Sarah K. Molina; Christina Vogel; Peter Groce v. City of St. Louis, Missouri; County of St. Clair, Illinois; John Doe, I-VI

United States Court of Appeals
For the Eighth Circuit
___________________________
No. 21-1830
___________________________
Sarah K. Molina; Christina Vogel; Peter Groce
Plaintiffs - Appellees
v.
City of St. Louis, Missouri; County of St. Clair, Illinois; John Doe, I-VI
Defendants
Daniel Book, in his individual capacity; Joseph Busso, in his individual capacity
Defendants - Appellants
Jason C. Chambers
Defendant
Lance Coats, in his individual capacity; Stephen Dodge, in his individual capacity;
Joseph Mader, in his individual capacity; Michael D. Mayo, in his individual
capacity; Mark S. Seper, in his individual capacity; William Wethington, in his
individual capacity
Defendants - Appellants
____________
Appeal from United States District Court
for the Eastern District of Missouri - St. Louis
____________

-- 1 of 22 --

-2-
Submitted: January 12, 2022
Filed: February 2, 2023
____________
Before BENTON, SHEPHERD, and STRAS, Circuit Judges.
____________
STRAS, Circuit Judge.
Officers in an armored police vehicle shot tear gas at three people near the
scene of a protest in downtown St. Louis. The district court concluded that all three
had a First Amendment retaliation claim. We agree that one of them does, but
qualified immunity shields the officers from the claims brought by the others. We
affirm in part, reverse in part, and remand.
I.
A large protest broke out in St. Louis in 2015. In the crowd were Sarah Molina
and Christina Vogel, both members of the National Lawyers Guild. In Molina’s
words, their goal was to “protect[] the right to protest,” not to participate in one. To
make their self-appointed role known, they wore bright green hats emblazoned with
the words “National Lawyers Guild Legal Observer.”
During the protest, St. Louis police officers formed a line and repeatedly
ordered the crowd to disperse. Instead of leaving, the protestors responded by
throwing rocks and bottles. The officers warned protestors about the possible use of
chemical agents, and when they refused to go, shot inert smoke canisters into the
crowd.
Vogel recorded these events as Molina stood nearby and watched. Once
officers switched to tear gas, the two women left. A few minutes later, they
reassembled with five to ten others on Molina’s property, located about 550 feet
away.

-- 2 of 22 --

-3-
Minutes later, an armored vehicle known as the BEAR barreled down the
street toward them. As it drove past, tear-gas canisters landed near Molina and
Vogel. Although the officers would later deny shooting chemicals from the BEAR,
an after-action report revealed otherwise.
As the BEAR continued down the street, Peter Groce followed on a bicycle.
Once it stopped, he approached and shouted, “[g]et the fuck out of my park.” The
officers responded by launching a tear-gas canister that allegedly hit him in the hip.
Molina, Vogel, and Groce sued the officers and their supervisor, Lieutenant
Stephen Dodge, under 42 U.S.C. § 1983 for, among other claims, First Amendment
retaliation. In the face of a summary-judgment motion seeking qualified immunity,
the district court ruled that the claims could proceed to a jury. The officers ask us to
determine whether the case should have ended there.
II.
In deciding whether the district court should have granted summary judgment,
we must answer two questions. First, did the officers violate a constitutional right?
Second, was the right clearly established? See Morgan v. Robinson, 920 F.3d 521,
523 (8th Cir. 2019) (en banc) (explaining that we may answer them in either order).
In answering these questions, “we [must] accept as true the facts that the district
court found were adequately supported, as well as the facts the district court likely
assumed.” Burbridge v. City of St. Louis, 2 F.4th 774, 779–80 (8th Cir. 2021)
(bracket and quotation marks omitted) (reviewing the summary-judgment
determination de novo); see Berry v. Doss, 900 F.3d 1017, 1021 (8th Cir. 2018)
(explaining that, in an appeal from a denial of qualified immunity, we review “purely
legal issue[s]” based on “the district court’s factual presumptions” (quotation marks
omitted)).

-- 3 of 22 --

-4-
A.
To prevail on their retaliation claim, the plaintiffs must show that “they
engaged in protected [First Amendment] activity.” Quraishi v. St. Charles County.,
986 F.3d 831, 837 (8th Cir. 2021); see Hoyland v. McMenomy, 869 F.3d 644, 655
(8th Cir. 2017), abrogated on other grounds by Nieves v. Bartlett, 139 S. Ct. 1715,
1722 (2019), as recognized in Laney v. City of St. Louis, 56 F.4th 1153, 1157 n.2
(8th Cir. 2023). If they can make that showing, then the focus shifts to whether the
officers “took [an] adverse action . . . that would chill a person of ordinary firmness
from continuing in the [protected] activity.” Hoyland, 869 F.3d at 655 (citation
omitted); see Eggenberger v. West Albany Township, 820 F.3d 938, 943 (8th Cir.
2016). If they did, then the next hurdle is causation: was the First Amendment
activity a “but-for cause” of the injury? Nieves, 139 S. Ct. at 1722 (quotation marks
omitted).
Establishing the violation itself, however, is only half the battle. Getting past
qualified immunity requires the plaintiffs to show that it would have been
“sufficiently clear [to] every reasonable official . . . that what [they were] doing
violate[d]” the First Amendment. Reichle v. Howards, 566 U.S. 658, 664 (2012)
(quotation marks omitted); Wilson v. Lamp, 901 F.3d 981, 986 (8th Cir. 2018)
(explaining that the burden remains with the plaintiffs, even at this step). “Existing
precedent,” in other words, must have put the issue “beyond debate.” Reichle, 566
U.S. at 664 (quoting al-Kidd, 563 U.S. at 741). Although Groce gets over each of
these hurdles, Molina and Vogel do not.
B.
The First Amendment provides that “Congress shall make no law . . .
abridging the freedom of speech.” U.S. Const. amend. I. It protects “symbolic or
expressive conduct as well as . . . actual speech.” Virginia v. Black, 538 U.S. 343,
358 (2003). Molina and Vogel claim that the First Amendment covers what they
did, which was observe and record police conduct during the St. Louis protest. Even

-- 4 of 22 --

-5-
if we were to assume they are correct, observing and recording police-citizen
interactions was not a clearly established First Amendment right in 2015.
Start with the Supreme Court’s 50-year-old decision in Colten v. Kentucky,
407 U.S. 104 (1972). After a group of college students left a political demonstration
in a “procession of [6] to 10” cars, a police officer pulled one of them over for an
expired license plate. Id. at 106. A student from another car then went over to
observe the traffic stop and ask questions. Eventually, other students joined him,
which prompted another trooper to repeatedly ask the group to “disperse.” Id. When
those efforts failed, the officer arrested one of the students for disorderly conduct.
Id. at 107.
Throughout trial and on appeal, the student claimed that Kentucky’s
disorderly-conduct statute was unconstitutionally overbroad. In concluding it was
not, the Supreme Court announced that individuals “[have] no constitutional right to
observe the issuance of a traffic ticket or to engage the issuing officer in
conversation.” Id. at 109. As for the student’s refusal to “move on,” it too was
unprotected, at least “without more.” Id. Colten suggests that observing police
conduct is not expressive.1
1 The dissent claims that two other cases cabin Colten, but neither undermines
our conclusion here. See post, at 4–5 (citing Hoyland, 869 F.3d at 656, and City of
Houston v. Hill, 482 U.S. 451 (1987)). The first one, Houston v. Hill, distinguished
protected expression, such as “words or conduct that annoy or offend,” from
unprotected activities like “stand[ing] near a police officer and persistently
attempt[ing] to engage the officer in conversation while the officer is directing traffic
at a busy intersection.” 482 U.S. at 462 n.11, 465 (citation omitted). Indeed, Hill
even touted the disorderly conduct statute in Colten as an example of the kind of
properly tailored statute that “infringe[d] no protected speech or conduct.” Id. at 465
n.14 (emphasis added) (quoting Colten, 407 U.S. at 111). Hoyland drew a similar
distinction. It contrasted a plaintiff’s “exercis[e of] his First Amendment rights” to
“verbally . . . oppose or challenge police action” with “simply ‘refusing to move on
after being directed to do so . . . without more.’” Hoyland, 896 F.3d at 656 (quoting
Colten, 407 U.S. at 109). Far from ignoring precedent, as the dissent alleges, we are
faithfully applying it.

-- 5 of 22 --

-6-
None of the plaintiffs’ cases clearly establish otherwise. Walker v. City of
Pine Bluff, 414 F.3d 989 (8th Cir. 2005), is an ordinary Fourth Amendment case.
When two police officers approached someone who had watched them conduct a
traffic stop, the bystander said he had been watching “Pine Bluff’s finest in action.”
Id. at 992. The officers arrested the bystander for “obstructing government
operations.” Id. at 993. Qualified immunity was unavailable, we concluded,
because the officers lacked “arguable probable cause to arrest . . . [him] in this
situation.” Id. (emphasis added). The opinion never mentions, much less discusses,
the First Amendment.
The same goes for the second case, Chestnut v. Wallace, 947 F.3d 1085 (8th
Cir. 2020). Like Walker, Chestnut involved a bystander who watched as a police
officer “perform[ed] traffic stops.” Id. at 1087. The officer eventually called for
backup because of the “suspicious person . . . following her to her car stops.” Id.
The arriving officer placed the bystander in handcuffs and detained him for about 20
minutes. See id. at 1087–88. We concluded there was no reasonable suspicion to
conduct an investigatory stop because the bystander was not doing anything illegal.
Id. at 1090 (stating that “no reasonable officer could conclude that a citizen’s passive
observation of a police-citizen interaction from a distance was criminal”).
It is true, as the plaintiffs note, that some of the language in Chestnut was
broad. Relying on a few out-of-circuit cases invoking the First Amendment, for
example, we stated that there is a “clearly established right to watch police-citizen
interactions at a distance and without interfering.” Id. at 1090. But we did so based
only on “the facts that existed when [the bystander] was seized”—a clear reference
to the Fourth Amendment issue we were deciding. Id. (emphasis added). And the
First Amendment cases only bolstered our (narrow) Fourth Amendment holding:
“[w]e merely hold that it was clearly established that [a police officer] could not
detain [the bystander] without more indication of wrongdoing.” Id. at 1091
(emphasis added).

-- 6 of 22 --

-7-
The point is that neither of these Fourth Amendment cases can clearly
establish a First Amendment right to observe police officers.2 See Colten, 407 U.S.
at 109. Nor did a clearly established First Amendment right to record them exist in
2015. See Frasier v. Evans, 992 F.3d 1003, 1019–20 (10th Cir. 2021) (holding that
a “purported First Amendment right to record [police officers] was not clearly
established in August 2014”); Fields v. City of Philadelphia, 862 F.3d 353, 362 (3d
Cir. 2017) (“[W]e cannot say that the state of the law at the time of our cases (2012
and 2013) gave fair warning so that every reasonable officer knew that, absent some
sort of expressive intent, recording public police activity was constitutionally
protected.”); Turner v. Lieutenant Driver, 848 F.3d 678, 687 (5th Cir. 2017)
(explaining that “there was no clearly established First Amendment right to record
the police” in 2015). The question is whether, as Colten put it, there is anything
“more” here.
C.
Molina and Vogel try to give us “more” in the form of three other theories.
The first is a peaceful-assembly theory: although the protest may have been unruly,
the gathering at Molina’s property was not. The second is based on the bright green
hats they wore, which they believe “proclaimed both their affiliation and their role
within the larger demonstration.” Their third theory is that the officers must have
mistakenly thought they were protestors, which they say is good enough to allow
2 It is not beyond the realm of possibility that a First Amendment right to
observe police exists, but our Fourth Amendment cases like Walker and Chestnut do
not clearly establish it. And it makes good sense why. It is one thing to conclude
that officers cannot arrest someone passively standing by and watching as they do
their job. After all, in the absence of interference, there is no crime in it. But it is
another matter to say that watching is itself expressive. Expressive of what? Not
even Molina and Vogel can provide a clear answer.

-- 7 of 22 --

-8-
them to sue on a First Amendment retaliation theory. In the end, none of these
theories work.3
1.
Timing is the basis for the peaceful-assembly theory. The officers did not fire
the tear-gas canisters until after Molina and Vogel had reassembled with five to ten
others. The argument is that the officers must have been reacting to their lawful
assembly, not to the protest itself. There are two problems with this argument.
The first is that not every gathering falls under the umbrella of the First
Amendment. The right of association presupposes that the purpose is to “engag[e]
in those activities protected by the First Amendment—speech, assembly, petition for
the redress of grievances, and the exercise of religion.” Roberts v. U.S. Jaycees, 468
U.S. 609, 618 (1984). And here, as the district court explained, the handful of people
in Molina’s yard were just trying to find “a safe meeting place away from the protest
site.” See URI Student Senate v. Town of Narragansett, 631 F.3d 1, 12–13 (1st Cir.
2011) (“The constitutionally protected right of association . . . has never been
expanded to include purely social gatherings.”). Not all reasonable officers would
have known the gathering was a protected assembly, particularly when they were
dodging rocks and bottles just a few minutes earlier. See Reichle, 566 U.S. at 664;
see also Grayned v. City of Rockford, 408 U.S. 104, 116 (1972) (explaining that
demonstrations “lose their protected quality as expression under the First
Amendment” when they “turn violent”).
The second is that, even assuming the gathering in Molina’s yard was
protected, there is no evidence to suggest that it had anything to do with the officers’
3 In their supplemental reply brief, Molina and Vogel argue they had a right of
“access to information about how our public servants operate in public.” We decline
to address this argument because the officers have never had a chance to respond to
it. See Berg v. Norand Corp., 169 F.3d 1140, 1146 (8th Cir. 1999) (“refus[ing] to
entertain [a] new argument” mentioned “for the first time in [the] reply brief”).

-- 8 of 22 --

-9-
decision to use tear gas. To succeed on their theory, Molina and Vogel must show
that the officers “singled [them] out because” they lawfully reassembled elsewhere,
regardless of what happened earlier. Baribeau v. City of Minneapolis, 596 F.3d 465,
481 (8th Cir. 2010) (quotation marks omitted). If something else was the motivation,
however, then the reassembly was not a “but-for cause” of their injuries. Nieves,
139 S. Ct. at 1722 (quotation marks omitted).
If Molina and Vogel were “singled out,” the district court suggested that, “at
the least,” it was because the officers “assumed” the gathering was a continuation of
the earlier unlawful assembly. When they were patrolling the streets surrounding
the protest, they were following orders from Lieutenant Dodge to break up the crowd
and prevent the protestors from doing further harm. Molina even admitted that she
“was assembled with [the protestors]” before moving to her house. The only
reasonable inference to draw from these facts is that the officers were “merely
carrying out their duty as they underst[ood] it.” Mitchell v. Kirchmeier, 28 F.4th
888, 897 (8th Cir. 2022).
It makes no difference that the officers may have made a mistake. As we have
explained, retaliatory animus cannot be the driving force whenever officers act based
on their “understanding—however mistaken—of [their] official duties,” even if the
mistake turns out to be “unreasonable.”4 Id. at 896. So even if the officers
“unreasonably believed” that the group was refusing to comply with their earlier
directions to disperse, their official orders—not retaliatory animus—caused them to
use the tear gas. Baribeau, 596 F.3d at 481.
2.
Returning to the protest itself, their second theory is that wearing the bright
green hats expressed a “pro-protest” message. Recall that the hats said, “National
4 An “unreasonable mistake,” by contrast, does not shield officials from Fourth
Amendment claims. Baribeau, 596 F.3d at 481.

-- 9 of 22 --

-10-
Lawyers Guild Legal Observer.” Although neither their color nor the words
emblazoned on them directly conveyed a pro-protest message, Molina and Vogel
claim that the act of wearing them sent a “particularized message.” Burnham v.
Ianni, 119 F.3d 668, 674 (8th Cir. 1997).
In First Amendment parlance, their theory is that wearing the hats was
“expressive behavior” that “constitutes speech.” Id. As we have recognized,
“nonverbal conduct constitutes speech if it is intended to convey a particularized
message and the likelihood is great that the message will be understood by those
who view it.” Id. And if wearing the hats is speech, qualified immunity is still
available unless “every reasonable official would know” that the act conveyed a pro-
protest (or other particularized) message. District of Columbia v. Wesby, 138 S. Ct.
577, 590–92 (2018).
Whether wearing the hats expressed a pro-protest message is a close call. On
the one hand, knowing a bit more about the National Lawyers Guild could lead a
reasonable officer to conclude that Molina and Vogel were there to support the
protestors. On the other, the words “legal observer” could lead someone less
knowledgeable to think they were neutral, there to make sure that neither the police
nor the protestors broke the law. Under the latter view, the hats would identify their
role, not express a “particularized message.” Burnham, 119 F.3d at 674. The point
is that not everyone would have understood the pro-protest message they were trying
to convey. See id.
To the extent courts have recognized that clothing can convey a particularized
message, the meaning was easily identifiable. Perhaps the most famous example
was the anti-war activist who wore a jacket with the words, “Fuck the Draft.” See
Cohen v. California, 403 U.S. 15, 16 (1971). The message was clear: he strongly
opposed “the Vietnam War and the draft” and wanted everybody to know it. Id.
The message in Baribeau v. City of Minneapolis was equally clear. 596 F.3d
at 470–71. As part of an elaborate protest, the participants “dressed as zombies” by

-- 10 of 22 --

-11-
wearing “white powder and fake blood on their faces and dark makeup around their
eyes” and “broadcasted announcements such as ‘get your brains here’ and ‘[b]rain
cleanup in Aisle 5.’” Id. at 470–71. During the protest, they “explained that they
meant their actions as an anticonsumerist commentary.” Id. at 471. So every
reasonable officer would have understood their anti-consumer-culture message.
The point is that some symbols and words carry a clear message. “Fuck the
Draft” unmistakably expresses an anti-war message. See Cohen, 403 U.S. at 16.
Displaying a swastika carries a different kind of message, though its import is
equally unmistakable. See Hurley v. Irish-American Gay, Lesbian and Bisexual
Grp. of Boston, 515 U.S. 557, 569 (1995). And in Tinker v. Des Moines Independent
Community School District, the “black armbands” were a “silent symbol . . . of
opposition to [the] Nation’s part in the conflagration in Vietnam” and the wearers’
“objections to the hostilities in Vietnam and their support for a truce.” 393 U.S. 503,
504, 510 (1969).
We cannot say the same thing about bright green hats that say “National
Lawyers Guild Legal Observer” on them.5 There is no obvious pro-protest message.
Or, at the very least, any pro-protest message is not “beyond debate,” which means
that this theory, like the others, cannot overcome qualified immunity. Wesby, 138
S. Ct. at 589 (quoting al-Kidd, 563 U.S. at 741).
3.
Their final theory is that, even if they did not actually engage in First
Amendment activity, the officers must have thought they did. There is support for
this argument, given that some of the officers thought they were protestors who had
remained together and just moved a couple of blocks away. Citing Heffernan v. City
5 Molina and Vogel also cite Hurley, but it involves “the protected expression
that inheres in a parade,” not in specific articles of clothing. 515 U.S. at 569. It
could not have clearly established a First Amendment right to wear the bright green
hats.

-- 11 of 22 --

-12-
of Paterson, 578 U.S. 266, 268 (2016), Molina and Vogel argue that even a mistaken
belief about what they did is good enough because perception is what counts for a
constitutional claim like this one.
Heffernan involved the “demot[ion] [of] an employee because [an] official
[incorrectly] believed . . . that the employee” had participated “in constitutionally
protected political [First Amendment] activity.” Id. at 268. The Supreme Court
concluded that the “demotion” was “what count[ed],” even though the employee
“had not engaged in . . . protected activities.” Id. at 270–71. Molina and Vogel
argue the same logic should apply here.
The problem with this theory is timing. The Supreme Court decided
Heffernan months after the events in this case took place. 578 U.S. at 266. Even if
the rule it adopted would otherwise apply here, no “controlling authority” or “robust
‘consensus of cases of persuasive authority’” clearly established the right before
early 2016—too late for it to matter here. See Wesby, 138 S. Ct. at 589–90 (quoting
al-Kidd, 563 U.S. at 741–42); see also Heffernan, 578 U.S. at 275 (Thomas, J.,
dissenting) (“[F]ederal law does not provide a cause of action to plaintiffs whose
constitutional rights have not been violated . . . .”).
* * *
Having considered multiple possibilities, we conclude that none of them
work. Qualified immunity prevents Molina and Vogel from recovering on their First
Amendment retaliation claims.
III.
Groce’s First Amendment retaliation claim, on the other hand, fares better.
Recall that he followed the BEAR on a bicycle and then yelled, “[g]et the fuck out
of my park.” His “[c]riticism of [the] officers, even with profanity, is protected
speech.” Thurairajah v. City of Fort Smith, 925 F.3d 979, 985 (8th Cir. 2019); see

-- 12 of 22 --

-13-
Hoyland, 869 F.3d at 656 (citing Hill, 482 U.S. at 461). And at least one case clearly
established the right to be free from retaliation in those circumstances. See Peterson
v. Kopp, 754 F.3d 594, 602 (8th Cir. 2014) (stating that “criticizing a police officer
and asking for his badge number is protected speech”); see also Hill, 482 U.S. at 461
(“[T]he First Amendment protects a significant amount of verbal criticism and
challenge directed at police officers.”).
A.
To counter what appears to be a clearly established constitutional right, the
officers claim this case is different. In their view, Groce acted in an aggressive and
threatening manner, which then gave them arguable probable cause to act. Even
assuming that the presence of arguable probable cause is an absolute defense to a
First Amendment retaliation claim—a question we need not decide today—their
position depends on reinterpreting the facts in a “light most favorable to” them, not
Groce. Engesser v. Fox, 993 F.3d 626, 629 (8th Cir. 2021) (stating that we must
view the facts “in [the] light most favorable to [the plaintiff]”).
Under the “plaintiff-friendly version of the facts,” by contrast, there was no
probable cause, arguable or otherwise, to take any action against Groce. N.S. v.
Kansas City Bd. of Police Comm’rs, 933 F.3d 967, 969 (8th Cir. 2019). The officers
cannot identify any crime that he allegedly committed. Yet they launched a tear-gas
canister at him a few minutes after video footage “depict[ed] a calm scene,” even
though they had no evidence “that [he] was armed or otherwise presented any threat
to the [officers] inside their armored vehicle.” We have no jurisdiction to reinterpret
these facts, which support the denial of qualified immunity. See Berry, 900 F.3d at
1021.
B.
Setting qualified immunity aside, the officers also challenge whether the
evidence is specific enough to allege that any of them individually violated Groce’s

-- 13 of 22 --

-14-
First Amendment rights. He cannot identify who launched the tear-gas canister, so
in their view, no one can be held liable for his injuries.
1.
Liability under section 1983 is “personal.” White v. Jackson, 865 F.3d 1064,
1080–81 (8th Cir. 2017). By personal, we mean that “a plaintiff must show each
individual defendant’s personal involvement in the alleged violation.” Id. It does
not follow, however, that a plaintiff must be able to “personally identify his
assailant[] to avoid summary judgment.” Id. at 1081 (emphasis added).
What we know from the evidence, viewing it in Groce’s favor, is that someone
launched a tear-gas canister from the BEAR. We also know that seven officers were
riding in it at the time: Michael Mayo, Joseph Mader, William Wethington, Mark
Seper, Daniel Book, Lance Coats, and Joseph Busso. All of them “had access to
[chemical] munitions,” which could “be released from one of many portholes of
[the] armored vehicle[,] either by hand or using a launcher.”
At this stage, there is enough evidence to establish the “personal involvement”
of everyone in the BEAR. Id. at 1081. To be sure, Groce could not see who launched
the tear-gas canister. But with multiple “officers present,” the jury could find that
each one of them participated in the decision or that one did it “while the other[s]
failed to intervene.” Velazquez v. City of Hialeah, 484 F.3d 1340, 1342 (11th Cir.
2007). Under these circumstances, the claims against the individual officers can
proceed.
2.
Not so for Lieutenant Dodge. Although Groce alleges that he was
“deliberately indifferent to or tacitly authorized the offending acts,” we disagree.
Barton v. Taber, 908 F.3d 1119, 1125 (8th Cir. 2018) (quotation marks omitted).

-- 14 of 22 --

-15-
Unlike the individual officers in the BEAR, Lieutenant Dodge had no
“personal involvement” in the violation. White, 865 F.3d at 1081. Supervisory
status on its own is not enough. See Ashcroft v. Iqbal, 556 U.S. 662, 676 (2009)
(stating that “vicarious liability is inapplicable to Bivens and § 1983 suits”); Tlamka
v. Serrell, 244 F.3d 628, 635 (8th Cir. 2001) (“A supervisor may not be held liable
under § 1983 for the constitutional violations of a subordinate on a respondeat
superior theory.”). And when the officers fired the tear-gas canister at Groce,
Lieutenant Dodge was in another vehicle patrolling a different area. To be sure, he
gave an order to “use chemical munitions to disperse” the crowd. But he had no
notice that his lawful order was “likely to result in a constitutional violation” or that
his “supervision [under the circumstances] w[as] inadequate.” Barton, 908 F.3d at
1125.
IV.
We accordingly affirm the denial of summary judgment on Groce’s First
Amendment retaliation claim, but only against the individual officers in the BEAR.
We otherwise reverse and remand for the entry of judgment on all other claims.
BENTON, Circuit Judge, dissenting in part and concurring in part.
The panel opinion here holds that observing and recording police-citizen
interactions was not a clearly established First Amendment right on August 19, 2015.
This conclusion violates this circuit’s “cardinal rule . . . that one panel is bound by
the decision of a prior panel.” Mader v. United States, 654 F.3d 794, 800 (8th Cir.
2011) (en banc). “One panel of this Court is not at liberty to disregard a precedent
handed down by another panel.” Drake v. Scott, 812 F.2d 395, 400 (8th Cir. 1987).
I.
Two cases bind this court: Chestnut v. Wallace, 947 F.3d 1085 (8th Cir.
2020), and Walker v. City of Pine Bluff, 414 F.3d 989 (8th Cir. 2005). Both cases

-- 15 of 22 --

-16-
found a clearly established First Amendment right to observe police officers that
existed before the events precipitating this case. See Chestnut, 947 F.3d at 1090;
Walker, 414 F.3d at 993.
Walker, the earlier case, held that observing police officers could not be
outlawed. Id. This court denied qualified immunity to an officer who arrested an
individual for “silently watching the [police] encounter from across the street with
his arms folded in a disapproving manner.” Id. at 992. The Walker case analyzed
the legal effect of only this silent observation. Id. This court held that the officer
lacked probable cause because it was “clearly established” that a peaceful onlooker
could not be arrested “for obstruction of governmental operations or for any other
purported crime.” Id. at 993 (emphasis added). The state, said this court, did not
just happen to permit officer-watching, it had to permit it.
This court detailed the nature and origin of Walker’s clearly established
substantive right in Chestnut v. Wallace, 947 F.3d 1085. This court there reaffirmed
the constitutional origin of the “clearly established right to watch police-citizen
interactions at a distance and without interfering.” Chestnut, 947 F.3d at 1091 (“[I]f
the constitution protects one who records police activity, then surely it protects one
who merely observes it.”). And it confirmed that the First Amendment gives rise to
this right by citing seven First Amendment cases, including two from this circuit.6
Id. at 1090–91, citing Thurairajah v. City of Fort Smith, 925 F.3d 979, 985 (8th
Cir. 2019) (“[T]he First Amendment prohibits government officials from subjecting
an individual to retaliatory actions . . . for speaking out,” quoting Hartman v. Moore,
547 U.S. 250, 256 (2006)); Hoyland v. McMenomy, 869 F.3d 644, 655 (8th Cir.
2017), overruled as to causation element by Nieves v. Bartlett, 139 S. Ct. 1715, 1722
(2019), as recognized by Laney v. City of St. Louis, ___ F.4th ___, 2023 WL
116837, at *3 n.2 (8th Cir. Jan. 6, 2023); Fields v. City of Philadelphia, 862 F.3d
353, 359 (3d Cir. 2017) (“The First Amendment protects the public’s right of access
6 The dissent “agree[d] with the court’s characterization of those [First
Amendment] cases” as establishing that the “right exists.” Chestnut, 947 F.3d at
1096 (Gruender, J. dissenting).

-- 16 of 22 --

-17-
to information about their officials’ public activities.”); id. (because “there is the
right for the eye to see or the ear to hear,” the First Amendment also protects the
right to record police); ACLU of Ill. v. Alvarez, 679 F.3d 583, 597 (7th Cir. 2012)
(First Amendment protects “gathering . . . information about the affairs of
government,” including secret audio recordings of police officers); id. (“[T]he First
Amendment goes beyond protection of the press and self-expression of individuals
to prohibit government from limiting the stock of information from which members
of the public may draw.” quoting First Natl Bank of Boston v. Bellotti, 435 U.S.
765, 783 (1978)); Glik v. Cunniffe, 655 F.3d 78, 81 (1st Cir. 2011) (“It is firmly
established that the First Amendment’s aegis extends further than the text’s
proscription on laws ‘abridging the freedom of speech, or of the press,’ and
encompasses a range of conduct related to the gathering and dissemination of
information.”); Smith v. City of Cumming, 212 F.3d 1332, 1333 (11th Cir. 2000)
(“The First Amendment protects the right to gather information about what public
officials do on public property.”); Fordyce v. City of Seattle, 55 F.3d 436, 439 (9th
Cir. 1995) (authorizing § 1983 claims against an officer who attempt[ed] to prevent
or dissuade [plaintiff] from exercising his First Amendment right to film matters of
public interest.).
Walker and Chestnut correctly attribute this right to the First Amendment.
The First Amendment protects the right to protest “by silent and reproachful
presence.” Brown v. Louisiana, 383 U.S. 131, 141–42 (1966) (opinion of Fortas,
J., announcing the judgment of the Court). See also id. at 148–49 (Brennan, J.,
concurring in the judgment); id. at 150–51 (White, J., concurring in the result).
Walker and Chestnut concerned silent and reproachful police-observation. See, e.g.,
Walker, 414 F.3d at 992 (plaintiff watched police “with his arms folded in a
disapproving manner”); Chestnut, 947 F.3d at 1087 (plaintiff’s desire to observe
police arose because “there had been a lot of difficulty in citizen/police interaction
as of late” (quotation omitted)). Cf. Hurley v. Irish-Am. Gay, Lesbian & Bisexual
Grp. of Boston, 515 U.S. 557, 568–69, 579 (1995) (a parade is First Amendment-
protected expressive conduct); Spence v. Washington, 418 U.S. 405, 410 (1974)
(context determines whether conduct is expressive, and the Kent State tragedy

-- 17 of 22 --

-18-
contextualized an upside-down flag with a peace sign); Baribeau v. City of
Minneapolis, 596 F.3d 465, 476, 477 (8th Cir. 2010) (the First Amendment
protected zombie costumes even though the anti-consumerism message was only
clear after protestors explained their meaning).7
The panel opinion here tries to avoid Chestnut and Walker’s First Amendment
conclusions based on what one Supreme Court case suggests. In Colten v. Kentucky,
407 U.S. 104 (1972), the Court found that “appellant was not engaged in activity
protected by the First Amendment” because “[h]e had no constitutional right to
observe the issuance of a traffic ticket” and, as the state court found, “appears to
have had no purpose other than to cause inconvenience and annoyance.” Colten,
407 U.S. at 109.
Again, this violates the prior-panel rule. This court has established Colten’s
reach in this circuit. In Hoyland v. McMenomy, this court articulated two facts—
and identified one later case—that distinguished Colten. Hoyland, 869 F.3d at 656
(“[T]the officers’ reliance on Colten is misplaced” and plaintiff was engaged in
“protected activity.”). Unlike the Colten appellant, the Hoyland plaintiff was at
home (rather than by a busy highway) and brought a § 1983 claim (rather than a
direct attack on a state statute). Id. This case contains both factors underlying
Hoyland’s conclusion that “reliance on Colten [was] misplaced.” Id. Molina and
Vogel were (allegedly) teargassed at their home where public safety was not
threatened, and they bring § 1983 claims rather than challenges to state criminal law.
7 The First Amendment’s broad sweep is “the proudest boast of [] free speech
jurisprudence.” Matal v. Tam, 137 S. Ct. 1744, 1764 (2017). It protects people of
all stripes. See, e.g., Masterpiece Cakeshop, Ltd. v. Colo. Civ. Rights Comm’n, 138
S. Ct. 1719, 1741–42 (2018) (Thomas, J., concurring in part and concurring in the
judgment). Narrowing its scope undermines “bedrock principles of [the Supreme
Court’s] free-speech jurisprudence” and “should not pass without comment.” Id. at
1740.

-- 18 of 22 --

-19-
More importantly, Hoyland recognized that the Supreme Court had already
cabined Colten in City of Houston v. Hill, 482 U.S. 451 (1987). The Court there
explained that the Colten ordinance “prohibit[ed] only disorderly conduct or fighting
words” and survived the First Amendment challenge only because it “infringe[d] no
protected speech or conduct.” Hill, 482 U.S. at 465 & n.14, quoting Colten, 407
U.S. at 108 (second alteration in original). Laws prohibiting “words or conduct that
annoy or offend” officers, by contrast, ran afoul of the First Amendment. Id. at 465.
Hill’s protection for criticizing officers—and Hoyland’s fidelity to Hill—
underpinned Chestnut’s holding that “if officers cannot seize someone who criticizes
or curses at them while they perform official duties, they cannot seize someone for
exercising the necessarily included right to observe the police in public from a
distance and without interfering.” Chestnut, 947 F.3d at 1091, citing Hoyland, 869
F.3d at 654.
Finally, Chestnut’s recognition of a clearly established First Amendment right
to observe police interactions was a holding, not dicta. This court said so:
Other legal authorities fully support our holding that the right here was
clearly established. Every circuit court to have considered the question
has held that a person has the right to record police activity in public.
[citation omitted]. Four circuits had so decided by the time of the
events in question here. [citations omitted]. This robust consensus of
cases of persuasive authority suggests that, if the constitution protects
one who records police activity, then surely it protects one who merely
observes it—a necessary prerequisite to recording.
Chestnut, 947 F.3d at 1090 (emphasis added).8
8 In Chestnut, this court did not confine this clearly established right to an
expressive-conduct strand of First Amendment jurisprudence. The First
Amendment captures more than expressive conduct. See Chestnut, 947 F.3d at
1090, citing Fields, 862 F.3d at 359 (holding that the First Amendment right to
record police stems from “the right for the eye to see or the ear to hear”). See also
Brown, 383 U.S. at 141–42, 148–49, 150–51 (Fortas, J., announcing the judgment

-- 19 of 22 --

-20-
II.
The panel opinion here attributes that clearly established right to the Fourth,
rather than First, Amendment. In doing do, it underappreciates the interplay between
First and Fourth Amendment rights in this circuit’s precedent.
Police may not seize a person without suspecting or believing that the person
committed or is about to commit a crime. See Waters v. Madson, 921 F.3d 725, 736
(8th Cir. 2019). The Fourth Amendment sets two thresholds, “reasonable suspicion”
for a detention and “probable cause” for an arrest. Id. Thus the Fourth Amendment
deems it “unreasonable” to seize someone solely for conduct not suggesting a crime.
See United States v. Cortez, 449 U.S. 411, 417 (1981) (“An investigatory stop must
be justified by some objective manifestation that the person stopped is, or is about
to be, engaged in criminal activity.”). Legal conduct generally does not suggest a
crime and therefore does not, without more, furnish the suspicion or belief necessary
for a seizure. See Illinois v. Wardlow, 528 U.S. 119, 125 (2000) (“[R]easonable
suspicion must be based on commonsense judgments and inferences about human
behavior.”).
Conduct can be legal—and therefore insufficient, standing alone, to justify a
seizure—in one of two ways: First, the conduct might happen to be legal in a certain
jurisdiction. In a city that permits skateboarding, for example, an officer cannot
seize an individual merely for riding a skateboard because, without more, that legal
conduct would not indicate “that criminal activity may be afoot.” See United States
v. Arvizu, 534 U.S. 266, 273 (2002) (quotation omitted). Second, some conduct
does not raise red flags because it must always be legal. In a traditional public forum,
an officer cannot seize a person merely for praying the Rosary, pleading the Fifth,
or even flying a Nazi flag because that conduct is constitutionally protected, cannot
of the Court, supported in concurrences by Brennan, J., and White, J.) (“[S]ilent and
reproachful presence” is protected by the First Amendment).

-- 20 of 22 --

-21-
be illegal, and does not, without more, suggest criminal activity.9 See R.A.V. v. St.
Paul, 505 U.S. 377, 381 (1992) (Nazi swastikas could not be outlawed “solely on
the basis of the subjects the speech addresses”). See also Baribeau, 596 F.3d at 479
(narrowing a state prohibition to exclude First Amendment-protected activity and
holding that “there was no probable cause to arrest the plaintiffs for engaging in
protected expressive conduct”).
Chestnut and Walker found it clearly established that peacefully observing a
police officer did not furnish reasonable suspicion (or probable cause) of criminal
activity and therefore did not justify a Fourth Amendment seizure. The panel
opinion today presumes that peaceful police-observation fits the first category of
legal conduct insufficient for a search or seizure—conduct that just happens to be
legal. But this circuit’s precedent clearly establishes that peaceful police-
observation must be legal—it is constitutionally protected First Amendment activity.
See Chestnut, 947 F.3d at 1090; Walker, 414 F.3d at 993.
Chestnut and Walker were indeed Fourth Amendment cases. But the Fourth
Amendment makes it unreasonable to arrest or detain someone for conduct that,
because the constitution protects it, could never be criminal. Chestnut and Walker,
in reaching their Fourth Amendment holdings, clearly established that the First
Amendment protects peaceful police-observation in this circuit. This panel is
“powerless” to overturn those holdings. Kostelec v. State Farm Fire & Cas. Co.,
64 F.3d 1220, 1228 n.8 (8th Cir. 1995). Because Molina and Vogel were engaged
in protected First Amendment activity, their First Amendment claims should be
allowed to proceed.
9 True, context may render suspicious otherwise-innocuous activity. Cortez,
449 U.S. at 417–18 (“[T]he totality of the circumstances—the whole picture—must
be taken into account. Based upon that whole picture, the detaining officers must
have a particularized and objective basis for suspecting the particular person stopped
of criminal activity.”).

-- 21 of 22 --

-22-
I dissent from Part II of the panel opinion here, but fully concur in Part III.
__________________________

-- 22 of 22 --