On Appellee's Petition for Rehearing

HARRY T. EDWARDS, Circuit Judge:

I often have been struck by Justice Stewart's concurring statement in Boys Markets, Inc. v. Retail Clerks Union, Local 770, 398 U.S. 235, 90 S.Ct. 1583, 26 L.Ed.2d 199 (1970), a case in which the Court reconsidered and overruled an earlier decision. Justice Stewart remarked that, "[i]n these circumstances the temptation is strong to embark upon a lengthy personal apologia." Id. at 255, 90 S.Ct. at 1595. This remark has special poignancy for me now, because it underscores the distress felt by a judge who, in grappling with a very difficult legal issue, concludes that he has made a mistake of judgment. Once discovered, confessing error is relatively easy. What is difficult is accepting the realization that, despite your best efforts, you may still fall prey to an error of judgment. Like Justice Stewart, I will take refuge in an aphorism of Justice Frankfurter:

Wisdom too often never comes, and so one ought not to reject it merely because it comes late.

Henslee v. Union Planters Nat. Bank Trust Co., 335 U.S. 595, 600, 69 S.Ct. 290, 293, 93 L.Ed. 259 (1949) (Frankfurter, J., dissenting).

In Moldea v. New York Times Co., 15 F.3d 1137 (D.C. Cir. 1994) ("Moldea (I)"), this panel was faced with an appeal brought by author and investigative journalist Dan E. Moldea in connection with his defamation action against the New York Times Company, Inc. ("Times"). Moldea's lawsuit alleged that appellee libeled him in a book review (the "Times review" or "review") published in the New York Times Book Review, a supplement to the Sunday edition of its daily newspaper. The review stated that Moldea's book, Interference: How Organized Crime Influences Professional Football (" Interference"), was marred by "too much sloppy journalism," and offered a number of examples of the work's alleged journalistic shortcomings. The District Court granted summary judgment in favor of the Times, ruling that the review in question was not actionable as a matter of law because it consisted only of unverifiable statements of the reviewer's opinion, or of law because it consisted only of unverifiable statements that no reasonable juror could find to be false. Moldea v. New York Times Co., 793 F.Supp. 335 (D.D.C. 1992). In a 2-1 decision, the panel reversed on the ground that some of the review's characterizations of Moldea's book were potentially actionable because they were verifiable, and could not be held to be true as a matter of law. Moldea (I), 15 F.3d at 1146-48.

After careful consideration of the Times' petition for rehearing and Moldea's response to that petition, we are persuaded to amend our earlier decision. The original majority opinion was generally correct in its statement of the law of defamation. Unfortunately, that opinion failed to take sufficient account of the fact that the statements at issue appeared in the context of a book review, a genre in which readers expect to find spirited critiques of literary works that they understand to be the reviewer's description and assessment of texts that are capable of a number of possible rational interpretations. While there is no per se exemption from defamation for book reviews, our initial resolution of this case applied an inappropriate standard to judge whether the Times review was actionable.

In light of our reconsideration of this case, we hold that the challenged statements in the Times review are supportable interpretations of Interference, and that as a matter of law the review is substantially true. Accordingly, we affirm the District Court's grant of summary judgment in favor of the Times.

I. BACKGROUND

The facts of this case are fully explained in Moldea (I), so we need only briefly sketch them here. The instant case grows out of a highly negative review of Interference written by New York Times sportswriter Gerald Eskenazi, and published in the New York Times Book Review on September 3, 1989. Moldea contends that prior to the publication of this review he was a respected author and journalist, and that both he and his publisher anticipated that Interference, his fourth book, would be a success. Appellant alleges that the review's harsh critique of Interference destroyed the book's prospects for commercial success, and effectively ended his career as a writer as well, because he is now unable to interest other publishers in his work. Appellant also claims that, because of the review, he can no longer obtain bookings for lectures and other public appearances, activities which formerly provided him with significant income.

On August 24, 1990, Moldea filed suit against the Times alleging defamation and false light invasion of privacy. The Times moved for summary judgment before either party had begun discovery, and, on January 31, 1992, the District Court granted the Times' motion based solely on the texts of the review and of Interference itself. The trial court ruled that Moldea's claim was not actionable as a matter of law because the portions of the Times review challenged in his suit either were statements of opinion about a literary work, or were so clearly true that no reasonable juror could find them to be false. Moldea, 793 F.Supp. at 338.

In the District Court and on appeal, Moldea alleged that six specific statements in the Times review had defamed him by accusing him of being an incompetent practitioner of his chosen profession, investigative journalism, and by supporting that accusation with false characterizations of his book. We held in Moldea (I) that one of these passages was a statement of opinion that implied defamatory facts because it accused Moldea of being an incompetent journalist. That statement read:

But there is too much sloppy journalism to trust the bulk of this book's 512 pages — including its whopping 64 pages of footnotes.

See Moldea (I), 15 F.3d at 1145-46. Moldea (I) went on to hold that the remaining statements Moldea challenged were offered by Eskenazi as factual examples of Interference's alleged "sloppiness," and that "[i]n order for the review to be nonactionable as a matter of law, the Times must show that it offered true facts in support of its judgment that served to support its statement of opinion." Id. at 1146.

Our earlier decision in this case held that three of the five remaining statements challenged by Moldea on appeal were not actionable in defamation. Of the three nonactionable passages, two were incontrovertibly true statements based upon facts revealed in the text of Interference, while the third was a supported statement of opinion. See id. at 1146-49. Moldea (I) held, however, that two of the challenged passages in the Times review were verifiable, and that a reasonable juror could conclude that they were false. First, the review stated:

Mr. Moldea tells as well of Mr. Namath's `guaranteeing' a victory in Super Bowl III shortly after a sinister meeting in a bar with a member of the opposition, Lou Michaels, the Baltimore Colts' place-kicker. The truth is that the pair almost came to blows after they both had been drinking; and Mr. Namath's well-publicized `guarantee' came about quite innocently at a Miami Touchdown Club dinner when a fan asked him if he thought the Jets had a chance. `We'll win. I guarantee it,' Mr. Namath replied.

Second, the review opined that:

[Moldea] revives the discredited notion that Carroll Rosenbloom, the ornery owner of the Rams, who had a penchant for gambling, met foul play when he drowned in Florida 10 years ago.

Our initial opinion in this case concluded that a reasonable juror could find that the Times review had mischaracterized Interference's portrayal of each of the foregoing two events. Accordingly, we held that it was error for the trial court to grant summary judgment at so early a stage of this litigation.

II. DISCUSSION

A. The Importance of Context

Moldea (I) noted that, "under the established case law, our analysis of this case is not altered by the fact that the challenged statements appeared in a `book review' rather than in a hard news story." Moldea (I), 15 F.3d at 1145-46. This statement is correct insofar as it suggests that there is no per se exemption from defamation for book reviews. Even the Times concedes this point in its Petition for Rehearing. See Petition for Rehearing at 4 ("No one doubts that a book review can be actionable."). A writer may not commit libel at will merely by labelling his work a "review." Moldea (I) is short-sighted, however, in failing to take account of the fact that the challenged statements were evaluations of a literary work which appeared in a forum in which readers expect to find such evaluations. As the Supreme Court has recognized, writers must be given some leeway to offer "rational interpretation" of ambiguous sources. See Masson v. New Yorker Magazine, Inc., 501 U.S. 496, ___, 111 S.Ct. 2419, 2434, 115 L.Ed.2d 447 (1991). Thus, when a reviewer offers commentary that is tied to the work being reviewed, and that is a supportable interpretation of the author's work, that interpretation does not present a verifiable issue of fact that can be actionable in defamation.

The fundamental framework established in Moldea (I) for defamation actions is sound, and we do not modify it in this decision. As we stated in our initial opinion, the Supreme Court's decision in Milkovich v. Lorain Journal Co., 497 U.S. 1, 110 S.Ct. 2695, 111 L.Ed.2d 1 (1990), and this court's decision in White v. Fraternal Order of Police, 909 F.2d 512 (D.C. Cir. 1990), make clear that there is no wholesale exemption from liability in defamation for statements of "opinion." Instead, statements of opinion can be actionable if they imply a provably false fact, or rely upon stated facts that are provably false. See generally Moldea (I), 15 F.3d at 1143-45.

In Milkovich, the Supreme Court rejected the argument that an accusation of perjury was nonactionable merely because it was offered as the writer's "opinion." In that case, a high school wrestling coach argued that an Ohio newspaper libeled him by printing a column which alleged that he had perjured himself in his testimony to a state court concerning his role in an altercation between his team and an opposing squad at a wrestling match. The column stated that: "Anyone who attended the meet . . . knows in his heart that Milkovich . . . lied at the hearing." Milkovich, 497 U.S. at 5, 110 S.Ct. at 2698. Although the statements at issue in Milkovich appeared in an "opinion column" in a newspaper sports section, the Court found no relevance in this fact in reaching its decision, apparently because an accusation of perjury is not the sort of discourse that even arguably is the usual province of such columns. Sports columnists frequently offer intemperate denunciations of coaches' play-calling or strategy, and readers know this and presumably take such railings with a grain of salt; but an accusation of criminal conduct is a classic libel, and so Milkovich did not even pause to assess the effect that the column's context may have had on those who read it.

In Moldea (I), this court observed that Milkovich made no mention of the fact that the statements at issue in that case appeared in a sports column, and took that fact to mean that context was irrelevant in the instant case. We now recognize, however, as has the First Circuit, that Milkovich did not disavow the importance of context, but simply "discounted it in the circumstances of that case." Phantom Touring, Inc. v. Affiliated Publications, 953 F.2d 724, 729 n. 9 (1st Cir. 1992) (holding newspaper theater column nonactionable in part because "the context of each article rendered the language not reasonably interpreted as stating `actual facts' about appellant's honesty."), cert. denied, ___ U.S. ___, 112 S.Ct. 2942, 119 L.Ed.2d 567 (1992). This conclusion is compelled by the logic of two Supreme Court cases expressly reaffirmed in Milkovich, and by the Court's decision in Masson, rendered the following term.

First, Milkovich reaffirmed the vitality of Greenbelt Cooperative Publishing Association v. Bresler, 398 U.S. 6, 90 S.Ct. 1537, 26 L.Ed.2d 6 (1970), and Letter Carriers v. Austin, 418 U.S. 264, 94 S.Ct. 2770, 41 L.Ed.2d 745 (1974). See Milkovich, 497 U.S. at 16-17, 110 S.Ct. at 2704-05. In Bresler, a real estate developer had engaged in negotiations with a city council for a zoning variance, while simultaneously negotiating with the city over other land that the city wished to purchase from him. A local newspaper account stated that some people had characterized the developer's tactics as "blackmail," and the developer sued for libel. The Court rejected the developer's argument that "blackmail" implied criminal activity, noting that "the word `blackmail' in these circumstances was not slander when spoken. . . ." Bresler, 398 U.S. at 13, 90 S.Ct. at 1541 (emphasis added). In Letter Carriers, the Court held that the use of the word "traitor" to define a "scab" in the context of a labor dispute could not be the basis for a defamation action. 418 U.S. at 271-72, 284-86, 94 S.Ct. at 2775, 2781-82. Both Bresler and Letter Carriers rely in large part on the notion that the speech at issue in each case was intended as hyperbole, see Milkovich, 497 U.S. at 20, 110 S.Ct. at 2706-07; however, this fact reinforces the importance of context, because it is in part the settings of the speech in question that makes their hyperbolic nature apparent, and which helps determine the way in which the intended audience will receive them. Thus, the "lusty and imaginative expression of the contempt felt by union members" for a "scab" may lawfully find hyperbolic expression during a strike, Letter Carriers, 418 U.S. at 286, 94 S.Ct. at 2782, because the context assures that no reader could understand the epithet "traitor" to be a charge that the "scab" has committed the criminal offense of treason. Id. at 285, 94 S.Ct. at 2781-82.

Second, Masson, handed down in the term following Milkovich, is further evidence that the Supreme Court has not abandoned the consideration of context in defamation actions. In Masson, the Court addressed the question whether a writer's alteration of quotations attributed to the subject of an interview could establish the "actual malice" required for a defamation suit by a public figure. Masson observed that whether quotations will be interpreted by readers as the actual statements of a speaker depends on context — for example, whether there is "an acknowledgment that the work is a so-called docudrama or historical fiction, or that it recreates conversations from memory, not from recordings. . . ." Masson, 501 U.S. at ___-___, 111 S.Ct. at 2430-31.

In Ollman v. Evans, 750 F.2d 970, 983 (D.C. Cir. 1984) (en banc), cert. denied, 471 U.S. 1127, 105 S.Ct. 2662, 86 L.Ed.2d 278 (1985), we recognized that courts have long "considered the influence that . . . well-established genres of writing will have on the average reader." Id. at 984 (emphasis in original). Given that Milkovich was decided against the backdrop of this settled principle, and that it expressly reaffirmed two of the Court's key precedents in this area, we are, on reflection, convinced that Moldea (I) erred in assuming that Milkovich abandoned the principle of looking to the context in which speech appears. The Court's decision in Masson appears to confirm this interpretation of Milkovich. While Milkovich could be interpreted as we read it in our initial decision, we are unwilling to assume that the Court meant to sweep away so much settled law without a clearer indication that this was indeed its intent.

B. Relevance of the Book Review Context

In contrast to the situation in Milkovich, the instant case involves a context, a book review, in which the allegedly libelous statements were evaluations quintessentially of a type readers expect to find in that genre. The challenged statements in the Times review consist solely of the reviewer's comments on a literary work, and therefore must be judged with an eye toward readers' expectations and understandings of book reviews. This would not be the case if, for example, the review stated or implied that Interference was a badly written book because its author was a drug dealer. In that situation, this case would parallel Milkovich: the reviewer would simply be employing the medium of a book review as a vehicle for what would be a garden-variety libel, and the review would thus potentially be actionable.

There is a long and rich history in our cultural and legal traditions of affording reviewers latitude to comment on literary and other works. The statements at issue in the instant case are assessments of a book, rather than direct assaults on Moldea's character, reputation, or competence as a journalist. While a bad review necessarily has the effect of injuring an author's reputation to some extent — sometimes to a devastating extent, as Moldea alleges is true here — criticism's long and impressive pedigree persuades us that, while a critic's latitude is not unlimited, he or she must be given the constitutional "breathing space" appropriate to the genre. New York Times Co. v. Sullivan, 376 U.S. 254, 272, 84 S.Ct. 710, 121, 11 L.Ed.2d 686 (1964).

We believe that the Times has suggested the appropriate standard for evaluating critical reviews: "The proper analysis would make commentary actionable only when the interpretations are unsupportable by reference to the written work." Petition for Rehearing at 8 (emphasis added). This "supportable interpretation" standard provides that a critic's interpretation must be rationally supportable by reference to the actual text he or she is evaluating, and thus would not immunize situations analogous to that presented in Milkovich, in which a writer launches a personal attack, rather than interpreting a book. This standard also establishes boundaries even for textual interpretation. A critic's statement must be a rational assessment or account of something the reviewer can point to in the text, or omitted from the text, being critiqued. For instance, if the Times review stated that Interference was a terrible book because it asserted that African-Americans make poor football coaches, that reading would be "unsupportable by reference to the written work," because nothing in Moldea's book even hints at this notion. In such a case, the usual inquiries as to libel would apply: a jury could determine that the review falsely characterized Interference, thereby libeling its author by portraying him as a racist (assuming the other elements of the case could be proved).

Our decision to apply the "supportable interpretation" standard to book reviews finds strong support in analogous decisions of the Supreme Court, all decided or reaffirmed after Milkovich. These cases establish that when a writer is evaluating or giving an account of inherently ambiguous materials or subject matter, the First Amendment requires that the courts allow latitude for interpretation. For example, in Bose Corp. v. Consumers Union of United States, Inc., 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984), a decision the Court discussed and reaffirmed in Masson, a reviewer writing for Consumer Reports magazine described the experience of listening to music through a pair of stereo speakers: "[I]ndividual instruments heard through the Bose system seemed to grow to gigantic proportions and tended to wander about the room." Bose, 466 U.S. at 488, 104 S.Ct. at 1953. Bose Corporation sued for defamation, alleging that the reviewer's unflattering portrayal was factually inaccurate. The Court held that the statements were not actionable, because they were not so obviously false as to sustain a finding of "actual malice." As the Court interpreted Bose in Masson:

[T]he result was not an assessment of events that speak for themselves, but "one of a number of possible rational interpretations of an event that bristled with ambiguities and descriptive challenges for the writer." We refused to permit recovery for choice of language which, though perhaps reflecting a misconception, represented "the sort of inaccuracy that is commonplace in the forum of robust debate to which the New York Times rule applies."

Masson, 501 U.S. at ___, 111 S.Ct. at 2434 (quoting Bose, 466 U.S. at 512, 513, 104 S.Ct. at 1966) (internal citation omitted).

The Court's opinion in Bose relied heavily on its earlier decision in Time, Inc. v. Pape, 401 U.S. 279, 91 S.Ct. 633, 28 L.Ed.2d 45 (1971). Pape reversed a libel judgment against a reporter who had summarized a report by the United States Commission on Civil Rights discussing civil rights abuses by police officers. The article quoted the Commission's summary of the facts of an alleged incident of police brutality, but failed to state that the Commission had qualified its remarks by noting that they were taken from a civil complaint. As in Bose, the Court held that the claim was not actionable because the publication was not sufficiently false to sustain a finding of "actual malice." See Masson, 501 U.S. at ___, 111 S.Ct. at 2434. Masson explained that Pape "distinguished between a `direct account of events that speak for themselves' and an article descriptive of what the Commission had reported. Time, Inc. v. Pape took into account the difficult choices that confront an author who departs from direct quotation and offers his own interpretation of an ambiguous source." Masson, 501 U.S. at ___, 111 S.Ct. at 2434 (quoting Pape, 401 U.S. at 285, 91 S.Ct. at 637) (internal citation omitted).

Finally, Masson itself noted that: "The protection for rational interpretation serves First Amendment principles by allowing an author the interpretive license that is necessary when relying upon ambiguous sources." 501 U.S. at ___, 111 S.Ct. at 2434. Masson concluded that in order to state a claim for defamation based upon the alteration of direct quotations, a plaintiff must show that the alterations resulted in "a material change in the meaning conveyed by the statement." Id. at ___, 111 S.Ct. at 2433. Although Masson, Bose and Pape all concerned the evidence necessary to establish "actual malice," those decisions are rooted in the question of a plaintiff's ability to prove falsity so as to show that a defendant presented information he or she knew to be false. Because of their focus on falsity, the reasoning of these decisions is fully applicable to the instant case. Masson, Bose and Pape recognized that some materials by their very nature require interpretation, and that the First Amendment affords latitude to those engaged in that task. Reasonable minds can and do differ as to how to interpret a literary work. Accordingly, as Masson counsels, we must allow a degree of "interpretive license." 501 U.S. at ___, 111 S.Ct. at 2434.

C. Application of the "Supportable Interpretation" Standard to the Times Review

As we noted in our initial decision, this appeal presents a pure question of law, which we review de novo: whether Moldea can in fact state a claim for defamation. Moldea (I), 15 F.3d at 1142. In this situation, we must determine as a threshold matter whether a challenged statement is capable of a defamatory meaning; and whether it is verifiable — that is, whether a plaintiff can prove that it is false. See generally, id. at 1142-45. The Times review is, as we previously held, capable of a defamatory meaning insofar as it tends to injure Moldea's reputation as a practitioner of his chosen profession, investigative journalism. The key to this case is the question of verifiability.

Although Moldea (I) held that the Times review's statement that Interference contained "too much sloppy journalism" was a verifiable assessment of the book, we now recognize that, in the context of a book review, it is highly debatable whether this statement is sufficiently verifiable to be actionable in defamation. Arguably, our decision in Moldea (I) failed adequately to heed the counsel of both the Supreme Court and our own precedents that "[w]here the question of truth or falsity is a close one, a court should err on the side of nonactionability." Liberty Lobby, Inc. v. Dow Jones Co., 838 F.2d 1287, 1292 (D.C. Cir.) (citing Philadelphia Newspapers, Inc. v. Hepps, 475 U.S. 767, 776, 106 S.Ct. 1558, 1564, 89 L.Ed.2d 783 (1986)), cert. denied, 486 U.S. 825, 109 S.Ct. 75, 102 L.Ed.2d 51 (1988). "The First Amendment requires that we protect some falsehood in order to protect speech that matters." Gertz v. Robert Welch, Inc., 418 U.S. 323, 341, 94 S.Ct. 2997, 3007, 41 L.Ed.2d 789 (1974). The Court has cautioned in several cases that the First Amendment preserves a "breathing space" essential to the exercise of freedom of the press. "To that end [the Supreme] Court has extended a measure of strategic protection to defamatory falsehood." Id. at 342, 94 S.Ct. at 3008.

However, we need not determine whether "too much sloppy journalism" is verifiable, as the statements that the Times review offers in support of this assessment are supportable interpretations of Interference. Thus, even if the review's assertion that the book contains "too much sloppy journalism" is verifiable, that assessment is supported by revealed premises that we cannot hold to be false in the context of a book review. As we stated in Moldea (I): "Because the reader understands that such supported opinions represent the writer's interpretation of the facts presented, and because the reader is free to draw his or her own conclusions based upon those facts, this type of statement is not actionable in defamation." Moldea (I), 15 F.3d at 1144-45.

As we noted above, Moldea (I) held that only two of the five challenged passages in the Times review could be proven to be false; the other three were held either to be true, or to be supported statements of opinion. In addition, Moldea did not challenge the review's assertion that Interference contains several spelling errors which the reviewer concluded "call into question [Moldea's] diligence at simple fact-checking." Thus, only two passages in the Times review are even potentially actionable: the statement that Interference "revives the discredited notion" that Carroll Rosenbloom was murdered, and the claim that Moldea described a meeting between opposing players just before the 1969 Super Bowl as "sinister."

Our initial decision in this case erred by basing its holding on a standard that failed to take into account the fact that the challenged statements appeared in the context of a book review, and were solely evaluations of a literary work. Moldea (I) considered whether a reasonable jury could find that the challenged statements were false because they mischaracterized Interference. See Moldea (I), 15 F.3d at 1146, 1147. Such a standard might be appropriate in the case of an ordinary libel such as that at issue in Milkovich, but it is an inappropriate measure of an interpretation of a book. Applying the "supportable interpretation" standard, the correct measure of the challenged statements' verifiability as a matter of law is whether no reasonable person could find that the review's characterizations were supportable interpretations of Interference. Applying this standard, we hold that the Times review is not actionable in defamation.

First, the Times review stated that:

[Moldea] revives the discredited notion that Carroll Rosenbloom, the ornery owner of the Rams, who had a penchant for gambling, met foul play when he drowned in Florida 10 years ago.

Moldea discusses Rosenbloom's drowning in pages 319 through 326 of his book, closing his account with quoted observations from several of Rosenbloom's friends, who speculate that he was murdered. Interference later reveals, on page 360, that Moldea has located previously unknown photographs, taken at Rosenbloom's autopsy, which make clear that he "died in a tragic accident and was not murdered." As we held in Moldea (I), a reasonable jury could conclude that the Times review's characterization of Interference's portrayal of Rosenbloom's death was false, and that the reviewer's account of the book creates the misleading impression that Moldea inadequately investigated this story. Moldea (I), 15 F.3d at 1147-48. However, given that Interference does not reveal that Rosenbloom's death was accidental until 35 pages after giving undeniably titillating hints of homicide, we cannot hold that a reviewer could not reasonably suggest that Moldea sought to "revive" the notion that Rosenbloom was murdered in order to build suspense before disproving that theory.

The second potentially actionable review passage states that:

Mr. Moldea tells as well of Mr. Namath's `guaranteeing' a victory in Super Bowl III shortly after a sinister meeting in a bar with a member of the opposition, Lou Michaels, the Baltimore Colts' place-kicker. The truth is that the pair almost came to blows after they both had been drinking; and Mr. Namath's well-publicized `guarantee' came about quite innocently at a Miami Touchdown Club dinner when a fan asked him if he thought the Jets had a chance. `We'll win. I guarantee it,' Mr. Namath replied.

Moldea (I) concluded that a reasonable jury could find that Moldea did not describe the meeting as a "sinister" rendezvous, but rather made clear that the meeting was "quite accidental and even confrontational." Interference at 197; see Moldea (I), 15 F.3d at 1146-47. Even applying the "supportable interpretation" standard, this review passage is close to the line. Interference not only states that the Namath-Michaels meeting was "accidental," but on the same page quotes Michaels as saying "What we talked about had no relationship to the game," and quotes another player present at the meeting as confirming that "`nothing technical' about the game was discussed." Interference at 197. The Times' petition for rehearing argues only that the review's characterization is supported by the fact that Interference's description of the Namath-Michaels meeting appears in a chapter of the book largely devoted to probing allegations that there was something "suspicious" about Super Bowl III. See Petition for Rehearing at 12 (citing Interference at 197).

We are troubled by the "sinister meeting" passage, but are constrained to conclude that it does not give rise to an actionable claim. The review offered at least six observations to support the charge of "sloppy journalism": the five challenged passages, plus the unchallenged claim that Moldea made several spelling errors. At least five of these observations could not be proved false at trial, either because they are true, are supported opinion, are reasonable interpretations, or are not challenged in this suit. Moldea is left with only the "sinister meeting" passage as a possible basis for his defamation claim, and this is a very weak basis indeed. For one thing, the "sinister meeting" passage is not defamatory on its face, but rather is simply one of the "interpretations" offered in support of the review's assessment of Moldea's book. Furthermore, even without the support of the "sinister meeting" passage, the review's assertion that Interference is marred by "too much sloppy journalism" is (as a legal matter) "substantially true," and so is not actionable in defamation.

As Moldea (I) noted, "substantial truth" is a defense to defamation. Moldea (I), 15 F.3d at 1150. "Slight inaccuracies of expression are immaterial provided that the defamatory charge is true in substance." Liberty Lobby v. Dow Jones, 838 F.2d at 1296 (citing RESTATEMENT (SECOND) OF TORTS § 581A cmt. f (1977)); accord Foretich v. CBS, Inc., 619 A.2d 48, 60 (D.C. 1993) (citing Liberty Lobby v. Dow Jones). The Supreme Court explained this defense in Masson by noting that: "Minor inaccuracies do not amount to falsity so long as the substance, the gist, the sting, of the libelous charge be justified." 501 U.S. at ___, 111 S.Ct. at 2433 (interpreting California law) (internal quotation omitted); accord Liberty Lobby v. Dow Jones, 838 F.2d at 1296 (a statement is nonactionable if "[t]he sting of the charge . . . is substantially true."). The difficulty here is that, as Moldea (I) pointed out, this circuit rejected the so-called "incremental harm rule" in Liberty Lobby, Inc. v. Anderson, 746 F.2d 1563, 1568 (D.C. Cir. 1984), vacated on other grounds, 477 U.S. 242, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). See Moldea (I), 15 F.3d at 1149-50. Application of the "substantial truth" test when "incremental harm" is not tolerated can be conceptually confusing. However, on reconsidering the instant dispute, we believe Moldea (I) read this court's rejection of the incremental harm rule much too broadly, and that Anderson's proscription is not applicable in this case.

Liberty Lobby v. Anderson rejected the defendants' claim that the plaintiffs in that case were "libel-proof" because "unchallenged portions of [defendants'] articles attribute[d] to the [plaintiffs] characteristics so much worse than those attributed in the challenged portions, that the latter could not conceivably do any incremental amage." 746 F.2d at 1568. Then-Judge Scalia observed that:

The law, however, proceeds upon the optimistic premise that there is a little bit of good in all of us — or perhaps upon the pessimistic assumption that no matter how bad someone is, he can always be worse. It is shameful that Benedict Arnold was a traitor; but he was not a shoplifter to boot, and one should not be able to make that charge while knowing its falsity with impunity."

Id. However, the opinion goes on to note that:

There may be validity to the proposition that at some point the erroneous attribution of incremental evidence of a character flaw of a particular type which is in any event amply established by the facts is not derogatory. If, for example, an individual is said to have been convicted of 35 burglaries, when the correct number is 34, it is not likely that the statement is actionable. That is so, however, not because the object of the remarks is "libel-proof," but because, since the essentially derogatory implication of the statement ("he is an habitual burglar") is correct, he has not been libeled.

Id. at 1568-69 n. 6. This latter point is dispositive of the instant case.

The disputed "sinister meeting" passage in the Times review is not inherently defamatory — i.e., it is not like calling Benedict Arnold a "traitor" and a "shoplifter," to cite the example used in Anderson. Rather, the discussion of the "sinister meeting" is but one of several interpretations of the book offered to support the claim of "sloppy journalism." As such, it does not come within the compass of "incremental harm." Because the review relies principally on statements that are true, supported opinions or supportable interpretations to justify the "sloppy journalism" assessment, we are constrained to find that it is substantially true and therefore not actionable.

III. INCIDENTAL ISSUES

Because we uphold the District Court's grant of summary judgment for the Times on Moldea's defamation claim, Moldea's related claim for false light invasion of privacy must also fail. As we noted in Moldea (I), "a plaintiff may not avoid the strictures of the burdens of proof associated with defamation by resorting to a claim of false light invasion." Moldea (I), 15 F.3d at 1151 (citing Cohen v. Cowles Media Co., 501 U.S. 663, ___, 111 S.Ct. 2513, 2519, 115 L.Ed.2d 586 (1991); Hustler Magazine, Inc. v. Falwell, 485 U.S. 46, 56, 108 S.Ct. 876, 882, 99 L.Ed.2d 41 (1988)). Similarly, our resolution of Moldea's defamation claim requires that we uphold the District Court's denial of permission to amend the Complaint in the instant case. See Moldea v. New York Times Co., 793 F.Supp. 338 (D.D.C. 1992). Appellant sought to add four causes of action to his Complaint, based upon the Supreme Court's decision in Cohen v. Cowles Media Co. In light of our resolution of Moldea's libel claim, we agree with the trial court's assessment that the amended Complaint could not withstand a motion to dismiss, and so would be futile. As Cohen itself held, a plaintiff may not use related causes of action to avoid the constitutional requisites of a defamation claim. See Cohen, 501 U.S. at ___, 111 S.Ct. at 2519.

IV. CONCLUSION

Moldea has made a number of allegations in this suit that Gerald Eskenazi's negative review of Interference was prompted in part by Eskenazi's allegiance to the National Football League ("NFL"). Moldea alleged in his original brief to this court that Eskenazi has covered the NFL as a correspondent for over thirty years, and that he was therefore biased against Interference because he was dependent on the league's goodwill in order to gain access to information necessary to report on its activities. See Brief for Appellant at 8-9. Indeed, Moldea's book predicts that the NFL's "loyal sportswriters" will try to discredit Interference. See Interference at 25. Even if true, however, these allegations do not make a case for appellant.

Any intelligent reviewer knows at some level that a bad review may injure the author of the book which is its subject. Indeed, some bad reviews may be written with an aim to damage a writer's reputation. There is nothing that we can do about this, at least not without unacceptably interfering with free speech. There simply is no viable way to distinguish between reviews written by those who honestly believe a book is bad, and those prompted solely by mischievous intent. To allow a plaintiff to base a lawsuit on claims of mischief, without some indication that the review's interpretations are unsupportable, would wreak havoc on the law of defamation. See McBride v. Merrell Dow Pharmaceuticals, Inc., 717 F.2d 1460, 1466 (D.C. Cir. 1983) ("Libel suits, if not carefully handled, can threaten journalistic independence. Even if many actions fail, the risks and high costs of litigation may lead to undesirable forms of self-censorship.").

"As James Madison pointed out in the Report on the Virginia Resolutions of 1798: `Some degree of abuse is inseparable from the proper use of everything; and in no instance is that more true than in that of the press.'" Gertz, 418 U.S. at 340, 94 S.Ct. at 3007. We are not insensitive to the fact that the "supportable interpretation" rule may permit some malicious reviews to go unchecked; however, "[b]ecause an ad hoc resolution of the competing interests at stake in each particular case is not feasible, we must lay down broad rules of general application." Id. at 343-44, 94 S.Ct. at 3009.

This is a difficult case, and we can easily understand the frustrations that have prompted Mr. Moldea's long legal battle. Upon reconsideration, however, we find that our first opinion in this case was misguided. Accordingly, we modify that opinion as indicated herein, and affirm the grant of summary judgment in favor of appellee New York Times Company.

So ordered.