Lawrence Owens v. Stephen Duncan , 781 F.3d 360 ( 2015 )

In the
United States Court of Appeals
For the Seventh Circuit
____________________
No. 14-1419
LAWRENCE OWENS,
Petitioner-Appellant,
v.
STEPHEN DUNCAN, Warden,
Respondent-Appellee.
____________________
Appeal from the United States District Court for the
Northern District of Illinois, Eastern Division.
No. 08 C 7159 — Thomas M. Durkin, Judge.
____________________
ARGUED MARCH 3, 2015 — DECIDED MARCH 23, 2015
____________________
Before POSNER, KANNE, and TINDER, Circuit Judges.
POSNER, Circuit Judge. Lawrence Owens was convicted of
first-degree murder in November 2000 after a bench trial in
the Circuit Court of Cook County. The judge sentenced him
to 25 years in prison; he has almost 11 years of his sentence
left to serve. His conviction and sentence were affirmed. He
then filed state claims for post-conviction relief that came to
naught eight years after having been filed, when the state
supreme court declined to hear an appeal from the affir-
2 No. 14-1419
mance by the intermediate appellate court of the trial court’s
denial of Owens’ petition for such relief.
He had already filed a petition for federal habeas corpus
relief, 28 U.S.C. § 2254, and, the state proceedings having fi-
nally wound up, the federal district court adjudicated the
petition—and denied it, precipitating this appeal. In it he ar-
gues that the state trial judge who convicted him based his
decision on evidence that did not exist, thus denying him
due process of law in violation of the Fourteenth Amend-
ment. Holbrook v. Flynn, 475 U.S. 560, 567 (1986) (“one ac-
cused of a crime is entitled to have his guilt or innocence de-
termined solely on the basis of the evidence introduced at
trial,” quoting Taylor v. Kentucky, 436 U.S. 478, 485 (1978));
Estelle v. Williams, 425 U.S. 501, 503 (1976) (“The right to a
fair trial is a fundamental liberty secured by the Fourteenth
Amendment. The presumption of innocence, although not
articulated in the Constitution, is a basic component of a fair
trial under our system of criminal justice”) (citation omitted);
see also United States v. Moore, 572 F.3d 334, 341 (7th Cir.
2009) (“Guilt beyond a reasonable doubt cannot be premised
on pure conjecture”); United States v. Garcia, 439 F.3d 363,
366–68 (7th Cir. 2006) (”The presumption [of innocence] is
violated . . . when the jury is encouraged (or allowed) to con-
sider facts which have not been received in evidence”).
In 1999, in the City of Markham (current population
12,508), 20 miles south of Chicago, a young man named Ra-
mon Nelson, while riding his bike away from a liquor store,
received a fatal blow to the head by a person, presumably
male, wielding a wooden stick that may have been a baseball
bat. Two eyewitnesses to the murder testified at Owens’ tri-
al. Maurice Johnnie identified Owens as the murderer from a
No. 14-1419 3
photo array of six persons and from a lineup—although
Owens was the only person in the line-up who also was in
the photo array, thereby diminishing the probative value of
the second identification. The other eyewitness, William Ev-
ans, identified Owens as the murderer from the same photo
array and a lineup. But at the trial, though Owens was pre-
sent in the courtroom, Evans twice pointed to a photo of
someone else in the photo array as being Owens. There were
other discrepancies between the two witnesses’ testimony.
Evans testified that there had been two assailants, Johnnie
that there had been only one. And Evans but not Johnnie tes-
tified that Nelson had spoken with the assailants before they
assaulted him.
Nelson had crack cocaine on his person when he was
killed that appeared to be packaged for individual sale, for
the cocaine was in 40 small plastic bags in his coat. No evi-
dence was presented that Owens had known Nelson, used
or sold illegal drugs, or had any gang affiliation. If Owens
had had any record of involvement in the illegal drug trade,
or in gangs, the prosecution would, one imagines, have pre-
sented evidence of that involvement; it did not. Also absent
was any physical evidence (such as fingerprints on the base-
ball bat) pointing to Owens as the murderer. Moreover, the
murder had taken place at 8:30 p.m. on September 22. Sunset
was at 6:47 p.m. that evening, and so it would have been
dark (“nautical twilight” as it is called—the hour after sun-
set— would have ended by 7:47 p.m., see WeatherSpark,
“Average Weather On September 22 For Chicago, Illinois,
USA: Sun,” https://weatherspark.com/averages/30851/9/22/
Chicago-Illinois-United-States (visited March 11, 2015))—
three quarters of an hour before the murder took place. Not
pitch black, though, and apparently the area in which the
4 No. 14-1419
murder was committed had been illuminated to an unde-
termined extent by light from street lamps and from a near-
by building.
The eyewitness identification (at least by Maurice John-
nie) could, we assume despite the substantial doubts that
have been raised concerning the reliability of eyewitness ev-
idence (see, e.g., National Academy of Sciences, Identifying
the Culprit: Assessing Eyewitness Identification (2014); Sandra
Guerra Thompson, “Beyond a Reasonable Doubt? Reconsid-
ering Uncorroborated Eyewitness Identification Testimony,”
4 U.C. Davis L. R. 1487 (2008); Nancy Steblay et al., “Eyewit-
ness Accuracy Rates in Police Showup and Lineup Presenta-
tions: A Meta-Analytic Comparison,” 27 Law and Human Be-
havior 523 (2003); Gary L. Wells et al., “Eyewitness Identifica-
tion Procedures: Recommendations for Lineups and Pho-
tospreads,” 22 Law and Human Behavior 603 (1998)), have
supported a finding beyond a reasonable doubt that Owens
had murdered Nelson. But it is highly uncertain whether the
judge (for remember that Owens’ trial was a bench trial, not
a jury trial) thought that any of the evidence that had been
presented sufficed to dispel reasonable doubt of Owens’
guilt. For at the end of the parties’ closing arguments the
judge said: “I think all of the witnesses skirted the real issue.
The issue to me was you have a seventeen year old youth on
a bike who is a drug dealer [Nelson], who Larry Owens
knew he was a drug dealer. Larry Owens wanted to knock
him off. I think the State’s evidence has proved that fact.
Finding of guilty of murder.”
That was all the judge said in explanation of his verdict,
and it was nonsense. No evidence had been presented that
Owens knew that Nelson was a drug dealer or that he want-
No. 14-1419 5
ed to kill him (we assume that by “knock him off” the judge
meant “kill him”), or even knew him—a kid on a bike. The
prosecutor had said in his closing argument that the case
“boils down to identification …, how they identified him
and where … and in this case identification equals recogni-
tion.” The judge seems not to have been convinced, for he
said nothing to suggest that he thought the real issue in the
case was identification. If one may judge from what he said,
which is the only evidence of what he thought, he thought
that Owens’ knowledge that Nelson was a drug dealer was
the fact that dispelled reasonable doubt of Owens’ guilt.
Otherwise, why would he have called the existence of this
fact the “real issue” in the case—the basis therefore of the
verdict?
Even if the proper way to determine whether the judge
would have convicted Owens had he not mistakenly be-
lieved that Owens knew Nelson and knew him to be a drug
dealer (and wanted to kill him for that reason) would be to
ask the judge what he was thinking when he sentenced Ow-
ens almost 14 years ago, that approach would be impossible.
The judge died two years ago. Chicago Tribune Obituaries,
“Joseph M. Macellaio,” April 3, 2013, www.legacy.com/
obituaries/chicagotribune/obituary.aspx?pid=164012815 (vis-
ited March 11, 2015).
On Owens’ appeal from his conviction, the state appel-
late court ruled that the trial judge’s belief that Owens knew
Nelson to be, or was himself, involved with drugs or gangs
was baseless, saying “there was no evidence presented that
defendant knew Nelson was dealing drugs, and there was
no evidence presented that defendant was involved with
gangs or the illegal drug trade.” The court also said that “the
6 No. 14-1419
reliability of Evans’ testimony is severely called into ques-
tion,” noting not only Evans’ two failures to identify Owens
in the photo array in the courtroom but also his having been
promised by the prosecution probation on two drug charges
in exchange for his testimony. Nevertheless the court ruled
2-1 that the trial judge’s error had been harmless because
Johnnie’s eyewitness identification of Owens was sufficient
to establish Owens’ guilt. The federal district court, in deny-
ing Owens’ petition for habeas corpus, added nothing to the
state courts’ discussion of the harmless-error issue.
The dissenting judge in the state appellate court pointed
out that the trial judge had “never stated that he relied on
that identification or other properly admitted evidence,” so
that “what we do have is a trial judge manufacturing, sup-
plying, and interjecting its own evidence into a trial and then
affirmatively stating on the record that this manufactured
evidence constituted the basis of its verdict.” She added that
the error was not harmless because the trial judge had “only
considered unsupported insinuations without any indication
that he was aware of this impropriety.” The trial judge
hadn’t actually said that he’d considered only evidence not
in the trial record, but the only ground for his finding Owens
guilty that he mentioned had no basis in that record (or
elsewhere for that matter).
Not every fact on which a verdict is based must be found
in or inferred from evidence introduced by a party. There is
judicial notice (for example of what time nautical twilight
begins and ends on a given day of the year); there are legis-
lative facts; there are stipulated facts. But there was no factu-
al basis of any sort, in the trial record or elsewhere, for the
No. 14-1419 7
judge’s finding that Owens knew Nelson, let alone knew or
cared that he was a drug dealer. The judge made it up.
In its brief in our court the state argues that the judge’s
determination that Owens knew Nelson and knew him to be
a drug dealer, etc. was a reasonable inference from Owens’
having killed him. The argument is that the judge, having
(without bothering to say so) concluded that Owens had
killed Nelson, was merely remarking that it was a drug-
related crime. And indeed it is highly plausible that Nelson
was murdered by someone in the drug business. But the crit-
ical issue in the case was whether Owens had murdered Nel-
son. The murderer’s motive could not have been “the real
issue” in the case—as the state keeps stressing, motive is not
an element of the crime of murder or of the prosecution’s
case and was not addressed by Owens’ trial counsel either,
and the prosecution had made no effort to prove that Owens
had known Nelson or was himself in the drug business. If as
the judge thought and is plausible in light of what was
found on Nelson’s person the murderer was a drug dealer,
there’s a good chance Owens was not the murderer.
The state mistakenly characterizes Owens as contending
that “inferring a defendant’s motive entitles petitioner to a
new trial if that inference is not fully supported by the evi-
dence at trial.” Had the judge said that he’d found the de-
fendant guilty beyond a reasonable doubt on the basis of the
evidence presented at the trial, and merely added that “by
the way my guess is that Owens knew Nelson and killed
him for reasons related to their both being drug dealers,”
Owens would have no case, because the judge’s observation
would not have been the basis of Owens’ conviction. In fact
the evidence of Owens’ guilt was as we’ve seen far from
8 No. 14-1419
conclusive, yet its uncertainty did not engage the judge’s at-
tention. The judge appears to have been thoroughly con-
fused—and likewise the state when it argues in its brief in
our court that the judge’s inference that Owens killed Nelson
for drug-related reasons “was arguably a reasonable infer-
ence from the” fact that “Nelson had forty bags of crack co-
caine on him at the time of the attack.” Nelson’s being a
drug dealer could not by itself have supported any such in-
ference about the identity of his murderer unless more had
been known about the murderer than the prosecution was
able to establish.
Nonetheless, to repeat, we can assume that if the evi-
dence of Owens’ guilt had been overwhelming, the judge’s
conjecture that Owens knew Nelson and knew him to be a
drug dealer and that Owens was (as the judge’s comment
implied) himself involved in the drug trade (why else would
he want to kill Nelson?) could be disregarded as goofy but
harmless. But evidence of Owens’ guilt was not overwhelm-
ing. Had it been, it is unlikely that the judge would have de-
scribed Owens’ supposed (but by only the judge) knowledge
of Nelson’s involvement in the drug business as “the real
issue” in the case. What may have made it the “real issue” to
the judge was the scantiness of the actual evidence of Ow-
ens’ guilt.
It remains to consider whether the judge’s blunder satis-
fies the test for whether a state court’s error, though of con-
stitutional magnitude (in this case, because the judge based a
verdict of guilty on ungrounded conjecture), can escape be-
ing deemed “harmless” when challenged in a federal habeas
corpus proceeding. The test “is whether the error had sub-
stantial and injurious effect or influence in determining the
No. 14-1419 9
jury’s verdict. Under this standard, habeas [corpus] petition-
ers may obtain plenary review of their constitutional claims,
but they are not entitled to habeas [corpus] relief based on
trial error unless they can establish that it resulted in actual
prejudice.” Brecht v. Abrahamson, 507 U.S. 619, 637 (1993) (ci-
tations and internal quotation marks omitted). Owens has
satisfied this standard. Given that the entire case pivoted on
two shaky eyewitness identifications, Owens might well
have been acquitted had the judge not mistakenly believed
that Owens had known Nelson to be a drug dealer and
killed him because of it. As we explained in Jones v. Basinger,
635 F.3d 1030, 1053–54 (7th Cir. 2011), the question that
Brecht requires us to answer is not whether a reasonable trier
of fact could have rendered the verdict that it (in this case
he) did, but whether the trier of fact committed an error that
had a substantial malign influence on the verdict. The trial
judge’s singling out as the only explanation for the verdict a
“fact” having no evidentiary support, and declaring it the
“real issue” in resolving the case, had to have had such an
influence.
We are mindful that only clearly established violations of
a defendant’s constitutional rights permit us to reverse a
state court decision challenged in a federal habeas corpus
proceeding. Nevada v. Jackson, 133 S. Ct. 1990 (2013). But
there’s no question that the right to have one’s guilt or inno-
cence adjudicated on the basis of evidence introduced at trial
satisfies that exacting standard. That is clear from the Su-
preme Court decisions in Holbrook, Taylor, and Estelle, and
our own Garcia and Moore decisions, from all of which we
quoted earlier. It’s true that we know of no case identical to
this one—unsurprisingly, given the combination of weak
proof with a verdict based on groundless conjecture. But
10 No. 14-1419
identity can’t be required. The Supreme Court has made
clear in the cases we’ve cited and quoted from that a judge
or a jury may not convict a person on the basis of a belief
that has no evidentiary basis whatsoever. Just imagine that
the judge in our case had said “I know there’s no evidence of
guilt, but I also know that prosecutors in the City of Mark-
ham never prosecute an innocent person.” The defendant
would be entitled to relief in a habeas corpus proceeding
even though that precise statement had never been uttered
by a judge before.
And so we reverse the judgment denying Owens relief
and give the state 120 days in which to decide whether to
retry him. If it does not decide within that period to retry
him, he must be released from prison.