Case: 19-60133 Document: 00515457870 Page: 1 Date Filed: 06/18/2020
IN THE UNITED STATES COURT OF APPEALS
FOR THE FIFTH CIRCUIT
United States Court of Appeals
Fifth Circuit
No. 19-60133
FILED
June 18, 2020
Lyle W. Cayce
JOSEPH THOMAS; VERNON AYERS; MELVIN LAWSON, Clerk
Plaintiffs - Appellees
v.
TATE REEVES, Governor of the State of Mississippi, all in the official
capacities of their own offices and in their official capacities as members of
the State Board of Election Commissioners; MICHAEL WATSON, Secretary
of State of the State of Mississippi, all in the official capacities of their own
offices and in their official capacities as members of the State Board of
Election Commissioners,
Defendants - Appellants
Appeal from the United States District Court
for the Southern District of Mississippi
Before OWEN, Chief Judge, and DAVIS, JONES, SMITH, STEWART,
DENNIS, ELROD, HAYNES, HIGGINSON, COSTA, WILLETT, HO,
DUNCAN, ENGELHARDT, and OLDHAM, Circuit Judges. *
PER CURIAM:
The en banc court unanimously agrees that this court no longer has
jurisdiction in this case because it has become moot. It is undisputed that the
*Judge Leslie H. Southwick and Judge James E. Graves are recused and did not
participate in the decision.
Case: 19-60133 Document: 00515457870 Page: 2 Date Filed: 06/18/2020
No. 19-60133
2019 general election has occurred, and as the State itself emphasized, the
current district lines will neither be used nor operate as a base for any future
election.
Therefore, the judgment of the district court is vacated, the appeal is
dismissed, and the case is remanded to the district court with instructions to
dismiss plaintiffs’ complaint for lack of jurisdiction. See U.S. v. Munsingwear,
Inc., 340 U.S. 36 (1950).
2
Case: 19-60133 Document: 00515457870 Page: 3 Date Filed: 06/18/2020
No. 19-60133
GREGG COSTA, Circuit Judge, joined by OWEN, Chief Judge, and DAVIS,
STEWART, DENNIS, and HIGGINSON, Circuit Judges, concurring:
The three-judge district court statute traces back more than a century.
In its long history, no court has applied the statute unless confronted with a
challenge to a law’s constitutionality. Mississippi asks our court to be the first.
What is the argument for disrupting the venerable understanding that
the extraordinary act of convening a three-judge trial court is limited to
constitutional cases? The statute allegedly contains an extra “the.” Despite
having gone undiscovered for decades in the high-stakes world of voting rights
litigation, the unnecessary “the” is supposedly such a glaring problem that it
requires us to read a law that contracted the reach of three-judge district courts
as one that for the first time extended the use of such courts to statutory cases.
An arguably redundant “the” cannot bear that weight. Indeed, when
considering Mississippi’s argument one cannot help but recall the wisdom of
Justice Scalia’s vivid point that “[Congress] does not . . . hide elephants in
mouseholes.” Whitman v. Am. Trucking Ass’ns, 531 U.S. 457, 468 (2001).
So although I join the per curiam opinion holding that this appeal is now
moot, I write to explain why a plain reading of the three-judge statute as well
as its ancestry reject the unprecedented notion that statutory challenges to
state legislative districts require a special district court.
I.
A.
As always, the starting place is the text. The general three-judge statute
states:
A district court of three judges shall be convened . . . when an
action is filed challenging the constitutionality of the
apportionment of congressional districts or the apportionment of
any statewide legislative body.
3
Case: 19-60133 Document: 00515457870 Page: 4 Date Filed: 06/18/2020
No. 19-60133
28 U.S.C. § 2284(a). It doesn’t take 30 pages to figure out what the statute
says. A person on the street would read it as requiring a three-judge court only
for constitutional challenges.
Courts have uniformly given the law that everyday meaning. See, e.g.,
Rural W. Tenn. African-Am. Affairs Council v. Sundquist, 209 F.3d 835, 838
(6th Cir. 2000) (noting reassignment of case to single judge after dismissal of
constitutional and Section 5 claims); Chestnut v. Merrill, 356 F. Supp. 3d 1351,
1357 (N.D. Ala. 2019) (rejecting argument that a single judge could not hear
Section 2 challenge); Bone Shirt v. Hazeltine, 336 F. Supp. 2d 976, 980 (D.S.D.
2004) (same as Rural West); Old Person v. Brown, 182 F. Supp. 2d 1002, 1003
(D. Mont. 2002) (single judge hearing Section 2 challenge). Their reading is
consistent with a judge’s duty to interpret the statutory language that
Congress approved and the President signed by “giving the words used their
ordinary meaning.” Levin v. United States, 568 U.S. 503, 513 (2013).
Against this backdrop, Mississippi offers the avant-garde view that the
law also requires three-judge courts for statutory-only challenges to state
legislative districts. 1 The novelty of the state’s reading does not merely suggest
that the question has “gone unasked,” Willett Op. 2; it shows that the ordinary
meaning of the statute is so clear that nobody ever bothered to ask the
question. 2
1 Before this case, Chestnut was the first to entertain an argument similar to the one
Mississippi makes. That was just last year. The disputed statutory language has been
around more than forty years.
2 Of course, when a plaintiff brings both constitutional and statutory challenges, the
constitutional hook for three-judge courts sweeps in the statutory claim. See Page v. Bartels,
248 F.3d 175, 191 (3d Cir. 2001). That makes sense given that section 2284(a) refers to
“action[s] . . . filed,” not individual claims. 28 U.S.C. § 2284(a). Notably, Page repeatedly
read section 2284(a)’s “constitutionality of” language to modify “the apportionment of any
statewide legislative body.” See 248 F.3d at 181, 185, 186, 188, 189 (quoting section 2284(a)
as requiring a three-judge district court for challenges to “the constitutionality of . . . the
4
Case: 19-60133 Document: 00515457870 Page: 5 Date Filed: 06/18/2020
No. 19-60133
And so it is. In common usage, a modifier like “constitutionality of”
usually applies to each term in a series of parallel terms. This principle is the
series-qualifier canon of construction. See ANTONIN SCALIA & BRYAN A.
GARNER, READING LAW: THE INTERPRETATION OF LEGAL TEXTS 147 (2012); cf.
Porto Rico Ry., Light & Power Co. v. Mor, 253 U.S. 345, 348 (1920) (“When
several words are followed by a clause which is applicable as much to the first
and other words as to the last, the natural construction of the language
demands that the clause be read as applicable to all.”). The canon is more of
“a matter of common English” than a hard-and-fast rule. See SCALIA &
GARNER, supra, at 147. It describes how people typically speak and write. See
Lockhart v. United States, 136 S. Ct. 958, 970 (2016) (Kagan, J., dissenting).
Consider a recent article in Mississippi’s leading newspaper. The article uses
a series modifier twice in three sentences when discussing how Mississippi
colleges are preparing for football season in the wake of the COVID-19
pandemic. It first refers to “testing all symptomatic athletes and staffers” and
then “educat[ing] returning students and employees on new protocols.” 3 Any
reader would understand that the modifiers—symptomatic and returning—
apply to both of the nouns that follow them.
The series-qualifier principle is just a fancy label for describing how a
normal person would understand section 2284(a). That is, the modifier
“constitutionality of” should apply to both of the parallel terms that follow it:
apportionment of any statewide legislative body” (omission in original)). The Third Circuit
did not confront the textual argument Mississippi makes today, but its instinctive reading
further demonstrates the statute’s plain meaning.
3 Nick Suss & Tyler Horka, What Needs to Happen for Ole Miss, Mississippi State to
Have Fans in Football Stadiums, MISS. CLARION-LEDGER (May 28, 2020, 2:38 PM),
https://www.clarionledger.com/story/sports/college/ole-miss/2020/05/28/ole-miss-msu-give-
thoughts-fan-attendance-football-season/5177694002/.
5
Case: 19-60133 Document: 00515457870 Page: 6 Date Filed: 06/18/2020
No. 19-60133
(1) challenges to “the apportionment of congressional districts” and
(2) challenges to “the apportionment of any statewide legislative body.” The
canon’s intuitive nature explains why the Supreme Court, other courts, and
leading treatises have taken that reading as a given. See, e.g., Harris v. Ariz.
Indep. Redistricting Comm’n, 136 S. Ct. 1301, 1306 (2016) (parenthetically
describing section 2284(a) as “providing for the convention of [a three-judge
district] court whenever an action is filed challenging the constitutionality of
apportionment of legislative districts”) 4; Armour v. Ohio, 925 F.2d 987, 989
(6th Cir. 1991) (en banc) (describing the test for section 2284(a) as whether
“there exists a non-frivolous constitutional challenge to the apportionment of
a statewide legislative body”); 22 JAMES W. MOORE, ET AL., MOORE’S FEDERAL
PRACTICE § 404.03[2], at 404-30 to -31 (3d ed. 2019) (explaining that section
2284(a) “is limited to federal constitutional claims”); 17A CHARLES ALAN
WRIGHT ET AL., FEDERAL PRACTICE AND PROCEDURE § 4235, at 202 (3d ed.
2007) (stating that section 2284(a) should apply to “all federal constitutional
challenges that could result in a reapportionment”).
B.
Mississippi says this straightforward reading of the three-judge statute
is wrong because of a redundant definite article it took more than forty years
for anyone to notice. Its argument rests on the statute’s use of the word “the”
before “apportionment of any state legislative body.” That determiner, the
state insists, indicates a break in the series modifier, so “constitutionality of”
modifies only what immediately follows it: challenges to “the apportionment of
congressional districts.” Because “constitutionality of” does not also modify the
4Judge Willett dismisses the Supreme Court’s language as dicta. Willett Op. 22 n.82.
But the importance of the Court’s aside is not its legal force. It instead shows what the
natural reading of the statute is when a lawyer is not parsing it to score a win for her client.
6
Case: 19-60133 Document: 00515457870 Page: 7 Date Filed: 06/18/2020
No. 19-60133
“apportionment of any statewide legislative body,” the argument goes, then
any challenge to state legislative districts—including Plaintiffs’ statutory
one—requires a three-judge court. Meanwhile, only constitutional challenges
to congressional districts would trigger the statute; an ordinary district court
would hear statutory attacks on the same target. 5
That reading creates a more convoluted statutory scheme than the clear-
cut distinction between constitutional and statutory claims that lawyers and
judges have long understood section 2284(a) to draw. I’ll return to that point.
For now, I focus on why Mississippi’s argument fails as a textual matter.
To be sure, “[t]he typical way in which syntax would suggest no carryover
modification” in a series is to repeat a determiner like “the” before one of the
series’ terms. SCALIA & GARNER, supra, at 148. Reading Law lists as one
example: “The charitable institutions or the societies (the presence of the second
the suggests that the societies need not be charitable).” Id. But again, the
series-qualifier canon is not a brightline rule to be applied mechanically.
“Perhaps more than most of the other canons,” it “is highly sensitive to
context.” Id. at 150. And here, the most natural reading of the statute is the
5 Mississippi did not make this argument until the very end of the district court
litigation. That would mean it forfeited the argument unless the three-judge statute is
jurisdictional. Nonbinding caselaw suggests section 2284(a) is jurisdictional. See, e.g.,
LULAC of Tex. v. Texas, 318 F. App’x 261, 264 (5th Cir. 2009) (per curiam). Yet a leading
authority observes that its jurisdictional status is uncertain. WRIGHT ET AL., supra, § 4235,
at 206–07. The statute’s “[a] district court of three judges shall be convened” language sounds
mandatory. 28 U.S.C. § 2284(a). But the Supreme Court has emphasized in recent years
that not all mandatory procedural rules are jurisdictional, and the distinction often turns on
whether Congress describes the requirement as jurisdictional. See, e.g., Fort Bend Cty. v.
Davis, 139 S. Ct. 1843, 1849–50 (2019). The word “jurisdiction” does not appear in section
2284(a). And the statute says that the procedure for convening a three-judge court begins
“[u]pon the filing of a request for three judges.” 28 U.S.C. § 2284(b)(1). If section 2284(a)
imposes a jurisdictional requirement, a district judge would be obligated to convene a three-
judge court regardless of whether a party asked for one. So I have serious doubts that the
statute is jurisdictional, but I nonetheless address the merits of the argument.
7
Case: 19-60133 Document: 00515457870 Page: 8 Date Filed: 06/18/2020
No. 19-60133
long-accepted interpretation that “constitutionality of” modifies both kinds of
apportionment challenges.
Most importantly, the use of “the” before each parallel term would not
cut off the modifier “constitutionality of” in everyday English. At the risk of
overdoing analogies to the sports pages, consider the hypothetical newspaper
line from the motions panel opinion: “The NCAA is investigating the recruiting
practices of the football program and the basketball program.” Thomas v.
Bryant, 919 F.3d 298, 306 (5th Cir. 2019). As with the three-judge statute, the
last “the” may not be necessary. But no reader would understand that last
“the” to mean that the investigation into the football program is limited to
recruiting violations while the investigation into the basketball program might
also look into point-shaving or ticket-scalping violations. See SCALIA &
GARNER, supra, at 176–77 (explaining that drafters sometimes repeat
themselves and add words that serve no legal function).
Judge Willett acknowledges that the commonsense reading of this
sentence is that “the recruiting practices of” applies to both sports. Willett Op.
7. But he suggests the way we read “informal, everyday phrasing” in a
newspaper is different than how we should read “formal, statutory phrasing”
in the law. Id. at 9. Justice Story thought otherwise:
Constitutions are not designed for metaphysical or logical
subtleties, for niceties of expression, for critical propriety, for
elaborate shades of meaning, or for the exercise of philosophical
acuteness or judicial research. They are instruments of a practical
nature, founded on the common business of human life, adapted to
common wants, designed for common use, and fitted for common
understandings. The people make them; the people adopt them;
the people must be supposed to read them, with the help of
common sense; and cannot be presumed to admit in them any
recondite meaning, or any extraordinary gloss.
8
Case: 19-60133 Document: 00515457870 Page: 9 Date Filed: 06/18/2020
No. 19-60133
JOSEPH STORY, COMMENTARIES ON THE CONSTITUTION OF THE UNITED STATES
157–58 (1833). So did Justice Scalia. Reading Law explains that words in all
legal instruments “are to be understood in their ordinary, everyday meanings.”
SCALIA & GARNER, supra, at 69. This command to follow ordinary meaning is
not just one among the many rules of statutory interpretation. It is “the most
fundamental semantic rule of interpretation.” Id. Indeed, the notion that
there are special, lawyers-only grammar rules for reading statutes is at odds
with the principle that, in a democracy, laws should be easily understood by
the people they govern. See generally Note, Textualism as Fair Notice, 123
HARV. L. REV. 542 (2009); cf. Sessions v. Dimaya, 138 S. Ct. 1204, 1225–26
(2018) (Gorsuch, J., concurring in part and concurring in the judgment)
(“Perhaps the most basic of due process’s customary protections is the demand
of fair notice,” id. at 1225).
Yet Judge Willett says that the statute’s “particular grammatical
construction” sets it apart from the newspaper line, so it requires more
legalistic scrutiny. Willett Op. 10. But is it so unusual that ordinary people
would not understand it? Take a look at the earlier example from Reading
Law: “The charitable institutions or the societies.” SCALIA & GARNER, supra,
at 148. That sentence’s natural reading is that “charitable” applies only to
“institutions.” The reason is that the modifier “charitable” appears between
the first “the” and that determiner’s noun, “institutions.” Because there is no
corresponding modifier between “societies” and its determiner, it is clear that
no modifier applies to “societies.” But the example’s structure is different than
that of the three-judge statute. Imagine it instead mirrored the statute’s
“constitutionality of” structure: “The charity of the institutions or the
societies.” An unusual construction, but not an ambiguous one. A reader
would assume that “charity” modifies both “institutions” and “societies.” The
9
Case: 19-60133 Document: 00515457870 Page: 10 Date Filed: 06/18/2020
No. 19-60133
difference is that, in this variant, the modifier precedes both parallel terms and
their determiners. 6 So too with the three-judge statute. 7
C.
Mississippi makes a second unconvincing textual argument. It asserts
that another canon subverts the natural reading of the statute that the series-
qualifier canon confirms: the canon against surplusage. That canon advises
that “a statute should be construed so that effect is given to all its provisions,
so that no part will be inoperative or superfluous, void or insignificant.” Corley
v. United States, 556 U.S. 303, 314 (2009) (cleaned up) (quoting Hibbs v. Winn,
542 U.S. 88, 101 (2004)). Mississippi points out that, if the three-judge statute
applied to constitutional challenges against both apportionment of
congressional districts and statewide legislative bodies, its second use of the
phrase “the apportionment of” would be unnecessary. The statute could have
the same meaning if it read as follows:
A district court of three judges shall be convened . . . when an
action is filed challenging the constitutionality of the
6 Some of Reading Law’s other examples have similar analogues. Take the sentence:
“A solid wall or a fence.” SCALIA & GARNER, supra, at 149. The determiner “a” cuts off the
modifier “solid,” so “the fence need not be solid.” Id. But change it to follow the structure of
section 2284(a): “The solidity of a wall or a fence.” Another unusual sentence structure. Yet
no lay reader would think “solidity” modified only “wall.”
7 Some markups illustrate the point. They emphasize the modifier, number the
parallel terms, and underline the parallel terms’ determiners.
• Original Reading Law Example—
[1] The charitable institutions or [2] the societies.
• Reading Law Variant—
The charity of [1] the institutions or [2] the societies.
• Section 2284(a)—
[T]he constitutionality of [1] the apportionment of congressional
districts or [2] the apportionment of any statewide legislative body.
10
Case: 19-60133 Document: 00515457870 Page: 11 Date Filed: 06/18/2020
No. 19-60133
apportionment of congressional districts or the apportionment of
any statewide legislative body.
That is true. But once again Mississippi treats a canon meant to describe
how people typically speak and write like an uncompromising rule. As our full
court recently emphasized, “the canon against surplusage yields to context as
it expresses courts’ general reluctance to treat statutory terms as surplusage.”
Latiolais v. Huntington Ingalls, Inc., 951 F.3d 286, 294 (5th Cir. 2020) (en
banc) (cleaned up) (quoting Bd. of Trs. of Leland Stanford Junior Univ. v.
Roche Molecular Sys., Inc., 563 U.S. 776, 788 (2011)). That general reluctance
“does not require us to avoid surplusage at all costs.” United States v. Atl.
Research Corp., 551 U.S. 128, 137 (2007).
We sometimes accept a little surplusage because we acknowledge that
the canon’s premise—that legislators do not include in statutes words that
have no effect—“is not invariably true.” SCALIA & GARNER, supra, at 176.
“Sometimes drafters do repeat themselves and do include words that add
nothing of substance . . . .” Id. at 176–77. Repetition and parallelism are
features of how ordinary people—and extraordinary ones—speak and write.
See, e.g., Abraham Lincoln, Gettysburg Address (Nov. 19, 1863)
(“[G]overnment of the people, by the people, for the people, shall not perish
from the earth.”). So when all other indicators support a plain reading of a
statute, we will not let minor repetition steer us toward a farfetched one. And
that is exactly what Mississippi asks us to do here: throw out a longstanding,
commonsense construction of the three-judge statute just to avoid making a
short phrase redundant.
Besides, Mississippi’s view of section 2284(a) suffers from surplusage
too. If “constitutionality of” modifies only “the apportionment of congressional
11
Case: 19-60133 Document: 00515457870 Page: 12 Date Filed: 06/18/2020
No. 19-60133
districts,” then the statute’s second “the” becomes unnecessary. We could
strike it out and the statute would carry the same meaning:
A district court of three judges shall be convened when . . . an
action is filed challenging the constitutionality of the
apportionment of congressional districts or the apportionment of
any statewide legislative body.
In fact, by using just one determiner before each type of suit, that version of
the statute would more clearly delineate between those suits and better
support Mississippi’s position. But Congress chose to include that second “the.”
So as it turns out, both parties’ readings leave the statute with an extra “the.”
Faced with two interpretations that each contain some surplusage, we should
give the statute its natural meaning.
Statutory interpretation is not a lawyer’s game to “divine arcane
nuances” and “discover hidden meanings.” See SCALIA & GARNER, supra, at 69;
see also id. at 177 (explaining that a “clever interpreter” can abuse the canon
against surplusages by “creat[ing] unforeseen meanings or legal effects” from
“stylistic” repetition). Our duty instead is to follow the natural, everyday
meaning of the words enacted into law. Id. at 33 (“The interpretive approach
we endorse is that of the ‘fair reading’: determining the application of a
governing text to given facts on the basis of how a reasonable reader, fully
competent in the language, would have understood the text at the time it was
issued.”). The plain reading of the three-judge statute is that it applies only to
suits alleging that a law is unconstitutional. Calling the contested “the” a
textual mousehole is being generous.
12
Case: 19-60133 Document: 00515457870 Page: 13 Date Filed: 06/18/2020
No. 19-60133
II.
Even if an extra definite article opens the door ever so slightly to some
ambiguity, section 2284(a)’s statutory history slams it back shut. 8 A brief
primer on the three-judge statute shows just how transformative Mississippi’s
interpretation would be.
Congress first enacted the three-judge statute in the aftermath of Ex
parte Young to require three judges to hear what it predicted would be an
increasing number of suits challenging state laws “upon the ground of the
unconstitutionality of such statute.” 28 U.S.C. § 2281 (1970); see generally
David P. Currie, The Three-Judge District Court in Constitutional Litigation,
32 U. CHI. L. REV. 1, 5–8 (1964). When courts later struck down many New
Deal reforms, one of the only aspects of President Roosevelt’s court-packing
plan to become law was a measure also requiring three-judge panels for suits
seeking to enjoin “any Act of Congress for repugnance to the Constitution of
the United States.” 28 U.S.C. § 2282 (1970); see 17A WRIGHT ET AL., supra,
§ 4234, at 194–95. Both laws’ focus on only constitutional challenges made
sense because striking down democratically enacted laws is “the gravest and
most delicate duty” courts are “called on to perform.” Nw. Austin Mun. Util.
Dist. No. One v. Holder, 557 U.S. 193, 204 (2009) (quoting Blodgett v. Holden,
275 U.S. 142, 147–48 (1927) (Holmes, J., concurring)). The idea behind
requiring three judges for this “class of cases of special importance” was to
“assure more weight and greater deliberation by not leaving the fate of such
litigation to a single judge.” Phillips v. United States, 312 U.S. 246, 249–50
(1941) (first quotation from Ex parte Collins, 277 U.S. 565, 567 (1928)); see also
8 Judge Willett agrees that statutory history—as opposed to more controversial
legislative history—can illuminate a statute’s meaning. See Willett Op. 13 n.45.
13
Case: 19-60133 Document: 00515457870 Page: 14 Date Filed: 06/18/2020
No. 19-60133
Swift & Co. v. Wickham, 382 U.S. 111, 118 (1965) (explaining that Congress
introduced the three-judge procedure because “if three judges declare that a
state statute is unconstitutional the people would rest easy under it” (quoting
45 CONG. REC. 7256 (1910) (statement of Sen. Overman))).
But in the mid-1970s, Congress scrapped most of the three-judge district
court regime because it was burdening the Supreme Court as well as lower
courts and had resulted in procedural complexities. See 17A WRIGHT ET AL.,
supra, § 4234, at 195–98; see also Kalson v. Paterson, 542 F.3d 281, 287 (2d Cir.
2008) (noting that the 1976 Act “vastly reduced the category of cases for which
a three-judge court is mandated”). It nonetheless retained the procedure for a
small set of important cases: constitutional challenges to redistricting for
congressional and state legislative seats, then-recent phenomena in the
aftermath of the revolutionary one person, one vote line of cases. See, e.g.,
Reynolds v. Sims, 377 U.S. 533 (1964).
Congress was thus narrowing the reach of the three-judge statute when
it added the current language. It is implausible (to put it mildly) that while
otherwise contracting the statute, Congress decided to expand it beyond
constitutional challenges for the first time. Indeed, neither Mississippi nor
Judge Willett explain why a law shrinking the three-judge statute should be
read to enlarge it. That unplugged hole in their argument is especially
damaging considering that Congress’s “one reason” for creating three-judge
courts was “to save state and federal statutes from improvident doom, on
constitutional grounds, at the hands of a single federal district judge.”
Gonzalez v. Automatic Emp. Credit Union, 419 U.S. 90, 97 (1974) (emphasis
added). It is particularly hard to believe that Congress would have made such
a significant and discordant change by merely adding an extra “the.”
14
Case: 19-60133 Document: 00515457870 Page: 15 Date Filed: 06/18/2020
No. 19-60133
Reading the statute in the way Mississippi urges does not make sense
for another reason: Why would Congress require three judges to hear statutory
claims challenging state legislative redistricting but not congressional
redistricting? Federalism concerns cannot explain the difference. While a
state of course has a strong interest in how it apportions its legislature, it also
has a strong interest in choosing how to divvy up its citizens into congressional
districts. In fact, contemporary critics of the initial judicial foray into review
of legislative apportionment viewed courts’ redrawing of congressional districts
as more intrusive on traditional state prerogatives than judicial redrawing of
state legislative districts. That is because of the view that Article I of the
Constitution grants “States . . . plenary power to select their allotted
Representatives in accordance with any method of popular election they
please, subject only to the supervisory power of Congress.” Wesberry v.
Sanders, 376 U.S. 1, 23 (1964) (Harlan, J., dissenting).
There is a further problem with concluding that the 1976 Congress
extended three-judge courts to cover statutory challenges: such claims were
not common then. It was not established in the 1970s that Section 2 of the
Voting Rights Act provided a private right of action. See City of Mobile v.
Bolden, 446 U.S. 55, 60 (1980) (assuming without deciding such a suit could
exist). A few years later, the Supreme Court held Section 2 did not prohibit
discriminatory effects, which meant it provided no guarantee beyond what the
Constitution already did. See id. And the typical pre-1982 Section 2 claim that
courts did consider was a challenge to at-large electoral systems, not the
drawing of single member districts. See, e.g., id. In other words, there was no
15
Case: 19-60133 Document: 00515457870 Page: 16 Date Filed: 06/18/2020
No. 19-60133
practice of statutory challenges to state legislative apportionment that
Congress needed to address in 1976. 9
Judge Willett and I agree on this history, but we differ on the takeaway.
See Willett Op. 13–17. I draw this lesson: Congress would not have used an
extra “the” to distinguish between constitutional and statutory apportionment
challenges when the latter kind of action was not even on its radar screen. The
idea that we should extend the three-judge statute to statutory challenges
because the 1976 Congress would have wanted that if only it had known what
the future held does not treat text as “the alpha and the omega” of statutory
interpretation. Willett Op. 1. Instead, the text fits neatly with what the
historical backdrop suggests. The Congress amending the three-judge statute
would have been focused only on constitutional challenges.
To sum up, Mississippi would give this much weight to the “the” that
comes before “reapportionment of any statewide legislative body”: Insertion of
that article would require three-judge panels for exclusively statutory claims—
followed by direct appeal to the Supreme Court, 28 U.S.C. § 1253—when the
three-judge regime Congress was paring down in 1976 never did. It would
require those three-judge panels only for statutory challenges to
apportionment of state legislative seats, not congressional ones. And it would
do all this to address statutory challenges to apportionment of state
legislatures when those claims hardly existed in 1976. An elephant indeed.
9 While Section 5 of the Voting Rights Act provided a statutory cause of action, it
already included its own three-judge district court requirement. 42 U.S.C. § 1973c (1976).
16
Case: 19-60133 Document: 00515457870 Page: 17 Date Filed: 06/18/2020
No. 19-60133
III.
We are now beyond belts and suspenders. The ordinary meaning of the
three-judge statute is the interpretation every court has given it. And the
statute’s history reinforces that it covers only constitutional challenges.
But even if the text of three-judge statute remains a “brain teaser,”
Willett Op. 10, open to more than one “plausible reading,” id. at 29, the
Supreme Court has told us how to solve the puzzle. 10 The tiebreaker is a
“canon of narrow construction” for three-judge court statutes. Gonzalez, 419
U.S. at 98 (“[W]e have stressed that the three-judge-court procedure is not ‘a
measure of broad social policy to be construed with great liberality.’” (quoting
Phillips, 312 U.S. at 251)). It reflects concerns over the “serious drain” that
three-judge district courts have on the judiciary’s resources as well as the
impact of an automatic appeal on the Supreme Court docket. Phillips, 312 U.S.
at 250. One treatise observed that those concerns were so pronounced in the
1960s and 1970s that Justice Frankfurter’s articulation of the strict
construction principle in Phillips became “the Court’s favorite quotation.” 17A
WRIGHT ET AL., supra, § 4234, at 197 (collecting cases).
This rule of construction no longer appears with such frequency. But
reduced citations do not allow us to jettison Supreme Court precedent. See
Ramos v. Louisiana, 140 S. Ct. 1390, 1416 n.5 (2020) (Kavanaugh, J.,
concurring) (“[F]ederal courts have a constitutional obligation to follow a
precedent of this Court unless and until it is overruled by this Court.”); cf.
Rodriguez de Quijas v. Shearson/Am. Exp., Inc., 490 U.S. 477, 484 (1989) (“If
a precedent of this Court has direct application in a case, yet appears to rest
10See also Thomas v. Bryant, 938 F.3d 134, 176 (5th Cir. 2019) (Willett, J., dissenting)
(describing Mississippi’s argument as “more confounding than convincing”); id. at 190 (“As
for the knotty jurisdictional question, I concede uncertainty.”).
17
Case: 19-60133 Document: 00515457870 Page: 18 Date Filed: 06/18/2020
No. 19-60133
on reasons rejected in some other line of decisions, the Court of Appeals should
follow the case which directly controls, leaving to this Court the prerogative of
overruling its own decisions.”). But see Willett Op. 21–22 (questioning the
continued application of the canon because the concerns that motivated it are
no longer as salient). Because it is still on the books, the rule of narrow
construction resolves this case even if Mississippi’s argument has cast any
doubt on the widely accepted meaning of the law.
* * *
The plain meaning of the statute’s text, uniform caselaw applying the
statute, the statutory history, and the rule that three-judge statutes should be
construed narrowly all favor the district court’s view that three judges are not
required for a suit raising only statutory challenges to state legislative
districts. To come to Mississippi’s contrary and unprecedented conclusion
would require us to wrench an elephant out of the tiniest of mouseholes.
18
Case: 19-60133 Document: 00515457870 Page: 19 Date Filed: 06/18/2020
No. 19-60133
DON R. WILLETT, Circuit Judge, joined by SMITH, ELROD, DUNCAN, and
ENGELHARDT, Circuit Judges, concurring in the judgment:
The en banc court is unanimous on the “what”—vacate the district
court’s judgment. But we have assorted views on the “why.”
Vacatur is the correct result, but for more than one reason. Putting aside
mootness (lack of a live controversy), there exists a separate problem (lack of
jurisdiction). The most forthright, text-centric reading of 28 U.S.C. § 2284(a) is
that a three-judge district court is required to decide apportionment
challenges—both statutory and constitutional—to statewide legislative bodies.
The wording of § 2284(a) may be imprecise. But its meaning—when read in
the light of blackletter syntactic and contextual canons—manifestly favors the
State.
I
“Text is the alpha and the omega of the interpretive process.” 1 And as
this is fundamentally a statutory-construction case, we begin (and end) with
the actual language that Congress chose. 2 Section 2284(a) states:
A district court of three judges shall be convened . . . when an
action is filed challenging the constitutionality of the
apportionment of congressional districts or the apportionment of
any statewide legislative body.
1 United States v. Maturino, 887 F.3d 716, 723 (5th Cir. 2018).
See Weatherly v. Pershing, L.L.C., 945 F.3d 915, 921 (5th Cir. 2019) (“[W]hen a
2
statute controls, our first stop (and usually our last) is the statutory text.”); United States v.
Lauderdale County, 914 F.3d 960, 964 (5th Cir. 2019) (“The task of statutory interpretation
begins and, if possible, ends with the language of the statute.”).
19
Case: 19-60133 Document: 00515457870 Page: 20 Date Filed: 06/18/2020
No. 19-60133
A divided merits panel labeled these words “clear.” 3 A divided motions panel
said they had a “natural reading.” 4 If only.
Section 2284(a) is not a paradigm of precision. It is inartfully worded,
and divining its meaning requires painstaking study. Our circuit is the first to
confront the issue, not because the answer is obvious, but because the question
has gone unasked—until now. Now raised, the question has managed to divide
16 federal circuit judges with hundreds of years of combined judicial
experience, sitting first as a motions panel, then as a merits panel, and now en
banc. The issue may be elemental—whether “constitutionality” modifies the
second phrase in the sentence—but it is anything but elementary. It’s a vexing
issue open to good-faith debate.
Here are the two competing interpretations:
Option A (Thomas)—a three-judge court is required to decide:
(1) the constitutionality of the apportionment of congressional
districts; or (2) the constitutionality of the apportionment of any
statewide legislative body.
Upshot: Only constitutional challenges to state and federal legislative maps
require three judges; purely statutory disputes can be heard by a single judge.
Option B (the State)—a three-judge court is required to decide:
(1) the constitutionality of the apportionment of congressional
districts; or (2) the apportionment of any statewide legislative
body.
3Thomas v. Bryant, 938 F.3d 134, 145 (5th Cir. 2019) (“The text of § 2284(a) is clear,
so we apply the statute as written.”), reh’g granted en banc, 939 F.3d 629 (5th Cir. 2019).
4Thomas v. Bryant, 919 F.3d 298, 306 (5th Cir. 2019) (determining that “context
supports the natural reading that courts have long given [§ 2284(a)]”).
20
Case: 19-60133 Document: 00515457870 Page: 21 Date Filed: 06/18/2020
No. 19-60133
Upshot: All challenges, both constitutional and statutory, to state maps require
three judges—but only constitutional challenges to federal maps require three
judges.
The core interpretive question is simply stated: Does “constitutionality
of” in § 2284(a) carry over to modify “the apportionment of any statewide
legislative body”? Thomas says yes, meaning a single judge could decide his
purely statutory challenge. The State says no, meaning three judges were
required. Here, the State requested a three-judge court, and the district court
refused, concluding that it could decide a standalone section 2 claim.
Language is often slippery, particularly legislative language. Chief
Justice Marshall famously made the point 201 years ago: “Such is the character
of human language, that no word conveys to the mind, in all situations, one
single definite idea . . . .” 5 Section 2284(a)’s phrasing is inexact. But its meaning
backs the State: “constitutionality of” does not carry over. Admittedly, this is
not the only possible reading, but it is the most textually plausible one.
A
Litigants in our adversarial system are loath to concede imprecision,
insisting forcefully, if not always convincingly, that the statutory text “plainly”
or “clearly” cuts their way. Here, each side insists that § 2284(a) has an
inescapably unambiguous reading—the one that favors their side.
The lion’s share of twenty-first century appellate judging is reading
legislative language and deciding what it means. In today’s statute-laden era,
how we decide—legisprudence: the jurisprudence of legislation 6—is as weighty
5 McCulloch v. Maryland, 17 U.S. (4 Wheat.) 316, 414 (1819).
6 Legisprudence, BLACK’S LAW DICTIONARY 1040 (10th ed. 2014) (“The systematic
analysis of statutes within the framework of jurisprudential philosophies about the role and
nature of law.”).
21
Case: 19-60133 Document: 00515457870 Page: 22 Date Filed: 06/18/2020
No. 19-60133
as what we decide. Methodology matters. So we must be mindful of our duty to
behave judicially by adjudicating and not politically by legislating. “Our
Constitution’s ingenious architecture demands that judges be sticklers when
decoding legislative text.” 7 Sticklers about not revising statutes under the
guise of interpreting them. Sticklers about not extolling judicial guesswork
over legislative handiwork. Sticklers about respecting a constitutional design
that, when conferring and confining power, leaves lawmaking to lawmakers.
“The law begins with language, and the foremost task of legal interpretation is
divining what the law is, not what the judge-interpreter wishes it to be.” 8
Principled, no-favorites textualism means tying oneself to the mast.
All to say, textual interpretation demands unswerving fidelity to text.
“Judges are minders,” after all, “not makers or menders.” 9 True, statutory
language is now and again imprecise—sometimes inadvertently, sometimes
intentionally. And while judges don’t have, and rarely need, secret decoder
rings to decrypt legislative text, we routinely use various tools to glean
meaning. “Statutory language, like all language, is suffused with age-old
interpretive conventions. And judges, like all readers, must be attentive not to
words standing alone but to surrounding structure and other contextual cues
that illuminate meaning.” 10
Time-honored canons of interpretation can help, provided the canons
esteem textual interpretation. True, the canons are not inexorable commands,
7 Reed v. Taylor, 923 F.3d 411, 415 (5th Cir. 2019).
8 Id.
9 Id.
10Id.; see also United States v. Graves, 908 F.3d 137, 141 (5th Cir. 2018) (“[T]ext may
not be divorced from context.” (quoting Univ. of Tex. Sw. Med. Ctr. v. Nassar, 570 U.S. 338,
356 (2013))).
22
Case: 19-60133 Document: 00515457870 Page: 23 Date Filed: 06/18/2020
No. 19-60133
but neither are they window dressing. They are venerable interpretive
presumptions about what smartly produced language means. The canons exist
to clarify meaning, not to cloak it. And no canon, however esteemed, can defeat
the obvious, non-absurd meaning of clearly drafted text. In this case, however,
the text is anything but clearly drafted. And as we try to solve the statutory
puzzle, some familiar canons, both syntactic and contextual, reveal § 2284(a)’s
semantic import. Notably, the canons point in the same interpretive direction,
underlining rather than undermining. This is not a Battle of the Canons case
pitting competing doctrines against each other. Here, we are not picking
between discordant canons; we are stacking harmonious ones.
B
Batting leadoff today, the series-qualifier canon. 11
The series-qualifier canon is a syntactic canon that looks to the
grammatical arrangement of words in a sentence. Specifically, when does a
modifier apply to a parallel series, and when does it not? The general rule:
“When there is a straightforward, parallel construction that involves all nouns
or verbs in a series, a prepositive or postpositive modifier normally applies to
the entire series.” 12 Consider these case examples, gathered in Reading Law:
• Charitable institutions or societies (charitable modifies both
institutions and societies).
• Internal personnel rules and practices of an agency (internal
personnel modifies both rules and practices).
• Intoxicating bitters or beverages (intoxicating modifies both
bitters and beverages).
11 See generally ANTONIN SCALIA & BRYAN A. GARNER, READING LAW § 19 (2012).
12 Id. § 19, at 147.
23
Case: 19-60133 Document: 00515457870 Page: 24 Date Filed: 06/18/2020
No. 19-60133
• Forcibly assaults, resists, opposes, impedes, intimidates, or
interferes with (forcibly modifies each verb in the list).
• Willfully damage or tamper with (willfully modifies both
damage and tamper with). 13
But there’s an important caveat to this carryover rule, one we recently
applied. 14 No carryover modification is suggested when “a determiner (a, the,
some, etc.) [is] repeated before the second element.” 15 Justice Scalia and
lexicographer Garner provide some no-carryover illustrations:
• The charitable institutions or the societies (the presence of the
second the suggests that the societies need not be charitable).
• A solid wall or a fence (the fence need not be solid).
• Delaware corporations and some partnerships (the partnerships
may be registered in any state).
• To clap and to cheer lustily (the clapping need not be lusty). 16
Now let’s return to our statute, § 2284(a): “the constitutionality of the
apportionment of congressional districts or the apportionment of any statewide
legislative body.” 17 The determiner “the” (or, perhaps more precisely, the
determining phrase “the apportionment”) sets the last phrase apart, indicating
that § 2284(a) requires three judges for all apportionment challenges to state
maps, not just constitutional challenges.
The district court cited the series-qualifier canon but misapplied it,
inexplicably failing to even acknowledge § 2284(a)’s crucial determiner. The
13 Id. § 19, at 148.
14 United States ex rel. Vaughn v. United Biologics, L.L.C., 907 F.3d 187, 195 (5th Cir.
2018) (quoting SCALIA & GARNER, supra note 11 § 19, at 148) (“The typical way to break the
series is to insert a determiner.”).
15 SCALIA & GARNER, supra note 11 § 19, at 148.
16 Id. § 19, at 148–49.
17 28 U.S.C. § 2284(a) (emphasis added).
24
Case: 19-60133 Document: 00515457870 Page: 25 Date Filed: 06/18/2020
No. 19-60133
motions panel majority declined to apply the series-qualifier canon at all,
lamenting (accurately) that the canon is “highly sensitive to context.” 18 And
the motions panel determined that context here supports a “natural reading,”
claiming that its reading of the statute (that “constitutionality of” carries over)
is how “ordinary people speak and write.” 19 To bolster its position, the motions
panel offered this illustration:
[C]onsider what a reader would think after seeing the following in
the newspaper: ‘The NCAA is investigating the recruiting
practices of the football program and the basketball program.’ As
with the three-judge statute, the final ‘the’ may not be necessary.
But would it make the reader think the investigation into the
basketball program is not limited to recruiting violations . . . ? 20
We do not disagree. But if anything, the NCAA illustration (which tellingly
uses a different structure from the one in § 2284(a)) simply bears out a
commonsensical and utterly uncontroversial notion: Interpretation is a human
enterprise, and no canon, including the series-qualifier canon, can override the
self-apparent meaning of written words. Sometimes, as in the NCAA example,
plain text is plain enough.
18 Thomas, 919 F.3d at 306. The merits panel majority determined that the series-
qualifier canon, “even if applicable,” confirmed its reading of the statute. Thomas, 938 F.3d
at 146. It supported this position only by suggesting that the canon’s exception for
determiners shouldn’t be applied because the canon is (as everyone agrees), “highly sensitive
to context.” Id. But that’s no reason to skirt the exception. Scrupulous concern for a text’s full
meaning requires scrupulous application of the full canon, not just the snippet that reinforces
one’s preferred interpretation.
19 Thomas, 919 F.3d at 306. (citing Lockhart v. United States, 136 S. Ct. 958, 970
(2016) (Kagan, J., dissenting). The majority cited to Justice Kagan’s dissent in Lockhart, but
ironically, Justice Kagan was urging the application of the series-qualifier canon, not its
rejection—precisely because the canon “reflects the completely ordinary way that people
speak and listen, write and read.” Lockhart, 136 S. Ct. at 970 (Kagan, J., dissenting). That’s
no less true when the statute contains a determiner.
20 Thomas, 919 F.3d at 306.
25
Case: 19-60133 Document: 00515457870 Page: 26 Date Filed: 06/18/2020
No. 19-60133
Used properly, the canons are neutral guides to reveal, not conceal,
meaning. “Often the sense of the matter prevails,” 21 and no sensible user of
English believes the canons should contort obvious meaning. As the NCAA
example is utterly clear on its face, no canons are needed to divine its meaning.
The gist of the language is instantly clear to the relevant linguistic community.
But § 2284(a) is hardly intuitive. It simply doesn’t have an obvious, self-evident
meaning and is capable of two distinct readings. 22 So we look to syntactic and
contextual canons, the shared background conventions that provide linguistic
context, to help choose among competing interpretations. 23
21 SCALIA & GARNER, supra note 11 § 19, at 150.
But notably, only one reading (Thomas’s) requires the deletion or insertion of words.
22
Thomas would either subtract words so that the statute reads this way:
“A district court of three judges shall be convened . . . when an action is filed
challenging the constitutionality of the apportionment of congressional
districts or the apportionment of any statewide legislative body.”
Or he would add words so that the statute reads this way:
“A district court of three judges shall be convened . . . when an action is filed
challenging the constitutionality of the apportionment of congressional
districts or the [constitutionality of the] apportionment of any statewide
legislative body.”
The former alteration treats part of the statute as surplusage. See infra I(C). And the
latter alteration collides with the Supreme Court’s admonition, recently (and
unanimously) reaffirmed, that judges “may not narrow a provision’s reach by inserting
words Congress chose to omit.” Lomax v. Ortiz-Marquez, No. 18-839, 2020 WL
3038282, at *3 (U.S. June 8, 2020).
23 Judge Costa’s opinion accuses us of employing “special, lawyers-only grammar rules
for reading statutes.” Costa Op. at 7. But the linguistic canons are not “special, lawyers-only
grammar rules.” The canons are traditional tools of interpretation routinely applied by both
the Supreme Court and by this court. Interestingly, Judge Costa’s opinion uses the same
canons we do. Moreover, he attacks a strawman. He scolds us for resorting to canons rather
than following the “natural” or “everyday” meaning of the text. But there is no disagreement
over what to do when faced with text whose obvious sense is immediately clear. Our
disagreement is simply over whether § 2284(a) is such a text. Readers can compare the
language of § 2284(a) with Judge Costa’s NCAA example and decide for themselves if
§ 2284(a) likewise carries such a “commonsense,” ordinary, and self-apparent meaning.
26
Case: 19-60133 Document: 00515457870 Page: 27 Date Filed: 06/18/2020
No. 19-60133
This approach accords with the principle that words “are to be
understood in their ordinary, everyday meanings.” 24 First, we start with the
plain meaning of the text, and if it’s obviously a spade, we call it a spade. But
sometimes obviousness doesn’t work. When it’s unclear whether it’s a spade or
a mattock, we consult the canons. If the canons also prove unavailing, and
we’ve made every effort to discern the meaning, then the statute is
ambiguous. Only then, after plain meaning and application of the interpretive
canons are found lacking, do the so-called substantive canons (here, “strict
construction” against three-judge panels, which I’ll return to later) come into
play.
This is precisely what a unanimous Supreme Court did earlier this year
in Shular v. United States. 25 Considering the rule of lenity, the Court explained
that the substantive canon “applies only when, after consulting traditional
canons of statutory construction, we are left with an ambiguous statute.” 26 Of
course, the canons of construction—linguistic, substantive, or otherwise—
cannot trump plain statutory language. But if a statute is not clear on its face—
if the ordinary meaning is not immediately self-apparent—then the canons can
help decipher the most textually plausible reading. Indeed, the law is rich in
interpretive conventions. 27 And in the realm of public lawmaking, when judges
Because we believe § 2284(a) isn’t instantly and facially self-evident, we consult the linguistic
canons for guidance.
24 SCALIA & GARNER, supra note 11 § 6, at 69.
25 140 S. Ct. 779 (2020).
26 Id. at 787 (quoting United States v. Shabani, 513 U.S. 10, 17 (1994)).
See SCALIA & GARNER, supra note 11, at xxvii (“In legal systems, there are linguistic
27
usages and conventions distinctive to private legal documents in various fields and to
governmental legislation. And there are jurisprudential conventions that make legal
interpretation more than just a linguistic exercise.”).
27
Case: 19-60133 Document: 00515457870 Page: 28 Date Filed: 06/18/2020
No. 19-60133
are grappling with enacted texts, linguistic canons apply commonsensical rules
of syntax to help us decode the meaning of language. 28
Back to the motions panel’s NCAA example, which uses informal,
everyday phrasing to smuggle in an assumed “natural” reading of formal,
statutory phrasing. This is classic question-begging. The premise (the canon is
inapplicable when the text is clear) assumes the truth of the conclusion (the
text is clear). But unlike the NCAA example, whose sense is a no-brainer,
§ 2284(a) isn’t run-of-the-mill, colloquial speech whose meaning is instantly
obvious. Its particular grammatical construction makes it a statutory brain
teaser, even for seasoned judges. And in such cases, we consult the linguistic
canons, salutary rules of thumb about how people use language.
The series-qualifier canon ought to be applied—together with its critical
“determiner” exception. We recently (and correctly) applied the canon—
determiner and all—in Vaughn, holding that a determiner divided a statute
into discrete parts, thus revealing its “correct and more natural” reading. 29 So
too here. The precise syntax of § 2284(a) indicates no carryover: A three-judge
court is required for all apportionment challenges to state maps.
C
But the series-qualifier canon isn’t the only relevant interpretive canon.
The surplusage canon, a contextual canon, also applies. And notably, while it
is true that one canon may be subject to defeasance by another, here, both the
series-qualifier canon and the surplusage canon point in the same direction.
The inferences drawn are not competing, but complementary.
28 Amy Coney Barrett, Substantive Canons and Faithful Agency, 90 B. U. L. REV. 109,
117 (2010).
29 Vaughn, 907 F.3d at 195.
28
Case: 19-60133 Document: 00515457870 Page: 29 Date Filed: 06/18/2020
No. 19-60133
The surplusage canon, lauded as a “cardinal principle of statutory
construction” by the Supreme Court, 30 teaches “it is no more the court’s
function to revise by subtraction than by addition.” 31 In other words, every
word should be given effect, and none should be read to duplicate another
provision or to have no consequence. Because we presume that legal drafters
don’t include words that have no effect, we “avoid a reading that renders some
words altogether redundant.” 32 The surplusage canon has a sterling pedigree
in the Supreme Court 33 and in this court. 34 And as the motions panel conceded,
the canon is implicated here because Thomas’s reading renders the phrase “the
apportionment” meaningless. 35 Thomas blue-pencils the statute to require a
three-judge panel only when a suit challenges:
the constitutionality of the apportionment of congressional
districts or the apportionment of any statewide legislative body.
Thomas avers that “the apportionment” “is simply a few redundant
words.” But that’s precisely what the surplusage canon seeks to avoid, the
notion that Congress’s words can be deemed idle, pointless, or nonoperative.
It’s the business of courts to take lawmakers at their word, and to presume
they meant what they said. On this vital point, the Supreme Court has been
30 Bennett v. Spear, 520 U.S. 154, 173 (1997).
31 SCALIA & GARNER, supra note 11 § 26, at 175.
32 Id.
33Nat’l Ass’n of Mfrs. v. Dep’t of Def., 138 S. Ct. 617, 632 (2018) (“As this Court has
noted time and time again, the Court is ‘obliged to give effect, if possible, to every word
Congress used.’ ”) (quoting Reiter v. Sonotone Corp., 442 U.S. 330, 339 (1979)).
34 Tex. Educ. Agency v. U.S. Dep’t of Educ., 908 F.3d 127, 133 (5th Cir. 2018)
(describing the canon as a “cardinal principle of statutory construction”) (quoting Bennett,
520 U.S. at 173).
35 Thomas, 919 F.3d at 305 n.5.
29
Case: 19-60133 Document: 00515457870 Page: 30 Date Filed: 06/18/2020
No. 19-60133
unsubtle: “[W]e must give effect to every word that Congress used in the
statute.” 36
Thomas also insists, quoting Justice Thomas, that we need not “avoid
surplusage at all costs. It is appropriate to tolerate a degree of surplusage
rather than adopt a textually dubious construction that threatens to render
the entire provision a nullity.” 37 But the canon’s application here wouldn’t
render any part of the provision a nullity; it would just lead to a result that
Thomas dislikes. It’s true we don’t avoid surplusage at all costs—but we do,
and we must, avoid it. Here, the surplusage canon counsels against slighting
even “a few redundant words” in the statute. Those words are given robust
meaning by Mississippi’s reading: If “constitutionality of” does not carry over,
then “the apportionment” is not rendered superfluous.
D
Thomas does ask us to employ one canon of construction—“that statutes
should not be construed in a way that leads to absurd results.” 38 Our precedent,
however, is not on Thomas’s side. As we recently held, “The absurdity bar is
high, as it should be. The result must be preposterous, one that ‘no reasonable
person could intend.’ ” 39 Justice Scalia and Garner essentially cabin its use to
“scrivener’s error.” 40 As Justice Story put it, the canon’s use must be limited to
situations “where the absurdity and injustice of applying the provision to the
36 Lowe v. SEC, 472 U.S. 181, 207 n.53 (1985).
37 See United States v. Atl. Research Corp., 551 U.S. 128, 137 (2007).
38 See Milavetz, Gallop & Milavetz, P.A. v. United States, 559 U.S. 229, 245–46 (2010)).
39Tex. Brine Co. L.L.C v. Am. Arbitration Ass’n., Inc., No. 18-31184, WL 1682777, at
*3 (5th Cir. Apr. 7, 2020) (quoting SCALIA & GARNER, supra note 11 § 37, at 237).
40SCALIA & GARNER, supra note 11 § 37, at 234. A scrivener’s error is an inadvertent
typo. For example, a court can properly interpret “third partly” as “third party.” Id.
30
Case: 19-60133 Document: 00515457870 Page: 31 Date Filed: 06/18/2020
No. 19-60133
case would be so monstrous, that all mankind would, without hesitation, unite
in rejecting the application.” 41
Thomas’s alleged absurdity: “state legislative redistricting cases brought
solely under § 2 would be heard by three-judge courts while similar
congressional redistricting cases would be heard by single judges.” But having
a different judicial mechanism for hearing challenges to federal districts than
to state districts seems a quintessential policy judgment for Congress. It may
be good (or bad) policy, but it’s light years away from absurd. 42
In sum, the absurdity doctrine is inapposite here. The only two relevant
canons of construction—the series-qualifier canon and the surplusage canon—
work in tandem in favor of the State’s “three-judge court” position.
II
The pertinent canons provide guidance enough to arrive at a conclusion:
§ 2284(a) requires a three-judge district court for any challenges to the
apportionment of statewide legislative bodies. One of the arguments against
relying on the canons here is that this reading conflicts with § 2284(a)’s
statutory history. The motions panel majority reasoned that, when Congress
amended § 2284(a) in 1976, Congress could not have “had a special concern
41 Id. § 37, at 237 (quoting 1 JOSEPH STORY, COMMENTARIES ON THE CONSTITUTION OF
THE UNITED STATES § 427, at 303 (2d ed. 1858)).
42 We need not theorize why Congress would treat state legislative and congressional
redistricting differently because the absurdity doctrine doesn’t require such speculation. But
as Judge Clement noted in her motions panel dissent, the answer may lie in federalism: “It
is entirely plausible that Congress wanted federal courts to show more deference to state
reapportionment plans that only affect state interests than to state reapportionment plans
which affect national interest.” Thomas, 919 F.3d at 323 (Clement, J., dissenting). This
reasoning is bolstered by Supreme Court precedent, which, as Judge Clement notes, affords
states “greater latitude in creating state legislative districts than in creating congressional
districts.” Id. (citing Gaffney v. Cummings, 412 U.S. 735, 742–44 (1973)).
31
Case: 19-60133 Document: 00515457870 Page: 32 Date Filed: 06/18/2020
No. 19-60133
with statutory challenges to the drawing of state legislative districts.” 43 That’s
because, “there was no practice of statutory challenges to state legislative
apportionment that Congress needed to address in 1976.” 44
First, our surest guide to what Congress pondered is what Congress
passed. 45 But if anything, the statutory history strengthens, rather than
weakens, the State’s reading of § 2284(a). Almost twenty years ago, the Third
Circuit conducted a helpful analysis of the statutory history of § 2284(a) in
Page v. Bartels. 46 In that case, the only circuit opinion addressing today’s issue
in any depth, the question was whether a single judge could hear a section 2
claim when a constitutional claim was also asserted. The Third Circuit ruled
that when a plaintiff brings both constitutional and statutory challenges, a
single district judge can’t decide the statutory claim while reserving the
constitutional claims for a three-judge court. The Page court did “not believe
that Congress made a deliberate choice to distinguish between constitutional
43 Thomas, 919 F.3d at 307.
44 Id.
45 We pause to note the important distinction between mining legislative history
(which is highly disfavored in the Fifth Circuit) and statutory history (which isn’t). See, e.g.,
id. at 306 (discussing statutory history while discounting legislative history); see also, BNSF
Ry. Co. v. Loos, 139 S. Ct. 893, 906 (2019) (Gorsuch, J., dissenting) (“[T]he statutory history
I have in mind here isn’t the sort of unenacted legislative history that often is neither truly
legislative (having failed to survive bicameralism and presentment) nor truly historical
(consisting of advocacy aimed at winning in future litigation what couldn’t be won in past
statutes). Instead, I mean here the record of enacted changes Congress made to the relevant
statutory text over time, the sort of textual evidence everyone agrees can sometimes shed
light on meaning.”); SCALIA & GARNER, supra note 11, at 440 (contrasting statutory history—
“[t]he enacted lineage of a statute”—with legislative history); In re Crocker, 941 F.3d 206,
213 (5th Cir. 2019) (same); Mei Xing Yu v. Hasaki Rest., Inc., 944 F.3d 395, 418 (2d Cir. 2019)
(Calabresi, J., dissenting) (same); Chhetri v. United States, 823 F.3d 577, 587 n.13 (11th Cir.
2016) (same).
46248 F.3d 175 (3d Cir. 2001). The Page court did not undertake a textual analysis of
§ 2284(a), likely because neither party advocated such an analysis.
32
Case: 19-60133 Document: 00515457870 Page: 33 Date Filed: 06/18/2020
No. 19-60133
apportionment challenges and apportionment challenges brought under § 2 of
the Voting Rights Act.” 47 This is because, when the relevant language from the
three-judge court statutes was revised in 1976, Ҥ 2 of the Voting Rights Act
was not available to litigants seeking to challenge apportionment.” 48 Most
challenges were constitutional, and the “established statutory basis for such
apportionment challenges was § 5 of the Voting Rights Act, whose own
statutory provisions required the convening of a three judge-court.” 49
As late as 1980, the Supreme Court “had not even definitely determined
whether § 2 of the Voting Rights Act created a private right of action for
voters.” 50 So when our contested language was drafted almost 45 years ago,
Congress would have expected all apportionment challenges to go to a three-
judge panel, as all cases attacking “the legitimacy of the state legislative
apportionment” are “highly sensitive matters.” 51 To the extent that voters had
a statutory private right of action, they did so under section 5, which, again,
provided for a three-judge panel.
The motions panel majority inferred from this statutory history that
Congress could not have had a “special concern” with statutory challenges in
1976. But it is Thomas’s reading of § 2284(a) that separates statutory and
constitutional challenges for disparate treatment when it comes to state maps.
We aren’t finding a special exception for statutory challenges. Our reading of
§ 2284(a) puts statutory and constitutional challenges to state maps on equal
47 Id. at 189.
48 Id.
49 Id. at 189–90 (citations omitted) (emphasis in original).
50 Id. at 189.
51 Id. at 190.
33
Case: 19-60133 Document: 00515457870 Page: 34 Date Filed: 06/18/2020
No. 19-60133
footing, which affirms what the Third Circuit found twenty years ago:
“Congress was concerned less with the source of the law on which an
apportionment challenge was based than on the unique importance of
apportionment challenges generally.” 52
The Third Circuit was correct. The enacted lineage of the Act, how the
statutory text changed over time, tracks § 2284(a)’s straight-ahead meaning in
light of the relevant interpretive canons. The mechanism of a three-judge
district court is meant to be a safety valve in “confrontations between state and
federal power or in circumstances involving a potential for substantial
interference with government administration.” 53 Those federalism concerns
are undiminished when a plaintiff opts for a purely statutory challenge to a
state reapportionment plan instead of a constitutional one. 54 When it comes to
52 Id. (emphasis in original).
53 Allen v. State Bd. of Elections, 393 U.S. 544, 562 (1969); see also Chapman v. Meier,
420 U.S. 1, 14 (1975) (describing apportionment challenges as “regular grist for the three-
judge court”); Gaffney, 12 U.S. at 742–44 (noting “fundamental differences” between state
and federal line-drawing and that “substantial state considerations” give states greater
latitude in drawing state maps than in drawing congressional maps).
54 In addition to looking at the statutory history of § 2284(a), the Page court also
examined its legislative history. So do the parties here. No thanks. Scouring legislative
detritus prone to contrivance is more likely to yield confusion than precision. Zedner v. United
States, 547 U.S. 489, 511 (2006) (Scalia, J., concurring in the judgment) (“[T]he use of
legislative history is illegitimate and ill advised in the interpretation of any statute . . . .”).
We decline any invitation to supplant (or to supplement) Congress’s chosen language, even
though a 1975 Senate Judiciary Committee Report supports our conclusion that three judges
are required, stating that “[t]hree-judge courts would continue to be required . . . in cases
under the Voting Rights Act of 1965, 42 U.S.C. section 1971g, 1973(a), 1973c and 1973h(c).”
Report at 9, 1976 U.S.C.C.A.N. at 1996. Section 1973(a) is none other than section 2(a), the
statutory basis for Thomas’s claim. On the first page of the Committee’s Report, entitled
“PURPOSE OF BILL,” the committee noted that “three-judge courts would be retained . . .
in any case involving congressional reapportionment or the reapportionment of any statewide
legislative body.” Report at 1, 1976 U.S.C.C.A.N. at 1988. The Committee declared that three-
judge courts should be preserved for “reapportionment of a statewide legislative body because
it is the judgment of the committee that these issues are of such importance that they ought
to be heard by a three-judge court.” Report at 9, 1976 U.S.C.C.A.N. at 1996. Again, no matter.
34
Case: 19-60133 Document: 00515457870 Page: 35 Date Filed: 06/18/2020
No. 19-60133
state maps affecting (and reflecting) state interests and state considerations,
§ 2284(a) treats all challenges—statutory and constitutional—equally.
Indeed, had Congress wanted single district judges to decide standalone
section 2 challenges to state maps, it certainly had better, simpler ways of
saying so. Congress could have made things crystal clear by writing simply
(and without surplusage) that a three-judge court is needed to decide only: 55
• Option 1: “the constitutionality of the apportionment of
congressional districts or [the apportionment] of any statewide
legislative body”
• Option 2: “the constitutionality of the apportionment of
congressional districts or [of] the apportionment of any
statewide legislative body”
• Option 3: “the constitutionality of [either] the
apportionment of congressional districts or the apportionment
of any statewide legislative body”
But Congress said no such thing. Instead Congress determined to use a
determiner. Our task is not to imagine what the 1976 Congress would have
wanted. It is to discern meaning in the words the 1976 Congress actually
passed. In sum, § 2284(a)’s on-the-books history—again, not the legislative
history, but the history of the legislation—reinforces what the canons reveal:
A single district judge cannot adjudicate a section 2-only challenge to a state
map.
The statute itself is what constitutes the law. And since we are a Nation of laws, not of
legislative histories, we decline the legal scavenger hunt that turns statutory construction
into statutory excavation.
55 Thomas, 919 F.3d at 322–23 (Clement, J., dissenting).
35
Case: 19-60133 Document: 00515457870 Page: 36 Date Filed: 06/18/2020
No. 19-60133
III
Another argument against Mississippi’s reading of § 2284(a) is the
principle that “congressional enactments providing for the convening of three-
judge courts must be strictly construed.” 56 This is a substantive canon of
construction, like the rule of lenity (that penal statutes should be given a
narrow construction) or constitutional avoidance (that statutory language
should be given a saving construction). 57 But unlike linguistic canons that look
for statutory meaning within the words themselves, “substantive canons
advance policies independent of those expressed in the statute.” 58 Exhibit A is
the “strict construction” gloss on three-judge statutes, born 80 years ago out of
the Supreme Court’s desire to keep its docket small, plus a concern that three-
judge jurisdiction “entails a serious drain upon the federal judicial system,
particularly in . . . all but the few great metropolitan areas.” 59
But there is a sequence that must be followed. As the Supreme Court has
repeatedly made clear, most recently a few months ago, substantive canons of
construction are not applied at the outset of textual inquiry. 60 Why? Because a
substantive canon (and the social policy it enhances) can never defeat concrete
text (and the congressional policy it enshrines). Moreover, even if a statute is
deemed ambiguous, traditional canons of interpretation must be invoked first.
And if those text-centric canons yield an answer, then that’s that; the thumb-
on-the-scale substantive canons have no role.
56 Allen, 393 U.S. at 561.
57 See generally, Barrett, supra note 28.
58 Id. at 110.
59 Phillips v. United States, 312 U.S. 246, 250 (1941).
60 Shular, 140 S. Ct. at 787.
36
Case: 19-60133 Document: 00515457870 Page: 37 Date Filed: 06/18/2020
No. 19-60133
Take the rule of lenity. Described by then-Professor Barrett as “a rule of
thumb for choosing between equally plausible interpretations of ambiguous
text,” 61 the rule of lenity has been around for at least half a millennium. 62 It
sometimes plays a role, but one more cameo than starring—and only in the
final act, if at all. Again, there’s a sequence to things. And that’s because there’s
a supremacy to things. Faithful statutory interpreters rightly insist,
vehemently so, on legislative supremacy: taking Congress at its word. Thus,
there must be ambiguity before there can be lenity. 63 So even if a less-harsh
result is grammatically possible, courts are duty-bound to seek the truest
meaning, not the tenderest one. 64 If a statute has an interpretation that is most
plausible, as opposed to dueling interpretations that are equally plausible,
then the lenity canon remains holstered.
As noted above, the Supreme Court made this exact point recently—and
unanimously—in Shular. Speaking through Justice Ginsburg, the Court
refused to apply the rule of lenity, stressing that it “applies only when, after
consulting traditional canons of statutory construction, we are left with an
ambiguous statute.” 65 The Shular Court put it simply: “Here, we are left with
no ambiguity for the rule of lenity to resolve.” 66 Justice Kavanaugh joined
Shular in full but wrote separately to underscore the primacy of interpretive
61 Barrett, supra note 28, at 109.
62 Id. at 128–29.
63Id. at 131 (“Courts repeatedly emphasized that lenity could never overcome the
ordinary meaning of a statute; on the contrary, the principle applied only in the event of
ambiguity.” (citations omitted)).
64Id. at 155 (“[T]he best interpretation of a penal statute should always trump a more
lenient but less plausible one.”).
65 Shular, 140 S. Ct. at 787 (quoting Shabani, 513 U.S. at 17).
66 Id.
37
Case: 19-60133 Document: 00515457870 Page: 38 Date Filed: 06/18/2020
No. 19-60133
canons over substantive canons, explaining that the rule of lenity plays a role
at the end of the interpretive process, not at the beginning, and only if, after
all the traditional tools have been employed, the statute remains not just
ambiguous, but “grievously ambiguous.” 67
The same principle applies here. As the motions panel conceded, Allen’s
strict-construction preference applies only “[t]o the extent there is
ambiguity.” 68 And if textual canons succeed in revealing § 2284(a)’s most
forthright meaning, as they do here, we need not—indeed, must not—indulge
malleable, atextual canons that beckon us to advance policies unexpressed in
the statute itself. Respectfully, we must not forgo § 2284(a)’s most
straightforward reading, gleaned from traditional linguistic tools, in favor of a
strained reading that stretches Congress’s words in order to further exogenous
judicial-policy goals. Unambiguous statutes must be left alone, neither
expanded (liberally construed) nor contracted (strictly construed).
But even if the traditional, text-focused canons (series-qualifier and
surplusage) did not erase ambiguity, Allen itself shows that invoking “strict
construction” does not guarantee checkmate. A central question in Allen was
whether section 5 of the Voting Rights Act authorizes three-judge courts only
in declaratory-judgment suits brought by States or also in section 5 suits
brought by private litigants. 69 The Supreme Court held that, even strictly
construing the three-judge language, Congress wanted all section 5 disputes to
be heard by three-judge courts. 70 Despite the difficulty of judicial
67 Id. at 789 (Kavanaugh, J., concurring).
68 Thomas, 919 F.3d 298 at 308.
69 Allen, 393 U.S. at 561.
70 Id. at 563. Moreover, in his partial concurrence in Allen, Justice Harlan noted that,
for section 5 cases, “there is no good reason to invoke the normal rule that three-judge court
38
Case: 19-60133 Document: 00515457870 Page: 39 Date Filed: 06/18/2020
No. 19-60133
administration that three-judge courts present, the Court reasoned that
“Congress has determined that three-judge courts are desirable in a number of
circumstances involving confrontations between state and federal power or in
circumstances involving a potential for substantial interference with
government administration.” 71 Importantly, the Court noted, “The Voting
Rights Act of 1965 is an example. Federal supervision over the enforcement of
state legislation always poses difficult problems for our federal system.” 72
The Court went on to emphasize that these problems are even more
severe when “the enforcement of state enactments may be enjoined and state
election procedures suspended because the State has failed to comply with a
federal approval procedure.” 73 To be sure, section 5 is distinct from section 2,
so Allen doesn’t directly control. But Allen is eminently instructive because its
holding did not hinge on preclearance. There’s certainly not the same degree
of confrontation between state and federal governments when suits are
brought by individuals. But the Allen Court determined that the “potential for
disruption of state election procedures remains” when individuals bring suit. 74
And suits by individuals still create the prospect, as here, of a single federal
judge overseeing state election procedures. These grave federalism concerns
are not erased just because the federal oversight is by a life-tenured Article III
judge rather than by an Article I department. 75
statutes should be construed as narrowly as possible.” Id. at 582 n.1 (Harlan, J., concurring
in part and dissenting in part). That’s because, generally, those bringing section 5 claims
could also bring constitutional claims. Id. The same is true here.
71 Id. at 562.
72 Id.
73 Id.
74 Id. at 563.
75 See In re Gee, 941 F.3d 153, 166–67 (5th Cir. 2019) (per curiam) (explaining that
39
Case: 19-60133 Document: 00515457870 Page: 40 Date Filed: 06/18/2020
No. 19-60133
Moreover, when Allen was decided in 1969, three-judge courts were
required to hear all claims for injunctive relief against States (and their
officials) and all constitutional claims seeking to enjoin a federal statute. 76
Appeals from those actions were taken directly to the Supreme Court. 77 This
imposed a heavy judicial-administration burden. And it’s against that
backdrop that the Supreme Court held that three-judge statutes were to be
strictly construed. Then, in 1976, Congress significantly narrowed the universe
of actions that required three-judge courts. 78 And after Congress stepped in, it
seems the Supreme Court has found little need to prophylactically apply the
strict-construction gloss. 79
When Allen is considered in light of the old three-judge-court regime
under § 2281 and § 2282, the fact that the Supreme Court insisted on three
judges for all section 5 challenges is even more telling. Despite the judicial-
administration burden imposed by the former three-judge-court statutes, the
Court repeated its serious federalism concerns with the potentially disruptive
oversight of State election procedures. Those concerns are no less relevant
here, especially now that the federal judiciary no longer faces the “serious
drain” of three-judge-jurisdiction that concerned the Phillips Court 80 years
federal oversight of state law by a federal court raises federalism concerns).
76 See 28 U.S.C. § 2281 (state actions); id. § 2282 (federal statutes).
77 See id. § 1253.
78See Pub. L. No. 94–381, 90 Stat 1119, 1119 (1976). Sections 2281 and 2282 were
repealed, and § 2284 was amended to require three-judge courts in only two scenarios: (1)
“when otherwise required by Act of Congress,” and (2) “when an action is filed challenging
the constitutionality of the apportionment of congressional districts or the apportionment of
any statewide legislative body.” Id.
79 As best we can tell, the Court hasn’t applied Allen’s strict-construction admonition—
at least for whether a three-judge court should be convened—since Congress acted in 1976.
40
Case: 19-60133 Document: 00515457870 Page: 41 Date Filed: 06/18/2020
No. 19-60133
ago. 80
IV
Finally, Thomas invokes custom, stressing that no circuit court has ever
interpreted § 2284(a) to require a three-judge court in a section 2-only
challenge to a state legislative map. The motions panel likewise emphasized
that many years of settled litigation practice “seems enough to prevent
[Mississippi] from showing a strong likelihood of succeeding on this issue.” 81
True, none of our sister circuits have ever confronted this question. 82
There’s a simple reason for that: No defendant has ever pressed it. 83 In most
Judge Costa’s opinion avers that our position “jettison[s] Supreme Court precedent.”
80
Costa Op. at 15. But we don’t suggest that courts should abandon strict construction of three-
judge statues. We simply follow the Supreme Court’s own strict-construction analysis in
Allen. The fact that the three-judge regime was so radically curtailed after Allen does not
mean that strict construction is now irrelevant. But it does make Allen more remarkable in
hindsight. Even when the Supreme Court had every reason to be concerned with the burdens
of convening three-judge courts, it still held that the potential disruption of state election
procedures counseled in favor of such courts.
81 Thomas, 919 F.3d at 305.
Thomas also cites Harris v. Arizona Independent Redistricting Commission, 136 S.
82
Ct. 1301, 1306 (2016), for the proposition that, in the § 2284(a) context, the Supreme Court
has “reviewed [single-judge] decisions without a doubt as to jurisdiction.” First, Harris was
reviewing the decision of a three-judge court, so Thomas’s assertion has zero support. Second,
Harris involved a constitutional challenge, so there was no dispute about whether the three-
judge court was proper. Thomas relies on a single parenthetical to a citation in Harris: “See
28 U.S.C. § 2284(a) (providing for the convention of [a three-judge court] whenever an action
is filed challenging the constitutionality of apportionment of legislative districts).” This
passing parenthetical citation merely provides the authority for the convening of the three-
judge court that heard that case before it arrived at the Supreme Court. Referring to this
citation as dicta is an insult to dicta.
While no other circuit courts have confronted this question, two district courts have,
83
both in 2019: Chestnut v. Merrill, 356 F. Supp. 3d 1351 (N.D. Ala. 2019) and Johnson v.
Ardoin, 2019 WL 4318487 (M.D. La. Sept. 12, 2019). Notably, Chestnut, which held that
§ 2284(a) did not require a three-judge court, fails to acknowledge the series-qualifier canon
or the superfluous language that results from its reading. Johnson reached the same
conclusion as Chestnut but was bound by the decisions of our motions and merits panels in
this case.
41
Case: 19-60133 Document: 00515457870 Page: 42 Date Filed: 06/18/2020
No. 19-60133
reapportionment cases, statutory claims are asserted alongside constitutional
claims, rendering moot the 3-judge vs. 1-judge question. But neither litigation
practice, however customary, nor “settled understanding,” 84 however inured,
can stand against the law’s demands. 85
This is a bedrock principle. Today’s case is, after all, a statutory-
interpretation dispute, and what matters most are Congress’s words, not the
until-now-unchallenged assumptions of litigants. The Supreme Court has put
it plainly: “a ‘long-established practice’ does not justify a rule that denies
statutory text its fairest reading.” 86 Our loyalty runs to Congress and its
commands. And as Congress’s faithful agents, we must choose fidelity to
explicit enactments over the continuity of implicit arrangements. Supreme
Court examples abound, such as the Court’s 1998 decision rejecting
Pennsylvania’s argument that Congress would never have imagined that the
Americans with Disabilities Act would apply to state prisoners. 87 The Court
explained that where statutory meaning is clear (as in today’s case once you
apply the linguistic canons), it’s “irrelevant” whether a specific application was
anticipated by Congress. 88 Our duty is to legislative text, not to litigation
habits that, until now, have gone merrily along, unexamined.
Just this week, the Supreme Court issued a landmark decision, holding
84 Thomas, 938 F.3d at 145.
85 One reason for the dearth of precedent: section 2 “results cases” are rarely pursued,
at least until recently, without accompanying constitutional claims under the Fourteenth or
Fifteenth Amendment.
86Armstrong v. Exceptional Child Ctr., Inc., 575 U.S. 320, 329 (2015); see also NLRB
v. Noel Canning, 573 U.S. 513, 572 (2014) (Scalia, J., concurring) (“We did not hesitate to
hold the legislative veto unconstitutional even though Congress had enacted, and the
President signed, nearly 300 similar provisions over the course of 50 years.”).
87 Pa. Dep’t of Corr. v. Yeskey, 524 U.S. 206 (1998).
88 Id. at 212.
42
Case: 19-60133 Document: 00515457870 Page: 43 Date Filed: 06/18/2020
No. 19-60133
that the 56-year-old Civil Rights Act forbids workplace discrimination on the
basis of sexual orientation or gender identity. 89 Specifically, the Court declared
that “because of sex” encompasses “because of sexual orientation or gender
identity.” The latter is not distinct from sex discrimination, but a form of it.
Hearteningly, all nine Justices applied textual analysis to Title VII, as we do
today with § 2284(a), but, just like us, they reached polar-opposite conclusions.
In dissent, Justices Alito and Thomas charged the majority with “disregarding
over 50 years of uniform judicial interpretation” 90 and protested that “there is
not a shred of evidence that any Member of Congress interpreted the statutory
text that way when Title VII was enacted.” 91 The Bostock majority did not—
indeed, could not—dispute those facts. 92 It just deemed them immaterial,
insisting that what matters (all that matters) is the literal text within a
statute’s four corners—what it called “Title VII’s plain terms.” 93
A time traveler from 1964 would doubtless express astonishment that
Congress had, unwittingly and unbeknownst to everyone, equated sex
discrimination with sexual orientation discrimination (much less with gender
identity discrimination)—and that it had done so by adopting a one-word
amendment (inserting “sex”) from a representative who was cynically trying to
89 Bostock v. Clayton County, 2020 WL 3146686, at *3 (U.S. June 15, 2020).
90 Id. at *39 (Alito, J., dissenting). Justices Alito and Thomas derided Bostock as pure
“legislation”—a “pirate ship” sailing under a false “textualist flag”—an “arrogant” and
“radical decision” that constitutes a “brazen abuse of our authority to interpret statutes. Id.
at *18–19, *21, *39.
91 Id. at *20.
92Nor did the Court dispute that the EEOC, the agency charged with enforcing Title
VII, went nearly a half-century before it began asserting such coverage. Id. at *20.
93 Id. at *8.
43
Case: 19-60133 Document: 00515457870 Page: 44 Date Filed: 06/18/2020
No. 19-60133
scuttle the entire Civil Rights Act. 94 But the Bostock majority focused on the
“broad language” that Congress adopted, not on the ripple effects, however
unforeseen, that flowed from it five decades later. 95 The Court thus gave no
interpretive weight to the fact that not a single drafter of Title VII in 1964
intended, noticed, or anticipated that “because of . . . sex” would cover
discrimination against homosexual or transgender persons. The Court
remarked that resorting to “expected applications” or only those “foreseen at
the time of enactment . . . seeks to displace plain meaning of the law in favor
of something lying beyond it.” 96 Text is paramount—“only the words on the
page constitute the law” 97—and if those words lead to “unexpected
consequences,” so be it. 98
Settled practices matter not, nor does the “unanimous consensus” among
the courts of appeals stretching across a half-century. 99 As the Court put it:
“Judges are not free to overlook plain statutory commands on the strength of
nothing more than suppositions about intentions or guesswork about
expectations.” 100 In the Bostock majority’s view, language codified by
lawmakers is like language coded by programmers. A computer programmer
may write faulty code, but the code will perform precisely as written,
regardless of what the programmer anticipated. Courts, no less than
94 Id. at *16. The term “transgender” actually wasn’t coined until the following decade.
Id. at *34 (Alito, J., dissenting).
95 Id. at *18.
96 Id. at *15.
97 Id. at *4.
98 Id. at *3, *43.
99 Id. at *43 (Alito, J., dissenting).
100 Id. at *18.
44
Case: 19-60133 Document: 00515457870 Page: 45 Date Filed: 06/18/2020
No. 19-60133
computers, are bound by what was typed, and also by what was mistyped.
What this means for Title VII: “When a new application emerges that is both
unexpected and important,” said the Court, it is no answer to “have us merely
point out the question, refer the subject back to Congress, and decline to
enforce the plain terms of the law in the meantime.” 101 We are bound by
Bostock, whose ascetic interpretive rules for Civil Rights Act cases apply with
equal force to Voting Rights Act cases.
Judge Costa’s opinion asserts virtually the same arguments as the
Bostock dissenters, appealing to “venerable” understandings, “widely accepted
meaning,” and “uniform caselaw.” 102 He labels the State’s § 2284(a) argument
“unprecedented,” invoking Justice Scalia’s colorful elephants-hiding-in-
mouseholes aphorism. 103 This image, born of nondelegation concerns but now
taking hold beyond its administrative law origins, is indeed vivid, like Justice
Scalia’s other renowned animalistic turns of phrase: “this wolf comes as a
wolf,” 104 or “jackals stealing the lion’s kill” (the latter about Article III
judges). 105 But pachyderms and rodents are in the eye of the beholder, leading
the Court to apply the maxim “seemingly haphazardly.” 106 What some Justices
see as “big game,” others see as “just a regular rodent.” 107 Reasonable judicial
101 Id. at *15.
102 Costa Op. at 1, 17.
103 Id. at 1.
104 Morrison v. Olson, 487 U.S. 654, 699 (1988) (Scalia, J., dissenting).
105 FCC v. Fox Television Stations, Inc., 556 U.S. 502, 526 (1988).
106 Jacob Loshin & Aaron Nielson, Hiding Nondelegation in Mouseholes, 62 ADMIN. L.
REV. 19, 45 (2010).
107 Id. at 46.
45
Case: 19-60133 Document: 00515457870 Page: 46 Date Filed: 06/18/2020
No. 19-60133
minds can, and do, differ. 108 Perhaps what lurks in the mousehole is, actually,
“a rather plump mouse.” 109 Or perhaps what some take for a mousehole is,
actually, “a rather cramped circus tent.” 110 The “elephants in mouseholes”
doctrine, while pithy, is beset with unpredictability: “[T]hose in the majority
one day are in the dissent the next, and vice versa.” 111 But there is a graver
problem than inconsistent, “I know it when I see it” application. The elephants-
hiding-in-mouseholes canon/doctrine/principle supplants textualism with
purposivism, summoning a statute’s overarching purpose to inform what the
law’s actual words mean (to the extent a singular legislative purpose can be
divined given the innumerable, untidy trade-offs baked into bill language). Our
textualist precedent, one that prizes “bright lines and sharp corners,” 112 is in
tension with this purposivist premise. The truest indication of what Congress
intended is what Congress enacted. 113 Thus, the biggest danger of mouseholes
is a methodological one—that the specter of them will be invoked to justify
elephantine departures from statutory text.
In any event, there is another elephant, this one in the room, and it is
not hiding. The voting-rights litigation landscape was transformed by the
Supreme Court’s 2013 decision in Shelby County. Any newness to the State’s
§ 2284(a) argument reflects the newness of post-Shelby County litigation
strategy. Shelby County upended the legal playing field, and voting-rights
lawyers (on both sides of the docket) are nothing if not adaptive. Once the
108 Id. at 46–48 (chronicling the oscillation across cases).
109 Id. at 45.
110 Id.
111 Id.
112 Reed, 923 F.3d at 415.
113 Id.
46
Case: 19-60133 Document: 00515457870 Page: 47 Date Filed: 06/18/2020
No. 19-60133
Supreme Court declared section 4’s outdated coverage formula facially
unconstitutional, thus rendering inoperative section 5’s prophylactic
preclearance regime (which required three-judge courts), the litigation-based
remedies of section 2 instantly took on outsized importance.
Immediately, scholars exhorted courts to give section 2 “special bite,”
such that “section 2 can be made to function like erstwhile section 5 in the post-
Shelby County world.” 114 The goal: “that the courts, in partnership with the
Department of Justice, could reform section 2 so that it fills much of the gap
left by the Supreme Court’s evisceration of section 5.” 115 All to say, any novelty
to Mississippi’s three-judge argument tracks the relative novelty of a section
2-only challenge, “as Voting Rights Act claims and constitutional claims are
usually asserted together,” 116 thus requiring a three-judge court. 117
Today’s question, fundamentally, is about jurisdiction—the very power
of federal courts. And “past practice does not, by itself, create power.” 118 The
power of Article III courts to hear cases is derived from statutes that Congress
enacts. And we must give those statutes their fairest reading, regardless of
how litigants have (or have not) tried their cases up to now, and regardless of
what may be more socially desirable as a policy matter. Judicial duty requires
us to revere, not revise, what Congress has passed.
114Christopher S. Elmendorf & Douglas M. Spencer, Administering Section 2 of the
Voting Rights Act After Shelby County, 115 COLUM. L. REV. 2143, 2147 (2015).
115 Id. at 2143.
116 Thomas, 919 F.3d at 324 (Clement, J., dissenting).
117 Costa Op. at 3 n.2 (“Of course, when a plaintiff brings both constitutional and
statutory challenges, the constitutional hook for three-judge courts sweeps in the statutory
claim.”) (citing Page, 248 F.3d at 191).
118Medellin v. Texas, 552 U.S. 491, 532 (2008) (brackets omitted) (quoting Dames &
Moore v. Regan, 453 U.S. 654, 686 (1981)).
47
Case: 19-60133 Document: 00515457870 Page: 48 Date Filed: 06/18/2020
No. 19-60133
V
Verbis legis tenaciter inhaerendum.
“Hold tight to the words of the law.” This medieval legal maxim, fittingly
the lead epigraph to Reading Law, 119 captures the paramount task of judges
when interpreting legal texts: giving enacted language its soundest, most
honest meaning. In this case, the language of § 2284(a), read in the light of
multiple blackletter interpretive canons, is both the start and the end of our
inquiry.
Requiring only a single judge to decide section 2-only challenges may be
wise policy, but it is not Congress’s enacted policy. While it is a plausible
reading of the statute, it is not the most plausible. The most sure-fire reading
of § 2284(a) is that a three-judge court must decide all challenges to the
apportionment of statewide legislative bodies. The syntax may be fuzzy, but its
sense is not. Twin esteemed canons of interpretation, plus § 2284’s enacted
history, point strongly in the State’s favor.
We have endeavored to give § 2284(a) its most forthright meaning,
discerning rather than distorting. And scrupulous fidelity to text, giving
Congress’s words not merely a bearable interpretation but the best one, the
most textually plausible one, leads us to conclude that the district court lacked
jurisdiction and that its judgment must be vacated.
119 SCALIA & GARNER, supra note 11, at v.
48
Case: 19-60133 Document: 00515457870 Page: 49 Date Filed: 06/18/2020
No. 19-60133
ANDREW S. OLDHAM, Circuit Judge, concurring:
It appears that the only thing left in this case is Plaintiffs’ application
for attorney’s fees. “Of course, a claim for attorney’s fees is not alone sufficient
to preserve a live controversy.” N.Y. State Rifle & Pistol Ass’n, Inc. v. City of
New York, 140 S. Ct. 1525, 1538 (2020) (Alito, J., dissenting) (citing Lewis v.
Cont’l Bank Corp., 494 U.S. 472, 480 (1990)). So I agree that the case is moot.
I write separately to explain why our Munsingwear vacatur also moots
Plaintiffs’ fee application.
I.
The effect of a Munsingwear vacatur is fairly well-known: It “eliminates
[the] judgment” below. United States v. Munsingwear, Inc., 340 U.S. 36, 40
(1950). What’s less appreciated is why.
Think of federal litigation like baseball. In baseball, a team wins nothing
by scoring the most runs in the first inning. Rather, we declare the winner
after nine. See Major League Baseball, Official Baseball Rules, R. 7.01(a)
(2018). Sometimes, uncontrollable circumstances—bad weather or a legal
curfew, for example—bring play to a halt and require calling the game early.
Id. R. 7.02(a). If, at the time the game is called, “five innings have [not] been
completed,” id. R.7.01(c), the umpire declares “No Game,” id. R. 7.01(e).
Neither team wins. Id. R. 1.06. As far as the League is concerned, it’s as if the
game never happened at all.
Litigation is functionally identical. When a party secures a contested
judgment in the district court, it has a lead. But it hasn’t won anything yet. It
must first protect its lead—the judgment—in the court of appeals. If certiorari
is granted, the party must protect it again before the Supreme Court. But if,
“through happenstance,” the case becomes moot before the Supreme Court can
review it, Munsingwear, 340 U.S. at 40, the umpire declares “No Game.”
Inferior-court decisions, like the first innings of baseball games, are “only
49
Case: 19-60133 Document: 00515457870 Page: 50 Date Filed: 06/18/2020
No. 19-60133
preliminary.” Ibid. So they must be vacated and “prevent[ed] . . . from
spawning any legal consequences.” Id. at 41.
Of course, not all cases make their way to the Supreme Court. But many
do reach ours. And we too have a “duty . . . to set aside the decree below” when
“the controversy has become entirely moot” before we can issue a decision.
Great W. Sugar Co. v. Nelson, 442 U.S. 92, 93 (1979) (per curiam) (quotation
omitted). Again, a first-inning lead is no victory.
And with no victory, a plaintiff cannot demand “prevailing party” fees.
For example, the Supreme Court has told us that, when a plaintiff wins a
preliminary injunction but ends up losing the case on the merits, the plaintiff
doesn’t get prevailing-party fees. See Sole v. Wyner, 551 U.S. 74, 78 (2007).
Why? Because, “at the end of the litigation, her initial success is undone and
she leaves the courthouse emptyhanded.” Ibid.
II.
Plaintiffs in this case secured a first-inning lead: judgment in the district
court. But, through no fault of either party, the game was called early without
review by our court. So, despite Plaintiffs’ preliminary success, today’s
Munsingwear vacatur means they leave the courthouse emptyhanded. We’ve
“strip[ped] the decision below of its binding effect.” Deakins v. Monaghan, 484
U.S. 193, 200 (1988).
This spells an end to Plaintiffs’ fee application. Plaintiffs told the district
court they would ask for “prevailing party” fees. See Unopposed Motion for
Extension of Time to Apply for Attorneys’ Fees & Costs, Thomas v. Bryant, No.
18-cv-441, Doc. 94 (S.D. Miss. Mar. 12, 2019). 1 But a plaintiff cannot demand
prevailing party fees without, well, having prevailed.
1 It’s not entirely clear what statute plaintiffs think could provide them “prevailing
party” fees. They told the district court that it was 42 U.S.C. § 1988. That’s wrong. That
statute provides “prevailing party” fees in actions under 42 U.S.C. §§ 1981, 1981a, 1982, 1983,
50
Case: 19-60133 Document: 00515457870 Page: 51 Date Filed: 06/18/2020
No. 19-60133
Plaintiffs in this case “prevailed” no more than the plaintiff in Sole.
Plaintiffs in both cases enjoyed the benefit of the district courts’ decisions for a
time. But, just as a loss at the end of a case terminates the benefit of temporary
injunctive relief, see Sole, 551 U.S. at 84–86, so too does a vacatur take away
whatever benefits were held by the winning party in the district court.
Plaintiffs in both cases had only a “transient victory.” Id. at 78. Plaintiffs in
both cases had that “transient victory” stripped of its legal effect. And plaintiffs
in neither case can demand prevailing party fees. See N.Y. State Rifle & Pistol
Ass’n, 140 S. Ct. at 1538 (Alito, J., dissenting) (noting that a successful [civil-
rights] plaintiff “is eligible for attorney’s fees,” but “dismissing the case as moot
means [plaintiffs] are stuck with the attorney’s fees they incurred”); S-1 v.
State Bd. Of Educ. of N.C., 21 F.3d 49, 51 (4th Cir. 1994) (en banc) (per curiam)
(“[T]he dismissal on appeal of an action . . . as moot operates to vacate the
judgment below, and prevents the plaintiffs from being found prevailing
parties by virtue of post-dismissal events.” (quotation omitted)).
Before the Supreme Court’s decision in Sole, we misunderstood the rules
of the game. We said, for example, “that a determination of mootness neither
precludes nor is precluded by an award of attorneys’ fees. The attorneys’ fees
question turns instead on a wholly independent consideration: whether
plaintiff is a ‘prevailing party.’ ” Murphy v. Fort Worth Indep. Sch. Dist., 334
F.3d 470, 471 (5th Cir. 2003) (per curiam) (quotation omitted). Obviously, the
1985, & 1986, Title IX, RFRA, RLUIPA, Title VI of the Civil Rights Act of 1964, and 34 U.S.C.
§ 12361. See 42 U.S.C. § 1988(b). Plaintiffs sued under none of these statutes. They sued
instead under Section 2 of the Voting Rights Act. See Compl. ¶ 37; Am. Compl. ¶ 37. The
Voting Rights Act has its own prevailing-party fee provision. See 52 U.S.C. § 10310(e). But it
applies only in an “action or proceeding to enforce the voting guarantees of the fourteenth or
fifteenth amendment . . . .” Ibid. Since plaintiffs forswore any relief under the Constitution—
which is the whole reason this case was not referred to a three-judge panel—they cannot
recover fees under § 10310(e) either. As noted in the body text, however, none of this matters
because plaintiffs are not “prevailing parties.”
51
Case: 19-60133 Document: 00515457870 Page: 52 Date Filed: 06/18/2020
No. 19-60133
Supreme Court has now told us that’s wrong: Plaintiffs’ prevailing-party status
is wholly dependent on whether they walk out the courthouse doors with an
enforceable judgment. See Sole, 551 U.S. at 78; Farrar v. Hobby, 506 U.S. 103,
113 (1992) (“No material alteration of the legal relationship between the
parties occurs until the plaintiff becomes entitled to enforce a judgment,
consent decree, or settlement against the defendant.”); cf. Staley v. Harris Cty.,
485 F.3d 305, 314 (5th Cir. 2007) (en banc) (holding plaintiff can be a prevailing
party where we refuse to order Munsingwear vacatur and plaintiff kept its
judgment). Our Munsingwear vacatur deprives Plaintiffs of such a judgment.
With that understanding, I concur.
52