Armstrong, Douglas Tyrone

IN THE COURT OF CRIMINAL APPEALS
OF TEXAS
NOS. WR-78,106-01

EX PARTE DOUGLAS TYRONE ARMSTRONG, Applicant

ON APPLICATION FOR A WRIT OF HABEAS CORPUS
FROM HIDALGO COUNTY

Y EARY, J., delivered the opinion of the court in which A LCALA, R ICHARDSON,
K EEL and W ALKER, JJ., joined. K ELLER, P.J., filed a dissenting opinion in which
K EASLER, J., joined. N EWELL, J., filed a dissenting opinion in which H ERVEY, J., joined.

OPINION

Applicant has filed four applications under Section 11.071 of the Texas Code of

Criminal Procedure, our statutory vehicle for post-conviction relief when a jury imposes the

death penalty. T EX. C ODE C RIM. P ROC. § 11.071. We filed and set for submission the initial

application, and held the subsequent applications pending our disposition of the initial

application. Among his other claims, Applicant contends in his initial application that his trial

counsel were constitutionally ineffective.1 More specifically, he alleges that his trial counsel

1
The State argues that Applicant’s ineffective assistance claim is not cognizable on habeas
under Ex parte Nailor, 149 S.W.3d 125, 131-32 (Tex. Crim. App. 2004), because we found on
direct appeal that the appellate record was developed at the hearing on a motion for new trial and we
addressed the claim on the merits. It is true that Appellant also raised this ground in a motion for new
 ARMSTRONG — 2

failed to conduct a constitutionally adequate investigation of mitigating evidence that could

have been introduced during the punishment phase of his trial. The application contained

affidavits from several witnesses who averred they could have told the jury about various

mitigating circumstances had they been asked to testify at Applicant’s trial. Applicant also

proffered reports from two psychological experts with respect to his mental health at the time

of the offense.

In a previously issued order, this Court already concluded that counsel’s failure to

investigate was constitutionally deficient, and noted that the affidavits attached to the

application contained evidence “similar to evidence that we have found to have mitigating

value.” Ex parte Armstrong, No. WR-78,106-01, 2015 WL 7354084, at *5 (Tex. Crim. App.

Nov. 18, 2015) (not designated for publication). We remanded the application to the

convicting court for findings of fact on the credibility and availability of the witnesses, both

lay and expert, and “any other findings of fact and conclusions of law that it deems relevant

and appropriate to the disposition of [A]pplicant's claim for habeas corpus relief.” Id. On

remand, the convicting court concluded that Applicant was not prejudiced by his defense

trial and on direct appeal. We rejected the argument on direct appeal because, though the record at
that time established that defense counsel’s mitigation investigation was somewhat cursory, it did
not establish what a more thorough mitigation investigation would have found. Armstrong v. State,
No. AP 75,706, 2010 WL 359020, at *6-7 (Tex. Crim. App. Jan. 27, 2010). Habeas counsel has
since conducted a more thorough mitigation investigation, and attached the fruits of that
investigation to the writ application. This case thus fits the Nailor exception for new evidence. See
149 S.W.3d at 131-32. We reach the merits of ineffective assistance claims where new relevant
evidence is presented on habeas, regardless of whether the merits were decided on appeal. E.g., Ex
parte Bryant, 448 S.W.3d 29, 34-35 (Tex. Crim. App. 2014).
 ARMSTRONG — 3

team’s failure to investigate. We hold that Applicant was prejudiced by a constitutionally

inadequate mitigation investigation. We will vacate Applicant’s death sentence, and remand

for a new punishment proceeding.

I. THE APPLICABLE LAW

“In all criminal prosecutions, the accused shall enjoy the right to have Assistance of

Counsel for his defence.” U.S. C ONST. amend. VI. “The right to counsel is the right to the

effective assistance of counsel.” McMann v. Richardson, 397 U.S. 759, 771 n.14 (1970). A

defendant’s own lawyer can “deprive the defendant of the right to effective assistance by

failing to render ‘adequate legal assistance.’” Strickland v. Washington, 466 U.S. 668, 686

(1984) (citing Cuyler v. Sullivan, 446 U.S. 335, 344 (1980)). “A convicted defendant’s claim

that counsel’s assistance was so defective as to require reversal of a . . . death sentence has

two components”: (1) deficient performance; and (2) prejudice. Id. at 687. Trial counsel may

provide constitutionally ineffective assistance if they fail to conduct an adequate

investigation of possible mitigating evidence to present at the punishment phase of a capital

murder trial. Wiggins v. Smith, 539 U.S. 510, 524 (2003). We have already determined that

Applicant’s counsel performed deficiently at the punishment phase of trial in this case by

failing to conduct an adequate mitigation investigation. Armstrong, 2015 WL 7354084, at

*5. Applicant is entitled to a new punishment hearing if he can show he was prejudiced.

To establish prejudice, a “defendant need not show that counsel’s deficient conduct

more likely than not altered the outcome of the case.” Strickland, 466 U.S. at 693. Prejudice
 ARMSTRONG — 4

is determined by whether there is a “reasonable probability that, but for counsel’s

unprofessional errors, the result of the proceeding would have been different.” Id. at 694. A

“reasonable probability” is simply “a probability sufficient to undermine confidence in the

outcome.” Id. In the context of a deficient mitigation investigation, Applicant need not

demonstrate that it is “more likely than not” that the jury would have assessed a life sentence;

instead, the proper question is whether there is a “reasonable probability” the jury would not

have sentenced Applicant to death if the post-conviction mitigation evidence had been

presented at trial. Id. at 694-95; see also Woodford v. Visciotti, 537 U.S. 19, 22 (2002)

(noting that Strickland “specifically rejected the proposition that the defendant had to prove

it more likely than not that the outcome would have been altered”). To answer that question,

we must consider “the totality of the evidence—‘both that adduced at trial, and the evidence

adduced in the habeas proceedings.’” Wiggins, 539 U.S. at 536 (quoting Williams v. Taylor,

529 U.S. 362, 397-98 (2000)). If, after introducing Applicant’s post-conviction mitigating

evidence into the punishment-phase evidentiary mix, we conclude that there is a reasonable

probability that at least one juror would have answered the mitigation special issue in his

favor, then he is entitled to relief. Id. at 537.2

2
Under Article 37.071, Section 2(e)(1), a jury that has found a capital murder defendant to
be eligible for the death penalty under Section 2(b) shall then answer the so-called mitigation special
issue:

Whether, taking into consideration all of the evidence, including the
circumstances of the offense, the defendant’s character and background, and the
personal moral culpability of the defendant, there is a sufficient mitigating
circumstance or circumstances to warrant that a sentence of life imprisonment
 ARMSTRONG — 5

This is not a case in which trial counsel failed altogether to present mitigating

evidence at trial. Applicant’s sister gave general testimony before the jury about Applicant’s

disadvantaged childhood. Applicant has now presented additional mitigating evidence that,

broadly speaking, consists of two types: (1) greater detail about his disadvantaged

background; and (2) entirely new expert testimony with respect to his mental state at the time

of the offense.

Twice before, this Court has addressed claims that trial counsel were ineffective for

conducting an investigation that failed to uncover all of the mitigating evidence available.

In Ex parte Martinez, 195 S.W.3d 713 (Tex. Crim. App. 2006), trial counsel presented

evidence at the punishment phase of a capital murder trial that the defendant had suffered

abandonment and neglect, and had been physically abused, as a child. Id. at 725. In post-

conviction proceedings, he was able to add “more extensive evidence” of physical abuse,

along with evidence that his own mother had also abused him sexually, to the evidentiary

mix. Id. at 725-26, 730. Although we characterized this added mitigation as “strong,” we

nevertheless reasoned that, because the aggravating evidence in the case was “severe,” and

“the jury was privy to some of the severe abuse [Martinez] suffered during his childhood,

there is not a reasonable probability that the unadmitted alleged mitigating evidence would

have tipped the scale in [his] favor.” Id. at 731. In so holding, we cited (among other cases)

without parole rather than a death sentence be imposed.

TEX . CODE CRIM . PROC. art. 37.071, § 2(e)(1).
 ARMSTRONG — 6

Hill v Mitchell, 400 F.3d 308, 319 (6th Cir. 2005), which supports the proposition that, in

order to satisfy the prejudice prong of Strickland/Wiggins, mitigating evidence presented at

the post-conviction stage “must differ in a substantial way—in strength and subject

matter—from the evidence actually presented at sentencing[.]” Martinez, 195 S.W.3d at 731;

see also Goodwin v. Johnson, 632 F.3d 301, 327 (6th Cir. 2011) (“To show prejudice

resulting from counsel’s deficient mitigation investigation, the petitioner must present new

evidence that differs both in strength and subject matter from the evidence actually presented

at sentencing, not just cumulative mitigation evidence.”).3

3
In neither Wiggins itself, nor in any subsequent case, has the United States Supreme Court
itself granted a new capital punishment hearing when the applicant developed new mitigating
evidence that was essentially just more of the same character of mitigation evidence that was
presented at trial. Compare Bobby v. Van Hook, 558 U.S. 4, 12 (2009) (new evidence that the
defendant’s mother was temporarily committed to a psychiatric hospital and that his father hit him
frequently and tried to kill his mother “would have added nothing of value” to the mother’s own trial
testimony that she had been under psychiatric care and that the father had been abusive), and Cullen
v. Pinholster, 563 U.S. 170, 200-01 (2011) (“The ‘new’ evidence largely duplicated the mitigation
evidence at trial. School and medical records basically substantiate the [trial] testimony of
Pinholster’s mother and brother.”), with Wiggins, 539 U.S. at 537 (“Wiggins’ sentencing jury heard
only one significant mitigating factor—that Wiggins had no prior convictions. Had the jury been able
to place [his] excruciating life history on the mitigating side of the scale, there is a reasonable
probability that at least one juror would have struck a different balance.”), Porter v. McCollum, 558
U.S. 30, 41 (2009) (“The judge and jury . . . heard about Porter’s turbulent relationship with [the
victim], his crimes, and almost nothing else.”), Rompilla v. Beard, 545 U.S. 374, 390 (2005) (“The
prison files pictured Rompilla’s childhood and mental health very differently from anything defense
counsel had seen or heard.”), and Sears v. Upton, 561 U.S. 945, 947-48 (2010) (“During the penalty
phase. . . counsel presented evidence describing his childhood as stable, loving, and essentially
without incident. . . . The mitigation evidence that emerged . . . postconviction . . . however,
demonstrates that Sears was far from privileged[.]”). But see Neal v. Puckett, 286 F.3d 230, 244 (5th
Cir. 2002) (in response to the State’s argument that new mitigating evidence was merely cumulative
of what was presented at trial, observing that, “with a more detailed and graphic description and a
fuller understanding of [petitioner’s] pathetic life, a reasonable juror may have become convinced
of [his] reduced moral culpability”).
 ARMSTRONG — 7

In Ex parte Gonzales, 204 S.W.3d 391 (Tex. Crim. App. 2006), decided only a few

months after Martinez, we concluded that the applicant did establish prejudice. At trial,

Gonzales produced some mitigating evidence in the form of testimony from his sister that

he had been bullied in school, had experienced difficulty learning, was of only borderline

intelligence, and suffered from epilepsy, attention deficit disorder, and depression. Id. at 395,

398. In post-conviction proceedings, Gonzales was able to demonstrate that additional

mitigating evidence had been available to show that he had suffered chronic physical and

sexual abuse at the hands of his father. Id. at 395, 399. Psychiatric evidence also established

that he suffered from Post-Traumatic Stress Disorder as a result of this abuse. Id. at 399.

Moreover, the psychiatrist also could have testified that Gonzales’s low intelligence “would

lead to poor processing of information and probably lower level of control of behaviors

which included antisocial behaviors and impulsive behaviors at an early age.” Id. We found

that “the mitigating evidence presented at the habeas hearing [was] substantially greater and

more compelling than that actually presented by the applicant at his trial.” Id. We concluded

that there was “at least” a reasonable probability of a different result had this additional

evidence been presented at trial. Id. at 399-400.

Trial counsel’s presentation of some mitigating evidence at the punishment phase does

not “foreclose an inquiry into whether a facially deficient mitigation investigation might have

prejudiced the defendant.” Sears v. Upton, 561 U.S. 945, 955 (2010). On the other hand,

when trial counsel has presented some mitigating evidence, the failure to conduct an
 ARMSTRONG — 8

adequate mitigation investigation will not prejudice the defendant if “the new mitigating

evidence ‘would barely have altered the sentencing profile presented’ to the

decisionmaker[.]” Id. at 954 (quoting Strickland, 466 U.S. at 700).

In the instant case, Applicant presented additional lay testimony to bolster the

mitigating evidence that his counsel produced at trial: his sister’s general account of his

disadvantaged childhood. Whether the new evidence from additional lay witnesses was

sufficiently different in strength and subject matter that we would conclude there is a

reasonable probability that at least one juror would have answered the mitigation special

issue “yes,” however, is a question we need not ultimately resolve in this case. Here, as in

Gonzales, Applicant has also presented mental health testimony that was not presented at the

punishment phase of his trial. We conclude that the addition of the expert testimony about

Applicant’s low intelligence and the various psychological maladies from which he suffered

at the time he committed this offense creates a reasonable probability of a different outcome

had it been presented at trial.

II. THE EVIDENCE PRESENTED AT TRIAL

A. The Guilt Phase Evidence

On the afternoon and evening of April 21, 2006, Applicant was drinking and buying

rounds of drinks for women in the Sunshine Bar in Donna, Texas. At some point, he ran out

of money and left the bar. A short time later, Raphael Castellan, the victim, was murdered

outside of his home a few blocks away. Two witnesses saw the murder from their minivan
 ARMSTRONG — 9

as they were leaving their home to go grocery shopping. They testified that they saw a large

black male attacking Castellan, who was smaller in stature. They drove toward the attacker

and honked the horn to stop the attack, and saw the attacker pulling Castellan back so that

he could not flee to their van. At some point, the attacker threw Castellan against the ground

with such force that they heard a thump as his head hit the ground. The attacker then slashed

Castellan repeatedly with a fold-out knife/box cutter, and went through his pockets.

Applicant was found a short time later in the bathroom back at the Sunshine Bar, borrowing

a shirt from someone else. Applicant’s bloody shirt was discarded on the floor by the urinal,

and his pants were covered in blood. One of the eyewitnesses from the van later identified

Applicant as the large black male she had seen attacking Castellan.

The jury saw photos of an especially bloody crime scene, along with autopsy photos

of clean incisions to Castellan’s face, neck, and hands. The medical examiner testified that

incisions are made by a slicing motion, and two of the incisions to Castellan’s neck were

especially long and deep. One of the deeper incisions, measuring six and a half inches long,

started on the lower left side of Castellan’s neck and ended by his right ear. The medical

examiner emphasized the depth of the wounds in explaining the cause of death to the jury:

Mr. Castellan died as a result of multiple incised cuts along [the right and left
sides of his neck] . . . of a such a severity that [they] cut the muscles, cut the
tendons, cut the soft tissue, cut a portion of the right hyoid bone, and also
slashed the right jugular vein, which is a good size vein that bleeds quite
profusely. Eventually, he [ex]sanguinated.

The State charged the offense as a capital murder on the theory that Applicant
 ARMSTRONG — 10

committed the murder in the course of a robbery or attempted robbery. T EX. P ENAL C ODE §

19.03(a)(2). The State argued that Applicant had committed the murder so he could get

money to buy more drinks at the bar. Five pieces of evidence supported the theory that a

robbery had occurred: First, the eyewitnesses testified that, after Applicant slashed Castellan,

he went through his pockets. Second, Castellan’s back pockets were torn off of his pants.

Third, a police officer testified that he found Castellan’s medicaid identification papers in

Applicant’s pocket at booking. Fourth, women for whom Applicant had been buying drinks

at the Sunshine Bar prior to the murder testified that Applicant had told them he was out of

money when he left the bar, but later came back and bought more drinks. Fifth, Castellan’s

blood was later found on money in the bar’s cash register. The jury found Applicant guilty

of capital murder.

B. The Punishment Phase Evidence

The State called two witnesses at the punishment phase of trial. First, it called a

fingerprint expert to authenticate the judgments in Applicant’s prior convictions for robbery,

arson, theft, and terroristic threat. Next it called Castellan’s brother to testify that Castellan

had children and grandchildren, that he “loved life” in the Rio Grande Valley, and that all of

Castellan’s friends from the adult daycare he attended had come to his funeral.

The defense team of first-chair Rogelio Garza and second-chair Nereyda Morales-

Martinez called three witnesses during the punishment phase: Applicant’s pastor, Applicant’s

common law wife, with whom he had a young son, and Applicant’s younger full-sister,
 ARMSTRONG — 11

Sheila Armstrong. Applicant’s pastor and common law wife did not meet Applicant until he

moved to Donna within a year of the murder. Applicant’s pastor testified that Applicant had

a job, had a positive attitude, and attended church regularly. To his pastor, Applicant

appeared to be “clearly in a depressed situation” when he counseled him at the jail.4 The

pastor was unaware that Applicant had prior offenses and that he had children by other

women in other states. The pastor acknowledged Applicant’s trouble with substance abuse,

and the fact that Applicant had not sought help for, or spiritual advice regarding, his apparent

addiction.

Applicant’s wife testified that she had lived with Applicant for only four-and-a-half

months prior to the murder. Applicant had never abused her in any way. She testified that

Applicant was generally a good guy, but she was “fed up” with his conduct and cocaine use.

The State introduced her statement to the police in which she had called Applicant “stupid

Tyrone,” but she denied using those words. The State also introduced a note that she admitted

to have written, which read:

Tyrone,
Thanks 4 everything. You stole, lied and cheated on me. There is nothing
between us so have @ nice life enjoy doing crack and please don’t look for
me.
enjoy the apartment
it all yours.
love, Cindy.

Sheila Armstrong, Applicant’s younger sister, was the only witness who had known

4
This was a lay opinion. It was not offered as a clinical diagnosis.
 ARMSTRONG — 12

Applicant before he moved to Donna. She gave general testimony about the hardships

Applicant faced growing up. Sheila testified that Applicant’s mother, Earline Armstrong

(“Earline”), had six children, but only Sheila shared a biological father with Applicant. They

grew up together in various housing projects, moving frequently. Sheila and Applicant

attended four different elementary schools. According to Sheila, Applicant was treated as the

“black sheep” of the family because he “was too much like” his biological father, Douglas

Clark (“Clark”). Clark abused Applicant’s mother and the children. Clark would lock

Applicant in an empty room, and Sheila would have to sneak him food. Clark also raped

Sheila, resulting in a pregnancy. Sheila and Applicant would often hide in the attic until their

mother returned, because they felt unsafe in the house alone with Clark. Applicant would do

“whatever he could” to protect Sheila. Sheila testified that Clark did not molest or rape her

as long as Applicant was home. Clark would beat Applicant bloody with whatever was

handy, including “old train tracks,” broom sticks, and extension cords. The children

witnessed Clark abusing their mother, including an incident in which he “knocked her eye

out, so she had to get help.” They had no food to eat, and Clark spent all of the money.

Sheila and Applicant had always been close. She was familiar with Applicant’s

struggles with the law and his out-of-state children. She explained that his terroristic threat

conviction resulted from an incident in which he had come home to find the mother of one

of his children with another man. Earline, their mother, had died while Applicant was serving

his one-year sentence for that conviction, and he was not able to attend her funeral. She
 ARMSTRONG — 13

testified that when Applicant was not incarcerated, he was able to hold down jobs and

support his numerous children.

Sheila admitted that even though she had been raped and abused by her father, she

herself had never committed capital murder. But she explained that she had thought about

committing murder. She credited counseling she received after she was raped for motivating

her to “change her life.” Applicant also received counseling as a child directed toward his

mental health and performance in school, but the counseling stopped after he went to juvenile

detention, which Sheila guessed occurred when he was twelve years old. She remembered

that, as a school-aged child, Applicant did things that were “just wrong,” like shaving his

eyebrows and plucking his eyelashes. She testified that Applicant was “slow,” and “not all

the way there,” and that he attended special education classes. Though Sheila was younger

than Applicant, she passed him in school, and “winded up being grades ahead of him.” On

cross examination, Sheila conceded that, as an adult, Applicant was capable of traveling

across the country by himself, holding down jobs, supporting children, deciding where to live

and work, choosing how to spend money, and paying his bills when he could. He was tried

as an adult for robbery at age fifteen or sixteen, and went to prison for ten years.

III. THE POST-CONVICTION MITIGATION EVIDENCE

A. The Lay Witnesses

At the writ hearing on remand, Applicant provided testimony from several of his
 ARMSTRONG — 14

family members who live in Alabama and Georgia.5 These family members witnessed the

conditions of Applicant’s childhood, and their testimony both corroborated and expanded

upon Sheila’s trial testimony. Sheila also testified, expanding upon her own trial testimony.

In addition to Sheila, two of Applicant’s older half-siblings, Pamlet Armstrong and Vincent

Armstrong, who had also lived with Earline and Clark, corroborated Sheila’s accounts.

Pamlet grew up in Clark and Earline’s house until she was removed and went to the blind

school, and later she came back. Vincent lived with his father in Talladega and attended the

blind school there, but he stayed with Clark and Earline during the summers and on holidays.

Benny Fields, Earline’s brother (Applicant’s maternal uncle), testified to the conditions as

he observed them on his regular visits to Applicant’s childhood homes.

Three other witnesses who had provided affidavits for the writ application had health

complications that prevented them from traveling to the writ hearing on remand. These health

complications, they claimed, would not have prevented them from traveling at the time of

trial. Clark’s brother, Johnny Clark (“Johnny”), and one of his sisters, Susie Dell Hall, gave

depositions instead. Clark’s youngest sister, Shirley Ann Clark (“Shirley Ann”), could not

give a deposition, but she gave a declaration under penalty of perjury reaffirming the contents

of her original affidavit and her availability to testify at the time of trial.6 The Clark siblings

5
Affidavits from each of these witnesses had been submitted with the writ application.
6
In that affidavit, Shirley Ann described her relationship with Clark as an adult, which
involved shooting at each other. She also explained that their mother and father (Applicant’s paternal
grandparents) were alcoholics who did not have money for food, and they would have to eat at other
people’s houses or not eat at all.
 ARMSTRONG — 15

provided background information on the Clark family of origin, including accounts of sexual

abuse and alcoholism pervading throughout the Clark family tree. They each also confirmed

the abusive conditions that were present in Applicant’s household during his youth.

1. Physical Abuse

Sheila testified that when Applicant was locked in a bedroom as a child, it was not an

isolated incident, and it would last anywhere from a few days to a week. During this time,

he would plead with her to bring him food. Sheila would then try to find food in the house.

If she found food, she would stack buckets on top of each other on the side of the house “to

build a little ladder” so that she could reach the window to pass the food through it to

Applicant. He would not have access to a bathroom, so he had no choice but to relieve

himself on the floor. Pamlet corroborated Sheila’s account that Applicant would be locked

in the empty room with no food, human contact, or a bathroom for days. The urine and feces

that Applicant left on the floor would not be cleaned up. Sheila described the effect of this

isolation on Applicant as severe. She testified that after their parents began locking him in

a room by himself in fourth grade, Applicant changed. He became “empty” and “blank,” and

it was around that time that he began getting into trouble with the law and first told Sheila

for the first time that he wanted to die.

Sheila, Pamlet, and Vincent Armstrong all testified that Clark was constantly drunk,

and whenever he was drunk, he was combative.7 Vincent testified that Clark would beat the

7
During her testimony at the punishment phase of trial, Sheila was never asked, and made
no mention of, the alcoholism and substance abuse that, according to her later testimony at the writ
 ARMSTRONG — 16

children. Vincent testified to receiving beatings when he was living in the house, and to

hearing the sounds of Clark taking Applicant into a room and striking him with a belt.

Sheila again testified that Clark physically abused their mother, Earline, at one point

knocking her eye out. At the writ hearing she went into further detail about this incident,

specifying that Clark had hit Earline over the back of the head with a tire iron, causing her

eye to pop out.8 Applicant witnessed this assault. Sheila testified that he would hit Earline

with his fists, a telephone, or anything else he had available that he could throw at her. Such

abuse happened regularly throughout Sheila and Applicant’s childhood, and Applicant

witnessed it as well. Other family members confirmed that Clark would abuse Earline and

that Applicant would witness the abuse.

Sheila testified that Earline would also beat the children. She would have them lie

down on the floor, put her feet across their necks to hold them still, and select a beating

utensil—a belt, extension cord, switch, piece of Hotwheels track, or whatever else was

available—and beat them. Applicant would get up from the beatings bruised and bloody. He

would be beaten in the head, and scars from those beatings remain on his head today. Vincent

hearing on remand, was pervasive in the home during Applicant’s childhood, fueling most of the
abuse and neglect he suffered and priming him for a life of substance abuse.
8
Pamlet recalled an incident where Clark broke into the children’s bedroom through the
window while they slept, snuck into their mother’s room, and hit her with a tire iron. She remembers
hearing her mother’s screams and her mother’s attempts to mop up the blood gushing from her head
with a white towel, which soon became red. This may well be the same incident that Sheila conveyed
to the jury, in abbreviated fashion, at the punishment phase of trial. Pamlet also recounted an incident
in which Clark cut Earline’s arm with a knife, which stood out in her mind because Clark would
normally only beat her, bloody her nose, hit her with his fists, stomp on her, and break her glasses.
 ARMSTRONG — 17

corroborated this testimony, and included “shoe” and “broom handle” in his list of Earline’s

weapons of choice.

Applicant was also abused by his siblings. Sheila testified that both she and Applicant

were treated differently among the siblings because they were Clark’s children. But

Applicant was viewed with exceptional disdain because he resembled Clark. Vincent and

Pamlet corroborated this general description and provided some specific instances of when

they abused Applicant.

The beatings caused the children to miss school and fall behind, because Earline and

Clark would not allow them to go to school with welts and bruises. Applicant could usually

go back to school sooner because his skin was darker, so the physical signs of abuse were not

as obvious to school employees who might otherwise have reported the abuse. The children

generally would not complete their homework because they were either absent when it was

assigned or absent when it was due. The probability that the class would still be on the same

topic when they returned to school was low enough that the children felt that it was not worth

ever attempting to complete homework.

2. Sexual Abuse

While Sheila testified at trial that Clark had been incarcerated for raping her, the post-

conviction investigation revealed pervasive sexual abuse in the Armstrong family.

Applicant’s extended family members testified to various incidents of molestation and sexual

abuse throughout the Armstrong family tree. Susie Dell Hall testified that the rampant sexual
 ARMSTRONG — 18

abuse in the family was ignored by her mother (Applicant’s paternal grandmother), and left

the abused and those who loved them angry, traumatized, and with very low self esteem.

Clark in particular engaged in sexual deviance at a young age. Johnny averred that he

overheard an argument between Clark and Applicant, in which Applicant was yelling at

Clark about how Clark had molested Sheila and Applicant, and “really messed” both of them

up. Both Sheila and Pamlet related details of particular instances in which Clark had sexually

abused (or tried to sexually abuse) them, and both described Earline’s apparent indifference

to these incidents.

3. Privation and Neglect

Sheila revealed for the first time at the writ hearing that Earline was an alcoholic, and

drank away all their money so they could not pay bills or buy food for the house. Her mother

drank all day, every day, and would get so drunk every day that she could not do anything to

care for the children. She would throw up—sometimes on the children—and pass out. Clark

also abused alcohol, and every time Sheila saw him, he was either drinking or drunk. Clark

would buy the children a single meal per month on the day he received his social security

check, but otherwise he spent all of the family’s money drinking, smoking, and “partying

up.” Vincent, Pamlet, Fields, Johnny, Shirley Ann, and Susie Dell Hall all provided evidence

supporting Sheila’s writ-hearing narrative of alcohol-fueled neglect. Fields testified that

Earline and Clark were alcoholics, and that they were drinking or drunk whenever he visited

the home, which was approximately twice per week when Applicant was an infant and
 ARMSTRONG — 19

toddler. Earline and Clark would drink whatever quantity of alcohol was in front of them

until it was gone.9 According to Fields, when the parents were drinking, the children were

not allowed to speak to them. Even if the children said something like “I’m hungry today,”

their parents would respond “Go back there in that back room!” or “Get out of my face!”

Susie Dell Hall agreed during her deposition that Clark was an alcoholic who was

always drunk. Vincent also testified that both parents used and abused alcohol. Though

Vincent was blind, he could tell when Earline was drunk in the house because “that’s when

. . . we’d get beatings lots of times.” The next day, she would stay in bed, “hollering” at the

children to bring her things and requiring them “to wait on her, hand and foot.” Other times,

she would leave the house to go drinking—sometimes for two or three days at a

time—leaving the children without adult supervision. Pamlet testified that both parents drank

every day, and when Earline drank, she would slur, cry, throw up, and lash out at the

children. Shirley Ann and Johnny each confirmed that Clark was a heavy drinker and a mean

drunk.

At the writ hearing on remand, Sheila elaborated on her trial testimony that Clark

would “take all the money” and that the family “had no food to eat.” She explained that Clark

would steal money from Earline when she had it and spend it on drugs and alcohol and that

the children were close to starving and never had groceries in the house. Earline and Clark

9
Fields explained further: “If they bought a fifth of whiskey, then they drunk a fifth of
whiskey. If they bought a gallon of whiskey—or a half gallon—I kind of stretched that when I said
‘a gallon’—a half a gallon of whiskey, they drunk a half gallon of whiskey.”
 ARMSTRONG — 20

would not feed the children. In addition to stealing Earline’s money, Clark would steal the

family’s food stamps but never bring food home. On good days, Sheila and Applicant would

make mayonnaise and ketchup sandwiches. Sheila said they sometimes had bologna. Other

times, they would make “pancakes” by combining sugar, water, and flour in a frying pan. If

they did not have flour or bread, they would go hungry unless Applicant stole food and

brought it home. Pamlet recalled that Applicant worked hard to try to get money. Sheila

testified that when Applicant was eleven or twelve, he took a job cleaning a nightclub across

the street to earn some money so he could buy “little snacks . . . or honey buns” for himself

and Sheila to eat. One time, Sheila recalled, she and the other children were “so happy”

because Applicant surprised them by bringing food home, but Earline was mad and beat him.

Despite having beaten Applicant for stealing it, Earline herself ate some of the food.

The children could get food from school during the school year, or from the recreation

center across the street in the summer. However, their parents did not wake them up for

school or help them with their homework. If they missed their bus, but there was a chance

they could still get to school by lunch time, they would attempt to walk the ten miles to

school so they could eat.10 If the children did not wake up in time to catch the bus on their

own, or if they were being held out of school until their bruises healed, they often did not eat.

Fields, Pamlet, and Vincent corroborated this testimony. Fields testified that, when

10
During some period of Sheila and Applicant’s childhood, a “special bus” would pick up
their disabled brother Patrick in the morning. During this period, when Sheila and Applicant would
miss the bus, they would help Patrick get on the special bus, and the special bus driver would take
them to school.
 ARMSTRONG — 21

he would visit, the refrigerator would sometimes be full of alcohol, but most of the time there

was not any food in it. The children, including Applicant, would say to Fields, “Uncle Benny,

I’m hungry today.” When he could, Fields would take them to McDonald’s or Burger King

or his parents’ home. If he was going to visit and knew he was not going to be able to take

them out of the house, he would try to remember to bring food for them. Fields had to feed

the children “a lot.” In hindsight, Fields wished he would have also bought them “proper

clothes.”

Pamlet also remembered eating mayonnaise and ketchup sandwiches without meat,

cheese, or anything else: “That was the meal.” Pamlet testified that their mother never

cleaned up when she was drinking, so any food they had would be left out for days. Even

though the food would be “hard, and matted,” with “roaches crawl[ing] through it,” they

would eat it. When the children stole money for food, Pamlet testified, they would eat all the

food they could and throw the rest away to avoid being caught. Pamlet appreciated

Applicant’s efforts to steal food for the rest of the children, and recalled an occasion where

Applicant had broken into “the school or something” and stolen a bunch of french fries,

enabling the children to eat french fries “for a good little while.” 11

Whether the children stole food or they got it from the community center, they had to

share it with Patrick, who was unable to get any food himself. The writ hearing on remand

11
Pamlet’s testimony is consistent with a statement in Vincent’s affidavit that Applicant
“used to steal food from school so that the family could eat.” Vincent was not asked about this when
he testified at the writ hearing on remand.
 ARMSTRONG — 22

revealed that Applicant’s half-brother Patrick was born with cerebral palsy and abandoned

in a room by Earline and Clark with only a broom to play with. According to Sheila, Pamlet,

and Vincent, Patrick could not communicate, walk, or do anything for himself. Earline did

not take care of Patrick, so the children medicated him, fed him, changed his diaper, and

bathed him until he eventually died. If they had running water, Sheila, Pamlet, and Applicant

would have to work together to get Patrick in and out of the bath tub to “soak” him so they

could scrub the feces off of him more easily. They had to puree and crush all of his food and

medicine “to the best of [their] abilities” as children. Though blind, Vincent knew that

Patrick was not well taken care of because Patrick would smell like urine, and his skin felt

rough. Vincent testified that sometimes Patrick would have a diaper on, but sometimes he

wore only a tee-shirt. Pamlet testified that they tried to change Patrick’s diapers, but they

often did not have a new diaper available, so they would find a rag or an old piece of clothing

lying in the filth on the floor and tie it around him. Other times they would let him crawl

around naked. The sheets on Patrick’s bed would not be changed. The odors associated with

Patrick’s condition caused the whole house to stink, and the other children at school claimed

to be able to smell the house when they walked past it.

4. Squalid and Dangerous Home Life

Even if they had not been abused by their own parents, Applicant and his siblings

never had a clean, safe place to live. Pamlet and Vincent testified that rats and roaches

infested all of the homes the children inhabited during their youth, but their parents never did
 ARMSTRONG — 23

anything to get rid of them. Pamlet and Sheila testified that their electricity and water were

sometimes turned off because their parents would not pay the bills. Pamlet also testified that

the children did not have toilet paper, toothbrushes, toothpaste, or soap. She and the other

Armstrong children, including Applicant, were the “laughing stock of the school,” and the

other children made fun of their unbrushed hair and teeth, their unwashed faces, and their

dirty clothes.

Sheila agreed that the children were never clean or dressed properly when they went

to elementary school, and they were bullied as a result. Their clothes were “dirty” and did not

fit. The kids at school called Applicant “fat boy” as a result of his tight clothing. Applicant

got into fights with people who made fun of him for the way he smelled. Pamlet confirmed

that the children’s clothes were dirty and too big or too small. Applicant’s clothes were the

worst: torn, ragged, and just “bad.” Fields confirmed that when he would visit the children,

they were not properly or sufficiently dressed much of the time. He would visit in winter and

the children would be playing outside without proper coats or shoes. Vincent testified that

the family did not have enough money to provide clothing for the children.

In addition to corroborating Sheila’s testimony that Clark and Earline spent all their

money on alcohol rather than food or clothes for the children, Pamlet testified that their

parents did not pay the rent. As a result, the family had to move constantly, and the children

never had a stable home. Three of the many places where the family lived were described in

detail at the hearing on remand:
 ARMSTRONG — 24

Happy Hills Projects. Happy Hills Projects is the first neighborhood Pamlet

remembered living in. It was not a safe place. Pamlet recalled an incident in which the entire

family, including then-three-year-old Applicant, was sitting in the living room watching

television, and Clark was shot when he answered the door. During the period the family lived

in Happy Hills, Earline and Clark were constantly having people over to the house to drink.

When Pamlet was about eight years old, some of the strange men that would come over

would look at her sexually. She started hiding from them, but they walked around the house

trying to find her and touch her. These men would also pass out in the house after drinking.

Pamlet recalled Clark asking her to “clip” the men—go through their pockets and take their

money—after they passed out. Pamlet was temporarily removed from the home at Happy

Hills after Clark hit Earline with the tire iron, and she recalled having difficulty packing her

things because the floor “was just like a carpet of clothing, and roaches, and rats.” At Happy

Hills, all of the children, including Patrick, shared one bed.

Orange Grove Projects. Orange Grove is the first home Sheila could remember. She

testified that the Orange Grove Projects were a “ghetto” and the home was “nasty,” “filthy,”

and “dirty.” Fields agreed that the inside of the house “wasn’t a clean environment.” Sheila

added that the children could not go outside because there was a lot of gun violence and

broken glass everywhere. Once, when Sheila and Applicant did go outside, they saw a man

on the roof of his house eating a dead cat, and they ran home afraid. When she was in first

or second grade, Sheila witnessed her mother beat a man with a baseball bat while the other
 ARMSTRONG — 25

children were trying to sleep. Fields also thought the neighborhood was unsafe because there

was “a lot of violence,” and people fought for no reason.

Pamlet testified that Orange Grove was no safer than Happy Hills. Earline and Clark

continued to have drunken parties in their home. The danger was compounded by the fact

that the parents sold alcohol out of the home to total strangers. This business was open at all

hours of the day. Pamlet did not feel it was safe to have strangers coming and going at all

hours. The whole house was available to the customers, from the kitchen, to the bathroom,

even the children’s bedroom. One night, Pamlet and her boyfriend were abducted at knife

point and taken out the back door. She was sexually assaulted in a car, and left on a bridge

to walk home “almost naked.” When she arrived home, the police were already at the house.

She gave her statement to the police in front of Applicant, who was six or seven years old

at the time. At Orange Grove, the three children shared a room with two beds. When Pamlet

and her boyfriend were kidnapped, they were taken from the same room where Sheila and

Applicant were asleep.

St. Stephens Road. All witnesses with knowledge of the house at St. Stephens Road

testified that it was the worst of the domiciles the children inhabited during their youth.

Sheila described it as “filthy” and more of a shack than a house. When she and Applicant

snuck downstairs one night to have some cake after their parents had a party and left the cake

out on the kitchen table, they stopped midway down the stairs upon seeing a “pool” of giant

rats swarming over their kitchen table, presumably on top of the cake that the parents had left
 ARMSTRONG — 26

out. Pamlet verified that the rats at the St. Stephens Road house were the largest, and all of

the witnesses with knowledge of the house described rats the size of squirrels. They had no

clothes to put in the drawers, but they would not have used the drawers for that purpose

anyway because there were cockroaches and baby rats nesting there. This was the house

where Applicant was locked in a room by himself in his own filth, and where Sheila first

heard him talk about wanting to die. There were dirty clothes and dishes everywhere, and the

house as a whole smelled strongly of urine. Sheila testified that the St. Stephens Road house

was eventually condemned.

5. Applicant’s Substance Abuse

Vincent testified that when he would visit the family as a young teenager, he would

give four-to-five year old Applicant enough alcohol to get him drunk. On these occasions,

Applicant would dance in a tipsy manner, causing everyone to laugh at him. The two

continued to drink together as Applicant grew older, and Applicant would tell Vincent that

he also used drugs. Specifically, Applicant said he would “snort that powder.”

Shirley’s affidavit states that Applicant stayed with her as an adult, about three years

before the murder. During this time, he was using cocaine in powder and crack form, as well

as methamphetamine. She characterized him as a functional drug addict. Johnny’s impression

of Applicant during that period was also that he had become a drug addict.

B. The Mental Health Witnesses

To the extent that the lay witnesses only essentially corroborated the substance of
 ARMSTRONG — 27

Sheila’s punishment-phase testimony, it might well be argued that they failed to provide

mitigating evidence that “differed in a substantial way—in strength and subject matter—from

the evidence actually presented at sentencing.” Martinez, 195 S.W.3d at 731 (citing Hill v

Mitchell, 400 F.3d at 319). But more has been shown to have been left out in this case than

simply the new lay testimony about Applicant’s disadvantaged childhood. Applicant also

presented testimony at the writ hearing from two psychological experts describing

Applicant’s mental condition at the time of the offense, stemming in part from their

perceptions of the deleterious effects that Applicant’s disadvantaged childhood had on his

mental makeup and stability.

The first expert was Dr. Phillip D. Harvey, a clinical psychologist, specializing in the

study and diagnosis of cognitive impairment. The second expert, Dr. Robert Lee Smith, was

a forensic psychologist and addiction specialist. Smith interviewed Applicant, conducted

collateral interviews with family members who had observed him, and reviewed Applicant’s

available mental health records. Both experts were engaged in this kind of work at the time

of trial, and said there would have been no barrier to evaluating Applicant and testifying at

trial in 2007 had they been contacted by trial counsel.12 Neither Smith nor Harvey found any

12
One of Morales-Martinez’s biggest concerns at the time that voir dire commenced was that
the defense team “had not been able to look into what kind of experts” might be needed. Morales-
Martinez knew “that [Applicant] had been getting counseling of some type,” and “he had some
mental health problems when he was young,” but the defense team “just didn’t have the time” to
track down Applicant’s records and investigate his mental health history. Morales-Martinez prepared
and filed a motion for continuance because she believed additional mitigation investigation was
called for, but she withdrew the motion before the trial court could rule on it at the behest of lead
counsel Garza, who declared himself ready to proceed to trial. Having withdrawn the motion for
 ARMSTRONG — 28

continuance, trial counsel failed to preserve that issue for appeal.

At about the time that voir dire began, Garza and Morales-Martinez did hire one expert, a
clinical psychologist, Dr. John Pinkerman, to evaluate Applicant. Unbeknownst to counsel,
Pinkerman’s professional license was on probation at the time for unprofessional conduct. Habeas
counsel found an unsigned draft report on Pinkerman’s letterhead in one of the mitigation specialists’
files. That report confirms that Pinkerman did not meet with Applicant until the middle of voir dire.
The report describes the results of four testing instruments: the Weschler Adult Intelligence Scale-III
(WAIS-III), the WRAT-3, the Millon Clinical Multilaxial Inventory-III (MCMI-III), and the MMPI-
2. Applicant’s “cooperation and motivation were appropriate” for the administration of the WAIS-III.
Applicant had a Verbal I.Q. score of 73, (which is borderline intellectual disability), a “Performance
I.Q.” score of 92, (which is average), and a “Full Scale” I.Q. score of 80 (which is “low average”).
Pinkerman also administered the WRAT-3, which ranked Applicant in the 0.6th percentile in
reading, the first percentile in Spelling, and the seventh percentile in Arithmetic. But Applicant could
not, according to the draft report, “answer the [MMPI-2 and MCMI-III] questionnaires in a valid
manner that allows clinical interpretation.” The reports goes on to explain that “[t]he nearly uniform
elevation of his obtained scores [on these instruments] could not be interpreted in a valid manner.”
It is not clear what the defense team hoped to learn from the MMPI-2, but Smith would later testify
that an MMPI-2 administered while Applicant was incarcerated would not have been an appropriate
instrument to determine his psychological profile before or during the offense.

At the motion for new trial hearing, the initial writ hearing, and writ hearing on remand,
Morales-Martinez provided different accounts of a conversation with Pinkerman from which she
drew the conclusion that Applicant was either exaggerating his answers to the MMPI-2, or simply
not cooperating. At the motion for new trial hearing, she testified that Pinkerman “never said
[Applicant] was being deceitful” during the MMPI-2, from which Morales-Martinez drew the
conclusion that Applicant had “cooperat[ed], but apparently the answers [had been] exaggerated, and
that was not good.” Later, at the initial writ hearing, she testified that Pinkerman “indicated” to her
that he “knew, based on his questions to [Applicant], that [Applicant] had lied to him.” By the time
of the writ hearing on remand, her testimony was that Pinkerman had indicated that Applicant “was,
obviously, lying to him.” Applicant told Morales-Martinez that he would not change his answers if
the test were administered again, so Morales-Martinez “chose not to have him retested by Pinkerman
because [Applicant] was not going to cooperate.” Based on her motion for new trial hearing
testimony, she never considered requesting a different expert or a more thorough mental health
evaluation. At the writ hearing on remand, her testimony was that “she just did not see how it would
be effective to get another person. For what? For him to lie to that person too? It’s just not logical.”

At the motion for new trial hearing, Gilda Bowen, the mitigation specialist, testified that, had
she not run out of time, she would have found an expert to explain to her and the lawyers what
Pinkerman’s test results meant, and “would’ve liked somebody else to look at them and interpret”
them. She asked the lawyers to retest Applicant, but the lawyers told her that they “didn’t have any
more time,” so the mental health investigation “pretty much died there.” According to Bowen,
 ARMSTRONG — 29

evidence that Applicant malingered, exaggerated, or manipulated any test results. Both

performed batteries of tests with built-in “red flags” to detect malingering, and none of those

red flags was triggered. Both also routinely test for malingering as part of their normal

scientific pursuits.13

Pinkerman was not asked to review Applicant’s school records or mental health records from prison.
Harvey testified at the hearing on remand that Pinkerman’s draft report mis-characterizes the
meaning of elevated validity scores. “[E]levated validity scores,” he explained, do not mean that the
patient cannot validly answer the questions, and “ethnic minorities, individuals with less education,
and individuals who are under true duress—such as being charged [with] murder—often have
elevations in their validity scales that need to be considered in context.” Smith likewise testified that
results can be invalid “for a number of different reasons,” but even if Applicant was malingering,
“that doesn’t mean that we can’t continue to work with that individual.” Smith added that, given
Applicant’s diminished intellectual functioning, “it would be difficult for [Applicant] to understand”
the questions on the MMPI-2, some of which involve double negatives. Harvey and Smith both
testified that such an invalid result indicates the need for further testing. Pinkerman’s draft report
concluded that the best option to obtain valid estimates of personality functioning is to “retest
[Applicant] enlisting his cooperation and compliance in the task.” Harvey testified that mental health
professionals who struggle with assessing people will often say that their patients are not
cooperative. Harvey could not independently evaluate the validity of Applicant’s scores on the tests
Pinkerman administered because Pinkerman had lost Applicant’s file, so Harvey did not have access
to the raw data.

The defense team’s failure to begin mental health evaluation testing until voir dire began,
hiring an expert with a probated license to conduct that testing, failing to seek further testing (and
any necessary continuances) when the expert was unable to get immediate valid results, and failing
to supply their expert with relevant school and prison mental health records, as we have already
concluded, all combined to amount to a deficient investigation of Applicant’s mental health. See
Armstrong, 2010 WL 359020, at *2.
13
Harvey gave an example of a federal case where he determined that “the person was clearly
malingering.” At the time of the writ hearing, Harvey was also involved in a federally funded study
of veterans with traumatic brain injury, where he was tasked with screening the veterans entering the
study to ensure they were not exaggerating their memory impairments.

Smith explained that licensed psychologists are trained to detect malingering, and that he has
participated in worker’s compensation cases in which it was important to determine whether workers
were “feigning or exaggerating their symptoms in order to receive benefits.”
 ARMSTRONG — 30

1. Dr. Harvey

Harvey evaluated Applicant for cognitive impairments.14 He diagnosed Applicant with

borderline intellectual functioning and acquired dementia. He also determined that

Applicant’s IQ and Wide Range Achievement Test (WRAT-4) scores suggested a learning

disability. The WRAT-4 put Applicant below the first percentile of the population in reading.

Harvey’s diagnosis of acquired dementia was based on Applicant’s score of “very impaired”

on other indices of cognitive functioning such as language, attention, immediate memory,

and delayed memory, which were significantly lower than his overall IQ. Harvey concluded

that Applicant’s impairments included impairments in judgment, reasoning, problem-solving,

attention, concentration, processing speed, and memory. He testified that Applicant’s

performance on these tests was consistent with cognitive disability associated with damage

to his fronto-striatal circuit. Applicant’s acquired dementia had several possible origins, any

one of which would have been sufficient, Harvey maintained, but that it was likely a

combination of substance abuse, including alcohol, crack cocaine and inhalants, repeated

traumatic brain injury, and “extreme life stress.” Moreover, he explained, Applicant’s

condition would have been fully formed at the time of the murder, and his symptoms may

have been even worse at that time than at the time of testing, because he was tested after

having lived for three years in a controlled prison environment.

14
Specifically, Harvey administered the third edition of the Wechsler Adult Intelligence
Scale, (WAIS-III), the Repeatable Battery for Assessment of Neuropsychological Syndromes (r-
BANS), the Hopkins Verbal Learning Test, Neuropsychological Assessment Battery Mazes Test,
Rey Complex Figure, and the fourth edition of the Wide Range Achievement Test (WRAT-4).
 ARMSTRONG — 31

2. Dr. Smith

Smith diagnosed Applicant with dysthymic disorder (long-standing depression),

substance dependence on alcohol and cocaine, and personality disorder-not otherwise

specified (which includes characteristics of paranoia and borderline personality disorder).

Smith determined that Applicant’s dysthymia was the result of environmental factors,

including his parents’ chronic intoxication, “violence between his parents, the emotional and

physical abuse of his sister, his own emotional and physical abuse by his father, neglect by

his mother, continual family relocations, changes in schools, and poverty.”

Smith used substance abuse screening tests to determine that Applicant was substance-

dependent. He found that Applicant’s childhood history of trauma and depression was a

significant risk factor for his development of addiction to alcohol and stimulants such as

cocaine and methamphetamine. His addiction, according to Smith, was the product of a

combination of genetic, psychological, and environmental factors.

Smith also diagnosed Applicant with a “personality disorder-not otherwise specified”

based on his background, including patterns of behavior during childhood and adolescence,

Applicant’s admissions regarding his beliefs about himself, others, and the community, the

opinions of family members, and documents. Based on that comprehensive review, Smith

testified that Applicant’s psychological characteristics fit into two diagnostic categories:

borderline and paranoia. Borderline, he explained, is a personality disorder in which an

individual is highly reactive to his environment, has very dysfunctional relationships, very
 ARMSTRONG — 32

low self esteem, depression, self-mutilation, and suicidal thoughts. Paranoia, according to

Smith, is characterized by suspicion of others, lack of trust, and fear of allowing people to

get too close. In Applicant’s life, Smith observed a pattern of unstable interpersonal

relationships, impulsiveness, recurrent suicidal behavior, labile mood, and chronic feelings

of emptiness. His “personality disorder-not otherwise specified” resulted, according to Smith,

from the same environmental factors as his depression.

Like Harvey, Smith recognized that Applicant had limited intellectual functioning.

Applicant’s school records demonstrated that Applicant had “significant problems” in school,

repeating many of the earlier grades. The pinnacle of Applicant’s academic achievement,

Smith testified, was seventh grade, and he failed a number of courses and struggled most in

verbal courses such as literature and language. Smith testified that Harvey’s assessment that

Applicant suffered from borderline intellectual functioning was consistent with this academic

record and Smith’s own observations of Applicant. Smith explained that limited intellectual

function is significant in the context of Applicant’s crime because individuals with limited

intellectual functioning have difficulty coping in stressful situations and also have difficulty

with problem solving.

Besides Applicant’s borderline intelligence, Smith explained, individuals with

Applicant’s history of traumatic physical abuse also tend to be suspicious, guarded, and

paranoid, and tend to overreact to situations in which they perceive that they may be at risk.

Smith further explained that the “counseling” Applicant received—mentioned by Sheila at
 ARMSTRONG — 33

trial and cited by the State in closing argument as evidence Applicant was intractable—was

extremely limited. As at trial, Sheila testified at the hearing that Applicant went to

counseling, but she was not sure for how long—she could only recall going with him “a

couple of times.” After analyzing Applicant’s mental health records, Smith determined that

Applicant participated in counseling at two points in his life: once at age eight, and again at

age thirteen. At age eight, he attended four counseling sessions because of behavioral

problems at school and at home, but he missed the other three appointments that were

scheduled. At age thirteen, Applicant had low-self esteem and had failed both first and

second grade. The records from ages eight and thirteen note that his mother had poor

parenting skills, and a treatment goal was to help her parent him better. After reviewing all

of Applicant’s mental health records, Smith concluded that any counseling Applicant

received was limited and directed toward improving Applicant’s environment by improving

his mother’s parenting skills rather than self-improvement of his mental health. Applicant

and his mother attended some sessions both times, but his mother cancelled the rest and

eventually decided to stop the counseling. The records corroborate Pamlet’s testimony that

Applicant went to counseling for a short time, maybe two or three times over a period of a

couple months. Thus, Smith concluded that Applicant’s mental health issues went largely

unaddressed throughout his childhood. Had this sort of testimony been presented at trial, it

would have corrected the impression Sheila left with the jury at the punishment phase of trial

that Applicant attended counseling somewhat regularly until he was twelve years old.
 ARMSTRONG — 34

Smith further explained that Applicant’s precise mental health issues caused him to

be highly reactive and impulsive, to lack judgment, to have difficulty coping with stressful

decisions, and to be unable to weigh the consequences of taking a particular action. All of

these qualities would have contributed to impair him psychologically from behaving

appropriately on the night of the offense. According to Smith, individuals with Applicant’s

history of “significant trauma,” where they have been “physically abused and mistreated,”

tend to be suspicious, guarded, paranoid, and tend to overreact in situations where they

perceive a risk. In addition, Smith opined, Applicant’s dysthymia also would have played a

role in his mental state on the night of the murder by enhancing his sense of despair and

entrapment, and causing him to believe he was being treated unfairly. Applicant’s history of

alcohol and drug abuse would have multiplied the effects of these disorders on cognition,

impulsivity, and judgment. He would have been “highly reactive, impulsive, demonstrating

poor judgment,” and would have “had difficulty coping” with the stress of the situation and

with “making decisions about weighing the pros and cons and what would be the best action

at the time.” Smith’s report concluded that “the culmination of each of these factors,

individually and collectively, played a significant role” in the commission of the offense.

IV. THE CONVICTING COURT’S CREDIBILITY DETERMINATIONS

We remanded this case to provide the convicting court an opportunity for further

factual development and additional findings of fact with respect to essentially two questions:

1) whether Applicant’s witnesses, both lay and expert, are credible, and 2) whether they
 ARMSTRONG — 35

would have been available to testify at trial, had they been called upon to do so at that time.

Armstrong, 2015 WL 7354084, at *4-5. Regarding the first question, the convicting court

declared that all of the substantive evidence discussed above, from both lay and expert

witnesses, was credible. With regard to the second question, however, the convicting court’s

findings are not as clear.

A. The Lay Witnesses

The three family witnesses who gave live testimony at the writ hearing pursuant to our

remand order (besides Sheila, who did testify at the punishment phase of Applicant’s trial)

were his two older half-siblings, Pamlet and Vincent, and his maternal uncle, Fields. Pamlet

and Vincent both averred that they would have testified at Applicant’s trial had they been

asked to do so. Fields was in prison at the time of Applicant’s trial, but he would have given

a deposition, he claimed, had one been requested of him. Without expressly rejecting these

assurances as lacking credibility, the convicting court has entered findings of fact that list

circumstances that suggest it ultimately finds them unconvincing. The convicting court found

it significant that: 1) Applicant’s family members do not presently talk to each other very

often; 2), most of them did not even realize that Applicant was on trial until after he was

convicted and sentenced to death; and 3) it was apparently only after he had been sentenced

to death that the importance of providing testimony that might help him became evident to

them. The convicting court apparently believed that these circumstances evince such a level

of indifference to Applicant’s fate—at least until a jury actually assessed a death
 ARMSTRONG — 36

sentence—as to engender doubt about the veracity of the family members’ present claims that

they would have made the effort to testify had they been asked to.15

The convicting court is the “original” fact-finder in post-conviction habeas corpus

proceedings, and this Court generally defers to its credibility determinations so long as there

is record support. Ex parte Reed, 271 S.W.3d 698, 727 (Tex. Crim. App. 2008). Even if we

were to defer to the convicting court’s apparent finding of incredibility in this instance, and

discount the additional family members’ testimony from the prejudice analysis, we would not

find it dispositive of Applicant’s present Wiggins claim. Assuming that none of Applicant’s

family members other than Sheila would have been willing to travel to Texas to testify on his

behalf until after the fact, there remains the question of whether the testimony of the mental

health experts would have made a substantial difference. Applicant’s family members

provided affidavits for habeas counsel readily enough. There is no particular reason to

believe that they would not have been willing to provide background information at the time

15
The only family member besides Sheila who was aware that Applicant was about to go to
trial for capital murder was Pamlet. There is some dispute in the record whether Pamlet was asked
to come to trial and testify. At the writ hearing on remand, she insisted that she would have come
if asked, and denied having ever told members of the defense team that she would not come because
she was sick, she did not want to miss work as a substitute schoolteacher, she is legally blind, and
she did not want to confess to having beaten Applicant during their childhood. She was not too sick,
blind, or busy with work to attend the writ hearing, however, at which she readily admitted her
mistreatment of Applicant during their youth. For her part, trial counsel Morales-Martinez gave
varying accounts of her efforts to recruit Pamlet, testifying at one point during the extended post-
conviction proceedings that Pamlet “would not have wanted to come” to Texas, but on another
occasion expressing uncertainty that she had ever even spoken to Pamlet prior to trial. A mitigation
specialist has stated in an affidavit that, about two weeks prior to trial, she spoke to Pamlet by
telephone and urged her to “come to Texas” to testify, but that Pamlet “absolutely could not.” The
convicting court’s findings of fact do not directly resolve this factual conflict other than to find
Morales-Martinez’s claim that Pamlet would not come to be “much more credible.”
 ARMSTRONG — 37

of trial to inform the diagnoses of experts such as Harvey and Smith, even if they did refuse

to come and testify in person. The information that goes into the formation of an expert

opinion need not itself take the form of admissible evidence to justify admission of that

expert opinion, T EX. R. E VID. 703, and Harvey and Smith could still have relayed their

opinions to the jury, as well as the factual bases of their opinions—to the extent that they

relied on information about the circumstances of Applicant’s upbringing thus supplied by

members of his immediate and extended family.16 Moreover, there is no basis in the record

whatsoever to doubt that Sheila would have been available to provide more expansive

testimony than she was asked to give at trial, since she made the journey twice to testify both

at the punishment phase and again at the writ hearing on remand.

B. The Expert Witnesses

Just as the convicting court found the lay witnesses’ substantive testimony to be

credible, it found the experts’ substantive testimony likewise to be, on the whole, credible.17

16
The underlying data supporting an expert opinion might be subject to exclusion “if their
probative value in helping the jury evaluate the opinion is outweighed by their prejudicial effect.”
TEX . R. EVID . 705(d). “But [this] limitation does not apply to an expert’s opinion itself.” Aguilar v.
State, 887 S.W.2d 27, 30 (Tex. Crim. App. 1994).
17
The trial court’s findings of fact and conclusions of law seem to discount Smith’s
testimony at least to a certain extent because Smith interviewed Applicant and Applicant’s family
members in addition to reviewing Applicant’s records, and Smith’s conclusions “are thus only as
good as that data upon which Dr. Smith based his conclusions.” It is unclear what the convicting
court meant by this, given that it found both the substance of the family’s various accounts of
Applicant’s childhood, and Smith’s substantive testimony with respect to Applicant’s mental status,
to be generally credible. In any event, whatever concern the convicting court may have about some
of the sources relied upon by Smith, it does not undermine Smith or Harvey’s ultimate assessments
of Applicant’s cognitive functioning. See Sears, 561 U.S. at 949 (“Whatever concern the dissent has
about some of the sources relied upon by Sears’ experts—informal personal accounts—it does not
 ARMSTRONG — 38

In fact, after Harvey finished testifying, the judge remarked that Harvey was more

“impressive” than most experts that had come through his courtroom. Both experts testified

that they would have been available to evaluate Applicant and testify at his trial and would

have done so had they been retained at that time. The convicting court nevertheless implicitly

decided that the experts would not have been available to testify at trial when it found “there

is no reason why Applicant’s trial attorneys would have found the out-of-state expert

witnesses whom habeas counsel later retained and it is unlikely that this court would have

granted requests for funds to pay said individuals, as opposed to qualified local experts.” It

is not lost on us that, though the convicting court may not have been willing to fund

Applicant’s particular habeas experts, it apparently would have been willing to fund

“qualified local experts.” Insofar as the convicting court determined that the substantive

testimony provided by Drs. Harvey and Smith was credible, there is no basis in this record

for concluding that qualified local experts would have testified any differently.18 Thus, we

undermine the well-credentialed expert’s assessment, based on between 12 and 16 hours of
interviews, testing, and observations, that Sears suffers from significant cognitive impairment.”).
18
Presiding Judge Keller’s dissent argues that Applicant has failed to satisfy his burden to
prove that he could have produced expert witnesses at trial who would have testified as Harvey and
Smith did at the writ hearing. It suggests that, because the convicting court was unwilling to fund
non-local experts, it is incumbent upon Applicant to demonstrate by a preponderance of the evidence
that Applicant could have presented the substance of Harvey and Smith’s testimony through local
experts whom the convicting court would have been willing to fund. Dissenting Opinion at 2.
“[N]othing in the record suggests that psychologists with similar credentials are unavailable
locally[,]” the dissent asserts. Id. In a supporting footnote, Presiding Judge Keller cites a concurring
opinion for the proposition that defense counsel does not perform deficiently for Strickland purposes
by relying on “a well-qualified, local, hands-on expert instead of a well-qualified, out-of-state,
research expert[.]” Id. n.2 (citing Ex parte Flores, 387 S.W.3d 626, 642 (Tex. Crim. App. 2012)
(Keller, P.J., concurring)). This is an unremarkable proposition. But it does not mean that defense
 ARMSTRONG — 39

will assume that the expert conclusions regarding Applicant’s mental health could have been

presented had trial counsel conducted an adequate mitigation investigation, and we will at

least consider whether Applicant was prejudiced inasmuch as the mental health testimony

was not presented at the punishment phase of his trial.

V. PREJUDICE

Once this Court has deferred to the convicting court’s resolution of all underlying

questions of historical fact and credibility, it is up to this Court, as the court of return in

capital post-conviction habeas corpus proceedings, to resolve the remaining legal question

de novo. See, e.g., Ex parte Navarijo, 433 S.W.3d 558, 567-68 (Tex. Crim. App. 2014) (once

underlying fact and credibility determinations are made, we apply a de novo standard of

review to the remaining legal question). A convicting court’s recommended conclusions to

this Court with respect to how the law ought to be applied to the facts as it has determined

them to be are just that—recommendations.19 We must decide for ourselves whether there

counsel is required to obtain an expert who is local before it may be said that he has performed
effectively under Strickland. We would err to convert this deficient-performance-prong proposition
into a requirement with respect to the prejudice prong of Strickland—that a post-conviction habeas
corpus applicant must establish that he could have presented a “well-qualified, local, hands-on
expert” to present his mental health mitigation testimony. And it would be no less of an error just
because the convicting court has declared itself unwilling to fund a non-local expert. (Never mind
that some localities may not even have a well-qualified expert.) Even assuming that a qualified local
expert were available, at least so long as the convicting court believes that the well-qualified non-
local expert has provided competent, reasonably objective, credible testimony, we believe it may be
inferred by a preponderance of the evidence for purposes of establishing prejudice that some well-
qualified expert, local or otherwise, would have been available to testify at trial essentially as that
non-local expert has.
19
While the convicting court made no overt recommendation with respect to the ultimate
question of prejudice on remand, it did make findings suggestive of a belief that no Hidalgo County
 ARMSTRONG — 40

is a reasonable probability that, had trial counsel performed an adequate mitigation

investigation, there is a reasonable probability of a different outcome, sufficient to undermine

confidence in the jury’s negative response to the mitigation special issue.

When some mitigating evidence has been presented at trial, and additional evidence

is developed during post-conviction proceedings, we consider what the marginal impact

might have been on the jury had both sets of evidence been presented as a package. See

Williams v. Taylor, 529 U.S. 362, 397-98 (2000) (state court’s prejudice determination was

unreasonable because it “failed to evaluate” the totality of the mitigating evidence because

it failed to consider how the post-conviction evidence would have complemented the

evidence and arguments counsel made at trial). Apart from the testimony from Applicant’s

pastor and wife, the only significant mitigating evidence presented at trial was Sheila’s broad

but sketchy account of the deprivation and abuse he suffered during his childhood. Her

jury would likely have been influenced by the kind of mitigating evidence that Applicant has
presented here. But the “assessment of prejudice should proceed on the assumption that the
decisionmaker is reasonably, conscientiously, and impartially applying the standards that govern the
decision[,]” and the prejudice inquiry “should not depend on the idiosyncracies of a particular
decisionmaker, such as unusual propensities toward harshness or leniency.” Strickland, 466 U.S. at
695. Our assessment of prejudice is therefore not contingent on the proclivities of a Hidalgo County
jury. It is also true that, before it will implicate the Sixth Amendment, an attorney’s deficient
performance must have “actually had an adverse effect on the defense” and that it must have had
more than “some conceivable effect on the outcome of the proceeding.” Id. at 693. But Strickland
itself makes it clear that a reviewing court’s assessment of “actual” prejudice depends upon how the
attorney deficiency would likely have impacted the deliberations of an objective decisionmaker
under the circumstances. Id. at 695. Indeed, the rules of evidence generally prohibit inquiries into
the actual deliberative process of a particular jury. TEX . R. EVID . 606(b). Consequently, we are
convinced that to give play to the convicting court’s perception of the particularly harsh proclivities
of a Hidalgo County jury in assessing “actual” prejudice would be tantamount to declaring that the
holding in Wiggins simply does not apply there.
 ARMSTRONG — 41

testimony at the writ hearing on remand added a great deal of texture and nuance, and

therefore contributed to the strength of her account. But the only additional subject matter

involved her descriptions (corroborated by the other family members) of the rampant and

destructive alcoholism that was pervasive on both sides of Applicant’s extended family, a

topic that was never broached during her punishment phase testimony. Yet this was an

important incremental addition. It helped lay the groundwork for making a case for

mitigation that could extend beyond a mere plea of sympathy for Applicant because of the

abuses and deprivations he suffered as a child, which the jury obviously felt was insufficient,

standing alone, to overcome the depravity of his offense.20

20
To provide a basis for assessing a sentence less than death, mitigating evidence need not
provide a “nexus” to specifically explain the capital offense. See Tennard v. Dretke, 542 U.S. 274,
287 (2004) (capital defendant need not prove that his crime was somehow “attributable to” the
mitigating circumstance he proffers in order to obtain a mitigation instruction). This is not to say,
however, that mitigating evidence that does provide such a nexus will not potentially have a greater
impact on the jury. Evidence that suggests that Applicant’s offense is in some sense “attributable to”
the circumstances of his upbringing will inevitably have a greater impact than evidence that merely
invites jury sympathy because of that deprived upbringing.

The evidence presented at the punishment phase of Applicant’s trial depicting his
disadvantaged childhood did not provide such a “nexus” to explain his offense; it merely served to
try to elicit a sympathetic response from the jury. With the addition of the expert testimony in the
case, however, Applicant could have described how the chaotic circumstances of his childhood,
fueled by a family legacy of alcoholism, contributed to a psychological diagnosis of mental
deficiency that helps explain how he could have perpetrated such an offense. To that extent, the jury
might have been willing to view his crime as at least partly the product of his dysfunctional
upbringing, over which he exerted no control whatsoever. Whether this renders Applicant’s new and
more detailed evidence of his disadvantaged childhood substantially different “in strength and
subject matter” from what was presented at the punishment phase of trial, Martinez, 195 S.W.3d at
731 (citing Hill, 400 F.3d at 319), is a question we need not resolve. The expert psychological
testimony itself, describing Applicant’s low intelligence and other mental deficiencies, is enough to
constitute mitigation of a substantially different character, regardless of whether it happens to derive
in part from an evaluation of the psychological impact of his disadvantaged past. And, unlike the
 ARMSTRONG — 42

The expert testimony that Applicant could have presented after a complete mitigation

investigation adds a significantly mitigating component because it explains the state of mind

from which he could have perpetrated the brutal offense for which he was convicted. Had

Applicant’s trial counsel conducted a more thorough mitigation investigation, they would

have been able to present evidence to show, and to argue, that his borderline intelligence, his

fronto-striatal damage and acquired dementia, his depression, the specific characteristics of

his personality disorder, and, particularly, his inherited propensity for alcoholism and

substance abuse, all combined to adversely impact his judgment and impulse control at the

time of the offense. Thus, the propensity of this new psychological testimony to reduce

Applicant’s moral culpability for the offense was considerable.

Jurors would not be required to credit the psychological evidence, of course, and

would still be justified in rejecting the mitigation special issue if they were to believe that the

mitigating value of Applicant’s disadvantaged childhood, along with his mental status at the

time of the offense, still failed to outweigh the aggravating circumstances: the obvious

brutality of the offense, the apparent depravity of the motive, and the criminal history that

preceded it. Moreover, the same psychological testimony that explains how Applicant could

have committed this capital offense could also serve somewhat to ameliorate the impact of

his criminal past.

naked evidence of childhood disadvantage presented at the punishment phase of Applicant’s trial,
it does provide a “nexus” that would tend, if believed, to ameliorate his moral culpability for the
offense itself.
 ARMSTRONG — 43

We need not conclude even by a preponderance of the evidence that the jury would

have answered the mitigation special issue in such a way as to spare Applicant’s life. We

need only conclude that there is a reasonable probability that at least one juror would have

insisted on voting affirmatively on the mitigation special issue. The addition of the expert

testimony to the existing picture of mitigation could, to that level of confidence, have tipped

the scale. When the additional mitigating evidence is added to the mix, the case for

mitigation is substantially “more compelling[,]” and “[w]e cannot say with confidence that

the facts of the capital murder and the aggravating evidence originally presented by the State

[in this case] would clearly outweigh the totality of [Applicant’s] mitigating evidence if a

jury had the opportunity to evaluate it again.” Gonzales, 204 S.W.3d at 399.

VI. CONCLUSION

Wiggins and its progeny compel us to hold that the deficient performance of

Applicant’s trial counsel in this case was prejudicial, entitling Applicant to relief in the form

of a new punishment hearing. Applicant’s sentence of death is vacated, and the cause is

remanded to the convicting court for a new punishment proceeding. All other relief with

respect to Applicant’s original writ application (No. WR-78,106-01) is denied. By separate

order, we dismiss Applicant’s second, third and fourth post-conviction applications for writ

of habeas corpus (Nos. WR-78,106-02 through WR-78,106-04) as abusive under Article

11.071, Section 5. T EX. C ODE C RIM. P ROC. art. 11.071, § 5.

DELIVERED: November 15, 2017
DO NOT PUBLISH