Procedural Issues in Workplace Harassment Law


The First Amendment does more than just guarantee substantive protection for speech; it also guarantees certain kinds of procedures aimed at making sure that protected speech remains practically secure.  Thus, for instance, the First Amendment:


1.      Prohibits judgments that are based even in part on protected speech — for instance, punishment based on a combination of offensive political statements and unprotected physical conduct.


2.      Allows employers to assert the free speech rights of their employees, even when the employees are not parties to the lawsuit.


1.      Demands independent appellate review in cases raising free speech questions (even when a substantively valid speech restriction — such as a ban on obscenity or fighting words — is involved).


2.      Prohibits laws that are unconstitutionally vague or overbroad.


3.      Prohibits prior restraints, such as overbroad injunctions.


4.      Restricts the availability of punitive damages.


5.      Prohibits strict liability for speech.


6.      Sometimes demands that certain facts be proven by clear and convincing evidence, rather than under the less demanding “preponderance of the evidence” standard.


1.                  Judgments Based in Part on Protected Speech


Many harassment lawsuits are based on a combination of core protected speech and unprotected conduct or unprotected speech (such as offensive touching, threats, or even one-to-one insults).  Consider, for instance, Makhayesh v. Great Lakes Steel, in which a Muslim employee of Syrian descent sued for national origin and religious harassment.  Part of the alleged harassment was direct, personal insults, but part was coworkers generally referring to Muslim religious leaders as “toilet seat[s]” and suggesting, in the context of the Gulf War, that the United States “nuke Iraq and Syria” and “go back [to Libya] and wipe them off the face of the earth.”[1]


The black-letter free speech law on this is quite clear:  A judgment cannot be based even in part on constitutionally protected speech.  If the plaintiff wants to sue based on the unprotected material, that’s just fine; the factfinder should then be allowed to consider only that material.  But a plaintiff cannot argue for a judgment based both on the unprotected matter and the constitutionally protected speech.


This is exactly the issue that the Court faced in NAACP v. Claiborne Hardware.[2]  The NAACP and other defendants organized a boycott of white-owned businesses in Claiborne County, demanding various concessions.  The businesses sued, on a tort theory of intentional interference with prospective business advantage, pointing to various behavior that interfered with their business — political statements exhorting solidarity, threats of social ostracism of boycott violators, threats of violence, and even actual violence.  The state courts held for the businesses, concluding that put together the behavior was indeed tortious.


The Court reversed.  Violence, the Court held, was indeed constitutionally unprotected; but much of the other behavior — including threats of social ostracism — was constitutionally protected speech.  And because “the nonviolent elements . . . are entitled to the protection of the First Amendment,” they did not lose this protection just because of the presence of unprotected behavior.  While the plaintiffs were entitled to sue based only on the unprotected behavior, they couldn’t sue based on a combination of the unprotected behavior and the protected speech.


The Court held the same in Street v. New York.[3]  Street was prosecuted for treating the flag disrespectfully; the prosecution’s theory rested both on Street’s flagburning and on certain things he said that expressed contempt for the flag.  In 1969, when the case was decided, the Court had not yet held that flagburning was constitutionally protected, but the Court concluded that it was unnecessary to pass on this here.  Even if flagburning was constitutionally unprotected conduct, the conviction couldn’t stand, because “[the] record [was] insufficient to eliminate the possibility . . . that appellant was convicted for both his words and his deed.”  Punishing someone based on a combination of constitutionally protected words and deeds would violate the First Amendment regardless of whether the deeds themselves would be protected.[4]  Eaton v. City of Tulsa[5] and Bachellar v. Maryland[6] take the same approach.


Likewise, in Mt. Healthy City School District Board of Education v. Doyle, the Court made clear that the First Amendment is violated whenever protected speech makes a difference in deciding a speaker’s fate.  If protected speech is a “motivating factor,” the Court held, in a government decision — even when other factors are also present — the decision is unconstitutional unless the government shows that it would have reached the same decision without considering the speech.[7]


This well-established law makes perfect sense, both theoretically and practically.


Theoretically, assume that someone is sued for a combination of constitutionally protected speech and unprotected conduct.  The plaintiff’s argument is that the speech and the conduct supposedly create a “hostile environment” or “interfere with prospective business advantage” or “intentionally inflict emotional distress” or some such.


Either the lawsuit would succeed based on the conduct alone, or it wouldn’t.  If it would succeed, then there’s no reason why the words need to be considered:  Might as well sue just based on the conduct.  But if it wouldn’t succeed without the words, then it’s the constitutionally protected speech that makes the difference between the defendant’s actions being lawful and unlawful.  If the defendant had engaged only in the unprotected conduct, he would have been acting lawfully, since by hypothesis the conduct alone wasn’t sufficient to support the plaintiff’s claim.  But saying the constitutionally protected statements on top of the conduct is what makes the behavior illegal.  The law is punishing the constitutionally protected speech, since it wouldn’t have punished the conduct standing alone.  That’s a First Amendment violation.


Practically, imagine how a cautious employer would react to a decision imposing liability in a harassment case like the one with which I began this discussion.  The employer can’t just say to its employees “It’s fine for you to make offensive political statements about Iraq, Syria, Libya, and Muslim religious leaders, unless some other people are also mistreating the offended worker in other ways (about which you, the employee, might not even know).”   So long as courts say that constitutionally protected speech can contribute to a hostile environment, the cautious employer would be wise to restrict it:  After all, it’s this constitutionally protected speech that might make the difference between a legally permissible, nonhostile environment, and an illegal hostile environment.  So — just as the Court has recognized — imposing liability based on a combination of constitutionally protected speech and unprotected conduct would unconstitutionally chill the speech.


2.                  Employers Asserting the Free Speech Rights of their Employees


Some harassment cases involve employers being punished for their own speech, or employees being enjoined from saying things.  Here, it’s clear that the defendant can raise his own First Amendment rights as a defense.


But say that a court is punishing the employer for its employees’ speech; the danger is such punishment will pressure the employer into censoring the speech of people — the workers — who are not represented in court.  Can the employer assert the workers’ rights as a defense against liability?


Definitely.  The government, by pressuring the employer with fear of liability, is restricting the speech of the employees.  (To see why government action is present, click here.)  This is a classic example of where “third-party standing” — the notion that a party can assert the constitutional rights of someone who isn’t present before the court — is appropriate.  In First Amendment cases, “litigants . . . are permitted to challenge a statute not because their own rights of free expression are violated, but because of a judicial prediction or assumption that the statute’s very existence may cause others not before the court to refrain from constitutionally protected speech or expression.”[8]


Thus, for instance, in Secretary of State of Maryland v. Joseph H. Munson Co., the Court held that a professional fundraiser could assert its clients’ rights in challenging a law that limited charitable fundraising:  Though the fundraiser “[was] not a charity and [did] not claim that its own First Amendment rights have been or will be infringed by the challenged statute,” the Court allowed the fundraiser to (successfully) argue that the law burdened the speech of its charity clients.[9]  Similarly, even if harassment law doesn’t restrict the employer’s own speech, the employer must be able to argue that harassment law burdens the speech of its employees.


Likewise, in Eisenstadt v. Baird, the Court held that contraceptive distributors could challenge a state anti-contraceptive law as violating the rights of contraceptive users.[10]  As with harassment law, such a contraceptive law restricts people’s rights indirectly, through an intermediary:  The law doesn’t directly ban contraceptive use or employee speech, but rather pressures distributors into not distributing contraceptives and pressuring employers into not tolerating certain kinds of speech.  But the court saw through that in Eisenstadt, as courts should with harassment law — because the law in ultimate effect restricts the behavior of employees, the intermediary that is charged with enforcing the law (the employer) has standing to assert the employees’ rights.


Finally, in City of Madison, Joint School District No. 8 v. Wisconsin Employment Relations Commission, a Wisconsin state agency ordered the Madison Board of Education to “cease and desist from permitting [certain classes of] employees” from saying certain things at Board of Education meetings.  The agency did not directly restrict the speech of the Board’s employees, but only tried to pressure the Board into restricting this speech.  Nonetheless, the Court concluded that even if the board had no independent First Amendment rights of its own implicated in the case, “the board may assert [the employees’ free speech] rights on behalf of [the employees].”[11]


3.                  Independent Appellate Review


Even if a court holds that harassment law is substantively constitutional — that any speech that creates a hostile work environment may indeed be punished — there’s still a great advantage to raising a First Amendment defense:  It provides an opportunity for independent appellate review.


Under the doctrine of Bose Corp. v. Consumers Union (a 1983 Supreme Court case), if the First Amendment is implicated at trial, then appellate courts must independently decide whether the offensive speech is indeed harassing.  Instead of deferring to the lower court’s finding that the speech was indeed “severe or pervasive” enough to create a “hostile or offensive” work environment — the usual rule — the appellate court would have to decide the matter for itself, thus giving the defendant a second chance.  For more on this point, click here.


4.                  Vagueness and Overbreadth


Harassment law can also be challenged on the grounds of vagueness and overbreadth; see Professor Kingsley Browne’s arguments regarding this in his Title VII as Censorship: Hostile-Environment Harassment and the First Amendment, 52 Ohio St. L.J. 481 (1991).


5.                  Prior Restraints and Overbroad Injunctions


Let’s assume, for argument’s sake, that a court rejects the substantive First Amendment arguments discussed on this site, and concludes that there’s a First Amendment exception for harassing speech:  If speech is severe or pervasive enough to create a hostile work environment, then it’s outside the First Amendment’s protections.


Even if this were true, this would limit the court to punishing and enjoining speech that’s actually harassing.  It wouldn’t allow the court to enjoin isolated bigoted statements:  Such statements aren’t by themselves severe or pervasive enough to create a hostile environment, and are therefore constitutionally protected.


Of course, a court might find it useful to enjoin even the protected bigoted statements, in order to better prevent future harassment.  This is probably why in harassment cases, courts often issue injunctions that bar even individual offensive statements:  For instance, injunctions that forbid all “racial, religious, ethnic, or other remarks or slurs contrary to their fellow employees’ religious beliefs” or all “racial, ethnic, or religious slurs whether in the form of ‘jokes,’ ‘jests,’ or otherwise” or “any and all offensive conduct and speech implicating considerations of race.”  (Click for details.)


But the prior restraint doctrine prevents such a solution, no matter how effective it might be.  It’s black-letter law that such overly broad injunctions, which sweep protected speech together with unprotected speech, are classic unconstitutional prior restraints.[12]  Even if harassing speech is unprotected, any judicial injunctions against harassment must be limited to speech that actually creates a hostile environment, not merely to isolated statements that may not be harassment at all.


A recent California Court of Appeal decision, Aguilar v. Avis Rent-A-Car System, Inc.,[13] disagrees with this analysis, but the California Supreme Court has agreed to rehear the case, and should render a decision some time in late 1997 or early 1998.


6.                  Punitive Damages


Punitive damages in free speech cases are especially troublesome, because they are especially likely to deter protected speech as well as unprotected speech.  The Court has, for instance, held that in some kinds of libel cases punitive damages are forbidden even when compensatory damages are allowed.[14]  The Court has reserved judgment on whether the same might be true for invasion of privacy cases.[15]  The Eleventh Circuit has suggested that punitive damages would be unconstitutional in cases involving negligent publication of criminal solicitations.[16]


Defense lawyers ought to argue that punitive damages should likewise be unavailable in workplace harassment cases, even if compensatory damages are available.  Neither I nor anyone else has written in detail about this subject; still, my sense is that many judges might find this sort of claim persuasive.


7.                  Strict Liability


The Free Speech Clause generally bans strict liability based on speech.[17]  Strict liability, the Court has held, unduly chills speakers and distributors, and the same may be true in harassment cases:  Imposing strict liability on employers for speech by their employees might lead them to suppress employee speech even more aggressively than they otherwise would.


Despite this, many state laws and the federal harassment regulations generally impose strict liability in cases where the hostile environment is created by a supervisor.[18]  Again, I’ve seen no detailed discussion of this issue, but defense lawyers should at least raise it.


8.                  Clear and Convincing Evidence


In some free speech contexts, courts have required that liability be proven not by a preponderance of the evidence, but by clear and convincing evidence.  The U.S. Supreme Court has so held for libel cases,[19] and the California Court of Appeal has so held for obscenity cases.[20]


The theory is simple:  Even if a certain kind of speech — such as libel or obscenity — is unprotected, a “preponderance of the evidence” standard makes it too likely that even protected speech will erroneously be found to fall into the unprotected category.  The “clear and convincing evidence” standard tries to minimize the chances of such an unconstitutional result.


The same argument can be applied to harassment.  Even if harassment is substantively unprotected, the argument would go, it’s impermissible for speech to be punished because it’s found to have been harassing under a “preponderance of the evidence” standard.  The speech can be restricted only if the plaintiff clearly and convincingly proves that the speech was indeed severe or pervasive enough to create a hostile environment.


As with the previous two points, I know of no article or case that fully explores this argument.  Still, there is enough precedent for this argument in other free speech contexts that defense lawyers ought to raise it.

[1]No. 91-108394-CZ (Mich. Ct. App. Apr. 10, 1995) (per curiam) (unpublished opinion).

[2]458 U.S. 886, 915 (1982).

[3]394 U.S. 576, 590 (1969).

[4]Of course, in all lawsuits about conduct, words can be introduced as evidence of some element of the offense — for instance, of the defendant’s intention.  Wisconsin v. Mitchell, 508 U.S. 476 (1993).  Even clearly protected speech, such as praise of Nazism, could be introduced as evidence when the question is, for instance, whether a defendant in a treason case acted with the intention to help the Nazis.  Haupt v. United States, 330 U.S. 631 (1947).  Likewise, in a Claiborne Hardware-like case, a person’s speech could be used as evidence in deciding whether he was indeed the perpetrator of a specific violent assault, or whether his acts were intentional or not.  But this does not mean that the words can themselves be made part of the offense, rather than introduced as evidence of some other conduct.  Perhaps this is a thin line in some cases, but a line that the Court has always respected, and that the Court must respect.

[5]415 U.S. 698, 699 n.* (1974) (holding that a conviction must be reversed even when it was based either “solely” or “in part” on protected speech).

[6]397 U.S. 564, 569-70 (1970) (holding that a conviction must be reversed when “the jury could have rested its verdict on any of a number of grounds,” and one of the possible grounds consisted of protected speech); id. at 671 (“since petitioners’ convictions may have rested on an unconstitutional ground, they must be set aside”).

[7]429 U.S. 274, 287 (1976).  See also Village of Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252, 270 n.21 (1976), which the Court decided the same day, which cited and was cited by Mt. Healthy, and which dealt with the same question in the Equal Protection Clause context.

[8]Secretary of State of Maryland v. Joseph H. Munson Co., 467 U.S. 947, 957 (1984); Broadrick v. Oklahoma, 413 U.S. 601, 612 (1973); see also, e.g., Thornhill v. Alabama, 310 U.S. 88, 97-98 (1940).

[9]467 U.S. 947, 955-57 (1984).

[10]405 U.S. 438, 445 (1972).  See also Barrows v. Jackson, 346 U.S. 249 (1953) (a seller of land may challenge a racially restrictive covenant on the grounds that enforcing the covenant violated the buyer’s equal protection rights).

[11]429 U.S. 167, 175 (1976).

[12]See, e.g., Vance v. Universal Amusement Co., 445 U.S. 308, 312 (1980).

[13]53 Cal. Rptr. 2d 599 (App. May 21), review granted, 921 P.2d 602 (Cal. 1996).

[14]See Gertz v. Robert Welch, Inc., 418 U.S. 323, 349-50 (1974) (punitive and presumed damages are impermissible in private-figure libel cases unless actual malice is shown).  But see Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 760 (1985) (generally allowing punitive damages and presumed damages in libel cases involving speech on a matter of purely private concern).

[15]Florida Star v. B.J.F., 491 U.S. 524, 541 n.9 (1989).

[16]Braun v. Soldier of Fortune Magazine, Inc., 968 F.2d 1110, 1119 n.7, 1121 n.13 (11th Cir. 1992).

[17]See, e.g., Smith v. California, 361 U.S. 147 (1959) (strict criminal liability for sales of obsnce literature impermissible); Manal Enterprises, Inc. v. Day, 370 U.S. 478 (1962) (suggesting same for civil penalties); New York Times, Inc. v. Sullivan, 376 U.S. 254 (1964) (rejecting strict liability for libel cases); New York v. Ferber, 458 U.S. 747 (1982) (suggesting strict liability is impermissible in child pornography cases).  But see Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749 (1985) (seemingly leaving open possibility of strict liability in private-concern libel cases).

[18]E.g., 29 C.F.R. §§ 1604.11(c), 1606.8(c); Cal. Gov’t Code § 12940(h)(1); College-Town v. Massachusetts Comm’n Against Discrimination, 400 Mass. 156, 166, 508 N.E.2d 587, 593 (1987); Lehmann v. Toys ‘R’ Us, 132 N.J. 587, 617, 626 A.2d 445, 460 (1993); Hanlon v. Chambers, 195 W. Va. 99, 108, 464 S.E.2d 741, 750 (1995).

[19]Rosenbloom v. Metromedia Inc., 403 U.S. 29, 51-52 (1971).

[20]People v. Mitchell Bros.’ Santa Ana Theaer, 180 Cal. Rptr. 728 (Cal. Ct. App. 1982).