The People of the State of New York, Respondent, v  Mauricio Altamirano, Appellant.

Argued November 20, 2019; decided December 17, 2019

People v Mairena, 160 AD3d 986, affirmed.

People v Altamirano, 61 Misc 3d 1, affirmed.

It is not disputed that the trial courts in both of these cases erred by reversing—after summations—their prior rulings on defendants' requests to charge the jury. Both courts erred by failing to charge the jury in accordance with their pre-summation rulings on defendants' charging requests. The question before us on these appeals is whether these errors mandate reversal. We hold that the error was harmless in both cases and, therefore, affirm.

After defendant David Mairena and the victim were asked to leave a restaurant where they had engaged in a late-night fight (referred to by defendant as the "first attack"), the victim returned with a machete, hit defendant with it and then chased him across the street (referred to by defendant as the "second attack"). The manager of the restaurant, armed with a baton, ran across the street, grabbed the victim and dragged him back to the front of the restaurant. As the victim was pulled away from defendant, he pointed the machete at defendant. The manager told [*2]defendant—who followed the victim back across the street—to go home. The manager then flagged down a passing police car, told the victim to drop the machete and, once he saw police coming, returned to his work. Another witness, a disc jockey, did not see the victim drop the machete, but{**34 NY3d at 478} he heard "a sound like when metal hits the floor" and did not see the victim holding the machete any longer as the victim walked away.

In what defendant refers to as the "third attack," he stepped onto the sidewalk and loudly stated something to the victim, who turned around and took off his jacket; defendant and the victim then began punching each other. The disc jockey saw the victim start to bleed from part of his right arm and fall to the sidewalk, while there was a "noise, like a bottle breaking." Defendant fled, and the police approached, observing the victim with blood "spraying" from his arm. An officer called an ambulance and attempted to stop the bleeding, but the victim lost consciousness shortly thereafter and was pronounced dead approximately 30 minutes after the officers arrived. All three interactions were captured on surveillance video, which was introduced into evidence at trial.

Officers returning to the scene the next morning found the machete on the street near "tiny bits and pieces of broken glass" that had no blood on them. When police spoke to defendant, he admitted that, before fleeing, he took a knife out of his pocket and swung it at the victim. Defendant was thereafter charged with manslaughter in the first degree, assault in the first degree and criminal possession of a weapon in the fourth degree. At trial, the medical examiner testified that the cause of the victim's death was blood loss from a severed brachial artery in the right arm. Although the medical examiner indicated that the four-inch long incised wound was "not inconsistent" with having been caused by a broken bottle, she testified that the wound was "certainly more consistent [with] having been sustained by [a] sharp instrument such as a knife." Defendant testified that, during the third attack, he feared for his life when the victim shed his jacket and approached. Therefore, although defendant did not see whether the victim had a weapon at that point, he pulled out the knife and swung it at the victim. Once home, defendant saw blood on the knife and washed it off. At trial, defendant's theory was that the victim's fatal wound was the result either of falling on a glass bottle or, if caused by a stabbing, that defendant's use of deadly physical force was justified.

At the charge conference, the People sought an expanded charge on intent—that the jury should consider whether the result of defendant's conduct was "the natural, necessary and{**34 NY3d at 479} probable consequence of that conduct"—and defendant requested that the People be precluded from arguing that, if the victim's injury was caused by falling on a glass bottle on the ground, his resulting death was a natural consequence of defendant's actions. As defendant noted, that theory of prosecution was not presented to the grand jury. Defendant also requested a specific instruction that, to convict him of first-degree manslaughter, the jury had to find that the victim's death was caused by a knife, specifically a box cutter. The court responded that it would charge that the jury had to find that defendant "[c]aused [the victim's] death with a dangerous instrument, to wit, a knife or a . . . box cutter." On summation, defendant argued that the jury had to acquit him of the manslaughter charge if the People failed to prove beyond a reasonable doubt that the instrumentality of death was the box cutter, as opposed to a bottle on the ground, and devoted a substantial portion of his summation to attempting to demonstrate that the victim died from falling on a broken bottle. The People argued in summation that defendant intended to cause the victim serious physical injury and that the victim's "death was actually caused by the defendant slashing [the victim] with that blade and not by anything else."

When charging the jury, the court failed to include the agreed-upon language that the jury could convict defendant of manslaughter only if it found that the victim's death was caused by a box cutter. Defense counsel objected on the ground that he had structured his summation in a particular way in anticipation of the promised charge. The court responded "[t]hat's what was argued"—that if the jurors found "it wasn't the box cutter, they have to find him not guilty." Defense counsel made no further objections or requests. The jury convicted defendant of first-degree manslaughter and fourth-degree criminal possession of a weapon. The Appellate Division affirmed, holding as relevant here that, "although it was error for the court to inform the parties, prior to summations, that it would instruct the jury on a specific instrumentality of death in its charge of manslaughter in the first degree, and then to subsequently remove that language from its charge following summations, . . . the error was harmless" (160 AD3d 986, 988 [2d Dept 2018]). A Judge of this Court granted defendant leave to appeal.{**34 NY3d at 480}

Defendant Mauricio Altamirano was convicted of criminal possession of a weapon in the fourth degree for storing another person's operable, but unloaded, .22 caliber revolver in his apartment. At trial, beginning with opening statements, counsel presented a temporary and lawful possession defense. Counsel argued that defendant was guilty only of "trying to do a favor for a friend," and that he did not know that his friend—who he knew only as[*3]"Columbia"—had a gun in a bag that he asked defendant to store in his home. Counsel asserted that, when defendant realized there was a gun in the bag, he twice asked Columbia to remove the gun from his home, to no avail, and later cooperated fully with police after they approached him to inquire about the gun.

The evidence at trial demonstrated that, after Columbia was arrested in connection with an unrelated assault, officers learned that he owned a gun and that defendant had a "connection to that" gun. Officers in plain clothes approached defendant at his place of business and asked if he was aware of or possessed a gun. Defendant, who was "extremely cooperative," offered to accompany police to his apartment to surrender the gun, and signed a consent to search form. He directed them to the weapon, which was wrapped in a blanket inside a garbage can in the single-room apartment. Officers took the gun and placed defendant under arrest. The People proffered defendant's statement to police, in which he explained that he was holding a bag for his acquaintance Columbia for at least three weeks. Defendant informed police that, after discovering that the package Columbia gave him contained the gun, he had told Columbia to remove it multiple times, but Columbia failed to do so.

Defendant requested that the court charge the jury on temporary and lawful possession of a weapon.^[FN1] Defendant argued that his statement, along with the evidence that he{**34 NY3d at 481} turned the weapon over to police at the first opportunity presented to him, provided a basis for the charge. The court denied the request on the ground that the evidence did not support lawful possession, but also refused the People's request to prohibit defendant from referring to temporary possession in summation. Counsel, relying on defendant's statements to police, then argued in summation that defendant innocently permitted Columbia to place a bag in his home without being aware that a gun was concealed inside, demanded that the gun be removed when he learned that the weapon was in the bag, and later fully cooperated with the police when they confronted him about the weapon, conduct which was consistent with that of someone who was innocent. Counsel also argued that there was insufficient evidence regarding the time period between defendant's discovery that the bag in his home contained a gun and when he turned the weapon over to police.

After summations, without first informing the parties, the court gave the temporary and lawful possession charge to the jury. Thereafter, outside the presence of the jury, the court explained that it decided to give the charge out of an abundance of caution, but it denied defendant's request to reopen his summation so that he could "reargue to the jury the charge." Defendant was convicted as noted above. Upon defendant's appeal, the Appellate Term affirmed, concluding that Criminal Court "deprived defendant of the right to an effective summation," but that "reversal of the judgment of conviction [was] not required" because "defendant was not entitled to a charge on the defense of temporary and innocent possession of a weapon based on the facts of the case" (61 Misc 3d 1, 5 [App Term, 2d Dept, 2d, 11th & 13th Jud Dists 2018]). A Judge of this Court granted defendant leave to appeal.

In Herring v New York, the Supreme Court of the United States explained that "[t]he Constitutional right of a defendant {**34 NY3d at 482}to be heard through counsel necessarily includes [the] right to have [defense] counsel make a proper argument on the evidence and the applicable law in [defendant's] favor" (422 US 853, 860 [1975] [internal quotation [*4]marks omitted]). The Herring Court ruled invalid a New York "statute that empower[ed] a trial judge to deny absolutely the opportunity for any closing summation at all" (id. at 863). The Court held that the complete denial of an opportunity to make a summation of the evidence is a violation of the Sixth Amendment because "closing argument for the defense is a basic element of the adversary factfinding process in a criminal trial" (id. at 858). That is, the "right to be heard in summation of the evidence from the point of view most favorable to" a defendant is implicit in the right to "the assistance of counsel that the [Federal] Constitution guarantees" (id. at 864-865). Indeed, "no aspect of . . . advocacy could be more important than the opportunity finally to marshal the evidence for each side before submission of the case to judgment" (id. at 862).

However, the rule "is not . . . that closing arguments in a criminal case must be uncontrolled or even unrestrained" (Herring, 422 US at 862). Rather, the Supreme Court explained in Herring that the trial "judge must be and is given great latitude in controlling the duration and limiting the scope of closing summations" (id.; see People v Ashwal, 39 NY2d 105, 109 [1976] ["It is fundamental that the jury must decide the issues on the evidence, and therefore fundamental that counsel, in summing up, must stay within the four corners of the evidence and avoid irrelevant comments which have no bearing on any legitimate issue in the case" (internal quotation marks and citation omitted)]). Moreover, although "[t]here is no way to know whether . . . appropriate arguments in summation might have affected the ultimate judgment in [a] case" (422 US at 864), the Supreme Court has since emphasized in Glebe v Frost that "[n]one of [its] cases clearly requires placing improper restriction of closing argument in [the] narrow category" of structural errors (574 US 21, 23 [2014]), as opposed to a "trial error . . . reviewable for harmlessness" (id. at 22). Of course, even "[m]ost constitutional mistakes call for reversal only if the government cannot demonstrate harmlessness" (id. at 23). It is "[o]nly the rare type of error—in general, one that infect[s] the entire trial process and necessarily render[s] [it] fundamentally unfair—[that] requires automatic reversal" (id. [internal quotation marks and citation omitted]). Although Glebe was decided pursuant{**34 NY3d at 483} to the deferential standard applied under the Antiterrorism and Effective Death Penalty Act of 1996 and, thus, leaves open the question of whether harmless error analysis applies to trial court rulings that operate to restrict counsel's strategic choices regarding summation, it nevertheless makes clear that Herring does not compel a conclusion that "the restriction of summation . . . amounts to structural error" (574 US at 24). Of course, regardless of whether the error is structural, "summation matters" as the dissent puts it (dissenting op at 499), and "the right of the defense to make a closing summary of the evidence to the trier of the facts" implicates the Sixth Amendment (Herring, 422 US at 860). Nevertheless, short of a complete deprivation of the right to present a summation (see id.; People v Harris, 31 NY3d 1183, 1185 [2018]), it cannot be said that Supreme Court precedent requires us to reconsider our well-established case law holding that the improper reversal of a prior charging ruling that impacts summation is subject to harmless error analysis.

In that regard, this Court's decisions in People v Miller (70 NY2d 903 [1987]), People v Greene (75 NY2d 875 [1990]), and People v Smalling (29 NY3d 981 [2017]) establish that, under New York law, harmless error analysis applies when the trial court fails to inform the parties, prior to summations, of the charge that will be given to the jury. In Miller, we expressly applied harmless error analysis to such an error. We explained that, where it could "[ ]not be said that defense counsel's summation would have been affected by knowledge that the petit larceny charge would be submitted to the jury," the error "was, in the circumstances of th[e] case, harmless" (70 NY2d at 907). While we did not use the specific phrase "harmless error" in Greene, we held that it was "error, prejudicial to defendant" when the trial court "reverse[d] its stance after assuring defendant that it would charge as he requested and after defendant had premised his summation on that theory" (75 NY2d at 877). Most recently, in Smalling, the Court relied on Greene in holding that, when the trial court "agreed to the People's request at the charge conference not to charge the jury on constructive possession, but then ultimately provided a constructive possession charge to the jury, resulting in prejudice to defendant," the "defendant [was] entitled to a new trial" (29 NY3d at 982). Smalling expressly applied harmless error analysis, stating that "[u]nder the unique circumstances of this case, the error is not harmless" (id.).

{**34 NY3d at 484}Our reliance on Greene in continuing to apply harmless error analysis in Smalling—decided just two years ago—completely refutes defendants' arguments that our precedent distinguishes between a determination of prejudice and a harmless error analysis in this context. Contrary to defendants' arguments, Miller, Greene and Smalling clearly establish that harmless error analysis applies and that the test for the prejudice component of harmless error analysis under these circumstances is whether it can be said that defense counsel's summation would have been materially affected by knowledge of the charge ultimately submitted to the jury (cf. Harris, 31 NY3d at 1185 [declining to apply harmless error analysis when the defendant was completely denied the opportunity to give a closing [*5]argument]).^[FN2] A showing of prejudice is necessary because it is not every post-summation change in instruction that amounts to a material misdirection about the legal framework of a case or that will have any sort of meaningful effect on defense counsel's summation (see Miller, 70 NY2d at 907). The fact that our prior cases did not expressly consider whether the evidence of guilt was overwhelming does not suggest that those cases were not, in fact, applying harmless error analysis, as the Court expressly stated that it was doing in both Miller and Smalling. That is because a determination that the evidence was overwhelming does not, in itself, answer the question of whether an error is harmless under People v Crimmins (36 NY2d 230 [1975]).

Under our traditional harmless error analysis, an appellate court does not reach the question of prejudice unless the evidence is overwhelming in the first instance. As we explained in Crimmins, "unless the proof of the defendant's guilt, without reference to the error, is overwhelming, there is no occasion for{**34 NY3d at 485} consideration of any doctrine of harmless error" (36 NY2d at 241). "That is, every error of law (save, perhaps, one of sheerest technicality) is, ipso facto, deemed to be prejudicial and to require a reversal, unless that error can be found to have been rendered harmless by the weight and the nature of the other proof" (id.). Under Crimmins, when "an appellate court has satisfied itself that there was overwhelming proof of the defendant's guilt, its inquiry does not end there. . . . Further inquiry must . . . be made by the appellate court as to whether, notwithstanding the overwhelming proof of the defendant's guilt, the error infected or tainted the verdict," and "[a]n evaluation must therefore be made as to the potential of the particular error for prejudice to the defendant" (id. at 242 [emphasis added]). When an error is not constitutional in nature, the "error is prejudicial . . . if the appellate court concludes that there is a significant probability . . . in the particular case that the jury would have acquitted the defendant had it not been for the error or errors which occurred" (id.). If the error at issue is of a constitutional dimension, it is harmless when "there is no reasonable possibility that the error might have contributed to defendant's conviction" (id. at 237). In Miller, Greene and Smalling, we expressly addressed only the second step of the Crimmins analysis, determining whether the error was prejudicial or harmless to defendant, having necessarily concluded, sub silentio, that the evidence was overwhelming. Our cases establish that prejudice in this context, for purpose of harmless error analysis, turns on whether defense counsel structured summation based on an anticipated charge that was not conveyed to the jury as promised, such that summation was materially affected by an alteration in that anticipated charge.

Our prior decisions do not clarify whether the constitutional or nonconstitutional harmless error standard applies in determining whether the defendant was prejudiced by a trial court's reversal, after summations, of a prior ruling on a charging request. However, Miller, Greene and Smalling represent this Court's consistent application of harmless error analysis in this context over the course of 30 years and, as explained above, these decisions are not contrary to the decisions of the Supreme Court in Glebe and Herring. In addition, defendants' assertion that the rule in all four Appellate Division Departments is that harmless error analysis does not apply to a broken charging promise is patently meritless. The First,{**34 NY3d at 486} Second and Fourth Departments of the Appellate Division all have expressly held that these types of errors may be "harmless" (see People v Gonzalez-Alvarez, 129 AD3d 647, 648 [1st Dept 2015], lv denied 27 NY3d 997 [2016]; People v Lugo, 87 AD3d 1403, 1404 [4th Dept 2011], lv denied 18 NY3d 860 [2011]; Matter of Jose R., 185 AD2d 819, 820 [2d Dept 1992]). The single case cited by defendants that supports their argument—People v Bacalocostantis (111 AD2d 991, 992 [3d Dept 1985] [holding that harmless error analysis as set forth in Crimmins cannot apply to the failure to inform counsel, prior to summations, that lesser included charge would be given])—is no longer good law inasmuch as it was decided before this Court expressly applied harmless error analysis in Miller to the same error that was present in Bacalocostantis. In addition, the Court that decided Bacalocostantis—the Appellate Division, Third Department—has held subsequent to Bacalocostantis that [*6]reversal is not required where there is no prejudice (see People v Carkner, 213 AD2d 735, 739 [1995], lv denied 85 NY2d 970 [1995]; People v Seiler, 139 AD2d 832, 834 [1988], lv denied 72 NY2d 924 [1988]).

In short, Miller, Greene and Smalling have consistently been applied by the appellate courts of this state and continue to be entitled to full precedential force. In those decisions, this Court meant what it expressly stated: a trial court's error in reversing a prior charging decision after summations have been completed is subject to harmless error analysis.

We conclude that the evidence of guilt in both of the instant cases was overwhelming. Thus, as in Miller, Greene and Smalling, whether the error was harmless turns on the question of whether defendants were prejudiced. Although those cases do not clarify whether the constitutional or nonconstitutional standard applies in evaluating prejudice, we need not resolve that question today because, under either standard, the error in each case was harmless.

In People v Mairena, defendant argues that, by revoking its promise to specifically instruct the jury as to the nature of the sharp instrument that caused the victim's injuries, the court undermined the strategic decisions counsel made when preparing his summation based on that promise. Defendant asserts{**34 NY3d at 487} that counsel argued that the People failed to prove that the victim died from a knife wound, rather than from falling to the ground on top of a glass bottle, and that, without the instruction, the jury could not have understood that his argument mandated acquittal. Defendant claims that, if he had known the court would not give the requested instruction, he may have avoided addressing the "bottle theory" altogether and focused solely on his justification defense.

To be sure, the facts of this case are superficially similar to those in Greene. The trial court in Greene agreed to charge the jury, with respect to manslaughter in the first degree, that it must acquit if it found that victim's injury was not caused by a shooting. In summation, counsel "emphasized the equivocal nature of the proof that a shooting occurred and told the jury that they should acquit defendant if there was a reasonable doubt that the victim's wound was caused by a gunshot" (75 NY2d at 877). We reversed because the trial court's ultimate instruction to the jury contravened its promised charge, and the defendant had premised his summation on the theory underlying the charge. Similarly here, the trial court promised a charge on the instrumentality of death with respect to manslaughter in the first degree, and then failed to deliver the specific language to which it had agreed.

Nevertheless, an affirmance is warranted. As the People note, defendant's primary motivation in seeking the charge was to prevent the People from arguing that he was guilty even if a fall onto glass bottles during the fight was the cause of death—a theory of prosecution that was not presented to the grand jury. The People not only refrained from arguing that theory, but they expressly argued to the jury that the victim's "death was actually caused by the defendant slashing [the victim] with that blade and not by anything else." Moreover, the jury charge on manslaughter, read as a whole, conveyed that the jury had to find the victim's death was caused by defendant's intentional actions in using the box cutter in order to convict; it did not permit the jury to convict defendant of manslaughter based on the bottle theory. With respect to manslaughter, the jury was instructed that it had to find that defendant caused the victim's death while specifically intending to cause serious physical injury to the victim. Serious physical injury was defined as "impairment of a person's physical condition which creates a substantial risk of death or which causes death or which causes serious and protracted disfigurement,{**34 NY3d at 488} protracted impairment of health or protracted loss or impairment of the function of any bodily organ." Had the victim's death been caused by an accidental fall on glass bottles during a fistfight that lasted only several seconds, the jury could not have found that defendant intended to cause serious physical injury.

Thus, although the specific language requested by defendant was not included in the jury charge, given the arguments actually made in summation by both parties that the jury had to find that the instrumentality of death was the box cutter in order to convict, and the conveyance of that instruction to the jury in the charge as a whole, any error in Mairena was harmless. Counsel essentially received the charge that he sought and both parties' closing arguments were entirely consistent with that charge. Under these circumstances, "it cannot be said that defense counsel's summation would have been affected by knowledge [of] the . . . charge [that] would be submitted to the jury" (Miller, 70 NY2d at 907).

In People v Altamirano, defendant maintains that the court's denial of his request for a jury instruction on temporary and lawful possession deprived him of the opportunity to argue that his possession of the weapon was not unlawful. He argues that, had counsel known the temporary and lawful possession charge would be given, he would [*7]have drawn the jury's attention to the factors set forth in the charge contained in the pattern criminal jury instructions for determining whether his possession of a weapon in his home was innocent. Specifically, counsel would have pointed to a lack of evidence of knowing possession, the manner in which the gun came into defendant's possession, the length of time it was in defendant's possession, whether defendant had the intent to use it or dispose of it, and whether defendant took advantage of any opportunity to turn it over to police. Instead, defendant contends, counsel was forced to make a nonsensical argument that he never had constructive possession of the gun.

Under Miller, Greene and Smalling, the relevant question is whether counsel's summation was materially affected in a manner that prejudiced defendant in light of the charge actually given. As the People correctly note, defense counsel did, in summation, highlight the same evidence that defendant now argues would have supported the temporary and lawful possession{**34 NY3d at 489} charge. Counsel argued innocent possession throughout the trial. During summation, he asserted that defendant did not know that a gun was concealed inside Columbia's bag, defendant had unsuccessfully demanded that the gun be removed once he found out that it was in the bag, and defendant's conduct was that of someone who was innocent—i.e., he was fully cooperative with police and turned the gun over to them at his first opportunity. Counsel also pointed out that there was insufficient evidence regarding the time period between defendant's discovery of the gun stored in his home and his turning it over to police. In other words, defendant made the argument in substance that he now claims he was forced to forgo due to the trial court's initial refusal to give the temporary and lawful possession charge; he connected the evidence at trial to the factors relevant to the defense of temporary and lawful possession as it was ultimately charged to the jury. Thus, under these circumstances, "defense counsel's summation would [not] have been [materially] affected by knowledge [of] the . . . charge [that] would be submitted to the jury" (Miller, 70 NY2d at 907) and, thus, the error was harmless.

We have considered defendant's remaining argument in Mairena and conclude that it is lacking in merit. Accordingly, the orders in both Mairena and Altamirano should be affirmed.

Fahey, J. (concurring). "Errors of law of nonconstitutional magnitude may be found harmless where 'the proof of the defendant's guilt, without reference to the error, is overwhelming' and where there is no 'significant probability . . . that the jury would have acquitted the defendant had it not been for the error' " (People v Byer, 21 NY3d 887, 889 [2013] [emphasis added], quoting People v Crimmins, 36 NY2d 230, 241-242 [1975]). By contrast, errors of law of constitutional dimension may only be found harmless where the evidence is overwhelming and "there is no reasonable possibility that the error might have contributed to [the] conviction" (id. at 237 [emphasis added]).

I agree with my colleagues in the majority to the extent they conclude that the errors here are harmless,^[FN*] but in my view the constitutional standard for harmless error should apply (cf.{**34 NY3d at 490} majority op at 486). These errors have a constitutional dimension and they must be reviewed under the heightened harmless error standard applicable to constitutional mistakes. Applying that standard, the evidence in each case is overwhelming and there is no reasonable possibility that the mistakes contributed to the convictions (see Crimmins, 36 NY2d at 237).

The opportunity to give an effective summation is an essential part of the fundamental right to counsel. Closing arguments sharpen the issues for resolution by the trier of fact and are "the last clear chance to persuade the trier of fact that there may be reasonable doubt of the defendant's guilt" (Herring v New York, 422 US 853, 862 [1975]). "The very premise of our adversary system of criminal justice is that partisan advocacy on both sides of a case will best promote the ultimate objective that the guilty be convicted and the innocent go free" (id.). A criminal proceeding is a fact-finding process, and few, if any, aspects of trial can be more important than a final summary of the evidence delivered on the precipice of deliberations and verdict (see id.).

The barriers to effective summations erected by the trial courts here implicated the right to counsel. The right to counsel is guaranteed by both the Federal and State Constitutions (see People v Benevento, 91 NY2d 708, 711 [1998]) and "extends to closing arguments" (Yarborough v Gentry, 540 US 1, 5 [2003]). Simply having counsel is not enough. The accused is entitled to counsel who provides what we have characterized as " 'effective' aid" (see Benevento, 91 NY2d at 711, quoting Powell v Alabama, 287 US 45, 71 [1932]). That assistance has alternately been described as "meaningful representation" (People v Baldi, 54 NY2d 137, 147 [1981]).

For the courts to hamstring counsel on summation was to limit their effectiveness at a critical stage of the proceedings and to strike at defendants' basic right to a fair trial (see US Const Amend VI; NY Const, art I, § 6). It is error for a court to fail to give an agreed-upon charge or to add a charge not discussed with counsel. The parties have a right to know what the court will charge in order to prepare their summations. Otherwise, counsel would proceed blindly into summations, leaving the accused in a position where their legal assistance in that potentially decisive segment of the proceedings could be meaningless (see generally Evitts v Lucey, 469 US 387, 396 [1985] ["a party whose counsel is unable to provide effective representation is in no better position than one who has no counsel at all"]).{**34 NY3d at 491}

Reason dictates that this created a "reasonable possibility that the error[s] might have contributed to [the] conviction[s]" (Crimmins, 36 NY2d at 237). It is only the overwhelming evidence in each case that allows these convictions to stand.

Rivera, J. (dissenting). The constitutional guarantees of the right to counsel and a fair trial include the opportunity to present an effective summation in support of acquittal. It is the only chance for counsel to marshal the evidence from the defense perspective based on the law as charged to the jury. Counsel's strategic choices of how best to weave legal arguments, facts and inferences for maximum persuasive impact are reflected in the closing statement. No other moment in a trial provides counsel the great leeway for exposition and rhetorical flourish grounded in logic and counsel's own voice as that allowed in this last chance for partisan advocacy. Whether a dispassionate recitation of the facts and law or a colorful display of oratorical skill, an effective summation can plant the seeds of reasonable doubt.

Counsel in these two appeals charted the course for their summations based on the respective trial judge's promised charge to the jury. In both appeals, we unanimously agree that the judges' change in course after summation constitutes error. The majority concludes the errors are harmless. However, the majority does not actually deploy our traditional harmless error analysis. Instead, it considers the impact of the trial judges' broken promises on counsel's summation. I agree with the substance of the majority's "harmless error" standard, if not its label. However, I conclude application of the standard here requires reversal of defendants' convictions.

Defendants contend that they were denied the right to an effective summation at their respective trials when their counsel prepared and delivered closing arguments in reliance on the judges' pre-summation charge determinations, which the judges subsequently failed to follow. In one case, the judge reneged on a promise to charge an element of the crime in the form requested by defendant and, in the other, the judge denied counsel's requested charge but gave it after summations.

Defendant David Mairena was tried for manslaughter in the first degree (Penal Law § 125.20 [1]), assault in the first degree {**34 NY3d at 492}(Penal Law § 120.10 [1])^[FN1] and criminal possession of a weapon in the fourth degree (Penal Law § 265.01 [2]) for the death of Miguel Jimenez, which resulted from an early morning fight outside a bar and restaurant. The fight that led to the fatal injury occurred after defendant and the victim were engaged in two other altercations, one of which included the use of a machete by the victim. The police and medical examiner concluded that the victim died from a cut in his right arm that severed his brachial artery. During the investigation, defendant admitted to the police that he had taken out and swung a box cutter during the fatal altercation. One of the witnesses, a disc jockey, also heard a "noise, like a bottle breaking" when the victim fell and saw blood streaming from the victim's right arm. At the charge conference, the People requested an expanded intent charge for the manslaughter count, in particular that the jury could convict defendant of manslaughter if the victim's fatal injury was caused by falling on broken glass bottles and this fall was the natural, necessary and probable consequence of defendant's actions. Defendant opposed the expanded intent charge because this theory of culpability was not presented to the grand jury. Defendant also requested that the court instruct the jury that they could only find defendant guilty of manslaughter if [*9]he caused the fatal wound with a knife, specifically a box cutter. The court agreed to charge that the jury must find defendant "[c]aused death with a dangerous instrument, to wit, a . . . box cutter."

In summation, counsel focused on two alternative defenses: (1) defendant did not use the box cutter to inflict the fatal injury in the victim's right arm and the victim died from falling on a broken bottle; or (2) if defendant slashed the victim in the arm, he was justified. Defense counsel spent approximately 15 pages of his summation arguing that the victim may have died from falling on a broken bottle, which created "reasonable doubt" necessitating a "not guilty" verdict. Counsel discussed the victim's injury in depth, replayed the surveillance video of the fatal altercation, and displayed the autopsy photo of the victim to the jury. Contrary to its promise, the judge did not charge the jurors that they could find defendant guilty of manslaughter only if they concluded the fatal injury was caused by{**34 NY3d at 493} a box cutter. The jury found defendant guilty of the submitted counts of manslaughter and criminal weapon possession.

Defendant Mauricio Altamirano was tried for criminal possession of a weapon in the fourth degree (Penal Law § 265.01 [1]) and attempted tampering with physical evidence (Penal Law §§ 110.00, 215.40 [2]). These charges stem from defendant's agreement to store a bag at his apartment as a favor for a friend named "Columbia." Unbeknownst to defendant at the time that he agreed to the favor, the bag contained a gun. After he learned about the weapon, defendant asked Columbia to remove the gun at least twice, but Columbia never did. While investigating an assault involving Columbia, plainclothes officers approached defendant at his work and inquired about the gun. Defendant was "extremely cooperative," admitted to having a gun wrapped in a blanket in a garbage can inside his home and took the officers to his apartment to surrender the gun. Defendant requested that the court charge the jury on temporary and lawful possession of a weapon, arguing that there was strong evidence presented at trial that defendant did not know that the bag given to him by Columbia contained a gun and that he turned the gun over to police at the first available opportunity. The judge denied the request but then gave the charge after summations and refused counsel's request to reopen summations so that he could address the lawful possession based on the instruction. Defendant was acquitted of the attempted evidence tampering count and convicted of the criminal possession of a weapon count.

The intermediate appellate courts affirmed defendants' convictions, and by separate orders of different individual Judges of this Court, each defendant was granted leave to appeal (People v Mairena, 160 AD3d 986 [2d Dept 2018], lv granted 31 NY3d 1150 [2018]; People v Altamirano, 61 Misc 3d 1 [App Term, 2d Dept, 2d, 11th & 13th Jud Dists 2018], lv granted 32 NY3d 1201 [2019]).

Before this Court, as they did below, defendants maintain that their claims have doctrinal roots in the fundamental constitutional rights to assistance of counsel by means of effective summation, and a fair trial.^[FN2] They contend judicial review of the claimed violation of those rights based on the judge's{**34 NY3d at 494} deviation from the promised charge must consider the unique role summation plays in our adversarial fact-finding system of justice. They further contend that a trial judge's sharp reversal from the agreed upon charge is bad policy, undermining both parties' ability to make informed strategic decisions and the legislature's intent in mandating jury charge determinations prior to summations pursuant to CPL 300.10 (4). The People counter that a judge's failure to abide by the jury charge promise is subject to traditional harmless error analysis.

Defendants propose a rule whereby (1) when a defendant relies on a court's charging promises in summation and the court does the opposite of what was promised after summation, a defendant's right to an effective summation has been violated and (2) reversal is required when the court does not or cannot give the defendant an opportunity to cure the prejudice, either by denying counsel the opportunity to reopen summation or by refusing to give the promised charge. I agree that these errors are constitutional in nature and appellate review should focus on whether the errors adversely impacted counsel's summation. This standard of review aligns with our prior approach (see People v Greene, 75 NY2d 875 [1990]). In contrast, traditional harmless error doctrine is both doctrinally inapplicable and ill-suited in practice to review the impact of a trial judge's decision not to follow the pre-summation charge determination.

The United States Supreme Court has eschewed "a narrowly literalistic construction" of the federal constitutional right to assistance of counsel (Herring v New York, 422 US 853, 857 [1975]).

"The right to the assistance of counsel has thus been given a meaning that ensures to the defense in a criminal trial the opportunity to participate fully and fairly in the adversary factfinding process.

"There can be no doubt that closing argument for the defense is a basic element of the adversary factfinding process in a criminal trial. Accordingly, it has universally been held that counsel for the defense has a right to make a closing summation to the jury, no matter how strong the case for the prosecution{**34 NY3d at 495} may appear to the presiding judge" (id. at 858 [footnote omitted]).

Why is summation of such importance to our system of criminal justice and the fact-finding process? Because, as the Supreme Court has explained,

"[C]losing argument serves to sharpen and clarify the issues for resolution by the trier of fact in a criminal case. For it is only after all the evidence is in that counsel for the parties are in a position to present their respective versions of the case as a whole. Only then can they argue the inferences to be drawn from all the testimony, and point out the weaknesses of their adversaries' positions. And for the defense, closing argument is the last clear chance to persuade the trier of fact that there may be reasonable doubt of the defendant's guilt.

"The very premise of our adversary system of criminal justice is that partisan advocacy on both sides of a case will best promote the ultimate objective that the guilty be convicted and the innocent go free. In a criminal trial, which is in the end basically a factfinding process, no aspect of such advocacy could be more important than the opportunity finally to marshal the evidence for each side before submission of the case to judgment" (id. at 862 [citation omitted]).

Our adversary system is thus built on the assumption that despite initial appearances and impressions of the evidence, "there will be cases where closing argument may correct a premature misjudgment and avoid an otherwise erroneous verdict" (id. at 863). It is precisely because we cannot know in advance which case outcome will turn on an effective summation that the opportunity to prepare and deliver a summation based on the trial evidence and the agreed upon charge is a fundamental right of every defendant.

Because summation is so fundamentally important to our adversarial justice system, holding the power to change verdicts and determine who goes free and who does not, it also obviously follows that the opportunity for an effective summation is essential to a fair trial (see Herring, 422 US at 859; Yarborough v Gentry, 540 US 1, 5 [2003] ["The Sixth Amendment guarantees criminal defendants the effective assistance of counsel"]; Chapman v California, 386 US 18, 23 and n 8 [1967]{**34 NY3d at 496} [recognizing "there are some constitutional rights so basic to a fair trial that their infraction can never be treated as harmless error," including right to counsel]; People v Etienne, 220 AD2d 446, 447 [2d Dept 1995]). The denial of the right to marshal facts and law in a manner that effectively capitalizes on the charge as promised undermines defendant's presentation of a defense in the strongest possible terms and in accordance with counsel's studied judgment (see People v Crimmins, 36 NY2d 230, 238 [1975]; People v DiPippo, 27 NY3d 127, 135 [2016] [" '(T)he Constitution guarantees criminal defendants a meaningful opportunity to present a complete defense' " (some internal quotation marks and citations omitted)], quoting Nevada v Jackson, 569 US 505, 509 [2013]). An altered charge denies counsel the opportunity to persuade the jury based on a version of facts that tracks the instructions counsel rightly believes will be forthcoming. It is fundamentally unfair to promise counsel that the jury will be instructed on the law in a particular way and then fail to so direct the jury.

New York has codified the right to a summation and the right to request specific jury instructions. CPL 300.10 provides the parties an opportunity to submit charge requests to be ruled on by the court and requires the court to inform the parties of the charge in advance of summations (CPL 300.10 [4], [5]).

This statutory provision recognizes that counsel brings to bear much skill, expertise and artistry in the process of crafting an effective summation. With the skill and technique of one trained in the law, a defense counsel, for example, may illuminate the facts from the defense perspective, holistically explain the significance of the [*11]evidence—much like putting together a jigsaw puzzle with bits and pieces of information—and eventually lead jurors to conclusions and inferences that support the defense theory of the case, all to persuade the trier of fact to acquit (see Raymond W. Bergan, Thoughts about Closing Argument, in American Bar Association, Litigation Manual: Jury Trials at 324 [2008] ["(T)he entire trial preparation should lead up to a closing argument. In closing argument you will draw together all those items of evidence that the jury heard and emphasize the significance they may not have fully grasped at the time"]; Patricia Lee Refo, Closing Argument: A String of Pearls, in American Bar Association, Litigation Manual: Jury Trials at 330 [2008] ["The distinguishing feature of closing argument is, well, argument. That means drawing inferences or conclusions from the facts; talking about witnesses'{**34 NY3d at 497} credibility, motivation, or demeanor while testifying; explaining the significance of the evidence; discussing why an event occurred the way it did"]; Henry G. Miller, Principles of Summation, 28 Am Jur Trials 599 [1981] ["Somehow all the fragments of the trial must be collected in summation. Now the advocate integrates the evidence which was introduced in imperfect order and often out of sequence. The puzzle is now solved for the jury. At the time testimony was given, the significance may have only been known to counsel. All the disparate elements must now be fused into one intelligible unit"]; 2 F. Lee Bailey & Kenneth J. Fishman, Criminal Trial Techniques § 47:1 [2019] ["A well-presented case can be lost by a listless summation. A spirited, effective summation can change a lost cause into an acquittal"]; see generally Henry B. Rothblatt, Defense Summation, in New York State Bar Association, Basic Criminal Practice [4th ed 1982]).

The significance of the jury instructions to counsel's summation cannot be understated or underestimated as the charge sets the ground rules for the deliberative process. As we have long recognized, "[j]urors are presumed to follow the legal instructions they are given" (People v Baker, 14 NY3d 266, 274 [2010], citing People v Guzman, 76 NY2d 1, 7 [1990]; People v Acevedo, 69 NY2d 478, 488 [1987]). Thus, when counsel designs summation in anticipation of the judge's promised instructions to be given, counsel maximizes the persuasive impact of the defense version of the case. As aptly explained by one commentator, "[b]y suggesting that the court's instructions of law, as well as the facts, support your side, a doubly effective argument can be crafted. Argue the facts and then argue that they support a particular legal principle" (W. Ray Persons, Preparing and Delivering the Defense Closing Argument, 16 Prac Litigator [No. 3] 55, 57 [May 2005]).^[FN3] In other words, an effective summation reinforces the advocate's view of the evidence and the inferences to be drawn by placing them within the legal framework to be explained by the judge.

The type of error addressed here should be measured by the impact on counsel's strategic choices of how to shape summation,{**34 NY3d at 498} and not on the evidentiary support for the verdict. Summation is an integral component of trial; it is the sole opportunity for counsel to speak directly to the trier of fact and present the evidence and arguments from the defendant's perspective. The majority arrives at that holding, though not by the path I take. Summation holds open the possibility that a juror will be persuaded by the sheer strength of counsel's oratory skills. The bench, bar and our diverse communities subscribe to the view that summation matters—even when the weight and apparent strength of the evidence leave little hope of acquittal—because of the power of an advocate to change minds. It is not possible to quantify the impact on the verdict of the summation counsel would have given if the court had charged the jury as promised.

Given the constitutional rights an effective summation serves and its critical role in our adversary fact-finding system, the standard of review in broken-charge-promise cases must focus on where the error has its greatest impact: the summation. Counsel cannot provide constitutionally mandated legal assistance in a just and fair setting when counsel's last appeal to the jury is based on a judge's unfilled representations about the law to be charged. Therefore, a judge's failure to charge in accordance with the agreed upon instructions is an error warranting reversal where (1) counsel structured the presentation of evidence and the core elements of the defense based on the anticipated charge, and (2) the court provides no satisfactory opportunity to cure the prejudice flowing from the changed charge. In other words, when summation reflects that counsel's strategic decisions about the best way to organize and advocate the [*12]defense's view of the case is based on the promised charge, and the judge's failure to provide the charge as promised calls for either recalibration of counsel's approach by additional summation or additional instructions to the jury, both of which the judge denies, a defendant is denied the right to counsel's full and comprehensive "partisan advocacy" that is essential to "the adversar[ial] factfinding process in a criminal trial" (Herring, 422 US at 862, 858).

Traditional harmless error analysis is inapplicable and ill-suited to assess the impact of the errors presented in these appeals.{**34 NY3d at 499} First, because the trial court's failure to charge as promised violates defendant's constitutional rights to assistance of counsel by presentation of an effective summation and to a fair trial, neither class of error is subject to harmlessness analysis (Herring, 422 US at 859; Yarborough, 540 US at 5; Chapman, 386 US at 23; Crimmins, 36 NY2d at 237-238; Etienne, 220 AD2d at 447). The federal standard for ineffective assistance requires consideration of the prejudice to defendant and the denial of a fair trial is reversible error (Strickland v Washington, 466 US 668, 684-689, 690-698 [1984]; Chapman, 386 US at 23; People v Felder, 47 NY2d 287, 295-296 [1979]). Second, the salient doctrinal underpinning of the harmless error analysis requires an appellate court to determine whether the evidence of guilt is overwhelming, but that is not the proper focus of the constitutional deprivation in these cases (People v Felder, 47 NY2d 287, 295-296 [1979] ["There can be no doubt that the right to assistance of counsel is an essential ingredient in our system of criminal jurisprudence, rooted deeply in our concept of a fair trial within the adversarial context. . . . It is precisely this fundamental nature of the right that renders harmless error analysis inapplicable"]; Crimmins, 36 NY2d at 238 ["The right to a fair trial is self-standing and proof of guilt, however overwhelming, can never be permitted to negate this right"]). Harmless error analysis is inapplicable because the error is inherently prejudicial where it affects the summation, regardless of whether the evidence appears overwhelming. That is why we must determine in the first instance whether the error impacted counsel's strategic choices about how to shape summation, and not the existence of evidentiary support for the verdict.

The majority's observation that the errors here are not structural misses the mark (majority op at 482-483). Our system presumes that summation matters. Limiting the analysis to whether there is wholesale denial of summation or improper restriction of summation is based on the disfavored "narrowly literalistic construction" of a defendant's constitutional trial rights. The majority's reliance on Glebe v Frost (574 US 21 [2014]) for this overly restrictive view is misplaced. In that case, the defendant filed a petition for writ of habeas corpus under 28 USC § 2254 after the Washington Supreme Court sustained the defendant's conviction, in part, by holding that improper restriction of summation was a trial error rather than a structural error and applied harmless error analysis{**34 NY3d at 500} (id. at 22-23). The Ninth Circuit Court of Appeals granted the defendant's habeas petition, holding that the Washington Supreme Court had unreasonably applied clearly established federal law in holding that the trial court's restriction of summation was not structural error (id. at 23). The United States Supreme Court reversed under the Antiterrorism and Effective Death Penalty Act of 1996, an entirely different statute governing habeas corpus adjudication that is inapplicable here, by holding that Herring "did not clearly establish that the restriction of summation also amounts to structural error" (id. at 24). At most, Glebe is a pronouncement that it remains an open question if improper restriction of summation is structural error. For our purposes here, Glebe does not establish that violation of the right to an effective summation mandates harmless error analysis as implied by the majority.

In any case, the issue here is not improper restriction of summation, but rather that counsel prepared a summation based on the trial judge's representations of the law to be charged to the jury. When a judge does not follow the charging determination, that is not a restriction of the summation in the sense that the judge tells counsel they cannot address an issue at all or in a particular manner. The damage is in the misdirection about the legal framework of the case.

What matters then is that a promise has been broken as to the legal rules that direct the jurors' deliberations and that served as the basis for strategic choices by counsel. Counsel might well have made different choices. We cannot easily measure the nuances because the error is not centered on limiting closing but about setting counsel's choices on the wrong track. For example, Mairena argues that his defense counsel would have not discussed the broken bottles theory in summation, including talking about the nature of the cut and showing the surveillance video and autopsy photo, had he known the charge on the instrumentality of death would not be given. Rather, counsel would have focused solely on his justification defense. Similarly, Altamirano contends that if his defense counsel had known the court would give the temporary and lawful possession charge, his defense counsel could have connected the evidence [*13]presented to the relevant factors in the temporary and lawful possession test and emphasized to the jury that Altamirano did not initially come into knowing possession of the weapon; he kept the weapon wrapped in a blanket in a garbage bin, showing lack of intent to use the gun; he was extremely{**34 NY3d at 501} cooperative with the police; and that he had several character witnesses testify to his honesty. While defense counsel was forced to make the nonsensical argument that there was no evidence that Altamirano constructively possessed the gun—when he had undisputedly led the police to his apartment and informed them of the gun's location—defense counsel could still have emphasized these salient facts had he been permitted to reopen summations, and mapped them to the elements of the temporary lawful possession charge.

Our prior case law does not mandate that we assess the impact of these errors under a traditional harmless error analysis. The cases on which the majority principally relies, People v Miller (70 NY2d 903 [1987]), People v Greene (75 NY2d 875 [1990]) and People v Smalling (29 NY3d 981 [2017]), either support the proposition that the effect on summation is crucial or are irrelevant to the analysis.

Greene applies the standard of review I have described as the proper measure of the type of error at issue in these appeals. Greene is on all fours with Mairena's case in that Greene also involved a manslaughter prosecution and a dispute as to the instrumentality of death, namely whether defendant was shot. The judge gave the charge requested by counsel that the People had to establish beyond a reasonable doubt that the defendant shot the victim. Then, in response to a jury request, over the defendant's objection, the court responded that the jurors could find the defendant guilty if they found that the defendant caused the death of the victim by means other than by a gun. This Court held that "it was error, prejudicial to defendant, for the court to reverse its stance after assuring defendant that it would charge as he requested and after defendant had premised his summation on that theory" (Greene, 75 NY2d at 877). In Greene, the People argued that traditional error analysis should apply. Instead, we described and applied a "prejudicial error" standard. Equating "prejudicial error" with harmless error conflates the impact on the verdict—the focus of harmless error analysis—with the impact on summation based on counsel's reliance on the promised charge—the reversible error identified by the Court in Greene. In fact, Greene expressly chose not to apply harmless error analysis by declining to consider the possible impact on the verdict: "we need not determine whether, in the abstract, defendant properly could{**34 NY3d at 502} have been convicted if the jury concluded that no shooting had occurred" (id. [citation omitted]).

Smalling and Miller do not require a different analysis. Smalling, although the most recent, is the least instructive. The decision is set forth in a one-paragraph memorandum, in which we ordered a new trial when the trial court initially complied with its promise not to charge on constructive possession but then provided the instruction in response to a jury note during deliberations. Defendant had argued that the additional charge was prejudicial and there was good reason to believe the jury convicted defendant on the erroneously given charge. Notably, the People argued that any error from the additional charge was harmless, relying on People v Badalamenti (27 NY3d 423, 439 [2016]). Defendant distinguished Badalamenti as a case involving a final instruction and not a response to a jury request. We concluded that defendant was prejudiced, citing Greene and CPL 300.10 (4), and held that "[u]nder the unique circumstances of this case, the error [was] not harmless" (29 NY3d at 982), citing People v Nevins (16 AD3d 1046 [4th Dept 2005]), a case which applied the Crimmins harmless error doctrine to a supplemental instruction. Thus, Smalling is distinguishable, as it did not determine the impact on the summation under a harmless error analysis, but as limited by the parties' arguments, Smalling focused on if and how "the error infected or tainted the verdict" under a traditional harmless error analysis (Crimmins, 36 NY2d at 242), which even the majority does not apply to the instant appeals.^[FN4]

Although the Court in Miller did conclude that any error due to the trial court's failure to inform counsel prior to summation{**34 NY3d at 503} that it would charge on a lesser included count was harmless, the Court did so by analyzing the impact of the error on counsel's summation. In Miller, the defendant, an assistant conductor employed by Metro-North, was charged with grand larceny in the third degree and 10 counts of offering a false instrument in the first degree for submitting 10 false Metro-North time slips for pay when he had not worked on those days. Defense counsel's summation explicitly focused on the time slips and the false instrument charges and did not touch on the grand larceny charge at all. Without first informing counsel, the trial court gave the charge for the lesser included offense of petty larceny, an offense distinguished from the grand larceny solely by the total value of the property stolen. On appeal this Court explained that under the circumstances, "it cannot be said that defense counsel's summation would have been affected by knowledge that the petit larceny charge would be submitted to the jury" (Miller, 70 NY2d at 907). Again, our Court did not focus on the ultimate impact on the outcome of the trial but rather on the impact on counsel's summation, notwithstanding the reliance on the trial court's promised and altered charge. This is not traditional harmless error but rather a measure of prejudice to the defendant based on what counsel would have done absent the judge changing course.

In any case, the majority does not employ traditional harmless error analysis because rather than consider whether the verdict would be different absent the trial judge's error, the majority focuses—just as I do—on the impact on defense counsel's summation. Indeed, the majority goes so far as to explain that "the relevant question is whether counsel's summation was materially affected in a manner that prejudiced defendant in light of the charge actually given" (majority op at 488). I agree, but part company with the majority in the way it answers that question as to the summations here.

In defendant Mairena's case, counsel's summation focused on his argument that the victim died due to injuries caused by falling on the broken bottle. In support, counsel discussed how the People's theory relied on the victim being cut before falling, but this was implausible because the victim "is punching as he{**34 NY3d at 504} goes down" while his wound would have led to "nerve damage and fairly quick loss of motor function and immediate blood spraying." Further, defense counsel pointed out how "there [was] no blood on the ground where [the victim] is punching," which showed "reasonable doubt that he was [not] cut back there because there is no blood on the ground where he is punching." Counsel also impeached the testimony of the police officers who investigated the crime, stating that they "[did not] follow up on this investigation of the bottles theory." In sum, counsel argued that the People "failed to prove beyond a reasonable doubt that the instrument of death was the knife" and that "the knife caused the injury beyond a reasonable doubt." Thus, counsel made an argument that even if the use of the box cutter led to the fatal fall on the bottle, defendant could not be found guilty. That argument conceded ground and supported a guilty verdict based on the court's instructions that stated only that the People had to establish defendant caused decedent's death, without stating, as promised, that the cause was a fatal wound inflicted by the box cutter.

The majority bases its conclusion that the court's failure to charge as promised did not affect counsel's summation on three points, none persuasive. First, the majority asserts that counsel merely sought to prevent the People from arguing guilt based on death caused by the fall on the broken bottle and the People never made that argument, arguing instead that the death was caused by defendant slashing the decedent. That is not the way the People saw it. When counsel objected to the charge, the People confirmed that the jury could find defendant guilty of manslaughter based on the charge as given—that the instrumentality of death is a dangerous instrument, which need not be the box cutter: "they can find that [defendant] used a dangerous instrument to cause those injuries and they happened to recover a box cutter. It [does not] have to be that box cutter that caused that injury." Thus, because the court gave a charge that permitted another view of the evidence—a charge we assume the jury followed (Baker, 14 NY3d at 274)—the fact that the People did not describe another theory for the death is beside the point. In any case, defense counsel essentially made the argument for the People. Thus, the People benefitted from both theories being presented to the jury because the charge allowed for a guilty verdict under either view of the evidence. Indeed, the People could comfortably mock the broken bottle theory as a means to undermine defense counsel's credibility,{**34 NY3d at 505} surely aware that if [*14]that strategy failed, the broken bottle theory would still be a basis for conviction. A win-win scenario for the People and a disastrous result for defendant.

The majority's second basis for concluding the error here did not impact the summation is similarly unsupported by the record and logic. According to the majority, "the jury charge on manslaughter, read as a whole, conveyed that the jury had to find the victim's death was caused by defendant's intentional actions in using the box cutter in order to convict; it did not permit the jury to convict defendant of manslaughter based on the bottle theory" (majority op at 487). Whatever else the charge "conveyed," it did not eliminate the return of a guilty verdict based on defendant intending to cause serious physical injury by cutting the decedent with the knife which led decedent to fall on broken glass, ultimately causing his fatal wounds. The majority claims that the jury could not have concluded that falling on glass caused serious physical injury, but the medical examiner testified to the contrary, stating that the wound "was not inconsistent" with being "caused by a broken bottle."

With respect to defendant Altamirano, the majority concludes that the court's denial of the temporary lawful possession charge did not affect the summation because counsel actually addressed the factors that the jury could consider since the judge actually gave the charge. Thus, no harm no foul. The harm, however, is that counsel was prevented from employing one of the essential components of summation: mapping the evidence against the elements of the defense for the jury's consideration. In addition, the majority mentions but glides over the fact that due to the denial of his charge request, counsel argued that defendant did not have possession of the weapon. That argument was unsupported by the evidence and the law, as the gun was found in a garbage can in his apartment, having been placed there with defendant's full knowledge. The fact that counsel referred to evidence supporting temporary lawful possession is not the same as crafting a summation deliberately based on a charge to which counsel may refer in order to reinforce the defense argument. Moreover, without knowledge that the charge would in fact be given, counsel could not rely on the charge to undermine the People's arguments that defendant knowingly possessed the firearm. Again, like in Mairena, the People, not defendant, benefitted from the summation choices counsel made based on the judge's error.{**34 NY3d at 506}

In both appeals, the trial judge's failure to abide by the agreed upon jury charge prejudiced defendant as counsel in each case prepared their summations based on the anticipated charge. I would reverse and grant a new trial for each defendant.^[FN5]

When defense counsel charts the course of summation based on the trial judge's description of the charge, and then after summation the judge fails to give the promised instruction or gives a previously denied instruction as requested by defense counsel, the judge should offer to reopen summations so that counsel may address the jury in light of the charge as actually given. Not only would reopening summations ensure the fairness of the proceeding, but it would also avoid a potential appealable issue. Absent such opportunity to mitigate the prejudice to the defense from the court's error, the conviction should be reversed and a new trial ordered. I dissent.

Chief Judge DiFiore and Judges Garcia and Feinman concur; Judge Fahey concurs in result in an opinion; Judge Rivera dissents in an opinion in which Judge Wilson concurs.

In each case: Order affirmed.

"[I]n certain circumstances, the possession of a weapon may be innocent and not criminal. . . .

"A person has innocent possession of a weapon when he or she comes into possession of the weapon in an excusable manner and maintains possession, or intends to maintain possession, of the weapon only long enough to dispose of it safely.

"There is no single factor that by itself determines whether there was innocent possession. In making that determination, you may consider any evidence which establishes that the defendant had knowing possession of a weapon, the manner in which the weapon came into the defendant's possession, the length of time the weapon remained in his/her possession, whether the defendant had an intent to use the weapon unlawfully or to safely dispose of it, the defendant's opportunity, if any, to turn the weapon over to the police or other appropriate authority, and whether and how the defendant disposed of the weapon" (see CJI2d[NY] Possession—Temporary).