Affirmed as Modified and Majority and Concurring Opinions filed August 27,
2015.
In The
Fourteenth Court of Appeals
NO. 14-14-00142-CR
ROGELIO AVILES-BARROSO, Appellant
V.
THE STATE OF TEXAS, Appellee
On Appeal from the 337th District Court
Harris County, Texas
Trial Court Cause No. 1364839
CONCURRING OPINION
In determining whether a voice identification is reliable, binding precedent
requires this court to consider the factors articulated in Neil v. Biggers.1 These
1
See 409 U.S. 188, 199–200, 93 S.Ct. 375, 382, 34 L.Ed.2d 401 (1972); Williams v.
State, 850 S.W.2d 784, 786–87 (Tex. App.—Houston [14th Dist.] 1993) (holding that indicia of
reliability in a voice identification include the witness’s opportunity to hear, the degree of
attention, the accuracy of the description, the level of certainty, and the time between the event
and the confrontation), aff’d on other grounds by, 895 S.W.2d 363 (Tex. Crim. App. 1994).
factors are: (1) the witness’s opportunity to hear the speaker at the time of the
event, (2) the witness’s degree of attention, (3) the accuracy of the witness’s prior
description of the speaker’s voice, (4) the level of certainty demonstrated by the
witness at the confrontation, and (5) the length of time between the event and the
confrontation.2 Applying these factors in a Biggers analysis, the majority
concludes that Diana’s voice identification is reliable. The precedent applying the
Biggers test supports that result. Because we are bound to apply the Biggers test, I
concur in the court’s judgment, but I write separately to suggest that it is time for
courts to reconsider the Biggers factors as applied to voice identification.
In the forty-three years since the Supreme Court of the United States
articulated the Biggers factors, scientists have been studying whether these factors
accurately predict the reliability of a witness’s identification. The findings raise
concerns. Studies are ongoing, but the research results in hand tend to undercut
confidence that the Biggers factors are truly indicia of reliability. The scientific
literature suggests that though some of the Biggers factors relate to the reliability
of a witness’s identification, others do not. Empirical research seems to be
revealing that some of the factors may not be good indicators of reliability.
Not surprisingly, scientists have confirmed that the opportunity to view or
hear a suspect and the length of time between the event and the confrontation
affect the reliability of a witness’s identification.3 But, unexpectedly, some
research has shown that a witness’s confidence in the identification does not
impact reliability.4 Likewise, some research seems to show that the relationship
2
See Biggers, 409 U.S. at 199–200, 93 S.Ct. at 382.
3
See Cindy E. Laub et al., Can the Courts Tell an Ear from an Eye? Legal Approaches
to Voice Identification Evidence, 37 LAW & PSYCHOL. REV. 119, 123 (2013).
4
Laub, supra note 3, at 124 & 142; Jason A. Cantone, Do you Hear What I Hear?, 17
2
between the characteristics of the witness’s verbal description of an individual and
the characteristics of the person identified does not affect reliability.5 If the
research results are valid, the lack of correlation undermines the validity of these
Biggers factors.
Questions about the efficacy of the Biggers test are more pronounced in the
context of a voice identification like the identification Diana made in today’s case.6
Research indicates that vocal identifications are less accurate than eyewitness
identifications, potentially because the human brain processes faces for later
recognition and speech for meaning.7 The fundamental differences between the
way the human brain processes faces and the way the human brain processes
speech may mean there are differences between the factors that indicate a reliable
visual identification and those that indicate a reliable vocal identification.8
Researchers have suggested, for example, that while a witness’s opportunity
to hear the speaker at the time of the event is important to the reliability of a voice
identification just as the opportunity to view the actor’s face is important to a
visual identification, the accuracy of a voice identification also depends on
similarities between the voice at the time of the event and the voice at the time of
identification.9 Research shows that an earwitness’s accuracy in identifying a
voice increases when the tone and emotion in the identified voice sample match the
tone and emotion of the voice during the event.10 Similarly, while research
TEX. WESLEYAN L. REV. 123, 130–31 (2011).
5
See Laub, supra note 4, at 124.
6
See id. at 120.
7
Id. at 124–25.
8
See id. at 123.
9
See id.
10
See Cantone, supra note 3, at 128.
3
suggests the length of time between hearing the voice and identifying it affects a
witness’s accuracy just as the length of time between the event and a witness’s
visual identification affects accuracy, the effect of the length of the retention
interval is different in vocal identifications.11 Researchers have suggested that with
unfamiliar voices, vocal identifications can deteriorate significantly within weeks
of the event.12 In one study, subjects asked to recall a voice had only a nine
percent accuracy rate after three weeks.13 It seems that the “reliability” window
with respect to the interval between the event and the confrontation may be much
narrower for voice identification.
Cross-lingual voice identification triggers additional concerns because this
species of voice identification implicates additional considerations. Although
existing studies appear to be somewhat in conflict, some research suggests that
people are more likely to hear the voices of speakers in a foreign or accented
language as homogeneous or similar, but are able to perceive diversity or
distinctive forms of speech among speakers of their own regional area.14
Anecdotally speaking, native Texans might easily distinguish among the voices of
other native Texans but might have difficulty distinguishing among speakers with a
Boston accent based on voice alone. Researchers and academicians alike warn of
the dangers inherent in cross-lingual voice identification.15
Though appellant’s expert witnesses discussed some of these considerations
11
Laub, supra note 3, at 124.
12
See Cantone, supra note 3, at 130; A. Daniel Yarmey, Earwitness Speaker
Identification, 1 PSYCHOL. PUB. POL’Y & L. 792, 805 (1995).
13
Cantone, supra note 3, at 130; Yarmey, supra note 11, at 805.
14
See Yarmey, supra note 11, at 798.
15
See Gary Edmond et al., ‘Mere Guesswork’: Cross-Lingual Voice Comparisons and
the Jury, 33 SYDNEY L. REV. 395, 405–08 (2011).
4
during the motion-to-suppress hearing, appellant did not fully develop the potential
problems in Diana’s voice identification. Appellant cross-examined investigator
Kerry Gillie about the twenty-year gap between the event and the confrontation,
and Dr. Philip Lyons noted that there are difficulties identifying a speaker who
spoke Spanish with a different dialect. Dr. Lyons testified, however, that he was
unaware of any studies that looked at dialects rather than identifying speakers of
foreign languages. In fairness, these are emerging issues in science as well as law.
The empirical research and scholarship are still developing, and the law may not be
keeping pace with the science.
Given the changing landscape, we must continue to ask whether the Biggers
factors are the best measure of reliability for voice identification. What is clear
from both scientific research and human experience is that familiarity plays a big
role in reliable voice identification. So, when the voice being identified is an
unfamiliar one, greater scrutiny is required for reliable voice identification, and
greater scrutiny still when an earwitness is not familiar with either the speaker’s
voice or the speaker’s dialect. Given the unfolding research in cross-lingual voice
identification, we should question whether these elements should be given greater
weight in a Biggers analysis and whether additional enumerated factors should be
added to the Biggers test when it is applied to voice identification.
To ensure fair proceedings, courts should take these considerations into
account in making reliability determinations, especially when the identification is
based on voice alone, unaccompanied by visual or other forms of identification.
Though the witness in today’s case did not get a look at the speaker’s entire face
during the criminal episode, the witness was able to recall and identify facial
features. Diana’s recognition of appellant’s facial features, though not powerful
visual-identification evidence, strengthens, to some degree, the identification.
5
The gaps in science and law, known and unknown, tell us that the Biggers
test may be falling short as a measure of reliability for voice identification. Now is
the time to consider what steps can be taken to improve reliability testing in voice-
identification cases.
Even though the scientific literature calls into question the validity of the
Biggers factors, this court is bound to consider them.16 Appellant did present
limited testimony about a witness’s inability to remember a voice over time. But,
the record also reveals the witness had an opportunity to hear the speaker’s voice
during the criminal episode and was keenly focused on his voice; the witness
provided a description of the Spanish dialect she detected in the speaker’s voice;
and she displayed unshakeable confidence in her identification. Under the
precedent applying the Biggers test, these four factors support the trial court’s
admission of the identification evidence.17 So, the result the majority reaches
applying the Biggers test is sustainable under current law as is the majority’s
conclusion that the trial court did not err in admitting the witness’s voice
identification into evidence, based on the evidence presented to the trial court. For
this reason, I respectfully concur in the judgment.
/s/ Kem Thompson Frost
Chief Justice
Panel consists of Chief Justice Frost and Justices Boyce and McCally. (Boyce, J.,
Majority).
Publish — TEX. R. APP. P. 47.2(b).
16
See Williams, 850 S.W.2d at 786–87.
17
See Davis v. State, 180 S.W.3d 277, 283–85 (Tex. App.—Texarkana 2005, no pet.).
6