Freedom of Speech and Appellate Review

in Workplace Harassment Cases

 

by Prof. Eugene Volokh, UCLA School of Law

 

 

Originally published in the Northwestern Law Review; abridged and updated.

Cite text as Eugene Volokh, Freedom of Speech and Appellate Review in Workplace Harassment Cases, 90 Nw. U. L. Rev. 1009 (1996).

 

 

 

The Facts

The Law

Your client's employee Mary is complaining about her coworker John. John has a Sports Illustrated swimsuit calendar hanging on his office wall; Mary sees it whenever she walks by.

 

John also sometimes complains in the lunchroom about "feminazis" who are ruining the nation. "The problem with this country," he says, "is that women don't stay at home, where they belong."

 

Mary thinks this constitutes sexual harassment, and asks your client to make John stop.

Workplace speech constitutes harassment if it's

 

·                    "severe or pervasive" enough to

 

·                    create a "hostile or abusive work environment"

 

·                    based on race, sex, religion, national origin, or several other categories,

 

·                    for the plaintiff and for a reasonable person.[1]

 

Your client wants your advice. What should he do? He could play it safe by just telling John to take down the calendar and keep quiet. But let's say your client sympathizes with John's speech. Or say he just doesn't like shutting his employees up. Is John's speech really unlawful harassment, he asks you? Will he really be liable if he doesn't make John stop?

 

Put the facts and the law side by side, and you don't get much of an answer. Are one sexually suggestive calendar and several sexist political comments "severe or pervasive"? Would a reasonable person find that they make the environment "hostile or abusive"? The language of the test doesn't really help.

 

Neither, unfortunately, does the case law. If a trial judge finds a hostile environment, most appellate courts will review this only for clear error.[2] If the finding is by a jury, the appellate court will ask only whether a reasonable jury could so conclude.[3] Likewise, a trial judge may grant summary judgment for a harassment defendant only if no reasonable jury could say the environment was hostile.[4]

 

Thus, instead of marking out two areas - hostile environment and nonhostile environment - courts mark out three areas:

 

1.      those environments that any reasonable factfinder would conclude are hostile,

 

2.      those that no reasonable factfinder would conclude are hostile, and

 

3.      those on which reasonable factfinders could disagree. If your case falls in the third area - a broad area indeed - all you can tell your client is that the result depends on the judge or jury he draws.

 

The basic thesis of this Article is simple: When a factfinder concludes that someone's speech has created a hostile environment, an appellate court is constitutionally bound to exercise its independent judgment on this point. The court must take an independent look at the record and determine for itself whether the environment the record discloses is "hostile." The Supreme Court's decision in Bose Corp. v. Consumers Union,[5] which requires independent judgment review in free speech cases, squarely controls.

 

And this is more than just the law - it's a good idea. Workplace harassment law[6] is a nationwide speech code. Outside the workplace, racist, sexist, and religiously bigoted statements are generally constitutionally protected. This is also true for sexually suggestive pictures, from Gauguins to Playboy centerfolds; suggestive jokes; even slurs and personal insults (so long as they aren't likely to provoke a fight).[7] But in the workplace, speech like this may lead to an injunction or a hefty damages award. (Click for more details.)

 

Though there's a hot debate over whether and to what extent this sort of speech restriction is constitutional (click) for this Article I'll set that question aside. I'll assume that, like obscenity, fighting words, and defamation, harassing speech - speech that creates a hostile environment - may constitutionally be restricted. Given this assumption, I'll argue, it's especially important that courts define the boundaries of this category as precisely as possible. And this sort of definition can happen only if courts exercise independent judgment in reviewing findings of a hostile environment.[8]

 

I'll also draw two other conclusions. First, even if harassing speech isn't constitutionally protected, this doesn't mean the Free Speech Clause is out of the picture in harassment lawsuits. So far, no‑one has discussed the extent to which "First Amendment Due Process"[9] doctrine - which can require, for instance, proof by clear and convincing evidence or a prohibition on punitive damages - applies to harassment cases. This issue deserves investigation.

 

Moreover, the lesson of all this for the defense lawyer is always take the First. Even if harassing speech is unprotected, raising the defense will give you a second chance with the appellate court. Maybe the court will independently find that the speech was harassing; but maybe it won't. If you don't raise the free speech defense, though, the appellate court won't even have a real opportunity to rule in your favor.

 

I. Why Independent Judgment Review Is Mandated

 

A.                 Freedom of Speech and Workplace Harassment

 

It's clear that harassment law restricts speech. What's not clear is which kinds of speech the law prohibits, and which kinds can still legally be said. As I document in exhaustive detail elsewhere, workplace harassment law is a speech restriction of remarkable breadth. It goes far beyond slurs, hardcore pornography, repeated vulgar sexual propositions, and the like, and can suppress, among other things,

·                    political statements,

·                    religious proselytizing,

·                    legitimate art (such as prints of Francisco de Goya paintings),

·                    sexually themed (perhaps not even misogynistic) jokes,

·                    and other kinds of speech that are generally seen as being entirely constitutionally protected.

 

No one knows, though, how broad the restriction goes - to what extent employers and employees may legally speak about religion in the workplace, make arguably bigoted (or even simply insensitive) political or social statements, say sexually suggestive jokes - even if they aren't at any particular employee's expense - or post "legitimate" art (whether it involves nudity or is merely sexually suggestive ) can lead to liability. Again, if you're skeptical that harassment law is this broad, just click here and you'll see how broad it really is.

 

I'm troubled by these speech‑restrictive results, troubled enough that I conclude harassment law is partly unconstitutional. But others disagree. Not all speech restrictions, they point out, are necessarily impermissible. When offensive speech is so severe or pervasive that it creates a hostile environment for coworkers, they argue, the speech should lose its constitutional protection.[10]

 

For purposes of this Article, I'll accept this position. I'll assume that speech which creates a hostile environment is unprotected. But the problem remains: How can employers tell whether speech creates a hostile environment? Should they bar all religious proselytizing, prohibit all arguably bigoted political statements, keep their employees from using all sex‑specific language, ban all suggestive jokes, and order that all sexually suggestive pictures be taken down? And if not, how can they know when they can say to the complaining employee, "We're sorry that you're offended by this speech, but it doesn't constitute harassment"?

 

B.                Freedom of Speech and Appellate Review

 

Fortunately, there's a well‑established solution to this problem. Harassing speech wouldn't be the first category of speech held to be unprotected by the Free Speech Clause. Fighting words, obscenity, and libel, for example, are also generally unprotected. For each of these categories, the Court has set forth rules that define the category's boundaries: For instance, defamatory statements about public figures are actionable only if they are made with "actual malice" - knowledge of their falsity or at least reckless disregard of the possibility of falsity.[11]

 

But the Court has recognized that these rules are often not self‑explanatory. "Providing triers of fact with a general description of the type of communication whose content is unworthy of protection has not, in and of itself, served sufficiently to narrow the category, nor served to eliminate the danger that decisions by triers of fact may inhibit the expression of protected ideas."[12] The content of many Free Speech Clause rules, the Court pointed out, "is not revealed simply by [the rule's] literal text."[13] Rather, the rules are "given meaning through the evolutionary process of common law adjudication."[14]

 

Because of this, the Court has held, appellate judges, "as expositors of the Constitution, must independently decide whether the evidence in the record is sufficient to cross the constitutional threshold."[15] They may not defer to the factfinder's conclusion. Instead, they must themselves review the record, "both to be sure that the speech in question actually falls within the unprotected category and to confine the perimeters of any unprotected category within acceptably narrow limits in an effort to ensure that protected expression will not be inhibited."[16]

 

This was said by the Court in Bose Corp. v. Consumers Union, a libel case, but the reasoning is equally applicable to harassment law:

 

·                    In both cases, some speech is protected and some isn't. In the public figure libel context, statements made without actual malice are protected. In the harassment context, speech that isn't harassing is protected.

 

·                    In both cases, there's a risk that the factfinder will misclassify the speech, either by finding actual malice where none exists, or by finding a hostile environment even though the speech wasn't severe or pervasive enough to create one.

 

·                    In both cases, the rule's literal text provides relatively little guidance in the absence of case‑by‑case elaboration by appellate courts.[17]

 

And the Bose Court made clear that it wasn't just announcing a libel rule: It specifically held that the same rule applies generally to judgments that a certain kind of speech is unprotected.[18]

 

Post‑Bose cases have faithfully applied Bose to alleged obscenity,[19] incitement,[20] a newspaper's negligent publication of criminal solicitation,[21] speech by lawyers that supposedly interferes with the administration of justice,[22] government employee speech,[23] speech in a possibly nonpublic forum,[24] and commercial speech.[25] The Court has reserved judgment on whether the Bose standard applies in the context of content‑neutral speech restrictions,[26] but harassment law is clearly content‑based.[27]

 

If anything, independent judgment review is especially appropriate in harassment cases. "Hostile environment" is an amorphous term. It's at least as vague as "prurient interest" and "patently offensive" (elements of the obscenity test), and in my view vaguer than "reckless disregard" (part of the libel test) and "likely to provoke the average person to retaliation" (part of the fighting words test).[28] There's no longstanding social consensus on the definition of "hostile environment" or "harassment." The terms have no intuitively obvious meaning. In fact, there's a vast amount of disagreement on what's harassment and what's not. Under the theory of Bose, this is precisely the sort of test that must be clarified through case‑by‑case appellate adjudication.[29] In fact, such clarification might be necessary to minimize constitutional vagueness problems.[30]

 

What grounds could there be for distinguishing libel, obscenity, and similar speech categories from harassing speech? Obviously the fact that harassing speech is (arguendo) unprotected can't be the distinction: False statements of fact made with actual malice, fighting words, incitement, and obscenity are unprotected as well. Independent review is required in such cases precisely to determine whether or not the speech falls within the unprotected category.

 

Nor should one's views about the importance of the state interest make a difference. Preventing incitement to imminent violence is surely an important interest, but the Bose rule applies to incitement cases also.[31] Moreover, Bose review doesn't prevent the interest from being served; it only requires that appellate courts have the final say.

 

It also can't matter that the speech in many harassment cases is supposedly "low‑value" pornography or slurs. The same is true in obscenity and fighting‑words cases, but Bose made clear that the independent judgment rule applies there, too.[32]

 

One could, in the words of my colleague Stephen Yeazell, view harassment law as part of a "continuing renegotiation of the social contract between the sexes," a process of social decisionmaking best carried on by jurors consulting their consciences on an ad hoc basis.[33] But Bose precludes such a conclusion. The holding of Bose is precisely that the definition of speech restrictions shouldn't be left - as is, for instance, the law of negligence - to the consciences of jurors.[34]

 

The only reason I can see for not applying Bose would be a judgment that speech on both sides of the line is unprotected. If even nonharassing workplace speech - speech that isn't severe or pervasive enough to create a hostile environment - were unprotected by the Constitution, the harassment/nonharassment line wouldn't have constitutional significance. If all bigoted political statements, religious proselytizing, suggestive jokes, or art containing nudity were unprotected in the workplace, it wouldn't matter for constitutional purposes whether a particular set of statements was "severe" or "pervasive" enough to create a "hostile" environment. The finding of a hostile environment might still be wrong, but it would be wrong only on statutory grounds.[35]

 

But surely this can't be correct. NLRB v. Gissel Packing Co.[36] made clear that workplace speech is generally protected by the Free Speech Clause (at least so long as the speech doesn't contain a "threat of reprisal or promise of benefit" by the employer or a union). (Click here for more details.) And this must be so: If, for instance, Congress tried to bar private employees from criticizing the war effort, or from saying things critical of returning soldiers, surely the law would be unconstitutional even if it applied only to workplace speech.[37]

 

Nor can the fact that employees in the workplace are arguably "captive" - unable to easily escape the offensive speech - lead to the conclusion that workplace speech is unprotected. The Court has never held that the mere presence of a captive audience justifies speech restrictions. People often can't avoid offensive speech. Courthouse employees couldn't easily avoid Cohen's "Fuck the Draft" jacket.[38] Strikebreakers are captive to picketers who march around with signs saying "Scab!"; they see the offensive speech every morning and every evening, and even more often if they have to come and go during the day, or if their desks face the street.

 

While the scope of the captive audience doctrine isn't particularly clear, I'm confident the Court wouldn't conclude that all workplace speech - or even all bigoted, proselytizing, offensive, or sexually suggestive workplace speech - is constitutionally unprotected. Certainly the Court has never used the captive audience doctrine as support for anything nearly this broad. And in light of Gissel, it seems clear that workplace speech is generally protected despite the presence of an arguably captive audience.

 

What makes harassment law constitutional under the Free Speech Clause, if anything does, is that the harassing speech is more than just offensive or ugly. When speech is so offensive, so pervasive that it creates a hostile environment, the government might arguably have a justification for restricting the speakers. But so long as the speech doesn't rise to this level - so long as it doesn't actually constitute harassment - there's no reason why the speech, which is clearly protected outside the workplace, should become unprotected.[39] And if nonharassing speech is constitutionally protected, the Bose principle demands that the line between harassing speech and nonharassing speech be reviewed independently by appellate courts.

 

II. How Independent Judgment Review Would Work

 

A. The Basics

 

Say a jury finds that the display of a few Gauguin nudes - recall the Naked Maja incident described above - constitutes sexual harassment, or perhaps even religious harassment.[40] I imagine that quite a few jurors don't have much more sympathy for sexually suggestive "serious art" than they do for Hustler centerfolds. And say the court of appeals, applying independent judgment, disagrees with the jury, and concludes that the speech was not harassment.[41]

 

Employers throughout the jurisdiction will then know that "artistic" nude pictures are safe, at least when there are only a few, and when they aren't combined with other speech or conduct. If in the future an employee complains about a similar display, the employer will know that it needn't fear liability.

 

Likewise, say a judge finds that religious proselytizing - for instance, religious articles in the company newsletter plus signs that say "There's no salvation except through Jesus" - has created a hostile environment. The court of appeals will then reconsider the issue. If it agrees with the trial judge, employers will know what's forbidden. On the other hand, if the appellate court concludes there's no harassment, that'll be binding precedent, and employers will know they can freely say similar things.

 

Granted, no two fact patterns are identical, and it's not even always possible to determine whether the speech in one situation is more or less severe than in another one. But the data points add up. With each new binding decision the rule becomes more explicit. That, at least, is the theory the Court adopted in Bose.

 

Independent judgment review needn't - and can't - be a reexamination of all the factual findings involved in the lower court's decision. The appellate court may, for instance, defer to the factfinder's judgments about witness credibility.[42] In jury cases, the court will generally have to assume that the jurors believed the winning side's factual claims. For instance, if what was said is in dispute - plaintiff asserts that a coworker said something offensive, but the coworker denies it - and plaintiff wins, the court will have to assume that plaintiff's story is correct.

 

But even if this is done, the question will remain: Did all the facts, even if viewed in the light most favorable to the winner, create an environment that a reasonable person would find hostile? This is a decision the appellate court can make at least as well as a jury or a trial judge.

 

Some skeptics suggest that in practice any standard of review doesn't matter much, and that judges will manipulate the standard to reach the result they want. I disagree. I'm sure such manipulation sometimes happens, but in my experience courts generally do take the standard of review seriously. Courts certainly say that standards of review matter, and I think it's fair to assume that standards of review can sometimes, even if not always, make a difference.[43]

 

Without independent judgment review, individual cases will give little guidance about what's allowed and what's forbidden.[44] Precedents would only be set when an appellate court concludes that no reasonable factfinder could find liability.[45] By definition, this will happen only in the most extreme cases - cases where a court could find that a jury would have to be, in one judge's words, "drunk or crazy" to conclude that the speech created a hostile environment.[46] That's a tough standard to meet.

 

In my view, the Bose Court, quoting Professor Leon Green, got it right:

 

[T]he judge has [a] distinct function . . . which though not frequently called into play, is of the utmost importance. It involves the determination of the scope of the general formula, or some one of its elements. . . . It requires the judge to say what sort of conduct can be considered as condemned under the rules . . . . It is the function through which the formulas and rules themselves were evolved, through which their integrity is maintained and their availability determined.[47]

 

B. The Symmetry Question

 

Bose did leave a significant question unresolved: Is independent judgment review proper if the defendant, who's making the free speech claim, wins at trial? The lower courts are split on this.

 

Some courts stress that independent judgment review is meant to let appellate courts develop and refine the constitutional rules. This development would happen regardless of which party won below, so under this approach independent judgment review should be applied symmetrically.[48]

 

But other courts stress a different basis for the Bose holding: that appellate review is necessary to decrease the chances that constitutionally protected speech would be erroneously punished. Under this view, when the free speech claimant wins below, there's no risk that the factfinder has erroneously abridged a constitutional right - at worst, it has erroneously failed to vindicate a statutory right.[49] Moreover, these courts say, courts can't adopt independent judgment review just for prudential reasons. Under Federal Rule of Civil Procedure 52(a), appellate courts must review factual findings for clear error, unless the Constitution commands otherwise.[50] The same is true for jury trials, under the Seventh Amendment.[51]

 

In my view, independent judgment review of the hostile environment question is generally both valuable and permissible under Rule 52(a) and the Seventh Amendment, whether or not the free speech claimant lost at trial. In either situation, appellate independent judgment should generally produce more refinement of the hostile environment standard. And, as I've argued above, judicial elaboration of the standard is very much needed.

 

Moreover, a symmetric rule is fairer to plaintiffs. Harassment plaintiffs' claims aren't of constitutional magnitude, but they're certainly important. I see no policy reason to treat these plaintiffs worse than defendants, assuming Bose's requirements are satisfied.

 

For review of bench trials, the symmetric approach can be used whether or not one concludes that Bose requires it as a constitutional matter. The hostile environment question is a question of application of law to fact - also known as a mixed question of law and fact - and most circuits hold that such questions may be reviewed de novo without running afoul of Rule 52(a).[52] The Ninth Circuit and, on one occasion, the Fifth Circuit have in fact treated hostile environment findings this way.[53] And the rationale for reviewing mixed questions de novo - that questions which involve "strik[ing] a balance between two sometimes conflicting societal values" and which are therefore "of clear precedential importance"[54] should be decided by appellate courts - is eminently applicable to the hostile environment context.

 

On the other hand, I don't think that independent judgment can work, regardless of how one reads Bose, when the court is reviewing a jury's general verdict for a defendant. Harassment claims involve many subsidiary factual inquiries. The jury must find both that the environment was objectively hostile, and that the plaintiff perceived the environment as hostile.[55] In most cases, there's also some dispute over what was said, over whether the employer knew or should have known about the misconduct, over what steps the employer took to try to remedy it, and so on.

 

When the defendant wins, it might be because (1) the jury concluded the environment wasn't objectively hostile, (2) the jury believed the defendant's factual story, or (3) the jury didn't think the plaintiff subjectively found the environment to be hostile. The court of appeals has no way of knowing the real reason. Even if it thinks the environment was objectively hostile, it can't reverse, because the jury might have held for defendant on one of the other grounds.

 

Review of special verdicts, however, is possible. In particular, the Seventh Amendment, which bars only reexamination of "fact[s] tried by a jury,"[56] probably doesn't prevent appellate courts from reviewing the conclusion that no hostile environment was created, which is a question of the application of law to fact.[57]

 

III. What Independent Judgment Review Means for Lawyers

 

Defense lawyers in workplace harassment cases almost never raise a Free Speech Clause defense.[58] This has always been a mystery to me, because raising the defense seems like a fairly low‑cost proposition, and because I think the defense will sometimes be a winner.[59]

 

Perhaps, though, I'm wrong on the latter point. Perhaps lawyers correctly predict that courts will be unwilling to extend constitutional protection to speech that causes a hostile environment. But even if the Free Speech Clause isn't substantively relevant in harassment cases, it is procedurally relevant. Defense lawyers in harassment cases involving speech must begin to raise a free speech defense because such a defense will trigger independent judgment review, and thus give the defendant a second chance on appeal.[60]

 

If no free speech defense is raised at trial and the employer loses, the appellate court might find that any right to independent judgment review was waived and might therefore review the harassment finding with great deference. But if the defendant had made a free speech defense - even if the defense was just "there's no hostile environment, so imposing liability would violate the Free Speech Clause"[61] - the court of appeals can draw its own conclusions about whether the speech actually created a hostile environment. This obviously won't guarantee a victory, but it will give the defendant another shot.

 

Conclusion

 

To summarize:

 

1.      When a factfinder concludes that speech has created a hostile environment, appellate courts must independently review this conclusion. Though most circuits don't do this now (and no circuit does this for jury trials), it's constitutionally required.[62]

 

2.      Such review might also be required if a trial judge, or a jury that returned a special verdict, concludes that a hostile environment has not been created.

 

3.      The only way to avoid the conclusion in point 1 is to argue that even nonharassing workplace speech is constitutionally unprotected, a position that's hard to defend.

 

4.      Lawyers must start making free speech defenses in harassment cases, even if they think that as a substantive matter harassing speech will probably be held to be unprotected.

 

This thesis is simple, and so is its proof. Still, I've found no articles or cases that even allude to it. And though it would be in many lawyers' interest to make this argument, they aren't making it.

 

This thesis also raises one other, broader point: even if harassing speech is constitutionally unprotected, the Free Speech Clause still matters in harassment cases. So far, the judicial and scholarly debate has been about the substantive rule: is harassment law constitutional? But free speech jurisprudence is also full of procedural rules.[63] For instance, sometimes the question whether speech is unprotected - for instance, whether it's libelous and possibly whether it's obscene - must be decided by clear and convincing evidence.[64] Likewise, in at least some situations punitive damages are barred even when compensatory damages are allowed.[65] Should these rules be applied to workplace harassment cases, or at least to some workplace harassment cases? To my knowledge, no one has ever addressed this question.



[1]See Harris v. Forklift Sys., Inc., 114 S. Ct. 367, 370 (1993). The other categories include at least age and disability, Eggleston v. South Bend Community Sch. Corp., 858 F. Supp. 841, 847-48 (N.D. Ind. 1994), and, in some states, sexual orientation, e.g., Leibert v. Transworld Sys., Inc., 39 Cal. Rptr. 2d 65, 67 (Cal. Ct. App. 1995).

I discuss here only hostile environment harassment; I don't purport to deal with quid pro quo sexual harassment, in which a supervisor demands sex in exchange for favorable treatment.

[2]See, e.g., Craig v. Y & Y Snacks, Inc., 721 F.2d 77, 79 (3d Cir. 1983); Spicer v. Virginia, 44 F.3d 218, 224, rev'd on other grounds, 66 F.3d 705 (4th Cir. 1995); Cortes v. Maxus Exploration Co., 977 F.2d 195, 198 (5th Cir. 1992); Rodgers v. Western-Southern Life Ins. Co., 12 F.3d 668, 674 (7th Cir. 1993); Ways v. City of Lincoln, 871 F.2d 750, 754 (8th Cir. 1989); Sauers v. Salt Lake County, 1 F.3d 1122, 1126 (10th Cir. 1993); EEOC v. Beverage Canners, Inc., 897 F.2d 1067, 1070 (11th Cir. 1990). But see Collins v. Baptist Memorial Geriatric Ctr., 937 F.2d 190, 195 (5th Cir. 1991) (de novo review, intra-circuit conflict with Cortes); EEOC v. Hacienda Hotel, 881 F.2d 1504, 1514 (9th Cir. 1989) (de novo review).

[3]See, e.g., MacArthur v. University of Tex. Health Ctr., 45 F.3d 890, 896 (5th Cir. 1995); Baskerville v. Culligan Int'l Co., 50 F.3d 428, 431 (7th Cir. 1995); Ingram v. Acands, Inc., 977 F.2d 1332, 1340 (9th Cir. 1992); Quick v. Peoples Bank, 993 F.2d 793, 797 (11th Cir. 1993); Therma-Tru Corp. v. Peachtree Doors Inc., 44 F.3d 988, 991 (Fed. Cir. 1995); Meyers v. Chapman Printing Co., 840 S.W.2d 814, 822-23 (Ky. 1992).

[4]See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248, 250-51, 252 (1986).

[5]466 U.S. 485 (1984).

[6]Most harassment cases are brought under Title VII, 42 U.S.C. § 2000e (1988); some are also brought under 42 U.S.C. § 1981 (1988), 42 U.S.C. § 1983 (1988), the Age Discrimination in Employment Act, 29 U.S.C. §§ 621-634 (1988), the Americans With Disabilities Act, 42 U.S.C. §§ 12101-12213 (1994), and various state antidiscrimination laws. I use "harassment law" to refer to the harassment prohibitions embodied in all these statutes; the standards for liability under all of them are similar. (Click here for more details.)

[7]See Gooding v. Wilson, 405 U.S. 518 (1972).

[8]Of course, if at least some harassing speech is substantively protected, my main thesis - that appellate courts must apply independent judgment in reviewing findings of harassment - remains valid a fortiori.

[9]See Henry P. Monaghan, First Amendment "Due Process," 83 Harv. L. Rev. 518 (1970).

[10]See, e.g., Robinson v. Jacksonville Shipyards, Inc., 760 F. Supp. 1486, 1535 (M.D. Fla. 1991); click here for references to commentators.

Some courts and commentators have suggested that harassment law doesn't even implicate the Free Speech Clause because no state action is involved. If speech is restricted, they argue, it's restricted by the private employer (albeit in response to the risk of liability), and not by the government. For a rebuttal to this, click here.

[11]Curtis Butts Publishing Co. v. Associated Press, 388 U.S. 130, 155 (1967); New York Times Co. v. Sullivan, 367 U.S. 254, 279-80 (1964).

[12]Bose Corp. v. Consumers Union, 466 U.S. 485, 505 (1984). Professor Henry Monaghan describes this quote and the text quoted infra at note 40 as the "core of the [Bose] opinion." Henry P. Monaghan, Constitutional Fact Review, 85 Colum. L. Rev. 229, 243 (1985).

[13]Bose, 466 U.S. at 502.

[14]Id.

[15]Id. at 511.

[16]Id. at 505; see also Harte-Hanks Communications, Inc. v. Connaughton, 491 U.S. 657, 685-86 (1989) (following Bose).

[17]Bose, 466 U.S. at 503.

[18]Id. at 504-08 (citing Supreme Court cases that apply this rule to fighting words, incitement, obscenity, and child pornography); see also Hurley v. Irish-American Gay, Lesbian & Bisexual Group, 115 S. Ct. 2338, 2344 (1995) (applying Bose to the question whether conduct was expressive); Miller v. Fenton, 474 U.S. 104, 114 (1985) (applying Bose by analogy in the Due Process Clause context to the question whether confession was voluntary); Murphy v. I.S.K.Con., 571 N.E.2d 340, 345 (Mass. 1991) (reading Bose as applicable to Free Exercise Clause issues).

Likewise, in Thompson v. Keohane, 116 S. Ct. 457 (1995), the Court held that independent review was appropriate when determining on habeas corpus whether a defendant was in custody for Miranda purposes. The Court based its decision partly on the grounds that "[t]he law declaration aspect of independent review potentially may guide [government decisionmakers], unify precedent, and stabilize the law." Id. at 467 (citing Henry P. Monaghan, Constitutional Fact Review, 85 Colum. L. Rev. 229, 243, 273-76 (1985), for the proposition that "norm elaboration occurs best when the Court has power to consider fully a series of closely related situations").

[19]E.g., Luke Records, Inc. v. Navarro, 960 F.2d 134, 138 (11th Cir. 1992).

[20]E.g., Herceg v. Hustler Magazine, Inc., 814 F.2d 1017, 1021 (5th Cir. 1987); Yakubowicz v. Paramount Pictures Corp., 536 N.E.2d 1067, 1071 (Mass. 1989).

[21]Braun v. Soldier of Fortune Magazine, Inc., 968 F.2d 1110, 1120-21 (11th Cir. 1992).

[22]Standing Comm. on Discipline v. Yagman, 55 F.3d 1430, 1443 (9th Cir. 1995).

[23]E.g., Swineford v. Snyder County, 15 F.3d 1258, 1265 (3d Cir. 1994); Mekss v. Wyoming Girls' Sch., 813 P.2d 185, 194 (Wyo. 1991); see also Rankin v. McPherson, 483 U.S. 378, 385-87 nn.8-9 (1987) (arguably dictum).

[24]E.g., AIDS Action Comm. v. Massachusetts Bay Transp. Auth., 42 F.3d 1, 6-7 (1st Cir. 1994); see also Brown v. Palmer, 915 F.2d 1435, 1441 (10th Cir. 1990) (independently reviewing factual findings underlying the determination of whether a forum is public), aff'd, 944 F.2d 732 (10th Cir. 1991) (en banc).

[25]E.g., Peel v. Attorney Registration and Disciplinary Comm'n, 496 U.S. 91, 108 (1990) (plurality); id. at 111-17 (Marshall, J., concurring) (engaging in independent review, but not citing Bose directly); Joe Conte Toyota, Inc. v. Louisiana Motor Vehicle Comm'n, 24 F.3d 754, 755-56 (5th Cir. 1994); Don's Porta Signs, Inc. v. City of Clearwater, 829 F.2d 1051, 1053-54 & n.9 (11th Cir. 1987).

Some cases reviewing federal administrative agency findings don't appear to follow Bose. There are two such cases that involve review of FTC findings that ads are false or misleading, and a line of cases (including a Supreme Court case) that involve review of NLRB findings that unionization-related speech by an employer or a union was impermissibly coercive. Both lines are grounded on a deference-to-expert-agencies rationale.

The FTC cases are Kraft, Inc. v. FTC, 970 F.2d 311, 316-17 (7th Cir. 1992), and FTC v. Brown & Williamson Tobacco Corp., 778 F.2d 35, 41 n.3 (D.C. Cir. 1985). (Both cases also argued that Bose was inapplicable to commercial speech, but that seems to be in considerable tension with the Supreme Court's position in Peel, as well as the circuit decisions in Joe Conte Toyota and Don's Porta Signs.) cf. Martin H. Redish, Product Health Claims and the First Amendment, 43 Vand. L. Rev. 1433, 1459-60 & n.144 (1990) (criticizing Brown & Williamson on Bose grounds). The NLRB cases follow NLRB v. Gissel Packing Co., 395 U.S. 575, 620 (1969), which held that "a reviewing court must recognize the Board's competence in the first instance to judge the impact of utterances made in the context of the employer-employee relationship." Since Gissel, lower courts have not applied independent judgment in this area but have instead reviewed NLRB findings for "substantial evidence." See, e.g., DTR Indus., Inc. v. NLRB, 39 F.3d 106, 114 (6th Cir. 1994).

Gissel came long before Bose, and no court has confronted the tension between them, though distinguished commentators have pointed to the discrepancy. Richard H. Fallon, Jr., Of Legislative Courts, Administrative Agencies, and Article III, 101 Harv. L. Rev. 915, 976, 990 (1988); Henry P. Monaghan, Constitutional Fact Review, 85 Colum. L. Rev. 229, 244 & n.84, 258 (1985). In any event, to my knowledge no court or commentator has suggested that substantial evidence review be transplanted from the expert agency setting to the review of findings made by judges and juries, where Bose is firmly entrenched.

The only non-agency case I could find that declined to follow