions, such as a frank interchange of fact and opinion. Accordingly, we find it difficult to believe that error-avoiding procedures will be terminated or stifled simply because there is liability for culpable error and because the editorial process will itself be examined in the tiny percentage of instances in which error is claimed and litigation ensues. Nor is there sound reason to believe that editorial exchanges and the editorial process are so subject to distortion and to such recurring misunderstanding that they should be immune from examination in order to avoid erroneous judgments in defamation suits. The evidentiary burden Herbert must carry to prove at least reckless disregard for the truth is substantial indeed, and we are unconvinced that his chances of winning an undeserved verdict are such that an inquiry into what Lando learned or said during the editorial process must be foreclosed.
This is not to say that the editorial discussions or exchanges have no constitutional protection from casual inquiry. There is no law that subjects the editorial process to private or official examination merely to satisfy curiosity or to serve some general end such as the public interest; and if there were, it would not survive constitutional scrutiny as the First Amendment is presently construed. No such problem exists here, however, where there is a specific claim of injury arising from a publication that is alleged to have been knowingly or recklessly false.23
*175Evidentiary privileges in litigation are not favored,24 and even those rooted in the Constitution must give way in proper circumstances. The President, for example, does not have an absolute privilege against disclosure of materials subpoenaed for a judicial proceeding. United States v. Nixon, 418 U. S. 683 (1974). In so holding, we found that although the President has a powerful interest in confidentiality of communications between himself and his advisers, that interest must yield to a demonstrated specific need for evidence. As we stated, in referring to existing limited privileges against disclosure, “[w]hatever their origins, these exceptions to the demand for every man’s evidence are not lightly created nor expansively construed, for they are in derogation of the search for truth.” Id., at 710.
With these considerations in mind, we conclude that the present construction of the First Amendment should not be modified by creating the evidentiary privilege which the respondents now urge.
IV
Although defamation litigation, including suits against the press, is an ancient phenomenon, it is true that our cases from New York Times to Gertz have considerably changed the profile of such cases. In years gone by, plaintiffs made out a prima facie case by proving the damaging publication. Truth *176and privilege were defenses. Intent, motive, and malice were not necessarily involved except to counter qualified privilege or to prove exemplary damages. The plaintiff’s burden is now considerably expanded. In every or almost every case, the plaintiff must focus on the editorial process and prove a false publication attended by some degree of culpability on the part of the publisher. If plaintiffs in consequence now resort to more discovery, it would not be surprising; and it would follow that the costs and other burdens of this kind of litigation would escalate and become much more troublesome for both plaintiffs and defendants. It is suggested that the press needs constitutional protection from these burdens if it is to perform its task,25 which is indispensable in a system such as ours.
Creating a constitutional privilege foreclosing direct inquiry into the editorial process, however, would not cure this problem for the press. Only complete immunity from liability for defamation would effect this result, and the Court has regularly found this to be an untenable construction of the First Amendment. Furthermore, mushrooming litigation costs, much of it due to pretrial discovery, are not peculiar to the libel and slander area. There have been repeated expressions of concern about undue and uncontrolled discovery, and voices from this Court have joined the chorus.26 But *177until and unless there are major changes in the present Rules of Civil Procedure, reliance must be had on what in fact and in law are ample powers of the district judge to prevent abuse.
The Court has more than once declared that the deposition-discovery rules are to be accorded a broad and liberal treatment to effect their purpose of adequately informing the litigants in civil trials. Schlagenhauf v. Holder, 379 U. S. 104, 114-115 (1964); Hickman v. Taylor, 329 U. S. 495, 501, 507 (1947). But the discovery provisions, like all of the Federal Rules of Civil Procedure, are subject to the injunction of Rule 1 that they “be construed to secure the just, speedy, and inexpensive determination of every action.” (Emphasis added.) To this end, the requirement of Rule 26 (b) (1) that the material sought in discovery be “relevant” should be firmly applied, and the district courts should not neglect their power to restrict discovery where “justice requires [protection for] a party or person from annoyance, embarrassment, oppression, or undue burden or expense . . . .” Rule 26 (c). With this authority at hand, judges should not hesitate to exercise appropriate control over the discovery process.
Whether, as a nonconstitutional matter, however, the trial judge properly applied the rules of discovery was not within the boundaries of the question certified under 28 U. S. C. § 1292 (b) and accordingly is not before us.27 The judgment of the Court of Appeals is reversed.
So ordered.
Criminal libel prosecutions are subject to the same constitutional limitations. Garrison v. Louisiana, 379 U. S. 64 (1964).
The Court of Appeals summarized the inquiries to which Lando objected as follows:
“1. Lando’s conclusions during his research and investigations regarding people or leads to be pursued, or not to be pursued, in connection with the ‘60 Minutes’ segment and the Atlantic Monthly article;
"2. Lando’s conclusions about facts imparted by interviewees and his state of mind with respect to the veracity of persons interviewed;
“3. The basis for conclusions where Lando testified that he did reach a conclusion concerning the veracity of persons, information or events;
“4. Conversations between Lando and Wallace about matter to be included or excluded from the broadcast publication; and
“5. Lando’s intentions as manifested by his decision to include or exclude certain material.” 568 F. 2d 974, 983 (CA2 1977).
Respondents’ petition for leave to appeal from an interlocutory order, which was granted, stated the issue on appeal as follows:
“What effect should be given to the First Amendment protection of the press with respect to its exercise of editorial judgment in pre-trial discovery in a libel case governed by New York Times Co. v. Sullivan, 376 U. S. 254 (1964)?”
See, e. g., Restatement of Torts §580 (1938); Pedrick, Freedom of the Press and the Law of Libel: The Modern Revised Translation, 49 Corn. L. Q. 581, 583-584 (1964); Developments in the Law — Defamation, 69 Harv. L. Rev. 875, 902-910 (1956). In Peck v. Tribune Co., 214 U. S. 185, 189 (1909), Mr. Justice Holmes summarized the prevailing view of strict liability in the course of reviewing a libel judgment rendered in a federal diversity of citizenship action:
“There was some suggestion that the defendant published the portrait by mistake, and without knowledge that it was the plaintiff’s portrait or was not what it purported to be. But the fact, if it was one, was no excuse. If the publication was libellous the defendant took the risk. As was said of such matters by Lord Mansfield, ‘Whatever a man publishes he publishes at his peril.' The King v. Woodfall, Lofft 776, 781. . . . The reason is plain. A libel is harmful on its face. If a man sees fit to publish manifestly hurtful statements concerning an individual, without other justification than exists for an advertisement or a piece of news, the usual principles of tort will make him liable, if the statements are false or are true only of some one else.”
The definition of fault was to be the responsibility of state laws. Gertz v. Robert Welch, Inc., 418 U. S. 323, 347 (1974).
See 388 U. S., at 156-169, where Mr. Justice Harlan, writing for a plurality of the Court, reviewed the record under the standard he preferred to apply to public figures, and upheld the verdict for the plaintiff. Mr. Chief Justice Warren independently reviewed the record under the “actual malice” standard of New York Times and also concluded in his concurring opinion *161that the verdict should be upheld. Id., at 168-170. The evidence relied on and summarized in both opinions included substantial amounts of testimony that would fall within the editorial-process privilege as defined by respondents. The record before the Court included depositions by the author of the defamatory article, an individual paid to assist the author in preparation, the sports editor of the Saturday Evening Post, and both its managing editor and editor in chief. These depositions revealed the Saturday Evening Post’s motives in publishing the story (Record, O. T. 1966, No. 37, pp. 706-717), sources {id., at 364, 662-664, 719-720, 729), conversations among the editors and author concerning the research and development of the article {id., at 363-367, 721-737), decisions and reasons relating to who should be interviewed and what should be investigated {id., at 666-667, 699-700, 734-736, 772-774), conclusions as to the importance and veracity of sources and information presented in the article {id., at 720, 732-735, 737, 771-772, 776), and conclusions about the impact that publishing the article would have on the subject {id., at 714^-716, 770). Mr. Justice BreNNAN, writing for himself and Mr. Justice White, also thought the evidence of record sufficient to satisfy the New York Times malice standard. It is quite unlikely that the Court would have arrived at the result it did had it believed that inquiry into the editorial processes was constitutionally forbidden.
The Court engaged in similar analysis of the record in reversing the judgments entered in a companion case to Butts, Associated Press v. Walker, 388 U. S., at 158-159; id., at 165 (Warren, C. J., concurring); and in Time, Inc. v. Hill, 385 U. S. 374, 391-394 (1967). In Hill, the record included the edited drafts of the allegedly libelous article and an examination and cross-examination of the author. During that examination, the writer explained in detail the preparation of the article, his thoughts, conclusions, and beliefs regarding the material, and a line-by-line analysis of the article with explanations of how and why additions and deletions were made to the various drafts. As in Butts, the editorial process was the focus of much of the evidence, and direct inquiry was made into the state of mind of the media defendants. Yet the Court raised no question as to the propriety of the proof.
A. Hanson, Libel and Related Torts ¶ 163 (1969); Developments in the Law — Defamation, supra n. 4, at 938; 50 Am. Jur. 2d, Libel and Slander § 352 (1970); 53 C. J. S., Libel and Slander § 260 (1955).
The Restatement originally provided in a separate section for the award of punitive damages for malicious defamations. Restatement of Torts § 1068 (Tent. Draft 13, 1936) :
“One who is liable for harm to another’s reputation caused by the publication of a libel or slander is also liable for punitive damages if the defamatory matter was published with knowledge of its falsity or if it was published in reckless indifference to its truth or falsity or solely for the purpose of causing harm to the plaintiff’s reputation or other legally protected interest.”
The provision was later omitted with the explanation that recovery of punitive damages would be determined by the rules in the Restatement with respect to damages in general. Restatement of Torts § 1068 (Proposed Final Draft 3, 1937).
Gertz v. Robert Welch, Inc., supra, at 350, limited the entitlement to punitive damages, but such damages are still awardable upon a showing of knowing or reckless falsehood.
As Mr. Justice Harlan noted, the jury had been instructed in considering punitive damages to assess “ ‘the reliability, the nature of the sources of the defendant’s information, its acceptance or rejection of the sources, and its care in checking upon assertions.’ ” 388 U. S., at 156 (emphasis added). The Justice found nothing amiss either with the instruction or the result the jury reached under it. Mr. Justice BreNNAN, dissenting in the Butts case, id., at 172-174, analyzed the instructions differently but raised no question as to the constitutionality of turning the award of either compensatory or punitive damages upon direct as well as circumstantial evidence going to the mental state of the defendant.
See n. 6, supra.
See Nalle v. Oyster, 230 U. S. 165, 179-180 (1913); White v. Nicholls, 3 How. 266, 286-292 (1845); T. Plucknett, A Concise History of the Common Law 502 (5th ed. 1956); Hallen, Character of Belief Necessary for the Conditional Privilege in Defamation, 25 Ill. L. Rev. 865 (1931). In White v. Nicholls, supra, at 290-291, the Court surveyed the common law and summarized the privilege as follows:
“We have thus taken a view of the authorities which treat of the doctrines of slander and libel, and have considered those authorities particularly with reference to the distinction they establish between ordinary instances of slander, written and unwritten, and those which have been styled privileged communications; the peculiar character of which is said to exempt them from inferences which the law has created with respect to those cases that do not partake of that character. Our examination, extended as it may seem to have been, has been called for by the importance of a subject most intimately connected with the rights and happiness of individuals, as it is with the quiet and good order of society. The investigation has conducted us to the following conclusions, which we propound as the law applicable thereto. 1. That every publication, either by writing, printing, or pictures, which charges upon or imputes to any person that which renders him liable to punishment, or which is calculated to make him infamous, or odious, or ridiculous, is prima jade a libel, and implies malice in the author and publisher towards the person concerning whom such publication is made. Proof of malice, therefore, in the cases just described, can never be required of the party complaining beyond the proof of the publication itself: justification, excuse, or extenuation, if either can be shown, must proceed from the defendant. 2. That the description of cases recognised as privileged communications, must be understood as exceptions to this rule, and as being founded upon some apparently recognised obligation or motive, legal, moral, or social, which may fairly be presumed to have led to the publication, and therefore prima facie relieves it from that just implication from which the general rule of the law is deduced. The rule of evidence, as to such cases, is accordingly *164so far changed as to impose it on the plaintiff to remove those presumptions flowing from the seeming obligations and situations of the parties, and to require of him to bring home to the defendant the existence of malice as the true motive of his conduct. Beyond this extent no presumption can be permitted to operate, much less be made to sanctify the indulgence of malice, however wicked, however express, under the protection of legal forms. We conclude then that malice may be proved, though alleged to have existed in the proceedings before a court, or legislative body, or any other tribunal or authority, although such court, legislative body, or other tribunal, may have been the appropriate authority for redressing the grievance represented to it; and that proof of express malice in any written publication, petition, or proceeding, addressed to such tribunal, will render that publication, petition, or proceeding, libellous in its character, and actionable, and will subject the author and publisher thereof to all the consequences of libel.”
Hallen, supra, at 866-867. In some jurisdictions a defendant forfeited his privilege if he published negligently or without probable cause to believe the statement was true. Id., at 867; see White v. Nicholls, supra, at 291.
See, e. g., 50 Am. Jur. 2d, supra n. 7, § 455:
“The existence of actual malice may be shown in many ways. As a general rule, any competent evidence, either direct or circumstantial, can be resorted to, and all the relevant circumstances surrounding the transaction may be shown, provided they are not too remote, including threats, prior or subsequent defamations, subsequent statements of the defendant, circumstances indicating the existence of rivalry, ill will, or hostility between the parties, facts tending to show a reckless disregard of the plaintiff’s rights, and, in an action against a newspaper, custom and usage with respect to the treatment of news items of the nature of the one under consideration. The plaintiff may show that the defendant had drawn a pistol at the time he uttered the words complained of; that defendant had tried to kiss and embrace plaintiff just prior to the defamatory publication; or that defendant had failed to make a proper investigation before publication of the statement in question. On cross-examination the *165defendant may be questioned as to his intent in making the publication.” (Footnotes and citations omitted.)
E. g., W. Odgers, A Digest of the Law of Libel and Slander *271-*288 (1st Am. ed. Bigelow 1881); 50 Am. Jur. 2d, supra n. 7, § 455; 53 C. J. S., supra n. 7, § 213.
Cf. Odgers, supra, at *271; F. Holt, The Law of Libel 57 (1st Am. ed. 1818); Billet v. Times-Democrat Publishing Co., 107 La. 751, 32 So. 17 (1902).
In scores of libel cases, courts have addressed the general issue of the admissibility of evidence that would be excluded under the editorial-process privilege asserted here and have affirmed the relevance and admissibility of the evidenoe on behalf of libel plaintiffs. See, e. g., Johnson Publishing Co. v. Davis, 271 Ala. 474, 124 So. 2d 441 (1960) (editor may be cross-examined on meaning intended to be conveyed by passages in magazine article) ; Freeman v. Mills, 97 Cal. App. 2d 161, 217 P. 2d 687 (1950) (malice may be established by direct proof of the state of mind of a person, or by evidence from which its existence may be inferred); Scott v. Times-Mirror Co., 181 Cal. 345, 184 P. 672 (1919) (all relevant circumstances concerning publication admissible); Sandora v. Times Co., 113 Conn. 574, 155 A. 819 (1931) (all relevant evidence including direct evidence on state of mind or surrounding circumstances — city editor and reporter called to stand and questioned extensively as to motives, circumstances of publication, and general practices) ; Rice v. Simmons, 2 Del. 309, 31 Am. Dec. 766 (1838) (where question of malice in issue, declarations of publisher at the time of publication admissible as part of the res gestae); Western Union Telegraph Co. v. Vickers, 71 Ga. App. 204, 30 S. E. 2d 440 (1944) (aE relevant evidence admissible, including direct evidence of state of mind and surrounding circumstances) ; Cook v. East Shore Newspapers, 327 Ill. App. 559, 64 N. E. 2d 751 (1945) (all relevant evidence concerning circumstances of publications admissible, including testimony by reporters and employees of defendant); Berger v. Freeman Tribune Publishing Co., 132 Iowa 290, 109 N. W. 784 *166(1906) (all relevant evidence); Thompson v. Globe Newspaper Co., 279 Mass. 176, 181 N. E. 249 (1932) (only evidence on state of mind of those agents of defendant .entrusted with determining what shall be published is admissible and material); Conroy v. Fall River Herald News Co., 306 Mass. 488, 28 N. E. 2d 729 (1940) (any relevant evidence on defendant’s malice); Cyrowski v. Polish-American Pub. Co., 196 Mich. 648, 163 N. W. 58 (1917) (testimony of individuals who advised reporter to question plaintiff before publishing defamatory article was admissible on the issue of malice); Friedell v. Blakely Printing Co., 163 Minn. 226, 203 N. W. 974 (1925) (any relevant evidence admissible); Cook v. Globe Printing Co., 227 Mo. 471, 127 S. W. 332 (1910) (evidence showing that defendant’s editorial manager knew an important fact to be false admissible on question of malice); Butler v. Gazette Co., 119 App. Div. 767, 104 N. Y. S. 637 (1907) (any evidence admissible to prove actual malice of defendant); Briggs v. Byrd, 34 N. C. 377 (1851) (express malice may be proved either by direct evidence or surrounding circumstances); McBurney v. Times Publishing Co., 93 R. I. 331, 175 A. 2d 170 (1961) (relevant evidence admissible to rebut testimony by reporters and editors that they published without malice); Lancour v. Herald & Globe Assn., 112 Vt. 471, 28 A. 2d 396 (1942) (any relevant evidence on malice); Farrar v. Tribune Publishing Co., 57 Wash. 2d 549, 358 P. 2d 792 (1961) (all circumstances surrounding publication relevant and admissible).
Similarly, the courts have uniformly admitted such evidence on behalf of the defendant. See, e. g., Bohan v. Record Pub. Co., 1 Cal. App. 429, 82 P. 634 (1905) (testimony on good faith); Hearne v. De Young, 119 Cal. 670, 52 P. 150 (1898) (testimony on sources, precautions taken, and good faith); Ballinger v. Democrat Co., 203 Iowa 1095, 212 N. W. 557 (1927) (testimony of reporter and editor on good faith admissible); Snyder v. Tribune Co., 161 Iowa 671, 143 N. W. 519 (1913) (testimony as to source of information and good faith of reporter admissible); Courier-Journal Co. v. Phillips, 142 Ky. 372, 134 S. W. 446 (1911) (testimony of reporter on good faith); Conner v. Standard Pub. Co., 183 Mass. 474, 67 N. E. 596 (1903) (testimony as to source of information); Davis v. Marxhausen, 103 Mich. 315, 61 N. W. 504 (1894) (testimony on good faith and proper precautions taken before publishing); Julian v. Kansas City Star Co., 209 Mo. 35, 107 S. W. 496 (1908) (testimony on thoughts and intentions at the time of publication admissible); Paxton v. Woodward, 31 Mont. 195, 78 P. *167215 (1904) (testimony as to motive, good faith, and sources); Las Vegas Sun, Inc. v. Franklin, 74 Nev. 282, 329 P. 2d 867 (1958) (testimony of publisher on good faith); Lindsey v. Evening Journal Assn., 10 N. J. Misc. 1275, 163 A. 245 (1932) (testimony on good faith); Kohn v. P&D Publishing Co., 169 App. Div. 580, 155 N. Y. S. 455 (1915) (source); Hains v. New York Evening Journal, 240 N. Y. S. 734 (Sup. Ct. 1930) (source); Goodrow V. Malone Telegram, Inc., 235 App. Div. 3, 255 N. Y. S. 812 (1932) (reporter’s testimony as to source); Goodrow v. Press Co., 233 App. Div. 41, 251 N. Y. S. 364 (1931) (defendant can testify and introduce evidence on his good faith at time of publication); Kehoe v. New York Tribune, 229 App. Div. 220, 241 N. Y. S. 676 (1930) (testimony on good faith admissible to prevent imposition of punitive damages); Varvaro v. American Agriculturist, Inc., 222 App. Div. 213, 225 N. Y. S. 564 (1927) (defendant may testify and introduce evidence on lack of malice); Van Arsdale v. Time, Inc., 35 N. Y. S. 2d 951 (Sup. Ct.), aff’d, 265 App. Div. 919, 39 N. Y. S. 2d 413 (1942); Weichbrodt v. New York Evening Journal, 11 N. Y. S. 2d 112 (Sup. Ct. 1939) (defendant may testify as to good faith and probable cause); Cleveland Leader Printing Co. v. Nethersole, 84 Ohio St. 118, 95 N. E. 735 (1911) (testimony on good faith); Cobb v. Oklahoma Pub. Co., 42 Okla. 314, 140 P. 1079 (1914) (defendant’s testimony as to lack of malice and source of information); Times Pub. Co. v. Ray, 1 S. W. 2d 471 (Tex. Civ. App. 1927), aff’d, 12 S. W. 2d 165 (1929) (testimony as to lack of malice); Pfister v. Milwaukee Free Press Co., 139 Wis. 627, 121 N. W. 938 (1909) (testimony as to absence of malice).
None of these eases as much as suggested that there were special limits applicable to the press on the discoverability of such evidence, either before or during trial.
As we stated in Tornillo, “no 'government agency — local, state, or federal — can tell a newspaper in advance what it can print and what it cannot.’ ” 418 U. S., at 255-256, quoting Pittsburgh Press Co. v. Human Relations Comm’n, 413 U. S. 376, 400 (1973) (Stewart, J., dissenting).
Two years later, in Time, Inc. v. Firestone, 424 U. S. 448 (1976), there was likewise no indication that the plaintiff is subject to substantial evi-dentiary restrictions in proving the defendant’s fault. As Mr. Justice Powell and Mr. Justice Stewart stated in concurrence, the answer to this question of culpability “depends upon a careful consideration of all the relevant evidence concerning Time’s actions prior to the publication of the 'Milestones’ article.” Id., at 465-466. They suggested that on remand all the evidence of record should be considered, which included evidence going to the beliefs of Time’s editorial staff. See id., at 467-470, and n. 5.
See, e. g., the cases collected in n. 15, supra, in which media defendants asserted, and courts upheld, the right to present this type of evidence at trial in order to establish good faith and lack of malice.
It was also suggested at oral argument that the privilege would cover questions in the “why” form, but not of the “who,” “what,” “when,” and “where” type. Tr. of Oral Arg. 32-34. But it is evident from Lando’s deposition that questions soliciting “why” answers relating to the editorial process were answered, e. g., Tr. of Deposition 21, L. 7; 1892, L. 18, and that he refused to answer others that did not fall into this category, e. g., id., at 666, L. 20; 774, L. 5; 877, L. 12; 880, L. 5; 1488, L. 3; 1893, L. 11; see Tr. of Oral Arg. 46.
The kind of question respondents seek to avoid answering is, by their own admission, the easiest to answer. See Tr. of Oral Arg. 31:
“[T]hey are set-up questions for our side. . . . [T]hese are not difficult questions to answer.”
Often it is the libel defendant who first presents at trial direct evidence about the editorial process in order to establish good faith and lack of malice. That was true in New York Times Co. v. Sullivan, see, e. g., Record, O. T. 1963, No. 39, p. 762, and in many of the cases cited in n. 15, supra.
They invoke our observation 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 and for their own interests to the detriment of the decisionmaking process.”
Mr. Justice BrenNAN would extend more constitutional protection to editorial discussion by excusing answers to relevant questions about in-house conversations until the plaintiff has made a prima facie ease of falsity. If this suggestion contemplates a bifurcated trial, first on falsity and then on culpability and injury, we decline to subject libel trials to such burdensome complications and intolerable delay. On the other hand, *175if, as seems more likely, the prima facie showing does not contemplate a minitrial on falsity, no resolution of conflicting evidence on this issue, but only a credible assertion by the plaintiff, it smacks of a requirement that could be satisfied by an affidavit or a simple verification of the pleadings. We are reluctant to imbed this formalism in the Constitution.
See Elkins v. United States, 364 U. S. 206, 234 (1960) (Frankfurter, J., dissenting): “Limitations are properly placed upon the operation of this general principle [of no testimonial privilege] only to the very limited extent that permitting a refusal to testify or excluding relevant evidence has a public good transcending the normally predominant principle of utilizing all rational means for ascertaining truth.” See also 8 J. Wigmore, Evidence §2192 (MeNaughton rev. 1961); 4 The Works of Jeremy Bentham 321 (J. Bowring ed. 1843).
It is urged that the large costs of defending lawsuits will intimidate the press and lead to self-censorship, particularly where smaller newspapers and broadcasters are involved. It is noted that Lando’s deposition alone continued intermittently for over a year and filled 26 volumes containing nearly 3,000 pages and 240 exhibits. As well as out-of-pocket expenses of the deposition, there were substantial legal fees, and Lando and his associates were diverted from news gathering and reporting for a significant amount of time.
Blue Chip Stamps v. Manor Drug Stores, 421 U. S. 723, 740-741 (1975); ACF Industries, Inc. v. EEOC, 439 U. S. 1081 (1979) (Powell, J., joined by Stewart and Rehnquist, JJ., dissenting from denial of certiorari); Burger: Agenda for 2000 A. D.: A Need for Systematic Anticipation, Address at the Pound Conference, 70 F. R. D. 83, 95-96 *177(1976). The Committee on Rules of Practice and Procedure of the Judicial Conference of the United States has proposed amendments to the Federal Rules of Civil Procedure designed to ameliorate this problem. Preliminary Draft of Proposed Amendments to the Federal Rules of Civil Procedure (1978).
Mr. Justice Stewart would remand to have the trial court rule once again on the relevance of the disputed questions. But the opinion of the *178trial judge reveals that he correctly understood that New York Times and Gertz required Herbert to prove either knowing falsehood or reckless disregard for truth. With the proper constitutional elements in mind, the judge went on to rule that the questions at issue were clearly relevant and that no constitutional privilege excused Lando from answering them. We hold that the judge committed no constitutional error but, contrary to Mr. Justice Stewart, find it inappropriate to review his rulings on relevancy.