Full opinion text
Opinion for the Court filed by Senior Circuit Judge MacKINNON. Opinion concurring in part and dissenting in part filed by Circuit Judge J. SKELLY WRIGHT. TABLE OF CONTENTS Page I. Background Facts................................................ 98 II. The Legal Standard.............................................. 103 III. Defamation and Falsity.......................................... 109 1. Defamatory Character of the November 30 Article.............. 109 2. Falsity........................................................ Ill IV. Reckless Disregard of Truth or Falsity — Actual Malice........ 114 1. The Christine Peterson Memorandum........................... 114 2. The Defendants’ Motivation.................................... 117 3. Resolution of Inferences Adverse to the Plaintiffs .............. 121 4. Suppression of Information Favorable to the Plaintiffs.......... 123 5. “Personally Dispatched”: The Checket Conversation............. 125 6. The SEC Paragraphs.......................................... 126 7. Reliability of Sources.......................................... 127 A. Piro...................................................... 128 B. Comnas .......:.......................................... 129 8. Other Indicia of Actual Malice................................. 131 9. The Defendants’ Evidence ..................................... 132 10. Some of Defendants’ Suggested Inferences from the Evidence ... 133 11. Summary..................................................... 134 V. The Verdict Against Piro........................................ 135 VI. The Verdict Against Golden..................................... 136 VII. Conclusion....................................................... 137 Appendix............................................................... 138 MacKINNON, Senior Circuit Judge. Plaintiffs William and Peter Tavoulareas brought suit against The Washington Post (“Post") and several other defendants for libel and against defendant Piro, a source for the story, for slander and its foreseeable republication. The Tavoulareases alleged that they were defamed by articles in the Post which stated, among other things, that William Tavoulareas, as President and Chief Executive Officer of Mobil Oil Corporation (“Mobil”), had used his influence to “set up” his son Peter in the shipping business, and then had diverted some of Mobil’s shipping business to him. The basic theme of the article was that William Tavoulareas had misused his position and corporate assets to benefit his son. The case was submitted to the jury, which was instructed, in accordance with standards constitutionally required for public figures, that defendants could be held liable only if they published false matter with “actual malice” — i.e., with knowledge of its falsity or reckless disregard of whether it was false or not. The jury returned verdicts for the plaintiffs. The trial judge then ruled on motion that the evidence was insufficient to support such a verdict and entered judgment notwithstanding the verdict (“n.o.v.”) for the defendants. Tavoulareas v. Washington Post Co., 567 F.Supp. 651 (D.D.C. 1983). After a careful and independent review of the entire record in the case, we conclude that the evidence adduced by the plaintiffs was sufficient to “establish [] actual malice with convincing clarity,” Bose Corp. v. Consumers Union of United States, — U.S.-,-, 104 S.Ct. 1949, 1967, 80 L.Ed.2d 502 (1984), and accordingly reverse the grant of the judgments n.o.v. as to the Post defendants and defendant Piro, and remand the case to the district court for further proceedings. We affirm the trial court’s judgment n.o.v. with respect to defendant Golden because of his lack of responsibility for the publication. I. Background Facts The business relationships that gave rise to this case began in 1974, when Mobil bought a 30 percent interest in the Saudi Arabian Maritime Company (“Samareo”), a joint venture that had been formed a short time before by Fairfield Maxwell Ltd. and members of a powerful Saudi Arabian family, the Alirezas. The Alirezas were the majority shareholders in Samareo. Mobil bought a minority interest in Samareo because, at that time, Mobil believed the Saudi government was going to give strong oil shipping preferences to Saudi-owned shippers. Ultimately, however, these preferences were never adopted by the Saudi government. The Samareo partners agreed to hire an outside firm to manage the ships that Samareo would operate. Mobil suggested Atlas Maritime Company, a new venture being formed by Greek shipping executive George Comnas. Comnas had previously worked for the Greek maritime firm of C.M. Lemos & Co., where Peter Tavoulareas, a 24-year-old graduate of the Master of Business Administration program at Columbia University, was also employed as a junior executive. Peter had first approached Lemos in August, 1972 with respect to employment and by January 1, 1973 had started working in the shipping business with C.M. Lemos & Co. (Trial Transcript (Tr.) 2385, 2387, 2451). Mobil owned a fleet of oil tankers. It wanted some of those ships to be operated under Samarco’s name in order to get the expected shipping preference. Mobil agreed to “bareboat charter” its ships {i.e., charter the ships empty and unmanned) to Samareo, which would in turn “time charter” them {i.e., charter the vessels complete with crews and provisions) back to Mobil. Mobil thus in a sense was “chartering” its own ships to itself and essentially obtaining its crews and provisions through Samarco. In admiralty law “for many, if not most, purposes the bareboat charterer is to be treated as the owner” of the chartered vessel. Reed v. The Yaka, 373 U.S. 410, 412, 83 S.Ct. 1349, 1352, 10 L.Ed.2d 448 (1963). Thus, the contractual arrangement placed considerable responsibility upon Samarco for Mobil’s ships. Samarco, for its part, obtained crews and provisions through its management firm, Atlas. Comnas, in setting up Atlas, asked two young coworkers at Lemos, Peter Tavoulareas and Ares Emmanuel, to join him as partners. When Peter joined Atlas in August 1974, his father William notified Mobil’s Conflict of Interest Committee and formally removed himself from taking part in decisions regarding Atlas. The Mobil Committee found no impropriety in the arrangement, which was fully disclosed to the Mobil Board of Directors. Comnas was forced out of Atlas a year later, in 1975, ostensibly because of self-dealing, fraudulent practices, and dissatisfaction with his performance on the part of Mobil and the Saudis. He was placed on Mobil’s payroll as a consultant, and Atlas reimbursed Mobil for his salary. Mobil sent Harmon Hoffmann, a Mobil shipping executive, to act as interim head of Atlas. Hoffmann was offered Comnas’ share of Atlas stock (Tr. 4151), but turned down the offer to become an equity partner, and after six months returned to Mobil. Peter Tavoulareas and Emmanuel then assumed the management duties of Atlas. Atlas prospered during the succeeding years; by 1979 it apparently was operating 17 ships, including 7 for Samarco, only 4 of which were Mobil’s. The Mobil-Samarco-Atlas relationship— and specifically the Tavoulareas-Atlas connection — was explained to Mobil shareholders in a 1976 letter. At about the same time, a party or parties sent anonymous letters to a number of newspapers, alleging that the arrangement violated the federal securities laws. One of those letters went to reporter Robert Woodward of the Post. Woodward, along with reporters for several other newspapers, investigated the anonymous charges; but none of them apparently found any wrongdoing, and no stories were published. The Securities and Exchange Commission, which obtained a copy of one of the letters, launched an informal investigation and called upon William Tavoulareas to testify in non-public proceedings. The SEC found no impropriety. Evidence presented by the plaintiffs showed that Mobil did not exert control over Samarco, in which it was a minority shareholder, and that all of the decisions relating to Atlas were approved by the controlling Saudi partners. The arrangement was fully disclosed to Mobil’s Conflicts of Interest Committee, Mobil’s Board of Directors, and to the company’s shareholders. Evidence indicated that Mobil and its shareholders had derived economic benefits from the Samarco-Atlas arrangement because Atlas operated the Mobil ships more cheaply than could Mobil. The article at issue in this case had its genesis five years after the creation of Atlas and three years after disclosure to the shareholders. In October, 1979, Samuel “Sandy” Golden, a reporter for The Montgomery [Maryland] Journal, met Philip Piro, a physician who was married to William Tavoulareas’ daughter. Piro was in the midst of a less than amicable divorce proceeding. Piro told Golden about the Atlas transactions. Golden testified that Piro told him that Mobil Oil president William Tavoulareas had set up his son Peter in the business, and that it had made Peter an “overnight millionaire” (Tr. 166). Piro mentioned that Woodward had previously investigated the story. Golden, who had been angling for a job at the Post for some time, sensed a story. He unsuccessfully called Woodward several times and finally left a message that he “had a story about the president of Mobil setting up his son to become an overnight millionaire” (Joint Appendix (J.A.) 786). Woodward, by then an editor at the Post, assigned reporter Patrick Tyler to return Golden’s call. Tyler called Golden. Golden told Tyler that he had a hot source for the Tavoulareas story. - Tyler agreed that if Golden’s source led to a big story they would share a Post byline. The two reporters met at The Owl restaurant in Baltimore, where they jointly interviewed Piro. Tyler apparently was pleased with Piro’s evidence; on leaving the restaurant he remarked that it’s “not every day you knock off one of the seven sisters.” At this point, the two reporters began following separate trails. Golden concentrated on getting more information from Piro and conducted several telephone interviews with him. Although Golden repeatedly assured Piro that he was not recording them, he taped several of the conversations. Golden also began seeking “possible enemies” of Tavoulareas, i.e., those with a “motive for revenge” and “with a reason to say bad things about Mr. Tavoulareas” (J.A. 815). In his search, he met Peter Stockton, an investigator for Rep. John Dingell, Chairman of the House of Representatives Subcommittee on Energy and Power. Stockton moonlighted as an employee of CBS’ “60 Minutes” television show. Golden believed that Stockton, who had something of a reputation as a crusader, “hate[d] people with money.” Golden played the Piro tapes for Stockton, who at some point relayed Golden’s information to Congressman Dingell. Stockton later contacted Tyler and passed other information to him. Dingell wrote to the SEC seeking an investigation. Tyler, meanwhile, tracked down Comnas and met with him in New York. Comnas, who initially requested anonymity, allegedly told Tyler about the Atlas arrangements. The Post contends that Comnas was the major source for its stories. (Since Comnas did not testify at trial and Tyler is a reporter who does not use tape recorders, only Tyler’s own account of the conversations is available.) Tyler also talked to John Kousi, an executive of Fairfield Maxwell, Ltd., one of the Samarco partners. Parts of Kousi’s deposition were placed into evidence. Tyler also tried to interview William Tavoulareas and other Mobil executives, but generally was rebuffed. Witnesses for the plaintiffs stated that they were wary of talking with Tyler because in an earlier story on the oil shortage he had misrepresented statements made by Mobil executives. Tavoulareas, however, agreed to give written answers to questions that Tyler would submit in writing, which was done. Tyler wrote the story, and the Post submitted it to its editing process. During editing, several pieces of material were deleted that were favorable to Mobil and Tavoulareas, chiefly opinions of various persons in and out of Mobil (1) that the Atlas arrangement had benefitted the company, and (2) that William Tavoulareas had not been directly involved in Atlas. The Post copy editor assigned to the story, Christine Peterson, read it, considered its implications, and wrote a memorandum to other editors that she found it “impossible to believe” that William Tavoulareas had arranged the Atlas affair in order to benefit his son. Tyler responded with a memo in which he acknowledged that “a good editor might say that part of our case against Tavoulareas seems tenuous,” and noted that a couple of “key points” were based on a single source — presumably Comnas. The Post editors did not alter the story in response to Peterson’s memo, and the story, after clearance from the lawyers, was published. When the article appeared, Tyler was given sole byline credit for it, but Golden received designation as a “Special Correspondent” who had “contributed” to the story. The first article — which is reprinted as an Appendix to this opinion — appeared on November 30, 1979. It bore the headline “Mobil Chief Sets Up Son in Venture,” and its essential theme is made clear from its beginning: Mobil Oil Corp. president William P. Tavoulareas set up his son five years ago as a partner in a London-based shipping management firm that has since done millions of dollars in business operating Mobil-owned ships under exclusive, no-bid contracts. The article detailed the transactions that led to the creation of Samarco and Atlas. Although the article contained a great deal of innuendo, the defamatory picture it conveyed is based upon certain factual allegations contained therein. Specifically, it alleged: 1. William Tavoulareas “set up his son ... as a partner” in Atlas, a company to which William and Mobil would divert a great deal of business. 2. “As the formation of Atlas was being planned in April 1974, [William] Tavoulareas personally urged [Comnas] that his son be included as an equity partner in Atlas.” 3. Although he had told the Mobile board that he recognized the potential conflict of interest and would not personally be involved in dealings with Atlas, it was nevertheless “the elder Tavoulareas [who] dispatched one of his senior shipping executives, Herman F. Hoffmann, [sic] to London to help run Atlas” after Comnas left the firm. The core allegations — baldly stated as facts in the article — were that William Tavoulareas set up Peter in Atlas and then dispatched a Mobil employee and used Mobil resources to bail out Atlas when Comnas was forced out. Without these “facts,” the story would be nothing more than a report that Mobil had business dealings with a small company in which its president’s son had a part interest — hardly page 1 material. These core “facts” formed the foundation on which the implications of the article rested. In a memo to his superiors at the Post (J.A. 2488-91), Tyler characterized the story as dealing with the “incredibly fancy corporate footwork” undertaken by Tavoulareas and Mobil to (1) “ingratiate itself with the Saudis,” and (2) “set up the son of Mobil’s president in a shipping business when business was bad and the business, therefore, stood little chance of prospering without Mobil’s help.” In the memo, he called it a “stor[y] of nepotism, reward without merit and favoritism by those who wield vast amounts of power.” He continued: ... Mobil — which originally turned down a chance to join Samarco — changed its mind after it dawned on Mobil’s president that such a partnership would justify creation of a small management firm at a time when Tavoulareas’ son was aspiring to such a career and was already at work with one of the Greek shippers. The story showed, wrote Tyler, “that Mobil’s decisions in this case were not made for the traditional business reasons, or for the reasons stated by Mobil.” In specific response to Peterson’s assertion that it was “impossible to believe that Tavoulareas alone pould put together [the Samareo-Atlas] scheme for the sake of his son’s business career, or that he would want to,” he wrote: Samarco can be seen as nothing more complicated than a slight diversion from Mobil in the way it moved its crude oil from point A to point B. The question of “why” Mobil’s president would want to orchestrate such a diversion when it would benefit his son’s business career is begging things a little. Tyler’s interpretation of the article’s implications was echoed by Peterson, who apparently read the story to suggest that “Tavoulareas alone ... put together [the Atlas-Samarco] scheme for the sake of his son’s business career.” As a whole, she wrote, the story was about the use of $680,000-a-year of Mobil’s assets as a “plaything for an indulged son.” The interpretations of Tyler and Peterson reflect fairly accurately the defamatory implications that an average reader would draw from the story: (1) that William Tavoulareas, using his position as a principal officer of Mobil, “set up” his son in Atlas; (2) that a primary motive for Mobil’s participation in the Atlas deal was to benefit Peter Tavoulareas; (3) that William Tavoulareas, although claiming to have isolated himself from business dealings with Atlas, nevertheless directly participated in important matters and personally channelled a Mobil employee and resources to help Atlas; and (4) that Tavoulareas and Mobil may have violated the securities laws by failing to disclose Peter’s involvement in Atlas. Ultimately, the basic implication was that Tavoulareas improperly used his position and the resources of Mobil to benefit his undistinguished 24-year-old son. All of these defamatory implications are based on the statements of “fact” identified above. After the first publication, William Tavoulareas denied the story and demanded a retraction. The Post has steadfastly refused to retract it and presently asserts that the “story was in fact true.” Brief of Post at 18. Other publications picked up the story, which enjoyed a brief flurry in the press. It does not appear, however, that either the Mobil shareholders or the Securities and Exchange Commission ever found any impropriety in the transactions at issue. William and Peter Tavoulareas brought two actions: one for libel against The Washington Post Company and several individuals (collectively, the “Post defendants”), including Tyler, Golden, and Woodward (No. 83-1605), and one for slander and its foreseeable republication against Piro (No. 83-1604). The two actions were consolidated. After trial, the jury returned verdicts in which they awarded damages that indicated they necessarily determined that: 1. In the November 30 article the Post defendants defamed William Tavoulareas. The premise of the article was false and was published with knowledge or reckless disregard of its falsity. The jury awarded Tavoulareas $250,000 in compensatory damages and $1,800,000 in punitive damages against the Post defendants. 2. The Post defendants did not defame Peter Tavoulareas. 3. Piro defamed both William and Peter Tavoulareas. The jury awarded compensatory damages of $5,000 to William and $1,000 to Peter. After receipt of the verdicts, the district court granted judgments n.o.v. for Piro and for the Post defendants against William Tavoulareas. The trial court found that Tavoulareas was a public figure and that there was no proof that the defendants had acted with actual malice: The article in question falls far short of being a model of fair, unbiased, investigative journalism. There is no evidence in the record, however, to show that it contained knowing lies or statements made in reckless disregard of the truth. 567 F.Supp. at 654. The district court affirmed the jury verdict in favor of Peter Tavoulareas against Piro. On appeal, the parties have argued two issues. First, whether William Tavoulareas is a public figure, requiring application of the actual malice standard of New York Times v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964)? And second, whether there is sufficient evidence in the record to support a finding of actual malice? On review, we consider the second of these issues and, in addition, inquire whether the charge that William Tavoulareas “set up” his son in the manner alleged, and certain implications reasonably flowing from the specifically pleaded defamatory allegations, were defamatory and false. Because we find, after review of the whole record, that the evidence was sufficient to demonstrate “clearly and convincingly” that the article was defamatory and false and that the defendants acted with actual malice — and hence to meet the New York Times standard — we do not reach the issue whether Tavoulareas was a “public figure.” II. The Legal Standard The constitutional standard of liability for public figures, as established by the Supreme Court in New York Times and restated in Gertz v. Robert Welch, Inc., 418 U.S. 323, 94 S.Ct. 2997, 41 L.Ed.2d 789 (1974), is well settled. A public figure cannot recover for defamatory falsehoods, no matter how damaging and how utterly baseless, unless he can show “actual malice” — i.e., establish by “clear and convincing proof that the defamatory falsehood was made with knowledge of its falsity or with reckless disregard for the truth.” Gertz, 418 U.S. at 342, 94 S.Ct. at 3008 (emphasis added). The Court consistently has refused to apply objective standards to determining actual malice; no matter how unreasonable the defamatory statement, the defendant cannot be held accountable unless he “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). At issue, then, is the state of mind of the defendants. Proof of mental state very often “depends ... on the credibility of the witnesses, which can best be determined by the trier of facts after observation of the demeanor of the witnesses during direct and cross-examination.” Morrison v. Nissan Motor Co., 601 F.2d 139, 141 (4th Cir.1979) (cases cited); accord, Mutual Fund Investors, Inc. v. Putnam Management Co., 553 F.2d 620, 624 (9th Cir.1977). Thus, in the ordinary case, federal courts are reluctant to decide state of mind issues in summary fashion. See, e.g., Morrison, 601 F.2d at 141. It is not that summary disposition is disfavored when the issue is the mental state of the defendant; rather, the very nature of the inquiry is frequently so bound up with the weighing of credibility that courts, as a practical matter, will only infrequently find the fact's regarding mental state so crystal-clear that a reasonable jury could reach only one conclusion. A few courts have argued that the public interest in fostering free debate requires that summary disposition be used especially freely in defamation cases — that it should, in fact, be the “rule,” rather than the “exception.” See, e.g., Oliver v. Village Voice, Inc., 417 F.Supp. 235, 237 (S.D. N.Y.1976); Guitar v. Westinghouse Electric Corp., 396 F.Supp. 1042, 1053 (S.D.N. Y.), aff'd, 538 F.2d 309 (2d Cir.1976). Summary judgment, in particular, is said to be necessary because the suit itself, not merely the unfavorable verdict, could “chill” protected speech: Due to the possibly overwhelming expense of even a successfully defended defamation action, the threat of litigation itself may cause potential defendants to steer far wider of the unlawful zone and act as self-censors. See, e.g., New York Times Co. v. Sullivan, 376 U.S. 254, 279, 84 S.Ct. 710, 725, 11 L.Ed.2d 686 (1964) (quoting Speiser v. Randall, 357 U.S. 513, 526, 78 S.Ct. 1332,1342, 2 L.Ed.2d 1460 (1958)) (“[Critics] tend to make only statements which ‘steer far wider of the unlawful zone.’ ”). The theory is that frivolous suits should be ended quickly by summary proceedings. By analogy, it could be argued that “[t]he policies governing the propriety of summary judgment presumably apply to motions for directed verdict and for judgment n.o.v.” R. Sack, Libel, Slander, and Related Problems 557 (1980). But that is not the case. Whatever value summary judgment has in sparing defamation defendants the expense of actually going to trial, there is no such value in judgments n.o.v., for they come into being after the trial is over and virtually all of the expense has been incurred. Strictly in terms of court costs and legal fees, the “chill” is the same whether judgment after trial is based on a jury verdict or a judgment n.o.v. In any event, the general proposition that summary proceedings are favored in defamation law has been seriously undercut by recent Supreme Court pronouncements. The Court, recognizing the critical role credibility frequently plays in assessing mental state, has acknowledged that “proof of ‘actual malice’ ... does not readily lend itself to summary disposition,” and expressed “some doubt” that summary judgment should be considered the “rule” rather than the “exception” in defamation cases. Hutchinson v. Proxmire, 443 U.S. 111, & 120 n. 9, 99 S.Ct. 2675, 2680 n. 9, 61 L.Ed.2d 411 (1979); see also Liberty Lobby, Inc. v. Anderson, 746 F.2d 1563,1570 (D.C. Cir.1984). In the recent case of Calder v. Jones, — U.S.-,-, 104 S.Ct. 1482, 1488, 79 L.Ed.2d 804 (1984), the unanimous Court remarked: [T]he potential chill on protected First Amendment activity stemming from libel and defamation actions is already taken into account in the constitutional limitations on the substantive law governing such suits____ We have already declined in other contexts to grant special procedural protections to defendants in libel and defamation actions in addition to the constitutional protections embodied in the substantive laws. The Calder Court cited Hutchinson for the proposition that “no special rules apply for summary judgment” in defamation cases. Id. Under Fed.R.Civ.P. 50(b), a motion for judgment n.o.v. is essentially a motion for directed verdict made after the jury has returned its verdict. Consequently, “the standard for awarding a judgment n.o.v. is the same as that applied when ruling on a motion for a directed verdict.” Vander Zee v. Karabatsos, 589 F.2d 723, 726 (D.C.Cir.1978), cert. denied, 441 U.S. 962, 99 S.Ct. 2407, 60 L.Ed.2d 1066 (1979); see Lester v. Dunn, 475 F.2d 983, 985 (D.C.Cir.1973); 5A Moore’s Federal Practice II 50.07[2] (2d ed. 1982). The standard usually applied was elaborated by this court in Alden v. Providence Hospital, 382 F.2d 163, 165 (D.C.Cir.1967): Unless the evidence, along with all inferences reasonably to be drawn therefrom, when viewed in the light most favorable to the plaintiff is such that reasonable jurors in fair and impartial exercise of their judgment could not reasonably disagree in finding for the defendant, the motion must be denied. Accord, Vander Zee, supra, 589 F.2d at 726 (emphasis added) (footnote omitted). The Ninth Circuit aptly summed up the application of these rules in a defamation action: [I]n a libel case, as in other cases, the party against whom ... a motion for a judgment notwithstanding the verdict is made is entitled to have the evidence viewed in the light most favorable to him and to all inferences that can properly be drawn in his favor by the trier of fact. We think, too, that in such cases it is not only not the duty of the judge, or this court of appeal, to weigh the credibility of the evidence, or to draw inferences in favor of the moving party (except, of course, when no contrary inference can legitimately be drawn) but that neither the judge nor this court on appeal has the authority to weigh credibility or to choose among legitimate inferences in such cases [emphasis added].!! The standard against which the evidence must be examined is that of New York Times and its progeny. But the manner in which the evidence is to be examined in the light of that standard is the same as in all other cases in which it is claimed that a case should not go to the jury. If the evidence, so considered, measures up to the New York Times standard, the case is one for the jury, and it is error to grant a directed verdict____ Guam Federation of Teachers, Local 1581 v. Ysrael, 492 F.2d 438, 441 (9th Cir.), cert. denied, 419 U.S. 872, 95 S.Ct. 132, 42 L.Ed.2d 111 (1974). As the Ninth Circuit held in Alioto v. Cowles Communications, Inc., 519 F.2d 777, 780 (9th Cir.), cert. denied, 423 U.S. 930, 96 S.Ct. 280, 46 L.Ed.2d 259 (1975): A district judge on motion for judgment n.o.v., or an appellate judge on review, must examine the evidence to see whether, if all permissible inferences were drawn in the plaintiff’s favor and all questions of credibility were resolved in his behalf, the evidence would then demonstrate by clear and convincing proof that the libelous material was published with actual malice. See also Yiamouyiannis v. Consumers Union of the United States, Inc., 619 F.2d 932, 940 (2d Cir.) (“[djefamation actions are, for procedural purposes, such as ... summary judgment, to be treated no differently from other actions”), cert. denied, 449 U.S 839, 101 S.Ct. 117, 66 L.Ed.2d 46 (1980); Maheu v. Hughes Tool Co., 569 F.2d 459, 464 (9th Cir.1977) (“traditional standard for appellate review of motions for a directed verdict and judgment n.o.v. should be applied in [a defamation] case”); Time, Inc. v. Ragano, 427 F.2d 219, 221 (5th Cir.1970) (“all inferences to be drawn from the underlying facts must be viewed in the light most favorable to the party opposing the motion for summary judgment”); Vandenburg v. Newsweek, Inc., 441 F.2d 378, 379 (5th Cir.) (same), cert. denied, 404 U.S. 864, 92 S.Ct. 49, 30 L.Ed.2d 108 (1971). The district court correctly stated this legal standard, 567 F.Supp. at 652-53, and even the defendants concede that “the evidence must be construed most favorably to William Tavoulareas.” Brief of Post at 29-30. The question thus is whether the evidence and the reasonable inferences derived from it, when examined in the light most favorable to the plaintiffs, establishes by clear and convincing evidence that the defendants acted with knowledge that the charge was false or with reckless disregard of its truth or falsity- But a court evaluating a defamation verdict which significantly involves First Amendment considerations must be far less deferential to the verdict of the jury. The Supreme Court held in New York Times v. Sullivan that in public figure defamation cases courts “must ‘make an independent examination of the whole record,’ ... so as to assure [themselves] that the judgment does not constitute a forbidden intrusion on the field of free expression.” 376 U.S. at 285, 84 S.Ct. at 729 (quoting Edwards v. South Carolina, 372 U.S. 229, 235, 83 S.Ct. 680, 683, 9 L.Ed.2d 697 (1963) (emphasis added). This principle has been consistently followed by the Court in subsequent cases. See, e.g., St. Amant v. Thompson, 390 U.S. 727, 732-33, 88 S.Ct. 1323,1326, 20 L.Ed.2d 262 (1968); Greenbelt Cooperative Publishing Ass’n v. Bresler, 398 U.S. 6, 11, 90 S.Ct. 1537, 1540, 26 L.Ed.2d 6 (1970). The rule has been recognized and followed by the courts of appeals. See Guam Federation, supra, 492 F.2d at 442 (recognizing duty to “make an independent evaluation of the whole record”) (quoting Edwards v. South Carolina, 372 U.S. 229, 235, 83 S.Ct. 680, 683, 9 L.Ed.2d 697 (1963)); Alioto, 519 F.2d at 780 (court must examine record and evaluate evidence itself on issue of actual malice). The Supreme Court recently reaffirmed that rule in Bose Corp. v. Consumers Union of United States, — U.S. -, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984). The court held: The question whether the evidence in the record in a defamation case is of the convincing clarity required to strip the utterance of First Amendment protection is not merely a question for the trier of fact. Judges, as expositors of the Constitution, must independently decide whether the evidence in the record is sufficient to cross the constitutional threshold that bars the entry of any judgment that is not supported by clear and convincing proof of “actual malice.” Id. 104 S.Ct. at 1965 (emphasis added). This is the standard that controls our review of the record in this case. The issue Bose presents in the present context is whether we are to apply our independent judgment to each separate fact determination that forms the basis for the ultimate conclusion of “actual malice,” or rather only to the ultimate conclusion of clear and convincing proof of “actual malice.” For a number of reasons, we think the latter is the case. First, the Court’s expression of its holding in Bose is phrased only in terms of the ultimate issue; and certiorari had been granted only “to consider whether the Court of Appeals erred when it refused to apply the clearly erroneous standard of Rule 52(a) to the District Court’s ‘finding’ of actual malice.” Id. at 1955. Admittedly, independent review of the ultimate issue, of course, could be thought to subsume independent review of the preliminary factual issues on which it is based. But as detailed below, the Supreme Court discussed extensively the distinction between purely factual findings and so-called ultimate facts to repel such a conclusion. Second, the court of appeals decision affirmed in Bose had not questioned the fact-finder’s preliminary factual determinations. The issue under review was whether it could properly be found defamatory, under an “actual malice” standard, for the defendant to print that the loudspeakers manufactured by the plaintiff caused musical instruments to sound as though they were wandering “about the room.” As the Supreme Court noted, the court of appeals had accepted all the district court’s factual findings, observing “that it ‘was in no position to consider the credibility of witnesses and must leave such questions of demeanor to the trier of fact.’ ” Id. at 1955 (quoting 692 F.2d at 195). The Supreme Court noted with approval this refusal of the court of appeals to second-guess credibility findings of the factfinder. Id. at 1958. The Court likewise did not question them: We may accept all of the purely factual findings of the District Court and nevertheless hold as a matter of law that the record does not contain clear and convincing evidence that Seligson or his employer prepared the loudspeaker article with knowledge that it contained a false statement, or with reckless disregard of the truth. Id. at 1967. What both the court of appeals and the Supreme Court held was that in the reviewing court’s own judgment, these accepted preliminary facts did not support the ultimate conclusion of clear and convincing proof of actual malice. It is entirely clear, then, that Bose does not hold that any factual determination other than the ultimate conclusion may be independently reviewed. The only question is whether it suggests so. One indication that it does not is the following statement in the opinion: Rule 52(a) commands that “due regard” shall be given to the trial judge’s opportunity to observe the demeanor of the witnesses; the constitutionally-based rule of independent review permits this opportunity to be given its due. Id. at 1959. Unless one attributes to this statement the unthinkable obfuscation that “permitting the opportunity to be given its due” includes ignoring the factfinder’s credibility determination and making an independent judgment, this must mean that de novo review does not apply to those preliminary factual determinations based on credibility. Since this typically includes most preliminary factual determinations, there would be little sense in preserving the “independent judgment” rule for the remainder. That Bose addresses only the ultimate question of actual malice is shown, thirdly, by its reliance upon the distinction between questions of fact (which are governed by Rule 52(a)) and questions of law (which are for the court). “At some point,” the Court says, “the reasoning by which a fact is ‘found’ crosses the line between application of those ordinary principles of logic and common experience which are ordinarily entrusted to the finder of fact into the realm of a legal rule upon which the reviewing court must exercise its own independent judgment.” Id. at 1960 n. 17. And again: “When the standard governing the decision of a particular case is provided by the Constitution, this Court’s role in marking out the limits of the standard through the process of case-by-ease adjudication is of special importance.” Id. at 1961 (emphasis added). Such descriptions, and such reliance upon the law-fact distinction, makes sense as applied to the ultimate issue of whether, given certain facts, the constitutional standard of “actual malice” has been met. The Court’s statements in Bose would make little sense if applied to such preliminary issues as whether the reporter was told thus-and-so by a particular source. In sum, if the Court’s statement in Bose that “First Amendment questions of ‘constitutional fact’ compel this Court’s de novo review,” id. at 1964 n. 27, is not limited to the ultimate constitutional fact (e.g., whether the writing was “obscene,” whether there was a “clear and present danger,” or, in the context of the present case, whether there was “actual malice”), then we see no rational stopping point short of holding that all factual issues in a First Amendment case are for the court. To embrace that kind of de novo review, as advocated by the dissent, would eviscerate the role of the jury and, in effect, vest appellate courts with original jurisdiction in libel actions. There is no indication that Bose was meant to set forth such a sweeping proposition. Indeed, its frequent references to the continuing applicability of Rule 52(a) would be incomprehensible on that basis. This searching review of the evidence is fully harmonious with the established rule that in evaluating such evidence, all reasonable inferences must be granted to the plaintiff. The Guam Federation-Alioto line of cases correctly applied the New York Times requirement of review of the entire record, and nothing in Bose indicates that they are no longer good law. Moreover, neither New York Times nor Bose suggests that a district court or a court of appeals is to substitute its assessment of credibility or its general appraisal of the evidence for that of the jury. Courts reviewing the determinations of a factfinder — be it judge or jury — are not to examine the evidence of the plaintiffs and that of the defendants and decide which to believe. That is the function of the fact-finder. Rather, they must carefully examine the evidence, draw all reasonable inferences in favor of the verdict of the jury, and then determine whether that evidence demonstrates clearly and convincingly that the defendant acted with actual malice. With the Supreme Court’s admonitions in mind, and cognizant of our duty to “exercise [our] independent judgment and determine whether the record establishes actual malice with convincing clarity,” Bose, at 1967, we turn to the required extended examination of the evidence as it relates to that issue. Such independent inquiry convinces us that the evidence is sufficient to clearly and convincingly establish actual malice, which the jury found existed on the part of the defendants. Before we begin our actual malice discussion, though, we first consider two issues necessary to reinstating the jury verdict, namely the defamatory character of the article and proof of falsity. III. Defamation and Falsity 1. Defamatory Character of the November 30 Article Statements do not constitute actionable defamation merely because they are false. They must be demonstrated to be defamatory to the plaintiff. As the district court instructed the jury, consistent with the standard in the District of Columbia, a publication is defamatory if it “tends to injure plaintiff in his trade, profession, or community standing, or lower him in the estimation of the community or subject him to scorn, ridicule, shame, contempt or embarrassment” (Tr. 4548). See Afro-American Publishing Co. v. Jaffee, 366 F.2d 649, 654 (D.C.Cir.1966). The Supreme Court has instructed us that “[a] publication claimed to be defamatory must be read and construed in the sense in which the readers to whom it is addressed would ordinarily understand it.” Washington Post Co. v. Chaloner, 250 U.S. 290, 293, 39 S.Ct. 448, 448, 63 L.Ed. 987 (1919). It is clear that while we must consider each portion of the article specified in the complaint, the words should not be considered in isolation; they must be ascribed their plain and natural meaning in the context of the article as a whole. In addition to the plain text, we consider what the article conveys through its structure, implications, and connotations. The theme of the Post article here is expressed in its headline — “Mobil Chief Sets Up Son In Venture” and in the lead paragraph — “Mobil Oil Corp. president William P. Tavoulareas set up his son five years ago as a partner in a London-based shipping management firm that has since done millions of dollars in business operating Mobil-owned ships under exclusive, no-bid contracts.” See Appendix. William Tavoulareas maintains that the article reasonably implies that he breached his fiduciary duties to Mobil, that he wasted and misused Mobil assets, and that he wrongfully directed those assets to his son Peter for his benefit (Tr. 4550). The Post attempts to trivialize the import of the phrase “set up” and argues that it was “just a simple way of describing how something came to be.” Brief of Post at 39. The district court, in ruling on the motion for judgment n.o.v., concluded that the term “set up” “may not be the most felicitous choice of vocabulary that could have been used to describe the situation but ... it cannot be said that it was a lie or a reckless untruth.” 567 F.Supp. at 659. The court similarly rejected, as unreasonable as a matter of law, the conclusion that the article implied that William Tavoulareas put together the whole Mobil-Atlas-Samarco arrangement solely for the sake of his son. Id. at 660. The November 30 article, the court said, also does not impliedly accuse Tavoulareas of misusing Mobil assets. Id. We conclude that the district court was in error and hold that both the phrase “set up,” in the context used, and the article in general, defame William Tavoulareas. Notwithstanding the fact that some statements in the article may imply that Mobil’s involvement with Atlas was purportedly justified on legitimate business grounds, the article could reasonably be read to suggest that William Tavoulareas became involved with Atlas-Samarco primarily to benefit his son. Certainly, this was the interpretation that Post editors Peterson and Tyler gave the article. Such interpretation by the jury would be wholly justified since in this case the article would have little or no news value if it merely intended to report a legitimate business act. At a minimum, the article implies that, irrespective of the motivation for creating the Mobil-Atlas-Samarco relationship, William misused Mobil assets and his position as president to advance his allegedly undeserving son — an act of corporate nepotism. Indeed, this interpretation of alleged nepotism was recognized by the district court as the basic theme of the article. 567 F.Supp. at 660. The caption and lead paragraph clearly connote this. This lead material sets the tone of the article which is not diminished by subsequent statements. On the contrary, the body of the story lends further support to such conclusion. While portions of the article are true, “the detain-er may be the more successful when he baits the hook with truth.” Afro-American, 366 F.2d at 655. The article’s seemingly detailed support, which gives the impression that it states facts, lends apparent truth to its accusations. The other statements specifically challenged as false, see falsity discussion infra, in similar fashion contribute to the defamatory premise of alleged nepotism. We thus conclude that the November 30 article, in its headline and considered as a whole, would reasonably be understood by the average reader in the community as being defamatory in its implication that William Tavoulareas was guilty of corporate nepotism, breached his fiduciary duty to Mobil, and misused Mobil assets. 2. Falsity As the district court properly instructed the jury, liability for defamation requires, as part of the plaintiff’s case and prior to a finding of actual malice, proof of falsity. Garrison v. Louisiana, 379 U.S. 64, 74, 85 S.Ct. 209, 215, 13 L.Ed.2d 125 (1964); New York Times v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964). Because the jury returned verdicts against the Post, Golden, and Piro in favor of William Tavoulareas, it necessarily found that the challenged statements in the November 30 article were false. In granting judgment n.o.v., the trial court found that there was insufficient proof of actual malice and thus did not consider it necessary to address directly the issue of falsity. 567 F.Supp. at 654 n. 9. Tavoulareas maintains that the jury’s finding of falsity “remains undisturbed” on appeal. Brief of Tavoulareas at 17a-17b. The defendants, while on appeal concentrating principally on the actual malice issue, rest on a general statement “that the story was in fact true.” Brief of Post at 18. In the circumstances of this case, the issue of actual malice, which is the major point on appeal, is intertwined with the issue of falsity of the defamatory allegations, since the evidence establishing falsity, and the degree of that evidence, overlaps with and constitutes much of the proof of actual malice. The issue of falsity was briefed and argued before the trial court, and constitutes part of the record that we have reviewed. The procedural posture of this case, reviewing a grant of judgment n.o.v., makes the question of proof of falsity one of law. We accordingly address the falsity issue, and our subsequent discussion of the issue in connection with malice will implicitly expand upon the same point. For purposes of our review we assume, without deciding, that a showing of falsity requires clear and convincing proof. Conclusions to be drawn from the evidence in this case, particularly with respect to the issues of falsity and actual malice, depend in large part on the determination of the credibility of the principal witnesses. As previously detailed, we must view the evidence, with all inferences reasonably to be drawn therefrom, in the light most favorable to the party against whom the judgment n.o.v. was ordered, i.e., the plaintiff. Guam Federation of Teachers, Local 1581 v. Ysrael, 492 F.2d 438, 439-41 (9th Cir.), cert. denied, 419 U.S. 872, 95 S.Ct. 132, 42 L.Ed.2d 111 (1974); see discussion and cases cited in Section II supra. On the key charge that William Tavoulareas “set up” his son in the shipping business, the evidence of truth is (1) Comnas’ alleged statement — as testified by Tyler — that Comnas had taken Peter into the firm only because William suggested it and (2) Piro’s assertion that William had once said that he had given Peter a “nudge” to get him into the shipping business. As to the first, William Tavoulareas’ testimony flatly contradicted Comnas’ alleged accusation (Tr. 1425), as did William’s prior testimony before the SEC (J.A. 2398-2459). The defendants at trial did not produce Comnas’ deposition testimony, though he was their principal source relied upon for the charges made in the article. Nothing uncovered in the Post’s massive pretrial discovery and none of the Post’s witnesses except defendants Piro and Tyler, provides any basis for the charge that William “set up” Peter. As to Piro, Golden testified that at his first meeting with him, Piro told him that “William Tavoulareas had set up his son in a shipping company that made him an overnight millionaire,” or “words to that effect” (Tr. 164-65). Piro never denied he said this at that meeting, but only denied using the words “set up” at the subsequent meeting at The Owl restaurant (Tr. 2897). As to defendant Tyler he is discussed below regarding the “second allegation.” As to the “nudge” statement, even if it can be considered a reference to Peter’s landing the job at Atlas instead of his first shipping-related job at Lemos, the jury had only Piro’s unsupported recollection of the statement. William flatly denied ever making such a remark to Piro (Tr. 1349). The jury in weighing credibility obviously could accept William’s testimony over Piro’s. The second allegation William Tavoulareas challenges as false is that he “personally urged” that Peter be included as an equity partner in Atlas. The statement is unattributed in the article. William Tavoulareas testified that he never urged that his son be included in Atlas (Tr. 1293-94,1296-97, 1433). Defendants rely on a statement allegedly made in a conversation between Tyler and Comnas for the truth of the statement, but the defendants never introduced Comnas’ deposition at trial The conversation was received through Tyler’s testimony (J.A. 979, 981, 1150, 2054, 2074-75, 2498, 2674, 3141, 3143, 3157), which the jury was entitled to weigh and to credit or discredit as it deemed appropriate. When pressed to relay what Comnas purportedly said to him, Tyler testified that “I don’t recall” whether Comnas said specifically that “William personally urged him [Comnas] to take his son into the [Atlas] operation,” but stated only that he used “words to that effect” (J.A. 981). Piro admitted that he had told Tyler he had “no knowledge” of whether William Tavoulareas had requested or urged that his son be included in Atlas (Tr. 292-93). The district court, in discussing “actual malice,” indicated that Kousi’s “nepotistic act” comment could be interpreted to support the “personally urged-set-up” allegation, 567 F.Supp. at 659. However, as to the issue of falsity, there was no evidence that this comment was based on personal knowledge. See note 45 infra. If the testimony of Tavoulareas is believed and that of the defendants and their witnesses disbelieved, as would have been reasonable for the jury to have found, the allegations are clearly false. The third allegation in the article was that “the elder [William] Tavoulareas [personally] dispatched ... Herman [sic] F. Hoffmann ... to help run Atlas.” The defendants rely on three sources for the truth of this statement: (1) Paul Wolfe’s testimony that William Tavoulareas “participated” in discussions where it was decided that Hoffmann would replace Comnas (Tr. 1193); (2) Mobil’s November 20, 1979 letter to Tyler stating that William Tavoulareas was involved in Comnas’ departure (J.A. 2345 (Plaintiffs’ Ex. 23)) (presumably, as defendants urge, “to open the door” for Hoffmann’s arrival); and (3) Piro’s testimony that he had overheard Tavoulareas state to Mobil Vice-President Checket “that he [Tavoulareas] had sent Harmon Hoffmann over to Atlas after Comnas had left” (Tr. 3720-21). The first two of these do not represent a clear assertion that Tavoulareas had any personal hand in deciding that Hoffmann would be dispatched, and are at best very weak circumstantial evidence to that effect. William Tavoulareas testified that he did not personally send Hoffmann to Atlas (Tr. 1440). Paul Wolfe testified that it was he who recommended to Mobil Chairman Rawleigh Warner (in the presence of William Tavoulareas) that Comnas be replaced (Tr. 1069, 1070, 1098, 1186). Checket denied that the alleged conversation between himself and Tavoulareas ever took place (J.A. 1844-48). Again, if the testimony of Tavoulareas, Wolfe, or Checket is believed, which we conclude reasonable inferences from the record permitted the jury to find, the statement that Tavoulareas personally dispatched Hoffmann is false. In conclusion, our review of the record reveals evidence that, when resolved in the light most favorable to the jury verdict, as it must be, demonstrates convincingly that the alleged defamatory statements in the November 30 article are clearly false. Additionally, our succeeding examination of the record on the “actual malice” question discloses further evidence which, under our prescribed standard of review, demonstrates that the challenged statements are clearly and convincingly false. Rather than to duplicate that discussion here, we refer to our actual malice discussion, infra. IY. Reckless Disregard of Truth or Falsity — Actual Malice We first note that the defendants have uniformly denied that they knew what they wrote was false or that they had serious doubts about its truthfulness. They have testified that they honestly believed the stories were true. The defendants argue that “[t]his unimpeached testimony as to the mental elements of actual malice is ‘dispositive.’ ” Brief of Post at 32. The defendants’ apparent claim is that actual malice can be proved only if the reporter or editor confesses to having had actual doubts regarding the truthfulness of the defamatory articles. But courts do not labor under such unreasonable rules regarding evidence. As the Supreme Court has noted: The defendant in a defamation action ... cannot ... automatically insure a favorable verdict by testifying that he published with a belief that the statements were true. The finder of fact must determine whether the publication was indeed made in good faith. St. Amant, supra, 390 U.S. at 732, 88 S.Ct. at 1326 (emphasis added). Hence, a defamation plaintiff is not required to accept the word of the defendant as to his state of mind any more than the government or a jury is required to take the word of a criminal defendant that he lacked the intent required for a criminal conviction. In any case where the mental state of the actor is at issue, it can be proved by the cumulation of circumstantial evidence: [I]n attempting to meet the test the plaintiff may rely upon “evidence of negligence, of motive and of intent ... for the purpose of establishing, by cumular tion and by appropriate inferences,” the requisite degree of culpability. Mere statements by the defendant of his belief in the truth of a publication will carry the day only if they are not overridden by evidence establishing knowing or “reckless” falsity. R. Sack, supra, at 214 (footnotes omitted) (emphasis added) (quoting Goldwater v. Ginzburg, 414 F.2d 324, 342 (2d Cir.1989), cert. denied, 396 U.S. 1049, 90 S.Ct. 701, 24 L.Ed.2d 695 (1970)). Although recklessness in the defamation arena requires a showing of subjective doubts, see St. Am-ant, supra, 390 U.S. at 731, 88 S.Ct. at 1325, it “is ordinarily inferred from objective facts.” Washington Post Co. v. Keogh, 365 F.2d 965, 968 (D.C.Cir.1966), cert. denied, 385 U.S. 1011, 87 S.Ct. 708, 17 L.Ed.2d 548 (1967). As the First Circuit has noted: The subjective determination of whether [a defendant] in fact entertained serious doubts as to the truth of the statement may be proved by inference, as it would be rare for a defendant to admit such doubts____ A court typically will infer actual malice from objective facts ____ These facts should provide evidence of negligence, motive, and intent such that an accumulation of the evidence and appropriate inferences supports the existence of actual malice. Bose Corp. v. Consumers Union of United States, 692 F.2d 189, 196 (1st Cir.1982), aff'd, — U.S. -, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984) (emphasis added) (citations omitted). The plaintiffs here attempted to demonstrate by a web of circumstantial evidence that the defendants acted with actual malice. Each of the major categories of evidence will be discussed in turn, though their order of appearance is not meant to reflect their relative importance. 1. The Christine Peterson Memorandum After Tyler wrote his first story, it entered into the Post’s editing process. During the course of that process it was given to copy editor Christine Peterson for editing. After reviewing the article, and prior to publication, Peterson sent a memorandum to assignment editor Peter Milius. The memorandum stated, in part: I’ve read the Mobil story several times, and while I’m impressed with the amount of work the reporter obviously did, I’m still left with an overwhelming sense of So What? Is there any way to give this story of high-level nepotism a dollars- and-cents angle? Did Mobil’s shareholders lose anything? Mobil’s customers? Parts of Tyler’s case against Tavoulareas seem tenuous, and the whole — a $680,00-a-year [sic] plaything for an indulged son, at worst — just seems like a withered peanut in an 84" gilded shell. A far more interesting angle, it seems to me, is Mobil’s concern about Saudi preference shipping — a concern so profound that it led to the formation of an entire dummy corporation. It’s impossible to believe that Tavoulareas alone could put together such a scheme for the sake of his son’s business career, or that he would want to. (J.A. 2486) (emphasis added). Peterson, who at the time was still in the employ of the Post, testified at trial that when she used the phrase “impossible to believe” she did not really mean that the story was “impossible to believe” — only that she felt that “the focus of the story should have been the formation of the company Samar-co in anticipation of preference shipping” (J.A. 1837). Peterson, as a Post employee, was an interested witness, and the jury was entitled to evaluate her testimony accordingly. It was not obliged to credit her subsequent protestations that she did not really mean what she wrote. Not surprisingly, the district court found the interpretation offered by the defendants to be “strained.” 567 P.Supp. at 655. A reasonable reading of the memorandum is that (1) Peterson read the article as saying that William Tavoulareas led Mobil into the Samarco-Atlas venture substantially as a means of benefitting his son, and (2) she found the charges in the article, as she wrote, “impossible to believe.” This is the district court’s independent evaluation of the facts as set forth in the article. The district court, relying heavily on what it felt to be Peterson’s relatively minor role in the editing process and the fact that she testified that she had done no independent research on the story, found that she was merely “expressing her uninformed opinion that she found the story hard to believe.” (Her actual appraisal was that the article was “impossible to believe.”) The district court continued: If Ms. Peterson had investigated the article’s substance and sources and thereafter had expressed a similar statement of doubt as to the article’s veracity, plaintiff’s argument might prevail. Plaintiff, however, has offered no evidence to this effect. The actual malice test focuses on the state of mind of the author and publisher of the allegedly libelous statements. Ms. Peterson’s memorandum does not show that those who were responsible for the article’s substance entertained any doubts whatsoever about its accuracy. 567 F.2d at 655 (emphasis added). The district court’s conclusion, however, depends entirely upon the court’s finding that copy editors at the Post have no responsible role in preparing articles for publication. The court’s opinion implicitly credited the defendants’ testimony and its reasonable inferences. Yet, the character of the Peterson memorandum and the responses it evoked within the Post organization indicate that her job involved more than mere proofreading. It is undisputed that copy editors are responsible for so-called “style” editing (which involves control over the exact language that will be used in the article) and for checking at least some of the facts in a story. See note 31 supra. Although at trial she described her position as having a very limited role,