Opinion for the Court filed by Circuit Judge STARR.

Concurring opinion filed by Circuit Judge BORK, with whom Circuit Judges WILKEY, GINSBURG and Senior Circuit Judge MacKINNON join.

Concurring opinion filed by Senior Circuit Judge MacKINNON.

Opinion dissenting in part, filed by Chief Judge SPOTTSWOOD W. ROBINSON, III, with whom Circuit Judge J. SKELLY WRIGHT joins.

Opinion dissenting in part filed by Circuit Judge WALD, with whom Circuit Judges HARRY T. EDWARDS and SCALIA join.

Statement concurring in part and dissenting in part filed by Circuit Judge HARRY T. EDWARDS.

Opinion dissenting in part filed by Circuit Judge SCALIA, with whom Circuit Judges WALD and HARRY T. EDWARDS join.

This defamation action arises out of the publication of a syndicated column by Rowland Evans and Robert Novak in May 1978. The question before us is whether the allegedly defamatory statements set forth in the column are constitutionally protected expressions of opinion or, as appellant contends, actionable assertions of fact. We conclude, as did the District Court, that the challenged statements are entitled to absolute First Amendment protection as expressions of opinion.

* Rowland Evans and Robert Novak are nationally syndicated columnists whose columns appear regularly in newspapers across the country. According to the complaint in this case, which was filed by plaintiff Bertell Ollman on February 15, 1979, an Evans and Novak column appeared on or about May 4, 1978 in The Washington Post and other newspapers across the Nation. Complaint p 5. Attached to the complaint as Exhibit A was a photocopy of the column, styled "The Marxist Professor's Intentions," as it appeared in The Washington Post on May 4, 1978. A copy of that column is reproduced as an Appendix to this opinion.

The plaintiff, Bertell Ollman, is a professor of political science at New York University. The complaint averred that Mr. Ollman "is widely esteemed among his colleagues and enjoys the highest possible reputation as a scholar of integrity and a teacher." Complaint p 2. In March 1978, Mr. Ollman was nominated by a departmental search committee to head the Department of Government and Politics at the University of Maryland. The committee's recommendation "was duly approved by the Provost of the University and the Chancellor of the College Park campus." Id. p 4.

With this professional move from Washington Square to College Park, Maryland thus in the offing, the Evans and Novak article appeared. Since the years of litigation that have followed revolve entirely around this single column, we will begin by describing its contents in some detail. In our description, we will highlight the specific portions that Mr. Ollman assails as false and defamatory. The column begins as follows:

What is in danger of becoming a frivolous public debate over the appointment of a Marxist to head the University of Maryland's department of politics and government has so far ignored this unspoken concern within the academic community: the avowed desire of many political activists to use higher education for indoctrination.

The column immediately goes on to state that:

[t]he proposal to name Bertell Ollman, Professor at New York University, as department head has generated wrong-headed debate. Politicians who jumped in to oppose Ollman simply for his Marxist philosophy have received a justifiable going-over from defenders of academic freedom in the press and the university. Academic Prince Valiants seem arrayed against McCarythite [sic] know-nothings.

With these opening two paragraphs as lead-in, the authors then pose what they deemed the pivotal issue in the debate: "But neither side approaches the crucial question: not Ollman's beliefs, but his intentions. His candid writings avow his desire to use the classroom as an instrument for preparing what he calls 'the revolution.' Whether this is a form of indoctrination that could transform the real function of a university and transcend limits of academic freedom is a concern to academicians who are neither McCarthyite nor know-nothing." (Emphasis added).

The columnists thus, in the first three paragraphs, articulated a view of what should be the central question in what they viewed as a fruitless debate. The authors then go on in the next paragraph to state: "To protect academic freedom, that question should be posed not by politicians but by professors. But professors throughout the country troubled by the nomination, clearly a minority, dare not say a word in today's campus climate."

With this observation, the authors turn in the following six paragraphs to a discussion of Mr. Ollman and his writings. Evans and Novak state that "[w]hile Ollman is described in news accounts as a 'respected Marxist scholar,' he is widely viewed in his profession as a political activist. Amid the increasingly popular Marxist movement in university life, he is distinct from philosophical Marxists. Rather, he is an outspoken proponent of 'political Marxism.' " (Emphasis added).

The authors next relate Mr. Ollman's two unsuccessful efforts to win election to membership on the council of the American Political Science Association. In these elections, the column states (and appellant does not dispute) that Professor Ollman ran as a candidate of the Caucus for a New Political Science and finished last out of sixteen candidates each time. "Whether or not that represents a professional judgment by his colleagues, as some critics contend, the verdict clearly rejected his campaign pledge: 'If elected ... I shall use every means at my disposal to promote the study of Marxism and Marxist approaches to politics throughout the profession.' "

Evans and Novak then direct the four ensuing paragraphs of the column to a summary of an article by Mr. Ollman, entitled "On Teaching Marxism and Building the Movement" in the Winter 1978 issue of New Political Science. Record ("R.") 3. In this article, Mr. Ollman claims that most students conclude his political science course with a " 'Marxist outlook.' " The authors go on:

Ollman concedes that will be seen "as an admission that the purpose of my course is to convert students to socialism."

That bothers him not at all because "a correct understanding of Marxism (as indeed of any body of scientific truths) leads automatically to its acceptance." * * * The "classroom" is a place where the students' bourgeois ideology is being dismantled. "Our prior task" before the revolution, he writes, "is to make more revolutionaries."1

Moving to a brief discussion of Mr. Ollman's principal work, Alienation: Marx's Conception of Man in Capitalist Society, the authors described the work as "a ponderous tome in adoration of the master (Marxism 'is like a magnificiently rich tapestry'). Published in 1971, it does not abandon hope for the revolution forecast by Karl Marx in 1848." This brings the columnists to the last statement specifically identified in the complaint as defamatory:

Such pamphleteering is hooted at by one political scientist in a major eastern university, whose scholarship and reputation as a liberal are well known. "Ollman has no status within the profession, but is a pure and simple activist," he said. Would he say that publicly? "No chance of it. Our academic culture does not permit the raising of such questions." (Emphasis added).

Evans and Novak then bring the column to a close, indicating in the penultimate paragraph that " '[s]uch questions' would include these: What is the true measurement of Ollman's scholarship? Does he intend to use the classroom for indoctrination? Will he indeed be followed by other Marxist professors? Could the department in time be closed to non-Marxists, following the tendency at several English universities?"

In the column's final paragraph, the authors return to their opening theme that "such questions" as set forth in the previous paragraph should not be raised by politicians, even if, as the anonymous political scientist claimed, they cannot be raised within the Academy. They conclude the column by calling upon academics to address these questions:

Here are the makings of a crisis that, to protect its integrity and true academic freedom, academia itself must resolve.

On May 19, 1978, Mr. Ollman's lawyer wrote to Evans and Novak demanding retraction of the allegedly defamatory statements in the column. Letter of I. Silver to R. Evans and R. Novak (May 19, 1978). R. 1. This Evans and Novak refused to do. On May 8, however, only four days after the Evans and Novak column appeared, The Washington Post published a letter from Mr. Ollman. In this letter, Professor Ollman rejected the allegation that he used the classroom to indoctrinate students and set the column's quotations from his writings in what he viewed as their proper context. Letter from B. Ollman to the Editors of The Washington Post (May 8, 1978). R. 3.

The District Court granted Evans and Novak's motion for summary judgment, concluding that the column simply reflected the columnists' opinion and their "interpretation of [Mr. Ollman's] writings." Memorandum Opinion at 5.2 Thus, the District Court held that the opinion was absolutely protected by the First Amendment. This appeal followed.

* This case presents us with the delicate and sensitive task of accommodating the First Amendment's protection of free expression of ideas with the common law's protection of an individual's interest in reputation. It is a truism that the free flow of ideas and opinions is integral to our democratic system of government. Thomas Jefferson well expressed this principle in his First Inaugural Address, when the Nation's memory was fresh with the passage of the notorious Alien and Sedition Acts:

If there be any among us who would wish to dissolve this Union or to change its republican form, let them stand undisturbed as monuments of the safety with which error of opinion may be tolerated where reason is left free to combat it.3

At the same time, an individual's interest in his or her reputation is of the highest order. Its protection is an eloquent expression of the respect historically afforded the dignity of the individual in Anglo-American legal culture.4 A defamatory statement may destroy an individual's livelihood, wreck his standing in the community, and seriously impair his sense of dignity and self-esteem.

The judiciary's task in accommodating these competing interests is by no means new: at common law, the fair comment doctrine bestowed qualified immunity from libel actions as to certain types of opinions in order that writers could express freely their views about subjects of public interest.5 However, since Gertz v. Robert Welch, Inc., 418 U.S. 323, 94 S.Ct. 2997, 41 L.Ed.2d 789 (1974), the nature of this accommodation has fundamentally changed. In Gertz, the Supreme Court in dicta seemed to provide absolute immunity from defamation actions for all opinions and to discern the basis for this immunity in the First Amendment. The Court began its analysis of the case by stating:

Under the First Amendment there is no such thing as a false idea. However pernicious an opinion may seem, we depend for its correction not on the conscience of judges and juries but on the competition of other ideas. But there is no constitutional value in false statements of fact. Neither the intentional lie nor the careless error materially advances society's interest in "uninhibited, robust, and wide-open debate on the public issues."6

By this statement, Gertz elevated to constitutional principle the distinction between fact and opinion, which at common law had formed the basis of the doctrine of fair comment.7 Gertz's implicit command thus imposes upon both state and federal courts the duty as a matter of constitutional adjudication to distinguish facts from opinions in order to provide opinions with the requisite, absolute First Amendment protection.8 At the same time, however, the Supreme Court provided little guidance in Gertz itself as to the manner in which the distinction between fact and opinion is to be discerned. That, as we shall see, is by no means as easy a question as might appear at first blush.

Indeed, Gertz did not focus on this distinction at all. Rather, assuming without lengthy discussion that the statements in that case could be construed as statements of fact, the Court held that the plaintiff, who was a private rather than public figure, could prove that the statements at issue there were libelous upon demonstrating that they were negligently made.9 The distinction in our law between public and private figures, however, does not directly bear on the distinction between fact and opinion.10 Expressions of opinion are protected whether the subject of the comment is a private or public figure. See Lewis v. Time, Inc., 710 F.2d 549, 555 (9th Cir.1983). In a word, Gertz's reasoning immunizes an opinion, not because the opinion is asserted about a public figure, but because there is no such thing as a "false" opinion.

While Gertz is mute with respect to the method of separating fact from opinion, two Supreme Court cases do provide guidance in this respect. Old Dominion Branch No. 496, National Association of Letter Carriers v. Austin, 418 U.S. 264, 94 S.Ct. 2770, 41 L.Ed.2d 745 (1974); Greenbelt Cooperative Publishing Association v. Bresler, 398 U.S. 6, 90 S.Ct. 1537, 26 L.Ed.2d 6 (1970).11 In Letter Carriers, decided by the Court on the same day as Gertz, three non-union employees of the Postal Service were included on a list of names circulated by the Letter Carriers union. To the list was appended a well-known piece of trade union literature attributed to Jack London, which defined the nature of a "scab." Drawing upon Biblical references and American history, the definition ended on the following unflattering note:

The scab sells his birthright, country, his wife and his children and his fellowmen for an unfulfilled promise from his employer.

Esau was a traitor to himself; Judas was a traitor to his God; Benedict Arnold was a traitor to his country; a SCAB is a traitor to his God, his country, his family and his class.

Letter Carriers, 418 U.S. at 268, 94 S.Ct. at 2773. Holding this allegedly defamatory language to be absolutely protected, the Supreme Court reversed a libel judgment in favor of the non-union employees. While the Court grounded its decision upon federal labor laws' protection of communications in a labor dispute, rather than the First Amendment, the Court's analysis derived from Gertz's proposition that opinions cannot be false. Id. 418 U.S. at 284, 94 S.Ct. at 2781 (citing Gertz, supra, 418 U.S. at 339-340, 94 S.Ct. at 3006-3007). To demonstrate that the union's "scab" description was indeed opinion, the Court considered both its specific linguistic context and its broader social setting. The Court found, for instance, that the epithet "traitor" in the context of a well-known piece of union literature was deployed in a "loose, figurative sense" and could not be taken for an assertion that the identified employees had "committ[ed] the criminal offense of treason." Id. at 284-85, 94 S.Ct. at 2781. Moving to the social context in which the statement was made, the Court further noted that this type of "exaggerated rhetoric was commonplace in labor disputes." Thus, the Court concluded, readers would be alerted by virtue of the broad context in which the statement was made that the statement was opinion, not an imputation of actual criminal conduct. Id. at 286, 94 S.Ct. at 2782.

Letter Carriers also relied upon Greenbelt Publishing, supra, for the proposition that the allegedly libelous language must be evaluated in its broader context to assess whether a reader would have understood the allegation to be a statement of fact. Id. at 284, 94 S.Ct. at 2781. Greenbelt Publishing was, of course, a pre-Gertz case, which may be seen in retrospect as an application of the distinction between fact and opinion subsequently delineated in Gertz. In Greenbelt Publishing, a developer was attempting to secure zoning variances to construct high density housing; at the same time, the city of Greenbelt, Maryland was trying to purchase land from the developer to build a school. During the course of these negotiations, some attendees at a public meeting characterized the developer's negotiating tactics as "blackmail." The developer thereafter brought a successful libel suit against a local newspaper that printed this colorful characterization. The Supreme Court reversed the judgment, concluding that "as a matter of constitutional law, the word "blackmail" in these circumstances was ... not libel when reported." Id., 398 U.S. at 13, 90 S.Ct. at 1541. The Court noted that the Greenbelt newspaper was performing a wholly "legitimate function as a community newspaper" and that it "accurately and fully" described the developer's negotiating proposals. The Court then held that in light of the full context of the articles a reader would have understood the "blackmail" characterization as a criticism of the developer's negotiating tactics rather than as an actual criminal charge. Under the circumstances, the remark was deemed to be merely "rhetorical hyperbole." Id. at 14, 90 S.Ct. at 1542.

There is, then, limited but helpful teaching from the Supreme Court to guide us in our inquiry. With largely uncharted seas having been left in Gertz's wake, the lower federal courts and state courts have, not surprisingly, fashioned various approaches in attempting to articulate the Gertz-mandated distinction between fact and opinion. We pause here, briefly, to examine the results of the efforts of our fellow laborers in this new constitutional vineyard.

Some courts have, in effect, eschewed any effort to construct a theory and simply treated the distinction between fact and opinion as a judgment call. See, e.g., Shiver v. Apalachee Publishing Co., 425 So.2d 1173 (Fla.Dist.Ct.App.1983). Other courts have concentrated on a single factor, such as the verifiability vel non of the allegedly defamatory statement. See, e.g., Hotchner v. Castillo-Puche, 551 F.2d 910, 913 (2d Cir.), cert. denied sub. nom. Hotchner v. Doubleday & Co., 434 U.S. 834, 98 S.Ct. 120, 54 L.Ed.2d 95 (1977). Still others have adopted a multi-factor test, attempting to assess the allegedly defamatory proposition in the totality of the circumstances in which it appeared. See, e.g., Information Control Corp. v. Genesis One Computer Corp., 611 F.2d 781 (9th Cir.1980).12

In formulating a test to distinguish between fact and opinion, courts are admittedly faced with a dilemma. Because of the richness and diversity of language, as evidenced by the capacity of the same words to convey different meanings in different contexts, it is quite impossible to lay down a bright-line or mechanical distinction.13 Judicial decisions, however, that represent mere ad hoc judgments or which, in contrast, lay down rules of excessive complexity may deter publication of the very opinions which the Gertz-mandated distinction is designed to protect, inasmuch as potential speakers or writers would, under such regimes, be at a loss to predict what courts will ultimately deem to be opinion. While this dilemma admits of no easy resolution, we think it obliges us to state plainly the factors that guide us in distinguishing fact from opinion and to demonstrate how these factors lead to a proper accommodation between the competing interests in free expression of opinion and in an individual's reputation.

In formulating this analysis, we agree with the overwhelming weight of post-Gertz authority that the distinction between opinion and fact is a matter of law. See, e.g., Lewis v. Time, Inc., supra, 710 F.2d at 553; Rinsley v. Brandt, 700 F.2d 1304, 1309 (10th Cir.1983); Orr v. Argus-Press Co., 586 F.2d 1108, 1114 (8th Cir.), cert. denied, 440 U.S. 960, 99 S.Ct. 1502, 59 L.Ed.2d 773 (1979). Although the Supreme Court has never directly addressed this issue, the Court has clearly ruled that questions as to other privileges derived from the First Amendment, such as the qualified privilege as to public officials and public figures, are to be decided as matters of law. See Gertz, supra, 418 U.S. at 346, 94 S.Ct. at 3010. Moreover, the predictability of decisions, which is of crucial importance in an area of law touching upon First Amendment values, is enhanced when the determination is made according to announced legal standards and when a body of public case law furnishes published examples of the manner in which these standards are to be applied.

While courts are divided in their methods of distinguishing between assertions of fact and expressions of opinion, they are universally agreed that the task is a difficult one. See, e.g., Rinsley v. Brandt, supra, 700 F.2d at 1309. To be sure, paradigm examples of statements of fact, on the one hand, and paradigm examples of expressions of opinion, on the other, can be contrasted. Clearly, in the former category are assertions that describe present or past conditions capable of being known through sense impressions. See Goodrich v. Waterbury Republican-American, Inc., 448 A.2d 1317, 1321 (Conn.1982) (citing 1 F. Harper & F. James, Torts Sec. 5.28, p. 458 n. 11, Sec. 7.8, p. 560). It is rather hard to imagine a context in which the statement, "Mr. Jones had ten drinks at his office party and sideswiped two vehicles on his way home," could be deemed to be a statement of opinion. At the other extreme are evaluative statements reflecting the author's political, moral, or aesthetic views, not the author's sense perceptions. A statement such as, "Mr. Jones is a despicable politician," is a paradigm of opinion.

It is a fitting illustration of the complexity of language and communication that many statements from which actions for defamation arise do not clearly fit into either category. These statements pose more subtle problems and are the stuff of which litigation is made. The principal difficulty arises from statements that on first analysis seem to be based upon perceptions of events, but are not themselves simply a record of those perceptions. Such statements may imply in some contexts the existence of facts not disclosed by the author.14 An example of such a statement, set forth in the Restatement (Second) of Torts, is: "Mr. Jones is an alcoholic."15 These statements obviously can be as damaging to reputation as statements which on their face describe particular historical events.

The degree to which such kinds of statements have real factual content can, of course, vary greatly. We believe, in consequence, that courts should analyze the totality of the circumstances in which the statements are made to decide whether they merit the absolute First Amendment protection enjoyed by opinion. To evaluate the totality of the circumstances of an allegedly defamatory statement, we will consider four factors in assessing whether the average reader would view the statement as fact or, conversely, opinion.16 While necessarily imperfect, these factors will, we are persuaded, assist in discerning as systematically as possible what constitutes an assertion of fact and what is, in contrast, an expression of opinion.

First, we will analyze the common usage or meaning of the specific language of the challenged statement itself. Our analysis of the specific language under scrutiny will be aimed at determining whether the statement has a precise core of meaning for which a consensus of understanding exists or, conversely, whether the statement is indefinite and ambiguous. See Buckley v. Littell, 539 F.2d 882, 895 (2d Cir.1976), cert. denied, 429 U.S. 1062, 97 S.Ct. 785, 50 L.Ed.2d 777 (1977). Readers are, in our judgment, considerably less likely to infer facts from an indefinite or ambiguous statement than one with a commonly understood meaning. Second, we will consider the statement's verifiability--is the statement capable of being objectively characterized as true or false? See, e.g., Hotchner v. Castillo-Puche, supra, 551 F.2d at 913. Insofar as a statement lacks a plausible method of verification, a reasonable reader will not believe that the statement has specific factual content. And, in the setting of litigation, the trier of fact obliged in a defamation action to assess the truth of an unverifiable statement will have considerable difficulty returning a verdict based upon anything but speculation. Third, moving from the challenged language itself, we will consider the full context of the statement--the entire article or column, for example--inasmuch as other, unchallenged language surrounding the allegedly defamatory statement will influence the average reader's readiness to infer that a particular statement has factual content. See Greenbelt Cooperative Publishing Association v. Bresler, supra, 398 U.S. at 13-14, 90 S.Ct. at 1541; cf. Restatement (Second) of Torts Sec. 563. Finally, we will consider the broader context or setting in which the statement appears. Different types of writing have, as we shall more fully see, widely varying social conventions which signal to the reader the likelihood of a statement's being either fact or opinion. See Old Dominion Branch No. 496, National Association of Letter Carriers v. Austin, supra, 418 U.S. at 286, 94 S.Ct. at 2782.

* The first factor of our inquiry is to analyze the common usage or meaning of the allegedly defamatory words themselves.17 We seek in this branch of our analysis to determine whether the allegedly defamatory statement has a precise meaning and thus is likely to give rise to clear factual implications.18 A classic example of a statement with a well-defined meaning is an accusation of a crime. To be sure, such accusations are not records of sense perceptions. Quite to the contrary, they depend for their meaning upon social normative systems. But those norms are so commonly understood that the statements are seen by the reasonable reader or hearer as implying highly damaging facts. Post-Gertz courts have therefore not hesitated to hold that accusations of criminal conduct are statements "laden with factual content" that may support an action for defamation. See, e.g., Cianci v. New Times Publishing Co., 639 F.2d 54, 63 (2d Cir.1980) (holding that an article which implied that the Mayor of Providence, R.I., had committed rape and which charged him with paying the alleged victim not to bring charges was not protected opinion). Even a somewhat less well defined accusation that a "judge is corrupt" has been held actionable. Rinaldi v. Holt, Rinehart & Winston, Inc., 42 N.Y.2d 1299, 397 N.Y.S.2d 943, 366 N.E.2d 1299, cert. denied, 434 U.S. 969, 98 S.Ct. 514, 54 L.Ed.2d 456 (1977). "Corruption," at least in the context of public service, was deemed to imply factual allegations of bribery or other official malfeasance.

On the other hand, statements that are "loosely definable" or "variously interpretable" cannot in most contexts support an action for defamation. See Buckley v. Littell, supra, 539 F.2d at 895. In that case, a writer in his book on the political right in the United States accused columnist and author William F. Buckley, Jr., of being a "fellow traveler" of "fascists." Noting that Mr. Buckley and the author of this particular tome embraced widely different definitions of "fascism" and different views as to which journals could be described as "fascist," the court declined to develop a "correct" definition of this pivotal term.19 The Second Circuit held, rather, that the use of such expressions "cannot be regarded as having been proved to be statements of facts, among other reasons, because of the tremendous imprecision of the meaning and usage of these terms in the realm of political debate, an imprecision which is similarly echoed in the book."20 Id. at 893. Pursuing a line of analysis similar to that found in Buckley, the same court that held actionable the term "corrupt" concluded that the term "incompetent" as applied to a judge was too vague to support a claim of libel. Rinaldi v. Holt, Rinehart & Winston, Inc., supra, 397 N.Y.S. at 947, 366 N.E.2d at 1303.

The use of indefinite terms is obviously not confined to the realm of politics and public policy. In Cole v. Westinghouse Broadcasting Co., Inc., 386 Mass. 303, 435 N.E.2d 1021, cert. denied, 459 U.S. 1037, 103 S.Ct. 449, 74 L.Ed.2d 603 (1982), the Massachusetts Supreme Judicial Court held that the statement that a reporter had engaged in "sloppy and irresponsible reporting" and had poor reporting technique was too "imprecise" to support a defamation action.21 Similarly, in Avins v. White, 627 F.2d 637 (3d Cir.), cert. denied, 449 U.S. 982, 101 S.Ct. 398, 66 L.Ed.2d 244 (1982), the former dean of a law school claimed that his academic ability and performance had been falsely disparaged in the summary evaluation of the school's first accreditation report. The summary bluntly stated: "[T]he most important deficiency [of the law school] is an intangible one; there is an academic ennui that pervades the institution. The intellectual spark is missing in the faculty and students." Id. at 642. Emphasizing that the statement itself described its criticism as "intangible," the Avins court classified the statement as an expression of opinion.

The straightforward but important principle to be drawn from cases such as Buckley, Rinaldi, Cole and Avins is that in all types of discourse, the courts must analyze the allegedly defamatory statement to determine whether it has a sufficiently definite meaning to convey facts.

In assessing whether the challenged statements are facts, rather than opinion, courts should, secondly, consider the degree to which the statements are verifiable--is the statement objectively capable of proof or disproof? See Goodrich v. Waterbury Republican-American, supra, 448 A.2d at 1319; Hotchner v. Castillo-Puche, supra, 551 F.2d at 913.22 The reason for this inquiry is simple: a reader cannot rationally view an unverifiable statement as conveying actual facts. Moreover, insofar as a statement is unverifiable, the First Amendment is endangered when attempts are made to prove the statement true or false. Lacking a clear method of verification with which to evaluate a statement--such a labelling a well-known American author a "fascist," see Buckley v. Littell, supra--the trier of fact may improperly tend to render a decision based upon approval or disapproval of the contents of the statement, its author, or its subject.

In making this observation, we imply no criticism of a jury's ability to find facts, if facts are to be found. The rule against allowing unverifiable statements to go to the jury is, in actuality, merely one of many rules in tort law that prevent the jury from rendering a verdict based on speculation. Cf. Hobson v. Wilson, 737 F.2d 1 at 62 (D.C.Cir.1984) (permitting First Amendment interests to be compensated "if they can be conceptualized and if harm can be shown with sufficient certainty to avoid damages based ... on pure speculation"). An obvious potential for quashing or muting First Amendment activity looms large when juries attempt to assess the truth of a statement that admits of no method of verification.

Needless to say, it will often be difficult to assay whether a statement is verifiable. Statements made in written communication or discourse range over a spectrum with respect to the degree to which they can be verified rather than dividing neatly into categories of "verifiable" and "unverifiable." But even if the principle of inquiring as to verifiability provides no panacea, this approach will nonetheless aid trial judges in assessing whether a statement should have the benefit of the absolute privilege conferred upon expressions of opinion. Trial judges have rich experience in the ways and means of proof and so will be particularly well situated to determine what can be proven.

In addition to evaluating the precision-indefiniteness and verifiability-unverifiability of a challenged statement, courts should, thirdly, examine the context in which the statement occurs. Readers will inevitably be influenced by a statement's context, and the distinction between fact and opinion can therefore be made only in context. As the Supreme Court's opinions in Greenbelt and Letter Carriers suggest, the context to be considered is both narrowly linguistic and broadly social.

The degree to which a statement is "laden with factual content" or can be read to imply facts depends upon the article or column, taken as a whole, of which the statement is a part. See Information Control v. Genesis One Computer, supra, 611 F.2d at 783. The language of the entire column may signal that a specific statement which, standing alone, would appear to be factual is in actuality a statement of opinion. An example of the power of context to transform an ostensibly factual statement into one of opinion is Greenbelt Publishing. See supra I A. Because the local newspaper in that case had described the substance of the land developer's negotiating proposals, the use of the term "blackmail" to characterize those proposals was quite plainly to be seen as an expression of opinion.23

An article or column, however, plainly does not have to include a complete set of facts to make it clear that a statement is being used in a metaphorical, exaggerated or even fantastic sense. In Myers v. Boston Magazine Co., Inc., 380 Mass. 336, 403 N.E.2d 376 (1980), the court held as protected opinion a magazine's statement that a television sports reporter was "the only newscaster in town who is enrolled in a course for remedial speaking." Id., 403 N.E.2d at 377. Although the statement on its face appears quite factual, the court emphasized in its analysis that the statement appeared in an article describing the best and worst sports personalities in a series of "one-liners." Id. For instance, the court noted that another item in the article described the Boston Bruins hockey team members as looking "like a gargoyle" and that the various descriptions had corresponding cartoons. The court concluded that the average reader would have been put on notice that he or she was reading opinions, and not being showered with facts. Id., 403 N.E.2d at 379.

Another consideration in this respect, of particular relevance to the case at hand and useful in distinguishing between fact and opinion, is the inclusion of cautionary language in the text in which the statement at issue is found, see Information Control, supra, 611 F.2d at 784 (noting that the allegedly libelous statement was preceded by the phrase, "In the opinion of Genesis' management" and that this favored treating the statement which followed as opinion), or framing the statement as an interrogatory ("Is it not true that ... ?"). The rationale typically advanced for this consideration is that cautionary language or interrogatories of this type put the reader on notice that what is being read is opinion and thus weaken any inference that the author possesses knowledge of damaging, undisclosed facts. See Pease v. Telegraph Publishing Co., 121 N.H. 62, 426 A.2d 463, 465 (1981). In a word, when the reasonable reader encounters cautionary language, he tends to "discount that which follows." See Burns v. McGraw-Hill Broadcasting Co., 659 P.2d 1351, 1360 (Colo.1983).

To be sure, there is authority against giving weight to cautionary or interrogatory language. Stating that "[i]t would be destructive of the law of libel if a writer could escape liability for accusations of crime simply by using, explicitly or implicitly, the words 'I think,' " Cianci, supra, 639 F.2d at 64, the Second Circuit in an opinion by Judge Friendly rejected the notion that cautionary language could immunize an otherwise defamatory statement. While Judge Friendly's argument is not without force, it may be overstated if applied outside the type of facts before the court in Cianci--the accusation of a crime--since cautionary language is only one of several factors to be considered in assessing an allegedly defamatory statement.24 Burns v. McGraw-Hill, supra, 659 P.2d at 1360 n. 4. When a statement is as "factually laden" as the accusation of a crime, which of course was the issue in Cianci, cautionary language is by and large unavailing to dilute the statement's factual implications. However, in statements less clearly factual, cautionary language may make a more substantial difference to the reader's understanding.

What is more, we cannot forget that the public has an interest in receiving information on issues of public importance even if the trustworthiness of the information is not absolutely certain. The First Amendment is served not only by articles and columns that purport to be definitive but by those articles that, more modestly, raise questions and prompt investigation or debate. By giving weight on the opinion side of the scale to cautionary and interrogative language, courts provide greater leeway to journalists and other writers and commentators in bringing issues of public importance to the public's attention and scrutiny.

Besides looking to the immediate context of the allegedly defamatory statement, courts should examine, finally, the broader social context into which the statement fits. Some types of writing or speech by custom or convention signal to readers or listeners that what is being read or heard is likely to be opinion, not fact.25 It is one thing to be assailed as a corrupt public official by a soapbox orator and quite another to be labelled corrupt in a research monograph detailing the causes and cures of corruption in public service. This observation reflects no novel principle. The Supreme Court has expressly recognized the importance of social context when, in finding as an expression of opinion the use of the word "traitor" as applied to an employee who crossed a picket line, the Court stated that "such exaggerated rhetoric was commonplace in labor disputes." Letter Carriers, supra, 418 U.S. at 286, 94 S.Ct. at 2782.26

Similarly, in Myers v. Boston Magazine, supra, the Massachusetts Supreme Judicial Court was even more explicit in focusing upon the reader's understanding of a particular type of writing. Emphasizing that the "magazine's statement partook of an ancient, lively tradition of criticizing, even lampooning, performers," the court concluded that the statement that a sportscaster was attending a course in remedial speaking constituted privileged opinion. Id., 403 N.E.2d at 381. In the lampooning tradition, the court emphasized, it is well understood that "a critic may resort to caricature and rhetorical license." Id. See also Pring v. Penthouse, Inc., 695 F.2d 438 (10th Cir.1982), cert. denied, --- U.S. ----, 103 S.Ct. 3112, 77 L.Ed.2d 1367 (1983) (finding that the imputation that the plaintiff had committed sexual acts on stage at the Miss America Pageant could not support a libel action when the writing in which the statement appeared was clearly a "fantasy").

Courts have, in the same vein, considered the influence that other well established genres of writing will have on the average reader. Of particular relevance in this respect to the case before us is Loeb v. Globe Newspaper Co., 489 F.Supp. 481 (D.Mass.1980). In that case, the court observed that the article containing the alleged defamations of the publisher of the Manchester Union-Leader was situated on the Boston Globe's editorial page. The court held that, in the specific context or setting at issue there, the statement to the effect that Mr. Loeb never backed a winner in a presidential election was protected opinion. Plainly, the general understanding of the nature of the statements on the editorial page was relevant to the decision; if the statement had appeared on the front page where news is reported, it would most likely have been treated as a statement of fact. See also National Rifle Association v. Dayton Newspapers, Inc., 555 F.Supp. 1299 (S.D.Ohio 1983) (holding that the statement in an editorial that the National Rifle Association "happily encourages ... murders and robberies" was protected opinion). In short, it is well understood that editorial writers and commentators frequently "resort to the type of caustic bombast traditionally used in editorial writing to stimulate public reaction." Id. at 1309. Hence, in analyzing the distinction between fact and opinion, the court will take fully into account the different social conventions or customs inherent in different types of writing.27

After deciding that a particular statement is opinion rather than fact, courts often undertake a second mode of analysis before wrapping the statement in the mantle of the First Amendment's opinion privilege. Relying upon the Restatement (Second) of Torts Sec. 566, the courts consider whether the opinion implies the existence of undisclosed facts as the basis for the opinion.28 If the opinion implied factual assertions, courts have held that it should not receive the benefit of First Amendment protection as an opinion.

We have no quarrel with the purpose of section 566. As we have already seen, categorizing a statement as fact or opinion is a difficult task. Many statements are not simple factual statements or simple opinions, but are statements that are "ladden with factual content." Cianci, supra, 639 F.2d at 63. The Restatement is obviously designed to address the problems posed by such statements. In our view, however, the tests already articulated are a sufficient aid in determining whether a statement implies the existence of undisclosed facts. The definiteness and verifiability of a statement (factors one and two) clearly bear on the ability of a statement to carry factual implications. The linguistic and social context of the statement (factors three and four) will also influence the average reader's readiness to infer from the statement the existence of undisclosed facts. Thus, once our inquiry into whether the statement is an assertion of fact or expression of opinion has concluded, the factors militating either in favor of or against the drawing of factual implications from any statement have already been identified. A separate inquiry into whether a statement, already classified in this painstaking way as opinion, implies allegedly defamatory facts would, in our view, be superfluous. In short, we believe that the application of the four-factor analysis set forth above, and drawn from the considerable judicial teaching on the subject, will identify those statements so "factually laden" that they should not receive the benefit of the opinion privilege.29

We are fortified in this respect by section 566's potential, on occasion, to mislead. Comments to that section may be taken to imply that only the disclosure of facts which form the basis of the statement will signal to the reader that the author is not employing an opinion to imply undisclosed facts.30 To be sure, we fully agree that in some contexts statements should be submitted to the trier of fact, unless the presence of facts surrounding the statement suggests that it is merely a characterization of those facts and thus is best classified, like the characterization in Greenbelt Publishing, supra, as an opinion. For instance, in the context of a front page news story or magazine article, the presence of such facts may be the only factor that would prevent the allegedly defamatory statement from being submitted to the jury.

However, in other contexts, as we have shown above, factors besides the disclosure of facts are relevant in determining whether a statement implies factual allegations to the reasonable reader. Here, for instance, as we shall discuss more fully, that the statements challenged by Professor Ollman were found in a column on the Op-Ed page suggests, among other factors, that the statements would be understood by the reasonable reader as opinion--even in the absence of full disclosure of facts signalling to the reader that the allegedly defamatory statement was a characterization. In a word, disclosure of facts in the surrounding text is not the only signal that hard facts cannot reasonably be inferred from a statement. We think that our four-factor test takes account of the insights provided by section 566, while not rejecting the other factors that may signal that a statement is to be read as opinion.31III

Now we turn to the case at hand to apply the foregoing analysis. As we have seen, Mr. Ollman alleges various instances of defamation in the Evans and Novak column. Before analyzing each such instance, we will first examine the context (the third and fourth factors in our approach) in which the alleged defamations arise. We will then assess the manner in which this context would influence the average reader in interpreting the alleged defamations as an assertion of fact or an expression of opinion.

From the earliest days of the Republic, individuals have published and circulated short, frequently sharp and biting writings on issues of social and political interest. From the pamphleteers urging revolution to abolitionists condemning the evils of slavery, American authors have sought through pamphlets and tracts both to stimulate debate and to persuade. Today among the inheritors of this lively tradition are the columnists and opinion writers whose works appear on the editorial and Op-Ed pages of the Nation's newspapers. The column at issue here is plainly part and parcel of this tradition of social and political criticism.32

The reasonable reader who peruses an Evans and Novak column on the editorial or Op-Ed page is fully aware that the statements found there are not "hard" news like those printed on the front page or elsewhere in the news sections of the newspaper. Readers expect that columnists will make strong statements, sometimes phrased in a polemical manner that would hardly be considered balanced or fair elsewhere in the newspaper. National Rifle Association v. Dayton Newspaper, Inc., supra, 555 F.Supp. at 1309. That proposition is inherent in the very notion of an "Op-Ed page." Because of obvious space limitations, it is also manifest that columnists or commentators will express themselves in condensed fashion without providing what might be considered the full picture. Columnists are, after all, writing a column, not a full-length scholarly article or a book. This broad understanding of the traditional function of a column like Evans and Novak will therefore predispose the average reader to regard what is found there to be opinion.33

A reader of this particular Evans and Novak column would also have been influenced by the column's express purpose. The columnists laid squarely before the reader their interest in ending what they deemed a "frivolous" debate among politicians over whether Mr. Ollman's political beliefs should bar him from becoming head of the Department of Government and Politics at the University of Maryland. Instead, the authors plainly intimated in the column's lead paragraph that they wanted to spark a more appropriate debate within academia over whether Mr. Ollman's purpose in teaching was to indoctrinate his students. Later in the column, they openly questioned the measure or method of Professor Ollman's scholarship. Evans and Novak made it clear that they were not purporting to set forth definitive conclusions, but instead meant to ventilate what in their view constituted the central questions raised by Mr. Ollman's prospective appointment. In the penultimate paragraph of the column, as we have already seen, the authors expressly posed the following "questions:"

What is the true measurement of Ollman's scholarship? Does he intend to use the classroom for indoctrination? Will he indeed be followed by other Marxist professors? Could the department in time become closed to non-Marxists, following the tendency at several English universities?

Prominently displayed in the Evans and Novak column, therefore, is interrogatory or cautionary language that militates in favor of treating statements as opinion.

* Having reviewed the context of the challenged statements, we turn next to the alleged defamation that, in our view, is most clearly opinion, namely that "[Ollman] is an outspoken proponent of political Marxism." This kind of characterization is much akin to the characterization, "fascist," found absolutely protected in Buckley v. Littell, supra. This is unmistakably a "loosely definable, variously interpretable statement[ ] of opinion ... made inextricably in the contest of political, social or philosophical debate ...." 539 F.2d at 895. It is obviously unverifiable. Since Mr. Ollman concedes that he is a Marxist, see Letter of B. Ollman to the Editors of The Washington Post (May 8, 1978), R. 3, the trier of fact in assessing the statement would have the dubious task of trying to distinguish "political Marxism" from "non-political Marxism," whatever that may be.

Nor is the statement that "[Mr. Ollman] is widely viewed in his profession as a political activist" a representation or assertion of fact. "Political activist" is a term, like "political Marxism," that is hopelessly imprecise and indefinite. It is difficult to imagine, much less construct, a means of deciding the quantum of political activity justifying the label "activist." While Mr. Ollman argues that this assertion is defamatory since it implies that he has no reputation as a scholar, we are rather skeptical of the strength of that implication, particularly in the context of this column. It does not appear the least bit evident that "scholarship" and "political activism" are generally understood to be incompatible. Moreover, Evans and Novak set out facts which signalled to the reader that this statement represents a characterization arising from the columnists' view of the facts. In the paragraph immediately following this statement, the column indicated that Mr. Ollman on no less than two occasions finished dead last among all candidates for election to the governing Council of the American Political Science Association, when he ran on the platform: "If elected ... I shall use every means at my disposal to promote the study of Marxism and Marxist approaches to politics throughout the profession." A reasonable reader would conclude that the authors' judgment that Mr. Ollman was "widely viewed as a political activist" was a characterization based upon the latter's unsuccessful electoral endeavors within his profession.

Next we turn to Mr. Ollman's complaints about the column's quotations from and remarks about his writings, and specifically his article, "On Teaching Marxism and Building the Movement."34 We note in this respect that even before the appearance of the constitutionally based opinion privilege in Gertz, commentary on another's writing was considered a privileged occasion at common law and therefore received the benefit of the fair comment doctrine.35 When a critic is commenting about a book, the reader is on notice that the critic is engaging in interpretation, an inherently subjective enterprise, and therefore realizes that others, including the author, may utterly disagree with the critic's interpretation.36 The average reader further understands that because of limitations of space, not to mention those limitations imposed by the patience of the prospective audience, the critic as a practical matter will be able to support his opinion only by rather truncated quotations from the book or work under scrutiny. The reader is thus predisposed to view what the critic writes as opinion. In this context, courts have rightly been wary of finding statements to be defamatory, unless the statements misquote the author, put words into the author's mouth or otherwise clearly go beyond the realm of interpretation.

Evans' and Novak's statements about Mr. Ollman's article clearly do not fall into the category of misquotation or misrepresentation. First, the plaintiff complains of the following statement: "Ollman concedes that [the fact that most students have a 'Marxist outlook' after taking his course] 'will be seen as an admission that the purpose of my course is to convert students to socialism.' " Tellingly, however, the quoted words are accurately reproduced from Mr. Ollman's article. See "On Teaching Marxism and Building the Movement" at 5. To be sure, the quotation has not been printed in its complete context.37 But that is neither here nor there; the quotation of remarks without the complete context in which the remarks appeared is entirely commonplace when summarizing a written work in a brief space. We are fully aware that this practice can be highly irritating when the context does not seem fully and fairly stated. The balm for the irritation, however, cannot be a libel suit, unless triers of fact are to sit in editorial judgment.38

Professor Ollman also objects to the column's posing the question, prompted in Evans' and Novak's view by Mr. Ollman's article, of whether he intended to use the classroom for indoctrination. As we noted previously, the column in no wise affirmatively stated that Mr. Ollman was indoctrinating his students. Moreover, indoctrination is not, at least as used here in the setting of academia, a word with a well-defined meaning. To paraphrase Justice Harlan in another context, see Cohen v. California, 403 U.S. 15, 25, 91 S.Ct. 1780, 1788, 29 L.Ed.2d 284 (1971), what is indoctrination to one person is merely the vigorous exposition of ideas to another. We therefore conclude that the column's statements concerning "indoctrination" constitute protected opinion.

Mr. Ollman also complains of the statement: "His candid writings avow his desire to use the classroom as an instrument for preparing what he calls the 'revolution'." This statement, unlike the column's remarks about indoctrination, is stated without any interrogatory language to allow the reader to discount it as opinion. However, it is clear from the context that the statement represents Evans' and Novak's interpretation of Mr. Ollman's writing. And, like the charge of indoctrination, this statement does not have a well-defined meaning or admit of a method of proof or disproof. What to one person is a patently improper use of the classroom for political purposes may represent to another no more than the imparting of ideas, in the faith that ideas have consequences.

Finally, we turn to the most troublesome statement in the column.* In the third-to-last paragraph, an anonymous political science professor is quoted as saying: "Ollman has no status within the profession but is a pure and simple activist." The District Court interpreted this remark as a statement that Mr. Ollman "lack[ed] a reputation in his field as scholar."39 Memorandum Opinion at 5.

Certainly a scholar's academic reputation among his peers is crucial to his or her career. Like the peripatetic philosophers of ancient Greece, modern scholars depend upon their reputation to enable them to pursue their chosen calling. We also acknowledge that at least one pre-Gertz case has held that the common-law privilege of fair comment does not extend protection to remarks which disparage one's status among one's peers. See Cepeda v. Cowles Magazines and Broadcasting, Inc., 328 F.2d 869 (9th Cir.1964) (holding that, inter alia, remarks that a baseball player had "doghouse status" with the San Francisco Giants' hierarchy was not protected by fair comment).40

We are of the view, however, that under the constitutionally based opinion privilege announced in Gertz, this quotation, under the circumstances before us, is protected. A confluence of factors leads us to this conclusion. First, as we have stated, inasmuch as the column appears on the Op-Ed page, the average reader will be influenced by the general understanding of the functions of such columns and read the remark to be opinion.41 The identical quotation in a newspaper article purporting to publish facts or in an academic publication which purported to rate status within a given discipline would, of course, be quite another matter. But here we deal with statements by well-known, nationally syndicated columnists on the Op-Ed page of a newspaper, the well-recognized home of opinion and comment. In addition, the thrust of the column, taken as a whole, is to raise questions about Mr. Ollman's scholarship and intentions, not to state conclusively from Evans' and Novak's first-hand knowledge that Professor Ollman is not a scholar or that his colleagues do not regard him as such.

Moreover, the anonymous professor's unflattering comment appears only after the columnists expressly state that Mr. Ollman is a professor at New York University, a highly respected academic institution, a fact which provides objective evidence of Mr. Ollman's "status." So too, the controversy itself was occasioned by Professor Ollman's nomination by the departmental search committee as chairman of an academic department at the University of Maryland, a fact stated in the column's opening paragraph which also plainly suggested to the average reader that Professor Ollman did in fact enjoy some considerable status in academia. Finally in this regard, the column expressly states that Professor Ollman's imminent ascension to the departmental chairmanship at Maryland was troubling only to a clear minority of academics. Thus, the charge of "no status" in this context would plainly appear to the average reader to be "rhetorical hyperbole" within the meaning of Greenbelt, and which in turn would lead the reader to treat the statement as one of opinion.42

We note especially in this respect that the anonymous academician quoted in the column goes on to say that he would not repeat his charge publicly, stating that: "[o]ur academic culture does not permit the raising of such questions." Thus, while Mr. Ollman's critic is asserting a proposition about Mr. Ollman, he is simultaneously implying that, in the contemporary academic environment, no evidence can publicly be adduced to support it. Whether right or wrong, this admission by the anonymous political scientist would clearly tend to make the reader treat this proposition as opinion.43

But most fundamentally, we are reminded that in the accommodation of the conflicting concerns reflected in the First Amendment and the law of defamation, the deep-seated constitutional values embodied in the Bill of Rights require that we not engage, without bearing clearly in mind the context before us, in a Talmudic parsing of a single sentence or two, as if we were occupied with a philosophical enterprise or linguistic analysis. Ours is a practical task, with elemental constitutional values of freedom looming large as we go about our work. And in that undertaking, we are reminded by Gertz itself of our duty "to assure to the freedoms of speech and press that 'breathing space' essential to their fruitful exercise." Gertz, supra, 418 U.S. at 342, 94 S.Ct. at 3008. For the contraction of liberty's "breathing space" can only mean inhibition of the scope of public discussion on matters of general interest and concern. The provision of breathing space counsels strongly against straining to squeeze factual content from a single sentence in a column that is otherwise clearly opinion.44 As the Ninth Circuit so succinctly put it, "[t]he court must consider all the words used, not merely a particular phrase or sentence." Information Control Corp. v. Genesis One Computer Corp., supra, 611 F.2d at 784.45

The judgment of the District Court is therefore

Affirmed.

What is in danger of becoming a frivolous public debate over the appointment of a Marxist to head the University of Maryland's department of politics and government has so far ignored this unspoken concern within the academic community: the avowed desire of many political activists to use higher education for indoctrination.

The proposal to name Bertell Ollman, professor at New York University, as department head has generated wrong-headed debate. Politicians who jumped in to oppose Ollman simply for his Marxist philosophy have received a justifiable going-over from defenders of academic freedom in the press and the university. Academic Prinve [sic ] Valiants seem arrayed against McCarythite [sic ] know-nothings.

But neither side approaches the central question: not Ollman's beliefs, but his intentions. His candid writings avow his desire to use the classroom as an instrument for preparing what he calls "the revolution." Whether this is a form of indoctrination that could transform the real function of a university and transcend limits of academic freedom is a concern to academicians who are neither McCarthyite nor know-nothing.

To protect academic freedom, that question should be posed not by politicians but by professors. But professors throughout the country troubled by the nomination, clearly a minority, dare not say a word in today's campus climate.

While Ollman is described in news accounts as a "respected Marxist scholar," he is widely viewed in his profession as a political activist. Amid the increasingly popular Marxist movement in university life, he is distinct from philosophical Marxists. Rather, he is an outspoken proponent of "political Marxism."

He twice sought election to the council of the American Political Science Association as a candidate of the "Caucus for a New Political Science" and finished last out of 16 candidates each time. Whether or not that represents a professional judgment by his colleagues, as some critics contend, the verdict clearly rejected his campaign pledge: "If elected ... I shall use every means at my disposal to promote the study of Marxism and Marxist approaches to politics throughout the profession."

Ollman's intentions become explicit in "On Teaching Marxism and Building the Movement," his article in the Winter 1978 issue of New Political Science. Most students, he claims, conclude his course with a "Marxist outlook." Ollman concedes that will be seen "as an admission that the purpose of my course is to convert students to socialism."

That bothers him not at all because "a correct understanding of Marxism (as indeed of any body of scientific truths) leads automatically to its acceptance." Non-Marxists students are defined as those "who do not yet understand Marxism." The "classroom" is a place where the students' "bourgeois ideology is being dismantled." "Our prior task" before the revolution, he writes, "is to make more revolutionaries. The revolution will only occur when there are enough of us to make it."

He concludes by stressing the importance to "the movement" of "radical professors." If approved for his new post, Ollman will have a major voice in filling a new professorship promised him. A leading prospect is fellow Marxist Alan Wolfe; he is notorious for his book "The Seamy Side of Democracy," whose celebration of communist China extols the beneficial nature of "brainwashing."

Ollman's principal scholarly work, "Alienation: Marx's Conception of Man in Capitalist Society," is a ponderous tome in adoration of the master (Marxism "is like a magnificently rich tapestry"). Published in 1971, it does not abandon hope for the revolution forecast by Karl Marx in 1848. "The present youth rebellion," he writes, by "helping to change the workers of tomorrow" will, along with other factors, make possible "a socialist revolution."

Such pamphleteering is hooted at by one political scientist in a major eastern university, whose scholarship and reputation as a liberal are well known. "Ollman has no status within the profession, but is a pure and simple activist," he said. Would he say that publicly? "No chance of it. Our academic culture does not permit the raising of such questions."

"Such questions" would include these: What is the true measurement of Ollman's scholarship? Does he intend to use the classroom for indoctrination? Will he indeed be followed by other Marxist professors? Could the department in time be closed to non-Marxists, following the tendency at several English universities?

Even if "such questions" cannot be raised by the faculty, they certainly should not be raised by politicians. While dissatisfaction with pragmatism by many liberal professors has renewed interest in the comprehensive dogma of the Marxists, there is little tolerance for confronting the value of that dogma. Here are the makings of a crisis that, to protect its integrity and true academic freedom, academia itself must resolve.

BORK, Circuit Judge, with whom WILKEY and GINSBURG, Circuit Judges, and MacKINNON, Senior Circuit Judge, join, concurring:

While I concur in the judgment of the court and in much of Judge Starr's scholarly exposition, I write separately because I do not think he has adequately demonstrated that all of the allegedly libelous statements at issue here can be immunized as expressions of opinion. The dissents, on the other hand, while acknowledging the importance of additional factors, seem actually premised on the idea that the law makes a clear distinction between opinions, which are not actionable as libel, and facts, which are. In my view, the law as enunciated by the Supreme Court imposes no such sharp dichotomy. Some lower courts have assumed, as do some members of this court, not only that this opinion vs. fact formula is controlling but that it is governed, at least primarily, by grammatical analysis. I think that incorrect. Any such rigid doctrinal framework is inadequate to resolve the sometimes contradictory claims of the libel laws and the freedom of the press.

This case illustrates that point. It arouses concern that a freshening stream of libel actions, which often seem as much designed to punish writers and publications as to recover damages for real injuries, may threaten the public and constitutional interest in free, and frequently rough, discussion. Those who step into areas of public dispute, who choose the pleasures and distractions of controversy, must be willing to bear criticism, disparagement, and even wounding assessments. Perhaps it would be better if disputation were conducted in measured phrases and calibrated assessments, and with strict avoidance of the ad hominem; better, that is, if the opinion and editorial pages of the public press were modeled on The Federalist Papers. But that is not the world in which we live, ever have lived, or are ever likely to know, and the law of the first amendment must not try to make public dispute safe and comfortable for all the participants. That would only stifle the debate. In our world, the kind of commentary that the columnists Rowland Evans and Robert Novak have engaged in here is the coin in which controversialists are commonly paid.

These reflections lead me to conclude that Professor Ollman cannot press a libel action. But I do not find it easy to reach that result through a blunt distinction between opinion and fact, which while sometimes useful in just that crude dichotomy, is not adequate to the task here.

This inadequacy is most apparent in dealing with what Judge Starr calls "the most troublesome statement in the column," that concerning Ollman's reputation. It will be well to place the statement more completely in its context. Toward the end of their column, Evans and Novak say this:

Ollman's principal scholarly work, "Alienation: Marx's Conception of Man in Capitalist Society," is a ponderous tome in adoration of the master (Marxism "is like a magnificently rich tapestry"). Published in 1971, it does not abandon hope for the revolution forecast by Karl Marx in 1848. "The present youth rebellion," he writes, by "helping to change the workers of tomorrow" will, along with other factors, make possible "a socialist revolution."

Such pamphleteering is hooted at by one political scientist in a major eastern university, whose scholarship and reputation as a liberal are well known. "Ollman has no status within the profession, but is a pure and simple activist," he said. Would he say that publicly? "No chance of it. Our academic culture does not permit the raising of such questions."

Judge Starr's opinion for the majority contends that, in the circumstances of this case and in the context of the column as a whole, the quoted statement that "Ollman has no status within the profession, but is a pure and simple activist" qualifies as an opinion and so is constitutionally protected. The dissents, on the other hand, suggest that an assertion about one's general reputation is an assertion of fact. If common usage were the test, and if we looked at the sentence standing alone, the dissent's characterization would certainly be correct. The challenged language is a statement that others hold a particular opinion. Whether or not they do is a question of fact, though, as I will try to show, it is a "fact" of a peculiar nature in the context of first amendment litigation. If placing the bare assertion in question into one of two compartments labelled "opinion" and "fact" were the only issue we were allowed to consider, I would join the dissent. But I do not think these simple categories, semantically defined, with their flat and barren descriptive nature, their utter lack of subtlety and resonance, are nearly sufficient to encompass the rich variety of factors that should go into analysis when there is a sense, which I certainly have here, that values meant to be protected by the first amendment are threatened.

The temptation to adhere to sharply-defined categories is understandable. Judges generalize, they articulate concepts, they enunciate such things as four-factor frameworks, three-pronged tests, and two-tiered analyses in an effort, laudable by and large, to bring order to a universe of unruly happenings and to give guidance for the future to themselves and to others. But it is certain that life will bring up cases whose facts simply cannot be handled by purely verbal formulas, or at least not handled with any sophistication and feeling for the underlying values at stake. When such a case appears and a court attempts nevertheless to force the old construct upon the new situation, the result is mechanical jurisprudence. Here we face such a case, and it seems to me better to revert to first principles than to employ categories which, in these circumstances, inadequately enforce the first amendment's design.

Viewed from that perspective, the statement challenged in this lawsuit, in terms of the policies of the first amendment, is functionally more like an "opinion" than a "fact" and should not be actionable. It thus falls within the category the Supreme Court calls "rhetorical hyperbole." See pp. 975-79, infra. I will try to set out the factors in this case that justify application of that concept.

Because Evans and Novak wrote that an anonymous political science professor said he had "no status" among political scientists, Ollman wants to ask a jury to award him $1,000,000 in compensatory damages and an additional $5,000,000 in punitive damages. In the field of journalism, these are enormous sums. They are quite capable of silencing political commentators forever. Unless the defamation was heinous and devastating, the amounts sought are entirely disproportionate. No one would think it appropriate for a state to levy such amounts as fines upon writers for statements of the sort made here. But, under current doctrine, lower courts have no way of saying that such sums may not be sought in libel actions, Gertz v. Robert Welch, Inc., 418 U.S. 323, 94 S.Ct. 2997, 41 L.Ed.2d 789 (1974), or, indeed, of saying that damages may not be awarded as punishment or that such components of compensation as psychological anguish are inconsistent with the first amendment when the libel occurs in a public, political dispute. Time, Inc. v. Firestone, 424 U.S. 448, 460, 96 S.Ct. 958, 968, 47 L.Ed.2d 154 (1976). Instead, unless we continue to develop doctrine to fit first amendment concerns, we are remitted to old categories which, applied woodenly, do not address modern problems.

The American press is extraordinarily free and vigorous, as it should be. It should be, not because it is free of inaccuracy, oversimplification, and bias, but because the alternative to that freedom is worse than those failings. Yet the area in which legal doctrine is currently least adequate to preserve press freedom is the area of defamation law, the area in which this action lies. We are said to have in the first amendment "a profound national commitment to the principle that debate on public issues should be uninhibited, robust, and wide-open." New York Times Co. v. Sullivan, 376 U.S. 254, 270, 84 S.Ct. 710, 721, 11 L.Ed.2d 686 (1964). That principle has resulted in the almost total abolition of prior restraints on publication; New York Times Co. v. United States, 403 U.S. 713, 91 S.Ct. 2140, 29 L.Ed.2d 822 (1971); Nebraska Press Association v. Stuart, 427 U.S. 539, 96 S.Ct. 2791, 49 L.Ed.2d 683 (1976); the curtailment of the possibility of criminal sanctions; Garrison v. Louisiana, 379 U.S. 64, 85 S.Ct. 209, 13 L.Ed.2d 125 (1964); and, in Sullivan itself, the construction of serious obstacles to private defamation actions by government officials. The cases that came afterward deployed similar obstacles to defamation actions by "public figures," Curtis Publishing Co. v. Butts, 388 U.S. 130, 87 S.Ct. 1975, 18 L.Ed.2d 1094 (1967); Rosenbloom v. Metromedia, Inc., 403 U.S. 29, 91 S.Ct. 1811, 29 L.Ed.2d 296 (1971); Gertz, 418 U.S. at 345, 94 S.Ct. at 3009. Thus, we have a judicial tradition of a continuing evolution of doctrine to serve the central purpose of the first amendment.

Judge Scalia's dissent implies that the idea of evolving constitutional doctrine should be anathema to judges who adhere to a philosophy of judicial restraint. But most doctrine is merely the judge-made superstructure that implements basic constitutional principles. There is not at issue here the question of creating new constitutional rights or principles, a question which would divide members of this court along other lines than that of the division in this case. When there is a known principle to be explicated the evolution of doctrine is inevitable. Judges given stewardship of a constitutional provision--such as the first amendment--whose core is known but whose outer reach and contours are ill-defined, face the never-ending task of discerning the meaning of the provision from one case to the next. There would be little need for judges--and certainly no office for a philosophy of judging--if the boundaries of every constitutional provision were self-evident. They are not. In a case like this, it is the task of the judge in this generation to discern how the framers' values, defined in the context of the world they knew, apply to the world we know. The world changes in which unchanging values find their application. The fourth amendment was framed by men who did not foresee electronic surveillance. But that does not make it wrong for judges to apply the central value of that amendment to electronic invasions of personal privacy. The commerce power was established by men who did not foresee the scope and intricate interdependence of today's economic activities. But that does not make it wrong for judges to forbid states the power to impose burdensome regulations on the interstate movement of trailer trucks. The first amendment's guarantee of freedom of the press was written by men who had not the remotest idea of modern forms of communication. But that does not make it wrong for a judge to find the values of the first amendment relevant to radio and television broadcasting.

So it is with defamation actions. We know very little of the precise intentions of the framers and ratifiers of the speech and press clauses of the first amendment. But we do know that they gave into our keeping the value of preserving free expression and, in particular, the preservation of political expression, which is commonly conceded to be the value at the core of those clauses. Perhaps the framers did not envision libel actions as a major threat to that freedom. I may grant that, for the sake of the point to be made. But if, over time, the libel action becomes a threat to the central meaning of the first amendment, why should not judges adapt their doctrines? Why is it different to refine and evolve doctrine here, so long as one is faithful to the basic meaning of the amendment, than it is to adapt the fourth amendment to take account of electronic surveillance, the commerce clause to adjust to interstate motor carriage, or the first amendment to encompass the electronic media? I do not believe there is a difference. To say that such matters must be left to the legislature is to say that changes in circumstances must be permitted to render constitutional guarantees meaningless. It is to say that not merely the particular rules but the entire enterprise of the Supreme Court in New York Times v. Sullivan was illegitimate.

We must never hesitate to apply old values to new circumstances, whether those circumstances are changes in technology or changes in the impact of traditional common law actions. Sullivan was an instance of the Supreme Court doing precisely this, as Brown v. Board of Education, 347 U.S. 483, 492-95, 74 S.Ct. 686, 690-92, 98 L.Ed. 843 (1954), was more generally an example of the Court applying an old principle according to a new understanding of a social situation. The important thing, the ultimate consideration, is the constitutional freedom that is given into our keeping. A judge who refuses to see new threats to an established constitutional value, and hence provides a crabbed interpretation that robs a provision of its full, fair and reasonable meaning, fails in his judicial duty. That duty, I repeat, is to ensure that the powers and freedoms the framers specified are made effective in today's circumstances. The evolution of doctrine to accomplish that end contravenes no postulate of judicial restraint. The evolution I suggest does not constitute a major change in doctrine but is, as will be shown, entirely consistent with the implications of Supreme Court precedents.

We now face a need similar to that which courts have met in the past. Sullivan, for reasons that need not detain us here, seems not to have provided in full measure the protection for the marketplace of ideas that it was designed to do. Instead, in the past few years a remarkable upsurge in libel actions, accompanied by a startling inflation of damage awards, has threatened to impose a self-censorship on the press which can as effectively inhibit debate and criticism as would overt governmental regulation that the first amendment most certainly would not permit. See Lewis, New York Times v. Sullivan Reconsidered: Time to Return to "The Central Meaning of the First Amendment," 83 Colum.L.Rev. 603 (1983).1 It is not merely the size of damage awards but an entire shift in the application of libel laws that raises problems for press freedom. See Smolla, Let the Author Beware: The Rejuvenation of the American Law of Libel, 132 U.Pa.L.Rev. 1 (1983).2 Taking such matters into account is not, as one dissent suggests, to engage in sociological jurisprudence, at least not in any improper sense. Doing what I suggest here does not require courts to take account of social conditions or practical considerations to any greater extent than the Supreme Court has routinely done in such cases as Sullivan. Nor does analysis here even approach the degree to which the Supreme Court quite properly took such matters into account in Brown, 347 U.S. at 492-95, 74 S.Ct. at 690-92. Matters such as the relaxation of legal rules about permissible recovery, the changes in tort law to favor compensation, and the existence of doctrinal confusion, see Smolla,supra, are matters that courts know well. Indeed, courts are responsible for these developments.

The only solution to the problem libel actions pose would appear to be close judicial scrutiny to ensure that cases about types of speech and writing essential to a vigorous first amendment do not reach the jury.3 See Bose Corp. v. Consumers Union of United States, Inc., --- U.S. ----, 104 S.Ct. 1949, 1965, 80 L.Ed.2d 502 (1984). This requires a consideration of the totality of the circumstances that provide the context in which the statement occurs and which determine both its meaning and the extent to which making it actionable would burden freedom of speech or press. That, it must be confessed, is a balancing test and risks admitting into the law an element of judicial subjectivity. To that objection there are various answers. A balancing test is better than no protection at all. Given the appellate process, moreover, the subjective judgment of no single judge will be controlling. Over time, as reasons are given, the element of subjectivity will be reduced. There is, in any event, at this stage of the law's evolution, no satisfactory alternative. Hard categories and sharply-defined principles are admirable, if they are available, but usually, in the world in which we live, they share the problem of absolutes, of which they are a subgenre: they do not stand up when put to the test of hard cases. In the process of "balancing," I will state my reasons fully so that it may be judged whether they are rooted adequately in central first amendment concerns and so that guidance may be given as to how I think cases should be decided in the future.

Two general considerations lead me to conclude that Professor Ollman should not be allowed to try his case to a jury. First, the state of doctrine in this area, if not precisely embryonic, is certainly still developing. Nothing in case law that is binding upon this court requires us to ignore context and the purposes of the first amendment and, instead, to apply a rigid opinion-fact dichotomy and to define the compartments of that dichotomy by semantic analysis. Indeed, the Supreme Court has indicated that we are not to do that. See pp. 975-79, infra. We are required, therefore, to continue the evolution of the law in accordance with the deepest rationale of the first amendment. Second, the central concerns of the first amendment are implicated in this case so that a damage award would have a heavily inhibiting effect upon the journalism of opinion. On the other hand, the statement challenged, in practical impact, is more like an expression of opinion than it is like an assertion of fact. It is the kind of hyperbole that must be accepted in the rough and tumble of political argument.

It is plain, I think, that the opinion-fact dichotomy is not as rigid as the various dissents suppose. There is no need to become caught up in a debate about the true nature of the allegedly libelous statement in terms of that dichotomy. The formalistic distinction between the two would be binding on us, sitting as an en banc court, only if the Supreme Court had required it. The thought that the Supreme Court has required it rests upon what I believe to be a misapprehension of dicta in Gertz v. Robert Welch, Inc., 418 U.S. 323, 94 S.Ct. 2997, 41 L.Ed.2d 789. The facts of that case are important, if only by contrast with other cases, to an understanding of still-evolving doctrine in this area. Plaintiff Gertz was a lawyer who represented the family of a youth killed by a policeman in civil litigation against the policeman. In his capacity as counsel, Gertz attended the coroner's inquest but otherwise did nothing more than press the civil suit. The defendant, which published a monthly magazine, ran an article that portrayed Gertz as "an architect of the 'frame-up' " against the police officer, implied that Gertz had a lengthy criminal record, called him a "Leninist" and a "Communist-fronter," and identified him as an official of an organization that advocated violent seizure of the government. 418 U.S. at 326, 94 S.Ct. at 3000. None of this was true. The Court introduced its discussion of the governing considerations with an observation that was not necessary to the decision:

We begin with the common ground. Under the First Amendment there is no such thing as a false idea. However pernicious an opinion may seem, we depend for its correction not on the conscience of judges and juries but on the competition of other ideas. But there is no constitutional value in false statements of fact. Neither the intentional lie nor the careless error materially advances society's interest in "uninhibited, robust, and wide-open" debate on public issues. New York Times Co. v. Sullivan, 376 U.S., at 270 [84 S.Ct. at 721]. They belong to that category of utterances which "are no essential part of any exposition of ideas, and are of such slight social value as a step to truth that any benefit that may be derived from them is clearly outweighed by the social interest in order and morality." Chaplinsky v. New Hampshire, 315 U.S. 568, 572 [62 S.Ct. 766, 769, 86 L.Ed. 1031] (1942).

Id., 418 U.S. at 339-40, 94 S.Ct. at 3007 (footnote omitted).

In Gertz, it was obvious that most of the assertions that were the subject of the action purported to be flat statements of fact. The two statements that might arguably have been statements of opinion were that Gertz was a "Leninist" and a "Communist-fronter." 418 U.S. at 326, 94 S.Ct. at 3000. The Court did not discuss their proper categorization. But as Judge Friendly said in Cianci v. New Times Publishing Co., 639 F.2d 54, 61 (2d Cir.1980), these assertions must have been "deemed sufficiently 'factual' to support an action for defamation," since the Supreme Court remanded the case for jury trial.

For this reason, it is instructive to compare the Court's treatment of an even more clearly "factual" assertion in Greenbelt Cooperative Publishing Association v. Bresler, 398 U.S. 6, 90 S.Ct. 1537, 26 L.Ed.2d 6 (1970). Plaintiff Bresler, a real estate developer and builder, engaged in negotiations with the City Council of Greenbelt, Maryland, for zoning variances so that he could build high-density housing on land he owned. Simultaneously, the city was trying to acquire another tract of land from Bresler to construct a high school. The concurrent negotiations gave each side bargaining leverage. Bresler, of course, could vary the price for the tract depending on the city's attitude toward the variances. A newspaper accurately reported the public debate at city council meetings at which Bresler's negotiating demands were denounced as "blackmail." Bresler sued, alleging that the articles imputed a crime to him. The Court held that this denunciation was a constitutionally protected statement since here the word "blackmail" was no more than "rhetorical hyperbole, a vigorous epithet used by those who considered Bresler's negotiating position extremely unreasonable." Id., 398 U.S. at 14, 90 S.Ct. at 1542. The context in which the words appeared was such that no reader could have thought that Bresler was charged with a crime.

The analytical approach of Bresler was reaffirmed in Old Dominion Branch No. 496, National Association of Letter Carriers v. Austin, 418 U.S. 264, 285-86, 94 S.Ct. 2770, 2781-82, 41 L.Ed.2d 745 (1974), a case argued and handed down on the same days as Gertz. In Letter Carriers, a union newsletter, Carrier's Corner, published the names of those, including plaintiffs, who had not joined the union under the heading "Li