I. Introduction

American politics is ever more polarized,[1] with both sides increasingly regarding each other as unreasonable and even dangerous.[2] One consequence of this alarming state of affairs concerns free speech in the workplace: anecdotal evidence suggests an uptick in adverse employment actions against employees who engage in controversial political expression.[3] And while keeping politics out of the workplace may sound like a good idea, such actions can be based on speech that took place outside the workplace and was unrelated to it.[4]

In any event, whatever is allowed to private employers, when it comes to public employers, the Federal Constitution may get involved: First Amendment doctrine limits the ability of government employers to punish employees for their speech. The doctrine governing these constitutional protections—known as the Pickering test—is seemingly simple but can become complicated and even confusing when we descend into its details.[5] This Article surveys the general contours of the Pickering test with an eye to some important controversies arising in the lower courts.

II. First Amendment Standards

Until the second half of the twentieth century, constitutional doctrine placed no restrictions on the government’s ability to punish public employees for their speech—a position epitomized in Justice Holmes’s quip that “[t]he petitioner may have a constitutional right to talk politics, but he has no constitutional right to be a policeman.”[6] As late as 1952, the U.S. Supreme Court stuck to this position when it rejected a challenge to a New York law that barred school teachers from membership in subversive political groups: “It is clear that such persons have the right under our law to assemble, speak, think and believe as they will,” said the Court, but “[i]t is equally clear that they have no right to work for the State in the school system on their own terms.”[7]

As with much else in First Amendment doctrine, this position was swept aside in the second half of the twentieth century.[8] The Warren Court extended free speech protections to the speech of government employees, although these protections are, for obvious reasons, less extensive than protections from government regulations of non-employee speech. Acting as an employer, the government can punish speech it cannot otherwise reach as a regulator; but its powers to do so have become constitutionally limited.[9]

A. The Types of Employee Speech

1. Speech Spoken as an Employee

In principle, speech spoken as an employee gets no constitutional protections (qua employee speech).

i. Employment Grievance of No Public Concern. One category of unprotected speech spoken as an employee is employment grievances of a mere personal interest. The seminal case dealing with employment grievance speech is Connick v. Myers (1983), a divided 5–4 opinion.[10]

Sheila Myers, an Assistant District Attorney, was told she would be transferred to another position and in response wrote and distributed a questionnaire regarding her office’s transfer policies and working conditions. Among the issues raised in the questionnaire were office morale, need for a grievance committee, lack of confidence in supervisors, and the concern that employees were pressured to work in political campaigns.[11] She was immediately fired from her job.[12]

In reviewing the constitutionality of the termination, the Supreme Court held that the questionnaire—but for one question—did not deserve constitutional free speech protections. According to Connick, “when a public employee speaks not as a citizen upon matters of public concern, but instead as an employee upon matters only of personal interest,” the speech is not constitutionally protected.[13] The Court then determined that “with but one exception, the questions posed by Myers to her co-workers [did] not fall under the rubric of matters of ‘public concern’” and therefore merited no constitutional protections.[14] The one exception concerned the alleged pressure on employees to work on political campaigns. That part was protected and subjected to constitutional scrutiny—but the Court concluded that it did not make the firing unconstitutional.[15] I will return to this decision shortly, when discussing the “public concern” inquiry.

ii. Speech Spoken as Part of the Employee’s Duties of Employment. Another form of speech spoken as an employee that gets no constitutional protection is speech spoken as part of the employee’s duties of office. The seminal Supreme Court case on this sort of speech is Garcetti v. Ceballos (2006)—another 5–4 decision dealing with another litigious prosecutor.[16] In Garcetti, a Deputy District Attorney, Richard Ceballos, wrote a memorandum recommending the dismissal of a criminal charge because a search warrant application contained inaccurate information. The recommendation was rejected, and Ceballos was demoted.[17] He sued, and the Supreme Court held that the memorandum was not entitled to constitutional free speech protections: “We hold that when public employees make statements pursuant to their official duties, the employees are not speaking as citizens for First Amendment purposes, and the Constitution does not insulate their communications from employer discipline.”[18] Employer retaliation against such speech “simply reflects the exercise of employer control over what the employer itself has commissioned or created.”[19] “Government employers,” added the Court, “like private employers, need a significant degree of control over their employees’ words and actions; without it, there would be little chance for the efficient provision of public services.”[20]

The dissenters argued that because the speech was about matters of public concern (i.e., about prosecution decisions at the local prosecutor’s office), it was constitutionally protected—even if it was part of the employee’s official duties. That did not necessarily mean that the demotion was unconstitutional; but it did mean that it required constitutional scrutiny.[21] The dissent pointed out that government whistleblowers often express themselves as part of their official duties[22] and that Supreme Court precedents have repeatedly emphasized that “[t]he interest at stake [in employee speech cases] is as much the public’s interest in receiving informed opinion as it is the employee’s own right to disseminate it.”[23]

In Pickering v. Board of Education, the case that gave us the constitutional test for public employees’ speech, a public school teacher was fired for publishing a “Letter to the Editor” that criticized the local school board’s allocation of school funds.[24] The Court, declaring the firing unconstitutional, noted that teachers were “the members of a community most likely to have informed and definite opinions” about public school expenditures.[25] Similarly, in City of San Diego v. Roe, the Court stated that “[w]ere [public employees] not able to speak on [the operation of their offices], the community would be deprived of informed opinions on important public issues.”[26] The Supreme Court has repeatedly “acknowledged the importance of promoting the public’s interest in receiving the well-informed views of government employees engaging in civic discussion.”[27] The majority—complained the Garcetti dissenters—was giving short shrift to this well-recognized public interest.[28]

The majority responded that the public’s interest in whistleblowing was sufficiently protected by statutes.[29] Accordingly, “when a public employee speaks . . . as an employee,” she has no First Amendment protections against adverse employment actions;[30] she speaks as an employee when the speech in question is about an employment grievance of a mere personal interest or when the speech owes its existence to the employee’s professional duties.[31]

Note that these two categories of speech spoken as an employee—employment grievances and speech spoken pursuant to the duties of office—do not occupy the same place in the flowchart of the constitutional analysis. Whether speech is uttered as part of the duties of office is a threshold question; if it is, it is recognized as unprotected employee speech and the inquiry is at an end. But employment grievances of purely personal concern are declared unprotected one step further down the line: they are removed from constitutional protection by virtue of the fact that, according to the Court, employment grievances of purely personal concern are not of public concern.

The fact that the two forms of speech spoken as an employee are recognized as such at different points in the constitutional analysis not only makes for a rather inelegant flowchart, it also creates some complications, as we will now see.

2. Speech Spoken as a Citizen on Matters of Public Concern

Speaking as a citizen and speaking as an employee appear to be mutually exclusive and mutually exhaustive categories: a speaker can speak either as one or the other.[32] An employee therefore speaks as a citizen whenever the speech is neither an employment grievance nor speech that owes its existence to the employee’s professional duties. And while speech spoken as an employee is unprotected, when an employee speaks as a citizen, any reactive adverse employment action would be subjected to a constitutional scrutiny—at least when the speech is about matters of public concern.

i. Public Concern. Two principal distinctions, already mentioned, are involved in drawing the line between protected and unprotected public employee speech: the first between speech spoken as an employee and that spoken as a citizen, the second between speech on matters of public concern and that not on matters of public concern. As we saw, speech spoken as an employee gets no constitutional protections (qua employee speech). We also saw that speech that is part of the duties of office is recognized as employee speech ab initio.[33]

By contrast, if the speech is not part of the duties of office, courts proceed to determine whether the speech is of public concern. If it is, then the speech is constitutionally protected, and the Pickering balancing test kicks into action. (We will examine the operation of the test in the next section.) If it isn’t, the speech may be unprotected (more on this below). The employee bears the burden of showing that the speech in question is on matters of public concern, and that question “is one of law, not fact.”[34]

What are matters of public concern? The Supreme Court has stated that a matter of public concern “is something that is a subject of legitimate news interest; that is, a subject of general interest and of value and concern to the public at the time of publication.”[35] And while “the boundaries of the public concern test are not well defined”—as can be seen in the occasional judicial disagreement[36]—there is no doubt that the concept is a broad one: speech is of public concern when it can “be fairly considered as relating to any matter of political, social, or other concern to the community.”[37] Indeed, speech on matters of public concern need not address “matters of transcendent importance, such as the origins of the universe or the merits of constitutional monarchy,”[38] and “[t]he inappropriate or controversial character of a statement is irrelevant to the question whether it deals with a matter of public concern. ‘[D]ebate on public issues should be uninhibited . . . [and] may well include vehement, caustic . . . attacks . . . .’”[39] (For example, the Supreme Court held that expressing approval for the assassination of the U.S. President qualified as speech on matters of public concern.[40])

But things can get a bit complicated: in Connick v. Myers, where the Supreme Court found that an employment grievance was not constitutionally protected, the Court announced that the “public concern” inquiry was not limited solely to the content of the speech.[41] Factors such as the employee’s purpose in uttering the speech, or the intended audience of the speech, must also be considered.

In Connick, the Court concluded that a questionnaire distributed to other employees in the District Attorney’s Office was not of public concern because of its purpose and distribution:

Myers did not seek to inform the public . . . . Nor did Myers seek to bring to light actual or potential wrongdoing . . . . [T]he questionnaire, if released to the public, would convey no information at all other than the fact that a single employee is upset with the status quo. . . . [T]he focus of Myers’ questions is not to evaluate the performance of the office but rather to gather ammunition for another round of controversy with her superiors. These questions reflect one employee’s dissatisfaction with a transfer and an attempt to turn that displeasure into a cause célèbre.[42]

In truth, the presence of an employment grievance may even cast doubt on the reliability of the allegations. Thus, in addition to the content of the speech, the public concern inquiry includes consideration of the “form, and context of a given statement . . . .”[43] What may be of public concern as a matter of pure content, may be of no public concern when raised in the context of an employment grievance: “a questionnaire not otherwise of public concern does not attain that status because its subject matter could, in different circumstances, have been the topic of a communication to the public that might be of general interest.”[44]

The dissenters objected to the introduction of these additional factors to the public concern inquiry. Whether speech was on a matter of public concern, they said, depended solely on its content:

[W]hether a particular statement by a public employee is addressed to a subject of public concern does not depend on where it was said or why.

. . . .

Myers’ questionnaire addressed matters of public concern because it discussed subjects that could reasonably be expected to be of interest to persons seeking to develop informed opinions about the manner in which the Orleans Parish District Attorney, an elected official charged with managing a vital governmental agency, discharges his responsibilities.[45]

Indeed, the record before the Court showed that the local press was very interested in the incident.[46] The dissenters thought that the presence of an employment grievance should impact the constitutional analysis only further down the line—when determining whether punishing the speech did or did not violate the First Amendment.[47] But the punishing needed to be subjected to constitutional scrutiny.

As we saw, the majority disagreed and made the presence of an employment grievance a factor in deciding whether the speech was of public concern—and therefore whether it received constitutional protections to begin with.[48] Note, however, that the Court had implicitly recognized that some employment grievances may be of public concern. It did so when it held that one part of the questionnaire—the one dealing with alleged pressure on employees to work on political campaigns—was in fact of public concern and therefore entitled to constitutional protections (even though the questionnaire, of which it was part, was “most accurately characterized as an employee grievance”).[49] Thus, employment grievances can, in principle, rise to the level of public concern—and hence be constitutionally protected.

Thus, lower courts have recognized some apparent employment grievances as matters of public concern and therefore protected.[50] In Johnson v. Multnomah County, for example, a public employee who had failed to obtain a promotion made critical allegations about the person who was promoted in her stead and was fired for it.[51] The Ninth Circuit extended constitutional protections for the speech:

[The] allegations did not constitute an “employee grievance” within the meaning of Connick. Although Johnson was embittered about not having been promoted, her allegations did not concern the “minutiae of workplace grievances” . . . but rather concerned information that is of inherent “relevance to the public’s evaluation of the performance of government agencies.”[52]

The fact that the speech was part of an employment dispute did not make it automatically disqualified from First Amendment protections. Indeed, such a rule would be unreasonable: it would defy logic to declare that all public employee speech uttered as part of an employment dispute is of no public concern.[53] Consider, for example, an employee’s grievance alleging racial discrimination in a government office.[54]

ii. Mere Private Concern.[55] Speech spoken as an employee as part of the duties of office is not constitutionally protected. Speech spoken as an employee as part of an employment grievance is not protected if of a mere personal interest but is protected if of public concern (and the fact that the speech is part of an employment grievance informs the public concern inquiry). Speech spoken as a citizen on matters of public concern is constitutionally protected. But what about speech spoken as a citizen on a matter that is not of public concern? Can a government employer fire an employee for expressing his pleasure at watching dog fights, or for calling her neighbor “a fat pig,” without any First Amendment scrutiny qua employee speech? This is no idle speculation: many government departments have “unbecoming” conduct rules that are often used as a basis for adverse employment actions.[56]

The better view is that speech spoken as a citizen on matters that are not of public concern should receive constitutional protections. After all, the abandonment of the doctrine that denied First Amendment protections to employee speech was rooted in the claim that public employees should not be made to surrender their individual rights when they join government service.[57] And for these employees, the interest in speech on matters of no public concern may be no different than the interest in speech about matters of public concern. Indeed, First Amendment protections have never been limited to speech on matters of public concern: while protecting a robust discourse on matters of importance to the body politic, free speech protections go much further and extend to expression of a purely personal nature—i.e., to “the individual’s right to speak his own mind.”[58] That right would be gravely harmed if millions of public sector employees (approximately 20 million—more than 12% of the workforce[59]) could be fired with impunity for speaking on private or personal matters in ways their employers disliked.

Moreover, there is something paradoxical in refusing to extend protections for speech spoken as a citizen on a matter of mere private concern. After all, the interest of government employers to punish speech is greatly reduced when employees speak on matters of mere personal interest.[60] As the Second Circuit put it,

[M]echanically applying a categorical public concern test to off-duty speech unrelated to Government employment would lead to the somewhat anomalous result that the Government would have far less latitude to dismiss an employee for a public display . . . involving public concerns than it has for, say, speech that was uttered in the privacy of the employee’s bedroom but was not on a matter of public concern.[61]

Unfortunately, Supreme Court statements appear to be to the contrary. A series of Supreme Court cases explicitly declared that employee speech not on matters of public concern was not constitutionally protected.[62] The idea stretches back to the seminal case in this area, Pickering v. Board of Education (1968), which extended constitutional protections for “the interests of the [employee], as a citizen, in commenting upon matters of public concern . . . .”[63] Connick v. Myers (1983) repeated the claim:

Pickering, its antecedents, and its progeny lead us to conclude that if Myers’ questionnaire cannot be fairly characterized as constituting speech on a matter of public concern, it is unnecessary for us to scrutinize the reasons for her discharge. When employee expression cannot be fairly considered as relating to any matter of political, social, or other concern to the community, government officials should enjoy wide latitude in managing their offices, without intrusive oversight by the judiciary in the name of the First Amendment.[64]

Similar statements can be found in Rankin v. McPherson (1987),[65] Waters v. Churchill (1994),[66] and Garcetti v. Ceballos (2006).[67] But Pickering dealt with speech spoken as an employee as part of an employment grievance, not speech spoken as a citizen. In fact, none of these Supreme Court cases dealt with speech spoken as a citizen on matters of no public concern. In other words, all these statements appear in dicta, in cases dealing with different forms of speech.[68]

The one exception is City of San Diego v. Roe (2004), a Supreme Court case that dealt specifically with speech spoken as a citizen on matters that are not of public concern.[69] In Roe, a police officer was fired for selling sexually explicit videos of himself stripping off a police uniform and masturbating.[70] The officer brought a lawsuit claiming, inter alia, that his termination violated his freedom of speech. A district court granted summary judgment to the City, but the Ninth Circuit reversed, holding that the speech was on matters of public concern and that the termination was unconstitutional.[71] The Supreme Court granted review and reversed. The short per curiam opinion is fairly confusing.[72] On the one hand, the Court stated that in order to get First Amendment protections, “a public employee’s speech must touch on a matter of ‘public concern,’” whereas “Roe’s expression does not qualify as a matter of public concern . . . .”[73] On the other hand, the Court upheld the firing by repeatedly relying on the fact that “[t]he speech in question was detrimental to the mission and functions of the employer”:

Although [Roe’s] activities took place outside the workplace and purported to be about subjects not related to his employment, the [SDPD] demonstrated . . . legitimate and substantial interests [of its own that were compromised by his speech]. [Far from confining his activities to speech unrelated to his employment,] Roe took deliberate steps to link his videos and other wares to his police work[, all in a way injurious to his employer] . . . .

. . . .

The use of the uniform, the law enforcement reference in the Web site, the listing of the speaker as “in the field of law enforcement,” and the debased parody of an officer performing indecent acts while in the course of official duties brought the mission of the employer and the professionalism of its officers into serious disrepute.[74]

However, if a public employee’s speech on matters of no public concern merited no constitutional protections, it was irrelevant to the constitutionality of the firing whether the speech was or was not related to the employee’s work, or whether it did or did not disrupt department operations (or, indeed, violated department regulations, as the Court was also careful to point out).[75]

Given that almost all Supreme Court statements on the subject were dicta, that the Roe holding seems confused, and that the intellectual pull is towards the opposite view, it is unsurprising that commentators and courts are divided on the status of such speech.[76] Some of the courts that favor extending constitutional protections draw a distinction between speech spoken as a citizen on matters that are of no public concern but is related to the employee’s job or was uttered on the job and speech that is unrelated to it—extending constitutional protections only to the latter.[77]

Most of these courts, which extend constitutional protections to speech on matters of no public concern, go on to employ the Pickering balancing test—the same test applicable to speech spoken as a citizen on matters of public concern. The advantage of this test—to which we shall now turn—is that it allows courts to take into account all the surrounding circumstances, including whether the speech in question touches on matters of public concern.[78] In other words, the fact that speech on matters that are not of public concern receives constitutional protections does not mean that the question of public concern loses its entire significance. Speech that is not of public concern receives less constitutional protection under the Pickering test itself, even if it is protected.[79] And as already noted, affording such speech constitutional protections does not mean that firing an employee for such speech is necessarily or even probably unconstitutional: the Pickering–Connick test allows the government substantial latitude in punishing public employees’ speech.

III. The Pickering Test—The Two Sides of the Scale

The constitutional test for employee speech cases was announced in Pickering v. Board of Education (1968).[80] The Pickering test is a balancing inquiry pitting “the interests of the [public employee], as a citizen” against “the interest of the State, as an employer, in promoting the efficiency of the public services it performs through its employees.”[81] While employees have a substantial interest in “speak[ing] out freely . . . without fear of retaliatory dismissal,”[82] this interest must be weighed against the government’s interest in “effective and efficient fulfillment of its responsibilities to the public.”[83] Whereas employees bear the burden of showing that the speech in question is protected (e.g., spoken by a citizen on matters of public concern), the government bears the burden of “show[ing] that its legitimate administrative interests outweigh the [employee’s] First Amendment rights.”[84] Both are questions of law, to be answered by courts rather than juries.

A. The Government’s Interests

The Supreme Court has identified various components of the government’s interest in “effective and efficient fulfillment of its responsibilities to the public.”[85]

1. Preventing Actual or Expected Workplace Disruption

First, the government can seek to justify adverse employment action by demonstrating that the statement in question “impairs discipline by superiors or harmony among co-workers, has a detrimental impact on close working relationships . . . or [otherwise] interferes with the regular operation of the enterprise.”[86] “Regular operation” includes smooth inter-office relations, effective employer control, and even relations with the public (especially for those departments, like the police, whose work depends on cooperation from the public). These and related concerns appear under the rubric of “workplace disruption.”

A government employer need not demonstrate actual workplace disruption in order to justify adverse employment action. The Connick Court did “not see the necessity for an employer to allow events to unfold to the extent that the disruption of the office and the destruction of working relationships is manifest before taking action.”[87] Instead, an adverse employment action can be constitutionally justified if the employer “reasonably believed [that the speech] would disrupt the office, undermine his authority, and destroy close working relationships.”[88] Relevant considerations include the content of the speech, as well as its “manner, time, and place”—for example, whether it took place in the office or elsewhere, whether it was delivered defiantly to a superior, or the extent of its distribution.[89] Government employers’ predictions of disruption are entitled to a measure of deference from the courts[90] but cannot be based on mere speculation and must be supported by “a substantial showing.”[91] In other words, notwithstanding the deference to the government’s prediction of disruption, “engaging in Pickering balancing,” as the Fifth Circuit has put it, “is not like performing rational basis review, where we uphold government action as long as there is some imaginable legitimate basis for it.”[92]

2. Unfitness for the Job

The government can also justify adverse employment action by showing that the speech in question demonstrates the employee’s unfitness for the job (Pickering speaks of the employee’s “proper performance of his daily duties”).[93] If a CIA agent expresses support for absolute government transparency, she may be fired irrespective of actual or likely disruption because the speech evidences unfitness for the position.

Claims of unfitness are obviously sensitive to the employee’s specific position within the government.[94] In Rankin, for example, the Supreme Court thought it significant that the employee, who had made a highly offensive political remark, did not work with the public.[95] As a general matter, it can be surmised that the higher the rank and authority of the employee in question, the easier it should be for the government to establish unfitness.[96]

B. The Employee’s—and the Public’s—Interests

The principal interest on the other side of the balancing scale is, broadly speaking, “the employee’s constitutionally protected interest in freedom of expression.”[97] But, as already touched upon, the employee’s interest is not the only one involved: the courts also consider “[t]he public interest in having free and unhindered debate on matters of public importance.”[98] Accordingly, courts factor in the social value of the employee’s speech in deciding the constitutionality of punishing it.

1. The Value of the Speech

Courts are usually barred from evaluating the merit of speech when calibrating its constitutional protections, and for a good reason. The government, of which courts are part, is supposed to treat all protected expression equally and abstain from suppressing speech it dislikes. The courts have enshrined this principle, inter alia, in the important content-based/content-neutral distinction, which seeks to minimize content assessment as a legitimate legislative or judicial consideration contemplating restrictions on speech.[99]

This principle is abandoned in the context of the Pickering balancing test: here, courts accord greater weight to the employee’s side of the scale the greater the social value of the speech in question. The main reason for this departure is that the employee’s side represents not only the employee’s interest in speaking but also the public’s interest in hearing. Thus, as the Supreme Court put it in Connick, “a stronger showing [on the government’s part] may be necessary if the employee’s speech more substantially involved matters of public concern.”[100] The Ninth Circuit elaborated:

[N]ot all statements of “public concern” are treated equally under the Pickering balancing test . . . [C]ourts . . . have effectively established a sliding scale for how much weight to give . . . when balancing the employee’s and the government’s competing interests.

. . . .

Courts have thus implicitly applied a sliding scale in which the “state’s burden in justifying a particular discharge or adverse employment action varies depending upon the nature of the employee’s expression.”[101]

Thus, in the limited context of government employment, the usual principle of courts’ neutrality gives way and courts are under a positive obligation to assess the value of speech when determining the constitutionality of adverse employment actions. The Ninth Circuit again:

At the apex of the First Amendment [protections for speech of government employees] rests speech addressing problems at the government agency where the employee works.

. . . .

On the other hand, at least one court has suggested that racially charged comments that have no connection to the government employee’s workplace arguably receive less First Amendment protection under the Pickering balancing test for government employees. See, e.g., Grutzmacher v. Howard Cty. [sic], 851 F.3d 332 (4th Cir. 2017) . . . .[102]

The same may be true, for example, for some sexually explicit speech, or speech that calls for unlawful violence.[103] On the other hand, speech about political matters may rank very high on this sliding scale: “speech on public issues occupies the ‘highest rung of the hierarchy of First Amendment values,’ and is entitled to special protection.”[104] But, of course, not all political speech is equally valuable: the endorsement of political violence or white supremacy may be political, but it is socially harmful and may therefore rank as having little to no value. In Hawkins v. Department of Public Safety & Correctional Services, for example, where Maryland’s highest court was disinclined to remove constitutional protections from speech spoken as a citizen that was of no public concern (racist epithets hurled at a bank teller), the court stated that “it nevertheless is clear that the comments are near the periphery, and not at the core, of [free speech] protection,” and “any first amendment interest . . . must be of a very low order.”[105] Thus it is here, when courts evaluate the value of speech under Pickering, that the distinction between matters of public concern and matters of no public concern finds its significance even if, as this Article has argued, the distinction does not spell the difference between protected and unprotected speech that is spoken as a citizen.

i. Limiting Strategies. In a radical break from ordinary First Amendment jurisprudence, the Pickering balancing test asks courts to evaluate the merit of the speech in question so that some contents receive greater protections than others. This exception to usual First Amendment principles is, of course, highly problematic: given the natural tendency to find value in one’s own opinions while discounting the value of opposing views, what’s to prevent courts—themselves organs of government—from favoring political or social orthodoxy at the expense of marginal or controversial views? Indeed, constitutional free speech protections are most important when it comes to the protection of unorthodox expression: orthodox opinions hardly need the protection of constitutional liberties.[106] It is the unorthodox, the offensive and despised expression, that is in most need of constitutional protections.[107] As a Maryland Court of Appeals justice put it:

A rule requiring the content of speech to be valued is fraught with many and, I fear, insurmountable complexities. The most difficult is that the process of valuation is inherently subjective and, thus, aside from the extremes, perhaps, there will be no uniformity; the value of the speech may depend upon which court hears the case.[108]

Unfortunately, it is hard to detect any constraining strategies in the caselaw. Although the Supreme Court declared that “[v]igilance is necessary to ensure that public employers do not use authority over employees to silence discourse . . . [merely] because superiors disagree with the content of the employees’ speech,” courts gauging the value of speech have had little to say to allay fears that favored views are being privileged at the expense of marginal or disliked ones.[109]

Still, long-standing First Amendment doctrines offer some obvious guideposts.

a. Viewpoint Discrimination. For one thing, while content-based evaluation is permissible (indeed mandated) under Pickering, viewpoint-based discrimination is a different matter. Viewpoint discrimination has long been regarded with greater suspicion than the one accorded content discrimination. As the Supreme Court put it in other First Amendment contexts, “where . . . the [government’s] suppression of speech suggests an attempt to give one side of a debatable public question an advantage in expressing its views to the people, the First Amendment is plainly offended.”[110] That principle appears to be equally applicable to the courts themselves.[111] In principle, all viewpoints on a given subject matter should be accorded equal weight by judges appraising the value of speech under Pickering.[112]

In fact, it seems that one of the very purposes of extending constitutional protections to public employees’ speech is to minimize the government employer’s ability to engage in viewpoint discrimination—i.e., to retaliate against employees for taking one side rather than another in a controversial but legitimate debate.[113]

b. Legitimate and Illegitimate Speech. Nevertheless, the prohibition on viewpoint-based discrimination when conducting a Pickering analysis is not absolute. Some viewpoints may in fact be ranked lower than others—with the result that the government’s burden to justify an adverse employment action is reduced. For example, racist speech is obviously less protected, under Pickering, than speech expressing support for racial equality. Indeed, while the Supreme Court referred to renunciations of racial discrimination as “a matter inherently of public concern,”[114] racist speech is sometimes classified as no matter of public concern at all—with the consequence that “[c]ourts routinely find that a plaintiff’s right to engage in racist off-duty speech that does not touch on any matter of public concern is outweighed by public agencies’ interest in a harmonious work environment and a positive relationship with the public they serve.”[115]

The operative concept here appears to be illegitimacy: viewpoint-based discrimination can be permissible vis-à-vis illegitimate positions.[116] To be sure, illegitimacy is a very high bar—lest the exception swallow the rule of viewpoint neutrality. Lots of speech with which we disagree, and even vehemently condemn, is perfectly legitimate. Thus, so-called pro-life and pro-choice advocates may regard their opponents as not only wrong but also dangerous and immoral; but almost all such expressions are part of a legitimate political discourse. For speech to be illegitimate it must be widely recognized not only as harmful but also as clearly beyond the pale, as outright unreasonable, and as clearly erroneous. This normative judgment is therefore informed by the perceived truth or falseness of a proposition, its underlying values (or their lack thereof), its potential harm, or the legality or illegality of the actions it endorses. The offensiveness of the message, however, is irrelevant.[117]

It is worth emphasizing just how high the threshold of illegitimacy must be, especially in regard to matters of public concern—particularly when it comes to partisan political expressions, which the Supreme Court has been especially keen to protect in public employees’ speech cases. The Court forbade the government to discharge employees for partisan affiliation unless it “can demonstrate that party affiliation is an appropriate requirement for the effective performance of the public office involved”—a condition that ordinarily limits such discharges to policy makers.[118] And it went out of its way to protect highly offensive partisan expressions. One of the earliest cases protecting public employees’ speech invalidated a New York statute that barred from public employment any person who willfully advocated “the doctrine that organized government should be overthrown by force or violence, or by assassination of the executive head or of any of the executive officials of government, or by any unlawful means.”[119] And in Rankin, the Supreme Court held unconstitutional the firing of an employee who expressed approval of the assassination attempt on President Reagan.[120]

To be sure, the question of illegitimacy can be difficult. At what point does the denunciation of abortion providers cross the line of illegitimacy—as it surely does with the advocacy of violence against doctors who perform abortion (even when such advocacy does not amount to true threats or incitement, which are wholly constitutionally unprotected)? Is it illegitimate to claim that COVID vaccines are dangerous to one’s health and are promoted solely for financial reasons? (I do not think so.) Still, the possibility of reasonable disagreements is, of course, the lot of many a constitutional doctrine.[121]

In short, while patently illegitimate speech does enjoy First Amendment protections against government regulation,[122] such speech receives lesser protections when it comes to the government’s ability to punish the speech of its employees. The concept of illegitimacy well captures the purpose of the Pickering ranking of speech—namely, identifying the sort of speech that government employers, who owe a duty of service to the entire community, may properly suppress. Indeed, the Supreme Court appealed to the concept of legitimacy in its very definition of “public concern,”[123] and “a stronger showing may be necessary if the employee’s speech more substantially involved matters of public concern.”[124]

It is worth remembering that even a determination of illegitimacy does not necessarily mean that government employers can retaliate against that speech. All it means is that, in the balancing of interests that makes up the Pickering inquiry, the speech is ranked lower than other viewpoints so that it should be easier for the government to demonstrate the disruption that would make its retaliation constitutional. But disruption, or reasonable fear of disruption, or unfitness for the job, must still be proven. The converse, of course, is also true: employers may be constitutionally permitted to punish perfectly legitimate expression—if the government’s proof of actual or expected disruption or unfitness is sufficiently strong.

Concern over the fact that government employers may be permitted to punish perfectly legitimate expression under the Pickering balancing test has been expressed by some judges through an appeal to the “heckler’s veto” doctrine.

ii. Heckler’s Veto. The reference to heckler’s veto is a familiar First Amendment trope expressing the concern that speakers may be silenced merely because some people dislike what they say. As one court of appeals put it, “Government’s instinctive and understandable impulse to buy its peace—to avoid all risks of public disorder by chilling speech assertedly or demonstrably offensive to some elements of the public—is a recurring theme in first amendment litigation.”[125] In the context of public employees’ speech, a Ninth Circuit concurring opinion provided some hypotheticals:

A measurable segment of the population, for example, is vigorously antagonistic to homosexual activity and expression; it could easily be encouraged to mobilize were a police officer discovered to have engaged, off duty and unidentified by his activity, in a Gay Pride parade, or expressive cross-dressing, or any number of other expressive activities that might fan the embers of antagonism smoldering in a part of the population.[126]

But courts are divided over the role of heckler’s veto in employee speech cases, with a number of courts simply dismissing that concern as inapposite to this area of the First Amendment.[127] For such courts, the government’s interest as an employer allows it to terminate public employees engaged in perfectly legitimate or even righteous speech if it can provide proof of disruption—which may include objections from the general public.

Wherever one stands in regard to the role of the heckler’s veto principle in employee speech cases, one thing is clear: the enhanced value of legitimate speech under Pickering serves to reduce the risk that government employees would be punished for legitimate expression. True, the government may still be able to provide proof of workplace disruption even in cases of legitimate content; but in such cases its burden is particularly high:

“[T]he amount of disruption a public employer has to tolerate is directly proportional to the importance of the disputed speech to the public.” “The more tightly the First Amendment embraces the employee’s speech, the more vigorous a showing of disruption must be made by the employer.”[128]

2. The Proportionality of the Adverse Employment Action

Adverse employment actions in response to employees’ speech run the gamut from reprimands, to demotions, to outright terminations.[129] Since the Pickering test “is a fact-intensive inquiry that requires consideration of the entire record and must yield different results depending on the relative strengths of the issue of public concern and the employer’s interest,” it seems that the severity of the sanction ought to be considered in the inquiry.[130]

Not everyone agrees that the severity of the government’s action should play a role. Justice Scalia, dissenting in Rankin, protested that:

We are not deliberating . . . (or at least should not be) about whether the sanction of dismissal was, as the concurrence puts it, “an . . . intemperate employment decision.” . . . [W]e are not sitting as a panel to develop sound principles of proportionality for adverse actions in the state civil service. We are asked to determine whether . . . [the employee] had a right to say what she did—so that she could not only not be fired for it, but could not be formally reprimanded for it, or even prevented from repeating it endlessly into the future.

. . . .

I emphasize once again that that is the issue here—and not, as both the Court’s opinion and especially the concurrence seem to assume, whether the means used to effect suppression (viz., firing) were excessive.[131]

But as Justice Scalia concedes, the majority of the Court believed otherwise. It is also implausible that the Rankin decision protected the employee’s right to voice support for the President’s assassination “endlessly into the future,” or that it forbade the government employer from imposing a milder sanction for such an incendiary on-the-job statement. The claim is also in tension with lower courts’ consideration of employees’ expression of regret for their speech—because expressions of regret may militate against outright termination, but they also demonstrate that the employee was out of line.[132]

Sound First Amendment principles also favor the consideration of the severity of the punishment. For one thing, as already mentioned, concern over chilling protected expression is a principal concern of First amendment doctrine; and allowing government employers to engage in the ultimate employment punishment—i.e., discharge—for any employee speech that may cause workplace disruption would cast a chilling pall over employees’ free speech.[133] Moreover, the degree to which intemperate or inappropriate speech harms government operations, or demonstrates unfitness for the job, is obviously a continuum, and so is the severity of adverse employment actions. Asking courts to close their eyes to the correspondence between these two in a balancing test makes little sense. And as already mentioned, speech that may have to be tolerated once ought not necessarily be tolerated a second or third time. Whatever the ability of government employers to discharge employees might be, their ability to reprimand or suspend them ought to be greater. Moreover, excluding the severity of the punishment from the constitutional calculus raises the concern that government employers would single out speech they disfavor for the ultimate punishment, while saving milder sanctions for speech they favor. Finally, taking into account the severity of the government employment action serves to accommodate the interests of both employees and their employers—who should be granted greater latitude when imposing milder sanctions.

IV. The Content of the Speech

Although the Pickering balancing is ultimately a legal question, to be resolved by courts rather than juries, its resolution often involves factual disputes. These can include, for example, the alleged workplace disruption, or the question of whether the adverse employment action was in retaliation for the speech. And sometimes there are factual questions about the content of the speech itself: the parties may disagree about what had been said and about the meaning of what had been said.

A. What Had Been Said

In Waters v. Churchill, a nurse was fired from a public hospital because of something she had allegedly said to a colleague, but the nurse disputed that she had said what the hospital claimed she did.[134] The district court granted the hospital’s request for a summary judgment after finding that it made no difference whether she said what the hospital claimed she said or what she claimed she did; but the Seventh Circuit reversed, holding that the two versions had different constitutional implications and that the district court therefore had to determine what had been actually said.[135] The Supreme Court granted review and reversed, holding, in a plurality opinion, that the dispositive question was not what a trier of fact believed had actually been said but what the government employer reasonably believed had been said—even if that belief may be wrong.[136] (A three Justice concurrence was even less protective of the employee, claiming that the government employer’s belief need not even be reasonable: all that mattered was what the employer believed, reasonable or not.[137])

The Court’s rationale for this surprising decision—which allowed a potentially mistaken factual belief to be a proper basis for punishing constitutionally protected speech—was its reluctance to impose judicial-like procedures on public employers engaged in adverse employment actions. Government employers, said the Court, “should be [able] to use personnel procedures that differ from the evidentiary rules used by courts, without fear that these differences [would] lead to liability.”[138]

As the dissent pointed out, this was an exceptionally weak argument:

The risk that a jury may ultimately view the facts differently from even a conscientious employer is not, as the plurality would have it, a needless fetter on public employers’ ability to discharge their duties. It is the normal means by which our legal system protects legal rights and encourages those in authority to act with care.[139]

There was nothing unusual about imposing on employers—or on anybody for that matter—the risk that juries may end up disagreeing with their factual assessments if an incident ends up in court. Public employers were free to adopt whatever pre-termination procedures they wished; but if they fired an employee for speech she did not in fact make, they could be in violation of the First Amendment. Allowing public employers to fire employees for speech they did not make—in order to save them from potentially cumbersome pre-termination procedures—was a surprisingly cavalier treatment of a fundamental constitutional right.

Still, Waters’s threat to the free speech of government employees remains minimal: the occasions for reasonable good faith mistakes about what had been said are rather limited. But another type of content-related factual dispute is far more common and more threatening: disputes about the meaning of what had been said.

B. The Meaning of the Speech

1. Speaker’s Meaning

In Rankin v. McPherson, a government employee remarked to a colleague about the failed assassination of President Reagan, “if they go for him again, I hope they get him.”[140] Then, when questioned about the comment by her boss, she allegedly told him that she “meant it.”[141] She was promptly fired. In reviewing the case, the Supreme Court referred to different potential interpretations of the employee’s statements:

We note that the question whether McPherson “meant” the statement is [itself] ambiguous. Assuming that McPherson told Rankin she “meant it,” McPherson might think she had said that she “meant” that she disliked the President and would not mind if he were dead, while Rankin might believe that McPherson “meant” to indicate approval of, or in any event hope for, political assassination. This ambiguity makes evident the need for carefully conducted hearings and precise and complete findings of fact.[142]

Ambiguity of meaning calls for meticulous judicial factfinding. Such ambiguity may have many causes—from the use of vague or ambiguous words, to the deployment of expressive devices like sarcasm, the use of satire, or a hyperbole (which was allegedly the case in Rankin). Note that both the government’s and the employee’s interpretations touched on matters of public concern; but the difference between the two interpretations was important for the constitutionality of the firing: expressing hope that the President be assassinated was a very different matter than conveying dislike for the President in a highly offensive manner. The two may differ in terms of the value of the speech in question—and hence in regard to the government’s burden of proof—and may carry different implications regarding the employee’s unfitness for office (the employee was fired after “the Constable . . . decided that she was not a suitable employee to have in a law enforcement agency.”).[143] An employee who declares her support for presidential assassinations may indeed be unfit for public employment in a constable’s office, but an employee who expresses her dislike to the president may not.

In Moser v. Las Vegas Metropolitan Police Department (2021), an officer commented on Facebook that it was a “shame” that a suspect who had shot a police officer did not have any “holes” in him. The Las Vegas Police Department claimed that the officer’s comment made him unfit for the position of a SWAT officer because he advocated the unlawful use of deadly force.[144] The officer countered that his comment did no such thing and was merely a hyperbole.[145] Deciding the case required the court to determine which of the two interpretations was correct: “If he encouraged unlawful violence, it certainly reflects poorly on his ability to serve as a SWAT officer. But if his comment was merely a hyperbolic statement on a public matter, then it may not.”[146]

The correct interpretation, in the context of unfitness for office, is mostly a function of the meaning that the speaker intended to convey—known as the “speaker’s meaning” or the “speaker’s intent” theory of meaning.[147] Moser’s fitness for the job, like the fitness of the employee in Rankin, depended on the message that these employees actually meant to express—not on the government’s interpretations of their words, even if the interpretations were reasonable.[148]

The meaning understood by a reasonable hearer is not irrelevant to the Pickering balancing test: it may come to bear, for example, on claims of actual or expected disruption or on fitness to a job that requires communicating with the public. If a police officer says jokingly that “we should have planted drugs on them” and the sarcasm is lost on the hearers, there may be disruption to police operations as a result of public outrage or objections from colleagues. But then again, even in such circumstances it is important for courts to determine the speaker’s meaning—if only because punishing an employee for a message she did not intend to convey may impact other relevant factors, like proportionality between the intended speech and the adverse employment action.

2. The Question of Deference

Unlike the Waters analysis, which held that courts should defer to the government’s reasonable position on what had been said, when determining the meaning of the employee’s speech, the government’s reasonable interpretation of that speech is not entitled to judicial deference.

In Moser, a Ninth Circuit panel reversed the district court’s grant of summary judgment.[149] The district court’s adoption of the government’s understanding of the meaning of the speech in question was in error, said the court, even though there was no dispute regarding “the reasonableness of [the government’s] understanding of what he said.”[150] Another Ninth Circuit decision put things as follows:

Under the plurality opinion in Waters v. Churchill, when conducting Pickering balancing, “courts look to the facts as the employer reasonably found them to be.” (Emphasis in original). Unlike Waters, the facts of this case are not in dispute. Here the parties agree precisely on the expression at issue—the six writings. Connick and Waters teach that in such a circumstance the application of Pickering to settled facts is the province and duty of the court. We are not to defer to the governmental employer’s analysis of the facts, just the governmental employer’s reasonable determination of the facts.[151]

There are sound constitutional reasons for these positions. As we already saw, the Waters decision was premised on the claim that “[g]overnment employers should be allowed to use personnel procedures that differ from the evidentiary rules used by courts, without fear that these differences will lead to liability.”[152] If courts did not defer to government employer’s reasonable factual findings, said the opinion:

[I]t would force the government employer to come to its factual conclusions through procedures that substantially mirror the evidentiary rules used in court. The government manager would have to ask not what conclusions she, as an experienced professional, can draw from the circumstances, but rather what conclusions a jury would later draw. If she relies on hearsay, or on what she knows about the accused employee’s character, she must be aware that this evidence might not be usable in court. If she knows one party is, in her personal experience, more credible than another, she must realize that the jury will not share that personal experience. If she thinks the alleged offense is so egregious that it is proper to discipline the accused employee even though the evidence is ambiguous, she must consider that a jury might decide the other way.[153]

Factfinding procedures were at the heart of Waters’s requirement of deference to the government’s reasonable position. The question of meaning, however—did the employee use hyperbole? Was she being sarcastic? Did she endorse political violence?—involves not a factual question but an interpretation of facts. And here there is no risk of imposing any cumbersome procedures on government employers.

Additionally, as already mentioned, one main concern of free speech doctrine is over chilling protected expression: the worry that speakers would refrain from expressing themselves for “fear of retaliatory dismissal” even though their speech is protected.[154] If government employers could dismiss employees for every conceivable reasonable interpretation of their speech, that could have a substantial chilling effect on public employees’ speech. But employees’ speech could hardly be chilled for fear that the government may believe, in good faith, that they said something different than what they actually did. The concern with chilling legitimate expression is therefore hardly implicated in the Waters scenario but is very much at play in questions of interpretation.

In fact, if courts had to accept the government’s reasonable interpretation of employees’ speech, courts’ preoccupation with linguistic devices like hyperbole or irony would be incomprehensible—because literalist interpretations are almost always reasonable. And yet, courts are careful to emphasize that determinations of meaning involve sensitivity to such linguistic devices.[155]

V. Conclusion

While the Pickering balancing inquiry looks like a model of simplicity and flexibility—pitting the respective interests of the government as an employer against those of government employees—there remain substantial areas of uncertainty and confusion in some critical junctures of this important constitutional doctrine. As we saw, these include the question of speech spoken as a citizen on matters that are not of public concern—which this Article argues is constitutionally protected employee speech. We examined the unusual and fraught practice of judges evaluating the merit of speech when gauging the scope of its constitutional protections and offered some limiting strategies for this anomalous if understandable deviation from standard free speech principles—including viewpoint neutrality and the distinction between legitimate and illegitimate speech. We also explained why the proportionality of the government’s adverse employment action is a proper factor in the Pickering balancing inquiry and why speaker’s meaning is often the dispositive theory of meaning under Pickering. Finally, we saw why courts defer to the government’s reasonable interpretation of what had been said but not of what was meant by it.

American labor laws can be found on the federal, the state, and the local levels, so it can be hard to generalize, but it is fair to say that American employers enjoy greater latitude in dealing with their employees than do their Canadian or European counterparts. Putting aside self-imposed human resources codes, as well as prohibitions on certain forms of discrimination, U.S. employers are relatively free to retaliate against their employees’ speech. But when the employer is a government entity, the First Amendment kicks into gear: adverse employment actions in response to protected speech are subjected to the Pickering balancing test.[156] This is not a serendipity of public employment: these constitutional protections go to the heart of the government’s obligations to its citizens. The government’s ability to discipline or fire its employees for their speech can have a substantial impact on the quality and nature of government services. Thus, the free speech rights of government employees are of interest not only to the employees themselves, which number in the millions, or to government employers, who wish to control their operations effectively, but also to the public at large.

On the one hand, government employers must be able to assure that their employees are suitable for their jobs and that their expressive activities do not harm managerial control, intraoffice relationships, the trust of the public, and the provision of government services. Such supervisory powers are essential for any organization’s ability to function properly. But the power of government to punish or dismiss employees for expressive activities must be closely monitored—not only because, as the Supreme Court put it, “a citizen who works for the government is nonetheless a citizen”[157] but also because the public’s interest in just and efficient government services is as much at play in the employee’s side of the balancing scale.

When government employers silence employees’ speech, they may be silencing opinions that could benefit the public or, worse, that wish to blow the whistle on government mismanagement or corruption. Moreover, at times of high political polarization, government employers may seek to impose a political orthodoxy on their employees. Such potentially partisan purges may harm or skew important government services—from public education at schools and universities, to welfare payments, to public security, to the criminal justice system or voting.

This important body of constitutional law is long overdue for tighter structuring and more uniformity, especially at these hazardous times of intense political tensions. This Article essayed to point to some of the pathways in this restructuring process.


  1. Claudia Deane & John Gramlich, 2020 Election Reveals Two Broad Voting Coalitions Fundamentally at Odds, Pew Rsch. Ctr. (Nov. 6, 2020), https://www.pewresearch.org/fact-tank/2020/11/06/2020-election-reveals-two-broad-voting-coalitions-fundamentally-at-odds/ [https://perma.cc/9ZSM-HCZ6].

  2. See, e.g., Jennifer McCoy & Murat Somer, Toward a Theory of Pernicious Polarization and How It Harms Democracies: Comparative Evidence and Possible Remedies, 681 Annals Am. Acad. Pol. & Soc. Sci. 234, 244, 257–58, 263–64 (2019).

  3. See Christine Hauser, Whole Foods Punished Workers for Black Lives Matter Masks, Suit Says, N.Y. Times (July 21, 2020), https://www.nytimes.com/2020/07/21/business/whole-foods-black-lives-matter.html [https://perma.cc/4WGA-6FMB]; Katie Way, Some Black Lives Matter Supporters Say Their Stance Got Them Fired, Vice (Oct. 1, 2020, 1:10 PM), https://www.vice.com/en/article/g5p4qj/some-black-lives-matter-supporters-say-their-stance-got-them-fired [https://perma.cc/3ZL3-G5GH]; David Ford, Publix Fires Five Black Employees Over Local BLM Protests, WFDD (July 23, 2020, 11:09 PM), https://www.wfdd.org/story/publix-fires-five-black-employees-over-local-blm-protests [https://perma.cc/A467-ECRN]; UPS Employee Fired for Wearing Blackface as Part of Costume, AP News (Nov. 2, 2018), https://apnews.com/article/457f1d6a66ed478db02ac3f582d92830 [https://perma.cc/AFB6-56AG]; Wisconsin Woman Claims She Was Fired for Supporting President Trump on Social Media, Fox 8 News (Jan. 16, 2020, 9:12 PM), https://fox8.com/news/wisconsin-woman-claims-she-was-fired-for-supporting-president-trump-on-social-media/ [https://perma.cc/K7LZ-U6E4]; Wilson Wong, Virginia Police Officer Fired After Donating to Kyle Rittenhouse Defense Fund, NBC News (Apr. 21, 2021, 12:57 PM), https://www.nbcnews.com/news/us-news/virginia-police-officer-fired-after-donating-kyle-rittenhouse-defense-fund-n1264783 [https://perma.cc/2RZL-EGKJ]; Rebecca Tan, Maryland State Employee Fired After Supporting Kenosha Shooting Suspect Kyle Rittenhouse on Facebook, Wash. Post (Aug. 30, 2020, 3:32 PM), https://www.washingtonpost.com/local/md-politics/maryland-state-employee-fired-after-posting-statements-in-support-of-kenosha-shooting-suspect-kyle-rittenhouse-on-facebook/2020/08/30/f6c49912-ead5-11ea-b4bc-3a2098fc73d4_story.html [https://perma.cc/CW29-38BS]; see also Emily Ekins, Poll: 62% of Americans Say They Have Political Views They’re Afraid to Share, Cato Inst. (July 22, 2020), https://www.cato.org/survey-reports/poll-62-americans-say-they-have-political-views-theyre-afraid-share [https://perma.cc/Y8N2-4DSQ].

  4. See, e.g., Bennett v. Metro. Gov’t, 977 F.3d 530, 533–36 (6th Cir. 2020), cert. denied, 141 S. Ct. 2795 (2021) (firing employee for using racially charged language when discussing presidential election results on a social networking platform); Locurto v. Giuliani, 447 F.3d 159, 163–68 (2d Cir. 2006) (firing city firefighters and a police officer for participating in a racist parade float); City of San Diego v. Roe, 543 U.S. 77, 78–79 (2004) (per curiam) (firing police officer for selling homemade, sexually explicit videos at auction).

  5. Pickering v. Bd. of Educ., 391 U.S. 563, 568–69, 572–74 (1968); Connick v. Myers, 461 U.S. 138, 140, 142, 146 (1983).

  6. McAuliffe v. Mayor of New Bedford, 29 N.E. 517, 517 (Mass. 1892). The decision, which dismissed the lawsuit of a police officer who was fired for engaging in political activities, was made under the Massachusetts Constitution—if only because the free speech provision of the Federal Constitution was not yet made applicable to the states. Id. at 517–18; see Slaughter-House Cases, 83 U.S. 36, 74 (1872) (refusing to apply the Bill of Rights to the states). Justice Holmes had yet to move to the U.S. Supreme Court, and the U.S. Supreme Court had yet to incorporate the Free Speech Clause of the First Amendment vis-à-vis the states; but federal law at the time was in fact no different. See infra note 7; see also Gitlow v. New York, 268 U.S. 652, 666 (1926) (incorporating freedom of speech right from the First Amendment to the states via the Due Process Clause of the Fourteenth Amendment).

  7. Adler v. Bd. of Educ., 342 U.S. 485, 489–92, 496 (1952); see also Garner v. Bd. of Pub. Works, 341 U.S. 716, 720 (1951) (holding that the Federal Constitution does not forbid a municipality to require its employees to execute affidavits disclosing whether or not they are or ever have been members of the Communist Party); United Pub. Workers of Am. v. Mitchell, 330 U.S. 75, 78, 102–03 (1947) (upholding the constitutionality of the Hatch Act, which forbade certain public employees from taking “any active part in political management or in political campaigns” (quoting 18 U.S.C. § 61(h))); United States v. Wurzbach, 280 U.S. 396, 397–99 (1930) (summarily dismissing a constitutional challenge to a federal statute forbidding congressmen or any officer of the United States “to directly or indirectly solicit [or] receive . . . any . . . contribution for any political purpose whatever” (quoting 18 U.S.C. § 208)).

  8. See Keyishian v. Bd. of Regents, 385 U.S. 589, 605–06, 608–10 (1967) (holding that, notwithstanding Adler, mere membership in a political group is insufficient to exclude from employment); Pickering, 391 U.S. at 568 (requiring courts to balance the free speech interests of the public employee against the interests of the government employer). These decisions followed earlier precedents that restricted the ability of the government to impose various speech-related limitations on public employment. See Wieman v. Updegraff, 344 U.S. 183, 186, 192 (1952) (holding that a state could not require its employees to establish their loyalty by extracting an oath denying past affiliation with communism); Cafeteria & Rest. Workers Union v. McElroy, 367 U.S. 886, 897 (1961) (recognizing that the government could not deny employment because of previous membership in a particular party); see also Shelton v. Tucker, 364 U.S. 479, 480, 490 (1960) (noting that a statute requiring disclosure of all of a teacher’s affiliations went “far beyond what might be justified in the exercise of the State’s legitimate inquiry into the fitness and competency of its teachers”); Torcaso v. Watkins, 367 U.S. 488, 489, 496 (1961) (finding Maryland’s test for public office requiring a declaration in a belief in God unconstitutional); Cramp v. Bd. of Pub. Instruction, 368 U.S. 278, 279, 287 (1961) (finding a Florida statute requiring public employees to swear that they had never lent “aid, support, advice, counsel or influence to the Communist Party” unconstitutional (quoting Fla. Stat. § 876.05 (2011))). By the time Sherbert v. Verner was decided, it was already “too late in the day to doubt that the liberties of religion and expression may be infringed by the denial of or placing of conditions upon a benefit or privilege.” Sherbert v. Verner, 374 U.S. 398, 404 (1963). It was therefore no surprise when, in Keyishian, the Court invalidated New York statutes barring employment on the basis of membership in “subversive” organizations, observing that the theory that public employment that may be denied altogether may be subjected to any conditions, regardless of how unreasonable, had been uniformly rejected. Keyishian, 385 U.S. at 605–06. “In all of these cases, the precedents in which Pickering is rooted, the invalidated statutes and actions sought to suppress the rights of public employees to participate in public affairs.” Connick, 461 U.S. at 144–45.

  9. Pickering, 391 U.S. at 568.

  10. Connick, 461 U.S. at 139.

  11. Id. at 140–41.

  12. Id. at 141.

  13. Id. at 147. The Court was careful to note that there may be other constitutional free speech protections that apply—constitutional protections for allegedly libelous speech, for example. Id. But so far as the specific doctrine governing public employees’ speech was concerned, no protections were afforded to speech that was not of public concern. Id.

  14. Id. at 148.

  15. Id. at 149–54.

  16. Garcetti v. Ceballos, 547 U.S. 410, 412 (2006).

  17. Id. at 413–15; see also id. at 443 n.14 (Souter, J., dissenting).

  18. Id. at 415, 421.

  19. Id. at 422.

  20. Id. at 418.

  21. Id. at 426–27 (Stevens, J., dissenting); id. at 427–31, 444 (Souter, J., dissenting); id. at 449–50 (Breyer, J., dissenting).

  22. See id. at 432–33 (Souter, J., dissenting); see, e.g., Nixon v. Fitzgerald, 457 U.S. 731, 733–34 (1982) (firing a management analyst with the Department of the Air Force over his testimony before a congressional subcommittee about cost overruns).

  23. Garcetti, 547 U.S. at 433 (Souter, J., dissenting) (quoting City of San Diego v. Roe, 543 U.S. 77, 82 (2004) (per curiam) (citation omitted)).

  24. Pickering v. Bd. of Educ., 391 U.S. 563, 566, 574–75 (1968).

  25. Id. at 572–75.

  26. Roe, 543 U.S. at 82 (per curiam); see also United States v. Nat’l Treasury Emps. Union, 513 U.S. 454, 470 (1995) (“[L]arge-scale disincentive to Government employees’ expression also imposes a significant burden on the public’s right to read and hear what the employees would otherwise have written and said.”).

  27. Garcetti, 547 U.S. at 419.

  28. Id. at 446 (Breyer, J., dissenting).

  29. Id. at 425 (“The dictates of sound judgment are reinforced by the powerful network of legislative enactments—such as whistle-blower protection laws and labor codes—available to those who seek to expose wrongdoing. See, e.g., 5 U.S.C.  § 2302(b)(8); Cal. Gov’t. Code Ann.  § 8547.8 (West 2005); Cal. Lab. Code Ann. § 1102.5 (West Supp. 2006).”).

  30. Connick v. Myers, 461 U.S. 138, 147 (1983); see also Rankin v. McPherson, 483 U.S. 378, 384 n.7 (1987) (“Even where a public employee’s speech does not touch upon a matter of public concern, that speech is not ‘totally beyond the protection of the First Amendment,’ Connick v. Myers, 461 U.S., at 147, but ‘absent the most unusual circumstances a federal court is not the appropriate forum in which to review the wisdom of a personnel decision taken by a public agency allegedly in reaction to the employee’s behavior.’ Ibid.”).

  31. See Connick, 461 U.S. at 147; Garcetti, 547 U.S. at 426.

  32. See Lane v. Franks, 573 U.S. 228, 239–40 (2014) (“[T]he mere fact that a citizen’s speech concerns information acquired by virtue of his public employment does not transform that speech into employee—rather than citizen—speech.”).

  33. See supra Sections I.A.1.i–ii.

  34. Connick, 461 U.S. at 148 n.7; see also Rankin, 483 U.S. at 385 n.8 (“Even assuming that the District Court can be viewed to have made any findings of fact on the public concern issue, it is unclear to what extent that issue presents a question of fact at all.”); Posey v. Lake Pend Oreille Sch. Dist. No. 84, 546 F.3d 1121, 1127 (9th Cir. 2008).

  35. City of San Diego v. Roe, 543 U.S. 77, 83–84 (2004) (per curiam).

  36. Id. at 83; see, e.g., Rowland v. Mad River Loc. Sch. Dist., 470 U.S. 1009, 1009–10, 1012 (1985) (Brennan, J., dissenting) (involving a nontenured school counselor who was fired after telling her secretary and fellow teachers, in private conversations, that she was bisexual, where the Sixth Circuit ruled that the petitioner’s speech was not about “a matter of public concern,” a determination disputed by the two Justices (quoting Rowland v. Mad River Loc. Sch. Dist., 730 F.2d 444, 451 (6th Cir. 1984), cert. denied, 470 U.S. 1009)); see also Flanagan v. Munger, 890 F.2d 1557, 1562–63 (10th Cir. 1989) (“[I]t is nearly impossible to logically apply the public concern test to the present case . . . The Supreme Court requires that the employee’s speech comment ‘upon matters of public concern.’ It is difficult to comprehend how each of the officer’s owning of a one-quarter interest in a video store which rents a small portion of sexually explicit videos is making a ‘comment’ on any subject, especially a subject of public concern.” (quoting Pickering v. Bd. of Educ., 391 U.S. 563, 568 (1967))).

  37. Connick, 461 U.S. at 146–48.

  38. Dishnow v. Sch. Dist. of Rib Lake, 77 F.3d 194, 197 (7th Cir. 1996).

  39. Rankin v. McPherson, 483 U.S. 378, 387 (1987) (quoting N.Y. Times Co. v. Sullivan, 376 U.S. 254, 270 (1964)).

  40. Id. at 381, 388.

  41. Connick, 461 U.S. at 147–48.

  42. Id. at 148.

  43. Id. at 147–49.

  44. Id. at 148 n.8.

  45. Id. at 160, 163 (Brennan, J., dissenting). The dissent also pointed to Supreme Court precedent recognizing speech to be on matters of public concern even where it was conveyed privately. Id. at 159 (concluding that a “public employee who arranges to communicate privately with his employer rather than to spread his views before the public” engages in speech about matters of public concern (quoting Givhan v. W. Line Consol. Sch. Dist., 439 U.S. 410, 415–16 (1979))).

  46. Id. at 160 n.2.

  47. Id. at 165, 169 (“The proper means to ensure that the courts are not swamped with routine employee grievances mischaracterized as First Amendment cases is not to restrict artificially the concept of ‘public concern,’ but to require that adequate weight be given to the public’s important interests in the efficient performance of governmental functions . . . .”).

  48. See id. at 154 (majority opinion).

  49. Id. at 149, 154.

  50. See Gilbrook v. City of Westminster, 177 F.3d 839, 866 (9th Cir. 1999) (finding that “[a]n employee’s motivation and the audience chosen for the speech also are relevant to the public-concern inquiry”); see also Liverman v. City of Petersburg, 844 F.3d 400, 410 (4th Cir. 2016) (“[P]laintiffs’ statements stand in stark contrast to the sort of ‘individualized concerns’ this court has characterized as personal grievances.” (quoting Brooks v. Arthur, 685 F.3d 367, 374 (4th Cir. 2012))).

  51. Johnson v. Multnomah Cnty., 48 F.3d 420, 421 (9th Cir. 1995).

  52. Id. at 425 (first quoting Havekost v. U.S. Dep’t of Navy, 925 F.2d 316, 319 (9th Cir. 1991); and then quoting McKinley v. City of Eloy, 705 F.2d 1110, 1114 (9th Cir. 1983)). The court went on to state:

    This case is not like Connick, which involved an employment dispute related to internal administrative procedures within a prosecutor’s office. . . . Rather, it is similar to Givhan v. Western Line Consolidated School District, . . . in which a teacher had been fired after complaining to her school’s principal about school policies and practices which she perceived as racially discriminatory.

    Id.

  53. Indeed, the Connick decision relied on the particularities of the employment dispute before it. See Connick, 461 U.S. at 154 (“[I]t would indeed be a Pyrrhic victory for the great principles of free expression if the Amendment’s safeguarding of a public employee’s right, as a citizen, to participate in discussions concerning public affairs were confused with the attempt to constitutionalize the employee grievance that we see presented here.” (emphasis added)).

  54. A Supreme Court precedent, Givhan v. Western Line Consolidated School District (discussed by both the majority and the dissent in Connick), held that First Amendment protections apply when a public-school employee complained to her employer about the school’s racially discriminatory policies. Givhan v. W. Line Consol. Sch. Dist., 439 U.S. 410, 411–14 (1979). See generally Connick, 461 U.S. 138. Although the complaint was “not tied to a personal employment dispute,” it is conceivable that even were it tied, such speech of “a matter inherently of public concern” (as the Connick majority described it) may very well enjoy constitutional protections. Id. at 148 n.8.

  55. The term “private,” when referring to speech, has been used by the Supreme Court to refer to speech that is not on matters of public concern. See, e.g., Waters v. Churchill, 511 U.S. 661, 674 (1994) (“[W]e have refrained from intervening in government employer decisions that are based on speech that is of entirely private concern.”); Connick, 461 U.S. at 147 (“We in no sense suggest that speech on private matters falls into one of the narrow and well-defined classes of expression which carries so little social value, such as obscenity, that the State can prohibit and punish such expression by all persons in its jurisdiction.”). But the Court also used the term to refer to speech that is conveyed in a private setting. See, e.g., Rankin v. McPherson, 483 U.S. 378, 389–91 (1987) (“Where, as here, an employee serves no confidential, policymaking, or public contact role, the danger to the agency’s successful functioning from that employee’s private speech [on matters of public concern] is minimal.”). The term is used in the first sense in this Article.

  56. For example, the Michigan Civil Service Commission Rules provide that “[a]n appointing authority may discipline a classified employee for just cause” and that “[j]ust cause includes, but is not limited to . . . [c]onduct unbecoming a state employee.” Michigan Civil Service Commission, Civil Service Rules, rule no. 2-6.1 (2022).

  57. See, e.g., Keyishian v. Bd. of Regents, 385 U.S. 589, 605 (1967) (“[C]onstitutional doctrine which has emerged since that decision has rejected its major premise. That premise was that public employment, including academic employment, may be conditioned upon the surrender of constitutional rights . . . .”).

  58. See, e.g., W. Va. State Bd. of Educ. v. Barnette, 319 U.S. 624, 634 (1943) (noting that the First Amendment “guards the individual’s right to speak his own mind”). For opposing views about which of the two—the public or the personal—is the chief original concern of the First Amendment, see, e.g., Jud Campbell, Natural Rights and the First Amendment, 127 Yale L.J. 246, 264–87 (2017). For the opposite view, see, e.g., Cass R. Sunstein, Democracy and the Problem of Free Speech 4–8 (1993).

  59. Elizabeth Dippold et al., Annual Survey of Public Employment & Payroll Summary Report: 2020 1 (2021), https://www.census.gov/content/dam/Census/library/publications/2021/econ/2020_summary_brief.pdf [https://perma.cc/LR6L-UTJR] (“In March 2020, state and local governments employed 19.8 million people . . . .”); see also U.S. Bureau Lab. Stat., Household Data Annual Averages 2, https://www.bls.gov/cps/cpsaat01.pdf [https://perma.cc/DN2J-KY4C] (showing a total civilian workforce of 160,742,000).

  60. It is true, of course, that because the employer’s interest is at its lowest ebb when it comes to speech spoken as a citizen on a matter of no public concern, firing employees for such speech may run into difficulties in relation to employment contracts and other employment regulations; but that may be true about speech on matters of public concern and, in any event, may not protect the interests with which this body of constitutional law is concerned: namely, the interest in protecting the ability of public employees to speak freely on matters of their choosing.

  61. Locurto v. Giuliani, 447 F.3d 159, 174–75 (2d Cir. 2006).

  62. See, e.g., Garcetti v. Ceballos, 547 U.S. 410, 418 (2006) (“Pickering and the cases decided in its wake identify two inquiries to guide interpretation of the constitutional protections accorded to public employee speech. The first requires determining whether the employee spoke as a citizen on a matter of public concern. . . . If the answer is no, the employee has no First Amendment cause of action based on his or her employer’s reaction to the speech.”).

  63. Pickering v. Bd. of Educ., 391 U.S. 563, 568, 574 (1968).

  64. Connick v. Myers, 461 U.S. 138, 146 (1983) (footnote omitted).

  65. Rankin v. McPherson, 483 U.S. 378, 384 n.7 (1987) (“Even where a public employee’s speech does not touch upon a matter of public concern, that speech is not ‘totally beyond the protection of the First Amendment.’ . . . [B]ut ‘absent the most unusual circumstances a federal court is not the appropriate forum in which to review the wisdom of a personnel decision taken by a public agency allegedly in reaction to the employee’s behavior.’” (quoting Connick, 461 U.S. at 147)); see also id. at 378 (Scalia, J., dissenting) (“[W]e have held that the First Amendment’s protection against adverse personnel decisions extends only to speech on matters of ‘public concern[.]’ . . . If, but only if, an employee’s speech falls within this category [is the speech constitutionally protected].” (quoting Connick, 461 U.S. at 147–49)).

  66. Waters v. Churchill, 511 U.S. 661, 674 (1994) (plurality opinion) (“[W]e have refrained from intervening in government employer decisions that are based on speech that is of entirely private concern. Doubtless some such speech is sometimes nondisruptive; doubtless it is sometimes of value to the speakers and the listeners. But we have declined to question government employers’ decisions on such matters.”).

  67. Garcetti, 547 U.S. at 419 (“So long as employees are speaking as citizens about matters of public concern, they must face only those speech restrictions that are necessary for their employers to operate efficiently and effectively.”).

  68. Two Supreme Court Justices explicitly stated that Connick—to which all subsequent Supreme Court cases making the point refer—actually left open the constitutional status of private speech. See Rowland v. Mad River Loc. Sch. Dist., 470 U.S. 1009, 1013 (1985) (Brennan, J., dissenting from a denial of certiorari) (“Moreover, even if petitioner’s speech did not . . . touch upon a matter of public concern, there remains a substantial constitutional question, reserved in Connick, whether it lies ‘totally beyond the protection of the First Amendment’ given its nondisruptive character.” (quoting Connick, 461 U.S. at 147)). Moreover, the circuit court cases to which Connick cited in support of the proposition that employee speech not on matters of public concern is unprotected also concerned only speech spoken as an employee, not as a citizen. Connick, 461 U.S. at 146 n.6. Both Clark v. Holmes, 474 F.2d 928, 929–30 (7th Cir. 1972) (per curiam), cert. denied, 411 U.S. 972 (1973), and Schmidt v. Fremont County School District, 558 F.2d 982, 983–84 (10th Cir. 1977), involved speech spoken as an employee. See also Hawkins v. Dep’t of Pub. Safety & Corr. Servs., 602 A.2d 712, 722 (1992) (Bell, J., dissenting).

  69. City of San Diego v. Roe, 543 U.S. 77, 78, 84 (2004) (per curiam).

  70. Id. at 78–79.

  71. Id. at 79.

  72. Id. at 77–78.

  73. Id. at 82–84 (quoting Connick, 461 U.S. at 143).

  74. Id. at 77, 81, 84.

  75. Id. at 84–85.

  76. For the proposition that such speech is unprotected, see, e.g., Laurence H. Tribe, American Constitutional Law 930–31 (2d ed. 1988) (“Only when a government employee engages in expression addressed to ‘matters of public concern’ does the first amendment protect him from termination.”); Toni M. Massaro, Significant Silences: Freedom of Speech in the Public Sector Workplace, 61 S. Cal. L. Rev. 1, 14 (1987) (“If the speech does not relate to a matter of public concern, then a federal court will not analyze the reasons for an employer’s decision to discipline the worker.”); Rowland v. Mad River Loc. Sch. Dist., 730 F.2d 444, 448–49 (6th Cir. 1984), cert. denied, 470 U.S. 1009 (1985) (holding that a guidance counselor being fired for telling her secretary and fellow teachers, in private conversations, that she was bisexual deserved no constitutional protection); Pruitt v. Howard Cnty. Sheriff’s Dep’t, 623 A.2d 696, 702 (Md. Ct. Spec. App. 1993) (“Because we find that the Pruitts’ speech did not concern a matter of public issue we do not reach the second prong of the analysis; we need not balance protected speech against the Sheriff’s department’s interest in protecting its credibility.”); Piscottano v. Murphy, No. 3:04CV682, 2005 WL 1424394, at *6 (D. Conn. June 9, 2005) (“Nothing in either Roe or [Treasury Union] suggests that the Supreme Court in those cases had decided to categorically exempt ‘off-duty’ public employee expression from the public concern requirement of Connick/Pickering.”). But see, e.g., Festa v. Westchester Med. Ctr. Health Network, 380 F.Supp. 3d 308, 316–18 (S.D.N.Y. 2019) (“The Parties agree that Plaintiff’s speech did not touch on a matter of public concern. . . . [E]ven assuming arguendo that Plaintiff’s Facebook post was protected by the First Amendment, Defendants’ ‘reasonable concern for the potentially disruptive effects’ of her speech could outweigh Plaintiff’s expressive interest.” (quoting Locurto v. Giuliani, 447 F.3d 159, 183 (2d Cir.2006))); Dible v. City of Chandler, 515 F.3d 918, 927 (9th Cir. 2008) (“If we determined that Ronald Dible’s activities were unrelated to his public employment, we would also have to apply a balancing test. Interestingly enough, it is not entirely clear whether the public concern concept would be a necessary threshold to that balancing. In Roe . . . , the Supreme Court did not exactly say that the public concern concept must be considered . . . .”); Locurto, 447 F.3d at 175 (“It is more sensible . . . to treat off-duty, non-work-related speech as presumptively entitled to First Amendment protection regardless of whether, as a threshold matter, it may be characterized as speech on a matter of public concern.”); Jean-Gilles v. Cnty. of Rockland, 463 F.Supp. 2d 437, 450 (S.D.N.Y. 2006) (“If the plaintiff can demonstrate his speech centered on a matter of public concern—or if the plaintiff’s speech concerned off-duty speech unrelated to his employment—the court must balance the parties’ competing interests.”); Hawkins v. Dep’t of Pub. Safety & Corr. Servs., 602 A.2d 712, 718 (Md. 1992) (“On a literal application of the Pickering–Connick rule, [the speech enjoyed no constitutional protections] because the threshold requirement of speech on a matter of public concern has not been met. Some doubt necessarily remains whether a literal application of the Pickering–Connick rule is the appropriate analysis in the instant matter where the speech was away from the job site, outside of business hours, and, in terms, did not discuss or comment upon any aspect of the public employment.”); Flanagan v. Munger, 890 F.2d 1557, 1562–65 (10th Cir. 1989) (“[T]he Connick public concern test is intended to weed out speech by an employee speaking as an employee upon matters only of personal interest. . . . The purpose of the public concern test is to avoid raising personal personnel grievances to constitutional cases. . . . In those cases [involving employee expression that is not at work nor about work] it makes little sense to ask whether this speech is of public concern.”).

  77. See, e.g., Dible, 515 F.3d at 926–27; Jean-Gilles, 463 F. Supp. 2d at 449–50.

  78. The location of the speech is often mentioned by courts, though its significance is highly context specific. See, e.g., Rankin v. McPherson, 483 U.S. 378, 388–89 (displaying a public employer taking adverse employment action against a public employee for making a statement that was unrelated to employment but was made at work); Givhan v. W. Line Consol. Sch. Dist., 439 U.S. 410, 415 n.4 (“When a government employee personally confronts his immediate superior, the employing agency’s institutional efficiency may be threatened not only by the content of the employee’s message but also by the manner, time, and place in which it is delivered.”); Connick, 461 U.S. at 152 (“Also relevant is the manner, time, and place in which the questionnaire was distributed.”). In principle, there is no doubt that employers have a greater interest in controlling on-site speech and employees have a greater interest in their freedom of speech in regard to off-site speech. Still, courts have upheld numerous adverse employment actions for off-work speech when such speech resulted either in demonstration of fitness for office or in disruption. See, e.g., Roe, 543 U.S. at 84–85. Moreover, the rise of social media means that speech expressed from the privacy of the home can prove as disruptive as speech expressed at the office. As the Supreme Court put the point recently in a case involving the constitutionality of public schools to discipline off-campus student speech, “[t]he school’s regulatory interests remain significant in some off-campus circumstances.” Mahanoy Area Sch. Dist. v. B.L., 141 S. Ct. 2038, 2045 (2021).

  79. See supra Section II.A.2.ii (dealing with the importance of the speech in question for the balancing inquiry).

  80. Pickering v. Bd. of Educ., 391 U.S. 563, 568 (1968).

  81. Id.

  82. Id. at 572.

  83. Connick, 461 U.S. at 150.

  84. Bauer v. Sampson, 261 F.3d 775, 784 (9th Cir. 2001).

  85. Connick, 461 U.S. at 150.

  86. Rankin v. McPherson, 483 U.S. 378, 388 (1987) (citing Pickering, 391 U.S. at 570–73).

  87. Connick, 461 U.S. at 151–52.

  88. Id. at 154 (emphasis added); see also Waters v. Churchill, 511 U.S. 661, 673 (1994) (“[W]e have given substantial weight to government employers’ reasonable predictions of disruption.”).

  89. See Givhan v. W. Line Consol. Sch. Dist., 439 U.S. 410, 415 n.4 (1979) (“When a government employee personally confronts his immediate superior, the employing agency’s institutional efficiency may be threatened not only by the content of the employee’s message but also by the manner, time, and place in which it is delivered.”).

  90. See Waters, 511 U.S. at 673 (“[W]e have consistently given greater deference to government predictions of harm used to justify restriction of employee speech than to predictions of harm used to justify restrictions on the speech of the public at large.”).

  91. Id. at 674; see Nunez v. Davis, 169 F.3d 1222, 1229 (9th Cir. 1999) (“[R]eal, not imagined, disruption is required.” (quoting McKinley v. City of Eloy, 705 F.2d 1110, 1115 (9th Cir. 1983))). A disruption claim must be supported by some evidence, not rank speculation or bald allegation. See Gustafson v. Jones, 290 F.3d 895, 909 (7th Cir. 2002) (“Pickering balancing is not an exercise in judicial speculation.”). An employer must provide some evidence by which we can measure whether its claims of disruption are reasonable. See Craig v. Rich Twp. High Sch. Dist. 227, 736 F.3d 1110, 1119 (7th Cir. 2013) (“[A]n employer’s assessment of the possible interference caused by the speech must be reasonable—‘the predictions must be supported with an evidentiary foundation and be more than mere speculation.’” (quoting Chaklos v. Stevens, 560 F.3d 705, 715 (7th Cir. 2009)); Nichols v. Dancer, 657 F.3d 929, 933–34 (9th Cir. 2011) (“[A]n employer cannot prevail under Pickering based on mere speculation that an employee’s conduct will cause disruption.”); Moser v. Las Vegas Metro. Police Dep’t, 984 F.3d 900, 909, 910 n.8 (9th Cir. 2021) (“[T]he government cannot rely on mere speculation that an employee’s speech will cause disruption. . . . [T]he government had to provide some evidence to support its prediction.”).

  92. Kinney v. Weaver, 367 F.3d 337, 363 (5th Cir. 2004).

  93. Pickering v. Bd. of Educ., 391 U.S. 563, 572–73 (1968).

  94. See Rankin v. McPherson, 483 U.S. 378, 392 (1987).

  95. Id. at 389.

  96. Id. at 390 (“The burden of caution employees bear with respect to the words they speak will vary with the extent of authority and public accountability the employee’s role entails.”).

  97. Connick v. Myers, 461 U.S. 138, 142 (1983).

  98. Pickering, 391 U.S. at 573; see also City of San Diego v. Roe, 543 U.S. 77, 82 (2004) (“Underlying the decision in Pickering is the recognition that public employees are often the members of the community who are likely to have informed opinions as to the operations of their public employers, operations which are of substantial concern to the public. Were they not able to speak on these matters, the community would be deprived of informed opinions on important public issues. . . . The interest at stake is as much the public’s interest in receiving informed opinion as it is the employee’s own right to disseminate it.” (citing Pickering, 391 U.S. at 572)).

  99. See, e.g., Police Dep’t of Chicago v. Mosley, 408 U.S. 92, 95–96 (1972) (“[The] government may not grant the use of a forum to people whose views it finds acceptable, but deny use to those wishing to express less favored or more controversial views.”).

  100. Connick, 461 U.S. at 152; see also Lane v. Franks, 573 U.S. 228, 242 (2014).

  101. Moser v. Las Vegas Metro. Police Dep’t, 984 F.3d 900, 905–06 (9th Cir. 2011) (citing Connick, 461 U.S. at 150).

  102. Id. at 906–07; see also Berger v. Battaglia, 779 F.2d 992, 999 (4th Cir. 1985) (“The court found the artistic expression involved [which consisted of performances in blackface] to be of less value than that found in speech on political matters and public affairs, the ‘core’ values protected by the first amendment. We do not disagree . . . .”).

  103. See, e.g., Dible v. City of Chandler, 515 F.3d 918, 928 (9th Cir. 2008) (“[I]t is a bit difficult to give [sexual expression] the same weight as the right to engage in political debate . . . .”); see also Roe, 543 U.S. at 84 (per curiam) (noting that sexually explicit tapes showing police officer engaged in sexual conduct were not matters of public concern).

  104. Connick, 461 U.S. at 145 (quoting NAACP v. Claiborne Hardware Co., 458 U.S. 886, 913 (1982)) (citing Carey v. Brown, 447 U.S. 455, 467 (1980)).

  105. Hawkins v. Dep’t of Pub. Safety & Corr. Servs., 602 A.2d 712, 719–20 (Md. 1992).

  106. See, e.g., McIntyre v. Ohio Elections Comm’n, 514 U.S. 334, 347 (1995) (“[A]dvocacy of a politically controversial viewpoint . . . is the essence of First Amendment expression.”).

  107. See, e.g., Young v. Am. Mini Theatres, Inc., 427 U.S. 50, 86 (1976) (Stewart, J., joined by Brennan, Marshall, and Blackmun, JJ., dissenting) (“[The Court] stands ‘Voltaire’s immortal comment’ on its head. For if the guarantees of the First Amendment were reserved for expression that more than a ‘few of us’ would take up arms to defend, then the right of free expression would be defined and circumscribed by current popular opinion.”).

  108. Hawkins, 602 A.2d at 724 n.5 (Bell, J., dissenting).

  109. Rankin v. McPherson, 483 U.S. 378, 384 (1987).

  110. First Nat’l Bank of Boston v. Bellotti, 435 U.S. 765, 785–86 (1978); see also Tinker v. Des Moines Indep. Cmty. Sch. Dist., 393 U.S. 503, 509 (1969) (“[F]or the State in the person of school officials to justify prohibition of a particular expression of opinion, it must be able to show that its action was caused by something more than a mere desire to avoid the discomfort and unpleasantness that always accompany an unpopular viewpoint.”).

  111. The Supreme Court denied that the Pickering analysis in fact sanctions viewpoint-based discrimination by allowing adverse employment actions that target disruptive (as opposed to nondisruptive) speech:

    Our Pickering cases only permit the Government to take adverse action based on employee speech that has adverse effects on “the interest of the State, as an employer, in promoting the efficiency of the public services it performs through its employees.” That certain messages may be more likely than others to have such adverse effects does not render Pickering’s restriction on speech viewpoint based.

    United States v. Nat’l Treasury Emps. Union, 513 U.S. 454, 467 n.11 (1995) (quoting Pickering v. Bd. of Educ., 391 U.S. 563, 568 (1968)).

  112. See, e.g., Austin v. Univ. of Fla. Bd. of Trs., 580 F. Supp. 3d 1137, 1170 (N.D. Fla. 2022), appeal filed Austin v. Univ. of Fla. Bd. of Trs., No. 22-10448 (11th Cir. 2022); see also U.S. Civ. Serv. Comm’n v. Nat’l Ass’n of Letter Carriers, 413 U.S. 548, 564 (1973) (“The restrictions so far imposed on federal employees are not aimed at particular . . . points of view . . . .”).

  113. See, e.g., Austin, 580 F. Supp. 3d at 1154.

  114. Connick v. Myers, 461 U.S. 138, 148 n.8 (“Mrs. Givhan’s right to protest racial discrimination—a matter inherently of public concern—is not forfeited by her choice of a private forum.” (quoting Givhan v. W. Line Consol. Sch. Dist., 439 U.S. 410, 415–16 (1979))).

  115. Festa v. Westchester Med. Ctr. Health Network, 380 F. Supp. 3d 308, 321 (S.D.N.Y. 2019).

  116. The concept of illegitimate speech is not foreign to First Amendment jurisprudence. See, e.g., Bantam Books, Inc. v. Sullivan, 372 U.S. 58, 66 (1963) (“The separation of legitimate from illegitimate speech calls for . . . sensitive tools . . . .”); Se. Promotions, Ltd. v. Conrad, 420 U.S. 546, 559 (1975) (“It is always difficult to know in advance what an individual will say, and the line between legitimate and illegitimate speech is often so finely drawn that the risks of freewheeling censorship are formidable.”); Lorillard Tobacco Co. v. Reilly, 533 U.S. 525, 565 (2001) (“A careful calculation of the costs of a speech regulation does not mean that a State must demonstrate that there is no incursion on legitimate speech . . . .”); Massachusetts v. Oakes, 491 U.S. 576, 577 (1989) (“The overbreadth doctrine serves to protect constitutionally legitimate speech . . . .”); Ashcroft v. ACLU, 542 U.S. 656, 666 (2004) (“[I]t is important to ensure that legitimate speech is not chilled or punished.”); Elonis v. United States, 575 U.S. 723, 767 (2015) (Thomas, J., dissenting) (“There is always a risk that a criminal threat statute may be deployed by the Government to suppress legitimate speech.”).

  117. See Rankin v. McPherson, 483 U.S. 378, 387 (1987) (“The inappropriate or controversial character of a statement is irrelevant . . . . ‘[D]ebate on public issues should be uninhibited, robust, and wide-open, and . . . may well include vehement, caustic, and sometimes unpleasantly sharp attacks . . . .’” (quoting N.Y. Times Co. v. Sullivan, 376 U.S. 254, 270 (1964))). Note, however, that offensiveness may be relevant to another factor in the Pickering balancing—viz, the government’s claims of operational disruptions.

  118. O’Hare Truck Serv., Inc. v. City of Northlake, 518 U.S. 712, 718–19 (1996) (quoting Branti v. Finkel, 445 U.S. 507, 518 (1980)).

  119. Keyishian v. Bd. of Regents, 385 U.S. 589, 609–10, 613 (1967).

  120. Rankin, 483 U.S. at 381–82, 392.

  121. For example, judges have disagreed on the value to be accorded the production of sexually explicit materials under Pickering. Dible v. City of Chandler, 515 F.3d 918, 932–33 (9th Cir. 2008) (Canby, J., concurring) (“I recognize that pornography, although apparently popular, is not a very respected subject of First Amendment protection in many quarters. The majority opinion here reflects that distaste, variously characterizing Dible’s expressive activities as ‘vulgar,’ ‘indecent,’ ‘sleazy,’ and ‘disreputable.’ But vigorous enforcement of the free speech guarantee of the First Amendment often requires that we protect speech that many, even a majority, find offensive.”).

  122. See, e.g., Snyder v. Phelps, 562 U.S. 443, 448–50, 461 (2011) (protecting speech celebrating the death of American soldiers at the funeral of a soldier killed in Iraq); United States v. Stevens, 559 U.S. 460, 481–82 (2010) (protecting videos of animal cruelty made for commercial purposes); Collin v. Smith, 578 F.2d 1197, 1200, 1210 (7th Cir. 1978) (protecting Nazi speech).

  123. City of San Diego v. Roe, 543 U.S. 77, 83–84 (2004) (“[P]ublic concern is something that is a subject of legitimate news interest; that is, a subject of general interest and of value and concern to the public at the time of publication.”).

  124. Connick v. Myers, 461 U.S. 138, 152 (1983).

  125. Berger v. Battaglia, 779 F.2d 992, 1001 (4th Cir. 1985).

  126. Dible, 515 F.3d at 933–34 (Canby, J., concurring).

  127. Compare Berger, 779 F.2d at 1001 (“Historically, one of the most persistent and insidious threats to first amendment rights has been that posed by the ‘heckler’s veto,’ imposed by the successful importuning of government to curtail ‘offensive’ speech at peril of suffering disruptions of public order.” (internal citations omitted)), and Flanagan v. Munger, 890 F.2d 1557, 1566 (10th Cir. 1989) (“The department cannot justify disciplinary action against plaintiffs simply because some members of the public find plaintiffs’ speech offensive and for that reason may not cooperate with law enforcement officers in the future. The Supreme Court has squarely rejected what it refers to as the ‘heckler’s veto’ as a justification for curtailing ‘offensive’ speech in order to prevent public disorder.”), with Dible, 515 F.3d at 928–29 (“We are not gallied by the [plaintiff’s claim that he] is being subjected to some kind of heckler’s veto. Worries about a heckler’s veto have generally dealt with the restriction of a citizen’s speech based upon the anticipated disorderly reaction by members of an audience. Those worries do not directly relate to the wholly separate area of employee activities that affect the public’s view of a governmental agency in a negative fashion, and, thereby, affect the agency’s mission.”).

  128. Munroe v. Cent. Bucks Sch. Dist., 805 F.3d 454, 472 (3d Cir. 2015) (first quoting Miller v. Clinton Cnty., 544 F.3d 542, 549 n.2 (3d Cir. 2008); and then quoting Dougherty v. Sch. Dist. of Phila., 772 F.3d 979, 991 (3d Cir. 2014)); see also Locurto v. Giuliani, 447 F.3d 159, 183 (2d Cir. 2006) (“A plaintiff alleging an adverse employment action in violation of the First Amendment can prevail even over an employer who took the action for legitimate reasons if the employee’s expressive interests outweigh the employer’s interest in preventing disruption.”); Curtis v. Okla. City Pub. Schs. Bd. of Educ., 147 F.3d 1200, 1213 (10th Cir. 1998) (“[T]he employer’s burden to justify its restriction on speech increases in proportion to the value of that speech in the public debate.” (quoting Ware v. Unified Sch. Dist. No. 492, 881 F.2d 906, 910 (10th Cir. 1989))); McKinley v. City of Eloy, 705 F.2d 1110, 1115 (9th Cir. 1983); Nunez v. Davis, 169 F.3d 1222, 1229 (9th Cir. 1999).

  129. See, e.g., Scott v. Flowers, 910 F.2d 201, 204 (5th Cir. 1990); Duke v. Hamil, 997 F. Supp. 2d 1291, 1293–94 (N.D. Ga. 2014); Dible, 515 F.3d at 922–23.

  130. Miller, 544 F.3d at 548.

  131. Rankin v. McPherson, 483 U.S. 378, 399, 401 (1987) (Scalia, J., dissenting).

  132. See, e.g., Moser v. Las Vegas Metro. Police Dep’t, 984 F.3d 900, 903, 907 (9th Cir. 2021) (noting that the officer recognized the inappropriateness of his Facebook post).

  133. Rankin, 483 U.S. at 384 (“‘[T]he threat of dismissal from public employment is . . . a potent means of inhibiting speech.’ Vigilance is necessary to ensure that public employers do not use authority over employees to silence discourse, not because it hampers public functions but simply because superiors disagree with the content of employees’ speech.” (quoting Pickering v. Bd. of Educ., 391 U.S. 563, 574 (1968))).

  134. Waters v. Churchill, 511 U.S. 661, 664–66 (1994) (plurality opinion).

  135. Id. at 667.

  136. Id. at 679–80.

  137. Id. at 686, 693–94 (Scalia, J., concurring).

  138. Id. at 676–77 (plurality opinion).

  139. Id. at 698 (Stevens, J., dissenting).

  140. Rankin v. McPherson, 483 U.S. 378, 379–81 (1987).

  141. Id. at 390.

  142. Id. at 382 n.4.

  143. Id. at 390. There were no allegations of workplace disruption:

    Constable Rankin testified that the possibility of interference with the functions of the Constable’s office had not been a consideration in his discharge of respondent and that he did not even inquire whether the remark had disrupted the work of the office. Nor was there any danger that McPherson had discredited the office by making her statement in public. . . . McPherson’s discharge [was] unrelated to the functioning of the office . . . .

    Id. at 389.

  144. Moser v. Las Vegas Metro. Police Dep’t, 984 F.3d 900, 902–04, 907 (9th Cir. 2021).

  145. Id. at 907.

  146. Id. at 909 n.6.

  147. See H.P. Grice, Meaning, 66 Phil. Rev. 377, 387 (1957).

  148. Moser, 984 F.3d at 907–08. As the dissent pointed out, the reasonableness of the police department’s interpretation of the statement did not matter: “Metro’s interpretation of Moser’s statement was by far more reasonable than Moser’s proffered alternative.” Id. at 912 (Berzon, J., dissenting).

  149. Id. at 912 (majority opinion).

  150. Id. at 912 (Berzon, J., dissenting).

  151. Bauer v. Sampson, 261 F.3d 775, 784 n.5 (9th Cir. 2001) (quoting Waters v. Churchill, 511 U.S. 661, 677 (1994)).

  152. Waters, 511 U.S. at 676–77 (plurality opinion).

  153. Id. at 676.

  154. Pickering v. Bd. of Educ., 391 U.S. 563, 572 (1968).

  155. See, e.g., Rankin v. McPherson, 483 U.S. 378, 396 (1987) (Scalia, J., dissenting); Am. Postal Workers Union v. U.S. Postal Serv., 830 F.2d 294, 310 (D.C. Cir. 1987); Bauer, 261 F.3d at 783–84; Moser, 984 F.3d at 909 n.6.

  156. Moser, 984 F.3d at 911–12 (“For private employers, it is their prerogative to take action against an intemperate tweet or a foolish Facebook comment. But when the government is the employer, it must abide by the First Amendment.”).

  157. Garcetti v. Ceballos, 547 U.S. 410, 419 (2006); see also Connick v. Myers, 461 U.S. 138, 147 (1983) (“Our responsibility is to ensure that citizens are not deprived of fundamental rights by virtue of working for the government.”).