Permissible Harassment Restrictions




Originally published in the UCLA Law Review; excerpted with modifications.


Cite text as Eugene Volokh, Freedom of Speech and Workplace Harassment, 39 UCLA L. Rev. 1791 (1992).





1.                  Harassment That Can Clearly Be Restricted


Some kinds of workplace harassment are clearly outside the protection of the First Amendment:


·                    Most obviously, unwanted physical conduct (whether fondling, hazing, vandalism, or what have you) is simply not speech, and is thus outside the First Amendment’s protection.

·                    Likewise, discrimination in job assignments -- which for complicated reasons is often treated under the rubric of “harassment” -- raises no free speech problems.

·                    Quid pro quo harassment (“sleep with me or you’re fired”) is just a form of extortion, and is also not constitutionally protected.

·                    Unprotected speech -- threats, fighting words, slander, and so on -- can be punished through workplace harassment law just as it can be through other kinds of laws.


These are all rather uncontroversial points.  The tough questions relate not to this, but to speech that is otherwise protected by the First Amendment.


2.                  The One-to-One / One-to-Many Distinction


I believe that one kind of otherwise protected speech can indeed be restricted by harassment law:  Unwanted “one-to-one” speech -- speech that is said to a particular employee because of her race, sex, religion, or national origin -- is of lower constitutional value and should be more readily restricable.  When someone is saying things to one listener that the listener clearly doesn’t want to hear, the speech is likely only to annoy and offend, not convince or edify.  If the main reason we value free speech is that it allows people to spread their ideas, then the value of one-to-one speech, aimed solely at an unwilling listener, is modest indeed.  A black employee who is told that he is a “nigger” is unlikely to become convinced of anything.And one-to-one speech can be restricted without interfering with the speaker’s right to communicate to other, potentially willing listeners.


Under my proposal, liability could thus be imposed not for any speech that creates a hostile work environment, but only for speech that (1) the speaker knows is unwanted,[1] that (2) is said to a particular employee because of her race, sex, religion, or national origin, and (3) that creates (together with whatever other nonspeech conduct might be present) a hostile work environment.


Personal face-to-face insults and persistent sexual propositions would thus generally be regulable, because they would be said to an individual offended employee and because it would be clear, either from the inherent offensiveness of the insult or from the listener’s responses, that the speech was unwelcome.  On the other hand, other speech, such as posters that are aimed at all one’s coworkers, or lunchroom conversations among willing listeners that are overheard by an unwilling one, will remain protected.


Of course, even restrictions on one-to-one speech do in some measure burden speakers:  Self-expression, to many, is an end in itself,[2] which may be one reason why even one-to-one speech, aimed at an unwilling listener, is often protected.[3]  A person may insult another to the other’s face, so long as the insult does not “have a direct tendency to cause acts of violence”;[4] a person may argue with a police officer, so long as the argument does not actually interfere with the police officer’s duties.[5]


Nonetheless, this burden -- unlike a burden on speech to potentially willing listeners -- is comparatively modest, as the Court has recognized elsewhere.  For instance, in Frisby v. Schultz, the Court upheld a ban on picketing of a private residence, while recognizing that general, one-to-many picketing in a residential area was protected.[6]  The ordinance, the Court found, was “narrowly tailored to protect only unwilling recipients of the communications . . . . [T]he picketing is narrowly one-to-one at the household, not the public.  The type of picketers banned by the [ordinance] generally do not seek to disseminate a message to the general public, but to intrude upon the targeted resident, and to do so in an especially offensive way.”[7]


Likewise, face-to-face insults that are likely to cause a fight can generally be banned, even though words communicated to the world at large on a jacket cannot be.[8]  A householder has veto power over the mail he gets, but this is a power that does not prevent the mailer from sending messages to others.[9]  And this distinction between one-to-one speech and other speech may justify the prohibitions on unwanted telephone calls, which have been largely upheld by lower courts.[10]


This analysis is equally true of all varieties of one-to-one speech, whether or not political, and whether or not intended to be offensive.  One-to-one vulgarities are certainly unlikely to convince the listener of anything except the speaker’s hostility.  The speaker may well get satisfaction from being able to express himself this way, but there is little reason to value the speaker’s satisfaction higher than the state interest in assuring equality of working conditions.  Explicitly sexual statements targeted at a listener, or even nonexplicit sexual propositions, should be just as restrictable.  Often the speaker will not intend to offend the listener (especially when he is trying to woo her), but once the listener lets the speaker know that these comments are unwelcome, then these comments, too, become unlikely to convince the listener of anything.


One-to-one political speech might be speech on a more “valuable” topic, but, again, once the listener has told the speaker that she does not want to hear more, the First Amendment value of the speech becomes negligible; in the words of Rowan v. Post Office Dep’t, “no one has a right to press even “good” ideas on an unwilling recipient.”[11]  Though it is possible that the speaker may eventually convince a hostile listener, it seems improbable.  One can imagine a female worker who every day is told a new reason why women should stay at home, and eventually becomes convinced, though she was initially offended by the idea.  More likely, however, the offended listener will every day get less and less receptive to the speaker’s message.  There should be nothing wrong with giving the employee the power to tell a coworker that she just does not want to hear anything more about this topic.


3.                  One-to-One Speech May Be Restricted Despite the Content-Based and Viewpoint-Based Nature of the Restriction


One of the reasons the one-to-one residential picketing ordinance in Frisby v. Schultz was upheld was its content-neutrality.  A similar ordinance that applied only to nonlabor picketing was struck down in Carey v. Brown[12] because “[the] permissibility of residential picketing under the [law was based] solely on the nature of the message being conveyed.”[13]  And harassment law, even when limited to one-to-one speech, is not only content-based, but viewpoint-based.  It can impose liability for an employee’s telling a black coworker that blacks are inferior, but not for the employee’s telling the coworker that blacks and whites are equal.  In a sense it is like the statute the Court struck down in Lamont v. Postmaster General, which banned, in the interests of preventing offense to the recipient, all mailings of Communist propaganda to anyone who had not explicitly requested it.[14]


Nonetheless, there are three reasons to sustain restrictions on offensive one-to-one speech despite their not being viewpoint-neutral.  First, unlike restrictions on one-to-many speech, they do not make it substantially harder for particular viewpoints to make their way into the “marketplace of ideas.”  Restrictions on one-to-one speech suppress only speech that is highly unlikely to convince the listener.  The spread of antiblack diatribes would not be substantially slowed by a requirement that they not be targeted at blacks.


Second, there may be no practical alternative to harassment law that still serves the state interest in ensuring equality of working conditions, but does it in a content-neutral or viewpoint-neutral way.  One can ban all picketing in front of a residence, but one certainly cannot ban all speech one-to-one at an employee in the workplace.  Harassment law’s restriction could be broadened to prohibit all offensive speech targeted at a particular employee,[15] regardless of whether the employee belongs to any particular group; this may make the restriction viewpoint-neutral.[16]  But it could also create an unacceptable burden for employers, because it would force them to intervene, on pain of possible liability, into every private disagreement or feud between two employees.[17]


4.                  Slippery Slope Concerns


In arguing for the protection of one-to-many speech, I cite the danger of the slippery slope:  Restrictions on such speech in the workplace can lead, and already are leading to restrictions in other places.  (Click here -- substanc.htm - SLIPPERY -- for details.)  The same argument can be made in response to my proposal here -- if we tolerate restrictions on unwanted one-to-one speech in the workplace, couldn’t these restrictions spread to schools, universities, public places, and so on?


I recognize that this danger exists, and I am troubled by the prospect that unwanted one-to-one speech might be banned -- especially in content- and viewpoint-based ways -- in many places.  For instance, I think that free inquiry and free expression in universities is so important that I would argue against restrictions on even one-to-one speech there.  Likewise, I think that sometimes citizens might sometimes have special rights to express their frustrations to government officials, such as police officers.[18]


Nonetheless, as I described above, one-to-one speech is already barred in certain ways even outside the workplace:  The government may ban unwanted one-to-one mail, phone calls, and in certain narrow circumstances even face-to-face insults and residential picketing.  I’m not sure these restrictions should be extended, and I’d certainly consider skeptically any proposals at extending them.  But if we do slide down the slippery slope to broad restrictions on unwanted one-to-one speech, it seems to me that the overall danger to free speech will be slight, much slighter than if we slide down the slope to broad bans on offensive one-to-many speech.

[1]The speaker might know this either because the listener has already told him that she does not want hear this sort of comment, or because the comment is generally known to be inherently offensive.

[2]See Steven Shiffrin, The First Amendment and Economic Regulation: Away from a General Theory of the First Amendment, 78 Nw. U. L. Rev. 1212, 1252 (1983) (“[T]he Court has been unwilling to confine the first amendment to a single value or even to a few values.”); Martin Redish, The Value of Free Speech, 130 U. Pa. L. Rev. 591, 626 (1982) (arguing that the main justification of free speech is self-expression).

[3]One-to-one insults might also sometimes be valuable in other ways, for instance, as safety valves for people to express their resentment at various political or social conditions, or as means of showing people what’s really in their coworkers’ heads.  See, e.g., Kingsley R. Browne, Title VII as Censorship: Hostile-Environment Harassment and the First Amendment, 52 Ohio St. L.J. 481, 541-42 (1991).  Nonetheless, unwanted one-to-one speech that’s valuable only for these reasons strikes me as less valuable than speech which is valuable both for these reasons and for its ability to communicate to willing listeners.

[4]Gooding v. Wilson, 405 U.S. 518, 523-24 (1972).

[5]Houston v. Hill, 482 U.S. 451, 461-63 (1987).

[6]487 U.S. 474, 486 (1988).  I don’t mean to imply that Frisby is a controlling precedent in the workplace; as I discuss elsewhere (click here -- substanc.htm - CAPTIVE -- for details), it was based on the special concern about the privacy of the home.  I cite Frisby only to show that the one-to-one-one-to-many distinction that I propose isn’t entirely unprecedented.

[7]Id. at 485-86 (emphasis added).

[8]Gooding v. Wilson, 405 U.S. 518, 523-24 (1972).

[9]Rowan v. Post Office Dep’t, 397 U.S. 728 (1970).  Cf. Organization for a Better Austin v. Keefe, 402 U.S. 415, 419 (1971) (striking down an injunction against a community group’s distributing leaflets criticizing a real estate broker, and distinguishing Rowan on the grounds that “respondent is not attempting to stop the flow of information into his own household, but to the public”).

[10]See generally Gormley v. Director, 632 F.2d 938 (2d Cir. 1980); State v. Anonymous, 34 Conn. Supp. 689, 389 A.2d 1270, appeal denied as State v. Gormley, 174 Conn. Supp. 803, 382 A.2d 1332 (1978); State v. Elder, 382 So. 2d 687 (Fla. 1980); Constantino v. State, 243 Ga. 595, 255 S.E.2d 710 (1979); United States v. Lampley, 573 F.2d 783 (3d Cir. 1978); People v. Smith, 392 N.Y.S.2d 968 (1977).  But see People v. Klick, 66 Ill. 2d 269, 362 N.E.2d 329 (1977) (striking down law); State v. Dronso, 90 Wis. 2d 110, 279 N.W.2d 710 (Wis. App. 1979) (same).  See generally Gormley v. Director, 449 U.S. 1023 (1980) (White, J., dissenting from denial of certiorari).

[11]397 U.S. 728, 738 (1970).

[12]447 U.S. 455 (1980).

[13]Id. at 461.

[14]381 U.S. 301, 310 (1965).

[15]This could not be done under Title VII as it now exists, but would require a new statute that banned all harassment, not just harassment based on race, sex, religion, or national origin.  It would also require a different state interest to justify it, because it would bar offensive speech regardless of its effect on equality in the workplace.

[16]See Boos v. Barry, 485 U.S. 312, 319 (1988).

[17]The law might also be made viewpoint-neutral by banning any discussion of racial or sexual matters, regardless of their point of view, in the workplace.  But this would obviously be a vastly overbroad and oppressive regulation, and might be no less harmful than an explicitly viewpoint-based rule.  See Geoffrey Stone, Restrictions of Speech Because of Its Content: The Peculiar Case of Subject-Matter Restrictions, 46 U. Chi. L. Rev. 81 (1978) (subject-matter restrictions that prohibit speech about a narrow range of political issues can have all the bad effects of viewpoint-based restrictions).

[18]See Lewis v. New Orleans, 408 U.S. 913 (1972) (Powell, J., concurring).