Opinion
SC 20727
07-16-2024
Edgar TATUM v. COMMISSIONER OF CORRECTION
Kara E. Moreau, assigned counsel, with whom was Emily C. Kaas-Mansfield, assigned counsel, for the appellant (petitioner). James A. Killen, senior assistant state’s attorney, for the appellee (respondent). Lisa J. Steele filed a brief for the Connecticut Criminal Defense Lawyers Association as amicus curiae. Robert J. Meredith filed a brief for the Innocence Project, Inc., et al. as amici curiae. Robinson, C. J., and McDonald, D’Auria, Mullins, Ecker, Dannehy and Seeley, Js.
Kara E. Moreau, assigned counsel, with whom was Emily C. Kaas-Mansfield, assigned counsel, for the appellant (petitioner).
James A. Killen, senior assistant state’s attorney, for the appellee (respondent).
Lisa J. Steele filed a brief for the Connecticut Criminal Defense Lawyers Association as amicus curiae.
Robert J. Meredith filed a brief for the Innocence Project, Inc., et al. as amici curiae.
Robinson, C. J., and McDonald, D’Auria, Mullins, Ecker, Dannehy and Seeley, Js.
McDonald, J.
736"[M]istaken eyewitness identification testimony is by far the leading cause of wrongful convictions."737 State v. Guilbert, 306 Conn. 218, 249–50, 49 A.3d 705 (2012). Recognizing the developments in the cognitive science of eyewitness identification, this court has recently established new rules for cases in which eyewitness identification evidence is proffered. Specifically, in Guilbert, we determined that "expert testimony on eyewitness identification is admissible upon a determination by the trial court that the expert is qualified and the proffered testimony is relevant and will aid the jury." Id., at 226, 49 A.3d 705. In doing so, we overruled earlier decisions from this court, which held that the factors affecting eyewitness identification were within the knowledge of an average juror. See id., at 226, 229, 251–53, 49 A.3d 705. We reasoned that our prior case law was "out of step with the widespread judicial recognition that eyewitness identifications are potentially unreliable in a variety of ways unknown to the average juror. This [broad-based] judicial recognition tracks a near perfect scientific consensus. The extensive and comprehensive scientific research, as reflected in hundreds of peer reviewed studies and meta-analyses, convincingly demonstrates the fallibility of eyewitness identification testimony and pinpoints an array of variables that are most likely to lead to a mistaken identification." (Footnotes omitted.) Id., at 234–36, 49 A.3d 705. We also noted that a trial court retains discretion to provide "focused and informative" jury instructions on the fallibility of eyewitness identification evidence. Id., at 257–58, 49 A.3d 705. Four years later, in State v. Dickson, 322 Conn. 410, 141 A.3d 810 (2016), cert. denied, 582 U.S. 922, 137 S. Ct. 2263, 198 L. Ed. 2d 713 (2017), we further developed protections against inherently suggestive identifications. In doing so, we overruled this court’s 1991 holding in this petitioner’s direct appeal related to a first-time, incourt cross-racial eyewitness identification. See id., at 434–36, 141 A.3d 810 (overruling in part State v. Tatum, 219 Conn. 721, 595 A.2d 322 (1991)). We concluded that "any [first-time] in-court identification by 738a witness who would have been unable to reliably identify the [petitioner] in a non-suggestive out-of-court procedure constitutes a procedural due process violation." (Emphasis in original.) State v. Dickson, supra, at 426 n.11, 141 A.3d 810. The sole issue in this certified appeal is whether the principles this court set forth in Guilbert and Dickson apply retroactively to the petitioner’s case on collateral review. We conclude that the principles articulated in Dickson do. Accordingly, we reverse in part the judgment of the Appellate Court.
In 1990, following a jury trial, the petitioner, Edgar Tatum, was convicted of murder in connection with the shooting death of the victim and sentenced to sixty years of incarceration. The state’s case against the petitioner was largely based on two cross-racial eyewitness identifications of the petitioner. Both eyewitnesses had previously identified the same person as the shooter, someone other than the petitioner. The eyewitnesses, who were both white, recanted these earlier identifications and, more than one year after the shooting, at the petitioner’s probable cause hearing, identified the petitioner, who was the only Black man seated at defense counsel’s table, as the shooter. Significantly, both eyewitnesses were heavy drug users, one admitting to using narcotics every day and the other admitting to "free-basing cocaine" on the evening of the shooting. The petitioner appealed his conviction to this court, challenging, among other things, the trial court’s admission of an unduly suggestive in-court identification and the eyewitness identification instructions given to the jury. See State v. Tatum, supra, 219 Conn. at 723, 595 A.2d 322. This court rejected his claims and upheld the judgment of conviction. Id., at 723, 742, 595 A.2d 322.
The petitioner has since filed four petitions for a writ of habeas corpus that are not relevant to this appeal. His fifth habeas petition, which is the subject of this appeal, was filed in 2016. In count six of the operative, 739amended petition, the petitioner alleged that the admission of unduly suggestive and unreliable eyewitness identification evidence in his underlying criminal case violated his due process rights under the fourteenth amendment to the federal constitution and article first, §§ 8 and 9, of the state constitution. He also argued that the jury instructions provided by the criminal trial court were insufficient to educate jurors about certain factors that could adversely impact eyewitness identification. Finally, he argued that this court’s decisions in Guilbert and Dickson should be retroactively applied to his case.
In count seven, the petitioner argued that advances in the science of eyewitness identification since his conviction highlight the unreliability of the eyewitness identifications that occurred in his criminal case and call into question the validity of his conviction. The habeas court interpreted this claim as a claim of actual innocence. In discussing the claim, the court explained that, "even giving the petitioner the benefit of the doubt the law requires, he is not actually claiming that there is ‘new’ evidence, as in a previously undiscovered witness, an unknown video of the incident, or bodily fluids not previously subject to DNA testing." The court stated: "What the claim really amounts to is that subsequent developments in the science of eyewitness identification have changed the information and instructions a jury can be given in a criminal trial, and, if the jurors in the petitioner’s trial were allowed to apply the ‘new’ science and instructions to the same ‘old’ evidence presented at the petitioner’s trial, they may have viewed the testimony of the eyewitnesses who identified the petitioner differently and come to a different conclusion."
The respondent, the Commissioner of Correction, moved to dismiss the operative petition in 2018. The habeas court granted the respondent’s motion to dismiss740 as to counts one (ineffective assistance of trial counsel), two (ineffective assistance of appellate counsel), three (ineffective assistance of first habeas counsel), six (due process), and seven (newly discovered evidence). As to counts six and seven, the habeas court construed count seven in conjunction with count six and explained that the petitioner already had litigated the identification procedures in his direct appeal and that the doctrine of res judicata prohibited "the petitioner from being able to relitigate this issue by changing the facts to focus on the identification procedures used in connection with witness [Tracy] LeVasseur because neither the grounds nor the requested relief is any different [from] the issue raised on appeal." The court emphasized that "the petitioner has not alleged a single new ‘fact’ related to his case." The court then went on to conclude that nothing in the Guilbert or Dickson decisions indicates that they were to be retroactively applied or intended to provide an avenue for collateral relief.
The habeas court denied the respondent’s motion to dismiss as to counts four (ineffective assistance of second habeas counsel) and five (ineffective assistance of third habeas counsel). The court held a hearing on those two claims, after which the parties filed posttrial briefs. The habeas court ultimately dismissed count four and denied count five of the habeas petition. The petitioner thereafter filed a petition for certification to appeal, which was granted by the habeas court. On appeal to the Appellate Court, the petitioner claimed, among other things, that the habeas court incorrectly determined that this court’s decisions in Guilbert and Dickson could not be applied retroactively to the identification claims raised in counts six and seven of the habeas petition. See Tatum v. Commissioner of Correction, 211 Conn. App. 42, 44, 272 A.3d 218 (2022). The 741Appellate Court disagreed and affirmed the judgment of the habeas court. Id., at 44, 76, 272 A.3d 218. We granted the petitioner’s petition for certification to appeal, limited to the following issue: "Did the Appellate Court incorrectly conclude that the habeas court had properly dismissed counts six and seven of the petitioner’s operative, amended habeas petition on the ground that State v. Dickson, [supra, 322 Conn. 410, 141 A.3d 810], and State v. Guilbert, [supra, 306 Conn. 218, 49 A.3d 705], both of which overruled this court’s rationale and holding Regarding in-court identifications in the petitioner’s direct appeal; see State v. Tatum, [supra, 219 Conn. 721, 595 A.2d 322]; did not apply retroactively to the petitioner’s case on collateral review?" Tatum v. Commissioner of Correction, 343 Conn. 932, 276 A.3d 975 (2022).
On appeal to this court, the petitioner claims that the Appellate Court should not have upheld the habeas court’s dismissal of counts six and seven of his petition on the basis that Guilbert and Dickson do not apply retroactively. He contends that both Guilbert and Dickson announced watershed rules of criminal procedure and, as such, should apply retroactively. Alternatively, even if this court were to conclude that Guilbert and Dickson do not apply retroactively to all criminal defendants and petitioners, the petitioner contends that justice requires that Guilbert and Dickson apply retroactively to his case because each case overruled the specific holdings in his direct appeal. The respondent contends that neither Guilbert nor Dickson applies retroactively. The respondent points to footnote 34 of Dickson, which he claims stated that the new constitutional rule announced in that case did not apply retroactively on collateral review. See State v. Dickson, supra, 322 Conn. at 451 n.34, 141 A.3d 810.
We begin with a discussion of this court’s recent eyewitness identification cases. First, in Guilbert, this 742court held, for the first time, that, because certain factors that bear on the reliability of eyewitness identifications are not within the knowledge of the average juror, expert testimony on those factors does not invade the province of the jury and is admissible. State v. Guilbert, supra, 306 Conn. at 226, 234–37, 251–52, 49 A.3d 705. We emphasized that "eyewitness identifications are potentially unreliable in a variety of ways unknown to the average juror," and this "recognition tracks a near perfect scientific consensus." Id., at 234–35, 49 A.3d 705. As a result, we recognized that the methods typically used to alert juries to the fallibility of eyewitness identifications— cross-examination, closing argument, and generalized jury instructions—often are not sufficient to alert the jury to the factors affecting the reliability of eyewitness identifications. Id., at 243, 49 A.3d 705. We also stated that a trial court retains the discretion to provide "focused and informative jury instructions on the fallibility of eyewitness identification evidence" that "reflect the findings and conclusions of the relevant scientific literature pertaining to the particular variable or variables at issue in the case …. " (Citations omitted; footnote omitted.) Id., at 257–58, 49 A.3d 705. Significantly, we noted that "broad, generalized instructions on eyewitness identifications, such as those previously approved by this court in [the petitioner’s direct appeal] … do not suffice." (Citations omitted.) Id., at 258, 49 A.3d 705.
We next had occasion to consider eyewitness identification evidence in Dickson. We held that, contrary to our prior case law on the topic, "in cases in which identity is an issue, in-court identifications that are not preceded by a successful identification in a nonsuggestive identification procedure implicate due process principles and, therefore, must be prescreened by the trial court." (Footnote omitted.) State v. Dickson, supra, 322 Conn. at 415, 141 A.3d 810. We reasoned that "we are hard-pressed to imagine how there could be a more suggestive identification procedure than placing a witness on the stand in open court, 743confronting the witness with the person whom the state has accused of committing the crime, and then asking the witness if he can identify the person who committed the crime." (Emphasis in original.) Id., at 423, 141 A.3d 810. We again revisited the petitioner’s direct appeal and concluded that our holding in that appeal that "it was ‘necessary’ for the state to present a [firsttime] in-court identification of the [petitioner] at the probable cause hearing must be overruled." Id., at 435–36, 141 A.3d 810. We explained that "[w]e simply can perceive no reason why the state cannot attempt to obtain an identification using a lineup or photographic array before asking an eyewitness to identify the [petitioner] in court. Although the state is not constitutionally required to do so, it would be absurd to conclude that the state can simply decline to conduct a nonsuggestive procedure and then claim that its own conduct rendered a [first-time] in-court identification necessary, thereby curing it of any constitutional infirmity." (Emphasis omitted.) Id., at 436, 141 A.3d 810. In short, "[t]he state is not entitled to conduct an unfair procedure merely because a fair procedure failed to produce the desired result." Id.
Finally, in State v. Harris, 330 Conn. 91, 191 A.3d 119 (2018), we explained that the due process provision of article first, § 8, of our state constitution affords greater protection than the federal constitution with respect to the admissibility of eyewitness identification testimony. See id., at 114–15, 131, 191 A.3d 119. In Harris, the defendant challenged the admission of an identification that was made while he was being arraigned in court on an unrelated case. See id., at 98–99, 191 A.3d 119. Although we concluded that the identification procedure was "overly suggestive by any measure" because "none of [the other Black, male] custodial arraignees was sufficiently similar to the defendant in height, weight and age," we also concluded that the identification was reliable in light of the circumstances of the case. Id., at 107–108, 191 A.3d 119.
[1–3] 744We now turn to the present case. The question of whether the principles this court set forth in Guilbert and Dickson apply retroactively to the petitioner’s case on collateral review is a question of law over which our review is plenary. See, e.g., Duperry v. Solnit, 261 Conn. 309, 318, 803 A.2d 287 (2002). In Teague v. Lane, 489 U.S. 288, 109 S. Ct. 1060, 103 L. Ed. 2d 334 (1989), the United States Supreme Court set forth the framework for evaluating whether a new rule applies retroactively on collateral review under the federal constitution. See id., at 299–314, 109 S. Ct. 1060 (plurality opinion); see also Thiersaint v. Commissioner of Correction, 316 Conn. 89, 112–13, 111 A.3d 829 (2015) (adopting Teague framework). Under Teague, a court "must [first] ascertain the legal landscape as it" existed at the time the petitioner’s conviction became final and "ask whether the [United States] [constitution, as interpreted by the precedent then existing, compels the rule …. That is, the court must decide whether the rule is actually new." (Citation omitted; internal quotation marks omitted.) Beard v. Banks, 542 U.S. 406, 411, 124 S. Ct. 2504, 159 L. Ed. 2d 494 (2004). A constitutional rule is "new" for purposes of Teague "if the result was not dictated by precedent existing at the time the [petitioner’s] conviction became final." (Internal quotation marks omitted.) Thiersaint v. Commissioner of Correction, supra, at 103, 111 A.3d 829.
[4] The respondent concedes that Guilbert and Dickson created "new" rules, within the meaning of Teague. As a result, we must decide whether these "new" rules apply retroactively. "With two exceptions, a new rule will not apply retroactively to cases on collateral review. … First, if the new rule is substantive, that is, if the rule places certain kinds of primary, private conduct beyond the power of the criminal lawmaking authority to proscribe … it must apply retroactively. Such rules apply retroactively because they necessarily carry a significant risk that a [petitioner] stands convicted745 of an act that the law does not make criminal or faces a punishment that the law cannot impose [on] him." (Citations omitted; internal quotation marks omitted.) Casiano v. Commissioner of Correction, 317 Conn. 52, 62–63, 115 A.3d 1031 (2015), cert. denied sub nom. Semple v. Casiano, 577 U.S. 1202, 136 S. Ct. 1364, 194 L. Ed. 2d 376 (2016).
[5] "Second, if the new rule is procedural, it applies retroactively if it is a watershed [rule] of criminal procedure … implicit in the concept of ordered liberty … meaning that it implicat[es] the fundamental fairness and accuracy of [a] criminal proceeding. … Watershed rules of criminal procedure include those that raise the possibility that someone convicted with use of the invalidated procedure might have been acquitted otherwise. … The United States Supreme Court has narrowly construed [the] second exception and, in the [more than thirty-five] years since Teague was decided, has [never] conclude[d] that a new rule qualifie[d] as watershed." (Citations omitted; internal quotation marks omitted.) Id., at 63, 115 A.3d 1031. Indeed, the United States Supreme Court has recently abolished the watershed rule. See Edwards v. Vannoy, 593 U.S. 255, 272, 141 S. Ct. 1547, 209 L. Ed. 2d 651 (2021). The court reasoned that "[c]ontinuing to articulate a theoretical exception that never actually applies in practice offers false hope to defendants, distorts the law, misleads judges, and wastes the 746resources of defense counsel, prosecutors, and courts. Moreover, no one can reasonably rely on an exception that is [nonexistent] in practice, so no reliance interests can be affected by forthrightly acknowledging reality. It is time—probably long past time—to make explicit what has become increasingly apparent to bench and bar over the last [thirty-five] years: New procedural rules do not apply retroactively on federal collateral review. The watershed exception is moribund. It must be regarded as retaining no vitality." (Internal quotation marks omitted.) Id.
Since Edwards, many scholars have advocated for the reimagining of how courts approach the retroactivity issue. See, e.g., J. Ho, Note, "Finality, Comity, and Retroactivity in Criminal Procedure: Reimagining the Teague Doctrine After Edwards v. Vannoy," 73 Stan. L. Rev. 1551, 1600 (2021) ("[i]n light of the weighty remedial interests—not just in accuracy but in human dignity and judicial integrity—a revised retroactivity framework should be more generous about granting retroactivity remedies for violations of constitutional rights"); T. Simkovic, Note, "Ramos Retroactivity and the False Promise of Teague v. Lane," 76 U. Miami L. Rev. 825, 830 (2022) ("[g]iven that Teagues exception for watershed rules is now extinct, the [United States Supreme] Court should rethink its entire retroactivity framework for new rules of criminal law on habeas review").
In the present case, the petitioner does not contend that the new rules articulated in Guilbert and Dickson are substantive. Rather, the petitioner argues that these new procedural rules are watershed ones. As we discussed, new procedural rules no longer apply retroactively on federal collateral review. See id. Nevertheless, we have explained that, although "federal decisions applying Teague may be instructive, this court will not be bound by those decisions in any particular case … but will conduct an independent analysis and application of Teague." Thiersaint v. Commissioner of Correction, supra, 316 Conn. at 113, 111 A.3d 829. We take this opportunity to clarify the viability of the watershed exception in Connecticut in light of the United States Supreme Court’s decision to abolish the watershed rule in Edwards.
We have applied the Teague framework "more liberally than the United States Supreme Court [might] otherwise apply it …. " (Internal quotation marks omitted.) Casiano v. Commissioner of Correction, supra, 317 Conn. at 64, 115 A.3d 1031; see also, e.g., Rhoades v. State, 149 Idaho 130, 139, 233 P.3d 61 (2010) (because comity concerns do not apply to state court’s review of state convictions, Idaho courts are "not required to blindly follow [the United States Supreme Court’s] view of … whether a new rule is a watershed rule"), cert. denied, 562 U.S. 1258, 131 S. Ct. 1571, 179 L. Ed. 2d 477 (2011); 747Stote v. Mares, 335 P.3d 487, 504 (Wyo. 2014) (applying Teague more broadly than United States Supreme Court when "a particular state interest is better served by a broader retroactivity ruling").
For example, we have applied the Teague analysis and concluded that the United States Supreme Court’s decision in Miller v. Alabama, 567 U.S. 460, 132 S. Ct. 2455, 183 L. Ed. 2d 407 (2012), applied retroactively on collateral review. Casiano v. Commissioner of Correction, supra, 317 Conn. at 62, 69, 71, 115 A.3d 1031. Miller forbade mandatory life without parole sentences for juvenile offenders. Miller v. Alabama, supra, at 465, 132 S.Ct. 2455. In Casiano, we reasoned that Miller created a new watershed rule of criminal procedure; Casiano v. Commissioner of Correction, supra, at 62, 69, 115 A.3d 1031; because, among other things, "the individualized sentencing prescribed by Miller is central to an accurate determination … that the sentence imposed is a proportionate one." (Citation omitted; internal quotation marks omitted.) Id., at 70, 115 A.3d 1031. Several months later, the United States Supreme Court similarly deemed Miller’s ruling retroactive in Montgomery v. Louisiana, 577 U.S. 190, 206, 136 S. Ct. 718, 193 L. Ed. 2d 599 (2016). The Supreme Court, however, based its decision on a different premise, namely, that Miller had created a new substantive rule, thus bypassing Teague’s watershed analysis. See id., at 206, 208–209, 136 S. Ct. 718. Given that we are not bound by the United States Supreme Court’s application of Teague, our conclusion in Casiano remains binding in Connecticut. See K. Kurland, "With Unanimity and Justice for All: The Case for Retroactive Application, of the Unanimous Jury Verdict Requirement," 17 Nw. J.L. & Soc. Policy 49, 75 (2021).
[6–8] Although Teague’s watershed rule may be "moribund" in the federal courts; Edwards v. Vannoy, supra, 593 U.S. at 272, 141 S.Ct. 1547; we conclude that it has continued vitality in Connecticut. The United States Supreme Court has explained that "the Teague rule of nonretroactivity was 748fashioned to achieve the goals of federal habeas while minimizing federal intrusion into state criminal proceedings. It was intended to limit the authority of federal courts to overturn state convictions—not to limit a state court’s authority to grant relief for violations of new rules of constitutional law when reviewing its own [s]tate’s convictions." Danforth v. Minnesota, 552 U.S. 264, 280–81, 128 S. Ct. 1029, 169 L. Ed. 2d 859 (2008). In other words, the particular concerns that serve to limit habeas relief "are unique to federal habeas review of state convictions." (Emphasis in original.) Id., at 279, 128 S. Ct. 1029. Whereas federal habeas review for state prisoners risks "rendering] the actions of state courts a serious disrespect"; Schneckloth v. Bustamante, 412 U.S. 218, 263, 93 S. Ct. 2041, 36 L. Ed. 2d 854 (1973) (Powell, J., concurring); state postconviction proceedings "merely [reflect] and [confirm] the [state] courts’ own inherent and discretionary power, firmly established in English practice long before the foundation of our [Republic, to set aside a judgment whose enforcement would work inequity." (Internal quotation marks omitted.) Plaut v. Spendthrift Farm, Inc., 514 U.S. 211, 233–34, 115 S. Ct. 1447, 131 L. Ed. 2d 328 (1995). Accordingly, the federalism rationales for Teague "simply do not apply" to state postconviction proceedings. State v. Precise, 129 N.J. 451, 475, 609 A.2d 1280 (1992). Whereas concerns about the relationship between state and federal courts warrant caution in the federal habeas context, these same concerns suggest that states should be particularly willing to provide fulsome postconviction procedures. See, e.g., Case v. Nebraska, 381 U.S. 336, 338–40, 85 S. Ct. 1486, 14 L. Ed. 2d 422 (1965) (Clark, J., concurring); id., at 344–47, 85 S. Ct. 1486 (Brennan, J, concurring).
[9] Considerations of finality are certainly very important in state habeas proceedings, but they are somewhat less important in state postconviction proceedings as compared to federal habeas proceedings, as the federal 749proceedings typically occur last and, therefore, must take into account the finality of the state proceedings. See, e.g., C. Lasch, "The Future of Teague Retroactivity, or ‘Redressability,’ After Danforth v. Minnesota: Why Lower Courts Should Give Retroactive Effect to New Constitutional Rules of Criminal Procedure in Postconviction Proceedings," 46 Am. Crim. L. Rev. 1, 4–5, 57 (2009). Moreover, there is good reason to conclude that the benefits of retroactivity on collateral review in appropriate cases outweigh finality concerns. The United States Supreme Court has observed that "[t]he finality interest is more at risk" in postconviction proceedings than on direct review and that "the costs and uncertainties of a new trial are greater because more time will have elapsed in most cases." Weaver v. Massachusetts, 582 U.S. 286, 302, 137 S. Ct. 1899, 198 L. Ed. 2d 420 (2017). "For the same reasons, those ‘costs and uncertainties’ are lower in state [postconviction] proceedings than in federal habeas, and place less of a thumb on the scale in favor of nonretroactivity. Furthermore, to the extent finality concerns refer to the finality of a state court decision (as opposed to a state court conviction), they are reduced in state [postconviction] proceedings, which are often the state courts’ first look at a constitutional claim." (Emphasis omitted.) J. Rutledge, "With Great (Writ) Power Comes Great (Writ) Responsibility: A Modified Teague Framework for State Courts," 59 Crim. L. Bull. 480, 494 (2023).
[10, 11] Again, we do not discount the importance of finality; we simply acknowledge that this court’s opportunity to review certain constitutional claims may arise for the first time in the habeas context, and the interest in finality plainly does not automatically outweigh interests in fairness and justice in every circumstance. We do not believe that we should follow the Supreme Court’s lead in Edwards by foreclosing the possibility of the retroactive application of new procedural rules 750in all cases. Cf. State v. Reddick, 351 So. 3d 273, 297 (La. 2022) (Griffin, J., dissenting) ("[t]he imperative to correct past injustices manifest in the deprivation of a constitutionally guaranteed right should not cede to reliance interests and administrative concerns," and courts should "not perpetuate something [they] all know to be wrong only because [they] fear the consequences—and costs—of being right"). As we have explained in a similar context, "in criminal matters, judicial economy must give way to the demand for the truth." State v. McDowell, 242 Conn. 648, 657, 699 A.2d 987 (1997); see, e.g., State v. Ellis, 197 Conn. 436, 471, 497 A.2d 974 (1985) ("the essentially public objectives of the criminal law advise against the uncritical adoption of [res judicata] concepts"). In short, finality "is less relevant in criminal cases [in which] the [preeminent] concern is to reach a correct result and [in which] other considerations peculiar to criminal prosecutions may outweigh the need to avoid repetitive litigation …. " (Internal quotation marks omitted.) State v. Ellis, supra, at 470, 497 A.2d 974.
Accordingly, we continue to see vitality in Teague’s watershed exception. Rather than "blindly follow" the United States Supreme Court’s application of Teague, this court will continue to "independently review cases when applying the Teague standard," including when determining whether a new procedural rule is watershed, notwithstanding the United States Supreme Court’s holding in Edwards v. Vannoy, supra, 593 U.S. at 272, 141 S.Ct. 1547. 751 Rhoades v. State, supra, 149 Idaho at 139, 233 P.3d 61; see also, e.g., Danforth v. State, 761 N.W.2d 493, 500 (Minn. 2009); State v. Mares, supra, 335 P.3d at 504.
State courts have regularly applied the Teague watershed exception, despite the United States Supreme Court’s aversion to doing so. They have done so for a variety of rules: rules protecting the right to counsel; see, e.g., Talley v. State, 371 S.C. 535, 544, 640 S.E.2d 878 (2007) (holding that Alabama v. Shelton, 535 U.S. 654, 122 S. Ct. 1764, 152 L. Ed. 2d 888 (2002), announced watershed rule); see also Alford v. State, 287 Ga. 105, 106–108, 695 S.E.2d 1 (2010) (same); rules protecting the right to trial by jury and the right not to be convicted without proof beyond a reasonable doubt; see, e.g., Powell v. Delaware, 153 A.3d 69, 70, 74, 76 (Del. 2016) (holding that Rauf v. State, 145 A.3d 430 (Del. 2016), announced watershed rule); People v. Beachem, 336 Ill. App. 3d 688, 693–700, 271 Ill.Dec. 67, 784 N.E.2d 285 (2002) (holding that Apprendi v. New Jersey, 530 U.S. 466, 120 S. Ct. 2348, 147 L. Ed. 2d 435 (2000), announced watershed rule), appeal denied, 203 Ill. 2d 552, 273 Ill.Dec. 139, 788 N.E.2d 730, cert. denied, 540 U.S. 897, 124 S. Ct. 243, 157 L. Ed. 2d 177 (2003); rules regarding sentencing in capital cases; see, e.g., State v. Zuniga, 336 N.C. 508, 512, 514, 444 S.E.2d 443 (1994) (recognizing that Mills v. Maryland, 486 U.S. 367, 108 S. Ct. 1860, 100 L. Ed. 2d 384 (1988), and McKoy v. North Carolina, 494 U.S. 433, 110 S. Ct. 1227, 108 L. Ed. 2d 369 (1990), announced watershed rules); rules announced by the United States Supreme Court; see, e.g., Alford v. State, supra, at 106–108, 695 S.E.2d 1; People v. Beachem, supra, at 693–700, 271 Ill.Dec. 67, 784 N.E.2d 285; State v. Zuniga, supra, at 512, 514, 444 S.E.2d 443; Talley v. State, supra, at 544, 640 S.E.2d 878; and rules announced by state high courts. Powell v. Delaware, supra, at 70, 74, 76. They have also done so even when the United States Supreme Court has reached the opposite conclusion. Compare State v. Zuniga, supra, at 512, 514, 444 S.E.2d 443 (recognizing that Mills and McKoy announced watershed rule under state version of Teague), with Beard v. Banks, supra, 542 U.S. at 408, 410, 419–20, 124 S.Ct. 2504 (recognizing that Mills and McKoy did not announce watershed rule under federal Teague standard).
Despite our disagreement with the conclusion reached in Edwards, the case did highlight the overly rigid strictures of the Teague watershed exception. As the United States Supreme Court has explained, "no new rules of criminal procedure can satisfy the watershed exception." Edwards v. Vannoy, supra, 593 U.S. at 271, 141 S.Ct. 1547. One scholar has even described the watershed exception as being so restrictive because "nothing is as important as Gideon [v. Wainwright, 372 U.S. 335, 83 S. Ct. 792, 9 L. Ed. 2d 799 (1963)], so nothing is retroactive." J. Marceau, "Gideon’s Shadow," 122 Yale L.J. 2482, 2488 (2013); cf. Howard v. United States, 374 F.3d 1068, 1081 (11th Cir. 2004) ("[a]t the risk of oversimplification, for purposes of the second Teague exception there are new rules, and then there are new Gideon-extension rules"). In light of the "confused and confusing" history of the retroactivity doctrine; Danforth v. Minnesota, supra, 552 U.S. at 271, 128 S.Ct. 1029; we think it prudent to continue to develop our application of Teague to ensure that its application is not so rigid that it "never actually applies in practice [and] offers [only] false hope to defendants, distorts the law, misleads judges, and wastes the resources of 752defense counsel, prosecutors, and courts." Edwards v. Vannoy, supra, at 272, 141 S.Ct. 1547. We take this opportunity to provide greater clarity regarding the independent review this court undertakes when determining whether a rule applies retroactively on collateral review.
[12] Specifically, in light of Edwards and the admittedly narrow applicability of the watershed exception, we think it necessary to adopt a third exception to the Teague rule of nonretroactivity. See Casiano v. Commissioner of Correction, supra, 317 Conn. at 64, 115 A.3d 1031 ("[w]e … remain free to apply the Teague analysis more liberally than the United States Supreme Court would otherwise apply it [when] a particular state interest is better served by a broader retroactivity ruling" (internal quotation marks omitted)); Thiersaint v. Commissioner of Correction, supra, 316 Conn. at 108, 111 A.3d 829 ("the United States Supreme Court … held in Danforth v. Minnesota, [supra, 552 U.S. at 282, 128 S.Ct. 1029], that the restrictions Teague imposes on the fully retroactive application of new procedural rules are not binding on the states" (internal quotation marks omitted)). We conclude that a new constitutional rule of criminal procedure must be applied 753retroactively on collateral review if the rule was a result of developments in science that persuaded us to reevaluate fundamental principles underlying judicial procedures, the rule significantly improves the accuracy of a conviction, and the petitioner advocated for the rule in the direct proceedings or in an earlier habeas petition.
State courts have adopted a variety of caveats to the Teague standard. See, e.g., J. Rutledge, supra, 59 Crim. L. Bull. 486–87; see also, e.g., id., 486–87 n.55, 487 nn.56–59 (citing cases).
Since Edwards was decided, only four states, namely, Colorado, Louisiana, Mississippi and Oklahoma, have explicitly addressed the future of the watershed exception. See People v. Melendez, 549 P.3d 1028, 1031 (Colo. App. 2024) (acknowledging that, in absence of ruling from Colorado Supreme Court, "the watershed rule remain[ed] embedded in Colorado jurisprudence" (internal quotation marks omitted)); State v. Reddick, supra, 351 So. 3d at 281 (explicitly rejecting Teague's watershed rule); Wess v. State, 348 So. 3d 333, 344 (Miss. App. 2022)) (same); State ex rel. Matloff v. Wallace, 497 P.3d 686, 688–89 (Okla. Crim. App. 2021) (court did not explicitly decide whether it would continue to apply watershed exception but spoke of exception in past tense and predicted that "such a rule is unlikely ever to be announced"), cert. denied sub nom. Parish v. Oklahoma, — U.S. —, 142 S. Ct. 757, 211 L. Ed. 2d 474 (2022); see also, e.g., Aili v. State, 963 N.W.2d 442, 448 n.4 (Minn. 2021) (declining to decide whether watershed exception continues to exist under state law because it would not apply to case at hand).
[13] As one scholar has explained, "for state [postconviction] proceedings to fulfill the traditional role of habeas corpus as the instrument by which due process [can] be insisted [on], they must allow prisoners to litigate the constitutional claims [the prisoners] were prevented from raising before their convictions became final." (Footnote omitted; internal quotation marks omitted.) J. Rutledge, supra, 59 Crim. L. Bull. 497. Courts have also reasoned that the possibility of overruling erroneous precedent may be a component of a meaningful opportunity to present a defense. See, e.g., McCarthan v. Director of Goodwill Industries -Suncoast, Inc., 851 F.3d 1076, 1087 (11th Cir.) (concluding that petitioner "had a meaningful opportunity to present his claim" because he had "the chance to have precedent overruled en banc or by the [United States] Supreme Court"), cert. denied sub nom. McCarthan v. Collins, 583 U.S. 1012, 138 S. Ct. 502, 199 L. Ed. 2d 385 (2017). Because retroactivity under Teague is a threshold question, petitioners advocating for a new constitutional rule of criminal procedure do not enjoy that possibility. "Even if the court wholeheartedly agreed that the [c]onstitution required overruling precedent, [a petitioner] could not receive the benefit of that overruling." J. Rutledge, supra, 498. As a result, scholars have encouraged state courts to "treat [a] petitioner’s first opportunity to raise a constitutional claim as a form of direct review for purposes of that claim." Id., 499. It strikes us as eminently reasonable, then, that a new constitutional rule of criminal procedure be applied retroactively on collateral review when the petitioner had previously raised 754that claim on direct appeal or in an earlier habeas proceeding. Just because a petitioner was ahead of scientific advancements that now call into question the fundamental principles underlying judicial procedures and the accuracy of a criminal conviction does not mean that the petitioner should be precluded from the application of that new rule.
A case from Louisiana highlights the injustice that occurs when a criminal defendant or petitioner is unable to obtain the benefit of a new constitutional rule for which he or she had previously argued. At his second jury trial, the defendant, Corey Miller, was found guilty by a jury vote of ten to two. State v. Miller, 83 So. 3d 178, 182 and nn.1–2 (La. App. 2011), writ denied, 89 So. 3d 1191 (La. 2012), cert. denied, 568 U.S. 1157, 133 S. Ct. 1238, 185 L. Ed. 2d 177 (2013). At the time, the Louisiana constitution permitted a nonunanimous guilty verdict so long as at least ten out of twelve jurors vote in favor of conviction. Id., at 204 and n.10; see also La. Const., art. I, § 17 (A) (2018) ("[a] case for an offense committed prior to January 1, 2019, in which the punishment is necessarily confinement at hard labor shall be tried before a jury of twelve persons, ten of whom must concur to render a verdict"). Miller was sentenced to life in prison without the possibility of parole. Stale v. Miller, supra, 83 So. 3d at 182. Miller claimed, both at the trial and appellate levels, that the nonunanimous verdict violated the federal constitution. See id., at 204. Both the trial court and the Louisiana Court of Appeal rejected this contention. See id., at 205. Miller thereafter sought review from the Louisiana Supreme Court and the United States Supreme Court on the unanimity issue, but both courts denied review. See Miller v. Louisiana, supra, 568 U.S. 1157, 133 S.Ct. 1238; State v. Miller, supra, 89 So. 3d 1191; see also, e.g., Petition for Writ of Certiorari, Miller v. Louisiana, 568 U.S. 1157, 133 S.Ct. 1238.
755Seven years after Miller’s conviction became final, the United States Supreme Court in Ramos v. Louisiana, 590 U.S. 83, 140 S. Ct. 1390, 206 L. Ed. 2d 583 (2020), determined that the conviction of a criminal defendant in state court following a nonunanimous jury verdict violates the federal constitution. See id., at 89–93, 140 S. Ct. 1390. The Supreme Court reasoned that the provision of the Louisiana constitution allowing for nonunanimous jury verdicts in criminal cases was adopted at a constitutional convention that had as its "avowed purpose … the supremacy of the white race"; (internal quotation marks omitted) id., at 87, 140 S. Ct. 1390; and that the provision was adopted "to ensure that [African American] juror service would be meaningless." (Internal quotation marks omitted.) Id., at 88, 140 S. Ct. 1390. Despite being "a Black man who did not match the witness description of the killer" and who was "convicted over the dissent of two Black jurors" in his criminal trial, Miller "remains imprisoned for life and cannot claim the retroactive benefit of the rule for which he advocated at every opportunity." J. Rutledge, supra, 59 Crim. L. Bull. 501; see also State v. Miller, supra, 83 So. 3d at 193. "The idea that like cases should be decided alike is a basic principle of justice. But that principle is sacrificed when those like … Miller receive no relief while others raising the same argument receive new trials simply because of the accident of when the [United States] Supreme Court chose to consider the question. The same is true with equal force when the court that eventually decides the issue is a state’s own high court." (Footnote omitted; internal quotation marks omitted.) J. Rutledge, supra, 502. Accordingly, when a petitioner has previously advocated for a rule in his direct proceedings or in an earlier habeas petition, and scientific advances subsequently persuade this court to reevaluate fundamental principles underlying judicial procedures that calls into question the accuracy of a conviction, we will apply that new constitutional rule retroactively on collateral review.
756This third exception to Teague is similar to the unavailability by exhaustion doctrine Texas courts apply in deciding whether to consider the merits of a petitioner’s habeas petition when the petitioner has previously filed one or more petitions. See Ex parte Hood, 211 S.W.3d 767, 776–77 (Tex. Crim. App.), cert. denied sub nom. Hood v. Texas, 552 U.S. 829, 128 S. Ct. 48, 169 L. Ed. 2d 43 (2007). The unavailability by exhaustion doctrine allows a court to consider a petition if it "is based on binding and directly relevant" precedent "decided after [the petitioner] had exhausted" the claim in a previous proceeding. Ex parte Martinez, 233 S.W.3d 319, 322 (Tex. Crim. App. 2007); see also Ex parte Hood, supra, at 776 ("[i]f we [had] decide[d] an issue adversely to a [petitioner] in a way that contradicts a later legal development, that later legal development constitutes a legal basis that was not presented and could not have been presented at the time [the petitioner’s prior habeas petition was filed]"). The Teague exception we adopt today balances the need to expand the circumstances in which retroactivity will work to prevent injustice with the importance of finality because it does not open the floodgates in a way that would seriously undermine finality. The limiting principles of unavailability by exhaustion applied by Texas courts would also apply to this exception. Namely, the petitioner must actually have raised the claim himself; see Ex parte Hood, supra, at 776; and must have done so in the court that eventually announces the rule. The intervening decision must come from the United States Supreme Court or this court. Cf. id. ("a change in the law under the exhaustion doctrine … must come from a binding authority, i.e. cases from [a state’s high court] and the United States Supreme Court" (internal quotation marks omitted)).
Having adopted this third exception to Teague, we turn to the issue of whether our decisions in Guilbert 757and Dickson apply retroactively on collateral review. We first note the unique requirements and history of the due process provisions under our state constitution. We have explained that the due process provision of article first, § 8, of our state constitution affords greater protection than the federal constitution with respect to the admissibility of eyewitness identification testimony. See State v. Harris, supra, 330 Conn. at 114– 15, 131, 191 A.3d 119. In Harris, this court disagreed with its earlier decision holding that the state constitution did not afford greater protection than the federal constitution in this area. See id., at 116–21, 131, 191 A.3d 119. We explained that our prior decision was "premised in part on our reservations about scientific studies that we now find persuasive." (Internal quotation marks omitted.) Id., at 119, 140 S.Ct. 1390. After conducting our analysis under State v. Geisler, 222 Conn. 672, 684–85, 610 A.2d 1225 (1992); see State v. Harris, supra, at 116–30, 191 A.3d 119; we concluded that "[Connecticut’s] precedent, persuasive federal and sister state precedent, and contemporary understandings of economic and sociological norms favor[ed] the defendant’s claim … [that] the state constitution [affords greater protection than the federal constitution with respect to eyewitness identification testimony]." Id., at 130, 191 A.3d 119. This militates in favor of retroactivity.
Our case law regarding eyewitness identification evidence has also progressed, steadily following scientific developments in the field. We now know that the accuracy of a criminal conviction based solely on eyewitness identification is not as strong as courts once believed. "Nationally, [approximately] 69 [percent] of DNA exonerations—252 out of 367 cases—have involved eyewitness misidentification, making it the leading contributing cause of these wrongful convictions. Further, the National Registry of Exonerations has identified at least 450 [non-DNA based] exonerations involving eyewitness misidentification." Innocence Project, How Eyewitness 758Misidentification Can Send Innocent People to Prison (April 15, 2020), available at https://innocenceproject.org/how-eyewitness-misidentification-can-send-innocent-people-to-prison/#:~:text=Eyewitness% 20misidentification% 20is%20a,cause%20of%20these%20wrongful%20convictions (last visited July 15, 2024). It is no wonder, then, that mistaken eyewitness identification testimony is "by far the leading cause of wrongful convictions." State v. Guilbert, supra, 306 Conn. at 249–50, 49 A.3d 705.
In recent years, we have "recognized that mistaken eyewitness identifications are a significant cause of erroneous convictions; [id.] (‘mistaken eyewitness identification testimony is by far the leading cause of wrongful convictions’); and the risk of mistake is particularly acute when the identification has been tainted by an unduly suggestive procedure. [See] United States v. Wade, 388 U.S. 218, 229, 87 S. Ct. 1926, 18 L. Ed. 2d 1149 (1967) (‘[t]he influence of improper suggestion [on] identifying witnesses probably accounts for more miscarriages of justice than any other single factor—perhaps it is responsible for more such errors than all other factors combined’ …)." (Footnote omitted.) State v. Dickson, supra, 322 Conn. at 425–26, 141 A.3d 810. Accordingly, this court has established new rules aimed at bringing our case law in line with our advanced understanding of eyewitness identifications. These rules strive to ensure that there is an accurate determination as to innocence or guilt.
[14] As to the retroactive application of Guilbert, we conclude that the principles articulated in Guilbert may not be applied retroactively because that case articulated an evidentiary rule, not a constitutional one. See State v. Guilbert, supra, 306 Conn. at 265 and n.45, 49 A.3d 705. Under either Teague’s watershed exception or the third exception we adopt today, in order to have retroactive application, the new rule must be of constitutional dimension. In Guilbert, we concluded that trial courts have the discretion759 to admit expert tes- timony on the reliability of eyewitness identifications. Id., at 257, 49 A.3d 705. Because this new rule is evidentiary, it cannot be applied retroactively to the petitioner’s case on collateral review.
[15, 16] There is no question, however, that Dickson announced a constitutional rule of criminal procedure. We reasoned that placing a witness on the stand, confronting the witness with the person whom the state has accused of committing the crime, and asking that witness to identify the person who committed the crime is likely the most suggestive identification procedure. State v. Dickson, supra, 322 Conn. at 423, 141 A.3d 810. In addition to the suggestive nature of an in-court identification, "[t]he pressure of being asked to make an identification in the formal courtroom setting and the lack of anonymity … create conditions under which a witness is most likely to conform his or her recollection to expectations .…" E. Mandery, "Due Process Considerations of InCourt Identifications," 60 Alb. L. Rev. 389, 417 (1997). Among other reasons, this is why eyewitness identification is among the least reliable forms of evidence. See, e.g., D. Medwed, "Anatomy of a Wrongful Conviction: Theoretical Implications and Practical Solutions," 51 Vill. L. Rev. 337, 358 (2006) ("[v]irtually all … pertinent studies … have pinpointed eyewitness misidentification as the single most pervasive factor in the conviction of the innocent"). There is no doubt that, in cases in which the identity of the person who committed the crime is at issue, first-time, in-court identifications are unnecessarily suggestive and, therefore, raise concerns regarding an accurate conviction, and "the fundamental fairness of [a] trial … is seriously diminished" in such a situation. (Citations omitted.) Teague v. Lane, supra, 489 U.S. at 312–13, 109 S.Ct. 1060 (plurality opinion). As a result, in Dickson, we concluded that "any [first-time] in-court identification by a witness who would have been unable to reliably identify the [petitioner] in a nonsuggestive 760out-of-court procedure constitutes a procedural due process violation." (Emphasis omitted.) State v. Dickson, supra, at 426 n.11, 141 A.3d 810. Thus, we conclude that it is necessary to apply Dickson retroactively to the petitioner’s case on collateral review to ensure the reliability of his criminal trial.
The respondent, however, claims that, in footnote 34 of Dickson, this court concluded that Dickson should not be applied retroactively on collateral review. See id., at 451 n.34, 141 A.3d 810. Although we agree that there is language in footnote 34 to that effect, we conclude that this comment was dictum. See, e.g., Cruz v. Montanez, 294 Conn. 357, 376–77, 984 A.2d 705 (2009) ("[d]ictum includes those discussions that are merely passing commentary … those that go beyond the facts at issue … and those that are unnecessary to the holding in the case" (internal quotation marks omitted)). First, we note that the parties in Dickson did not significantly discuss retroactivity concerns in their briefing. The issue of retroactivity was not necessary to determine the outcome in Dickson, which was a direct appeal. Moreover, we disagree with this court’s assertion in Dickson that "the rule [requiring prescreening of a first-time, in-court identification] is merely an incremental change in identification procedures." State v. Dickson, supra, 322 Conn. at 451 n.34, 141 A.3d 810. As we have explained, "[t]he influence of improper suggestion [on] identifying witnesses probably accounts for more miscarriages of justice than any other single factor—perhaps it is responsible for more such errors than all other factors 761combined .… " (Internal quotation marks omitted.) Id., at 426, 141 A.3d 810. We cannot conclude that prescreening an eyewitness prior to a first-time, in-court identification is merely an "incremental change" that serves only to remove "some remote possibility" of a wrongful conviction. (Internal quotation marks omitted.) Id., at 451 n.34, 141 A.3d 810. Thus, we disagree with the respondent’s contention that footnote 34 in Dickson is dispositive of the retroactivity issue.
In passing, the state addressed retroactivity in the context of this court’s using its supervisory authority rather than announcing a constitutional prophylactic rule. See State v. Dickson, Conn. Supreme Court Records & Briefs, December Term, 2015, State’s Brief p. 50; see also id., State’s Supplemental Brief p. 5; id., State’s Supplemental Reply Brief pp. 1–2, 5. The defendant did not address retroactivity in his primary brief but briefly discussed the issue in his supplemental reply brief. See id., Defendant’s Supplemental Reply Brief p.
In sum, the rule articulated in Dickson is "central to an accurate determination of innocence or guilt"; Teague v. Lane, supra, 489 U.S. at 313, 109 S.Ct. 1060 (plurality opinion); such that the rule’s absence would create an impermissibly large risk that innocent persons will be convicted. We agree with the brief of the amici curiae, the Innocence Project, Inc., and the Connecticut Innocence Project, that "[t]he issue of inaccurate eyewitness identification testimony … strikes at the heart of whether a criminal proceeding is fair and accurate." We therefore conclude that the rule set forth in Dickson must apply retroactively on collateral review because the rule was a result of developments in science that persuaded us to reevaluate the fundamental principles underlying eyewitness identification evidence, the application of the rule significantly improves the accuracy of the petitioner’s conviction, and the petitioner advocated for the rule in his direct appeal.
In the petitioner’s criminal trial, there were two eyewitnesses. LeVasseur, a white female, initially identified someone other than the Black petitioner from a photographic array as the shooter. State v. Tatum, supra, 219 Conn. at 724, 595 A.2d 322. It was not until almost three months after the shooting, during a subsequent array, that she identified the petitioner. See id. The second witness also identified someone other than the petitioner as the shooter—the same individual the first witness had identified—but later declined to identify anyone until he 762could see the suspect in person. Id. More than one year after the shooting, at the probable cause hearing, both witnesses identified the petitioner, who was the only Black man seated at defense counsel’s table. See id., at 724–25, 595 A.2d 322. Notwithstanding the prior identifications of another individual, both eyewitnesses later testified that they had no doubt about their identifications of the petitioner.
On appeal, the petitioner contends that, because he was arrested approximately two months prior to the probable cause hearing and remained incarcerated through the hearing, it is "highly likely that he was wearing prison clothes, rather than street clothes, at the time of [the witness’] identification." We note, however, that the record indicates that the petitioner was wearing a "green, plaid shirt" at the probable cause hearing.
One additional fact weighs heavily in favor of applying Dickson. retroactively in this particular case. More than thirty years ago, in his direct appeal, the petitioner challenged the procedures related to an eyewitness identification used in his criminal case. See id., at 723, 725, 728, 595 A.2d 322. At that time, this court concluded that the first-time, in-court identification of the petitioner at the probable cause hearing was not unnecessarily suggestive because it was "necessary for the prosecution to present evidence at the preliminary hearing to establish probable cause to believe that [the petitioner] had committed the crimes charged." (Emphasis omitted.) Id., at 728. Twenty-five years later, recognizing the inherent suggestiveness of a first-time, in-court identification, this court overruled the holding in the petitioner’s direct appeal regarding the procedure that was used to identify the petitioner, calling the first time, in-court identification of the petitioner "unfair …. " State v. Dickson, supra, 322 Conn. at 435–36, 141 A.3d 810. Despite this conclusion in 763 Dickson, which explicitly rejected the eyewitness identification procedures used in his criminal case, the petitioner has not yet had the opportunity to raise the claim that, in light of our decision in Dickson, the identification procedures used in his criminal case violated his right to due process.
During his direct appeal, the petitioner also raised a challenge to the eyewitness identification jury instructions given in his criminal case. See State v. Tatum, supra, 219 Conn. at 732, 595 A.2d 322. This court concluded that the instructions were "adequate to alert the jury to the dangers inherent in eyewitness identification." Id., at 734, 595 A.2d 322. Of course, more than two decades later, on the basis of developed science on the reliability of eyewitness identifications, we recognized that eyewitness identifications are potentially unreliable in a number of ways unknown to the average juror; State v. Guilbert, supra, 306 Conn. at 234, 49 A.3d 705; and, as such, the jury instructions used in the petitioner’s criminal trial were not sufficient to alert the jury to factors affecting the reliability of the eyewitness identifications. See id., at 258, 49 A.3d 705; see also id., at 247 n.27, 49 A.3d 705. Many of the factors that could adversely impact eyewitness identifications and that are unknown to the average juror were present in the petitioner’s criminal case. For example, both eyewitnesses in the petitioner’s criminal case testified that they had no doubt regarding their identification of the petitioner, notwithstanding their earlier identification of another individual as the shooter. Cf. id., at 253–54, 49 A.3d 705. Both eyewitnesses also had only a limited opportunity during the high stress situation to view the individual they later identified; we now understand that high stress situations involving weapons can impact the reliability of an identification. See id., at 253, 49 A.3d 705. Moreover, both eyewitness identifications were cross-racial, which we know are "considerably less accurate than identifications involving the same race .…" Id. Nevertheless, given our conclusion that Guilbert announced an evidentiary rule, not a constitutional one, Guilbert does not apply retroactively on collateral review.
[17, 18] The central purpose of a criminal trial is "to ascertain the truth which is the sine qua non of a fair trial." Estes v. Texas, 381 U.S. 532, 540, 85 S. Ct. 1628, 14 L. Ed. 2d 543 (1965). Mistaken eyewitness identifications are the leading cause of wrongful convictions. See, e.g., State v. Guilbert, supra, 306 Conn. at 249–50, 49 A.3d 705. The risk of mistaken eyewitness identifications is particularly acute when the identification has been tainted by an unduly suggestive procedure. See, e.g., United States v. Wade, supra, 388 U.S. at 229, 87 S.Ct. 1926. As a result, unduly suggestive and unreliable eyewitness identifications undermine the truth seeking function of the criminal justice system. Given the developments in the science of eyewitness identification, the heightened risk of a wrongful conviction, and the fact that the petitioner raised eyewitness identification claims in his direct appeal, we conclude that the 764rule articulated in Dickson must be applied retroactively on collateral review in the petitioner’s case. See, e.g., Colwell v. State, 118 Nev. 807, 820, 59 P.3d 463 (2002), cert. denied, 540 U.S. 981, 124 S. Ct. 462, 157 L. Ed. 2d 370 (2003). Accordingly, we conclude that the Appellate Court, which lacked the benefit of our newly expanded formulation of the Teague rule, should not have upheld the habeas court’s dismissal of counts six and seven of the petitioner’s operative petition on the ground that Dickson does not apply retroactively to the petitioner’s case on collateral review.
CONCLUSION
Thirty-three years ago, the petitioner argued before this court that the trial court had improperly admitted an in-court eyewitness identification of him at his criminal trial that was tainted by an unnecessarily suggestive pretrial identification procedure. See State v. Tatum, supra, 219 Conn. at 723, 725, 595 A.2d 322. At that time, this court rejected that claim; see id., at 723, 595 A.2d 322; and the petitioner has served decades in prison as a result. In recent years, however, this court’s jurisprudence has benefitted from significant developments related to the cognitive science associated with eyewitness identifications. In light of those scientific developments, in Dickson, we recognized that this court was clearly wrong when it rejected the petitioner’s original claim in his direct appeal regarding the unnecessarily suggestive in-court, pretrial identification, and we overruled the holding in the petitioner’s direct appeal. See State v. Dickson, supra, 322 Conn. at 434–36, 141 A.3d 810. We do not lightly overrule holdings in prior cases. It is only "[w]hen a prior decision is seen so clearly as error that its enforcement [is] for that very reason doomed" that we will overrule it. (Emphasis omitted; internal quotation marks omitted.) Conway v. Wilton, 238 Conn. 653, 659, 680 A.2d 242 (1996); see, e.g., Kluttz v. Howard, 228 Conn. 401, 406, 636 A.2d 816 (1994) ("a court should not overrule its earlier decisions 765unless the most cogent reasons and inescapable logic require it" (internal quotation marks omitted)). Had the petitioner made the same claim today that he raised in his direct appeal more than thirty years ago, he would prevail. It is long past time that the petitioner be afforded the opportunity to challenge the procedures related to the eyewitness identification used in his criminal case in light of the principles we articulated in Dickson.
The judgment of the Appellate Court is reversed insofar as it upheld the habeas court’s dismissal of counts six and seven of the petitioner’s operative habeas petition and the case is remanded to the Appellate Court with direction to reverse the judgment of the habeas court with respect to those counts and to remand the case to that court for a trial on counts six and seven and with direction to apply the holding of Dickson retroactively to the petitioner’s case.
In this opinion the other justices concurred.