Reply Brief — John Q. Hamm, Commissioner, Alabama Department of Corrections, Petitioner v. Joseph Clifton Smith

No. 24-872

Supreme Court of the United States

──────────────────────────

JOHN Q. HAMM,

COMMISSIONER OF THE ALABAMA

DEPARTMENT OF CORRECTIONS,

Petitioner,

v.

JOSEPH CLIFTON SMITH,

Respondent.

──────────────────────────

ON WRIT OF CERTIORARI TO THE

UNITED STATES COURT OF APPEALS

FOR THE ELEVENTH CIRCUIT

──────────────────────────

REPLY BRIEF FOR PETITIONER

──────────────────────────

Steve Marshall

Attorney General

STATE OF ALABAMA

Office of the Att’y Gen.

501 Washington Ave.

Montgomery, AL 36130

(334) 242-7300

Edmund.LaCour@

AlabamaAG.gov

October 17, 2025

Edmund G. LaCour Jr.

Solicitor General

Counsel of Record

Robert M. Overing

Deputy Solicitor General

Dylan Mauldin

Ass’t Solicitor General

Benjamin M. Seiss

Ass’t Attorney General

Counsel for Petitioner

TABLE OF CONTENTS

TABLE OF AUTHORITIES ...................................... iii

REPLY BRIEF ............................................................. 1

I. Smith’s claim fails because he did not prove

an IQ of 70 or less using any approach to the

cumulative effect of multiple IQ scores. ............. 2

A. Alabama law requires proof of an IQ of 70

or below by a preponderance of evidence. ..... 2

B. Smith did not prove that his five IQ scores

together yield a likely IQ of 70 or below. ...... 3

1. The cumulative effect of Smith’s five IQ

scores must be considered. ....................... 4

2. Smith’s “holistic approach” ignores the

cumulative effect of multiple IQ scores ... 8

3. Smith did not prove that his IQ is 70 or

below using any method for jointly

analyzing his scores. ............................... 10

II. The Court should not expand Atkins with

Smith’s rule denying the cumulative effect of

multiple IQ scores. ............................................ 12

A. Smith concedes that the constitutional

text and history are independent reasons

to reject his position. .................................... 12

B. No nationwide consensus mandates a rule

denying the cumulative effect of IQ scores . 14

ii

C. No precedent requires eroding prong one

with a rule denying the cumulative effect

of IQ scores. .................................................. 15

III.Smith’s IQ score of 78 precludes relief. ............ 19

CONCLUSION .......................................................... 21

iii

TABLE OF AUTHORITIES

Cases

Albarran v. State,

96 So. 3d 131 (Ala. Crim. App. 2011) ................ 2, 3

Atkins v. Virginia,

536 U.S. 304 (2002) ................1, 2, 10, 12-15, 17-20

Brumfield v. Cain,

576 U.S. 305 (2015) ........................................ 16, 19

Bucklew v. Precythe,

587 U.S. 119 (2019) .............................................. 12

Bush v. State,

92 So.3d 121 (Ala. Crim. App. 2009) ..................... 2

Byrd v. State,

78 So.3d 445 (Ala. Crim. App. 2009) ................. 2, 3

City of Grants Pass v. Johnson,

603 U.S. 520 (2024) .................................. 12, 13, 15

Ex parte Jerry Smith,

213 So.3d 214 (Ala. 2003) ...................................... 2

Ex parte Perkins,

851 So.2d 453 (Ala. 2002) .................................. 1-3

Ford v. Wainwright,

477 U.S. 399 (1986) .............................................. 14

Hall v. Florida,

572 U.S. 701 (2014) ............................. 2, 3, 6, 7, 10,

................................................................... 11, 14-21

Hamm v. Smith,

604 U.S. 1 (2024) ...................................... 2, 6, 9, 16

iv

Jones v. Mississippi,

593 U.S. 98 (2021) .....................................15, 19-20

Kahler v. Kansas,

589 U.S. 271 (2020) .............................................. 21

Kennedy, v. Louisiana,

554 U.S. 407 (2008) ........................................ 10, 13

Moore v. Texas,

581 U.S. 1 (2017) ........................................ 8, 13, 19

Penry v. Lynaugh,

492 U.S. 302 (1989) .................................. 13, 14, 21

Reeves v. State,

226 So.3d 711 (Ala. Crim. App. 2016 .................... 8

Roper v. Simmons,

543 U.S. 551 (2005) .................................. 13, 20, 21

Stanford v. Kentucky,

492 U.S. 361 (1989) .............................................. 13

United States v. Wilson,

170 F. Supp. 3d 347 (E.D.N.Y. 2016) .................... 9

Rules

Ala. R. Crim. P. 32.3.................................................... 3

Other Authorities

Am. Ass’n on Intellectual and Dev. Disabilities,

Intellectual Disability: Definition, Diagnosis,

Classification, and Systems of Supports

(12th ed., 2021) (“AAIDD-12”) ............................ 18

v

Am. Psychiatric Ass’n,

Diagnostic and Statistical Manual of Mental

Disorders (5th ed., Text Revision 2022)

(“DSM-5-TR”) ............................................. 5, 17-19

Brief of APA et al. as Amici Curiae,

McCarver v. North Carolina,

No. 00-8727 (U.S. June 8, 2001) ..................... 17-18

Brief of Appellant, Smith v. Campbell,

No. 14-10721 (11th Cir. Oct. 20, 2014).................. 3

D. Watson, Intelligence Testing, in The Death Penalty

and Intellectual Disability,

(E. Polloway ed., 2015) ......................................... 11

R. Floyd et al., Theories and Measurement of

Intelligence, in 1 APA Handbook of Intellectual

and Developmental Disabilities

(L. Glidden et al. eds., 2021) ................................ 11

W. J. Schneider, Principles of Assessment of Aptitude

and Achievement, in The Oxford Handbook of

Psychological Assessment of Children and

Adolescents

(D. Saklofske et al. eds., 2013) ...................... 11, 12

REPLY BRIEF

The Court can resolve Smith’s claim with “the fruit

of a multi-century scientific enterprise to ascertain

the best possible estimate of intelligence.” APA.Br.9.

Or it can adopt Smith’s novel, opaque, and unscientific

rule that “where an Atkins claimant has multiple

scores with SEM ranges above and below 70,” “a court

must consider all relevant evidence” of his “actual

functioning” like whether he has “been in a taxi” or

can “make sense of a map.” Br.24-25, 36-37.

The choice is clear. Smith’s brief bobs and weaves,

but he cannot avoid the facts: 75, 74, 72, 78, and 74.

Smith’s five IQ scores together are decisive evidence

that he did not satisfy his burden to prove that his IQ

is 70 or below. There’s a reason none of Smith’s amici

write that he proved his IQ to be 70 or less. He did not.

So Smith tries to change his burden. First, he says

that Alabama doesn’t demand proof of an IQ below 70,

as if Ex parte Perkins, 851 So.2d 453, 456 (Ala. 2002),

does not exist. He identifies no Alabama inmate with

an IQ above 70 who won an Atkins claim.

Second, Smith says that IQ scores are inconclusive

if “the range” reaches 70. But if “the range” starts from

the lowest score minus five, as he suggests, this is just

the lowest-score-wins rule that the Eleventh Circuit

swiftly disavowed after the Court first considered

Smith’s claim. If “the range” accounts for all scores, as

his amici suggest, then Smith loses: He never proved

with any method that the cumulative effect of his

scores yields a range that reaches 70. Incanting

“clinical judgment” cannot transmute his failure into

a viable claim without relieving Smith of his burden,

expanding Atkins, and straying further from the

original meaning of the Eighth Amendment.

2

I. Smith’s claim fails because he did not prove

an IQ of 70 or less using any approach to the

cumulative effect of multiple IQ scores.

A. Alabama law requires proof of an IQ of 70

or below by a preponderance of evidence.

Smith concedes that Ex parte Perkins defined the

first prong for Atkins claims as “significantly subaverage intellectual functioning (an IQ of 70 or below).”

Br.7 The Eleventh Circuit also recognized that the

first prong “turns on” Smith’s IQ. Pet.App.35a (citing

Perkins); see also Hamm v. Smith, 604 U.S. 1, 2 (2024)

(“Smith’s claim … depended in part on whether his IQ

is 70 or below.”). Smith cannot prove his IQ is below

70, so he offers three ways to misread Alabama law,

none of which are constitutionally required (infra §II).

First, Smith claims that “IQs somewhat higher

than 70” can qualify, Br.43 n.5, but “Alabama has not

adopted [that] definition, Albarran v. State, 96 So. 3d

131, 198 n.8 (Ala. Crim. App. 2011) (citing Perkins).

While a single score above 70 with an error range

including 70 cannot defeat a claim under Hall, Smith

still has the “burden to establish that it is more likely

than not that his IQ is 70 or below.” Byrd v. State, 78

So.3d 445, 452 (Ala. Crim. App. 2009); accord, e.g.,

Bush v. State, 92 So.3d 121, 151 (Ala. Crim. App.

2009) (with scores of 74, 75, 69, claim failed under “the

most liberal definition”); Ex parte Jerry Smith, 213

So.3d 214, 225 (Ala. 2003) (counting 72 against claim).

Second, Smith asserts that only “conclusive” IQ

evidence can sink his claim, Br.2, 5, 19, 24, which

sounds a lot like the lower courts’ burden-shifting

framework and not at all like Alabama law. It also

begs the question, for Smith and friends never say

when the IQ evidence is “conclusive.” Perkins does: the

3

claim fails when the offender has not proven that his

IQ is 70 or less. And because that is a “fact[] necessary” for relief, it must be more likely true than false,

Ala. R. Crim. P. 32.3, not a mere “possibility,” Byrd,

78 So.3d at 452; contra Pet.App.6a (requiring that

Smith’s IQ scores “rule out the possibility”); id. at 31a,

37a-40a, 70a (similar).

Third, Smith asserts that Alabama law boils down

“to convinc[ing] the court by a preponderance” that

under today’s clinical criteria, “the best expert” would

say his intellect is significantly subaverage. Br.49-50.

This is also not the law. No doubt experts play a key

role in administering, scoring, and interpreting IQ

tests. But whatever the “best expert” would say about

him, Smith still must prove that his IQ is 70 or less.

See, e.g., Albarran, 96 So.3d at 200.

Smith did not find an Alabama case where an

offender with an IQ above 70 satisfied prong one based

on expert opinion or any other evidence. There is not

an Alabama decision remotely like the one below.1

B. Smith did not prove that his five IQ scores

together yield a likely IQ of 70 or below.

Smith’s 1998 IQ test score of 72 ± 3 is ambiguous,

according to Hall, and does not give us confidence that

1 It’s a bit rich for Smith to claim the mantle of Alabama law

after complaining (successfully) in his first federal appeal that

Alabama set a “bright-line … at 70” and “determined that Smith

could not qualify … because [of] his scores.” App’t.Br.53, Smith

v. Campbell, No. 14-10721 (11th Cir. Oct. 20, 2014). The court

refused to defer precisely because the state courts had held that

Smith’s 75, 74, and 72 “conclusively established” that he “could

never meet Perkins[.]” App.411. The idea that Alabama law

would be more favorable to him on this record—after he scored

78 and 74—is baseless.

4

his true IQ is above or below 70. But with five IQ

scores above 70, the evidence was not ambiguous. The

odds that Smith’s IQ is 70 or below plummeted. For

Smith, the “IQ scores alone” were “conclusive.” Br.2.

Whether and how to consider this “cumulative

effect” is the question before the Court. Smith and the

Eleventh Circuit favor a constitutional rule that

would force courts to evaluate scores in isolation. See,

e.g., Br.34 (“several IQ tests inside and outside the

range” cannot “conclusively resolve the first prong”).

But common sense, state law, the science, and his own

amici favor estimating IQ based on the combined

effect of multiple scores. Smith had the burden to

prove that his scores collectively indicate an IQ of 70

or below. He did not carry that burden, and the Court

need not “move on” to consider other evidence.

1. The cumulative effect of Smith’s five IQ

scores must be considered.

There is uniform support for the principle that

multiple valid, consistent IQ scores above 70 decrease

the risk of error and increase the likelihood that the

test-taker’s true IQ is above 70. Because federal law

permits denying an Atkins claim when the offender’s

true IQ is likely above 70, evidence or inferences about

the cumulative effect must be considered, and the

Eleventh Circuit’s refusal to do so must be reversed.

Experts on both sides below expressed support for

the cumulative-effect principle. The State’s expert Dr.

King testified that “multiple sources of IQ over a long

period of time … contribute[] to … the construct of

validity indicating what a true IQ score is for an individual.” J.A.271. Smith’s expert Dr. Reschly agreed

that error is “much reduced when you have more than

one IQ score.” J.A.105; J.A.104 (agreeing that Smith’s

5

“remarkably consistent” scores “corroborate each

other”).2 And Smith’s expert Dr. Fabian supported the

cumulative effect too. J.A.167 (testifying that the “convergent validity” of Smith’s five “pretty consistent” IQ

scores “trump … one administration”).

Likewise, the APA and the AAIDD emphatically

endorse the cumulative effect. Smith castigated the

idea that multiple scores can “diminish” or “reduce”

error, Br.33, 42, hoping that his amici would agree.

They did not. As the APA explained, having multiple

IQ test scores “allows for a more refined, accurate

assessment,” “providing a more complete and holistic

picture.” APA.Br.16, 17; accord DSM-5-TR at 41. The

AAIDD agreed (at 23): “Multiple IQ test scores … can

provide additional information for an evaluation of

intellectual disability.” Converging scores “can be very

informative.” Id. at 23-24. Smith mistakenly points to

guidance that “each individual IQ score” has an error

range, Br.33, which no one disputes and does not

answer the question presented.

Even the courts below seemed (at times) to admit

the cumulative effect of multiple IQ scores. According

to the Eleventh Circuit, “consistent scores” “reflect a

person’s intellectual ability as opposed to random

chance.” Pet.App.4a; accord Pet.App.5a (this “logic

‘leans in favor of finding that Smith does not [meet

prong one]’” (quoting Pet.App.70a)).

But the panel strayed from sense and science when

it adopted Hall’s vague and unsourced dicta that “the

test” may be “flawed, or administered in a consistently

2 Dr. Reschly did say that “you can’t use poor data to increase

the reliability of good data.” Br.43 n.4. But Smith omits the next

question and answer in which the expert explained that Smith’s

five IQ scores were not “poor data.” J.A.105; see J.A.875.

6

flawed manner” such that even consistent scores are

not determinative. Pet.App.4a (quoting Hall v. Florida, 572 U.S. 701, 714 (2014)).

Parroting that line repeatedly, Br.5, 18, 23, Smith

calls it “Hall’s reasoning (if not its holding)” without

engaging the State’s argument. Compare Br.25 with

Pet.Br.40. First, Hall cannot be read to reject any

“cumulative-score rule” (Br.25-26) because it never

“specified how” to handle multiple scores, Hamm, 604

U.S. at 2. Because Hall reached no holding on the

cumulative effect, the metaphysical possibility of a

“consistent flaw” played no role. Second, Smith did not

argue that his five IQ tests over forty years were

“consistently flawed.” Rather, the unrebutted expert

testimony is that “all of these examiners over all this

period of time giving different tests[] basically [came]

up with the same result.” J.A.271 (State’s expert).

Smith offers nothing to infer that the tests were

consistently biased or skewed against him (rather for

him or not at all). This purely “speculative reasoning”

(Br.5) that multiple scores could be consistently

flawed should play no role.

Smith’s other deflections fail too. No, crediting the

cumulative effect does not mean using “only” the raw

“numerical results” without any “expert testimony.”

Br.42; see also Br.25. Experts can interpret scores and

cast doubt on their reliability. But that’s irrelevant in

this case because Smith’s scores are undisputed. See

Br.12-13, 33; J.A.105. Nor does a cumulative approach

mean “simply tallying up IQ scores” above a threshold.

Br.2. Expert testimony can help assess the cumulative

effect of multiple scores, but again, in this case, Smith

has not identified how any method could prove his IQ

to be 70 or less.

7

Smith also misses the mark when he attacks “a

categorical rule” that “multiple scores above 70” end

the inquiry. Br.25-27. While his scores are decisive,

another inmate with two scores near 70 might bring

evidence that the joint effect is truly indeterminate,

i.e., it yields “a range within which … [his] true IQ

score lies” that reaches 70 or below. Hall, 572 U.S. at

713. But unlike Hall, who proved with the SEM that

a 71 was just as good as a 70, id. at 724, Smith has not

proven that his five scores together “span 70,” Br.25.

Smith’s binary—the notion that he accounts for

the SEM while the State would “erase” it—is bogus.

Br.33. Considering the cumulative effect does not

mean pretending that IQ tests are errorless. The State

agrees (at 41) with the APA (at 21) that “multiple IQ

tests can reduce error” but not “eliminate it.” By denying that basic fact, Br.33, 42, Smith not only misuses

the SEM on its own terms; he has no basis to believe

that his “multiple scores, accounting for the SEM,

span 70.” Br.25. This case is not Hall.

It’s not just “Dr. King’s view” (Br.42) or that of “a

few non-peer-reviewed articles” (Br.33). It’s the “broad

scientific and professional agreement” that obtaining

multiple IQ scores “addresses error and bias in particular test administrations” and produces “greater

accuracy.” APA.Br.3, 5, 17; see Pet.Br.40-41 (citing a

major textbook, an APA publication, other expert

sources); U.S.Br.18-19; CJLF.Br.5-6; AAIDD.Br.24.

The risk of error is never zero, but that nonzero risk

provides no path to conclude that Smith’s true IQ is

likely 70 or below.

8

2. Smith’s “holistic approach” ignores the

cumulative effect of multiple IQ scores.

Instead of grappling with the cumulative effect of

his scores, Smith repeats the word “holistic.” But the

“‘holistic’ rhetoric” is “just window dressing” for a

novel and indefensible change in constitutional law.

Pet.Br.36; see also U.S.Br. 24-25. What Smith means

(and needs to prevail) is a rule that “if the range of

[IQ] scores … reaches 70,” then “courts must consider

other evidence.” Br.18-19; see also Br.25 (“[W]here

multiple scores, accounting for the SEM, span 70, a

court must consider … actual functioning,” rather

than IQ.); BIO.i, 1, 20.

While Smith tries to run from a lowest-score rule,

Br.29-30, 47-48, his brief illustrates how he thinks

“the range” is constructed by simply reducing the

lowest score by five. For instance, IQ was inconclusive

in Moore, Smith says, because the “scores of 74 and

78” made “a range from 69 to 83.” Br.5. He reads

Reeves the same way, asserting that the range for

scores of 68, 71, and 73 went “as low as 63” to “as high

as 78.” Br.43; see Reeves v. State, 226 So.3d 711 (Ala.

Crim. App. 2016). It’s no mischaracterization (contra

Br.47) to call this a “‘per se rule’ that a single score

[range] dipping below 70 is ‘dispositive’” of IQ. 3

Smith’s answer to “whether” courts may consider

the cumulative effect of multiple IQ scores is “never.”

The lowest score range “reaches 70,” or it does not.

3 Smith’s brief is opaque, but he may intend to advance a rule

that requires “multiple scores with SEM ranges … below 70,”

Br.24 (emphasis added). Of course, that would just transform the

lowest-score rule into a two-lowest-scores rule. Both rules are

insensitive to the cumulative effect.

9

Br.18. Needless to say, this is not the “complicated

endeavor.” Hamm, 604 U.S. at 2.

Yet it’s also how the Eleventh Circuit analyzed IQ.

The panel thought that Smith “needed to prove only”

“one valid IQ test score” with a range that includes 70.

Pet.App.42a, 44a.4 So even after vacatur, the court

emphasized “that Smith had an IQ test score as low

as 72, which, according to expert testimony, meant his

true IQ could be as low as 69.” Pet.App.5a. But that

testimony was about a single score, see J.A.364, not

the cumulative effect of multiple. The courts thus refused to credit the cumulative effect because they

thought it would mean “throw[ing] out as an outlier”

“the lowest score.” Pet.App.6a. That remark, which

Smith endorses, is very confused (no score is “thrown

out” by the cumulative approach), and it reveals the

special weight given to the lowest end of the lowest

score’s error range.

Nor does the panel’s observation on remand that

“four out of Smith’s five IQ scores” are below 76 evince

a cumulative analysis. Br.42. Just the opposite. Fixing

the line at 75 assumes a universal error range of ± 5

points, which is an express rejection of any cumulative

effect to narrow the risk of error. There’s nothing

“holistic” about any of this. See Pet.Br.35-36, 38-41.

Another way to see that Smith and the Eleventh

Circuit reject the cumulative approach is to consider a

hypothetical inmate with ten or twenty or any number

of scores of 75. On any method of aggregating scores,

Here, the court’s vacated (but holistic) opinion relied on

United States v. Wilson, 170 F. Supp. 3d 347 (E.D.N.Y. 2016), in

rejecting the “assertion that a district court can consider anything other than the lower end of [the] standard-error range.”

Pet.App.42a. Tellingly, Smith now disavows Wilson. Br.29-30.

4

10

the odds of an IQ of 70 or less would be de minimis.

Yet on Smith’s rule, when one or two scores of 75 ± 5

“span 70,” the court must consider non-IQ evidence.

Likewise, on the Eleventh Circuit’s rule, every 75

would be “consistent with mild intellectual disability”

and count in favor of the inmate. Pet.App.7a.

Far from “[t]ying all of the scores together,” Br.13,

or considering “the relationship of the scores to one

another,” Br.32, Smith and the Eleventh Circuit

would have courts evaluate each score “individually”

and then “move on.” Pet.App.4a, 6a.5

3. Smith did not prove that his IQ is 70 or

below using any method for jointly

analyzing his scores.

Nowhere does Smith’s brief explain how his five IQ

scores combine to prove an IQ of 70 or below. That

should be the end of the case, for Atkins protects those

offenders “known to have an IQ below 70,” Kennedy,

v. Louisiana, 554 U.S. 407, 425 (2008). At most,

Smith’s response implies this is a case like Hall where

IQ is “inconclusive,” but he never proved that.

Considered jointly, the five scores vanquish the

likelihood that Smith qualifies for Atkins relief. One

simple rule is the median, which is endorsed by two

AAIDD and APA publications cited in both briefs. The

5 Even then, Smith’s claim should fail, for he never assessed

“each separate score … using the SEM” with test-specific ranges.

Br.26. His experts did not “provide[] information” about the

range for his WISC-R scores. Contra Br.40 n.3. If he means one

can assume an error range of ± 5, that “contradicts the test’s own

design,” Hall, 572 U.S. at 724, and the record. See Pet.Br.16-18,

36-38 & nn.26-27.

11

APA Handbook views the median as “Option 1.”6 And

an AAIDD chapter endorses it as the “best practice”

when the inputs for a composite score are not

available.7 Smith’s median score is 74.

A similar method focuses on the overlap among

each score’s error range, which clinicians “often” use

to estimate IQ. AAIDD Br.24 (citing Watson, supra at

124). Even assuming a range of ± 5, the “overlap” for

Smith’s scores is a range of 73 to 77. This is

“compelling evidence that the best estimate lies

within the convergent range.” APA.Br.20.8

More complicated is the composite-score method

championed by Dr. Schneider. See APA.Br.18; AAIDD

Br.25; U.S.Br.19. The composite score treats each

score as a subtest of a “Mega-IQ Test.”9 But it requires

data that is hard to find and “simply may not exist,”

AAIDD.Br.25 n.17; accord CJLF.Br.7-9. Smith’s amici

support the method, but tellingly, none would crunch

the numbers to show how it might help Smith’s claim.

It would not. See CJLF.Br.10.

Although taking the average is “not quite right,” it

still gives a “rough approximation” of the composite

6 R. Floyd et al., Theories and Measurement of Intelligence, in

1 APA Handbook of Intellectual and Developmental Disabilities

415 (L. Glidden et al. eds., 2021).

7 D. Watson, Intelligence Testing, in The Death Penalty and

Intellectual Disability 124 (E. Polloway ed., 2015).

8 AAIDD seems to undermine its support for the overlap by

suggesting (at 26) that “the ‘true’ score is beyond the grouping.”

If it means to imply that Smith’s IQ is below 72, it stretches its

lone citation (to an example using two 70s) well beyond its moorings. Cf. CJLF.Br.13-14.

9 W. J. Schneider, Principles of Assessment of Aptitude and

Achievement, in The Oxford Handbook of Psychological

Assessment of Children and Adolescents 290 (D. Saklofske et al.

eds., 2013).

12

score. Schneider, supra at 290; see also CJLF.Br.6.

Smith’s average IQ score is 74.6.

*

*

*

The State does not advocate “a single, mechanical

rule … for aggregating multiple IQ scores.” Br.34. The

State does advocate that Smith must “demonstrate—

one way or another—that his collective IQ test scores

betray deficient intellectual functioning.” Id. Though

it is not the State’s burden to prove anything with

respect to Smith’s IQ, the foregoing demonstrates that

it is not even theoretically possible that the Eleventh

Circuit held Smith to his constitutional burden. The

only conclusion is the one drawn by the State’s expert

that Smith’s many “sources of IQ” together suggest a

“true IQ score” well above 70. J.A.271. There is

certainly no “contrary evidence” (Br.47) that rebuts

what Smith’s scores plainly prove.

II. The Court should not expand Atkins with

Smith’s rule denying the cumulative effect of

multiple IQ scores.

A. Smith concedes that the constitutional

text and history are independent reasons

to reject his position.

Smith’s new rule egregiously departs from the

original meaning of the Eighth Amendment, which

was about punishments that intentionally “superadd

terror, pain, or disgrace” and had “fallen out of use.”

City of Grants Pass v. Johnson, 603 U.S. 520, 542

(2024) (citation modified); see Bucklew v. Precythe, 587

U.S. 119, 130-31 (2019); Pet.Br.3, 29-32.

In response to the State’s argument that no

method for assessing multiple IQ scores could be

“cruel” as an original matter, Smith is mum. He offers

13

“nothing in the Eighth Amendment” that would give

the Court “lawful authority to extend” Atkins and its

progeny. City of Grants Pass, 603 U.S. at 550. Smith

cites a single dissenting opinion not even endorsing

the “state consensus” approach, Br.26, and he ignores

that the Constitution’s “text, history, meaning, and

purpose,” Kennedy, 554 U.S. at 421, are “independent”

“reason[s] to disagree” with his position regardless of

the evolving standards, cf. Moore v. Texas, 581 U.S. 1,

27 (2017) (Roberts, C.J., dissenting) (quoting Atkins v.

Virginia, 536 U.S. 304, 321 (2002)); see also id. at 22;

Pet.Br.29-32; Idaho.Br.4-17; U.S.Br.11-12, 26-32.

There is no contention that Smith’s “sentence constitutes one of ‘those modes or acts of punishment that

had been considered cruel and unusual at the time

that the Bill of Rights was adopted.’” Stanford v. Kentucky, 492 U.S. 361, 368 (1989), abrogated by Roper v.

Simmons, 543 U.S. 551 (2005). While the common law

exempted those with an “IQ of 25 or below” from all

punishments, Penry v. Lynaugh, 492 U.S. 302, 333

(1989), the rule that the mildly intellectually disabled

“should be tried and punished” but not capitally is a

20th-century innovation, Atkins, 536 U.S. at 306; see

id. at 340-41 (Scalia, J., dissenting). The Constitution

as interpreted by this Court even permitted their execution until some time between Penry in 1989 and

Atkins in 2002. Thus, the original meaning cannot

support Smith’s test; at most, it supports exemptions

for those who lack the capacity “to form criminal intent or to understand the difference between good and

evil.” Penry, 492 U.S. at 333; see also Atkins, 536 U.S.

at 340-41 (Scalia, J., dissenting). Smith is nowhere

near that line.

14

The closest Smith comes to an Eighth Amendment

argument is his assertion that any “fact must be

determined with the high regard for truth that befits

a decision affecting … life or death.” Br.24. But that

dictum from Ford v. Wainwright was about the test

for granting a hearing, 477 U.S. 399, 410-12 (1986), it

comes with no support in text or history, and it begs

the question whether Smith’s non-cumulative rule

better seeks “truth.” It does not. Supra §I.B.1.

B. No nationwide consensus mandates a rule

denying the cumulative effect of IQ scores.

Smith’s claim requires the Court to expand the

immunity from capital punishment to include inmates

whose IQ scores do not prove a likely IQ below 70.

No nationwide consensus supports that expansion.

Smith does not dispute that the vast majority of state

laws dealing with IQ specify 70. See Pet.Br.12 n.7.

Nor does Smith identify state laws that focus on

the lowest score or reject the cumulative effect of

scores. If no consensus supports any “categorical rule,”

Br.27, then no consensus supports a categorical rule

that “where an Atkins claimant has multiple scores

with SEM ranges above and below 70,” “a court must

consider … actual functioning,” Br.24-25.

A few state courts have misapplied Hall, even

granting relief to inmates who did not seem to prove

an IQ of 70 or below, Pet.Br.14 n.10; Ky.Br.6-14, but

their errors do not have nationwide support, nor are

they the “most reliable objective evidence of contemporary values,” which is “legislation.” Penry, 492 U.S.

at 331. Yes, state decisional law is state law, Br.27-28,

but it reflects public morality only to the extent that

courts interpret and apply the will of the people—not

the meaning of Hall’s dicta. Cf. Jones v. Mississippi,

15

593 U.S. 98, 127 (2021) (Thomas, J., concurring in

judgment) (States “have spent years chasing the everevolving definitions of mental incompetence promulgated by this Court and its preferred experts.”). That

state courts take claims “case-by-case,” Br.27, sheds

no light on what the American people consider “cruel.”

There is a world of difference between what Smith

brings and what Hall proved: that “an individual with

an IQ score of 71” would not lose due to his 71 in most

States. 572 U.S. at 716. In contrast, Smith does not

prove by a consensus that an inmate like him—with

scores that together conclusively prove his IQ is above

70—would not lose for that reason in most States. At

minimum, Smith needed to prove assent to his rule

that failure to meet his IQ burden is excused by facts

like his facility with “subway lines,” Br.37. He did not.

C. No precedent requires eroding prong one

with a rule denying the cumulative effect

of IQ scores.

Alabama makes IQ the touchstone of prong one

because IQ is the “best single representation of intellectual functioning,” AAIDD.Br.15, and “the best

possible estimate of intelligence,” APA.Br.9. This is

one of many “appropriate ways to enforce” Atkins, 536

U.S. at 321, and “nothing in the Eighth Amendment

gives federal judges the authority or guidance they

need” to demand a different approach, City of Grants

Pass, 603 U.S. at 559 n.8.

1. The Court is free to reject Smith’s invitation to

diminish the role of IQ. Hall did not reach the

“complicated endeavor” because Florida barred relief

for “an individual with an IQ score of 71” regardless of

other scores. 572 U.S. at 714, 716. That “strict IQ test

score cutoff” was “the issue.” Id. at 712. Smith mostly

16

ignores Hall’s limited holding but happens to get it

right (at 22) when he states that Hall applies to “a

given IQ score.” Extrapolating wildly, he asserts that

all IQ evidence amounts to “a single factor” that can

never be “dispositive” or “bar” other evidence. Id.

That rule cannot be derived from Hall. Again,

we’re here because Hall did “not specif[y] how” to treat

multiple IQ scores, yet Smith’s claim “depend[s]” on

multiple IQ scores. Hamm, 604 U.S. at 2. Reading

Hall to make IQ non-dispositive conflicts with both of

those holdings. Second, Hall’s recognition that some

IQ evidence is inconclusive implies that other IQ

evidence can be conclusive. For example, Hall

declined to address “a bright-line cutoff at 75 or

greater,” 572 U.S. at 715, which the parties agreed

would be perfectly constitutional, see Oral.Arg.Tr.9,

13-14, 25. If Smith’s rule were Hall’s rule, Hall would

have been an easy case, and so would a “cutoff at 75.”

Smith makes no attempt to square his rule with

Brumfield v. Cain, the Court’s only decision in which

the presence or absence of multiple scores made a

difference. 576 U.S. 305 (2015). There, it was crucial

that the inmate’s second IQ test was not “sufficiently

rigorous”; if it had been, it could have “preclude[d]

definitively” any “possibility” of relief. Id. at 316.

Brumfield did not answer the question presented, but

it did take Smith’s answer—that IQ alone cannot preclude a claim, see, e.g., Br.22-23—off the table.

2. Smith repeats ad nauseum that intellectual

disability is “not a number,” id., but prong one “has

always required an IQ score” because IQ is the “single

best” measure of intellect, AAIDD.Br.10 n.4, 15. The

choice is not between “truth” and falsehood, Br.24, but

two ways to approach the same body of evidence. On

17

the one hand, Smith proposes that the Constitution

requires courts to “move on” from IQ if the lowest

score is ambiguous. Rather than estimating IQ, courts

would focus on facts like whether he has insurance

(BIO.13) or skill at a childhood game (AAIDD Br.28).

On the other, the State has reasonably allowed the

“legal determination of intellectual disability,” which

“is distinct from a medical diagnosis,” Hall, 572 U.S.

at 721, to turn on the very best evidence of intelligence.

That yields a consistency that avoids “inequities” in

criminal justice, id. at 737 (Alito, J., dissenting), and

preserves meaningful appellate review.

For Alabama’s discretion under Atkins to matter,

it must be permitted to identify signal in the noise.

Hall characterizes the relationship between prongs

one and two as “conjunctive and interrelated.” 572

U.S. at 723. But the AAIDD says they are “distinct

and separate.” AAIDD-12 at 33. Smith labels the two

prongs “independent but relevant to one another.”

Br.35 (capitalization altered). The APA says they are

“inherently interrelated.” APA.Br.7. States need clarity, not doublespeak.

Smith puts much weight (at 4, 15, 25, 34, 41) on

the DSM’s “example” that a person “whose IQ score is

somewhat above 65–75 may nevertheless have such

substantial adaptive behavior problems … that the

person’s actual functioning is clinically comparable to

that of individuals with a lower IQ score.” DSM-5-TR

at 42. This is not the law, and its adoption would

greatly expand Atkins, illustrating one of many

reasons that States do not need to adhere to the latest

medical guide. Cf. Pet.Br.31 & n.21; Cert.Pet.14 n.4;

contra Br.38 (assuring that the definition will “always

remain a consistent proportion of the population”).

18

Whether Smith casts his rule as weighing IQ

against prong two (as if Atkins were a balancing test),

or abandoning IQ in favor of “actual functioning,” the

result is the same: an amorphous and subjective inquiry that abdicates enforcement of the first prong.

Every murderer on death row can allege maladaptive

behavior. AFLF Br.27-28. Every inmate would get a

hearing and a “battle[] of experts,” exactly what the

APA promised would “not result” from Atkins. See Br.

of APA et al. as Amici Curiae at 16, McCarver v. North

Carolina, No. 00-8727 (U.S. June 8, 2001).

A “close case” on IQ is not made easier by injecting

adaptive deficits, Br.15, which are “only moderately

correlated” with intelligence. AAIDD-12 at 33. Those

with an IQ of “75 to 85 often function on a daily basis

similarly to or, sometimes, lower than persons” in the

“range of 55 to 75.” J.A.866. Adaptive deficits can

indicate many “conditions or mental disorders,” DSM5-TR at 42, like ADHD, autism, anxiety, and

depression. That’s why the prongs are separate.

Adaptative deficits should be “confirmed” by IQ tests,

DSM-5-TR at 37, not the other way around.

And it’s easy to see why in practice. For offenders

like Smith, prong two means asking his mother and

childhood friends how well he socialized, cooked food,

or bought groceries decades ago. Pet.App.87a. It

means asking a death-row prisoner about his “money

management” and “transportation.” Pet.App.88a. He

could “make barbecue,” “soup,” and “fried chicken,”

but courts are supposed to discount his 78 IQ score

because he “did not use a bus” or take “a taxi.” J.A.801.

Perhaps the alchemy of clinical judgment can turn

these tidbits into a diagnosis, but the Constitution

does not obligate States to prioritize that evidence

19

over the cumulative effect of IQ. In truth, these are

“other sources of imprecision” that cannot “narrow”

the range of functioning disclosed by IQ testing.

Moore, 581 U.S. at 14. As Dr. Reschly put it, “you can’t

use poor data to increase the reliability of good data.”

J.A.105. To assess intellectual functioning, “individualized, standardized intelligence testing” is the good

data. DSM-5-TR at 37. It’s also the data “that most

directly relates to the concerns” that motivated Atkins

in the first place. Hall, 572 U.S. at 737 (Alito, J., dissenting); accord Atkins, 536 U.S. at 320.

III.

Smith’s IQ score of 78 precludes relief.

Smith’s claim fails independently because he

obtained a valid IQ score of 78. Assuming the most

favorable error range, Smith’s 78 yields an interval of

73 to 82 “within which one may say [his] true IQ score

lies” with “95% confidence.” Hall, 572 U.S. at 713.

When the entire error range lies above the state-law

criterion for deficient intellect, the Constitution does

not bar capital punishment.

A. Smith addresses this argument in one sentence,

asserting that it would not “comport” with precedent,

science, or nationwide practice to deny Smith’s claim

based on his highest score. Br.48. But Hall conceded

that a high-score rule would be constitutional, and the

Court did not identify such a rule as forbidden by national consensus. 572 U.S. at 715. Then in Brumfield,

which Smith ignores, the Court contemplated that a

score above 75 would “preclude definitively the possibility” of disability. 576 U.S. at 316. That makes sense.

If taking confidence intervals seriously means that an

inmate whose sole score range dips to 70 can proceed,

it should also mean that an inmate with a valid score

range entirely above 70 cannot.

20

Smith, whose entire theory rests on the low end of

his error range, supra §I.B.2, should not be heard to

complain that it is unscientific to focus on one end or

the other. And focusing on the high end has a scientific

warrant: There is no way for a test-taker to fake being

smarter than he is. Pet.Br.14-15 & n.11. By contrast,

there are many ways for a test to underestimate IQ. A

prisoner’s score may be artificially deflated by his poor

effort, stress, distraction, fatigue, health conditions, or

the obvious incentive to underperform in a capital

case. But Smith ignores the obvious.

If Smith’s 78 were “tainted” by error or bias, the

district court could have discarded it as “[in]valid and

[un]reliable.” APA.Br.22. An “outlier.” APA.Br.20. But

there is no claim that Smith’s expert erred, and no one

contested the score’s validity. To the contrary, Smith’s

other scores bolster the 78 because all the ranges (at

95% confidence) “overlap[].” AAIDD.Br.24.

B. A “simple and static” test would be best. Jones,

593 U.S. at 127 (Thomas, J., concurring in judgment);

cf. Roper, 543 U.S. at 594 (O’Connor, J., dissenting)

(“especially desirable in this sphere”). Atkins itself

was prophylactic, protecting not only those who do not

deserve their sentences but anyone “less likely” to be

culpable. 536 U.S. at 320; Hall, 572 U.S. at 709. Any

later rule-making is a proxy on top of a proxy for the

ultimate issue of moral desert. No line can be perfect.

Nonetheless, a “line must be drawn,” Roper, 543

U.S. at 574, and the Court has already rejected the

contention that it needs to be a “medical diagnosis”

using “clinical judgment,” Hall, 572 U.S. at 721, 723.

Legal and medical judgments are “distinct,” id. at 721,

as they must be. Judges are not psychiatrists versed

in the “uncertainties about the human mind,” and

21

even psychiatrists “disagree widely and frequently.”

Kahler v. Kansas, 589 U.S. 271, 280 (2020). Rather

than mimicking clinical judgment when a valid IQ

score shows that the inmate is highly unlikely to be

disabled, and rather than using a diagnosis as a rough

proxy for blameworthiness, there is a better and more

familiar approach: Let the sentencer express a “reasoned moral response to the defendant’s background,

character, and crime,” including his intelligence, after

an individualized sentencing determination. Penry,

492 U.S. at 328. That’s what the doctrine provides for

a defendant who cannot meet prong three, for example, or one who murders on his eighteenth birthday.

And it’s what was provided for Smith in 1998.

If state legislatures wish to provide immunity for

offenders with scores above 75, they are free to do so.

But the judgment below vacated Smith’s sentence

based on the Eighth Amendment. If the Eighth

Amendment does not immunize an offender with a

valid IQ score of 78, then Smith’s Eighth Amendment

claim fails. That is true regardless of how Alabama

courts “implement” the “principles of Atkins.” Hall,

575 U.S. at 709-10. Contra Br.30, 48.

CONCLUSION

The Court should reverse.

22

Respectfully submitted,

Steve Marshall

Attorney General

Edmund G. LaCour Jr.

Solicitor General

Counsel of Record

Robert M. Overing

Deputy Solicitor General

Dylan Mauldin

Assistant Solicitor General

Benjamin M. Seiss

Assistant Attorney General

Office of the Att’y Gen.

501 Washington Avenue

Montgomery, AL 36130

(334) 242-7300

Edmund.LaCour@

AlabamaAG.gov

Counsel for Petitioner

October 17, 2025