The opinion
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 1 of 56
[PUBLISH]
In the
United States Court of Appeals
For the Eleventh Circuit
____________________
No. 23-10478
____________________
UNITED STATES OF AMERICA,
Plaintiff-Appellee,
versus
JOSEPH DELEON,
a.k.a. Joseph Nieves,
Defendant-Appellant.
____________________
Appeal from the United States District Court
for the Middle District of Florida
D.C. Docket No. 8:18-cr-00325-MSS-JSS-1
____________________
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 2 of 56
23-10478 Opinion of the Court 2
Before ROSENBAUM, NEWSOM, and ABUDU, Circuit Judges.
ROSENBAUM, Circuit Judge:
Our prior-precedent rule requires us to follow Eleventh Cir-
cuit precedent—even if we disagree with it or think that prior pan-
els have overlooked important arguments—unless and until the
Supreme Court or our court sitting en banc abrogates the prece-
dent. This is one of those cases where the prior-precent rule con-
trols the outcome.
The United States Sentencing Guidelines impose a two-level
enhancement, as relevant here, for an armed robbery, “if any per-
son was physically restrained to facilitate commission of the offense
or to facilitate escape.” U.S. Sent’g Guidelines Manual §
2B3.1(b)(4)(B). This case requires us to determine whether the vic-
tim was “physically restrained” under the meaning of the Guide-
lines when Defendant Joseph Deleon walked into a store, pointed
a gun at the cashier while demanding money from the register, re-
ceived the money, and then left, all within about one minute.
After careful review of the record and case law, and with the
benefit of oral argument, we are bound to affirm.
I. BACKGROUND
Deleon walked into a convenience store, went up to the
counter, and asked for cigarettes. He handed the cashier $10 to pay
for the cigarettes. But when the cashier opened the register,
Deleon took out a handgun and pointed it at the cashier. A video
of the offense shows that the counter put little—but some—
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 3 of 56
23-10478 Opinion of the Court 3
distance between Deleon and the cashier. The cashier gave Deleon
all the money from the register.
Deleon wasn’t satisfied. So he repeatedly signaled for the
cashier to look again in the register, including by reaching over the
counter to point into the register. So the cashier kept looking. Out
of cash, the cashier gave Deleon $40’s worth of United States postal
stamps. At no point does the record indicate that Deleon actually
touched the cashier. Deleon then left the store. Based on the
video, the entire incident lasted about one minute.
A grand jury indicted Deleon for Hobbs Act robbery, in vio-
lation of 18 U.S.C. § 1951(a) and (b) (Count 1), and brandishing a
firearm during and in relation to a crime of violence, in violation of
18 U.S.C. § 924(c)(1)(A)(ii) (Count 2). After a trial, a jury convicted
Deleon on both counts. The district court sentenced Deleon to 240
months’ imprisonment (156 months on Count 1 and 84 months on
Count 2, running consecutively). Over Deleon’s objection, this
sentence included a two-level physical-restraint enhancement un-
der section 2B3.1(b)(4)(B) of the Guidelines. To support this en-
hancement, the district court found that Deleon had physically re-
strained the cashier by pointing a gun at him and forcing him to
empty the register.
Deleon appealed that sentence but did not challenge the
physical-restraint enhancement. The Eleventh Circuit vacated the
sentence and remanded for resentencing on the ground that the
district court had added a career-offender enhancement even
though Hobbs Act robbery is not a crime of violence for purposes
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 4 of 56
23-10478 Opinion of the Court 4
of that enhancement. United States v. Deleon, 812 F. App’x 948, 952
(11th Cir. 2020) (per curiam).
On remand, the probation officer again recommended the
two-level enhancement for physical restraint under section
2B3.1(b)(4)(B). Deleon again objected to the enhancement, argu-
ing that his conduct did not rise to the level of “physical restraint.”
And the district court again rejected Deleon’s arguments and im-
posed the physical-restraint enhancement.
In support of its determination, the court explained,
[Deleon] did pull out a gun. He held it to the victim.
He forced the victim to comply by passing him the
goods he sought to secure, and the victim clearly in
this case did not feel at liberty to leave in light of the
presence of the weapon. In fact, as I recall, the victim
was so entranced by and terrorized by the weapon
that that was all he could remember. And so the
[c]ourt finds under the circumstances of this case, the
use of the weapon did ensure compliance by the vic-
tim and did impede the victim from fleeing to another
part of the store and restrained the victim to the space
where he was standing in order to meet the demands
of Mr. Deleon.
Sent’g Hr’g 10, Sept. 15, 2021.
Then the court resentenced Deleon to 135 months in prison
(51 months as to Count 1 and 84 months as to Count 2, running
consecutively). Relevant here, the total offense level for Count 1
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 5 of 56
23-10478 Opinion of the Court 5
(robbery) under the Guidelines was twenty-two. That total offense
level included a base level of twenty plus the two-level physical-
restraint enhancement. Count 2 (using a firearm during a crime of
violence) added an additional seven years under 18 U.S.C. §
924(c)(1)(A)(ii).
Deleon timely appealed.
II. DISCUSSION
On appeal, Deleon challenges his sentence based solely on
the district court’s application of the two-level sentencing enhance-
ment for physical restraint under section 2B3.1(b)(4)(B). He argues
that if the enhancement applies here, it necessarily applies in all
armed robbery cases, effectively raising the base level of armed rob-
bery to twenty-two instead of twenty.
We first addressed the scope of the physical-restraint en-
hancement in United States v. Jones, 32 F.3d 1512 (11th Cir. 1994).
There, we considered the commentary. See id. at 1518. We noted
that the commentary defined “physically restrained” as “the forci-
ble restraint of the victim such as by being tied, bound, or locked
up.” Id. (citation omitted). Then we reasoned that “such as” in
that definition “indicates that the illustrations of physical restraint
are listed by way of example rather than limitation.” Id. (citation
and quotation marks omitted). With that in mind, we adopted the
Seventh and Eighth Circuits’ understanding of the definition of
“physically restrained”: “a defendant physically restrains his victims
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 6 of 56
23-10478 Opinion of the Court 6
if he creates circumstances allowing the persons no alternative but
compliance.” Id. at 1519 (citation and quotation marks omitted).
Several years later, in United States v. Victor, 719 F.3d 1288
(11th Cir. 2013), we built on what we said in Jones. There, the de-
fendant pretended he had a gun in his pocket and pointed it at a
bank employee. Id. at 1289. Then he herded the employee to the
teller line and yelled that he had a gun and would kill any employee
who did not comply with his demands. Id. We affirmed applica-
tion of the enhancement. In doing so, we explained that “by threat-
ening the [bank] employee with what the employee believed to be
a gun to prevent her from escaping,” the defendant physically re-
strained her under Jones. Id. at 1290. Even though he did not actu-
ally have a gun, we reasoned, the victim believed he did “so that
she was forced to comply.” Id. And when the defendant argued
that he had not moved the teller far, we noted that section
2B3.1(b)(4)(b) “contains no requirement that the victim be moved
at all.” Id. In short, we said, the enhancement applies when “the
defendant’s conduct ensured the victims’ compliance and effec-
tively prevented them from leaving a location.” Id. (citation and
quotation marks omitted).
Most recently, in United States v. Ware, 69 F.4th 830 (11th Cir.
2023), cert. denied, No. 23-5946, 2024 WL 1706044 (Apr. 22, 2024),
we invoked Jones and Victor and again said that section
2B3.1(b)(4)(B) applies “where a defendant creates circumstances al-
lowing [his victims] no alternative but compliance.” Id. at 854 (ci-
tation and quotation marks omitted).
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 7 of 56
23-10478 Opinion of the Court 7
In dicta, we offered that “[w]e would have a closer case if
[the defendant] had entered these establishments, pointed a gun at
the victim behind the welcome counter, demanded, ‘Your money
or your life,’ obtained money from the victim, and left without fur-
ther incident.” Id. at 855. But under the test we again approved in
Ware—the enhancement applies “where a defendant ‘creates cir-
cumstances allowing [his victims] no alternative but compli-
ance’”—we don’t see how the enhancement wouldn’t apply to the
then-hypothetical scenario we raised in Ware. Id. at 854. In that
case, the robber would have created circumstances allowing his
victim no realistic alternative but compliance. After all, who would
think compliance was optional in that situation?
And that hypothetical situation from Ware is nearly identical
to the fact pattern here. As we’ve noted, here, Deleon pointed a
gun at the cashier and demanded the cashier provide him with
money. Though he didn’t say, “Your money or your life,” that
message surely came through loud and clear, all the same. So
Deleon “create[d] circumstances allowing [the cashier] no alterna-
tive but compliance.” And under our precedent, the district court
properly applied section 2B3.1(b)(4)(B)’s enhancement for “physi-
cal restraint.”
Deleon argues that, despite our precedent, the enhancement
should not apply here. First, he asserts that applying the enhance-
ment to these facts means it will apply to virtually every armed
robbery, effectively raising the base offense level to twenty-two in-
stead of twenty. Deleon also argues that applying the physical-
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 8 of 56
23-10478 Opinion of the Court 8
restraint enhancement when a defendant demands money at gun-
point during an armed robbery punishes him a second time for the
same conduct as his sentence for his conviction under §
924(c)(1)(A)(ii) for brandishing a firearm during a crime of violence.
In other words, Deleon makes a double-counting argument.
The short answer to both of Deleon’s arguments is this: as
we’ve explained, our precedent binds us to conclude that the en-
hancement applies to conduct like Deleon’s, and “we have categor-
ically rejected an overlooked reason or argument exception to the
prior-panel-precedent rule.” In re Lambrix, 776 F.3d 789, 794 (11th
Cir. 2015) (per curiam). End of story.
To the extent that Deleon argues that applying the physical-
restraint enhancement on the facts of his particular case wrongly
double-counts the harms from his crime, we must also reject that
argument.
In support of his double-counting argument, Deleon points
to the Guidelines commentary under section 2K2.4 that precludes,
in part, application of any specific offense characteristic for bran-
dishing or using a firearm if the defendant is also sentenced under
§ 924(c). He points out that the probation officer did not recom-
mend applying the six-level enhancement for when “a firearm was
otherwise used” under section 2B3.1(b)(2)(B) because of the section
2K2.4 commentary. By the same logic underlying the section
2K2.4 commentary, Deleon reasons that we shouldn’t apply the
physical-restraint enhancement because a § 924(c) conviction al-
ready addresses the conduct of brandishing a gun.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 9 of 56
23-10478 Opinion of the Court 9
Double counting can be permissible or impermissible.
We’ve explained that “[i]mpermissible double counting occurs
only when one part of the Guidelines is applied to increase a de-
fendant’s punishment on account of a kind of harm that has already
been fully accounted for by application of another part of the Guide-
lines.” United States v. Dudley, 463 F.3d 1221, 1226–27 (11th Cir.
2006) (emphasis added) (citation omitted).
In contrast, “double counting is permissible” when “(1) the
Sentencing Commission intended the result; and (2) each guideline
section in question concerns a conceptually separate consideration
related to sentencing.” United States v. Grushko, 50 F.4th 1, 16 (11th
Cir. 2022). And we “presume that the Sentencing Commission in-
tended to apply separate guideline sections cumulatively unless we
are specifically directed otherwise.” Id. For these reasons, we’ve
recognized that the same conduct can implicate “multiple guide-
lines sections.” Id.; see also United States v. Suarez, 893 F.3d 1330,
1337 (11th Cir. 2018) (“It is irrelevant that [two sentencing] adjust-
ments will often be triggered by the same conduct.”).
Looking to the guidelines at issue here, the sentence for
brandishing a gun and the enhancement for physically restraining
a victim do not necessarily and fully cover the same conduct. For
instance, an armed robber who enters a store with a gun on his
waistband and heads straight for the cash register without blocking
any exit or demanding assistance with the cash register may have
brandished a gun, but he likely has not “physically restrained” his
victims—even under our Circuit’s broad definition of the phrase.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 10 of 56
23-10478 Opinion of the Court 10
And though the commentary to section 2K2.4 directs courts
not to apply enhancements “for possession, brandishing, use, or
discharge” of a firearm when a separate sentence for a § 924(c) vio-
lation applies, the Guidelines include no similar commentary di-
recting us not to apply the physical-restraint enhancement in this
circumstance. Nor do we see any other indication that the Sen-
tencing Commission specifically directed us not to apply these dis-
tinct sentence components cumulatively. See United States v. Ste-
vens, 580 F.3d 718, 722 (8th Cir. 2009) (“The specific offense charac-
teristic punished by § 2B3.1(b)(4)(B) is the use of physical restraint,
not the presence of a firearm. The firearm was merely a tool used
to effect the physical restraint accomplished by [defendants] . . . .
Consequently, . . . the challenged sentencing enhancement does
not amount to double counting.”); United States v. Wilson, 198 F.3d
467, 472 n.1 (4th Cir. 1999) (explaining that “‘double counting’ is
permissible under the sentencing guidelines except where it is ex-
pressly prohibited,” and the Guidelines do not expressly prohibit a
physical-restraint enhancement on top of a sentence for a § 924(c)
conviction).
For all these reasons, we must reject Deleon’s claim that sec-
tion 2B3.1(b)(4)(B)’s physical-restraint enhancement does not apply
to his conduct. And we therefore affirm the district court.
III. CONCLUSION
We AFFIRM the district court’s application of the physical-
restraint enhancement.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 11 of 56
23-10478 ROSENBAUM, J., Concurring 1
ROSENBAUM, Circuit Judge, joined by ABUDU, Circuit Judge, con-
curring:
The issue in this case is ripe for en banc review. As the Ma-
jority Opinion explains, the Jones/Victor line of cases binds us to af-
firm the application of the physical-restraint enhancement. That’s
because our precedent has understood the enhancement to apply
any time “the defendant’s conduct ensured the victims’ compliance
and effectively prevented them from leaving a location.” United
States v. Victor, 719 F.3d 1288, 1290 (11th Cir. 2013) (citation and
quotation marks omitted). But if I were writing on a clean slate, I
would conclude that section 2B3.1(b)(4)(B) does not reach Deleon’s
conduct. Quite simply, a plain reading of the text of section
2B3.1(b)(4)(B) does not support our precedent’s standard or appli-
cation of that guideline here.
When we construe a guideline, we start with its text. See
United States v. Dupree, 57 F.4th 1269, 1277 (11th Cir. 2023) (en
banc). The version of section 2B3.1(b)(4)(B) that applied to
Deleon’s conduct (2018) provides, “[I]f any person was physically
restrained to facilitate commission of the offense or to facilitate es-
cape, increase by 2 levels.”
To determine the meaning of this phrase, we “apply[] our
traditional tools of statutory interpretation,” id. at 1277. Only if the
text remains ambiguous after that do we then look to the Guide-
lines commentary. Id. at 1276–77, 1279. 1
1 As the Majority Opinion notes, we relied in part on the commentary to sec-
tion 2B3.1(b)(4)(B) when we adopted the broad definition of “physically
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 12 of 56
23-10478 ROSENBAUM, J., Concurring 2
Following this roadmap, I begin with the ordinary meaning
of “physically restrained,” § 2B3.1(b)(4)(B), the phrase that drives
the meaning of the guideline here. To determine the ordinary
meaning of a term, we often turn to dictionary definitions. In re
Walter Energy, 911 F.3d 1121, 1143 (11th Cir. 2018). “Restrain”
means “[t]o restrict, limit, confine.” 2 Other definitions include the
following: “to prevent from doing, exhibiting, or expressing some-
thing,” “to limit, restrict, or keep under control,” “to moderate or
limit the force, effect, development, or full exercise of,” and “to de-
prive of liberty.” 3
restrained” in Jones and Victor. Yet under Dupree, because the guideline itself
is unambiguous, we shouldn’t have gotten to the commentary. We issued
United States v. Ware, 69 F.4th 830 (11th Cir. 2023), after Dupree, so perhaps we
could have corrected this aspect of our precedent in Ware. But we didn’t do
that. Of course, “when circuit authority is in conflict, a panel should look to
the line of authority containing the earliest case, because a decision of a prior
panel cannot be overturned by a later panel.” MacPhee v. MiMedx Grp., Inc., 73
F.4th 1220, 1250 (11th Cir. 2023) (citation and quotation marks omitted). But
we can’t follow Dupree instead of Ware now as the earlier-issued case because
Ware’s holding—though arrived at in a way contrary to Dupree’s stated meth-
odology—does not conflict with Dupree’s holding. See Washington v. Howard,
25 F.4th 891, 900 (11th Cir. 2022) (explaining that we must reconcile prior
panel decisions whenever possible, and that “only the holdings of prior deci-
sions bind us” and “legal principles set forth outside of the decision’s holding
do not bind us”).
2 Restrain, Oxford English Dictionary, https://www.oed.com/dictionary/re-
strain_v1?tab=meaning_and_use#25713964 [https://perma.cc/9FW3-5UJ2].
3 Restrain, Merriam-Webster Dictionary, https://www.merriam-web-
ster.com/dictionary/restrain [https://perma.cc/TV96-CHJ8].
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 13 of 56
23-10478 ROSENBAUM, J., Concurring 3
All these definitions contemplate some restriction, limita-
tion, or confinement. But none specify the type of restriction, lim-
itation, or confinement.
That’s where “physically” comes in. Section 2B3.1(b)(4)(B)
qualifies the word “restrained” with “physically.” “Physically”
means “[w]ith regard to the body; in bodily terms or by bodily
means,” 4 “in a physical manner: in accord with physical laws,” 5 or
“in respect to the body.” 6 Similarly, the relevant definition of
“physical,” the adjective form of “physically,” is “[o]f or relating to
natural phenomena perceived through the senses (as opposed to
the mind); of or relating to matter or the material word; natural;
tangible; concrete.” 7
Putting it all together, then, under the guideline, a re-
straint—or limitation or confinement—must concern the body for
the enhancement to apply. A restraint that is solely mental or psy-
chological in nature should not qualify under section
2B3.1(b)(4)(B)’s plain text.
4 Physically, Oxford English Dictionary, https://www.oed.com/diction-
ary/physically_adv?tab=meaning_and_use#30459183
[https://perma.cc/S4BN-FN6D].
5 Physically, Merriam-Webster Dictionary, https://www.merriam-web-
ster.com/dictionary/physically [https://perma.cc/6263-7SBF].
6 Id.
7 Physical, Oxford English Dictionary, https://www.oed.com/diction-
ary/physical_adj?tab=meaning_and_use#30455352
[https://perma.cc/Q7QV-XDDK].
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 14 of 56
23-10478 ROSENBAUM, J., Concurring 4
Not only does a non-physical restraint fail to comport with
the definition of “physically,” but construing “physically re-
strained” to include non-physical restraints—no matter how intim-
idating a non-physical restraint may be—makes “physically” mean-
ingless in the guideline.
Indeed, if the framers of the guideline wanted it to apply
whenever “any person was . . . restrained” in either a physical or
non-physical way, they wouldn’t have included the qualifier “phys-
ically.” But the guideline contains the modifier “physically” before
“restrained.” That adverb has meaning. And by its plain meaning,
“physically restrained” should not include psychologically or emo-
tionally “restrained.” After all, “[i]t is ‘a cardinal principle of statu-
tory construction’” that, “‘if it can be prevented, no clause, sen-
tence, or word shall be superfluous, void, or insignificant.’” TRW
Inc. v. Andrews, 534 U.S. 19, 31 (2001) (quoting Duncan v. Walker,
533 U.S. 167, 174 (2001)). See Antonin Scalia & Bryan Garner, Read-
ing Law: The Interpretation of Legal Texts 176 (2012) (“Because legal
drafters should not include words that have no effect, courts avoid
a reading that renders some words altogether redundant.”).
When we give every word its meaning in section
2B3.1(b)(4)(B), I understand this enhancement to reflect Congress’s
view that physical bounding adds another dimension to the intim-
idation a victim of an armed robbery endures. In other words, it’s
bad enough to be robbed at gunpoint, but being tied up or other-
wise physically bound is invasive, dangerous, and victimizing in yet
another way.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 15 of 56
23-10478 ROSENBAUM, J., Concurring 5
Given that the text is clear, if I were writing on a clean slate,
I would end with the text as well as starting with it. See Dupree, 57
F.4th at 1276–77. And I would not apply the enhancement in
Deleon’s case. But even if the text were ambiguous (it’s not), the
application notes and commentary are wholly consistent with and
fully support the plain meaning of the text. See id. 8
The Application Notes define “physically restrained” as “the
forcible restraint of the victim such as by being tied, bound, or
locked up.” § 2B3.1 cmt. n.1 (citing § 1B1.1 cmt. n.1(L)). And sep-
arately, the commentary states that the “guideline provides an en-
hancement for robberies where a victim . . . was physically re-
strained by being tied, bound, or locked up.” § 2B3.1 cmt. back-
ground. So just as the text itself conveys, the commentary and
8 Dupree, 57 F.4th 1269, sets forth the process for interpreting the Guidelines.
It relied on the deferential scheme afforded to agency interpretations of their
own legislative rules when those rules are truly ambiguous. Id. at 1275–77.
Recently, the Supreme Court held that, where Congress has not conferred dis-
cretionary authority on an agency to interpret a federal statute, the Adminis-
trative Procedure Act, 5 U.S.C. § 706, does not permit courts to “defer to an
agency interpretation of the law simply because a statute is ambiguous.” Loper
Bright Enters. v. Raimondo, 144 S. Ct. 2244, 2263, 2268, 2273 (2024). To be sure,
that is not squarely on point with Dupree, which addressed the proper inter-
pretation of the Sentencing Guidelines and their commentary and application
notes, or the cases Dupree relied on. But to the extent Loper Bright casts any
doubt on Dupree’s methodology, that may be another good reason for us to
consider rehearing this case en banc.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 16 of 56
23-10478 ROSENBAUM, J., Concurring 6
application notes reflect that section 2B3.1 contemplates enhance-
ment upon only the use of physical restraints. 9
Were we to apply this understanding of section
2B3.1(b)(4)(B) to Deleon’s case, I would agree with Deleon that the
enhancement does not apply. Deleon pointed a gun at the cashier
to obtain his compliance. I can certainly understand why the cash-
ier did not try to leave under those circumstances. But I cannot say
that the pointed gun “physically” restrained the cashier to facilitate
Deleon’s crime because it did not physically bind, confine, or apply
any force to the cashier’s body. So by its plain meaning, section
2B3.1(b)(4)(B) should not apply. 10
9 As the Majority Opinion notes, in United States v. Jones, 32 F.3d 1512 (11th
Cir. 1994), we based our decision to so broadly define “physically restrained”
on the commentary’s use of “such as.” Id. at 1518. We said that phrase indi-
cates that the “illustrations of physical restraint are listed by way of example
rather than limitation.” Id. at 1518. I agree that the phrase “such as” means
that a person can engage in conduct that warrants the enhancement other than
through the modes the commentary lists. But most respectfully, I don’t see
how the phrase “such as” somehow permits broadening the scope of the
guideline beyond “physical” restraint. The doctrine of noscitur a sociis—that
is, a word is known by the company it keeps—tells us that a word is “given
more precise content by the neighboring words with which it is associated.”
Fischer v. United States, 144 S. Ct. 2176, 2183 (2024) (citation and quotation
marks omitted). So while the phrase “such as” denotes that the enhancement
contemplates the inclusion of conduct other than that listed, other qualifying
conduct must be similar to the conduct that the commentary enumerates.
And all that conduct is physical in nature.
10 Of course, that would not prevent a sentencing court from considering
Deleon’s gun-pointing conduct in crafting an appropriate sentence; the
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 17 of 56
23-10478 ROSENBAUM, J., Concurring 7
At least five of our sister circuits construe section
2B3.1(b)(4)(B) more narrowly than we do and require some type of
“physical” restraint. Take United States v. Anglin, 169 F.3d 154 (2d
Cir. 1999). There, the Second Circuit held that “displaying a gun
and telling people to get down and not move, without more, is in-
sufficient to trigger the ‘physical restraint’ enhancement.” Id. at
164.
The court gave three reasons. First, it explained that the
conduct there was “materially different from the Guidelines exam-
ples” of physical restraint, namely, being “tied,” “bound,” or
“locked up.” Id. Second, it said that the plain meaning of “physical
restraint” requires more than pointing a gun and giving directions
not to move. Id. The court reasoned that the Guidelines would
not have included the word “physical” in section 2B3.1(b)(4)(B) if
the enhancement applies even when the victim believes they are re-
strained but are not actually “subjected to physical restraint.” Id.
at 164–65. Third, the court found that if the enhancement applied
in that case, “virtually every robbery would be subject to the 2-level
enhancement for physical restraint unless it took place in unoccu-
pied premises.” Id. at 165. In the Second Circuit’s view, that would
be a “problematic effect for a provision drafted to deal with a spe-
cial circumstance.” Id.
Similarly, in United States v. Garcia, 857 F.3d 708 (5th Cir.
2017), the Fifth Circuit held that the physical-restraint
sentencing court simply could not apply section 2B3.1(b)(4)(B)’s two-point en-
hancement.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 18 of 56
23-10478 ROSENBAUM, J., Concurring 8
enhancement did not apply when the defendants held firearms, one
defendant pointed a gun at an employee and told them to get on
the ground, and another stood by the exit. Id. at 712. The court
explained that the defendants “never forced [the victims] to move
to a confined space” or subjected the victims to “the type of physical
restraint that victims experience when they are tied, bound, or
locked up.” Id. at 713. Instead, the court reasoned that the “de-
fendants did not do anything with their firearms that goes beyond
what would normally occur during an armed robbery,” so the en-
hancement did not apply. Id. at 713–14. The Fifth Circuit cited
Anglin with approval. Id. at 713.
The Seventh, Ninth, and D.C. Circuits also require some-
thing more than pointing a gun and giving a command to justify the
physical-restraint enhancement. See United States v. Herman, 930
F.3d 872, 875–77 (7th Cir. 2019) (holding that pointing a gun at a
victim and ordering them not to move does not qualify as physical
restraint); United States v. Parker, 241 F.3d 1114, 1118–19 (9th Cir.
2001) (holding that the enhancement requires “more than briefly
pointing a gun at a victim and commanding her once to get down
to constitute physical restraint, given that nearly all armed bank
robberies will presumably involve such acts”); United States v. Drew,
200 F.3d 871, 880 (D.C. Cir. 2000) (holding the enhancement did
not apply even where the defendant forced the victims to walk at
gunpoint because, even though the victims felt restrained, they
were not physically restrained, and “[a]ny other interpretation
would effectively add the two-level adjustment to almost any
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 19 of 56
23-10478 ROSENBAUM, J., Concurring 9
attempted murder because presumably any victim would feel re-
strained if directed to move at gunpoint”).
In contrast, by my count, four other circuits (besides ours)
construe the physical-restraint enhancement broadly to apply
when a defendant restricts a victim’s movement by using a gun,
even in the absence of bodily contact, confinement, or other phys-
ical force.11 See United States v. Wallace, 461 F.3d 15, 34–35 (1st Cir.
2006) (affirming application of the physical-restraint enhancement
where the defendants pointed their guns at the victims, told them
not to move, and blocked one victim’s path); United States v. Di-
mache, 665 F.3d 603, 608–09 (4th Cir. 2011) (affirming application
of the physical-restraint enhancement where the defendant used a
gun to force bank tellers to the floor because they “were prevented
from both leaving the bank and thwarting the bank robbery”);
United States v. Howell, 17 F.4th 673, 692 (6th Cir. 2021) (affirming
application of the physical-restraint enhancement where the de-
fendant ordered the victim at gunpoint to lie on the floor and
threatened to kill her unless her coworkers complied with orders);
United States v. Miera, 539 F.3d 1232, 1235–36 (10th Cir. 2008) (af-
firming application of the physical-restraint enhancement where
11 The Third Circuit takes yet another approach when considering whether
conduct qualifies as physical restraint. See United States v. Bell, 947 F.3d 49, 56
(3d Cir. 2020) (adopting five factors to balance when deciding whether to im-
pose the enhancement, namely, “[1] Use of physical force; [2] Exerting control
over the victim; [3] Providing the victim with no alternative but compliance;
[4] Focusing on the victim for some period of time; and [5] Placement in a
confined space”).
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 20 of 56
23-10478 ROSENBAUM, J., Concurring 10
the defendant waved a gun around, told people not to move, and
blocked the bank doors).
As the circuit split deepens on the proper interpretation of
the physical-restraint enhancement, Deleon’s case provides a good
opportunity to reexamine our jurisprudence. Given the heavier
emphasis on textualism in the last several years—and especially
against the background of Dupree—we should align our jurispru-
dence with what section 2B3.1(b)(4)(B) actually says.
Because our jurisprudence on § 2B3.1(b)(4)(B) is not faithful
to the text of the physical-restraint enhancement, I respectfully sub-
mit that we should revisit it en banc.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 21 of 56
23-10478 NEWSOM, J., Concurring 1
NEWSOM, Circuit Judge, concurring:
This is a sequel of sorts to my separate opinion in Snell v.
United Specialty Insurance Co., 102 F.4th 1208 (11th Cir. 2024) (New-
som, J., concurring). There, I floated the following suggestion,
which I suspected (rightly, as it turns out) would be fairly provoca-
tive: “Those, like me, who believe that ‘ordinary meaning’ is the
foundational rule for the evaluation of legal texts should consider—
consider—whether and how AI-powered large language models like
OpenAI’s ChatGPT, Google’s Gemini, and Anthropic’s Claude
might—might—inform the interpretive analysis.” Id. at 1221. With
the benefit of a little perspective, and incorporating by reference
here all the caveats that I expressed there, I stand by what I said.
The LLM research that I conducted in connection with this
case, though, raised a question that I hadn’t squarely confronted in
Snell, one that seemed worth flagging and exploring: What should
we make of the fact that the models sometimes provide subtly dif-
ferent answers to the exact same question? I’ll confess that the var-
iation initially spooked me, but it now (for reasons I’ll get into)
seems not only unremarkable but perhaps even expected. And sig-
nificantly from an ordinary-meaning perspective, it accurately re-
flects real people’s everyday speech patterns.
Let me explain.
I
First, a bit of background. As the majority opinion explains,
the district court here enhanced Deleon’s sentence on the ground
that he “physically restrained” one of his victims during the course
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 22 of 56
23-10478 NEWSOM, J., Concurring 2
of a robbery. U.S.S.G. § 2B3.1(b)(4)(B). I agree with the majority
that our decisions in United States v. Jones, 32 F.3d 1512 (11th Cir.
1994), and United States v. Victor, 719 F.3d 1288 (11th Cir. 2013),
control this case and require us to affirm the enhancement. See
Maj. Op. at 5–7. Having said that, I also agree with Judge Rosen-
baum that Jones and Victor were wrongly decided and that we
should rehear this case (or another one like it) as an en banc court
to overrule them. See Rosenbaum Conc. Op. passim.
In Jones and Victor, panels of this Court held that
§ 2B3.1(b)(4)(B)’s enhancement applies where, even in the absence
of any physical contact, the defendant’s possession of a gun “en-
sured the victims’ compliance and effectively prevented them from
leaving,” Jones 32 F.3d at 1519, or the defendant “threaten[ed a vic-
tim] with what [she] believed to be a gun to prevent her from es-
caping,” Victor, 719 F.3d at 1290. I disagree. For reasons that I’ll
explain—and that complement those Judge Rosenbaum has set out
in her separate opinion—neither circumstance constitutes “physi-
cal[] restrain[t]” within the ordinary meaning of that phrase. See
Rosenbaum Conc. Op. at 1–3.
One thing that makes the interpretive issue in this case in-
teresting—and perhaps a little bit harder than usual—is that it turns
on the meaning not of a single word, but rather of a composite,
multi-word phrase: “physically restrained.” That poses a challenge
because so far as I’m aware, the entire phrase—“physically re-
strained”—isn’t defined in any reputable dictionary, the tool to
which plain-language interpreters typically turn. Given that
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 23 of 56
23-10478 NEWSOM, J., Concurring 3
limitation, I think Judge Rosenbaum’s concurrence nicely exempli-
fies the standard approach. Specifically, we break the phrase into
its constituent parts, use conventional interpretive tools to gauge
the ordinary meaning of each, and then piece those definitions back
together into a coherent whole. See, e.g., Garland v. Cargill, 602 U.S.
406, 415–16 (2024) (using the breaking-and-repiecing method to de-
fine the phrase “function of the trigger”); Barton v. U.S. Att’y Gen.,
904 F.3d 1294, 1298–99 (11th Cir. 2018) (Newsom, J.) (same, defin-
ing the phrase “renders admissible”). So, Judge Rosenbaum ex-
plains, “physically restrained” means, in essence, the state of being
“restrained”—i.e., “restrict[ed], limit[ed], confine[d]”—by “physi-
cal[]” means—i.e.,“[w]ith regard to the body.” See Rosenbaum
Conc. Op. at 2–3 (citing dictionary definitions of both terms). 1
1 Worth filing away for another day: The breaking-and-repiecing approach
won’t always work so well. To be sure, sometimes a phrase really is just the
sum of its parts—say, “strictly enforced” or “accurately measured.” But that’s
not uniformly true. Consider, for instance, common idioms like “kick the
bucket” or “spill the beans.” Or metaphors like “falling in love” or “drowning
in work.” Or, even beyond those, plain old linguistic combinations like “white
noise” or “dark matter.” Most have a sense for what those phrases denote, but
their meanings wouldn’t be self-evident from splicing together dictionary def-
initions of each constituent word. “White noise,” for example, isn’t a “[l]oud,
confused, or senseless shouting or outcry,” or even a “sound of any sort,” that
is “[o]f the color of pure snow, milk, or sunlight.” Webster’s New International
Dictionary 1658, 2915 (2d ed. 1944) (defining both terms). Again, I think Judge
Rosenbaum has shown that the breaking-and-repiecing method works well
enough for “physically restrained.” My point is simply that it won’t always do
the trick.
In those instances where a phrase is more than the sum of its compo-
nent parts, I think that LLMs may well help to fill the gaps left by word-centric
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 24 of 56
23-10478 NEWSOM, J., Concurring 4
That seems about right to me. But because there’s no ready
dictionary definition of the composite phrase, because by their very
nature LLMs aim to capture and reflect how real people ordinarily
dictionaries. The reason: They aren’t dependent on human beings going to
the trouble of identifying and then defining commonly used phrases. There
are no doubt others who can explain the technical details better than me, but
as I understand things, the models’ ability to tackle this task—i.e., to make
predictions about the meaning of phrases, in context—is hard-wired into their
DNA, inherent in what they’re designed to do. See Snell, 102 F.4th at 1226–30
(Newsom, J., concurring); see also Yonathan Arbel & David A. Hoffman, Gen-
erative Interpretation, 99 N.Y.U. L. Rev. 451, 479 (2024) (explaining how the at-
tention mechanism helps LLMs understand phrases like “he swung the bat”).
Why? Well, consider several of the uses to which people frequently put these
models in professional settings: (1) summarizing information contained in
books or articles; (2) brainstorming ideas for a business meeting; and (3)
providing feedback on written work. Cf. Kevin Roose, How Should I Use AI
Chatbots Like ChatGPT?, N.Y. Times (Mar. 30, 2023), https://www.ny-
times.com/2023/03/30/technology/ai-chatbot-chatgpt-uses-work-life.html.
In order for an LLM’s output to be useful to everyday consumers, the model
has to be able to understand and distinguish between the different uses of a
phrase like, say, “passed out.” Suffice it to say, an LLM that can’t ascertain
that “passed out” means something different in the sentences (1) “The teacher
passed out the writing assignment” and (2) “The teacher passed out on the floor”
wouldn’t be very useful. And yet—and this is the important point—the mod-
els can (with uncanny accuracy) draw that distinction because they’ve
crunched enough raw linguistic data to “understand,” based on contextual
clues, how the phrase is being used in a particular circumstance.
Anyway, perhaps I’ll revisit this issue down the road, in a case in which
the breaking-and-repiecing method of explaining a compound phrase’s mean-
ing comes up short.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 25 of 56
23-10478 NEWSOM, J., Concurring 5
use both individual words and multi-word phrases, and because,
well, I couldn’t help myself, I queried ChatGPT. 2
Here’s what happened.
II
Me: “What is the ordinary meaning of ‘physically re-
strained’?”
ChatGPT:
The ordinary meaning of “physically restrained” re-
fers to the act of limiting or preventing someone’s
movement by using physical force or some kind of
device. This can involve holding someone down, us-
ing handcuffs, straps, or other mechanisms to restrict
their ability to move freely. The term implies an ac-
tive and physical intervention to control or confine a
person’s movements.
So far, so good. ChatGPT’s response basically squared with what
I had assumed the phrase “physically restrained” means. And as it
turned out, its answer also corresponded with the result of the tra-
ditional breaking-and-repiecing method exemplified in Judge Ros-
enbaum’s concurrence. See Rosenbaum Conc. Op. at 2–3.
For reasons I’ve explained before, though, I didn’t want to
seize on ChatGPT’s explanation just because it matched my priors.
See Snell, 102 F.4th at 1225 (Newsom, J., concurring). So, I asked
2 To be more precise, I queried the free online version, which uses GPT4-o.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 26 of 56
23-10478 NEWSOM, J., Concurring 6
Claude 3 the same question: “What is the ordinary meaning of
‘physically restrained’?” Reassuringly, Claude’s response largely
mirrored ChatGPT’s:
The ordinary meaning of “physically restrained” gen-
erally refers to:
1. The act of using physical force or methods to limit
or control someone’s movements or actions.
2. A state in which a person is prevented from moving
freely or acting as they wish due to physical barriers
or interventions.
This can involve various forms of restraint, such as:
• Holding someone back with one’s hands or
body
• Using devices like handcuffs, straps, or re-
straining jackets
• Confining someone to a specific space, like a
locked room
The term is often used in contexts related to law en-
forcement, medical settings, or situations where
someone’s freedom of movement needs to be re-
stricted for safety or security reasons. However, it can
3 Specifically, Claude 3.5 Sonnet—again, the free version.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 27 of 56
23-10478 NEWSOM, J., Concurring 7
also be used in more general situations where physi-
cal limitation is applied.
Then, for reasons I don’t specifically recall—but that can
presumably be chalked up to a “better safe than sorry” instinct—I
asked Claude the exact same question again. To my surprise, its
answer the second time around was ever-so-slightly different:
The ordinary meaning of “physically restrained” typ-
ically refers to:
Using physical means to limit or prevent someone’s
movement or actions.
This usually involves:
• Holding someone back
• Using restraining devices (e.g. handcuffs)
• Confining someone to a limited space
It implies direct physical intervention to restrict a per-
son’s freedom of movement.
Hmm. Claude’s responses were basically the same—but not
exactly, not verbatim. The substance seemed pretty much identi-
cal, but the style and structure were just a smidgen off. Most obvi-
ously, the first was longer and more detailed than the second. For
instance, unlike the first response, the second didn’t include a “state
in which” explanation. In the “holding someone back” bullet, the
first specified a means—“with one’s hands or body”—while the
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 28 of 56
23-10478 NEWSOM, J., Concurring 8
second didn’t. Similarly, in the “devices” bullet, the first referred
to handcuffs, straps, and restraining jackets as examples, while the
second referred only to handcuffs. Finally, the first response con-
cluded with a short paragraph about the various “contexts” in
which the phrase “physically restrained” is often used—law en-
forcement, medicine, etc.—while the second closed with a more
generic statement that the phrase “implies direct physical interven-
tion to restrict a person’s freedom of movement.”
Honestly, Claude’s varying responses rocked me back on
my heels a little bit. Whatever’s going on inside the LLM’s black
box, I worried about the same model’s failure (inability? refusal?)
to give the same answer to identical back-to-back queries. Might it
suggest that the LLM wasn’t accurately communicating its predic-
tion about how ordinary people use a phrase like “physically re-
strained”—or worse, that it didn’t even really have an accurate pre-
diction? After all, standard printed dictionaries don’t evolve or
change their minds—certainly not within the span of a few sec-
onds. At least within a particular bound edition, their definitions
are static.
Having done some sleuthing, I’m now less concerned than I
once was. There is, I think, both a valid technical explanation for
the variation and, perhaps more importantly for our purposes, a
pretty powerful analogy to the way that ordinary people ordinarily
use phrases in ordinary conversation. I’ll unpack both the explana-
tion and the analogy shortly. First, though, let me describe the
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 29 of 56
23-10478 NEWSOM, J., Concurring 9
results of a little (and I do mean little) experiment that I ran, which
will show, descriptively, the sort of variability that I encountered.
III
Surprised by Claude’s second, slightly different response to
my query, I decided to go both a little broader and a little deeper.
I queried the three leading models—GPT, Claude, and Gemini4—
ten times apiece. Each time, I opened a new chat and asked the
same question: “What is the ordinary meaning of ‘physically re-
strained’?” Not perfectly scientific, I know, but better than a rifle
shot.
My hope was that my humble little mini-experiment might
begin to reveal how “confident” a model is (and we can be) about
its response. Cf. Yonathan Arbel & David A. Hoffman, Generative
Interpretation, 99 N.Y.U. L. Rev. 451, 487–488 (2024); Christoph En-
gel & Richard H. McAdams, Asking GPT for the Ordinary Meaning of
Statutory Terms 15–16 (Max Planck Inst. Discussion Paper No.
2024/5), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=
4718347 [https://perma.cc/7A2K-VLBC]. A series of wildly differ-
ent answers would probably suggest (and would certainly inspire)
a lack of confidence. 5 If the model instead returned essentially
4 Gemini 1.5 Flash—which, again, is free.
5 Because I used the free chat-based interfaces available online, I couldn’t (so
far as I know) adjust the LLMs’ “temperature” settings as some have done in
more formal studies. See, e.g., Arbel & Hoffman, supra, at 487–488; Engel &
McAdams, supra, at 15–16. (“Temperature” is a setting that can influence the
predictability—or alternatively, “creativity”—of a model’s responses.) I opted
for the less technical approach because I think it’s worth exploring how LLMs
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 30 of 56
23-10478 NEWSOM, J., Concurring 10
consistent answers—responses that coalesced around a common
core—then one could probably say, with a higher degree of confi-
dence, that the model was getting at the “correct” response.6
Again reassuringly, the 30 results I received—10 apiece from
each of the three leading LLMs—largely echoed the initial response
can be used in a low-cost, user-friendly way—and thereby, I hope, best approx-
imate how judges, lawyers, and ordinary citizens might typically employ
them.
Having said that, I was able—check that, one of my genius law clerks
was able—to figure out how to use the (more sophisticated, and paywalled)
Open AI Playground. He asked the same question, “What is the ordinary
meaning of ‘physically restrained’?”, and he set the temperature to zero. By
doing so, we were able, I think, to get even a wee bit closer to the model’s
“best guess” at the answer to my question. See Engel & McAdams, supra, at
40 n.128. Playground’s response basically tracked ChatGPT’s:
The ordinary meaning of “physically restrained” refers to the
act of limiting or controlling someone's movements through
physical means. This can involve using physical force, devices,
or other methods to prevent a person from moving freely. Ex-
amples include handcuffs, ropes, straps, or simply holding
someone down with one’s hands. The key aspect is that the
restriction is achieved through physical measures rather than
verbal commands or psychological tactics.
6 I realize that my mini-experiment differed from others’ empirical work in
another respect. Rather than asking about a word’s or phrase’s “ordinary
meaning,” as I did, those studies (at least in part) asked LLMs to apply their
predictions about ordinary meaning to a case’s facts and then render final de-
cisions—e.g., In the circumstances presented by Deleon’s case, “Was anyone
physically restrained?” See, e.g., Arbel & Hoffman, supra, at 487; Engel &
McAdams, supra, at 25. As I’ve said before, nailing down the “proper ques-
tion” is critically important. See Snell, 102 F.4th at 1232–33 (Newsom, J., con-
curring). For better or worse, I tend to default to asking, more generally, for
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 31 of 56
23-10478 NEWSOM, J., Concurring 11
that I got from ChatGPT. If you’re interested in the nitty gritty, all
the responses are available in the Appendix. But here’s the gist:
When defining “physically restrained,” the models all tended to
emphasize “physical force,” “physical means,” or “physical barri-
ers.” ChatGPT and Claude specifically used one (or more) of those
phrases in every one of their responses. For whatever reason,
Gemini was a little different. It didn’t invariably employ one of
those terms explicitly, but even when it didn’t, the concept of what
I’ll call corporeality (via either human touch or a tangible object)
pervaded and tied together its example-laden answers. 7
To be sure, the models’ responses exhibited some minor
variations in structure and phrasing. ChatGPT’s answers, for ex-
ample, tended to fluctuate in length by a sentence or two. For its
part, Claude altered the number of examples it provided from one
response to the next. But for reasons I’ll explain in the next part,
these subtle, marginal divergences were probably (and should have
been) expected. Far more importantly, I think, the responses did
coalesce, substantively, around a common core—there was an ob-
jectively verifiable throughline. For our purposes, what matters is
that the LLMs consistently defined the phrase “physically re-
strained” to require the application of tangible force, either
a term’s ordinary meaning—probably because that’s what I’m doing when I
consult dictionaries. If there’s a better way—a better question—I’m fine with
that.
7 So, for instance, Gemini often mentioned “holding, tying, and locking” as
examples of physical restraints.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 32 of 56
23-10478 NEWSOM, J., Concurring 12
through direct bodily contact or some other device or instrument.
And that, again, squares comfortably with the results obtained
through the traditional, dictionary-driven breaking-and-repiecing
method.
IV
So, what to make of the slight variations among the answers
that the models returned in response to my query? For present
purposes, I think there are two important points. First, there’s a
technical explanation for the variation, which, upon reflection,
doesn’t much concern me—or, upon further reflection, even much
surprise me. Second, there is, upon even further reflection, a sense
in which the substantively-identical-and-yet-marginally-different
answers (perhaps ironically) underscore the models’ utility in the
ordinary-meaning analysis—namely, in that they pretty closely
mimic what we would expect to see, and in fact do see, in everyday
speech patterns.
A
First, the technical details—so far as I understand them. As
it turns out, the LLMs that many of us use—often through a chat-
based interface like ChatGPT—aren’t even designed to produce the
exact same answer every time. Instead, and I’ll confess that I’m
simplifying (and perhaps oversimplifying) things here, the models
have settings that can introduce variation—“creativity,” as many
people call it—into their responses. Again, this is all pretty tech-
nical, but the upshot is that an LLM’s response reflects its best sta-
tistical, probabilistic prediction about the answer to the user’s
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 33 of 56
23-10478 NEWSOM, J., Concurring 13
query. See, e.g., Mathew Renze & Erhan Guven, The Effect of Sam-
pling Temperature on Problem Solving in Large Language Models 1–2
(June 14, 2024) (unpublished manuscript), https://arxiv.org/
pdf/2402.05201. If a model’s creativity dial is turned down, its suc-
cessive responses to identical questions should be very similar, if
not perfectly repetitive. 8 See, e.g., François Chollet, Deep Learning
with Python 374–75 (2d ed. 2021). As it turns out, though, many
LLMs’ default creativity settings are dialed up—which, in a nut-
shell, can subtly adjust the underlying probability distribution that
drives the models’ answers. All of which is to say, the very thing
that I was concerned about—the LLMs’ subtly different responses,
particularly in structure and phrasing—was a feature, not a bug.
It’s how the models are meant to perform—at least for those of us
who (perhaps unwittingly) are using their default settings. 9
Wait, but why? Why are LLMs’ default parameters set to
introduce subtle variation? Doesn’t that seem weird, or even a
8 I’m thinking here principally about an LLM’s “temperature” setting, which
appears to be the most frequently discussed creativity parameter. See supra at
9–10 n.5.
9 I’ll confess that I couldn’t find an authoritative explanation of what the lead-
ing LLMs’ default creativity settings are for those of us who interact with the
models via a chat-based interface like ChatGPT. Based on the subtle variation
I saw, though, I’m assuming they’re not zero—i.e., that the models are intro-
ducing some variation into their responses. And it seems reasonable to as-
sume, as well, that the settings are similar for those who plug into the models
directly. See, e.g., Anthropic, Create a Text Completion, API Reference,
https://docs.anthropic.com/en/api/complete [https://perma.cc/KHX3-
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 34 of 56
23-10478 NEWSOM, J., Concurring 14
little suspicious? Initially, I certainly thought so—I prefer my data
nice and neat, crisp and clean. But upon reflection, I shouldn’t have
been surprised. Think about how and for what purposes most peo-
ple—which is to say, not judges—currently use LLMs. If you’re
brainstorming business ideas, summarizing large masses of infor-
mation, or (dare I say) “writing” a term paper, then variety is, so to
speak, the spice of life. Mindless repetition is bad; a little imagina-
tion is good. So, while I’ll confess that I find it a little frustrating
that I can’t explain, at a granular technical level, exactly why the
models changed their successive responses, however slightly, it
now makes sense to me that they did so. Again, that’s how they’re
designed. 10
5PKA] (last visited Sep. 4, 2024) (explaining that temperature “[d]efaults to
1”).
10 One more thing I should mention briefly before moving on: Without think-
ing about it one way or the other, when I initially queried Claude for the sec-
ond time, I did so within the same chat “string.” I’m afraid, therefore, that I
might have unwittingly tapped into the chatbot’s “conversational history.”
See, e.g., Akash Gupta et al., LLM Task Interference: An Initial Study on the Impact
of Task-Switch in Conversational History 1–2 (Feb. 28, 2024) (unpublished manu-
script), https://arxiv.org/pdf/2402.18216. The basic idea, so far as I under-
stand it—I keep feeling the need to say that—is that an LLM will adjust its
behavior based on the questions the user feeds it, offering clarifications and
refinements along the way. (Rather than claiming that the model is “learning”
from repeated questions, I think it’s probably safer and more accurate to say
that its answers are becoming more “precise” or “useful.” An LLM’s
knowledge base is essentially fixed, based on the underlying data on which it
trains; it doesn’t “update” in response to questions.) Indeed, figuring out how
best to use prompts to better guide LLMs to more accurate results is a new
and valuable skill. See, e.g., Jack Kelly, The Hot, New High-Paying Career Is an AI
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 35 of 56
23-10478 NEWSOM, J., Concurring 15
B
Now, and even more interestingly to me, the upshot for or-
dinary-meaning interpretation: Technical explanations aside, why,
in retrospect, should I have been so unnerved by the models’
slightly (though again, not substantively) different answers? Given
the claims that I’ve made about LLMs’ utility to the interpretive
enterprise, shouldn’t a little variation have been exactly what I ex-
pected?
Remember, our aim is to discern “ordinary meaning.” Pre-
sumably, the ideal gauge of a word’s or phrase’s ordinary meaning
would be a broad-based survey of every living speaker of American
English—totally unrealistic, but great if you could pull it off. Im-
agine how that experiment would go: If you walked out onto the
street and asked all umpteen million subjects, “What is the ordi-
nary meaning of ‘physically restrained’?”, I think I can confidently
guarantee that you would not get the exact same answer spit back
at you verbatim over and over and over. Instead, you’d likely get
a variety of responses that differed around the margins but that,
when considered en masse, revealed a common core. And that
common core, to my way of thinking, is the ordinary meaning.
Prompt Engineer, Forbes (Mar. 6, 2024), https://www.forbes.com/sites/jack-
kelly/2024/03/06/the-hot-new-high-paying-career-is-an-ai-prompt-engi-
neer/ [https://perma.cc/45VA-YJC4]. In any event, as explained in text, I
tweaked my approach for my mini-experiment, opening a new chat for each
successive question.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 36 of 56
23-10478 NEWSOM, J., Concurring 16
So, as it turns out, the very thing that had initially given me
pause—namely, that the LLMs were returning subtly different re-
sponses to the same question—has instead given me (more) hope
that the models have something significant to offer the interpretive
enterprise. The fact is, language is an organic thing, and like most
organic things, it can be a little messy. So too, unsurprisingly, are
our efforts to capture its ordinary meaning. Because LLMs are
trained on actual individuals’ uses of language in the real world, it
makes sense that their outputs would likewise be less than perfectly
determinate—in my experience, a little (but just a little) fuzzy
around the edges. What’s important, though—and I think encour-
aging—is that amidst the peripheral uncertainty, the LLMs’ re-
sponses to my repeated queries reliably revealed what I’ve called a
common core.
Now, I’m not at all sure that we’ve yet come up with the
perfect method for tapping into the models’ seeming ability to
quantify and communicate that core. I’m not even sure that exist-
ing LLMs (warts and all) can realistically accomplish that goal. But
it does strike me that some marginal uncertainty is inherent in the
assessment of ordinary meaning, and I (now) view the fact that
LLMs appear to reflect that uncertainty as a virtue rather than a
vice.
V
So, what are the takeaways?
1. I continue to believe—perhaps more so with each inter-
action—that LLMs have something to contribute to the ordinary-
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 37 of 56
23-10478 NEWSOM, J., Concurring 17
meaning endeavor. They’re not perfect, and challenges remain,
but it would be myopic to ignore them.
2. An important (and perhaps underappreciated) benefit of
incorporating LLMs into our interpretative practice is their ability
to decipher and explain the meaning of composite, multi-word
phrases in a way that standard tools like dictionaries can’t always
do.
3. We should give careful thought to how we assess and
account for LLMs’ sometimes varying answers to user queries. For
reasons I’ve explained, I think (a) that there’s a valid technical ex-
planation for that phenomenon and (b) that the peripheral varia-
tion reflects everyday speech patterns, and thus may actually make
the models more (rather than less) accurate predictors of ordinary
meaning.
4. A final coda: No one should mistake my missives for a
suggestion that AI can bring scientific certainty to the interpretive
enterprise. As I’ve been at pains to emphasize, I’m not advocating
that we give up on traditional interpretive tools—dictionaries, se-
mantic canons, etc. But I do think—and increasingly so—that
LLMs may well serve a valuable auxiliary role as we aim to trian-
gulate ordinary meaning.
Again, just my two cents. I remain happy to be shouted
down.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 38 of 56
23-10478 NEWSOM, J., Concurring 18
APPENDIX
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 39 of 56
23-10478 NEWSOM, J., Concurring 19
ChatGPT
Query: What is the ordinary meaning of “physically restrained”?
Date of query: August 9, 2024
Response 1:
The term “physically restrained” generally refers to the act of lim-
iting or controlling someone’s physical movement or freedom
through the use of physical force, mechanical devices (such as
handcuffs or straps), or by holding them in place. This can occur in
various contexts, such as in law enforcement, healthcare, or even
everyday situations where someone is being prevented from mov-
ing freely. The goal of physical restraint is usually to prevent harm
to the restrained person or to others, but it is often considered a
serious measure and is typically used as a last resort.
Response 2:
The term “physically restrained” generally refers to the act of using
physical force or a device to limit or control someone’s movement
or actions. This can include being held down, tied up, or confined
in a way that restricts the ability to move freely. Physical restraint
is often used in contexts like law enforcement, healthcare, or care-
giving, where it might be necessary to prevent harm to the person
being restrained or to others.
Response 3:
The term “physically restrained” generally refers to the act of lim-
iting or controlling someone’s physical movements or freedom by
using physical force or devices. This can involve methods like tying
someone down, using handcuffs, holding someone back with
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 40 of 56
23-10478 NEWSOM, J., Concurring 20
hands or arms, or employing mechanical devices such as straps or
belts to prevent a person from moving freely. It is often used in
contexts such as law enforcement, medical care, or safety situations
where there is a need to prevent harm to the individual or others.
Response 4:
The ordinary meaning of “physically restrained” refers to the act
of limiting or restricting someone’s movement through physical
means. This could involve the use of physical force, devices like
handcuffs or straps, or even holding someone in place with one’s
hands. The purpose of physical restraint is typically to control or
prevent a person from moving freely, often in situations where they
may pose a danger to themselves or others.
Response 5:
The term “physically restrained” generally refers to the act of lim-
iting or restricting someone’s physical movement through the use
of physical force, devices, or restraints. This can involve anything
from holding someone down with one’s hands, to using handcuffs,
straps, or other mechanical restraints to prevent them from mov-
ing freely.
The context in which the term is used can influence its precise
meaning, but it typically conveys the idea that a person’s ability to
move or act is being intentionally controlled or limited by another
person or mechanism.
Response 6:
The ordinary meaning of “physically restrained” refers to the act
of limiting or controlling someone’s movement by using physical
force or devices. This could involve using handcuffs, straps, or
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 41 of 56
23-10478 NEWSOM, J., Concurring 21
simply holding someone down to prevent them from moving
freely. The term is often used in legal, medical, or security contexts
to describe situations where a person is being restricted from mov-
ing or acting on their own accord.
Response 7:
The ordinary meaning of “physically restrained” refers to the act
of limiting or preventing someone’s physical movement through
the use of physical force, devices, or barriers. This could involve the
use of handcuffs, straps, or simply holding someone in place to pre-
vent them from moving freely. The term is often used in legal,
medical, and security contexts to describe situations where a per-
son is being controlled or confined to prevent harm to themselves
or others.
Response 8:
The ordinary meaning of “physically restrained” refers to a situa-
tion where a person’s physical movement is restricted or controlled
by external forces. This could involve the use of physical barriers,
devices (like handcuffs or straps), or the physical presence of an-
other person who holds or confines the individual to limit their
ability to move freely. The term is often used in contexts where
someone is being held back, prevented from moving, or otherwise
kept under control through physical means.
Response 9:
The ordinary meaning of “physically restrained” refers to the act
of limiting or controlling a person’s movements through physical
means. This can involve the use of physical force, devices, or other
methods to prevent someone from moving freely. For example,
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 42 of 56
23-10478 NEWSOM, J., Concurring 22
this could include being held down by another person, being tied
or bound with ropes or handcuffs, or being confined in a space
where movement is restricted. The term generally implies that the
person is unable to move or act according to their own will due to
these physical limitations.
Response 10:
“Physically restrained” generally means being prevented from
moving freely due to physical barriers or constraints. This could
involve being tied up, handcuffed, or otherwise restricted in move-
ment by physical means. It’s often used in contexts involving secu-
rity, medical care, or legal situations where someone’s movement
is limited to control or protect them.
* * *
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 43 of 56
23-10478 NEWSOM, J., Concurring 23
Claude
Query: What is the ordinary meaning of “physically restrained”?
Date of query: August 9, 2024
Response 1:
The ordinary meaning of “physically restrained” typically refers to
the use of physical force or barriers to limit or control a person’s
movements or actions. This can include:
1. Holding someone back with one’s hands or body
2. Using devices like handcuffs, straps, or restraints
3. Confining someone to a specific space, such as a room or cell
4. Preventing someone from moving freely by blocking their
path
The term implies a direct, tangible interference with a person’s abil-
ity to move or act as they wish. It’s often used in contexts related to
law enforcement, healthcare, or situations where someone’s free-
dom of movement needs to be restricted for safety or security rea-
sons.
The exact interpretation may vary slightly depending on the con-
text, but generally, it involves a clear, physical limitation on some-
one’s freedom of movement. Would you like me to elaborate on
any specific aspect of this definition?
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 44 of 56
23-10478 NEWSOM, J., Concurring 24
Response 2:
The ordinary meaning of “physically restrained” typically refers to:
1. The use of physical force or mechanical devices to limit or
control a person’s bodily movements.
2. Preventing someone from moving freely by holding them,
tying them up, or confining them in a space.
3. Applying physical measures to restrict an individual’s ability
to act or move as they wish.
This term is often used in contexts such as:
• Law enforcement situations where officers may need to con-
trol a suspect
• Medical settings, where patients might need to be restrained
for their own safety
• Instances of assault or abduction
• Certain therapeutic or protective interventions
The concept generally implies a temporary and situational applica-
tion of physical control, rather than a permanent condition.
Would you like me to elaborate on any specific aspect of this defi-
nition or its applications?
Response 3:
The ordinary meaning of “physically restrained” typically refers to
the act of using physical force or methods to restrict or control a
person’s movements or actions. This can include:
1. Holding someone back with hands or arms
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 45 of 56
23-10478 NEWSOM, J., Concurring 25
2. Using restraints like handcuffs, straps, or rope to limit
movement
3. Confining someone to a specific space, such as a room or
cell
4. Using one’s body weight to pin someone down
5. Employing devices designed for restraint, such as strait
jackets or leg shackles
The term implies a direct, tangible limitation on a person’s physical
freedom, usually against their will or as a means of control. It’s
often used in contexts related to law enforcement, medical settings,
or situations where someone’s movements need to be restricted for
safety or security reasons.
Would you like me to elaborate on any specific aspect of this defi-
nition or its applications?
Response 4:
The ordinary meaning of “physically restrained” typically refers to
the use of physical force, barriers, or implements to restrict or con-
trol a person’s bodily movements or actions. This could include:
1. Holding someone back with hands or arms
2. Using restraint devices like handcuffs, straps, or rope
3. Confining someone to a limited space, such as a room or cell
4. Preventing someone from moving freely by blocking their
path
5. Using one’s body weight to immobilize another person
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 46 of 56
23-10478 NEWSOM, J., Concurring 26
The term generally implies a deliberate action to limit someone’s
physical freedom of movement, often against their will or as a
means of control. This could occur in various contexts, such as law
enforcement, medical settings, or during conflicts between individ-
uals.
It’s important to note that the specific interpretation may vary
slightly depending on the context in which the term is used. Would
you like me to elaborate on any particular aspect of this definition?
Response 5:
The ordinary meaning of “physically restrained” typically refers to
the act of using physical force or barriers to limit or control some-
one’s movements or actions. This can include:
1. Holding someone back with your hands or body
2. Using restraints like handcuffs, straps, or rope to restrict
movement
3. Confining someone to a specific space, such as a room or cell
4. Using physical objects to block or impede movement
The term implies a direct, tangible intervention to prevent a person
from moving freely or taking certain actions. It’s often used in con-
texts like law enforcement, healthcare (particularly in psychiatric
settings), or in describing forceful interpersonal interactions.
It’s important to note that the use of physical restraints can have
legal, ethical, and safety implications depending on the context and
manner in which they are applied.
Would you like me to elaborate on any specific aspect of this defi-
nition or its applications?
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 47 of 56
23-10478 NEWSOM, J., Concurring 27
Response 6:
The ordinary meaning of “physically restrained” typically refers to
the use of physical force or mechanical means to limit or control a
person’s movements or actions against their will. This can include:
1. Holding someone back with hands or arms
2. Using handcuffs, straps, or other restraining devices
3. Confining someone to a specific space, like a room or chair
4. Blocking someone’s path to prevent movement
The term implies a direct, tangible intervention to restrict a per-
son’s physical freedom of movement. It’s often used in contexts
like law enforcement, healthcare (particularly in psychiatric set-
tings), or in describing forceful interpersonal interactions.
It’s important to note that the specific interpretation can vary
slightly depending on the context in which it’s used. Would you like
me to elaborate on any particular aspect or application of this
term?
Response 7:
The ordinary meaning of “physically restrained” typically refers to:
1. The use of physical force or mechanical means to limit a per-
son’s movement or ability to act freely.
2. Holding someone back or preventing them from moving us-
ing bodily contact or restraining devices.
3. Restricting an individual’s physical mobility against their
will.
This can include actions such as:
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 48 of 56
23-10478 NEWSOM, J., Concurring 28
• Holding someone’s arms or body to prevent them from
moving
• Using handcuffs, straps, or other restraining devices
• Confining someone to a specific space, like a room or cell
• Forcibly keeping someone in a particular position
The term is often used in contexts related to law enforcement,
healthcare settings, or situations where someone’s freedom of
movement needs to be limited for safety or security reasons. How-
ever, it can also be used in broader contexts to describe any situa-
tion where someone’s physical movement is forcibly restricted.
Would you like me to elaborate on any specific aspect of this defi-
nition?
Response 8:
The ordinary meaning of “physically restrained” typically refers to
the use of physical force, barriers, or implements to restrict or con-
trol a person’s movement or actions against their will. This can in-
clude:
1. Using bodily force to hold someone in place
2. Applying handcuffs, straps, or other restraining devices
3. Confining someone to a restricted space, like a locked room
4. Using physical barriers to prevent movement
5. Forcibly limiting someone’s range of motion
The term implies a direct, tangible interference with a person’s abil-
ity to move freely, often in the context of law enforcement, medical
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 49 of 56
23-10478 NEWSOM, J., Concurring 29
situations, or instances where someone’s actions need to be con-
trolled for safety reasons.
It’s important to note that the specific interpretation may vary
slightly depending on the context in which the term is used. Would
you like me to elaborate on any particular aspect of this definition?
Response 9:
The ordinary meaning of “physically restrained” typically refers to
the act of using physical force or mechanisms to limit or control a
person’s movement or actions. This can include:
1. Holding someone back with one’s hands or body
2. Using devices like handcuffs, straps, or restraints
3. Confining someone to a specific space, such as a room or
chair
4. Preventing someone from moving freely by blocking their
path
The term implies a direct, tangible intervention to restrict a per-
son’s physical mobility or actions, usually against their will or as a
means of control. This could occur in various contexts, such as law
enforcement, medical settings, or during conflicts.
It’s important to note that the specific interpretation may vary
slightly depending on the context in which the term is used. Would
you like me to elaborate on any particular aspect of this definition?
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 50 of 56
23-10478 NEWSOM, J., Concurring 30
Response 10:
The ordinary meaning of “physically restrained” typically refers to
the act of using physical force or barriers to restrict or limit some-
one’s movement or actions. This can involve:
1. Holding someone back with your hands or body
2. Using restraints like handcuffs, straps, or ropes to immobi-
lize a person
3. Confining someone to a specific space, such as a room or cell
4. Using physical objects to block or impede movement
The term is often used in contexts like law enforcement, medical
settings, or situations where someone needs to be prevented from
harming themselves or others. It implies a direct, hands-on ap-
proach to limiting a person’s physical freedom of movement, as
opposed to other forms of restraint like verbal commands or legal
restrictions.
Would you like me to elaborate on any specific aspect of this defi-
nition or provide examples of its use in different contexts?
* * *
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 51 of 56
23-10478 NEWSOM, J., Concurring 31
Gemini
Query: What is the ordinary meaning of “physically restrained”?
Date of query: August 9, 2024
Response 1:
Physically restrained means that a person’s freedom of movement
is restricted by another person or object against their will. This can
be done in a variety of ways, such as:
• Holding: Someone physically preventing another person
from moving.
• Tying: Using ropes, chains, or other materials to bind a per-
son.
• Locking: Confining a person in a space with no means of
escape.
Essentially, it means a person is being physically controlled and pre-
vented from moving freely.
Would you like to know the meaning of “physically restrained” in
a specific context, such as legal or medical?
Response 2:
Physically restrained means that a person’s freedom of movement
is restricted by another person or object against their will. This can
be done through:
• Direct physical force: Holding, grabbing, or pushing some-
one to prevent them from moving.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 52 of 56
23-10478 NEWSOM, J., Concurring 32
• Physical barriers: Confining someone in a room or space
with locked doors or windows.
• Mechanical devices: Using handcuffs, shackles, or similar
restraints.
Essentially, it means someone is being held back or prevented from
moving freely.
Would you like to know more about the legal or specific context in
which this term is used?
Response 3:
Physically restrained means that a person’s freedom of movement
is restricted by another person or object against their will. This can
be done in a variety of ways, such as:
• Holding: Someone physically preventing another person
from moving.
• Tying: Using ropes, chains, or other materials to bind a per-
son.
• Locking: Confining a person in a space with limited or no
exit.
Essentially, it means someone is being held against their will and
unable to move freely.
Would you like to know more about the legal or specific context of
this term?
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 53 of 56
23-10478 NEWSOM, J., Concurring 33
Response 4:
Physically restrained means that a person’s freedom of movement
is restricted by another person or object against their will. This can
be done in a variety of ways, such as:
• Holding: Someone physically holding another person to
prevent them from moving.
• Tying: Using ropes, chains, or other materials to bind a per-
son.
• Confining: Locking someone in a room or space.
• Immobilizing: Using physical force or devices to prevent
movement.
Essentially, it involves limiting a person’s ability to move freely
through physical means.
Would you like to know the legal definition of physically re-
strained, or do you have a specific context in mind?
Response 5:
Physically restrained means that a person’s freedom of movement
is restricted by another person or object against their will. This can
be done through:
• Direct physical force: Holding someone down, pushing
them back, or preventing them from moving.
• Physical barriers: Confining someone in a room, tying
them up, or using handcuffs.
Essentially, it means someone is being forcibly prevented from
moving freely.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 54 of 56
23-10478 NEWSOM, J., Concurring 34
Example: A person who is handcuffed to a chair is physically re-
strained.
Would you like to know the legal definition of “physically re-
strained” in a specific context?
Response 6:
Physically restrained means that a person’s freedom of movement
is restricted by another person or object against their will. This can
be done through:
• Direct physical force: Holding someone down, pushing
them back, or preventing them from moving.
• Physical barriers: Confining someone in a locked room, us-
ing handcuffs, or tying someone up.
Essentially, it means someone is being physically prevented from
doing what they want to do.
Would you like to know the legal definition in a specific context?
Response 7:
Physically restrained means that a person’s freedom of movement
is restricted by another person or object against their will. This can
involve:
• Holding or grabbing: Using physical force to prevent some-
one from moving.
• Tying or binding: Using ropes, chains, or similar items to
restrict movement.
• Confining: Keeping someone in a limited space, such as a
room or cage.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 55 of 56
23-10478 NEWSOM, J., Concurring 35
Essentially, it means someone is being physically prevented from
doing what they want to do.
Would you like to know about the legal or medical implications of
physical restraint?
Response 8:
Physically restrained means that a person’s freedom of movement
is restricted by another person or object against their will. This can
be done through:
• Direct physical force: Holding someone down, pushing
them back, or preventing them from moving.
• Physical barriers: Confining someone in a locked room, us-
ing handcuffs or shackles, or tying someone up.
Essentially, it involves limiting someone’s ability to move freely and
independently.
Would you like to know the legal definition or the term’s use in a
specific context?
Response 9:
Physically restrained means that a person’s freedom of movement
is restricted by another person or object against their will. This can
be done through:
• Direct physical force: Holding, grabbing, or pinning some-
one down.
• Physical barriers: Using handcuffs, zip ties, or other devices
to restrict movement.
USCA11 Case: 23-10478 Document: 53-1 Date Filed: 09/05/2024 Page: 56 of 56
23-10478 NEWSOM, J., Concurring 36
• Environmental restrictions: Confining someone in a
locked room or space.
Essentially, it means someone is being forcibly prevented from
moving freely.
Would you like to know more about the legal or specific context of
the term?
Response 10:
Physically restrained means that a person’s freedom of movement
is restricted by another person or object against their will. This can
involve:
• Holding: Someone physically holding another person to
prevent movement.
• Tying: Using ropes, chains, or other materials to bind a per-
son.
• Locking: Confining a person in a space with no means of
escape.
Essentially, it means a person is prevented from moving freely by
external forces.
* * *