State v. Ayala (Dissent)Annotate this Case
*********************************************** The “officially released” date that appears near the beginning of each opinion is the date the opinion will be published in the Connecticut Law Journal or the date it was released as a slip opinion. The operative date for the beginning of all time periods for filing postopinion motions and petitions for certification is the “officially released” date appearing in the opinion. All opinions are subject to modification and technical correction prior to official publication in the Connecticut Reports and Connecticut Appellate Reports. In the event of discrepancies between the advance release version of an opinion and the latest version appearing in the Connecticut Law Journal and subsequently in the Connecticut Reports or Connecticut Appellate Reports, the latest version is to be considered authoritative. The syllabus and procedural history accompanying the opinion as it appears in the Connecticut Law Journal and bound volumes of official reports are copyrighted by the Secretary of the State, State of Connecticut, and may not be reproduced and distributed without the express written permission of the Commission on Official Legal Publications, Judicial Branch, State of Connecticut. *********************************************** STATE v. AYALA—DISSENT D’AURIA, J., with whom PALMER and ECKER, Js., join, dissenting. Lawyers learn early in law school what we all know instinctively to be true: conclusions built on hearsay can be inherently unreliable and unfair. When a witness testifies to what another has said—unsworn and out of court—there is a heightened potential that the evidence is inaccurate, fabricated or lacking context. In criminal cases particularly, when the out-ofcourt declarant is unavailable for cross-examination, admitting certain hearsay testimony against a defendant can render the trial constitutionally infirm. See Crawford v. Washington, 541 U.S. 36, 68, 124 S. Ct. 1354, 158 L. Ed. 2d 177 (2004) (sixth amendment confrontation clause bars admission of testimonial hearsay against criminal defendant unless defendant had prior opportunity for cross-examination and witness is unavailable); see also Michigan v. Bryant, 562 U.S. 344, 370 n.13, 131 S. Ct. 1143, 179 L. Ed. 2d 93 (2011) (due process clauses of fifth and fourteenth amendments may bar admission of ‘‘unreliable evidence’’). As a general matter, then, in pursuit of their truthseeking function, courts do not permit hearsay. See Conn. Code Evid. § 8-2. Under certain circumstances, however, and also in pursuit of the truth-seeking function, our evidence code makes exceptions to this general rule and permits courts to admit hearsay. These exceptions are most often justified when the statements at issue were made ‘‘under circumstances that tend to assure reliability and thereby compensate for the absence of the oath and opportunity for cross-examination.’’ Chambers v. Mississippi, 410 U.S. 284, 299, 93 S. Ct. 1038, 35 L. Ed. 2d 297 (1973).1 In the present case, both the victim, Thomas L. Mozell, Jr., and the defendant, Vincente Ayala, were members of a street gang known as Piru, which had a presence in the New Haven area. No physical evidence or firsthand eyewitness evidence linked the defendant to the victim’s murder in 2012. It is therefore no understatement to say that all of the pertinent testimony identifying the defendant as the murderer was either hearsay or hearsay statements ostensibly offered to prove something other than the truth of the matter asserted. Moreover, each of the hearsay statements was testified to by two other gang members, Timothy Thomas and Jordan Richard, whom the trial court determined were ‘‘significantly compromised individuals’’ who provided ‘‘extremely limited and self-serving’’ testimony. None of this testimony was corroborated by anything other than more hearsay. In the trial court’s words, there was a ‘‘complete lack of any evidence supporting’’ their claims. In my view, these are hardly ‘‘circumstances that tend to assure reliability . . . .’’ Chambers v. Mississippi, supra, 410 U.S. 299. At issue in this appeal is a particular hearsay statement delivered to the court through Richard, in which one of the gang’s leaders, known as ‘‘Terror,’’ claimed to have seen the defendant shoot the victim. The trial court admitted the statement under the coconspirator exception to the rule against hearsay. See Conn. Code Evid. § 8-3 (1) (E). The majority assumes that this was error but concludes that it was not harmful. Given the scant evidence in this case, including the state’s heavy reliance on uncorroborated hearsay that came into evidence through witnesses the trial court found to be unreliable, I am left without ‘‘a fair assurance’’ that the admission of the coconspirator evidence ‘‘did not substantially affect the verdict.’’ State v. Favoccia, 306 Conn. 770, 808, 51 A.3d 1002 (2012). I must therefore confront the question of whether the trial court erroneously admitted Terror’s statement, as the majority assumes. I conclude it did because there is a lack of evidence that the statement was made ‘‘in furtherance’’ of an ongoing conspiracy, as required by the coconspirator exception. I would therefore reverse the trial court’s judgment. Accordingly, I respectfully dissent.2 I Clearly, someone in the Piru gang wanted the victim dead. He was believed to be thinking of leaving the Piru gang to join another gang and would perhaps retaliate against Piru. The main question at trial was which Piru member carried out the victim’s murder. I have no disagreement with the majority’s recitation of facts that the jury could have reasonably found to be true. I simply summarize them here to explain why I cannot agree with the majority that the state’s case was ‘‘sufficiently strong’’ so as to leave me with a fair assurance that admission of the challenged statements did not substantially affect the verdict. In light of this conclusion, I further explain why I do not find the defendant’s claim of error harmless, if indeed it was error. A The victim apparently knew he was not in good standing with Piru. We know this from a statement he supposedly made to his friend, Tavaris Wylie, on the night of the murder. Wylie testified that the victim told him he had a ‘‘funny vibe about everybody’’ in the gang; he ‘‘felt like they was rocking him to sleep’’; acting like ‘‘a sheep in [wolf’s] clothing’’; and that he had to ‘‘[w]atch [his] back’’ because ‘‘[t]hey was playing [him]’’ and ‘‘[c]oming for [him].’’ Although the victim expressed that he had this ‘‘vibe’’ about everybody in the gang, never mentioning the defendant individually, the defendant was the only one on trial. As many as forty members of Piru were in the New Haven area at that time, however. The trial court nonetheless determined that Wylie’s testimony was relevant to establish the defendant’s motive to kill the victim and admitted it into evidence under either the state of mind exception to the hearsay rule or as nonhearsay. See Conn. Code Evid. §§ 8-3 (4) and 8-1 (1). The defendant has challenged this ruling on appeal. See part II of the majority opinion. The victim’s fear of the gang appears to have been well founded. We know this because Terror apparently had ordered him murdered. Terror did not testify, and we know virtually nothing about him other than that he was a leader of an out-of-state faction of the Piru gang. And yet, despite this detachment, what Terror supposedly said both before and after the murder— namely, his order that the victim be murdered and his statement recounting how the murder happened— played a prominent part in the case against the defendant. In particular, Terror’s purported description of how the murder occurred was the only eyewitness evidence of the killing. No testifying witness saw the murder, nor could a testifying witness say the defendant committed it, without relying on the alleged admissions from the defendant himself. Terror’s description of the killing is unquestionably at issue in this appeal. Thomas testified that Terror had initially ordered him to kill the victim. This hearsay statement was admitted under either the coconspirator exception to the hearsay rule or as nonhearsay, and its admissibility is not challenged on appeal. See Conn. Code Evid. §§ 8-3 (1) (E) and 8-1 (1). Thomas refused the order because the victim was a close friend. According to Thomas, the defendant then volunteered to carry out the murder. This hearsay statement is an admission by the defendant, and its admissibility is not challenged on appeal. See Conn. Code Evid. § 8-3 (1) (A). Later that evening, Thomas claimed, he warned the victim that the gang posed a threat to his life, although he did not mention the defendant specifically. The police found the victim dead in his vehicle the next morning, a bullet to the back of his head. There seems little doubt that someone in the gang committed the murder, but because none of the testifying witnesses claimed to have seen the shooting, it is less clear which gang member pulled the trigger. Richard gave the most contemporaneous firsthand testimony, although he claimed to have no personal knowledge as to who shot the victim. He testified that on the evening of the murder, the victim sat in the driver’s seat of his vehicle smoking marijuana with Terror, Richard and one other gang member, Montese Gilliams. Richard claimed that the defendant then showed up and entered the vehicle. At that point, Richard said, the other occupants all told him to leave, which he did. No one who remained in the vehicle testified at trial. Rather, the evidence that it was the defendant who killed the victim came in the form of three other hearsay statements. The first two statements allegedly came from the defendant. Years after the murder, when Thomas and Richard faced unrelated criminal charges, each pointed the finger at the defendant. Both testified that the defendant had admitted to them that he had shot the victim. Richard said that the defendant had confessed on the night of the murder that he had shot the victim. Thomas said that the defendant had told him the next day that he felt badly about killing the victim but that Terror had ordered him to do so. These hearsay statements are also party admissions and are not challenged on appeal. See Conn. Code Evid. § 8-3 (1) (A). The third statement is the subject of this appeal. It allegedly came from Terror, a leader of the New York faction of the gang. Richard testified that he and Terror went to the home of Terror’s mother in the Bronx, New York, the day after the murder and remained there for about one week. At some point while they were in New York, Richard claimed, Terror described to him how the defendant had murdered the victim: ‘‘[Terror] told me that all four of them were in the car smoking. . . . And in the midst of, you know, the joint being passed around, he somehow handed [the defendant] the weapon, and the entire time while [the defendant] is, you know, playing with the—you know, fixing the gun, cocking it back or whatever, Terror is blocking [the defendant] so [the victim] can’t see him in the rearview mirror, and Terror sat—[a]fter that, Terror sat back, he looked at [the defendant], and he fired.’’ The trial court admitted Terror’s hearsay statement, over the defendant’s objection, under the coconspirator exception to the hearsay rule. See Conn. Code Evid. § 8-3 (1) (E). The state’s case against the defendant can therefore, I believe, be fairly summarized as follows: The victim said that he feared the entire gang. Members of that gang, who were themselves involved in the victim’s murder, then blamed the defendant for the murder via hearsay statements. One gang member (Thomas) testified that the defendant agreed to kill the victim. Two gang members (Thomas and Richard) claimed the defendant admitted that he killed the victim. And a leader of the gang (Terror) told the second gang member (Richard) how the defendant had killed the victim. This evidence was perhaps sufficient to convict the defendant.3 But in a case in which every member of the Piru gang—including everyone who was in the vehicle with the victim that night—had the same motive (to carry out the gang leader’s order), in my view, the evidence was hardly overwhelming. I am not the only one who thinks so. The defendant was acquitted of the charges of carrying a pistol without a permit and criminal possession of a firearm, which he elected to have tried to the court. Although these charges differ from those the jury considered, the court’s finding turned on precisely the same factual issue. As the court stated: ‘‘The only real issue in the dispute before me is whether the defendant shot [the victim] on March 16, 2012, and therefore was carrying and possessing a pistol when he did so.’’ The court explained that the ‘‘evidence against the defendant rested primarily on the testimony of two witnesses,’’ Thomas and Richard, ‘‘significantly compromised individuals’’ who gave testimony that was ‘‘extremely limited and self-serving.’’ The court further stated: ‘‘The testimony of the two gang members [Thomas and Richard] who had pointed the finger at the defendant—that testimony was simply unconvincing to me.’’ In light of the ‘‘complete lack of any evidence supporting’’ their claims, the court concluded: ‘‘After considering all the evidence, and most importantly in this case the lack of evidence, I am left with an honest and reasonable uncertainty in my mind that the defendant sat in the backseat of the victim’s motor vehicle, held a pistol in his hand, and shot [the victim].’’ In ruling on the defendant’s postverdict motion for a judgment of acquittal on the jury’s guilty verdict on the charges of murder and conspiracy to commit murder, the trial court appropriately observed that it could not substitute its assessment of the witnesses’ credibility for that of the jury and therefore denied the motion, notwithstanding that the jury verdict was inconsistent with its own finding on the weapons charges. But the trial court sat through the trial and observed the witnesses, and, therefore, in my view, its assessment of the evidence should not be discounted in assessing the strength of the state’s case without Terror’s statement. E.g., State v. Kendrick, 314 Conn. 212, 223, 100 A.3d 821 (2014) (deference to fact finder appropriate when ‘‘assessment of the credibility of the witnesses . . . is made on the basis of its firsthand observation of their conduct, demeanor and attitude’’ [internal quotation marks omitted]). The majority apparently believes that in evaluating the strength of the state’s case for harmless error purposes we must ignore this inconsistent verdict. I acknowledge that in evaluating the strength of the case, we must begin with the assumption that the jury found the state’s witnesses credible, which the trial court did not. I further acknowledge the unique challenge of evaluating the ‘‘strength’’ of the state’s case when the jury has found the defendant guilty beyond a reasonable doubt. Viewed in this light, all cases on appeal are strong cases. The question we confront—as required by our case law—is whether the case was strong enough to survive the removal of the contested evidence, which the majority assumes—and I conclude—to have been erroneously admitted. See State v. Favoccia, supra, 306 Conn. 809 (‘‘ ‘overall strength of the prosecution’s case’ ’’ among several factors considered in evaluating harm- fulness). On the basis of my own review of the tenuous record of hearsay evidence in this case, including the fact that no forensic evidence implicated the defendant, that there was no firsthand testimony of any eyewitness who saw the shooting, and no other circumstantial evidence of any kind, I have no trouble concluding that this was not a strong case, notwithstanding that the jury returned a verdict against the defendant. And its strength would have diminished with the removal of the Terror statement, which, in addition to being the only evidence of a purported eyewitness who claimed the defendant committed the murder, served to corroborate the other hearsay that remained in this case. The majority will admit only that the case against the defendant was not ‘‘ironclad . . . .’’ Although the different conclusions reached by different fact finders is not dispositive to me in light of my own independent evaluation, I can think of no better, objective example of a ‘‘close case’’—and therefore, in my view, not a strong one—than a case in which two fact finders come to different conclusions, particularly where one of them found the state’s main witnesses ‘‘simply unconvincing . . . .’’ I cannot agree with the majority that the state’s case can objectively be described as strong, even when considering all of the evidence, let alone if Terror’s statement to Richard is excluded. B The majority assumes error in the admission of Terror’s out-of-court statement to Richard under the coconspirator exception to the hearsay rule; see Conn. Code Evid. § 8-3 (1) (E); but proceeds to find that the defendant has not carried his burden of proving that this error harmed him. In large part because I disagree with the majority that the state’s case was sufficiently strong, I disagree with the majority’s conclusion on harm. An evidentiary error is harmless only ‘‘when an appellate court has a fair assurance that the error did not substantially affect the verdict.’’ (Internal quotation marks omitted.) State v. Favoccia, supra, 306 Conn. 809. In searching for this fair assurance, we consider several factors, including ‘‘the importance of the witness’ testimony in the prosecution’s case, whether the testimony was cumulative, the presence or absence of evidence corroborating or contradicting the testimony of the witness on material points, the extent of crossexamination otherwise permitted, and, of course, the overall strength of the prosecution’s case.’’ (Internal quotation marks omitted.) Id. ‘‘Perhaps the single most significant factor in weighing whether an error was harmful . . . is the strength of the case against the defendant. . . . [A] court should be especially loath to regard any error as harmless in a close case, since even the smallest error may have been enough to tilt the balance in favor of a conviction.’’ 3B C. Wright et al., Federal Practice and Procedure (4th Ed. 2019) § 854. For reasons previously detailed, I believe that, at best, this was a ‘‘close case.’’ Four points persuade me that if Terror’s out-of-court statement to Richard was erroneously admitted, the defendant has shown that its inclusion was harmful: the absence of physical or eyewitness evidence corroborating the testimony of both key witnesses; the state’s admitted reliance on unreliable witnesses; the fact that Terror’s statement was not entirely cumulative of other evidence; and the fact that the declarant was not subject to cross-examination. First, no physical evidence connected the defendant to the crime scene. Neither the defendant’s fingerprints nor his DNA were found on or near the vehicle. Although ‘‘the absence of conclusive physical evidence . . . does not automatically render [the state’s] case weak, that same absence surely does not strengthen the state’s case against the defendant.’’ State v. Ceballos, 266 Conn. 364, 416, 832 A.2d 14 (2003). Physical evidence did, however, tie Richard to the crime scene. The police found Richard’s fingerprint on the vehicle’s rear right interior door handle—the handle next to the seat the shooter occupied. In fact, this evidence helped establish a sufficient connection to Richard as the culprit that the trial court permitted the defendant to make a third-party culpability argument and instructed the jury on that theory. To warrant a third-party culpability instruction, the evidence must ‘‘directly [connect] the third party to the crime.’’ (Emphasis in original; internal quotation marks omitted.) State v. Francis, 267 Conn. 162, 174, 836 A.2d 1191 (2003). This is a ‘‘high standard.’’ Id., 175; see also Johnson v. Commissioner of Correction, 330 Conn. 520, 564, 198 A.3d 52 (2019) (‘‘proffered evidence [of thirdparty culpability] [must] establish a direct connection to a third party, rather than raise merely a bare suspicion’’ [internal quotation marks omitted]). According to the trial court, ‘‘a wealth’’ of evidence pointed to Richard. Apart from his fingerprint being found next to the shooter’s seat, Richard’s presence in the vehicle on the night of the murder and his membership in Piru meant he had the same opportunity and motive to kill the victim as the defendant did. Richard also tampered with evidence when he removed a cell phone from the victim’s vehicle and destroyed it,4 fled to New York after the murder, where he remained for one week, and lied to the police for several years about his involvement in the incident. In evaluating the strength of the state’s case, I simply do not find it impressive that Richard was able to deliver testimony consistent with Terror’s hearsay and the testimony of the prosecution’s forensics expert, James R. Gill, the state’s chief medical examiner. Specifically, the majority finds it remarkable that Richard was able to place the defendant in the very seat in the car where, according to the expert’s testimony, the killer must have sat when he fired the fatal gunshot. But if Richard were the killer—and a wealth of evidence suggested he was— of course he would know where the killer was sitting! He was in that very seat, by his own admission. His testimony would have been more remarkable if anyone besides Richard testified at trial that he left the car that night so the defendant could get in. Instead, he explains that his fingerprint was in the car (the defendant’s wasn’t) because he was in the car before and after the shooting. And the only corroboration for his story comes from hearsay—testified to by himself. In all, forensic evidence placed the killer in a particular seat. An admission and fingerprints placed Richard in that same seat. Unlike the majority, I make very little of the fact that Richard (a seriously compromised witness) claimed (without corroboration) that the defendant also occupied that seat. In addition to the absence of physical evidence implicating the defendant, no firsthand eyewitness testimony implicated him, either. The only statement of an eyewitness to the murder was Terror’s unsworn, out-of-court statement, delivered through Richard. Even if I assume that this statement was admitted in error, as the majority does for purposes of assessing harm, the state would be left with no physical evidence or eyewitness testimony connecting the defendant to the crime. Only one eyewitness, Richard, even placed the defendant in the vehicle with the victim near the time of the killing. This compels the state to concede that its case rested on the testimony of Thomas and Richard.5 But as the trial court aptly observed, both suffered from serious credibility problems. This leads to my second point: although Thomas’ and Richard’s testimony, if believed, was perhaps sufficient to uphold the verdict, the harmlessness inquiry is more searching than a sufficiency of the evidence inquiry. E.g., Kotteakos v. United States, 328 U.S. 750, 765, 66 S. Ct. 1239, 90 L. Ed. 1557 (1946) (‘‘[t]he [harmlessness] inquiry cannot be merely whether there was enough to support the result’’); United States v. Ince, 21 F.3d 576, 583 (4th Cir. 1994) (‘‘more [stringent]’’ harmless error inquiry ‘‘does not ask simply whether we believe that irrespective of the error there was sufficient untainted evidence to convict,’’ but ‘‘whether we believe it highly probable that the error did not affect the judgment’’ [internal quotation marks omitted]). In deference to the jury, the majority declines to ‘‘second-guess’’ the credibility of Thomas and Richard. Although for some issues raised on appeal, we must defer to the jury’s credibility determinations, our cases make clear that we may consider witness credibility in a harmless error analysis.6 E.g., State v. Ritrovato, 280 Conn. 36, 57, 905 A.2d 1079 (2006) (error not harmless when state’s case lacked independent physical evidence or witnesses and ‘‘[victim’s] credibility was crucial to successful prosecution of the case’’); State v. Cortes, 276 Conn. 241, 256, 885 A.2d 153 (2005) (error not harmless ‘‘in a case that essentially turned on the jury’s crediting [the complainant’s] version of the events’’). Indeed, we have acknowledged that cases that present the jury with a ‘‘credibility contest characterized by equivocal evidence . . . [are] far more prone to harmful error.’’ (Internal quotation marks omitted.) State v. Favoccia, supra, 306 Conn. 816–17. Similarly, in other contexts, we have recognized the weakness of a case that is based largely on a defendant’s own incriminating statements, when the veracity of the statements was questionable and the state presented little physical or eyewitness evidence to support the defendant’s conviction. See, e.g., State v. A. M., 324 Conn. 190, 213–14, 152 A.3d 49 (2016) (discussing weakness of state’s case in adjudicating claimed violation of defendant’s fifth amendment right to remain silent when state presented ‘‘no physical evidence,’’ and key witness gave ‘‘inconsistent statements’’ and ‘‘refused to answer certain questions during her direct testimony’’); Lapointe v. Commissioner of Correction, 316 Conn. 225, 323, 112 A.3d 1 (2015) (discussing weakness of state’s case in [context of claim pursuant to Brady v. Maryland, 373 U.S. 83, 87, 83 S. Ct. 1194, 10 L. Ed. 2d 215 (1963)] when state presented minimal physical evidence and no eyewitnesses, and its case ‘‘rested almost entirely on [the petitioner’s own] incriminating statements’’ that were made in unreliable circumstances); Gaines v. Commissioner of Correction, 306 Conn. 664, 691, 51 A.3d 948 (2012) (discussing weakness of state’s case in context of claim pursuant to [Strickland v. Washington, 466 U.S. 668, 104 S. Ct. 2052, 80 L. Ed. 2d 674 (1984)] when primary evidence implicating petitioner was testimony of two witnesses who relayed purposed admissions by petitioner, and witnesses were ‘‘subject to substantial impeachment evidence that they had implicated the petitioner only to serve their own needs—either by directing suspicion away from their own involvement in the murders or by procuring more favorable outcomes in their other, unrelated criminal matters’’). Because of the broad language of our harmful error test, requiring that we examine the ‘‘overall strength of the prosecution’s case,’’ and because of the wide variety of cases in which courts have considered ways in which a key witness’ credibility has been undermined, I respectfully disagree with the majority to the extent it suggests we may consider witness credibility only when the claimed error relates to the admission or exclusion of impeachment evidence. Here, rather than impeachment, the claimed error relates to corroborative evidence. But both types of evidence bear on whether a jury would be likely to find the state’s witnesses believable. That, I believe, is the larger point to consider when assessing the overall strength of the state’s case and the ultimate question of whether we, as an appellate court, have ‘‘a fair assurance that the error’’ in admitting unreliable, corroborative evidence ‘‘did not substantially affect the verdict.’’ (Internal quotation marks omitted.) State v. Favoccia, supra, 306 Conn. 809; accord Holmes v. South Carolina, 547 U.S. 319, 330, 126 S. Ct. 1727, 164 L. Ed. 2d 503 (2006) (‘‘where the credibility of the prosecution’s witnesses or the reliability of its evidence is not conceded, the strength of the prosecution’s case cannot be assessed without making the sort of factual findings that have traditionally been reserved for the trier of fact’’). Thus, although the jury still could have found the defendant guilty without Terror’s hearsay statement by believing Thomas’ and Richard’s other testimony, that is not the question. Removing Terror’s statement from the evidentiary calculus removes critical corroboration from Thomas’ and Richard’s testimony, thus making them even less believable than they already were. This leaves me, at least, without a fair assurance that the trial court’s error in admitting Terror’s hearsay statement did not substantially affect the verdict. See State v. Favoccia, supra, 809. On the basis of the self-serving nature of Thomas’ and Richard’s testimony and the discrepancies in material aspects of their stories, I am persuaded that their credibility was subject to significant challenge. The trial court’s specific conclusion that they were not credible fortifies me in my view. Both witnesses’ testimony was clearly self-serving. The jury heard that Thomas only came forward about the victim’s murder two years later, just after he had been arrested in connection with other crimes and in exchange for a potential sentence reduction as part of a plea agreement with federal authorities. Similarly, Richard admitted he was an accomplice in the victim’s murder and was seeking consideration from the state in return for his testimony. He apparently was never charged. Moreover, as previously described, Richard exhibited enough suspicious behavior to warrant a third-party culpability charge. He also conceded that he had lied about the incident several times, offering three versions of his story over the years.7 There were also material discrepancies in the testimony of Thomas and Richard regarding encounters that preceded the victim’s murder. One encounter involved Terror’s apparent desire to have the victim killed. In short, Thomas and Richard agreed that a meeting occurred in the bedroom of the home of Davon Youmans, another Piru member, at which the victim was discussed, but Thomas and Richard disagreed on the meeting’s timing (morning versus evening), attendees (whether Thomas and the defendant were there), and results (whether Terror asked and whether the defendant volunteered to kill the victim).8 In my view, these discrepancies cast doubt on material facts: whether and when the conspiracy to kill the victim formed, and whether and when the defendant formed the intent to kill him. The other encounter involved the witnesses’ final interaction with the victim. Both Thomas and Richard agreed that the victim showed up at Youmans’ house in the early evening prior to the murder, but their stories diverged from there. Thomas testified that the victim walked to Youmans’ front porch only and did not enter the house. Thomas said that he and Richard then went onto the front porch to warn the victim that his life was in danger. At that point, the victim became upset and later got into his vehicle and left. Thomas then left with Gilliams and went home for the night. Once again, however, Richard’s version of this encounter substantially differed. Richard testified that the victim walked into Youmans’ kitchen, purchased marijuana from someone in the house, and left with Richard, Terror, and Gilliams to smoke it in the victim’s vehicle, where he ultimately was killed. Richard testified that the victim had not been to the house earlier in the day. Notably, Richard did not mention a conversation in which he and Thomas warned the victim of the threat on his life. In sum, either Thomas misremembered or fabricated Richard’s presence at his last face-to-face interaction with his close friend or Richard hid his involvement in this meeting from the court. Either explanation bears on the witnesses’ credibility. Third, I consider the role Terror’s hearsay statement (purporting to describe the murder from inside the vehicle as it occurred) played in relation to Thomas’ and Richard’s testimony (purporting to describe events from well before and after the murder occurred). Unlike the majority, I am not given a ‘‘fair assurance’’ that the erroneous admission of Terror’s statement was harmless simply because it contained some of the information already testified to or because it came in through Richard. Certainly, Terror’s statement was cumulative of Thomas’ and Richard’s statements in a general sense—all three identified the defendant as the shooter. But to the extent their statements were cumulative, they were also corroborative. Because Thomas’ and Richard’s credibility was central to the case—and, as previously described, in doubt—I cannot say that having a third voice identify the defendant as the shooter did not make Thomas’ and Richard’s firsthand accounts more believable to the jury, even if they were not believable to the trial judge. Terror’s statement was also distinct from Thomas’ and Richard’s, though. It brought the jury far closer to the murder in time and space. It provided the only account that allowed the jury to visualize the defendant receiving the gun, preparing it to fire and pulling the trigger. And the prosecutor alluded to it in closing argument to the jury: ‘‘Richard indicated to you he got kicked out [of the vehicle] because there was not enough room or not enough pot. The reality is, is Terror and the defendant do not want him there. They need to be able to be in the backseat and be able to maneuver . . . .’’ Fourth, I note the absence of cross-examination. Although the defendant cross-examined Richard, the conduit, he could not question Terror, the declarant. Although no more was legally required, I cannot conclude that the defendant’s inability to ask Terror himself about the out-of-court statement was irrelevant when considering its effect on the jury. The defendant was unable to probe Terror’s veracity when the statement in question exonerated Terror at the expense of the defendant. Especially in a situation like this, I am mindful of Justice Marshall’s admonition that ‘‘[t]he conspirator’s interest is likely to lie in misleading the listener . . . . It is no victory for common sense to make a belief that criminals are notorious for their veracity the basis for law.’’ (Internal quotation marks omitted.) United States v. Inadi, 475 U.S. 387, 404–405, 106 S. Ct. 1121, 89 L. Ed. 2d 390 (1986) (Marshall, J., dissenting). Ultimately, I disagree with the majority’s assertions that Terror’s firsthand account of the murder ‘‘added very little’’ and ‘‘simply was not pivotal’’ to the state’s case. To summarize, Terror’s statement was the sole eyewitness account of the murder in a case bereft of physical evidence pointing to the defendant. It served to counter ‘‘a wealth’’ of evidence pointing at Richard. It corroborated Thomas’ and Richard’s self-serving testimony, which, according to the trial court, otherwise went uncorroborated. It added details to the state’s case, which the prosecutor referenced in closing argument, that Thomas’ and Richard’s secondhand accounts did not provide. Finally, because Terror’s statement was hearsay, it largely went untested by cross-examination. Given my view that this was a close case, even when considering all of the evidence, I cannot agree that the erroneous admission of Terror’s supposed statement to Richard was harmless. To the contrary, I would hold that the defendant carried his burden of demonstrating that the admission of this statement was harmful. In light of this conclusion, I must address the merits of the defendant’s evidentiary claim. II The defendant challenges the admission of a certain hearsay statement allegedly made by Terror to Richard one week after the victim’s murder. That statement, which purported to describe how the defendant carried out the murder, is recounted in part I A of this dissenting opinion. The trial court admitted the statement under the coconspirator exception to the hearsay rule. See Conn. Code Evid. § 8-3 (1) (E).9 I conclude that this was an abuse of discretion because there was not a reasonable basis to determine whether the statement was made ‘‘in furtherance’’ of a conspiracy, as required by our rules of evidence. A A hearsay statement of a party opponent is admissible on the basis of the very nature of the adversary system itself. ‘‘[O]ne cannot claim that his own statement should be excluded because it was not made under oath or subject to cross-examination or in view of the trier of fact. . . . [T]he party himself is present and can explain, deny, or rebut any such statement.’’ 4 C. Mueller & L. Kirkpatrick, Federal Evidence (4th Ed. 2019) § 8:44; see also Fed. R. Evid. 801, advisory committee notes, 28 U.S.C. app., p. 1063 (‘‘[a]dmissions by a party-opponent are excluded from the category of hearsay on the theory that their admissibility in evidence is the result of the adversary system’’). Evidence codes extend this theory in several ways. Most relevant here, they rely on agency concepts to justify admitting hearsay on the basis of the declarant’s relationship to a party in the case at issue. See, e.g., Conn. Code Evid. § 8-3 (1) (C) (admissions of individual authorized, under substantive agency law, to speak on party’s behalf); Conn. Code Evid. § 8-3 (1) (D) (admissions of party’s agent or employee). The coconspirator exception is one type of agency based extension. See J. Levie, ‘‘Hearsay and Conspiracy: A Reexamination of the Co-Conspirators’ Exception to the Hearsay Rule,’’ 52 Mich. L. Rev. 1159, 1163 (1954) (‘‘[t]he usual reason given for the co-conspirators’ exception is the classical agency rationale that conspirators are co-agents and, as such, liable for each other’s declarations’’). Its roots in agency theory are tenuous, though, making it ‘‘[t]he most controversial extension,’’ according to some prominent commentators.10 30B C. Wright et al., Federal Practice and Procedure (2019 Ed.) § 6777. As another commentator has stated, ‘‘the exception is fraught with problems. In terms of theory, it is an embarrassment.’’ C. Mueller, ‘‘The Federal Coconspirator Exception: Action, Assertion, and Hearsay,’’ 12 Hofstra L. Rev. 323, 324 (1984). Even the drafters of rule 801 of the Federal Rules of Evidence conceded that agency theory only offers limited justification for a coconspirator statement: ‘‘the agency theory of conspiracy is at best a fiction and ought not to serve as a basis for admissibility beyond that already established.’’ Fed. R. Evid. 801, advisory committee notes, 28 U.S.C. app., p. 1064. The ‘‘in furtherance’’ requirement plays an important role in justifying the coconspirator exception to the hearsay rule. It is meant to ensure that the hearsay statement at issue is sufficiently tied to a party in the case so that the agency theory plausibly holds: ‘‘To fall within the [coconspirator exception], the co-conspirator’s statement had to be made ‘in furtherance of’ the conspiracy, a requirement that arose from the agency rationale that an agent’s acts or words could be attrib- uted to his principal only so long as the agent was acting within the scope of his employment.’’ Bourjaily v. United States, 483 U.S. 171, 188–89, 107 S. Ct. 2775, 97 L. Ed. 2d 144 (1987) (Blackmun, J., dissenting); 30B C. Wright et al., supra, § 6780 (‘‘[o]nly when the conspirator is trying to further the conspiracy can the [agency] fiction be maintained’’). Indeed, although the drafters of the Model Code of Evidence in 1942 eliminated the ‘‘in furtherance’’ requirement, the drafters of the Federal Rules of Evidence retained it ‘‘because they adjudged it a useful device for protecting defendants from the very real dangers of unfairness posed by conspiracy prosecutions.’’ (Internal quotation marks omitted.) United States v. Perez, 989 F.2d 1574, 1577–78 (10th Cir. 1993). This decision ‘‘should be viewed as mandating a construction of the ‘in furtherance’ requirement protective of defendants, particularly since the Advisory Committee was concerned lest relaxation of this standard lead to the admission of less reliable evidence.’’ (Internal quotation marks omitted.) United States v. Lang, 589 F.2d 92, 100 (2d Cir. 1978). ‘‘[S]ome courts construe this aspect of the rule so broadly that anything related to the conspiracy is found to be in furtherance of its objectives. This, of course, is precisely the result the Advisory Committee sought to avoid by retaining the ‘in furtherance’ requirement.’’ Garlington v. O’Leary, 879 F.2d 277, 283 (7th Cir. 1989); see also 30B C. Wright et al., supra, § 6777 (‘‘the rule, and particularly its ‘in furtherance’ requirement, can best be viewed as an effort to limit the admission of co-conspirator statements, rather than an invitation to let such statements flow unchecked through the courthouse doors’’). Thus, the ‘‘in furtherance’’ element of the exception ‘‘is a limitation on the admissibility of coconspirators’ statements that is meant to be taken seriously.’’ (Emphasis omitted.) Garlington v. O’Leary, supra, 283.11 To determine whether a statement is in furtherance of a conspiracy, courts ask ‘‘whether some reasonable basis exists for concluding that the statement furthered the conspiracy.’’ (Internal quotation marks omitted.) State v. Carpenter, 275 Conn. 785, 845, 882 A.2d 604 (2005), cert. denied, 547 U.S. 1025, 126 S. Ct. 1578, 164 L. Ed. 2d 309 (2006). A statement furthers a conspiracy if it ‘‘in some way [has] been designed to promote or facilitate achievement of the goals of the ongoing conspiracy . . . .’’ (Internal quotation marks omitted.) Id., 844. This includes a statement ‘‘prompting the listener— who need not be a coconspirator—to respond in a way that promotes or facilitates the carrying out of a criminal activity . . . .’’ (Internal quotation marks omitted.) Id. The declarant does not have ‘‘to ask a third party expressly to do something to further the conspiracy’’; (internal quotation marks omitted) id., 845; and the statement need not actually further the conspiracy. State v. Peeler, 267 Conn. 611, 631, 841 A.2d 181 (2004). Whether a statement was made ‘‘in furtherance’’ of the conspiracy is ultimately a question of the declarant’s intent. See id., 632 (‘‘[i]t is enough that [the statement is] intended to promote the conspiratorial objectives’’ [internal quotation marks omitted]). Although, as stated, a statement from a coconspirator to a nonconspirator may be ‘‘in furtherance’’ of the conspiracy in some cases, ‘‘statements to non-conspirators informing them of the conspiracy will generally not qualify for admission because such statements are rarely in furtherance of the conspiracy.’’ 30B C. Wright et al., supra, § 6780. ‘‘[A] statement that merely . . . spills the beans, with no intention of recruiting the [nonconspirator] into the conspiracy does not further the conspiracy . . . .’’ (Citation omitted; internal quotation marks omitted.) State v. Carpenter, supra, 275 Conn. 845. Examples of statements that evidence a lack of intent behind them include a ‘‘merely narrative’’ description of events underlying the conspiracy; e.g., United States v. Beech-Nut Nutrition Corp., 871 F.2d 1181, 1199 (2d Cir.), cert. denied sub nom. Lavery v. United States, 493 U.S. 933, 110 S. Ct. 324, 107 L. Ed. 2d 314 (1989); idle chatter; e.g., id.; and bragging; e.g., United States v. Warman, 578 F.3d 320, 339 (6th Cir. 2009). In assessing a statement to a nonconspirator, a court must often undertake a ‘‘careful examination of the context in which it was made.’’ United States v. Shores, 33 F.3d 438, 444 (4th Cir. 1994), cert. denied, 514 U.S. 1019, 115 S. Ct. 1365, 131 L. Ed. 2d 221 (1995). B At trial, the defendant argued that Terror’s statement was not made ‘‘in furtherance’’ of the conspiracy because there was not enough evidence about Terror’s intent in making the statement. The state offered several possible explanations, but little proof, of what Terror’s intentions might have been. The trial court conceded the question was ‘‘problematic’’ but nonetheless admitted the statement. It inferred Terror’s intent almost entirely from Richard’s knowledge of incriminating information. In ruling on the defendant’s objection, it noted that Richard was a member of Piru. Although the trial court determined, based on the evidence presented up to that point, that Richard was not a participant in the murder of the victim, it noted that he ‘‘was involved later in assisting the conspirators,’’ thereby learning ‘‘extremely prejudicial information’’ about the circumstances surrounding the murder. From this, the court extrapolated that Terror more likely than not intended to make this statement ‘‘to further involve [Richard]’’ in the incident and ‘‘to discourage [Richard] from going to the police and revealing what he knew, thereby bringing him into the conspiracy to conceal the murder. In my view, on the basis of the very little we know about the statement itself, the context in which it was uttered and the actors involved, this is a huge leap of logic with virtually no evidence to support it. I believe the trial court should have excluded the statement because there was not a reasonable basis to determine whether Terror intended it to ‘‘further’’ the conspiracy and other indicia of reliability were lacking. In discerning a declarant’s intent, our cases suggest that a third party’s knowledge of incriminating information may be a factor. But we have typically relied on evidence more closely tied to the declarant himself, such as a declarant’s requests for help with the conspiracy, the declarant’s membership in a related conspiracy with the third party, the difficulty the declarant would have in hiding the conspiracy from the third party, a specific statement of intent by the declarant to bring the third party into the conspiracy, or a series of later events to suggest the declarant was ultimately successful in recruiting the third party to join or help with the conspiracy. For instance, the state cites State v. Camacho, 282 Conn. 328, 924 A.2d 99, cert. denied, 552 U.S. 956, 128 S. Ct. 388, 169 L. Ed. 2d 273 (2007), in which the declarant’s statements to his live-in girlfriend were intended ‘‘to secure her continued cooperation in the concealment of the crimes,’’ similar to Terror’s alleged intention of convincing Richard not to go to the police. Id., 356. But in Camacho, we considered as evidence of this intent more than just the fact that the girlfriend had seen the declarant behave suspiciously after a murder (e.g., hiding a gun, scrubbing clothes). We also observed that the declarant had asked her to assist him in covering up the murder (e.g., to avoid talking to the police). Id. We noted the fact that the girlfriend was an active coconspirator with the declarant in a related drug-selling operation. Id. Finally, we relied on the fact that it would have been ‘‘ ‘difficult, if not impossible’ ’’ for the declarant to hide the conspiracy from his girlfriend any longer, given their romantic relationship and that they lived together. Id., 357. The defendant points to two other cases, and in each we cited evidence of the declarant’s intent beyond the third party’s knowledge of incriminating evidence. In State v. Pelletier, 209 Conn. 564, 552 A.2d 805 (1989), the declarant’s statements to his wife were intended ‘‘to lessen any emotional trauma the [conspiracy] would cause [his wife],’’ which would ensure her ‘‘further cooperation . . . .’’ Id., 578. As evidence of this intent, this court relied not only on the fact that the wife’s house was being used as the base for hiding the fruits of the conspiracy, but also on the fact that the declarant had specifically ‘‘indicated that he wanted to tell [his wife] about the [conspiracy] before she heard about it on the news.’’ Id., 577. In State v. Carpenter, supra, 275 Conn. 785, the declarant’s statements to his son were intended to be ‘‘the first step in gaining . . . cooperation, moral sup- port, future assistance and guaranteed silence in the aftermath of the [conspiracy].’’ Id., 846. As evidence of this intent, we relied on the fact that it would have been ‘‘difficult, if not impossible, to conceal the conspiracy’’ because the declarant and his son lived together. Id., 847. Moreover, we could discern ‘‘no other logical explanation’’ for why the declarant would have made the statements. Id., 846. We also cited a series of ‘‘[s]ubsequent events’’ where the declarant and his son worked together to carry out the conspiracy, a contract murder (e.g., the son’s knowing assistance in finding the intended victim of the conspiracy, the declarant’s offer to let his son kill the victim, and the son’s help in disposing of the murder weapon). Id., 847. Notably, we also concluded in Carpenter that other statements by the declarant to his longtime girlfriend were not made in furtherance of the conspiracy. Id., 851–52. Initially, the declarant had disclosed the existence of the conspiracy to his girlfriend. Id., 849. In a later conversation, he told her that he felt threatened by the man who had hired him to commit the murder. Id. Noting that the girlfriend did not portray the conversation as a ‘‘serious discussion’’ and that she did not believe him, we concluded that the declarant’s comments in this later conversation were ‘‘more akin to a casual reference to what was happening in his life than an attempt to induce [her] to take part in the conspiracy.’’ Id., 852. There appeared to be no other evidence of the declarant’s intent in making the statement. In the present case, the trial court based its inference about Terror’s intent on Richard’s membership in the gang and his knowledge of incriminating information. I do not believe this is a reasonable basis for concluding that Terror’s statement was made with the intent to further the conspiracy. Three reasons persuade me that it is simply too far of an inferential leap: the statement largely repeated what Richard already knew, the statement was presented to the court without any context, and Richard himself declined to suggest that Terror intended to bring him further into the conspiracy. On this record, I believe any inference that Terror feared Richard would inform the police about the murder and told him details about the murder in an effort to pull him into the conspiracy is too speculative to be reasonable. First, the statement itself did not give Richard any new material information and, thus, there was little reason to believe it would have actually drawn him any further into the conspiracy. There was evidence that Richard already knew of the plan to kill the victim, what the murder weapon was, where it had come from, and who had pulled the trigger. He admitted he was in the vehicle just before the murder, that he heard the defendant confess just after the murder, and that he went to the scene a few minutes after that, got into the vehicle with the victim’s body and tampered with evidence. The trial court agreed with the defendant that this was enough evidence against Richard to charge the jury on third-party culpability. In this light, the only new facts in Terror’s statement were that Terror leaned over the defendant while he prepared the gun and looked at the defendant before he fired. I cannot conclude that, given Richard’s already extensive knowledge of and involvement with the conspiracy, this new information was of such significance that we can reasonably infer that Terror’s imparting it to Richard was ‘‘designed to promote or facilitate’’ the goal of preventing him from going to the police, thereby furthering the conspiracy.12 Second, context does not help the state’s case here, either, because, in short, we have none. We know only that Terror supposedly made this statement at some point during Richard’s five or six days in New York (anywhere from two to eight days after the murder) and that this was the only conversation they had regarding the murder. When asked what he and Terror did in New York the rest of the time, Richard testified: ‘‘Nothing. Just walked around, and that was pretty much it.’’ None of this limited context suggests that Terror’s statement was intended to further the conspiracy to kill the victim. To the contrary, if Terror had intended to bring Richard further into the conspiracy by telling him more about it, presumably he would have mentioned it more than once over the course of Richard’s near weeklong stay. Third, Richard did not provide any insight into what Terror might have been thinking. The state failed to show that Richard believed Terror wanted to bring him further into the conspiracy, which was part of its proffer to the trial court.13 In sum, we barely know anything about Terror, whose intent is the subject of the inquiry. And without any context, we know even less about his intent in making this statement to Richard—certainly nothing that would take Terror’s statement beyond a ‘‘merely narrative’’ description of events, idle chatter or bragging. Instead, we know only that Richard already knew all of the material facts surrounding the murder and that he did not think Terror’s conduct in New York was ‘‘in furtherance’’ of anything in particular. Because I believe the ‘‘in furtherance’’ requirement ‘‘is a limitation on the admissibility of coconspirators’ statements that is meant to be taken seriously’’; (emphasis omitted) Garlington v. O’Leary, supra, 879 F.2d 283; I conclude that this falls short of a reasonable basis on which to admit this evidence against the defendant. Because of my previous conclusion that the defendant has demonstrated that the admission of this evidence harmed him, I would reverse the judgment of the trial court and remand the case for a new trial. I therefore respectfully dissent. 1 In some instances, courts have made exceptions because the hearsay declarant was a party and, under our adversary system, ‘‘it was thought that a party could not complain of the deprivation of the right to cross-examine himself (or another authorized to speak for him) or to advocate his own, or his agent’s, untrustworthiness.’’ Bourjaily v. United States, 483 U.S. 171, 190, 107 S. Ct. 2775, 97 L. Ed. 2d 144 (1987) (Blackmun, J., dissenting). Although such exceptions generally do not require an independent showing of reliability; e.g., Fed. R. Evid. 801, advisory committee notes, 28 U.S.C. app., p. 1063 (‘‘[n]o guarantee of trustworthiness is required in the case of an admission’’); they nonetheless are justified, at least in part, by their supposed trustworthiness. See 4 C. Mueller & L. Kirkpatrick, Federal Evidence (4th Ed. 2019) § 8:44 (‘‘important is the fact that the party himself is present and can explain, deny, or rebut any such statement’’). 2 I ask the reader to forgive my reversal of the usual order in opinion writing. Because the majority affirms the judgment of conviction on the coconspirator issue by assuming error and finding harmlessness, I meet the majority on its terms by addressing harm first. Because I conclude that the trial court erroneously admitted the coconspirator evidence and that this error was harmful, I do not reach the question of whether the trial court abused its discretion in admitting, under the state of mind exception to the hearsay rule, the victim’s statement to his friend, Tavaris Wylie, that he feared Piru. See part I A of this dissenting opinion. I note, however, that this court and other courts have traditionally viewed such statements skeptically. See, e.g., State v. Smith, 275 Conn. 205, 218, 881 A.2d 160 (2005). In my view, the facts of this case provide even more reason to doubt the worth of the victim’s hearsay. Therefore, in considering the strength of the state’s case, I consider the victim’s statement to be of limited value as proof of the defendant’s motive to murder him. The theory of admissibility in this context reasons that a victim’s statement of fear of a defendant is evidence that their relationship is collapsing. If the jury infers that their relationship is indeed collapsing, it may further infer that the defendant is so affected by the collapse that he has a plausible motive to commit murder. See, e.g., State v. Thomas, 205 Conn. 279, 285, 533 A.2d 553 (1987) (‘‘the purpose of the evidence was to show that the relationship had broken down, and that . . . the victim’s estrangement from the defendant supplied the motive for him to commit murder’’). A jury is likely to misunderstand or go beyond this permissible inference. See E. Prescott, Tait’s Handbook of Connecticut Evidence (6th Ed. 2019) § 8.16.2, p. 560 (‘‘[t]he failure to connect a declarant’s subjective state of mind to the issues in the case is not an uncommon error in logical analysis’’); see also 2 K. Broun, McCormick on Evidence (7th Ed. 2013) § 276, pp. 401–402 (‘‘the most likely inference that jurors may draw from the existence of fear, and often the only logical inference that could be drawn, is that some conduct of the defendant, probably mistreatment or threats, occurred and caused the fear’’). We have also recognized ‘‘the heightened potential for prejudice in cases such as this,’’ in which we ‘‘[allow] surrogates to speak for the victim pointing back from the grave.’’ (Internal quotation marks omitted.) State v. Smith, supra, 275 Conn. 218. In this case, any inference connecting the victim’s statement of fear to the defendant is particularly tenuous. We know almost nothing about the nature of the relationship between the victim and the defendant other than that they were members of the same gang. Moreover, because the victim had no apparent reason to fear the defendant specifically, the victim’s statement—that he feared ‘‘everybody’’ in the gang—was equally applicable to all forty members of Piru in New Haven at that time. Thus, although I do not reach the question of whether the trial court abused its discretion in admitting the statement, in assessing the strength of the state’s case, I believe it does little to prove that the defendant had a motive to murder the victim. 3 On appeal, the defendant does not challenge the sufficiency of the evidence to convict him. 4 Even Richard’s explanation about retrieving the cell phone raised questions. Specifically, he testified that about thirty minutes after he heard a gunshot, he met up with the defendant at the home of Davon Youmans, another Piru member. According to Richard, the defendant claimed to have left his cell phone in the victim’s car and asked Richard to get it for him. Richard went to the car, saw the victim, and removed a cell phone from the center console. On his way back to the house, he realized that he mistakenly had taken the victim’s phone from the car and therefore destroyed it. When Richard arrived back at the house, the defendant was gone. The idea that Richard would not only risk going to the crime scene, but also would get into the vehicle, seems highly counterintuitive because he could not know when the police might arrive. Nor did Richard explain how he later realized he had taken the wrong phone, why the defendant had left the house without his phone without waiting for Richard to return, or, most notably, why the police never found the defendant’s phone in the victim’s car. The questionable nature of this testimony not only diminished Richard’s credibility, but it also supported the defendant’s third-party culpability defense by insinuating that Richard had fabricated this testimony in order to inculpate the defendant and exculpate himself. 5 In closing argument to the jury, the prosecutor conceded that the state’s case depended on the testimony of Thomas and Richard: ‘‘[T]his case is not about the physical evidence. . . . There are two crucial witnesses in this case, and there’s no doubt about that. That is, Mr. Thomas and Mr. Richard. . . . [I]n this case, it’s whether or not you find those two individuals to be believable and credible.’’ On appeal, the state relies exclusively on their testimony as evidence of the defendant’s guilt. 6 We have also noted that deference to the fact finder is most appropriate when an ‘‘assessment of the credibility of the witnesses . . . is made on the basis of its firsthand observation of their conduct, demeanor and attitude.’’ (Emphasis added; internal quotation marks omitted.) State v. Kendrick, supra, 314 Conn. 223. Here, however, the evidence undermining the witnesses’ credibility—namely, various forms of self-interest, including the desire to lessen or eliminate their criminal liability—is apparent not from subjective firsthand observation, but objectively from the transcript and exhibits offered by the parties. We also have the determination of another fact finder—the trial judge—who observed the witnesses’ testimony and determined that Thomas and Richard were not credible. 7 Because Richard did not reside in Connecticut, he was flown here for trial by the state, which also provided him room and board. 8 The majority concludes that these inconsistencies are not ‘‘irreconcilable’’ by speculating that there were perhaps two meetings. Or that Thomas was mistaken about Richard’s presence. Anything is possible. But in considering the strength of the state’s case, I cannot simply overlook multiple discrepancies on significant details in the testimony of two crucial witnesses because an innocent explanation is ‘‘entirely possible . . . .’’ This conjecture does not leave me, as it does the majority, with any more of a fair assurance that the erroneous admission of the Terror statement did not substantially affect the verdict. 9 Section 8-3 of the Connecticut Code of Evidence provides in relevant part: ‘‘The following are not excluded by the hearsay rule . . . (1) . . . A statement that is being offered against a party and is . . . (E) a statement by a coconspirator of a party while the conspiracy is ongoing and in furtherance of the conspiracy . . . .’’ 10 Although agency is the most common justification offered for the coconspirator exception, other common justifications include reliability (i.e., a coconspirator’s statement made to further a conspiracy is a verbal act in the best interests of the conspiracy and, therefore, likely true) and necessity. See C. Mueller, ‘‘The Federal Coconspirator Exception: Action, Assertion, and Hearsay,’’ 12 Hofstra L. Rev. 323, 335 (1984) (‘‘[s]ince conspiracies are dangerous to society and hard to prove at trial, a relaxation of the hearsay doctrine is required’’). 11 ‘‘Case law suggesting that courts interpret the phrase in furtherance of the conspiracy broadly, should not be viewed as adding anything to the rule, but instead as respecting its broad verbiage. Indeed, the Advisory Committee’s comments on the rule suggest the opposite intent. Courts, consequently, often interpret the rule’s requirements, and particularly the in furtherance requirement, strictly.’’ (Footnotes omitted; internal quotation marks omitted.) 30B C. Wright et al., supra, § 6780. Even so, many argue that the ‘‘in furtherance’’ requirement ‘‘seems an imperfect measure’’ of reliability. C. Mueller, supra, 12 Hofstra L. Rev. 335. Justice Thurgood Marshall has written: ‘‘That a statement was truly made in furtherance of a conspiracy cannot possibly be a guarantee, or even an indicium, of its reliability. . . . The conspirator’s interest is likely to lie in misleading the listener . . . . It is no victory for common sense to make a belief that criminals are notorious for their veracity the basis for law.’’ (Citations omitted; internal quotation marks omitted.) United States v. Inadi, supra, 475 U.S. 404–405 (Marshall, J., dissenting); see also C. Mueller, supra, 357 (‘‘[a] statement may actually further a conspiracy simply by being plausible to its audience, which means that it may well fit within the circumstances without being true’’); D. Davenport, ‘‘The Confrontation Clause and the CoConspirator Exception in Criminal Prosecutions: A Functional Analysis,’’ 85 Harv. L. Rev. 1378, 1387 (1972) (‘‘[m]any statements actually in furtherance of an alleged conspiracy will be quite unreliable in whole or in part’’). 12 Unlike the majority, I consider there to be a fundamental difference between (1) evaluating, for evidentiary purposes, whether Terror’s one hearsay statement describing the murder furthered the conspiracy by recruiting Richard, and (2) evaluating, for harmless error purposes, the weight Terror’s two hearsay statements ordering the murder and describing the murder carried with the jury. See part I B of this dissenting opinion. 13 In initially arguing against the defendant’s motion in limine to preclude Terror’s statement, the prosecutor suggested that bringing Richard to New York manifested Terror’s concern that he might go to the police: ‘‘Richard, at this point, has left Connecticut and has now been in New York with the head Piru guy [Terror], and there’s a concern on behalf of that head Piru guy. And, I think, it could be inferred, there’s a concern on him on whether or not they are going to let [Richard] go back to . . . Connecticut or whether he’s going to talk to the police.’’ The state offered no proof at that time to support the inference that Terror was concerned about Richard going to the police. The next day, however, the state tried to bolster its claim by asking Richard about whether Terror and other members of Piru made Richard fear for his safety while in New York. The defendant renewed his objection. At this point, the state proffered that Richard would testify that he felt compelled to go to New York out of fear for his safety, which was connected to his loyalty to the gang: The prosecutor argued to the court: ‘‘The witness has been very clear that he felt pressured by the defendant to do certain acts and felt that he had to go to New York with Terror. It’s [the] same sides of the same coin. It’s this loyalty, but also concerns for his safety.’’ But Richard never said this on the witness stand. Instead, when asked why he went to New York, he responded only that it was ‘‘to prove [his] loyalty’’ to Terror, his superior. When asked to draw the connection between gang loyalty and safety that the state had relied on, he did not do so and replied only in general terms: ‘‘Well, I mean, when you’re in a gang, you know, or when you’re in the streets, period; there’s certain people, you know, you got to watch because it’s dangerous.’’ Contrary to the state’s representation, at no point did Richard say or even imply that he felt frightened of Terror or that Terror tried to bring him further into the conspiracy.