I Was Sued Over A Joke Tweet And I Won
The Third Circuit Court of Appeals comes through for free speech and raises questions about the scope of the NLRB
Longtime readers will be familiar with the fact that I have, since 2019, been sued by the National Labor Relations Board, at the behest of leftists Matt Bruenig and subsequently Joel Fleming, over a joke I made on Twitter:
For an explanation in video form about what happened next, watch this:
The accusation was laughable. No employee had expressed a desire to unionize. If anyone had, my joke wouldn’t have stood in the way. Mr. Fleming was an interloper anyway—he had no association with the Federalist. But according to the NLRB, anyone can file such a complaint against any company.
The NLRB proceeded to invade our publication, heedless of the freedom of the press. Members of my staff were subpoenaed to testify in New York, where none of them lived and we had no office. The NLRB attempted to subpoena all emails and communications between staff members going back years—including about editorial decisions, hiring decisions, and confidential sources during our coverage of the Russia-collusion hoax.
Help arrived in the form of the New Civil Liberties Alliance, a nonprofit set up by legal scholar Philip Hamburger that defends constitutional rights against overreach by the administrative state.
The NLRB proposed a settlement: I delete the joke, I post information on the rights of employees to unionize, and the complaint goes away. I said no…
It is my good fortune to know many brilliant lawyers who are willing to stand up against abuses of the administrative state. Most people aren’t so lucky. What happens when another small-business owner on social media makes a similar joke? When a neighbor decides to make a federal case out of an impolitic Facebook post? What happens if freedom of speech is only worth what the common business owner is willing to pay in legal fees, missed work and the cost of flying employees to other states to testify in front of bureaucrats?
It’s understandable that those who can’t afford to fight often bend the knee. But in that America, the bureaucrats, and the trolls who use them to shut down speech they don’t like, will keep rolling on until someone stands up and says no.
After working our way through the NLRB’s process before an administrative law judge or ALJ (bureaucrats judging bureaucrats, that’s how it works), we finally ended up at the Third Circuit, before Circuit Judges Hardiman, Matey, and Scirica.
I’ve pasted the bulk of the opinion below, cleaned up a bit, but I want to particularly draw your attention to this paragraph, which seems to indicate the key question that would need resolution should the NLRB choose to appeal to the Supreme Court (emphasis mine):
The NLRA is remarkably broad in scope and power. Its jurisdictional provision, written in the passive voice, places no limits on who may file a charge. The Supreme Court has recognized—and has done nothing to cabin—this capacious authority. So the circuit courts of appeals (including this one) have consistently upheld the Board’s rule that “any person” may precipitate an unfair labor practice investigation. Unfortunate as it may be, the Act as written and interpreted empowers a politically-motivated busybody as much as a concerned employee or civic-minded whistleblower.
Here’s a portion of the decision (again, emphasis mine):
When considering an alleged unfair labor practice, an employer’s conduct must be examined “in light of all the existing circumstances.” Wheeling-Pittsburgh Steel Corp. v. NLRB, 618 F.2d 1009, 1020 (3d Cir. 1980) (emphasis added) (citations omitted); see also NLRB v. Va. Elec. & Power Co., 314 U.S. 469, 479 (1941) (The Board’s finding of an unfair labor practice must be based “upon the whole course of conduct revealed by [the] record.”). Context is an important part of language, and that’s especially true where, as in this case, pure speech is at issue.
The ALJ found that Domenech’s tweet was “an obvious threat” that “had no other purpose except to threaten the FDRLST [Media] employees with unspecified reprisals.” FDRLST Media, 370 N.L.R.B. at 5. The Board agreed. In adopting the ALJ’s finding, the Board disclaimed any reliance on the tweet’s timing or The Federalist’s editorial content, leaving only the words of the tweet, devoid of any context, as support. But the Board erred when it disregarded relevant contextual evidence. ImageFIRST, 910 F.3d at 736 (citation omitted). Even more problematic than the timing and editorial content the Board ignored are the circumstances surrounding the tweet that the Board and the ALJ never considered. Had the Board considered the tweet’s full context, it could not have concluded that a reasonable FDRLST Media employee would view the tweet as a threat of reprisal.
For starters, FDRLST Media is a tiny media company. Its six employees (not including Domenech) are writers and editors. The tweet’s suggestion that these employees might be sent “back” to work in a “salt mine” is farcical. The image evoked—that of writers tapping away on laptops in dimly-lit mineshafts alongside salt deposits and workers swinging pickaxes—is as bizarre as it is comical. So from the words of the tweet alone, we cannot conclude that a reasonable FDRLST Media employee would view Domenech’s tweet as a plausible threat of reprisal.
The ALJ rightly noted that the salt mine idiom “most often . . . refer[s] to tedious and laborious work,” FDRLST Media, 370 N.L.R.B. at 5, but the Board failed to realize that “[t]hreatening statements are not usually made in bantering terms” like these. NLRB v. Champion Labs., Inc., 99 F.3d 223, 229 (7th Cir. 1996) (citing NLRB v. Windemuller Elec., Inc., 34 F.3d 384, 392 (6th Cir. 1994)). To conclude that Domenech’s tweet is a “thinly veiled statement concerning adverse consequences,” Garry Mfg. Co., 630 F.2d at 940, requires some additional evidence of the tweet’s coercive tendency. But the Board points to none. The Supreme Court has instructed, “[a]ny assessment of the precise scope of [impermissible] employer expression, of course, must be made in the context of its labor relations setting.” Gissel Packing Co., 395 U.S. at 617.
Yet neither the Board nor the ALJ discussed the labor environment at FDRLST Media when Domenech posted his tweet. Significantly, this Court has never affirmed a finding of an unfair labor practice based on employer speech alone absent any indicia of labor friction. And in this case, the Board points to no history of labor strife, no evidence of antagonism, nor even a single example of labor-management tension. The Board cited only one brief tweet, posted from a supervisor’s personal Twitter account. The record contains no sign— indeed, no inkling—of any circumstance at FDRLST Media that leads us to conclude that a reasonable employee would interpret Domenech’s tweet as a veiled threat.
We also think it significant that The Federalist “publishes commentary on a wide variety of contemporary newsworthy and controversial topics,” including matters involving politics and labor relations, AR 69, and that Domenech used his personal Twitter account to promote and discuss the magazine’s commentary. The record does not show that Domenech ever used this account to communicate with employees or that employees were required to follow it. Taken together, a reasonable FDRLST Media employee who became privy to Domenech’s tweet—posted the same day as the Vox Media walkout—would be far more likely to view the tweet as “commentary on a . . . contemporary newsworthy and controversial topic” than as a threat that implicated her status with the Employer. See AR 69.
The Board also went too far when it opined that employees’ “subjective interpretations of [an employer’s conduct] are irrelevant to determining whether” the employer committed an unfair labor practice. FDRLST Media, 370 N.L.R.B. at 1 n.3. We have held that employees’ subjective impressions are not dispositive; not that they are irrelevant. See, e.g., Hedstrom Co., 629 F.2d at 316–17 (affirming the Board’s finding that a supervisor’s statement was coercive, despite employee testimony suggesting otherwise, because therecord supported both interpretations). We have acknowledged that subjective impressions can sometimes be helpful to determine how a reasonable employee would objectively view her employer’s conduct. See, e.g., Garry Mfg. Co., 630 F.2d at 941 & n.6 (considering employee testimony that removal of plant machinery before a union election caused “panic and fear among the employees” to support a finding that the removal was reasonably “understood as a threat of reprisal should the Union win the election”).
Employees’ subjective impressions are especially helpful where, as here, the employer claims his statement was made in jest. Humor is subjective. What is funny to a fisherman may be lost on a farmer. A quip about New England winters is unlikely to get a laugh in Alaska. The propensity for jokes to fall flat for want of context or audience understanding has given rise to idioms like “I guess you had to be there” and “too soon?”
Excluding context and viewing a statement in isolation, as the Board did here, could cause one to conclude that “break a leg” is always a threat. But when expressed to an actor, singer, dancer, or athlete, that phrase can reasonably be interpreted to mean only “good luck.” Break a leg, McGrawHill’s American Idioms Dictionary (Richard A. Spears ed., 4th ed. 2000). Consistent with these commonsensical observations, some of our sister courts have considered employees’ subjective responses when evaluating whether employer speech or expressive conduct was reasonably viewed as a joke or a threat…
The Board discounts Federal-Mogul and Windemuller by noting that neither case “holds that threats of retaliation are lawful as long as someone laughs.” Board Br. 17. The Board is right to that extent. But the Fifth and Sixth Circuit analyses also show that employees’ subjective responses can be relevant to determining whether a reasonable employee—who is familiar with her employer and the context of a remark—would tend to be coerced. This is particularly true where, as here, a third party with no connection to the employer or the employees—and who lacks knowledge of relevant context—files a charge against an employer with no history of labor problems.
The record contains no evidence that any FDRLST Media employee perceived Domenech’s tweet as a threat, and the Board failed to even acknowledge that noteworthy gap in the record. Because the charge was filed by an unrelated thirdparty, the alleged unfair labor practice was pure speech, and the meaning of the employer’s statement is open to question, the “silence of the record” is significant and should have been considered, see Windemuller Elec., Inc., 34 F.3d at 393.
(The Board acted within its discretion when it excluded from consideration the affidavits offered by the Employer. See NLRB v. Domsey Trading Corp., 636 F.3d 33, 37 (2d Cir. 2011); 29 C.F.R. § 101.10(a). But the exclusion was also expedient, since the two employee affidavits contradict the Board’s conclusion: the employees stated that they did not view Domenech’s tweet as a threat, but as a “funny,” “obviously sarcastic,” “pithy,” and “satirical” expression of Domenech’s personal views. AR 156, 158)
Finally, the mode of communication also weighs against finding that Domenech’s tweet would tend to coerce a reasonable FDRLST Media employee. Domenech posted his message on Twitter, a public platform that limits tweets to 280 characters, which encourages users to express opinions in exaggerated or sarcastic terms. Domenech sent his message to the timelines of his more than eighty thousand Twitter followers, not to the email inboxes of his FDRLST Media employees. And he made the tweet available to the public—a peculiar choice indeed for a threat supposedly directed at six employees. These characteristics of Domenech’s tweet would give a reasonable FDRLST Media employee even more reason to read the tweet as mocking a rival internet media company or commenting on a timely socio-political issue than as threatening reprisal.
We “recognize the Board’s competence in the first instance to judge the impact of utterances made in the context of the employer-employee relationship.” Gissel Packing Co., 395 U.S. at 620. But in this case, the Board’sfailure to consider the tweet’s context dooms its finding of a veritable threat. Considered in context, the tweet was not an unfair labor practice. So we will set aside the Board’s order to the contrary.
Our conclusion is buttressed by the fact that the alleged unfair labor practice consists of the Employer’s words alone. In protecting employees’ statutory labor rights, neither we, nor the Board, can violate an employer’s right to free speech under the First Amendment. Gissel Packing, 395 U.S. at 617. The Act distinguishes prohibited employer conduct from protected employer speech in Section 8(c), which provides that [t]he expressing of any views, argument, or opinion, or the dissemination thereof, whether in written, printed, graphic, or visual form, shall not constitute or be evidence of an unfair labor practice under any of the provisions of this subchapter [29 U.S.C. §§ 151–169], if such expression contains no threat of reprisal or force or promise of benefit. 29 U.S.C. § 158(c); see Gissel Packing, 395 U.S. at 617 (“[Section] 8(c) . . . implements the First Amendment.”). Section 8(c) “manifest[s] a congressional intent to encourage free debate on issues dividing labor and management,” Chamber of Com. of the U.S. v. Brown, 554 U.S. 60, 67 (2008) (cleaned up), and the Act favors “uninhibited, robust, and wide-open debate in labor disputes, stressing that freewheeling use of the written and spoken word . . . has been expressly fostered by Congress and approved by the [Board].” Id. at 68 (cleaned up). Section 8(c) reinforces the “open marketplace” created by the First Amendment, “in which differing ideas about political, economic, and social issues can compete freely for public acceptance without improper government interference.” Knox v. Service Employees International Union, Local 1000, 567 U.S. 298, 309 (2012).
To give effect to Congress’s intent and avoid conflict with the First Amendment, we must construe the Act narrowly when applied to pure speech, recognizing that only statements that constitute a true threat to an employee’s exercise of her labor rights are prohibited. Cf. Graham Architectural Prods. Corp. v. NLRB, 697 F.2d 534, 541 (3d Cir. 1983) (“What the Act proscribes is only those instances of true “interrogation” which tend to interfere with the employees’ right to organize.”). To interpret the Act otherwise risks expanding Section 8(a)(1)’s prohibition beyond its constitutionally permissible bounds.
We thus must be “vigilant to see that the [Board] does not read elements of interference, restraint or coercion into speech that is in fact nonthreatening and that would not strike a reasonable person as threatening.” Windemuller, 34 F.3d at 392. We need not determine here the precise contours of prohibited “threat[s] of reprisal or force.” 29 U.S.C. § 158(c). We merely observe that the Board must be careful to distinguish “[w]hat is a threat . . . from what is constitutionally protected speech.” Watts v. United States, 394 U.S. 705, 707 (1969) (per curiam).
At the very least, any finding of a prohibited threat must be made by examining the employer’s statement in its full context, with due consideration of the audience and accompanying circumstances. Cf. id. at 707–08 (relying on the “context, . . . expressly conditional nature of the statement and the reaction of the listeners,” who laughed, in concluding the petitioner’s statement was “political hyperbole” and not a criminally prohibited threat). The Board’s acontextual analysis of Domenech’s tweet falls far short of this standard.
The National Labor Relations Act grants the National Labor Relations Board vast authority to investigate charges of unfair labor practices, even when charges are filed by parties who are not personally aggrieved by the alleged practice. But the Board’s authority to find an unfair labor practice is not unlimited.
Here, the Board spent its resources investigating an online media company with seven employees because of a facetious and sarcastic tweet by the company’s executive officer. Because the Board lost the forest for the trees by failing to consider the tweet in context, it misconstrued a facetious remark as a true threat. We will accordingly grant FDRLST Media’s petition, set aside the Board’s order, and deny the Board’s petition for enforcement.
The concurrence, from Judge Matey, is even better:
I agree with the majority’s thoughtful analysis of the threats to dialogue and debate, not to mention farce, in the National Labor Relations Board’s decision. In its haste to join the tedious chorus of disapproval against whatever disfavored view has most recently appeared somewhere on the internet, the Board shelved serious supervision of the protections for America’s employees. Rightly, the majority concludes the Board’s efforts to restrict speech cannot be squared with what “the Founders recognized [as] an inalienable natural right to express one’s thoughts, sometimes described as the ‘freedom of opinion.’” Jud Campbell, Natural Rights and the First Amendment, 127 Yale L.J. 246, 267 (2017).
But I see an earlier problem with the Board’s action. As best understood, the National Labor Relations Act cabins charging authority to those who suffered some adverse effect in the workplace. Not, as here, someone who took offense to something seen while scrolling Twitter. As the filer was not aggrieved under the NLRA, the Board lacked jurisdiction to launch this case. And because neither precedent nor deference alters the best reading of the NLRA, I respectfully concur only in the judgment.
Here’s his portion on Chevron:
Finally, the Board turns to deference, urging us to put aside the statute because it has “consistently reaffirmed” Indiana in its regulations. (Response Br. at 36–37.) But just as the Board cannot transform day into night, it cannot import Indiana’s pre-1947 statutory analysis into the post-1947 NLRA. A point made clear by the Supreme Court’s repeated instructions on the limited deference given to agency statutory interpretations.
Under the familiar framework of Chevron, a court must first determine whether the statute is “ambiguous.” Chevron, U.S.A., Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 843 (1984). A phrase that, while once shrouded in mystery, is now understood to turn not on possible meanings, but ordinary understanding. See Kisor, 139 S. Ct. at 2415 (courts must “empty” the “legal toolkit”); City of Arlington v. FCC, 569 U.S. 290, 296 (2013) (“First, applying the ordinary tools of statutory construction, the court must determine whether Congress has directly spoken to the precise question at issue.” (quotations omitted)).
Finding ambiguity where none exists is, as Chevron itself states, contrary to the judicial duty to act as “the final authority on issues of statutory construction.” 467 U.S. at 843 n.9; see also Gun Owners of America, Inc. v. Garland, 19 F.4th 890, 926 (6th Cir. 2021) (en banc) (Murphy, J., dissenting) (“A finding of ambiguity can occur only at the end of our usual interpretative process . . . . [A] court must give the relevant words their ordinary meaning.”); Forrest Gen. Hosp. v. Azar, 926 F.3d 221, 228 (5th Cir. 2019) (“Chevron deference must be reflective, not reflexive.”); Arangure v. Whitaker, 911 F.3d 333, 338 (6th Cir. 2018); Voices for Int’l Bus. & Educ., Inc. v. NLRB, 905 F.3d 770, 780–81 (5th Cir. 2018) (Ho, J., concurring) (“Finding ambiguity where it does not exist . . . misuse[s]. . . Chevron.”).
FDRLST notes that Chevron deference is constitutionally suspect, a practice that requires courts to “bow to the nation’s most powerful litigant, the government, for no reason other than that it is the government.” Egan v. Delaware River Port Auth., 851 F.3d 263, 278 (3d Cir. 2017) (Jordan, J., concurring). That suspicion is decreased with a searching application of the statutory text, after which a “court will almost always reach a conclusion about the best interpretation,” leaving “no need to adopt or defer to an agency’s contrary interpretation.” Kisor, 129 S. Ct. at 2448 (Kavanaugh, J., concurring). “In other words,” the interpretation requirement of Chevron, “taken seriously, means that courts will have no reason or basis to put a thumb on the scale in favor of an agency.” Id.; see also Pauley v. BethEnergy Mines, Inc., 501 U.S. 680, 707 (1991) (Scalia, J., dissenting) (“Chevron is a recognition that the ambiguities in statutes are to be resolved by the agencies charged with implementing them, not a declaration that, when statutory construction becomes difficult, we will throw up our hands and let regulatory agencies do it for us.”).
In short, deference arises in the rare case when no superior statutory reading can be found, not when an inferior construction competes with a best reading. And as explained, the best reading of the NLRA grants charging power only to those aggrieved. With our tools unpacked, the NLRB’s plea for deference is unwarranted.
If we needed to reach whether Domenech’s words constituted an unfair labor practice, the majority’s reasoned holding is correct. But we need not, because the best reading of the NLRA trims the NLRB’s jurisdiction and prevents unaffiliated parties from searching the internet for wisecracks to transform into workplace violations that unleash the NLRB’s sweeping power. Precedent does not require today’s jurisdictional holding, and the NLRA’s text marshals against it. For those reasons, I respectfully concur only in the judgment.
The full opinion should be posted here later today. For now, I’m just thankful to the NCLA, pleased with the outcome, and happy that the humorless trolls and their army of bureaucrats lost.
The Transom is a reader-supported publication. Please consider becoming a paid subscriber.