Anthony Novak v. City of Parma

RECOMMENDED FOR FULL-TEXT PUBLICATION
Pursuant to Sixth Circuit I.O.P. 32.1(b)
File Name: 19a0170p.06

UNITED STATES COURT OF APPEALS
FOR THE SIXTH CIRCUIT

ANTHONY NOVAK, ┐
Plaintiff-Appellee, │
│
│
v. > No. 18-3373
│
│
CITY OF PARMA; KEVIN RILEY, Individually and in his │
Official Capacity as an Employee of the City of │
Parma, Ohio; THOMAS CONNOR, Individually and as │
an Employee of the City of Parma, Ohio, │
Defendants-Appellants. │
┘

Appeal from the United States District Court
for the Northern District of Ohio at Cleveland.
No. 1:17-cv-02148—Dan A. Polster, District Judge.

Argued: June 20, 2019

Decided and Filed: July 29, 2019

Before: MERRITT, THAPAR, and READLER, Circuit Judges.

_________________

COUNSEL

ARGUED: Steven D. Strang, GALLAGHER SHARP LLP, Columbus, Ohio, for Appellants.
Subodh Chandra, THE CHANDRA LAW FIRM LLC, Cleveland, Ohio, for Appellee. ON
BRIEF: Steven D. Strang, Monica A. Sansalone, GALLAGHER SHARP LLP, Columbus,
Ohio, for Appellants. Subodh Chandra, Patrick S. Kabat, THE CHANDRA LAW FIRM LLC,
Cleveland, Ohio, for Appellee.
 No. 18-3373 Novak v. City of Parma, et al. Page 2

_________________

OPINION
_________________

THAPAR, Circuit Judge. Apple pie, baseball, and the right to ridicule the government.
Each holds an important place in American history and tradition. So thought Anthony Novak
when he created a Facebook page to mock the Parma Police Department. He styled his page to
look like the department’s official Facebook page. But the similarities ended there. Novak
shared posts like an advertisement for a “Pedophile Reform event,” at which pedophiles would
receive honorary police commissions.

Novak’s page delighted, disgusted, and confused. Not everyone understood it. But when
it comes to parody, the law requires a reasonable reader standard, not a “most gullible person on
Facebook” standard. The First Amendment does not depend on whether everyone is in on the
joke. Neither is it bothered by public disapproval, whether tepid or red-hot.

Novak’s Facebook page was either a protected parody in the great American tradition of
ridiculing the government or a disruptive violation of state law. Maybe both. At this stage, we
decide only whether the officers are entitled to qualified immunity. For some of Novak’s claims
they are, but for others they are not.

I.

This case comes to us after a motion to dismiss, so we take the facts as Novak alleges
them and draw reasonable inferences in his favor. Courtright v. City of Battle Creek, 839 F.3d
513, 520 (6th Cir. 2016). Novak created a “farcical Facebook account” designed to look like the
police department’s official page. R. 6, Pg. ID 1239. The page was up for twelve hours and
published several posts. Among the posts was a recruitment advertisement that “strongly
encourag[ed] minorities to not apply.” Id. at 1250. Novak also posted an apology from the
department for “neglecting to inform the public about an armed white male who robbed a
Subway sandwich shop,” while promising to bring to justice an “African American woman” who
was loitering outside the Subway during the robbery. Id.
 No. 18-3373 Novak v. City of Parma, et al. Page 3

The page was polarizing. Some of its about 100 followers thought it was “the funniest
thing ever.” Id. at 1253. Others were angry. And yet others were confused, wondering whether
this was the actual Parma police official Facebook page. A handful of people were so angry or
confused that they called the police station. In all, the station received twelve minutes of calls.
Others continued to enjoy the page, which soon “became a platform for a wide range of citizens
to air their grievances about the Department.” Id. at 1259. The officers later testified that they
worried the page would confuse the public and that the “likely result is that people would call.”
Id. at 1271.

One of the page’s audiences—the Parma Police Department—did not find the page
funny. Once the officers got wind of Novak’s page, they “all stopped what [they] were doing to
take a look at it, and a couple of [them] tried to figure out who did it.” Id. at 1253. One officer
said they “just wanted it down.” Id. at 1254. They took several steps to make that happen.

A Facebook battle ensued. First, the department posted a warning on its official
Facebook page. The warning alerted the public to the fake page and assured them that the matter
was “currently being investigated.” Id. at 1255–56. Then Novak reposted the exact same
warning on his own page. He claims he did this to “deepen his satire.” Id. at 1259. For the
same reason, Novak deleted “pedantic comments” on his page explaining that the page was fake,
as these “clumsy explication[s]” only “belabored the joke.” Id. at 1253.

After that, the conflict moved offline and into the real world. Officer Kevin Riley
assigned Officer Thomas Connor to the case and tasked him with finding out who ran the page.
So Connor sent a letter to Facebook requesting that the page be shut down immediately. He also
sent an email to a different Facebook representative asking that the page be taken down. The
police also informed local news outlets of the investigation. The case of the fake police page
even appeared on the nightly news. At that point, Novak decided to delete his creation. He had
heard of the department’s investigation and was worried about the consequences.

Though Novak was done posting, the police department was not done investigating.
They still wanted to find the person behind the laptop. So Connor subpoenaed records from
Facebook. Riley directed Connor to go further and obtain a search warrant for Facebook. Novak
 No. 18-3373 Novak v. City of Parma, et al. Page 4

alleges that Connor made several “material misrepresentations and omissions” to obtain that
warrant. Id. at 1260. The warrant still issued, and Facebook disclosed that Novak was the one
behind the fake account.

Once the department realized that Novak was the cyber culprit, Riley directed Connor to
obtain two more warrants—one to search Novak’s apartment and one to arrest him. The
warrants said that Novak unlawfully impaired the department’s functions, in violation of Ohio
Rev. Code § 2909.04(B). Novak responds that, other than twelve minutes of phone calls to the
department, the police department suffered no disruption to its functions. And Novak claims the
officers were unaware of the twelve minutes of call time when they obtained the warrants. But,
once again, the warrants still issued, and the department arrested Novak. The case went to trial,
and Novak was acquitted.

After he was acquitted of the criminal charge, Novak sued the City of Parma and Officers
Riley and Connor. He alleged (in over thirty claims) that the city and its officers violated his
constitutional and statutory rights under federal and Ohio law. The defendants moved to dismiss
his thirty-plus claims. The district court granted the motion in part and denied it in part, with
twenty-six claims left standing. On appeal, the police claim that qualified immunity shields them
from Novak’s lawsuit. We review de novo whether the officers are entitled to qualified
immunity and issues “inextricably intertwined” with that question. Courtright, 839 F.3d at 517–
18, 523.

II.

Qualified immunity protects government officials like the Parma police officers from
being liable for money damages if their conduct did not violate “clearly established statutory or
constitutional rights of which a reasonable person would have known.” Harlow v. Fitzgerald,
457 U.S. 800, 818 (1982). The doctrine balances “the need to hold public officials accountable
when they exercise power irresponsibly” with “the need to shield officials from harassment,
distraction, and liability when they perform their duties reasonably.” Pearson v. Callahan, 555
U.S. 223, 231 (2009).
 No. 18-3373 Novak v. City of Parma, et al. Page 5

Officers Riley and Connor are entitled to qualified immunity “unless (1) they violated a
federal statutory or constitutional right, and (2) the unlawfulness of their conduct was clearly
established at the time.” District of Columbia v. Wesby, 138 S. Ct. 577, 589 (2018) (internal
quotation marks omitted). For a right to be “clearly established,” the “constitutionality of the
officer’s conduct” must have been “beyond debate” in the “particular circumstances before him.”
Id. at 589–90 (quoting Ashcroft v. al–Kidd, 563 U.S. 731, 741 (2011)). The Supreme Court has
cautioned that “clearly established” must not be defined “at a high level of generality.” Id. at
590. Instead, we must be sensitive to the fact that police officers work in the real world, which is
often messier than law books would have us believe. Nieves v. Bartlett, 139 S. Ct. 1715, 1725
(2019) (“Police officers conduct approximately 29,000 arrests every day—a dangerous task that
requires making quick decisions in circumstances that are tense, uncertain, and rapidly
evolving.” (internal quotation marks and citation omitted)). So when it comes to holding police
officers liable for heat-of-the-moment decisions they make in the line of duty, abstract legal
principles will not do the trick.

On both the facts and the law, specificity is our guiding light. But we must also be
mindful of the stage of the proceedings. This case reaches us early, after a motion to dismiss.
And while we always hope to resolve qualified immunity claims at the earliest possible point in
the litigation, we cannot resolve such claims when we need more factual development to do so.
Phillips v. Roane Cty., 534 F.3d 531, 538 (6th Cir. 2008) (noting that an appeal of a denial of
qualified immunity must be “premised not on a factual dispute, but rather on ‘neat abstract issues
of law’” (quoting Johnson v. Jones, 515 U.S. 304, 317 (1995))). We consider each of Novak’s
claims under these standards.

III. Retaliation

Novak argues that the officers retaliated against him because of his protected speech.
The retaliation claim turns on two issues: (1) whether Novak’s Facebook page was a parody and
(2) whether the Parma police had probable cause to arrest Novak for his page. Because resolving
both issues involves questions of fact, the claim survives. Greene v. Barber, 310 F.3d 889, 898
(6th Cir. 2002) (“Because of the fact-intensive nature of the requisite inquiry, . . . we would be
 No. 18-3373 Novak v. City of Parma, et al. Page 6

usurping the role of the jury were we to attempt to [resolve it] . . . at this stage of the
proceeding.”).

To allege a retaliation claim, Novak must show that: (1) he “engaged in a
constitutionally protected activity,” (2) the officers’ adverse actions caused Novak “to suffer an
injury that would likely chill a person of ordinary firmness” from continuing that activity, and
(3) the officers were motivated, at least in part, by his exercise of his constitutional rights. Paige
v. Coyner, 614 F.3d 273, 277 (6th Cir. 2010). At this stage, the parties dispute whether Novak’s
Facebook page was protected speech.

a. Parody

Was Novak’s speech protected? The Supreme Court has repeatedly reminded us that
almost all speech is protected other than “in a few limited areas.” United States v. Stevens, 559
U.S. 460, 468 (2010) (internal quotation marks omitted). These “limited areas” include speech
expressed as part of a crime, obscene expression, incitement, and fraud. See United States v.
Alvarez, 567 U.S. 709, 717, 720 (2012). It is clearly established, though, that parody does not
fall in one of these “limited areas.” Hustler Magazine, Inc. v. Falwell, 485 U.S. 46, 56–57
(1988). It is protected speech. Id.

The question, then, is whether Novak’s page was a parody. The officers claim that his
Facebook page was false and meant to mislead the public, not a parody. But they are wrong to
think that we just look to a few confused people to determine if the page is protected parody.

Our nation’s long-held First Amendment protection for parody does not rise and fall with
whether a few people are confused. Instead, we must apply a “reasonable reader” test. Id.
Speech that “could not reasonably have been interpreted as stating actual facts” is a parody, even
if “patently offensive.” Id. The test is not whether one person, or even ten people, or even one
hundred people were confused by Novak’s page. Indeed, the genius of parody is that it comes
close enough to reality to spark a moment of doubt in the reader’s mind before she realizes the
joke. “The germ of parody lies in the definition of the Greek parodeia . . . as a song sung
alongside another.” Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569, 580 (1994) (internal
quotation marks omitted). And masterful parody may skirt that line even closer. Benjamin
 No. 18-3373 Novak v. City of Parma, et al. Page 7

Franklin’s 1784 satirical essay in the Journal de Paris came so close to the truth that it
anticipated reality before it happened. Franklin spoke of the benefits of daylight and joked that
the French should consider waking up earlier to save money on candles. In his tongue-in-cheek
proposal, Franklin recommended several measures for the implementation of his plan.
He suggested that: “Every morning, as soon as the sun rises, let all the bells in every church be
set ringing; and if that is not sufficient?, let cannon be fired in every street, to wake the sluggards
effectually, and make them open their eyes to see their true interest.”
Benjamin Franklin, An Economical Project, Letter to the Editor of the Journal of Paris
(1784), http://www.webexhibits.org/daylightsaving/franklin3.html. Through his satire, Franklin
predicted the reality of daylight saving time, which would come a century and a half later.

And a parody need not spoil its own punchline by declaring itself a parody. “Parody
serves its goals whether labeled or not, and there is no reason to require parody to state the
obvious (or even the reasonably perceived).” Campbell, 510 U.S. at 583 n.17. Imagine if The
Onion were required to disclaim that parodical headlines like the following are, in reality, false:
Presidential Debate Sidetracked By Booker, De Blasio Arguing About Best Place In Lower
Manhattan To Get Tapas, or, John Bolton Urges War Against the Sun After Uncovering
Evidence It Has Nuclear Capabilities. News in Brief, The Onion (June
26, 2019), https://politics.theonion.com/presidential-debate-sidetracked-by-booker-de-blasio-ar-
1835870332; News in Brief, The Onion (June 10, 2019), https://politics.theonion.com/john-
bolton-urges-war-against-the-sun-after-uncovering-1835805360. The law of parody does not
require us to strain credulity so far. And that is not because everyone always understands the
joke. Susanna Kim, All the Times People Were Fooled by The Onion, ABC News (June 1,
2015), https://abcnews.go.com/International/times-people-fooled-onion/story?id=31444478.

Instead, the test for parody is whether a reasonable reader would have seen Novak’s
Facebook page and concluded that the posts stated “actual facts.” Hustler, 485 U.S. at 50. Our
nation boasts a long history of protecting parody and satire. “[F]rom the early cartoon portraying
George Washington as an ass down to the present day, . . . satirical cartoons have played a
prominent role in public and political debate.” Id. at 54. And parody, like all protected speech,
need not be high-minded or respectful to find safe haven under the First Amendment. “One of
 No. 18-3373 Novak v. City of Parma, et al. Page 8

the prerogatives of American citizenship is the right to criticize public men and measures—and
that means not only informed and responsible criticism but the freedom to speak foolishly and
without moderation.” Baumgartner v. United States, 322 U.S. 665, 673–74 (1944) (Frankfurter,
J.). “The art of the cartoonist is often not reasoned or evenhanded, but slashing and one-sided.”
Hustler, 485 U.S. at 54. We uphold this right, even where parody shocks us, because “[o]ur trust
in the good sense of the people on deliberate reflection goes deep.” Baumgartner, 322 U.S. at
674.

Whether Novak’s page was a protected parody is a question of fact that we cannot answer
at this stage. See Hustler, 485 U.S. at 57 (“The Court of Appeals interpreted the jury’s finding to
be that the ad parody was not reasonably believable, and in accordance with our custom we
accept this finding.” (internal quotation marks and citation omitted)). Instead, the jury will have
to answer that question. At this stage, though, Novak has alleged enough facts that a reasonable
jury could find that his page was a parody.

b. Probable Cause

Since we accept for now that the page was protected speech, we move to the second
question: did the Parma police have probable cause to arrest Novak? Probable cause exists
where there is a “fair probability” or “substantial chance” that officers will discover evidence of
criminal activity. See Safford Unified Sch. Dist. No. 1 v. Redding, 557 U.S. 364, 370 (2009). To
answer this question, we need more facts. We need to know whether the citizen calls to the
police station gave the officers probable cause to think there was a “disruption” or “interruption”
of police operations under Ohio law. Thus, whether the police had probable cause to arrest
Novak is an issue of fact, which we do not have jurisdiction to decide. Swiecicki v. Delgado,
463 F.3d 489, 498 (6th Cir. 2006) (“Probable cause is an issue of fact for the jury to resolve if
there are any genuine issues of material fact that are relevant to the inquiry.”), abrogated on
other grounds by Wallace v. Kato, 549 U.S. 384 (2007). Of course, a retaliation claim is like a
flow chart—once you decide one issue, it leads to the next. So, we move on. In the probable
cause inquiry that follows, we assume that Novak’s page was protected speech.
 No. 18-3373 Novak v. City of Parma, et al. Page 9

If the police did not have probable cause to arrest Novak, then he may bring a claim of
retaliation. Nieves, 139 S. Ct. at 1725. To prevail on this claim, Novak will need to show that
the officers arrested him based on a “forbidden” retaliatory motive. Id. at 1722–23. But
retaliatory motive is often difficult to prove. After all, “protected speech is often a ‘wholly
legitimate consideration’ for officers when deciding whether to make an arrest.” Id. at 1723–24
(quoting Reichle v. Howards, 566 U.S. 658, 668 (2012)) (explaining that the “content and tone”
of a suspect’s speech may indicate whether he presents a threat). A plaintiff alleging retaliatory
arrest must disentangle these “wholly legitimate” considerations of speech from any wholly
illegitimate retaliatory motives.

To do so, the threshold question Novak must answer is whether “retaliation was a
substantial or motivating factor” for his arrest. Id. at 1725. Novak bears the burden of making
that showing. If he does, the next question is whether the officers would have arrested him
absent that retaliatory motive. Id. (quoting Lozman v. City of Riviera Beach, 138 S. Ct. 1945,
1952–53 (2018)). The burden to answer that lies on the officers. Id. If they show that they
would have arrested Novak even if he had not criticized the police department, his retaliatory-
arrest claim fails. Id. So the questions will be: (1) Can Novak show that the officers were
motivated by retaliatory animus, not legitimate motivations? (Novak’s burden); if yes, (2) Can
the officers justify Novak’s arrest based on something other than retaliation—i.e., a mistaken but
honest belief that there was probable cause? (Officers’ burden).

If the officers did have probable cause, on the other hand, they are entitled to qualified
immunity. The Supreme Court has said as much. “This Court has never recognized a First
Amendment right to be free from a retaliatory arrest that is supported by probable cause.”
Reichle, 566 U.S. at 664–65. The Supreme Court said that in 2012, and it remains true today.

The Supreme Court decided two retaliation cases after Reichle. Neither case clearly
established Novak’s right to be free from a retaliatory arrest based on probable cause. First, the
Supreme Court decided Lozman v. City of Riviera Beach. There, the Court held that a plaintiff
can bring a retaliation claim if the police had probable cause to arrest but only against official
municipal policies of retaliation. 138 S. Ct. at 1954–55. So Lozman does not apply where, as
here, the plaintiff sues individual officers. Nieves, 139 S. Ct. at 1722 (noting that the facts in
 No. 18-3373 Novak v. City of Parma, et al. Page 10

Lozman were “far afield from the typical retaliatory arrest claim” and recognizing that Lozman’s
holding was “limited . . . to arrests that result from official policies of retaliation”). Second, the
Court held most recently in Nieves that a plaintiff generally cannot bring a retaliation claim if the
police had probable cause to arrest. Id. at 1725. Though Nieves also created an exception to that
general rule that we will discuss later, the exception does not apply here because the officers
would not have been aware of it at the time of Novak’s arrest since the case was decided later.

Nor has our circuit clearly established the law on this issue. In Sandul v. Larion, the
Sixth Circuit denied an officer qualified immunity for a First Amendment retaliation claim and
held that “protected speech cannot serve as the basis for a violation of any of the . . . ordinances.”
119 F.3d 1250, 1256 (6th Cir. 1997). But in that case, the ordinance criminalized the plaintiff’s
speech directly, and there was little question whether the speech was protected. Id. at 1255–56
(“These cases should leave little doubt in the mind of a reasonable officer that the mere words
and gesture ‘f—k you’ are constitutionally protected speech.”). Plus, it is not clearly established
how we reconcile the apparent holding in Sandul that protected speech cannot be the basis for
probable cause with the rule that protected speech can be a “wholly legitimate consideration” for
officers when they decide whether to arrest someone. Reichle, 566 U.S. at 668. “[I]f judges thus
disagree on a constitutional question, it is unfair to subject police to money damages for picking
the losing side of the controversy.” Id. at 669–70 (citation omitted). Simply put, Ohio’s statute
appears to punish the effects of speech (interruptions), not the speech itself, and whether
enforcing such a statute in these circumstances violates the First Amendment is not clearly
established. So the officers would be entitled to qualified immunity.

To sum up, to resolve the retaliation claim, the factfinder below will have to decide:
(1) whether Novak’s Facebook page was a parody, and thus protected speech, and; (2) whether
the officers had probable cause to arrest Novak under the Ohio statute. If the officers did not
have probable cause, they are not entitled to qualified immunity, and Novak can attempt to show
the arrest was retaliatory. If the officers did have probable cause, they are entitled to qualified
immunity even if Novak’s page was protected speech because the law at the time did not clearly
establish that charging Novak under the statute would violate his constitutional rights.
 No. 18-3373 Novak v. City of Parma, et al. Page 11

c. Future Issues

At this stage, we have jurisdiction to review only whether the officers are entitled to
qualified immunity. But a few interesting issues remain. They do not bear on the qualified
immunity analysis above because, as with most interesting legal issues, the law is not clearly
established.

Issue 1. The Supreme Court held recently in Nieves that to bring a First Amendment
retaliatory arrest claim, a plaintiff must generally show that there was no probable cause for the
arrest. 139 S. Ct. at 1727. But the Nieves Court also recognized a narrow exception to this rule
“where officers have probable cause to make arrests, but typically exercise their discretion not to
do so.” Id. For example, “[i]f an individual who has been vocally complaining about police
conduct is arrested for jaywalking . . . it would seem insufficiently protective of First
Amendment rights to dismiss the individual’s retaliatory arrest claim on the ground that there
was undoubted probable cause for the arrest.” Id. It is plausible that Novak’s arrest under Ohio
Rev. Code § 2909.04(B), or one like it, would trigger the exception—i.e., if officers never or
rarely arrested someone under this statute. Unfortunately for Novak, this exception was not
clearly established before Nieves.

Issue 2. Even if Novak’s case would not fall within the narrow exception of Nieves, 139
S. Ct. at 1727, there is good reason to believe that in the future probable cause alone may not
protect the officers.

First, this case may not be subject to the general rule of Nieves because the sole basis for
probable cause was speech. Besides posting to his Facebook page, Novak committed no other
act that could have created probable cause. In other First Amendment retaliation cases on point,
by contrast, the defendant’s conduct was a mix of protected speech and unprotected conduct.
That is, the defendants both said something and did something. See, e.g., id. at 1720–21
(defendant made remarks to police officers (protected speech) and acted aggressively toward
them in an intoxicated state (unprotected conduct)); Reichle, 566 U.S. at 660–61 (defendant
made political remarks (protected speech) and unlawfully touched the Vice President
(unprotected conduct)); Swiecki, 463 F.3d at 491–92 (defendant made comments to the officer
 No. 18-3373 Novak v. City of Parma, et al. Page 12

(protected speech) and engaged in disorderly conduct while intoxicated (unprotected conduct)).
Here, we have nothing like that. Novak did not create a Facebook page criticizing police and use
his computer to hack into police servers to disrupt operations. The sole basis for probable cause
to arrest Novak was his speech. And there is good reason to believe that, based on the reasoning
underlying the First Amendment retaliation cases, this is an important difference.

This is important because in Nieves and its predecessors, the Court based its reasoning on
the thorny causation issue that comes up in cases with both protected speech and unprotected
conduct. The idea is that in cases where the plaintiff both did something and said something to
get arrested, the factfinder will not be able to disentangle whether the officer arrested him
because of what he did or because of what he said. “[R]etaliatory arrest cases . . . present a
tenuous causal connection between the defendant’s alleged animus and the plaintiff’s injury.”
Reichle, 566 U.S. at 668. For example, in Mt. Healthy, the Court held there was no retaliation “if
the same decision would have been reached absent [plaintiff’s] protected speech.” Nieves, 139
S. Ct. at 1722 (quoting Mt. Healthy City Sch. Dist. Bd. of Educ. v. Doyle, 429 U.S. 274, 285
(1977)). Here, that inquiry gets us nowhere because “absent [Novak’s] protected speech,” there
would be no basis for probable cause. So, in this case, the causal connection is not so tenuous.
And the reason for requiring that plaintiff show an absence of probable cause where probable
cause is based only on protected speech is not so clear.

Second, this case strikes at the heart of a problem the Court has recognized in the recent
retaliation cases. “[T]here is a risk that some police officers may exploit the arrest power as a
means of suppressing speech.” Lozman, 138 S. Ct. at 1953. The Court also recognized this risk
in Nieves. The jaywalking exception acknowledges that officers can use probable cause as a
pretext for retaliation. “In such a case, . . . probable cause does little to prove or disprove the
causal connection between animus and injury . . . .” Nieves, 139 S. Ct. at 1727. Novak’s case is
prime ground for the pretext that the Supreme Court has worried about.

For one, potential probable cause was based on protected speech alone. That is not
dispositive because the officers’ consideration of his protected speech may have been “wholly
legitimate.” Id. at 1723–24. But the fact that the arrest was made based only on protected
 No. 18-3373 Novak v. City of Parma, et al. Page 13

speech at least raises a concern that probable cause “does little to prove or disprove the causal
connection” between Novak’s criticism of the police and his arrest. Id. at 1727.

Issue 3. Finally, the vague language of the Ohio statute further heightens the concern
raised in Issue 2. That statute makes it a crime to “use any computer . . . or the internet so as to
disrupt, interrupt, or impair the functions of any police . . . operations.” Ohio Rev. Code §
2909.04(B). To see how broad this statute reaches, consider an example. An activist tweets the
following message: “The police are violating our rights #TakeAction #MakeYourVoiceHeard.”
People in the community see the tweet and begin calling the police department to share their
views. A small protest even forms in the town square. Police station employees spend time
fielding the calls, and a couple of officers go down to monitor the protest. Under the plain text
of the Ohio statute, have these acts of civic engagement “interrupt[ed]” police operations?
Taken at face value, the Ohio law seems to criminalize speech well in the heartland of First
Amendment protection. This broad reach gives the police cover to retaliate against all kinds of
speech under the banner of probable cause. Critical online comments, mail-in or phone bank
campaigns, or even informational websites that incite others to “disrupt” or “interrupt” police
operations could violate the law. See id.

Where a statute gives police broad cover to find probable cause on speech alone,
probable cause does little to disentangle retaliatory motives from legitimate ones. Thus, this case
raises new questions under Nieves. It may be that, based on the Supreme Court’s reasoning in
that case and others, the general rule of requiring plaintiffs to prove the absence of probable
cause should not apply here. We need not decide that now.

IV. Other Claims

Prior Restraint. Novak alleges that the Parma police imposed a prior restraint on his
speech. This claim survives, for now.

A prior restraint is an “administrative” or “judicial order[]” that forbids protected speech
in advance. Alexander v. United States, 509 U.S. 544, 550 (1993). An action taken after the
speech is expressed, like a punishment for disfavored speech, is not a prior restraint. Id. at 554
(admonishing that courts not “blur the line separating prior restraints from subsequent
 No. 18-3373 Novak v. City of Parma, et al. Page 14

punishments” for speech). The First Amendment guarantees “greater protection from prior
restraints.” Id. Indeed, we generally presume prior restraints are unconstitutional. See Se.
Promotions, Ltd. v. Conrad, 420 U.S. 546, 558 (1975). But the question is whether Novak has
alleged a prior restraint. He alleges that the police issued a press release threatening to prosecute
him, sent a letter and an email to Facebook demanding the page be taken down, and confiscated
some of his computer equipment. He says the letter and email to Facebook “demanded” that the
page be taken down with an “implicit threat of adverse governmental action” against Facebook if
they refused. R. 6, Pg. ID 1257. This question turns on whether these communications were an
administrative order. Alexander, 509 U.S. at 550.

First, in light of our long history of guarding against prior restraints on speech, cf.
Respublica v. Oswald, 1 U.S. (Dall.) 319, 325 (Pa. 1788); see also 4 William Blackstone,
Commentaries on the Laws of England *151, we should not be overly formalistic in defining
what counts as an administrative order. But courts have not always been clear about what counts
as an administrative order, and that poses a problem when we are talking about what is clearly
established. Take Alexander. That case held that a prior restraint must raise a “legal
impediment” to speech and described the “classic examples of prior restraints” as temporary
restraining orders, permanent injunctions, and court orders. Alexander, 509 U.S. at 550–51. But
the formality of these classic cases should be a sufficient condition for prior restraint, not a
necessary one. See id. at 575 (Kennedy, J., dissenting) (“Though perhaps not in the form of a
classic prior restraint, the application of the forfeiture statute here bears its censorial cast.”). A
government official should not have to declare his order official or jump through certain
procedural hoops to create a prior restraint. Such a rule would allow government officials to
cloak unconstitutional restraints on speech under the cover of informality. To borrow a concept,
when an officer “carr[ies] a badge of authority of the government and represent[s] it in some
capacity,” Edmonson v. Leesville Concrete Co., 500 U.S. 614, 632 (1991) (O’Connor, J.,
dissenting), his order to a private party to take a specific action may fairly be called an
“administrative order.” This is true even if the order is not on its terms binding. Bantam Books,
Inc. v. Sullivan, 372 U.S. 58, 70 (1963) (recognizing “a system of prior administrative restraints”
even where affected parties were free to ignore the notices). And because taking down the page
 No. 18-3373 Novak v. City of Parma, et al. Page 15

would mean Novak could no longer post critical comments about the police on his page, the
letter and email to Facebook could be administrative orders that constituted a prior restraint.

Novak also plausibly alleges that the officers created a prior restraint with their press
release threatening to take legal action. In the release, the department announced that it had
opened a criminal investigation into Novak’s page. Under Bantam Books, a threat of
prosecution can trigger a prior restraint, even if the threat is non-binding. 372 U.S. at 60, 71.
True, the facts in Bantam Books were more extreme than what we have here. There, Rhode
Island created an eerily titled “Commission to Encourage Morality in Youth” to investigate
obscene or impure literature. Id. at 59. The commission sent notices to publishers saying certain
books were too objectionable to sell and that violators may be prosecuted. Id. at 60–62. Police
then visited the publishers’ book distributors to see if the objectionable books had been removed.
Id. at 63. But the facts of Bantam Books need not be perfectly analogous for the rule to apply.
Rather, we need more facts to determine whether the facts in this case are close enough.

These issues were not briefed here or decided below. And the officers do not argue that
their Facebook communications were not an administrative order. So, we leave this decision in
the first instance in the capable hands of the district court. The prior restraint claim goes on.

Additional First Amendment Claims. Novak argues that when Officers Riley and Connor
deleted comments on the official police Facebook page, they unlawfully censored speech in a
public forum and violated his right to receive information. These claims fail because they are
not based on clearly established law.

The First Amendment no doubt applies to the wild and “vast democratic forums of the
Internet.” Packingham v. North Carolina, 137 S. Ct. 1730, 1735 (2017). But when it comes to
online speech, the law lags behind the times. And rightly so. “The forces and directions of the
Internet are so new, so protean, and so far reaching that courts must be conscious that what they
say today might be obsolete tomorrow.” Id. at 1736; see also id. at 1744 (Alito, J., concurring in
the judgment) (noting that courts should “proceed circumspectly, taking one step at a time” in
applying “free speech precedents” to the Internet).
 No. 18-3373 Novak v. City of Parma, et al. Page 16

Courts have not reached consensus on how First Amendment protections will apply to
comments on social media platforms. So far, the courts that have considered the issue have
taken different approaches. See Morgan v. Bevin, 298 F. Supp. 3d 1003, 1012–13 (E.D. Ky.
2018) (denying preliminary injunction regarding the deletion of Facebook and Twitter comments
in a case of first impression). But see Davison v. Randall, 912 F.3d 666, 687–88 (4th Cir. 2019)
(holding that a government official violated the First Amendment by banning a critical
constituent from a Facebook page). No doubt, any right Novak or the commenters may have to
post or receive comments was not “beyond debate” at the time the officers deleted the comments.
al–Kidd, 563 U.S. at 741. Riley and Connor are entitled to qualified immunity from these
claims.

Anonymity. Novak argues that the officers violated his right to speak anonymously. This
claim does not survive because Novak does not allege a violation of clearly established law.

The right to speak anonymously is deeply rooted in American political tradition and in
First Amendment doctrine. McIntyre v. Ohio Elections Comm’n, 514 U.S. 334, 342 (1995); id.
at 371 (Thomas, J., concurring in the judgment). From Thomas Paine’s “Common Sense”
(originally published anonymously during the Revolution) to the debates between Federalists and
Anti-Federalists during Ratification, anonymity was core at the Founding. Id. at 368 (Thomas,
J., concurring in the judgment). When some Federalists encouraged newspapers to ban
anonymous speech, the Anti-Federalists defended their right to remain anonymous. Id. at 364–
66 (Thomas, J., concurring in the judgment). Free speech was originally understood to include
the right to speak without being known. Consistent with this original understanding, the
Supreme Court has upheld the right by striking down laws banning anonymous speech. Id. at
357; id. at 371 (Thomas, J., concurring in the judgment); Talley v. California, 362 U.S. 60, 65
(1960).

But Novak is not contesting a law or policy that bans anonymous speech. Instead, he
argues that the police officers disclosed his identity as part of their criminal investigation. Yet he
has pointed to no law clearly establishing that investigative actions by police can violate the right
to speak anonymously. Investigations are often public events. So too are criminal trials. See
Craig v. Harney, 331 U.S. 367, 374 (1947). True, the Supreme Court has recognized that courts
 No. 18-3373 Novak v. City of Parma, et al. Page 17

can reduce a criminal trial’s publicity, but only under the right to a fair trial, not under a right to
remain anonymous. See Neb. Press Ass’n v. Stuart, 427 U.S. 539, 554 (1976).

It is not clearly established that announcements made in an ongoing criminal
investigation can violate Novak’s First Amendment right to speak anonymously. The officers
are entitled to qualified immunity on this claim.

Search and seizure and malicious prosecution. Novak alleges that the officers
unlawfully searched him and seized his property. He also alleges wrongful arrest and malicious
prosecution. These claims survive as well at this stage of the litigation. Ultimately, Novak will
have to show that Officer Connor lied to get a warrant (for unlawful seizure) or lied in the course
of his prosecution (for malicious prosecution). And Novak will have to make this showing in
light of the Ohio statute that grants broad discretion to officers.

To prove malicious prosecution, Novak must show (1) that the officers’ “deliberate or
reckless falsehoods result[ed] in arrest and prosecution without probable cause” and (2) that the
officers did more than passively participate in the decision to prosecute or to keep prosecuting
him. Newman v. Twp. of Hamburg, 773 F.3d 769, 772 (6th Cir. 2014); see also Johnson v.
Moseley, 790 F.3d 649, 654–55 (6th Cir. 2015). Novak must also show the absence of probable
cause. See Sykes v. Anderson, 625 F.3d 294, 305 (6th Cir. 2010).

Usually, a warrant from a neutral magistrate, like the ones Connor got in this case, would
be a “complete defense” to these § 1983 claims. Id. at 305, 310 & n.8; Knott v. Sullivan, 418
F.3d 561, 568–69 (6th Cir. 2005). Not so here. Warrants are typically a defense because they
demonstrate probable cause. But warrants do not demonstrate probable cause if the officer
“ma[de] false statements and omissions to the judge” and if probable cause would not exist but
for those false statements or omissions. Sykes, 625 F.3d at 305 (quoting Vakilian v. Shaw, 335
F.3d 509, 517 (6th Cir. 2003)). In these limited circumstances, officers may be held liable for
their searches, seizures, and arrests even though they obtained a warrant. Id. at 308.

Thus, these claims turn on whether Officer Connor made false statements or omissions.
According to Novak, Officer Connor falsely represented that Novak (1) “disrupted and
impaired” the functioning of the Parma Police Department “by knowingly posting false
 No. 18-3373 Novak v. City of Parma, et al. Page 18

information,” (2) “altered or affected” the department’s official page, and (3) falsely represented
he was “a representative of the Parma Police Department.” R. 6, Pg. ID 1265–66. Further,
Novak says that Connor knew there was no interruption or disruption. As for malicious
prosecution, he alleges that Connor and Riley lied at trial by testifying that Novak’s page caused
a disruption when they knew it did not. See Moseley, 790 F.3d at 655 (noting a plausible
allegation of malicious prosecution has been made when officers testify at trial and provide
“false statements [and] flagrant misrepresentations, or fail[] to disclose key items of evidence”
(citing Sykes, 625 F.3d at 301–02, 306–07, 311–17)). For now, those allegations are enough.

Privacy Protection Act. Novak alleges a violation of the Privacy Protection Act. This
claim too depends on whether the officers lacked probable cause to search Novak’s apartment
and seize his property. Because Novak has alleged facts that make it plausible that the officers
lacked probable cause, the claim survives for now.

The Privacy Protection Act makes it unlawful for a government officer to “search for or
seize any work product materials possessed by a person reasonably believed to have a purpose to
disseminate” information to the public. 42 U.S.C. § 2000aa(a). But the statute has a “suspect
exception.” S.H.A.R.K. v. Metro Parks Serving Summit Cty., 499 F.3d 553, 567 (6th Cir. 2007).
The Act does not apply if the officers have “probable cause to believe that the person possessing
such materials has committed or is committing the criminal offense to which the materials
relate.” 42 U.S.C. § 2000aa(a)(1); see S.H.A.R.K., 499 F.3d at 567. Novak has alleged that the
officers lacked probable cause to search and seize the contents of his apartment, so the “suspect
exception” does not apply at this stage. The claim goes forward.

Supervisory liability. Novak seeks to hold Riley, Connor’s supervisor, liable for
Connor’s alleged constitutional violations. But Novak sues under § 1983, and under that law, a
plaintiff cannot sue for vicarious liability or respondeat superior. Ashcroft v. Iqbal, 556 U.S.
662, 676 (2009). So Riley can be held responsible only for his own actions, not for his
supervision of anyone else. Id. Novak must allege that Riley “encouraged the specific
[unconstitutional conduct] or in some other way directly participated in it.” McQueen v. Beecher
Cmty. Sch., 433 F.3d 460, 470 (6th Cir. 2006) (quoting Shehee v. Luttrell, 199 F.3d 295, 300 (6th
Cir. 1999)). He has so alleged, so the supervisory liability claim survives.
 No. 18-3373 Novak v. City of Parma, et al. Page 19

Novak alleges that Officer Riley first assigned Officer Connor to investigate the
Facebook page and then directed Connor to take the allegedly unconstitutional actions. Novak
also attaches a transcript from his criminal trial where Riley testified that he “contacted Detective
Connor, asked him to look into [Novak’s page], [and] assigned the case to him.” R. 6-1, Pg. ID
1369. At this stage, these allegations are enough. Novak may proceed against Riley for his own
actions and for any of Connor’s actions that Riley directed or supervised.

Conspiracy. Novak brings a claim against Riley, Connor, and “John Doe” of the Ohio
Internet Crimes Against Children Task Force for conspiring to shut down his Facebook page.
John Doe allegedly told Connor how to contact Facebook and shut down the page.

In this circuit, the test for conspiracy is simple. “All that must be shown [for conspiracy]
is that there was a single plan, that the alleged coconspirator shared in the general conspiratorial
objective, and that an overt act was committed in furtherance of the conspiracy that caused
injury . . . .” Hooks v. Hooks, 771 F.2d 935, 943–44 (6th Cir. 1985). The officers respond that
Novak’s allegations are conclusory. But allegations alone, even if conclusory or improbable,
may suffice for this early stage of litigation. Novak names the coconspirators, suggests that they
came to an agreement, and alleges that they acted against his Facebook page. These are “enough
fact[s] to raise a reasonable expectation that discovery will reveal evidence of illegal agreement.”
Bell Atl. Corp. v. Twombly, 550 U.S. 544, 556 (2007).

But that is not the end of the conspiracy inquiry. In the time since the district court
denied the officers’ motion to dismiss, our circuit has changed its law on conspiracy. We held
that the “intracorporate conspiracy doctrine” applies to § 1983 lawsuits like this one. Jackson v.
City of Cleveland, 925 F.3d 793, 818 (6th Cir. 2019). That doctrine holds that “members of the
same legal entity cannot conspire with one another as long as their alleged acts were within the
scope of their employment.” Id. at 819 (citation omitted). At this time, the intracorporate
conspiracy doctrine does not apply here. Novak alleges that Doe is a member of a different
agency than Riley and Connor—the Ohio Internet Crimes Against Children Task Force. So they
are not in the same “legal entity.” With more facts, the district court should consider whether
Doe does work for the Task Force and whether the Task Force is a different agency for purposes
of the intracorporate conspiracy doctrine under Jackson.
 No. 18-3373 Novak v. City of Parma, et al. Page 20

Municipal liability. Novak brings several claims against the City of Parma. The district
court denied Parma’s motion to dismiss on these claims, and the city now appeals. We do not
have jurisdiction over these claims.

This appeal is limited to qualified immunity and issues “inextricably intertwined” with it.
Courtright, 839 F.3d at 523–24. Two claims are “inextricably intertwined” if resolving one
claim will “necessarily determine” the other. Id. Here, the officers’ liability depends on their
actions against Novak. The city’s liability, in contrast, depends on a separate analysis of “its
municipal policies, training programs, and customs.” Id. So Parma’s liability is not
“inextricably intertwined” with qualified immunity, and we do not have jurisdiction to consider
municipal liability at this stage. Id.

State claims. Finally, Novak brings several state law claims. The officers raise one
defense. They argue that they are protected by an Ohio statute that insulates police officers from
liability unless their actions were taken “with malicious purpose, in bad faith, or in a wanton or
reckless manner.” Ohio Rev. Code § 2744.03(A)(6)(b). The officers say that Novak has not
shown their conduct was in bad faith, wanton, or reckless. Here again, the stage of proceedings
informs this question. As the district court rightly noted, to dismiss Novak’s complaint at this
stage, we must find that it is “devoid of [allegations] tending to show that the [officers] acted” as
Novak alleges. Irving v. Austin, 741 N.E.2d 931, 934 (Ohio Ct. App. 2000); see also Range v.
Douglas, 763 F.3d 573, 586 (6th Cir. 2014). The complaint is not devoid of such allegations.
Indeed, it is filled to the brim with them. Novak alleges that the officers lied to Facebook to take
down his page, lied to secure warrants to arrest him, and lied on the witness stand about their
actions. At this early motion-to-dismiss stage, that is enough to plausibly allege that the officers
acted with a “dishonest purpose” constituting bad faith. Cook v. Hubbard Exempted Vill. Bd. of
Educ., 688 N.E.2d 1058, 1061 (Ohio. Ct. App. 1996). Novak’s state law claims live to fight
another day.
 No. 18-3373 Novak v. City of Parma, et al. Page 21

* * *

Though Novak’s Facebook page mocking the Parma Police Department has since left the
cyber world, several of his legal claims will live on. Others will end here. We REVERSE the
district court’s decision to deny the motion to dismiss on Novak’s claims related to anonymous
speech, censorship in a public forum, and the right to receive speech. We AFFIRM the district
court’s decision with respect to all other claims except municipal liability, over which we lack
jurisdiction.