OPINION OF THE COURT

Under New York’s capital punishment scheme, a person who commits an intentional (second degree) murder is eligible for a death sentence if any one of 13 aggravating factors is proved (see Penal Law § 125.27 [1] [a] [i]-[xiii]), and if within 120 days after arraignment the prosecution files a notice of intention to seek the death penalty (CPL 250.40 [2]). In the case before us, a jury found defendant guilty of two counts of first degree murder, based on two aggravating factors (witness elimination murder, Penal Law § 125.27 [1] [a] [v], and intentional murder in the course of and in furtherance of second degree burglary, Penal Law § 125.27 [1] [a] [vii]). At the penalty phase of the trial, the jury determined that defendant should be sentenced to death on both counts. For reasons that follow, we conclude that neither of the aggravating factors was proved. This being so, the penalty phase was conducted without legal foundation and the resulting death sentences must be vacated. Defendant’s guilt for intentional murder, however, was proved beyond a reasonable doubt, and we therefore reduce defendant’s conviction to murder in the second degree and remit the case to the trial court for resentencing.

In early April 1998, defendant and his wife, Jill, signed a separation agreement but continued to live under the same roof at their home in Spafford, Onondaga County. On April 21, during a predawn heated argument, defendant struck. Jill repeatedly on the head with a baseball bat. The couple’s two young children were nearby and Jill called out, urging them to call the police because their father was trying to kill her. After the attack, defendant phoned his parents for help. They soon arrived at the Cahill residence, along with defendant’s brother and a family friend who was a physician. Having been summoned to the scene, the police found Jill lying on the kitchen floor. She was covered in blood, writhing in pain and moaning incoherently. Her left temple was indented from the injury.

Defendant and victim were taken to different hospitals. After hospital personnel treated defendant for minor injuries, the po*36lice brought him to the station house for questioning. At first, he stated that Jill had instigated the argument and attacked him with a knife, causing some cuts and scratches on his body. He claimed he struck her in self-defense. Defendant later admitted that he struck Jill with the bat when she was unarmed and that he cut himself, making it look like self-defense. Defendant added that after the assault, he taped a length of garden hose to the tailpipe of his car in order to poison himself with carbon monoxide, but decided against suicide when he saw a rosary in his vehicle.

In June 1998, a grand jury indicted defendant for assault in the first degree and criminal possession of a weapon in the fourth degree. In the months that followed, he and his attorney prepared for trial. In the meantime, there were custody proceedings in Onondaga Family Court, which placed the children with their maternal grandparents and aunt. In addition, Family Court and Onondaga County Court issued orders of protection prohibiting defendant from seeing his children or entering University Hospital, where Jill was confined.

By one medical estimate, Jill had been hit at least four times in the skull. At the hospital, she underwent emergency surgery to remove a blood clot from her brain. In the ensuing weeks, Jill suffered from brain swelling and a number of life-threatening infections. She began to improve and eventually moved from intensive care to the coma rehabilitation unit, and later to the general rehabilitation unit. Her recovery was slow and by no means complete. By October of 1998—six months after the assault—Jill was able to recall the names of her children and had regained some ability to speak, but could use only short, simple words.

On October 27, 1998, after the hospital was closed to visitors, defendant entered the premises, in disguise. According to several members of the staff, defendant wore a wig and glasses, posing as a maintenance worker, complete with a mop and falsified name tag.1 Shortly after 10:00 p.m., a nurse detected a strong odor in the room and saw Jill having trouble breathing. The nurse also observed a waxy-looking substance on Jill’s chest *37and that Jill’s hospital gown caused a burning sensation when touched. Despite efforts to revive her, Jill died the next morning. She had been poisoned. Am autopsy revealed that potassium cyanide was administered through her mouth or feeding tube.

Police promptly arrested defendant for Jill’s murder. Employing search warrants, they recovered data from the hard drive of the Cahill home computer revealing Internet searches that used the words “cyanide” and “ordering potassium cyanide.” The “slack”2 also yielded letters composed on the computer. The letters were purportedly sent from an East Syracuse company called General Super Plating to Bryant Laboratories, placing orders for potassium cyanide. In the area near the shed on the Cahill property police found a half-burned wig and a bottle containing potassium cyanide. Further investigation produced eyewitnesses who saw defendant intercept the delivery of cyanide in the vicinity of General Super Plating in July of 1998.

On November 19, 1998, while the assault charges were pending, a grand jury indicted defendant on two counts of first degree murder and related offenses. One murder count charged defendant with having murdered Jill to prevent her from testifying against him at his trial for the April 1998 assault (Penal Law § 125.27 [1] [a] [v]), the other with intentionally murdering Jill in the course of and in furtherance of a burglary (Penal Law § 125.27 [1] [a] [vii]). The grand jury also indicted defendant on two counts of murder in the second degree, burglary in the second degree, aggravated criminal contempt and criminal possession of a weapon in the fourth degree. On December 30, 1998, the District Attorney filed a CPL 250.40 (2) notice of intention to seek the death penalty. In addition, the prosecution moved to consolidate the murder and assault indictments.3 The trial court granted the motion in January 1999.

Pursuant to CPL 400.27, the court conducted the jury trial in two phases.4 In the first phase, the jury found defendant guilty of both counts of first degree murder, first degree assault (based *38on the April 1998 beating) and related charges. The penalty phase followed, in which the jury returned with verdicts of death under both first degree murder counts. Pursuant to article VI, § 3 (b) of the State Constitution and CPL 450.70 (1), defendant has appealed directly to our Court.

Although we address several of the 38 points briefed by defendant, our determination turns on three primary issues: jury selection, weight of the evidence supporting first degree (witness elimination) murder under Penal Law § 125.27 (1) (a) (v) and legal sufficiency of first degree murder (based on burglary) under Penal Law § 125.27 (1) (a) (vii). We conclude that the trial court erred in its rulings concerning potential jurors Nos. 23 and 855; that the first degree murder conviction based on witness elimination was against the weight of the evidence; and that the first degree murder conviction premised on burglary was legally insufficient to support a conviction.

I. Pretrial Issues

A. Pretrial Publicity

Defendant argues that extensive, prejudicial pretrial publicity denied him a fair trial before an impartial jury. He contends that the adverse publicity warranted a change of venue, either before or during voir dire. We disagree.

CPL 230.20 authorizes a change of venue when either party shows “reasonable cause to believe that a fair and impartial trial cannot be had in such county” (CPL 230.20 [2]). A motion for change of venue must be made before the Appellate Division department embracing the county in which the superior court is located. In the exercise of its discretion, the Appellate Division can order removal to the superior court of another county (CPL 230.20 [2] [a]) or direct the commissioner of jurors (in consultation with the administrative judge of the judicial district in which the county is located) to expand the jury pool to include people from jury lists of geographically contiguous counties within the judicial district (CPL 230.20 [2] [b]).

It is imperative that prospective jurors be open-minded and unbiased, but they need not “be totally ignorant of the facts and issues involved” (Irvin v Dowd, 366 US 717, 722 [1961]). As recognized in Irvin, “[i]t is sufficient if the juror can lay aside *39his impression or opinion and render a verdict based on the evidence presented in court” (id. at 723). The Supreme Court has identified a number of conditions that may justify a change of venue, such as televised confessions (see Rideau v Louisiana, 373 US 723 [1963]), a media-generated “carnival atmosphere” in the courtroom (Sheppard v Maxwell, 384 US 333, 358 [1966]), or a close temporal proximity between the media coverage and the jury selection (see Patton v Yount, 467 US 1025,1032 [1984]).

Disinclined to presume prejudice, our courts have rarely granted motions for change of venue before jury selection.5 To succeed, a party must show not only extensive publicity and comment but also that media coverage has aroused “a deep and abiding resentment” in the county (People v Boudin, 90 AD2d 253, 259 [2d Dept 1982]; Preiser, Practice Commentaries, McKinney’s Cons Laws of NY, CPL 230.20, at 311). Here, the Appellate Division denied defendant’s pre-voir dire motion for a change of venue, stating that if during voir dire a fair and impartial jury could not be drawn, the defense could renew the motion (261 AD2d 972 [4th Dept 1999]). Defendant did so after jury selection, arguing that the voir dire revealed a level of prejudice that infected the jury pool. The Appellate Division denied that motion as well.

This case did not warrant a pre-voir dire change in venue. Onondaga County was not “deluged by a tidal wave of prejudicial publicity to such an extent that even an attempt to select an unbiased jury would be fruitless” (Boss, 261 AD2d at 4). Thus, the Appellate Division properly denied defendant’s first CPL 230.20 motion. Following the denial of a change of venue motion on the ground of prematurity, the proper procedure is to undertake voir dire and attempt to select an impartial jury (People v Parker, 60 NY2d 714, 715 [1983]; People v Smith, 63 NY2d 41, 69 [1984]). The trial court did so here, and for the reasons that follow we conclude that the Appellate Division did not abuse its discretion in denying defendant’s post-voir dire motion.

This Court has several times addressed the problem of pretrial publicity in first degree murder cases. In People v DiPiazza (24 NY2d 342, 347 [1969]), we upheld the denial of a change of venue, finding it significant that defendant did not exhaust his *40peremptory challenges, but noting that “the court’s discretion will not be disturbed unless the newspaper articles are of such a sensational character as to excite local popular passion and prejudice so that the defendant will not be able to have the fair trial to which he is entitled.” In that case, however, the Herkimer County media coverage was “surprisingly objective” (id.). Moreover, less than 25% of the prospective jurors expressed an opinion as to DiPiazza’s guilt (id. at 346).

In People v Culhane (33 NY2d 90, 110 [1973]), an Ulster County case in which three prisoners killed a deputy sheriff, we observed that the highly localized and incessant nature of the prejudicial publicity surrounding the case “probably” warranted a change of venue. Nevertheless, we reversed the defendant’s conviction due to the bias of prospective jurors who did not make expurgatory oaths required under former Code of Criminal Procedure § 376 (2). In People v Lynch (23 NY2d 262, 272 [1968]), this Court affirmed the denial of defendant’s motion for change of venue where voir dire resulted in selection of a jury free of media influence.

This case generated a good deal of pretrial publicity in Onondaga County. According to the juror questionnaires, 86% of the prospective jurors had heard of the case from media accounts (including 8 of the 12 jurors). The test for removal, however, is not based on the number of prospective jurors who heard of the case. If that were the test, juries in highly publicized cases would necessarily consist only of the most reclusive and uninformed segment of the population. What counts is not knowledge of the accusation, but whether that knowledge has shaped the jurors’ attitudes and predispositions. Merely having heard of the case is not enough, if the jurors come into the courtroom knowing no more than they learn when they are told of the charge in open court. The question is whether media or other accounts have been so inflammatory as to thwart the selection of a fair-minded jury.

A review of the media coverage in this case shows that the publicity was not so sensational or prejudicial as to taint the jury pool. Most of the media coverage tended to be objective, including “police blotter” reports and news reports on the court proceedings. The publicity was similar to that in DiPiazza, in which we upheld the denial of a change in venue: “The victim’s funeral and the members of her family were sympathetically portrayed and the defendant’s action was described as having caused a widespread reaction and aroused deep feeling. But *41there was very little that could be said to be affirmatively hostile to him” (id. at 347). Here, although 52% of the jurors came to court with an opinion as to defendant’s guilt or innocence, the voir dire successfully culled out jurors who may have been biased by pretrial publicity. Thus, we hold that, in denying defendant’s motions, the Appellate Division acted within its discretion.

B, Search Warrant

Police found a container of potassium cyanide in a hollow space of a cinder block on the exterior base of the shed on defendant’s property. Defendant argues that the warrant referred only to his house and that the police conducted an impermissible search of the shed. Contrary to defendant’s contention, the search warrant authorized a search of the shed.

When conducting searches, search warrants are preferred, because they contemplate an orderly procedure and authorization by a neutral and detached magistrate (see United States v Jeffers, 342 US 48 [1951]; Johnson v United States, 333 US 10 [1948]). For that reason, reviewing courts should accord the process proper deference and not defeat search warrants (or discourage law enforcement officials from seeking them) by imposing overly technical requirements or interpreting them incompatibly with common sense (see United States v Ventresca, 380 US 102, 108 [1965]). Thus, a court may review the supporting documents to clarify any ambiguity (see People v De Lago, 16 NY2d 289, 290-291 [1965]). Also, and especially relevant here, material previously submitted to a judge may be incorporated by reference in a subsequent warrant application “so long as the earlier information was given under oath, is either available to the Magistrate or sufficiently fresh in the Magistrate’s memory so that he or she can accurately assess it and it is available in a form which can be reviewed at a later date” (People v Tambe, 71 NY2d 492, 502 [1988]).

The Fourth Amendment requires a search warrant to describe particularly the place to be searched (US Const Fourth Amend; see NY Const, art I, § 12), so that the right of privacy is protected from arbitrary police intrusion. To meet the particularity requirement, the warrant must be specific enough to leave no discretion to the police (see People v Brown, 96 NY2d 80, 84 [2001]).

Here, the particularity requirement has been met. The warrant that led to the seizure explicitly notes that it is an ad*42dendum to a warrant issued three days before, which included searches for “toxic/caustic materials” in defendant’s home or “within any unattached garages or storage sheds” (emphasis ours). Thus, the warrant incorporated previously submitted materials known to the issuing Judge, who, over four consecutive days, signed five search warrants related to this case (see Tambe, 71 NY2d at 502). The search was valid and the court properly denied suppression of the evidence.

C. Defendant’s Right to a Bench Trial

Defendant argues that he should have had the option of a bench trial to alleviate the alleged prejudicial effects of pretrial publicity and consolidation of the murder and assault indictments. His argument fails. As noted, we find no prejudicial impact from the pretrial publicity, nor did any result from consolidation of the indictments.6 Moreover, “[t]he history of our jurisprudence reveals that the fundamental right is the right to a trial by jury” (People ex rel. Rohrlich v Follette, 20 NY2d 297, 301 [1967] [emphasis in original]). This tenet is reflected in article I, § 2 of the New York State Constitution, which prohibits waivers of a jury trial in capital cases. Our State Constitution’s ban on jury waivers in capital cases is longstanding and purposeful.

Before 1938, no defendant in a criminal case was allowed to waive trial by jury (see People v Cosmo, 205 NY 91, 96 [1912]; Cancemi v People, 18 NY 128, 137 [1858]). The 1938 Constitution—our present Constitution—allows bench trials in criminal cases, but pointedly preserves the prohibition in capital cases (People v Page, 88 NY2d 1, 6 [1996]). The Bill of Rights Subcommittee to that Constitutional Convention drafted the provision, explaining that “its determination to withhold the jury waiver right from capital defendants derived from the notion ‘that the Constitution will still not permit this choice to a defendant in a capital case, and regards such a prohibition against waiver as a measure for the protection of the defendant.’ ”7

Consistent with constitutional commands, CPL 320.10 excludes capital cases from the provisions that govern jury trial *43waivers. Moreover, there is no federal constitutional right to a bench trial (see Singer v United States, 380 US 24, 36 [1965]). Thus, defendant was not entitled to one (see also People v McIntosh, 173 Misc 2d 727 [Dutchess County Ct 1997]).8

D. Consolidation of the Indictments

Defendant contends that the court should not have granted the prosecution’s motion to consolidate the assault and murder indictments for trial. He further asserts that after consolidation he had the right to plead guilty to the assault charges over the prosecution’s objection and proceed to trial on the murder charges only.

The trial court did not abuse its discretion in allowing consolidation (see People v Lane, 56 NY2d 1 [1982]). CPL 200.20 (2) (b) permits joinder of offenses based on different criminal transactions when “such offenses, or the criminal transactions underlying them, are of such nature that either proof of the first offense would be material and admissible as evidence in chief upon a trial of the second, or proof of the second would be material and admissible as evidence in chief upon a trial of the first.” The April 1998 assault and October 1998 murder of Jill Cahill fit this provision because the prosecution based one of its murder charges on Jill’s potential testimony in the assault trial, making proof of the assault relevant to the murder case. Thus, the offenses were joinable and consolidation was proper (see People v Bongarzone, 69 NY2d 892 [1987]).

We also reject defendant’s argument that he should have been allowed to plead guilty to only the assault portion of the consolidated indictment. From the time defendant was indicted for assault until the court consolidated the assault and murder indictments—a period of approximately seven months—defendant could have pleaded guilty to the assault indictment. Once the court consolidated the indictments, defendant could plead guilty to the entire indictment (if the prosecution determined *44that it would, no longer seek the death penalty)9 but not to only a portion of it (see CPL 220.10 [2], [4]).10 Unless the People and the court consented, such a plea was untenable.

II. Jury Selection

Defendant raises a number of arguments concerning jury selection, and in particular assails the trial court’s rulings concerning 16 prospective jurors. We will address only the rulings as to prospective jurors Nos. 23 and 855. 11

A. Defendant’s For-Cause Challenge of Prospective Juror No. 23

In capital punishment jurisprudence, jurors must be death qualified and life qualified if they are to serve. Broadly put, a juror who is “death qualified” may have misgivings about the death penalty but does not rule out voting for it. Conversely, a juror who is “life qualified” may generally favor the death penalty but does not rule out voting against it.

Defendant claims that prospective juror No. 23 was overly prone toward the death penalty and thus not properly life qualified. He contends that the trial court erroneously denied his challenge for cause under CPL 270.20 (1) (f), requiring the defense to waste a peremptory challenge. Having exhausted his peremptory challenges before the completion of jury selection, defendant seeks redress under CPL 270.20 (2). We agree that the trial court erred in not dismissing prospective juror No. 23 for cause.

In his questionnaire, prospective juror No. 23 said he was “not a total 100% supporter” of capital punishment. He added, however, that a sentence of life without parole could mean that “in some cases the defendant got off easy,” and he indicated that he could not consider a sentence of life without parole *45because he thought defendant should die, just as the victim did. When questioned individually, however, he made a number of statements attesting to his ability to deliberate fairly. As the prosecution points out, the juror said he was “willing to go along and listen now to hear, you know, what prompted such actions, what happened, what really happened, and see what the story really is.” This seemed reassuring for the moment, but a later exchange was particularly revealing and cast doubt on his previous assertions. When examined by the defense attorney concerning his ability to consider both death and life without parole, juror No. 23 related an incident in his own marriage:

“I had a terrible argument with my wife once, real bad. Almost a knock down drag out. I raised my [h]and to her and I didn’t hit her but I grabbed her and threw her on the couch. You know, I felt sick for doing that because that’s my generation. My generation doesn’t do that. You know, I have never—we have had arguments since, we have had discussions, but, you know, when it comes to that point, it’s time to walk. It’s time to go out the door, get some air, get the hell away. So it’s possible that that could come to mind. As much as I’m sitting here and I’m trying to promise everybody that I’d be fair, I’m still human, I’m still a man, that could come to mind. That’s being honest. . . .

“I’d love to be able to say I could just blank myself out but if the gentleman, if we find the gentleman guilty of First Degree Murder and now comes the penalty phase, as much as I would say, you know, I have to just deal with this, that situation will come up because it still haunts me. It still bothers me.”

Following up on the point, defense counsel asked him if “that life experience of yours ... is causing you some concern as to whether you’d be able to consider both penalties fairly?” and whether “[t]hat would cause you some problems with being able to consider life without parole as the appropriate penalty other than death?” The juror answered yes to both questions. Defendant immediately challenged him for cause, contending that the juror’s life experience would amount to “an inability at the sentencing phase, the penalty phase, to be able to engage in the statutory weighing process that New York law mandates a juror to engage in.” The trial court denied the challenge.

The test for life and death qualification is prescribed in CPL 270.20 (1) (f). Pursuant to that statute, either party in a capital *46case may challenge a prospective juror for cause if the juror “entertains such conscientious opinions either against or in favor of such punishment as to preclude such juror from rendering an impartial verdict or from properly exercising the discretion conferred upon such juror by law in the determination of a sentence pursuant to section 400.27” (emphasis supplied). In People v Harris (98 NY2d 452, 484 [2002]), we interpreted “preclude” to embody the Wainwright v Witt (469 US 412 [1985]) “prevent or substantially impair” standard, and explained that under this criterion jurors must be dismissed if they “express an inability to set aside their personal views on the death penalty in deference to the court’s instructions” (Harris, 98 NY2d at 484). The proponent of the challenge must demonstrate through questioning the juror’s inability to fulfill his or her oath (see Morgan v Illinois, 504 US 719, 733 [1992]).

Prospective juror No. 23 divulged that an incident in his life involving domestic violence—an issue highly relevant to this case—would strongly color, if not dictate, his views as to the penalty. The People urge that the juror’s earlier statements regarding his ability to serve impartially established his suitability. The juror’s expressions of open-mindedness, however, were undermined by his later responses. Once he acknowledged that his views on domestic violence would impede his ability to consider one of the two sentencing options, his earlier expressions of fitness rang hollow. Rather than grant defendant’s challenge, or at least question the juror as to whether these deeply held views would preclude him from serving impartially in a penalty phase, the court simply allowed the juror to remain on the panel.12 This was error.

B. The People’s For-Cause Challenge of Prospective Juror No. 855

The People successfully challenged prospective juror No. 855 based on her attitude toward the death penalty. We agree with defendant that this juror was death qualified (see Wainwright v Witt, 469 US 412 [1985]) and that the court improperly granted the People’s challenge.

In her questionnaire, prospective juror No. 855 equivocated about her feelings toward capital punishment. In some respon*47ses she indicated she could not consider death as a sentencing option, but in others said she would consider both death and life without parole. When questioned individually, she said that “[i]n most instances” the death penalty was wrong, but allowed that she could accept it in a few cases. The prosecutor questioned her further on this point, pressing her to give examples of cases in which she thought the death penalty appropriate. She agreed with the prosecutor when he suggested that a case involving multiple victims might justify capital punishment, and said she would consider the death penalty in cases that involved “[something very, very very wrong,” or were “bizarre” or “brutal[ ].” She also referred to an instance in which she thought the death penalty warranted, identifying “the case of a mother and daughters that were out in the west.” She declined to give further examples, explaining that she would base her decision on “prayerful consideration and hearing about the situation.”

Furthermore, prospective juror No. 855 answered yes when at different times the prosecutor, defense counsel and the court asked her if she could consider both death and life without parole as possible punishments. Her voir dire concluded when the court asked her if she considered death a possible penalty in this case, and she replied that she “[saw] it as a possible one, sir, yes.” The prosecutor challenged her for cause, arguing that “she refuses to answer a dispositive question,” and the trial court granted the challenge, stating “I don’t think she is in a right mind to be a juror in this case. She is out.”

On this record, juror No. 855 should not have been dismissed for cause. CPL 270.20 (1) (f) and Harris (98 NY2d at 484) require dismissal of a juror whose views “prevent or substantially impair” the juror from rendering an impartial verdict or properly exercising sentencing discretion. Drawing on United States Supreme Court precedent dealing with death and life qualification of capital jurors,13 we explained that “[w]here jurors express conscientious views concerning the death penalty yet still make clear that they are able to follow their oaths to act impartially, they cannot be excluded for cause from participating on the jury” (Harris, 98 NY2d at 484). Prospective juror No. 855 may have expressed reservations about the death penalty but, after doing so, made it clear that she could follow *48her oath, act impartially and consider both the death penalty and life without parole. Moreover, the prosecutor made no correlation between this juror’s views on capital punishment and his argument that she refused to answer a question (see id. at 486-487). In advancing his challenge for cause, the prosecutor maintained that this juror’s views on the death penalty were “unknown.”

The contrast between prospective jurors Nos. 855 and 23 illustrates the trial court’s uneven standard in addressing the challenges based on death or life qualification. Juror No. 23 candidly stated that his experience with domestic violence “still haunts [him]” and that he could not “blank his mind” when considering possible penalties. Though the juror “expressed] an inability to set aside [his] personal views” (id. at 484), the court denied defendant’s challenge. On the other hand, juror No. 855 revealed no inability to consider both punishments. She gave examples of cases where she might consider the death penalty, said she would make her decision on penalty after hearing the case, and repeatedly told the parties she would consider both death and life without parole. Nonetheless, the court granted the People’s challenge.14 We caution that the “prevent or substantially impair” standard applies equally to both death and life qualification (see id. at 484; Morgan, 504 US at 733-734) and that trial courts should employ the test meticulously. The prosecutor based his challenge for cause, at least in part, on his assertion that the juror’s views on the death penalty were “unknown.” The court dismissed the juror, stating only that “I don’t think she is in a right mind.” We caution that a trial court should make every effort to couch its ruling with reference to an articulated standard and, most importantly, create *49a record that permits appellate review. Here, the recorded exchanges among the juror, the court and the attorneys did not justify dismissal.15

Lastly, and most emphatically, we underscore that the chances of reversal are too great to make risky voir dire rulings that could occasion retrial, with all the needless effort and expense that goes with it (see e.g. People v Heard, 31 Cal 4th 946, 966-967, 75 P3d 53, 66 [2003]). We have issued this caution in non-capital criminal cases (People v Nicholas, 98 NY2d 749 [2002]), and it is all the more compelling in capital cases.

C. Remedy for Penalty-Related Juror Selection Errors

Defendant urges that pursuant to CPL 270.20 (2) the trial court’s errors pertaining to prospective jurors Nos. 23 and 855 compel us to overturn not only the death sentence but the guilt phase verdict as well. We disagree. The errors as to both jurors related to their ability to serve impartially only during the penalty phase. Errors of that type do not infect the guilt phase and by no means warrant a reversal of the entire trial.

CPL 270.20 (2) states that “[a]n erroneous ruling by the court allowing a challenge for cause by the people does not constitute reversible error unless the people have exhausted their peremptory challenges at the time or exhaust them before the selection of the jury.” Subdivision (2) goes on to state that the denial of a defendant’s for-cause challenge is not reversible error unless the defendant has exhausted all peremptory challenges or uses a peremptory against the disputed juror and later exhausts all such challenges. In laying out these requirements, the statute contemplates remedial action by the appellate court when the trial court improperly grants or denies a challenge for cause.

Indeed, the loss of a peremptory challenge constitutes harm enough to trigger the statutory remedy contemplated in CPL 270.20 (2) (see e.g. People v Bludson, 97 NY2d 644 [2001]; People v Arnold, 96 NY2d 358 [2001]). An erroneous denial of a defendant’s challenge for cause is not rendered “harmless” merely because the defense later excuses the juror peremptorily (see People v Culhane, 33 NY2d 90, 97 [1973]; People v Chambers, 97 NY2d 417 [2002]). To the contrary, the defendant’s loss of the peremptory challenge constitutes the harm. Where a *50defendant’s peremptory challenges are thereafter exhausted, erroneous denial of the prior challenge for cause constitutes reversible error warranting corrective action (CPL 270.20 [2]).

The precise corrective action, however, depends on the type of trial involved. Capital trials are divided into guilt and penalty phases (see CPL 400.27). In the traditional, single-phase criminal trial, the jury decides only the defendant’s guilt and not the sentence. Thus, when a trial court rules in violation of CPL 270.20 (1), a new trial is the only possible remedy. The unique, bifurcated structure of a capital trial, however, not only affords the defendant broader protections but also provides alternative remedies for improper rulings. In reviewing the evolution of our statutory law, a number of historical markers support our conclusion that a guilt phase retrial is not required if a juror’s bias goes only to the sentencing phase.

Bifurcated capital trials began in New York in 1963, with the amendment of sections 1045 and 1045-a of the former Penal Law (see L 1963, ch 994, §§ 1, 2).16 The bill was proposed because, at the time, New York State was the only American jurisdiction that had retained the mandatory death penalty for murder. Proponents concluded that by providing the jury with a life imprisonment option, the bill would achieve the “worthy objective” of terminating New York’s solitary adherence to the mandatory death penalty for murder (see Mem of Commn on Revision of Penal Law and Crim Code in Support of L 1963, ch 994, 1963 McKinney’s Session Laws of NY, at 2019). It would eliminate “the illogical and wasteful situation arising when a jury determination of guilt is negated by failure to agree upon the penalty or recommendation aspect. Through severance of the two issues and prescription of separate verdicts for each, the primary verdict of guilty stands final and recorded regardless of any farther proceedings or determinations with respect to sentence’’ (id. at 2020 [emphasis added]).

*51Effective 1967, the Legislature revised the Penal Law to make murder a degreeless crime (L 1965, ch 1030, as amended by L 1967, ch 791, § 9), treating intentional killing, depraved indifference murder and felony murder as the same level of offense. Under then-existing Penal Law § 125.30, a defendant convicted of intentional or felony murder17 was subjected to a second proceeding to determine whether the sentence should be death or life imprisonment.18

At the time of these enactments, the former Code of Criminal Procedure had been in effect for over a century, well before the United States Supreme Court decided a number of cases that shaped our statutory development. A Historical Note following the text of section 374 of the Code (Cons Laws of NY Ann, Book 66, at 824 [Edward Thompson Co 1958]) states that “[a] major consideration in enacting sections 374-378 [on challenges for cause] was to define that bias which should be the grounds for eliminating a juror, confining it to those degrees of prejudice which would ‘endanger the substantial rights of the prisoner’ ” (citing Report of Commissioners on Practice and Pleadings, at 196, submitted Dec. 31, 1849). A prospective juror could be challenged for “actual bias” when the juror expressed a state of mind such that the juror could not “try the issue impartially and without prejudice to the substantial rights of the party challenging” (Code Crim Pro § 376 [2]).

Additionally, as the Court said in People v Carolin (115 NY 658 [1889]), a challenge for “implied bias” could also be made in a capital case when a prospective juror expressed “such conscientious opinions as could preclude his finding the defendant guilty; in which case he shall neither be permitted nor compelled to serve as a juror” (Code Crim Pro § 377 [8]).

In 1968, the United States Supreme Court decided Wither-spoon v Illinois (391 US 510 [1968])—a case that influenced our *52statutoiy scheme relating to juror challenges. Witherspoon held that in a capital case no state could exclude jurors merely because they expressed general objections to the death penalty. The Court stated that, “[i]n its quest for a jury capable of imposing the death penalty, [Illinois] produced a jury uncommonly willing to condemn a man to die” (id. at 520-521).

Soon after, the Legislature added Criminal Procedure Law § 270.20 (1) (f) (L 1970, ch 996, § 1), permitting challenges for cause on the ground that:

“[tjhere is a possibility that the crime charged may be punishable by death and the prospective juror entertains such conscientious opinions either against or in favor of the death penalty as to preclude him from rendering an impartial verdict or from properly exercising the discretion conferred upon him by law in the setting of the penalty upon a proceeding conducted pursuant to section 125.35 of the penal law” (id.).19

It also modernized Code of Criminal Procedure § 376 (2), replacing it with CPL 270.20 (1) (b) (L 1970, ch 996, § 1).

At the same time, the Legislature enacted new CPL 270.20 (2), setting out the circumstances in which an erroneous ruling by the trial court on a challenge for cause would amount to reversible error. In People v Culhane (33 NY2d 90 [1973]),20 the Court focused on prospective jurors who believed defendants guilty primarily based on media accounts, observing that:

“Although the veniremen did not sit on the jury, because the defendants exercised peremptory challenges, this is of no consequence. It is well settled that an erroneous ruling by the court, denying a challenge for cause, constitutes reversible error when the defendant peremptorily challenges the prospective juror and his peremptory challenges are exhausted before the jury selection process is *53complete (People v. Casey, 96 N. Y. 115, 123; People v. Flaherty, 162 N. Y. 532, 537, 538). This rule of long standing, derived from the common law, has recently been codified in CPL 270.20 (subd. 2).”

It is important to recognize that this common-law rule embodied within CPL 270.20 (2) developed independently, and did not envision challenges for cause pertaining to penalty phase jurors. Indeed, until 1963, there was no such thing as a two-stage capital trial. Rather, CPL 270.20 (2) was concerned with actual bias that could affect a jury’s finding of guilt. By contrast, it is evident—and critical to our determination—that in enacting CPL 270.20 (1) (f), the Legislature was responding only to United States Supreme Court constitutional jurisprudence relating to sentencing phase bias.

This federal-state dialogue concerning the death penalty continued after the passage of CPL 270.20. In People v Fitzpatrick (32 NY2d 499, 509-513 [1973]), relying on the Supreme Court’s recent pronouncement in Furman v Georgia (408 US 238 [1972]), we held Penal Law § 125.35 (5) unconstitutional as violative of the Cruel and Unusual Punishment Clause because it permitted the jury to impose the penalty with unfettered discretion (see also Culhane, 33 NY2d at 95).

In response, the Legislature amended CPL 270.20 (1) (f) to delete the language pertaining to the juror’s exercise of discretion in setting a penalty. A subdivision (1) (f) for-cause challenge was thus limited to whether the juror’s conscientious opinions for or against the death penalty would preclude the juror from rendering an impartial verdict (L 1974, ch 367, § 14) .21

In an about-face from the course it took in 1963, and in response to Furman, the Legislature added Penal Law § 60.06, requiring mandatory death sentences for defendants convicted of murder in the first degree (L 1974, ch 367, § 2).22

That brings us to the present statute. In 1995, the Legislature amended CPL 270.20 (1) (f) (L 1995, ch 1, § 15), embracing Supreme Court standards for life/death qualification (Harris, 98 *54NY2d at 482-485). The Legislature sought to ensure that capital defendants receive the same protections afforded in federal prosecutions. To that end, and to craft an additional means of challenging a capital juror for cause, the lawmakers chose language akin to Witherspoon’s requirements regarding life and death qualifications. Indeed, the Legislature has continually responded to Supreme Court rulings in framing the standards for jury selection in capital punishment cases. Based on this progression, and particularly on the passage of the original CPL 270.20 (1) (f) in 1970 in response to Witherspoon, we conclude that the remedy intended in the case of an unqualified penalty phase juror is the same as that mandated by Supreme Court jurisprudence—the reversal of the sentence, as opposed to the entire trial.

This conclusion is buttressed by the explicit language of CPL 270.20 (1) (b) and (f). A subdivision (1) (b) challenge pertains to juror views related only to guilt. Inexorably, reversible error necessitates a new trial. By contrast, a subdivision (1) (f) challenge pertains only to jurors’ capital sentencing views, and reversible error in that context necessitates a new sentencing proceeding. Indeed, in People v Harris (98 NY2d 452 [2002]), we interpreted CPL 270.20 (1) (f)’s “preclude” language to embody the federal “prevent or substantially impair” standard of Wainwright v Witt (469 US 412 [1985]). Considering that our interpretation of the standards governing life and death qualification is derived exclusively from federal precedent, a violation should likewise parallel the federal remedy. Moreover, there is no reason to grant a defendant a windfall by ordering a new guilt phase trial when the jury selection error pertains only to sentence.

Here, the erroneous rulings pertaining to prospective jurors Nos. 23 and 855 were based on challenges that went only to their ability to deliberate fairly and impartially at the penalty phase. As to prospective juror No. 23, defense counsel specifically drew the connection between the juror’s experience with domestic violence and his difficulty in considering life without parole. Additionally, the court dismissed prospective juror No. 855, erroneously concluding that she was unsuitable for the penalty phase. This error does not justify a new guilt phase trial. Moreover, because we conclude that the penalty phase should not have taken place, these errors are of no consequence.

Procedurally, as Professor Preiser explains, a case is in limbo following reversal “absent specification of corrective action . . . *55required to either finally dispose of the case (e.g., dismissal of the indictment) or prescribe the next step or steps to be taken (e.g., new trial)” (Preiser, Practice Commentaries, McKinney’s Cons Laws of NY, Book 11 A, CPL 470.10, at 527). Simply put, in a reversal, a reviewing court may employ various types of corrective action. Here, vacatur of the sentence alone is not only apt, but also the norm (see Gray v Mississippi, 481 US 648 [1987]; see also Heard, 31 Cal 4th at 966-967, 75 P3d at 66; Farina v State, 680 So 2d 392, 396 n 3 [Fla 1996]; Morgan, 504 US at 739).

D. Defendant’s Absence at Sidebar Conferences During Jury Selection

Citing People v Antommarchi (80 NY2d 247 [1992]), defendant argues that his absence from two sidebar conferences during the group voir dire violated his statutory right to be present at the trial (see CPL 260.20). We conclude that defendant waived his right to be present at the bench conferences.

CPL 260.20 requires that “[a] defendant must be personally present during the trial of an indictment.” As we recently noted in People v Foster (1 NY3d 44 [2003]), this right extends to “sidebar and robing room conferences with prospective jurors regarding possible bias or hostility because they may give counsel input ‘in making discretionary choices during jury selection, based on impressions gained from seeing and hearing the juror’s responses on voir dire’ ” (id. at 47, quoting People v Roman, 88 NY2d 18, 26 [1996]). “Moreover, although the right is fundamental, it may be waived” (Foster at 48, citing People v Vargas, 88 NY2d 363, 375-376 [1996]).

Defendant waived his presence twice. First, at a March 1999 court appearance, defense counsel stated that *56Defendant answered yes. The court assented and asked to be told if defendant changed his mind. On another occasion, defense counsel advised the court that defendant waived his right to be present. The court asked defendant if this was true, and he answered yes. The court emphasized the point: “You’ve talked this over with your lawyer? You know you have an absolute right to be present at every stage of your trial?” Defendant replied: “Yeah, I understand that and I have talked it over with Mr. Priest.” Finally, at the start of group voir dire, defense counsel told the court that defendant “doesn’t wish to participate at the bench conferences.” The conferences in question occurred soon thereafter.

*55“I would request now for the record to show that [defendant] does not have to be present, unless a decision is made by his attorneys that he would be present. We think it’s highly prejudicial to bring him in front of the cameras, not allowed to dress in his proper clothing as he has been before, and I see no reason for it and just adds to the problems of this case.

“So I ask at this time the record should reflect, Mr. Cahill, you understand that you’re going to waive your right to be present in future appearances, except for hearings?”

*56Although “a trial court need not engage the defendant in an on-the-record colloquy to ensure the requisite voluntary, knowing and intelligent nature of the waiver” (Foster at 49, citing People v Spotford, 85 NY2d 593, 598 [1995]; People v Epps, 37 NY2d 343, 350-351 [1975]), defendant twice asserted on the record that he accepted and understood the waiver.23

III. Guilt Phase

A. Defendant’s Conviction Under Penal Law § 125.27 (1) (a) (v)

Witness elimination murder is committed when a defendant intentionally kills a victim who “was a witness to a crime committed on a prior occasion and the death was caused for the purpose of preventing the intended victim’s testimony in any criminal action” (Penal Law § 125.27 [1] [a] [v] [emphasis added]). As a threshold matter, the parties dispute the thrust of the words “for the purpose of.” Defendant argues that the crime is not made out unless a defendant kills the victim for the sole purpose of preventing the victim’s testimony. He argues, in essence, that the verdict cannot stand if there was proof that he had any motive for the murder other than his desire to prevent Jill’s testimony. The People contend that evidence of multiple motives may support a conviction for witness elimination murder.

Both parties cite the legislative memorandum, which states:

“Killings must be committed ‘for the purpose’ of preventing, influencing or retaliating for prior testimony. Thus, this provision is applicable when *57there is both a defined victim characteristic, (witness, family member) and when it can be proven that the defendant’s motivation for committing a killing was to prevent or influence the actual testimony of a victim in a criminal proceeding” (Mem of Codes Comm, at 2, Bill Jacket, L 1995, ch 1).

Like the statutory language, however, this recitation does not address mixed motives. We conclude that the statute would have to be read too expansively to authorize a conviction when a defendant’s motivation to eliminate a witness is insubstantial or incidental. Conversely, we cannot imagine that the Legislature intended to exclude a defendant—whose motivation to eliminate a witness was a substantial reason for the murder—merely because the defendant may have had other reasons or motives for the murder. Accordingly, the statute is satisfied if defendant’s motivation to eliminate Jill as a witness was a substantial factor in murdering her, even though he may have had mixed motives. Applying this standard, we address defendant’s claim that his conviction for murder in the first degree based on witness elimination (see Penal Law § 125.27 [1] [a] [v]) is both legally insufficient and against the weight of the evidence. We conclude that the evidence adduced on this count is legally sufficient, but that the verdict is against the weight of the evidence.

CPL 470.30 directs that criminal appeals taken directly to our Court are governed by CPL 470.15 and 470.20, which in turn address the scope of our review and the corrective action to be taken upon reversal or modification. Pursuant to CPL 470.15 (2) and (4), reversal or modification may be based on a determination that the evidence adduced at trial is not legally sufficient to establish the defendant’s guilt. On the other hand, CPL 470.15 (5) allows for reversal or modification when a verdict is, in whole or in part, against the weight of the evidence.

Legal sufficiency review and weight of the evidence review involve different criteria. In assessing legal sufficiency, a court must “determine whether there is any valid line of reasoning and permissible inferences which could lead a rational person to the conclusion reached by the jury on the basis of the evidence at trial” (People v Bleakley, 69 NY2d 490, 495 [1987], citing Cohen v Hallmark Cards, 45 NY2d 493, 499 [1978]). By contrast, weight of the evidence review recognizes that “[e]ven if all the elements and necessary findings are supported by some credible evidence, the court must examine the evidence further” (Bleak*58ley, 69 NY2d at 495). An appellate court weighing the evidence “must, like the trier of fact below, ‘weigh the relative probative force of conflicting testimony and the relative strength of conflicting inferences that may be drawn from the testimony’ ” (id., quoting People ex rel. MacCracken v Miller, 291 NY 55, 62 [1943]). If “based on all the credible evidence a different finding would not have been unreasonable” and if the “trier of fact has failed to give the evidence the weight it should be accorded,” the appellate court may set aside the verdict (id.).24 When an appellate court performs weight of the evidence review, it sits, in effect, as a “thirteenth juror” (Tibbs v Florida, 457 US 31, 42 [1982]).

Our dissenting colleagues are critical of our weight of the evidence analysis, claiming that this review is not an “open invitation” to substitute our own judgment for that of the jury. Of course that is true. But on the other hand, weight of the evidence review does not connote an invitation to abdicate our responsibility. A guilty verdict based on a legally sufficient case is not the end of our factual analysis but the beginning of our weight of the evidence review. Indeed, we are not only authorized to conduct this review but also constitutionally compelled to do so (NY Const, art VI, §§ 3, 5). Under the Constitution of our state, “in capital cases in which the sentence of death has been imposed, this court is vested with the power to and must review the facts” (People v Davis, 43 NY2d 17, 36 [1977], citing NY Const, art VI, §§ 3, 5; People v Carbonaro, 21 NY2d 271, 274 [1967]). In People v Crum (272 NY 348 [1936]), we underscored the importance of weight of the evidence review, noting that “[w]e are obliged to weigh the evidence and form a conclusion as to the facts. It is not sufficient, as in most of the cases with us, to find evidence which presents a question of fact; it is necessary to go further before we can affirm a conviction and find that the evidence is of such weight and credibility as to convince us that the jury was justified in finding the defendant guilty beyond a reasonable doubt.” (Id. at 350.)25

*59Thus, because a death sentence appeal comes directly to our Court, we conduct this type of analysis, which is routine in Appellate Division review of criminal cases (CPL 470.15 [5]). There is nothing the least bit novel about Appellate Division weight of the evidence reversals.26

In arguing for witness elimination murder, the People drew upon certain events in defendant’s life and matched them with steps he took toward planning Jill’s murder. The People linked (1) the May 1998 Family Court appearances with defendant’s Internet searches for cyanide; (2) the June and July 1998 developments in defendant’s assault case with his procurement of cyanide; (3) defendant’s October 1998 conversation with Patricia Cahill, defendant’s mother, about her visit to Jill in the hospital with defendant’s first disguised entrance into the hospital (in which a nursing assistant discovered defendant in Jill’s room); and (4) defendant’s upcoming Huntley hearing in the assault case with his having committed the murder. Viewed in the light most favorable to the People (see People v Carr-El, 99 NY2d 546, 547 [2002]), this timeline theory establishes a legally sufficient case. But it is barely that, and is decidedly against the weight of the evidence.

First, the May 1998 Family Court appearances primarily concerned the Cahill children. At the May 11 hearing, the Assistant District Attorney stated that the criminal court had issued an order of protection on behalf of the children and that they were potential witnesses in the assault case. The only evidence pertaining to the May 19 hearing involved psychological testing for the children. Jill’s status as a witness was not mentioned during these hearings—in fact, Jill was not mentioned at all. The better part of the evidence reveals that defendant was motivated to poison his wife because their marriage and family *60life were being destroyed, not because he wanted to kill a witness to the assault case.27

Using the prosecution’s timeline, a critical feature refutes its witness elimination theory: defendant procured potassium cyanide long before there was any possible belief—on anyone’s part—that Jill could ever testify at a trial, given her condition.28 After the assault, Jill suffered from a host of medical complications that utterly incapacitated her as a witness. Further, one doctor testified that as late as August 1998, when Jill entered the coma rehabilitation unit, her cognitive abilities ranked at a five or six on a 25-point scale. Moreover, at the corresponding criminal proceedings, there was no mention that Jill might or possibly could testify against defendant. Indeed, there is not a shred of evidence in this record that Jill had even retained a memory of the assault. Most compellingly, an assistant district attorney testified that up to the date of Jill’s murder no law enforcement official had even interviewed her about the assault.

In seeking to prove that defendant knew Jill was able to speak and was afraid of what she would say, the People relied largely on the testimony of Patricia Cahill, defendant’s mother. She *61merely testified, however, that defendant told her that “ T hope that when this is all concluded that she can tell the truth about what led to the break-up’ of their marriage.” This statement does not evince defendant’s fear of Jill’s possible testimony in the assault trial; it is more a comment about the dissolution of his family. In addition, Patricia Cahill testified that, during a visit, Jill “said something” to the nurses, and also that she told defendant about Jill’s physical condition. But we do not know what Jill said, or what Patricia Cahill told defendant about Jill’s status. Furthermore, Fred Russell, Jill’s father, stated that “there was not a word spoken” during Patricia Cahill’s visit to the hospital. Thus, Patricia Cahill’s testimony does not even suggest, let alone reveal, what defendant knew about Jill’s condition and speaking ability. It does not begin to show that he was afraid that Jill would testify against him at the assault trial.29

Additional factors rebut the witness elimination theory. The very brutality of the April 1998 assault permits the compelling inference that, even as of then, defendant wanted to kill his wife and that ultimately doing so fulfilled his previously formed intent, which sprang from the impending divorce.30

Beyond that, defendant had fully confessed to the assault and the prosecution had a powerful case without Jill’s testimony. In confessing to the assault, defendant admitted that he hit Jill on the head at least three or four times with the baseball bat. He further admitted that he struck her while she was unarmed, thus foreclosing any plausible claim of self-defense. He also confessed to having wounded himself, a deception designed to make it look as if he acted in self-defense, and that he attempted *62suicide after the assault. In light of these candid disclosures, there is scant basis to believe that defendant thought he could avoid an assault conviction by murdering Jill.

In weighing the conflicting inferences that can be drawn from the facts, the proof leads us to conclude that defendant, to put it plainly, wanted to kill Jill at the hospital for reasons that had virtually nothing to do with her ability to testify against him. The weight of the evidence does not support witness elimination as a substantial motive for the murder, and we therefore hold that the conviction for murder in the first degree under Penal Law § 125.27 (1) (a) (v) must be vacated.

B. Defendant’s Conviction Under Penal Law § 125.27 (1) (a) (vii)

In addition to witness elimination murder, defendant was convicted under Penal Law § 125.27 (1) (a) (vii). That section elevates intentional murder to capital-eligible murder when a defendant with “intent to cause the death of another person . . . causes the death of such person . . . and . . . the victim was killed while the defendant was in the course of committing or attempting to commit and in furtherance of . . . burglary in the first degree or second degree.” As a matter of statutory interpretation, we conclude that the conviction cannot stand because the burglary carried no intent other than to commit the murder.

Penal Law § 125.27 (1) (a) (vii) is best understood in light of the Legislature’s purpose in devising aggravating factors as predicates for the death penalty. In Tuilaepa v California (512 US 967, 971-972 [1994]), the Supreme Court held that “[t]o render a defendant eligible for the death penalty in a homicide case . . . the trier of fact must convict the defendant of murder and find one ‘aggravating circumstance’ (or its equivalent) at either the guilt or penalty phase. See, e.g., Lowenfield, v. Phelps, 484 U.S. 231, 244-246 (1988); Zant v. Stephens, 462 U.S. 862, 878 (1983).”

The Legislature drew up a list of aggravating factors to create a subclass of defendants who, in contrast to others who commit intentional murder, it thought deserving of the death penalty. By this device, the lawmakers saw to it that the death penalty could not fall randomly on all murder defendants. The Legislature’s factors govern the discretion of courts and juries by limiting capital punishment to certain enumerated categories of *63intentional killings, ensuring that the State follows its “constitutional responsibility to tailor and apply its law in a manner that avoids the arbitrary and capricious infliction of the death penalty” (Godfrey v Georgia, 446 US 420, 428 [1980]; see also Zant, 462 US at 876-878).

Penal Law § 125.27 (1) identifies death-eligible defendants as those who commit intentional murders in the context of one or more of 13 aggravating factors. Five aggravating factors relate to the killing of a member of a specific group (police officers, peace officers, corrections employees, witnesses and judges [sub-pars (i)-(iii), (v), (xii)]), two relate to the present or past circumstances of the offender (defendants serving life sentences and defendants previously convicted for murder [subpars (iv), (ix)]), four address the circumstances of the killing or criminal transaction (murder committed in furtherance of certain enumerated felonies, multiple murders as part of the same criminal transaction, murder by torture and terrorism [subpars (vii), (viii), (x), (xiii)]). The remaining two involve contract killing and serial murder (subpars [vi], [xi]) (see Donnino, Practice Commentary, McKinney’s Cons Laws of NY, Book 39, Penal Law § 125.27, at 386-391).31

Among other arguments, defendant contends that he cannot be convicted under Penal Law § 125.27 (1) (a) (vii) because the statute requires that the underlying felony (here, burglary in the second degree) have an objective apart from the intentional murder and that the burglary was merely an act that enabled the murder, one of many anticipatory steps along the way.32

By the terms of the indictment, defendant was charged with having killed his wife “in the course of committing or attempting to commit and in the furtherance of the crime of Burglary in the Second Degree.”33 Because our analysis turns on whether defendant’s burglary qualifies as an aggravating factor, we must address the crime of burglary in the setting before us.

*64Burglary is part of a larger category of criminal behavior that involves intrusion upon property (Penal Law art 140). The statutory hierarchy is relevant. The lowest degree of intrusion is criminal trespass (a violation), by which a person knowingly enters or remains unlawfully in or upon premises (Penal Law § 140.05). From there, a trespass becomes more serious, depending on the nature of the premises and whether the trespasser possesses certain weapons (Penal Law §§ 140.10, 140.15, 140.17). The critical distinction between burglary and trespass is that a trespass in a building or dwelling is complete when a person knowingly enters or remains unlawfully in those premises. Burglary requires more. There can be no burglary unless the trespasser intends to commit a separate crime when entering or remaining unlawfully in a building (see People v Gaines, 74 NY2d 358 [1989]). Burglary is thus an aggravated form of criminal trespass, in which the aggravating factor is the trespasser’s intent to commit a separate crime (see People v Henderson, 41 NY2d 233 [1976]; see also CJI2d [NY] Penal Law § 140.25 [1] [b]).

With that in mind, we may better appreciate how burglary fits into the design of Penal Law § 125.27 (1) (a) (vii). The statute begins by declaring that every first degree murder must include an intentional (second degree) murder. An additional aggravating factor—murder “plus”—raises the crime to murder in the first degree. A candidate for first degree murder is the burglar who enters a dwelling to steal or rob or rape and in addition kills someone intentionally, in the course and furtherance of the burglary. It is this double crime—murder “plus”—that is the defining core of Penal Law § 125.27 (1) (a) and renders the offender eligible for the death penalty.

The case before us does not fit this statutory paradigm. Burglary requires an intent to commit a crime in the burglarized premises, and here the prosecutor uses defendant’s “intent to kill” to satisfy the burglary definition. But the very same mens rea—the intent to kill—also defines intentional murder (Penal Law § 125.25). Thus, the prosecution employs the identical mens rea both to define burglary and to elevate defendant’s intentional murder to murder in the first degree. The defense argues that this circularity is impermissible and that the capital murder statute contemplates a felonious intent independent of the murder itself. We agree.

*65A burglar who intends, for example, both to rob and murder is committing two crimes, both felonies, whose intents are purposively independent of each other. The robbery may be committed in connection with the murder, but as a substantive crime it is distinct from the murder and can be aptly characterized as an aggravating factor that fulfills Penal Law § 125.27 (1) (a) (vii). It is the “plus” (in the “murder plus” formulation) that is necessary to make it a death-eligible crime.

Burglary, however, is different because it is a trespass—a misdemeanor—that becomes a felony only if the trespasser intends to commit a separate crime when entering a building (see Gaines). If the burglar intends only murder, that intent cannot be used both to define the burglary and at the same time bootstrap the second degree (intentional) murder to a capital crime. To do so would not narrow the class of those eligible for the death penalty, but would widen it. In promulgating the list of aggravators, the Legislature did not expressly sweep within Penal Law § 125.27 all killings in which the murderer unlawfully entered the victim’s home. We decline to imply such an intent, let alone write one into the statute, in the face of the unswerving legislative goal of narrowing rather than expanding the class of defendants eligible for the death penalty. Nor will we, in the absence of legislative intent or expression, have life or death hinge on whether a defendant engaged in conduct that simply enabled the intended murder and had no point of its own. To do so would spurn rather than follow the Legislature’s objectives.

In arguing that defendant’s burglary satisfies Penal Law § 125.27 (1) (a) (vii), the prosecutors and our dissenting colleagues rely heavily on People v Miller (32 NY2d 157 [1973]). They assert that Miller applies and that if we read Penal Law § 125.27 (1) (a) (vii) as requiring a felony independent of the murder, we will be overruling a body of felony murder jurisprudence that extends back three decades. This is not so. Miller does not govern this case, and the reason is plain: in Miller, the Court interpreted Penal Law § 125.25 (3), the felony murder statute. Here, we are reviewing not felony murder but Penal Law § 125.27 (1) (a) (vii), a capital punishment statute directed at those who commit intentional murder, and more. Miller is distinguishable on the facts and in its legal premise. By today’s decision, we leave our body of felony murder jurisprudence intact.

In Miller, the defendant knocked on the door of Fennell’s apartment intending to assault Fennell. When Fennell opened *66the door, Miller stabbed him in the arm. Fennell’s roommate, Aleem, intervened and Miller killed Aleem. Miller was convicted of felony murder and manslaughter in the second degree as to Aleem and first degree assault as to Fennell. The Court upheld the felony murder conviction as against Miller’s contention that neither the assault on Fennell nor the killing of Aleem could serve as a felony to satisfy the felony murder doctrine.

The defense argues that Miller stands at most for the proposition that the defendant’s conviction rested on his having killed the roommate, Aleem, to advance an independent, qualifying felony (burglary) and that the rest of the opinion is dicta, pointing out that Judges Jones and Gabrielli concurred in the result. The concurring Judges stated that Miller’s conviction for felony murder could properly be sustained on the basis of his having killed Aleem and that the Court need not have addressed the question whether the assault upon Fennell could qualify as a felony under the felony murder statute.34 The prosecution disagrees with defendant’s characterization of the case as partly dicta.

The obvious response to this dispute over the reach of Miller is that in the case before us we are not in the felony murder arena and need not determine how we would rule if we were. We therefore have no occasion to decide whether in a felony murder case—which this is not—Miller should be extended to a defendant who enters a building to murder the victim (in contrast to having the intent to assault the victim, as in Miller). We did not address that question in Miller, let alone decide it, and it would be improvident for us to render an opinion in a hypothetical case under a statute involving concepts and purposes different from the one before us.35

*67In contrast to Penal Law § 125.27 (1), felony murder covers nonintentional killings. The very purpose of the felony murder doctrine is to utilize the underlying felony as a substitute for the defendant’s murderous intent and thereby raise an unintentional killing to the level of murder (see People v Chico, 90 NY2d 585 [1997]; People v Lytton, 257 NY 310 [1931]). As we said in People v Hernandez (82 NY2d 309, 317 [1993]), “The basic tenet of felony murder liability is that the mens rea of the underlying felony is imputed to the participant responsible for the killing. By operation of that legal fiction, the transferred intent allows the law to characterize a homicide, though unintended and not in the common design of the felons, as an intentional killing” (citations omitted).

The felony murder concept was derived from the common law, at which no intent to kill was necessary. It was enough that the victim was killed while the accused was engaged in the commission of a felony.36 Under the common law, the felonious intent was imputed to the committed act, and, if it were homicide, made it murder (see People v Enoch, 13 Wend 159 [1834]).37

Penal Law § 125.27 (1) (a) (vii) borrows language from the felony murder statute but is critically different because it deals with intentional, not unintentional, killings. The purposes of *68the capital statute and the felony murder statute are distinct, and the felonies covered by them are not the same. For example, under the felony murder statute, the killing need not be committed by one of the people engaged in the commission of the underlying crime (.People v Hernandez, 82 NY2d 309 [1993]), whereas Penal Law § 125.27 (1) (a) (vii) does not apply where the defendant’s liability is based on someone else’s conduct (unless the defendant commanded the murder). Moreover, felony murder liability for the death of a victim has been broadly construed (see e.g. People v Ingram, 67 NY2d 897 [1986]; People v Matos, 83 NY2d 509 [1994]), whereas the Legislature crafted Penal Law § 125.27 (1) (a) to narrow the class of eligible offenders. Conceptually, Penal Law § 125.27 (1) (a) begins with murder in the second degree and builds on it. Conversely, felony murder builds toward it. Thus, felony murder ends up as murder, whereas Penal Law § 125.27 (1) (a) begins with murder. The two concepts share certain components but have entirely different objectives and constituents, and were statutorily constructed to reach different types of homicides and different categories of defendants.38

For these reasons, the jurisprudence underlying the Miller felony murder statute cannot be equated with the goals of Penal Law § 125.27 (1) (a). Furthermore, Miller deals with assault and we have not been made aware of a single case in which this Court (or any appellate court in New York) ever discussed whether a felony murder conviction may be based upon a burglary with an underlying intent to kill, let alone held that way.39

The parties cite out-of-state cases dealing with whether the intent to kill can serve as a predicate for burglary when the same intent is used to elevate a murder to a capital offense. It is appropriate that we consider these decisions. The prosecution cites State v Tillman (750 P2d 546 [Utah 1987]) and Smith v *69State (499 So 2d 750 [Miss 1986]), in which the highest courts of Utah and Mississippi upheld capital sentences where the intentional murders were each aggravated by having been carried out during a burglary with the killing as its objective.40

The defense counters with Parker v State (292 Ark 421, 731 SW2d 756 [1987]) and People v Green (27 Cal 3d 1, 59-62, 609 P2d 468, 504-506 [1980]). In Parker, the Arkansas Supreme Court held that the defendant’s burglary was a facilitating step along the way in fulfilling his intent to commit murder. Because the underlying burglary had no objective independent of the murder, the court refused to elevate the murder to capital offense status. The Parker court stressed that the Arkansas statute required that the murder be committed in the course of and “in furtherance of’ the felony. The court pointed out that Parker did not commit the murder “in furtherance of’ the burglary and that the very opposite was true: he committed the burglary in furtherance of his intent to kill his victim.

We note also that in the most recent case on this subject the Supreme Court of Delaware reached the same conclusion {see Williams v State, 818 A2d 906 [Del 2002]). In reversing a death sentence, the court held that “if the intent of the burglary was to commit murder, the death that occurred was not 'in furtherance of the burglary” (id. at 908). Although it may have been carried out “ ‘in the course of the burglary,” it was not “carried out ‘in furtherance of it” (id. at 913).

Putting aside the possible differences in the language and legislative histories of the statutes in Utah, Mississippi and New York, we respectfully disagree with the rationale expressed in Tillman and Smith.41 As for Parker and Williams, we note that our statute—like those in Arkansas and Delaware— *70requires that the murder be in the course of and “in furtherance of’ the burglary.42 The Supreme Courts of Arkansas and Delaware based their analyses chiefly on the “in furtherance of’ requirement and held it cannot be satisfied where the burglary is committed to further the murder, as opposed to the converse.

Because we reach the same result on slightly different grounds, we need not and do not determine whether the prosecution is correct in its assertion that defendant killed his wife “in furtherance of’ the burglary. It is certain that the converse is true—defendant committed the burglary to further his intent to kill his wife. That being so, the burglary was not meaningfully independent from the murder. Defendant’s trespass on the hospital premises was merely a prerequisite to his committing the murder—an enabling measure that had no purpose or substance other than to serve his only goal, to kill his victim.

In Green (27 Cal 3d at 59-62, 609 P2d at 504-506), the Supreme Court of California addressed a question similar to the one before us. Green is pertinent because the California capital punishment statute does not require that the murder be “in furtherance of’ the felony. The court therefore did not rely on the “in furtherance” language that influenced the results in Parker and Williams. Although Green involved a robbery, the analysis is germane and bears repetition, considering that we, like the California Supreme Court, hold that the felonious intent must be independent of the murder, if it is to qualify as an aggravating factor:

“The Legislature’s goal is not achieved, however, when the defendant’s intent is not to steal but to kill and the robbery is merely incidental to the murder—‘a second thing to it,’ as the jury foreman here said—because its sole object is to facilitate or conceal the primary crime. In the case at hand, for example, it would not rationally distinguish between murderers to hold that this defendant can be subjected to the death penalty because he took his victim’s clothing for the purpose of burning it later to prevent identification, when another defendant who committed an identical first degree murder could not be *71subjected to the death penalty if for the same purpose he buried the victim fully clothed—or even if he doused the clothed body with gasoline and burned it at the scene instead. To permit a jury to choose who will live and who will die on the basis of whether in the course of committing a first degree murder the defendant happens to engage in ancillary conduct that technically constitutes robbery or one of the other listed felonies would be to revive ‘the risk of wholly arbitrary and capricious action’ condemned by the high court plurality in Gregg.” (27 Cal 3d at 61-62, 609 P2d at 505-506 [citations omitted].)43

In sum, we reject the approach adopted by Mississippi and Utah and, while aware of the differences among the respective statutes, join Delaware, Arkansas and California in refusing to elevate a case such as this to capital murder status. We acknowledge that Penal Law § 125.27 (1) (a) (vii) could be read differently and conceivably be given the interpretation urged by the prosecution. But to do so we would have to stretch the statute’s meaning and bend the language to encircle the case before us. Requiring an intent independent of the intentional murder of the victim, as we do, is both the most sensible reading of the statute and the one most consonant with the Legislature’s intent in establishing the statute’s aggravating factors.44

*72We are also aware of the dissent’s contention that our interpretation could engender results that, on the surface, appear incongruous. The defendant who breaks into a home with the joint intent of killing the occupant and stealing an appliance would, under our interpretation, be death-eligible, but a defendant who breaks into the same home for the sole purpose of killing the occupant would not. This might appear as a surface flaw, but on further analysis the result is fully in keeping with the statutory plan. In the former case, the defendant who burglarizes in order to steal and commits intentional murder would be punished more severely under Penal Law § 125.27 (1) (a) (vii) for having committed both murder and a burglary intended apart from the murder. This result is neither arbitrary nor unjust, and is more faithful to the Legislature’s language and design.45

IV Conclusion

Accordingly, the judgment of Onondaga County Court should be modified by reducing defendant’s conviction of two counts of murder in the first degree to one count of murder in the second degree and remitting to that court for resentencing on the second degree murder as well as the remaining counts and, as so modified, affirmed.

G.B. Smith, J. (concurring). I concur with the Court’s conclusions that the trial court’s failure to grant appropriate cause challenges went to the sentencing phase of the trial and does not require a new trial, and that the convictions for capital murder, murder in the first degree with burglary as the underlying felony and murder in order to eliminate a witness from testifying must be reversed. I join the Court’s decision insofar as it concludes that defendant was convicted of intentional mur*73der beyond a reasonable doubt and remits for sentencing for murder in the second degree and associated charges.

I address two other issues—the deadlock jury instructions which are required to be given by CPL 400.27 (10) and the arbitrariness of the death penalty statute. Both issues have been raised by the defendant. Moreover, the dissent, while concluding that the judgment should be affirmed, does not address all of the arguments raised by the defendant. While I would prefer that the entire Court, majority and dissents, deal with these issues, I address them because they are central to a determination of the validity of a death penalty prosecution.

Because the provision requiring the jury to receive a particular deadlock instruction is coercive, I would reach that issue. Judge Ciparick joins me in concluding that the deadlock instruction is coercive. I also disagree with the failure of the Court to address the arbitrariness of the death penalty scheme, and would reach that issue.

On April 21, 1998, two weeks after signing a separation agreement with his wife, defendant James F. (Jeff) Cahill assaulted his wife, Jill Cahill, with an aluminum baseball bat. The incident left her unconscious, with severe head injuries for which she spent months in a hospital. Defendant first claimed that she came at him with a knife and he hit her in self-defense. Later he admitted that he continued to hit her after she dropped the knife. “When Jill Cahill fell down in the mud room after [I] hit her, she dropped the knife. After helping her back into the kitchen, [I] retrieved the knife from the mud room and put it on the kitchen floor. Back in the kitchen, the struggle resumed, and [I] hit her with the bat at least twice when she was not holding the knife.” Defendant further admitted that his wounds were self-inflicted. “[I] used a nail to make the scratches on [my] shoulder and arm.”

Defendant was arrested and indicted for assault in the first degree (Penal Law § 120.10) and criminal possession of a weapon in the fourth degree (Penal Law § 265.01). As a result of the arrest and indictment, defendant was not allowed to have contact with his wife or his two children. On October 27, 1998, while Jill was still in the hospital recovering from the assault, defendant entered her room and poisoned her with cyanide.1 Just before Jill was poisoned, several nurses and other hospital workers saw defendant in the hospital disguised as a janitor. *74About a week earlier, one of the nursing assistants recalled seeing defendant in the hospital so disguised.2

Defendant was arrested at home on October 27th for the incident at the hospital. The next day the police seized a computer from the Cahills’ house. From the computer, the police were able to ascertain evidence of Internet searches for information about cyanide, and a letter on letterhead from a local company ordering cyanide. It was later determined that the letter and the letterhead were forgeries. Additionally, the police recovered a container of cyanide which was hidden near the shed on the property.

Defendant was indicted for one count of murder in the first degree (Penal Law § 125.27 [1] [a] [vii]), with burglary as the underlying felony (Penal Law § 140.25); one count of murder in the first degree (Penal Law § 125.27 [1] [a] [v]), a killing with the intent to prevent a witness from testifying; two counts of murder in the second degree, burglary in the second degree, criminal possession of a weapon in the fourth degree and aggravated criminal contempt (Penal Law § 215.52).3 In January 1999 the two indictments were consolidated.

During the guilt/innocence phase of the trial (August 3 to August 9, 1999), after the People presented their case, defendant called one witness. On August 10, 1999, at the conclusion of the trial, defendant was found guilty of one count of murder in the first degree (Penal Law § 125.27 [1] [a] [vii]), with burglary as the underlying felony (Penal Law § 140.25); one count of murder in the first degree (Penal Law § 125.27 [1] [a] [v]), killing to prevent a witness from testifying; two counts of criminal possession of a weapon in the fourth degree (Penal Law § 265.01); and one count of assault in the first degree (Penal Law § 120.10).

During the penalty phase, August 17 to August 19, 1999, defendant, after the People’s case, called 21 witnesses. On August 20, 1999, the jury determined that defendant should be sentenced to death as to each count of murder in the first degree. On October 5, 1999, defendant was sentenced to I2V2 to *7525 years in prison for assault in the first degree and a concurrent sentence of one year on the two counts of criminal possession of a weapon. Pursuant to CPL 450.70 (1), defendant filed his appeal directly to this Court.

The Current Disposition

Four Judges of this Court conclude that defendant was properly convicted of intentional second degree murder. The dissents, without addressing all of the arguments raised by the defendant, conclude that he was properly convicted of both counts of capital murder. In People v Harris (98 NY2d 452 [2002]) and again in this case, defendants have raised issues concerning the constitutionality of the capital murder legislation. I believe those issues should be addressed.

I continue to believe that heightened scrutiny must be applied to all aspects of a capital murder case {see People v Harris, 98 NY2d at 497-506). Even if that doctrine is not applied here, there are arguments raised by defendant which go to the heart of any case under the capital murder legislation. These include the court’s instructions on a deadlocked jury, instructions required by CPL 400.27 (10), and the arbitrariness of the death penalty legislation.

Challenges for Cause to Some Prospective Jurors

A prospective juror must give an unequivocal assurance that he or she can be fair and impartial (People v Johnson, 94 NY2d 600 [2000]). In People v Johnson this Court stated:

“Thus, from the statute [CPL 270.20] and case law, the guiding principles are perfectly plain: when potential jurors reveal knowledge or opinions reflecting a state of mind likely to preclude impartial service, they must in some form give unequivocal assurance that they can set aside any bias and render an impartial verdict based on the evidence. Obviously, when potential jurors themselves openly state that they doubt their own ability to be impartial in the case at hand, there is far more than a likelihood of bias, and an unequivocal assurance of impartiality must be elicited if they are to serve” (94 NY2d at 614 [emphasis in original]).

Thus, a trial court must excuse a prospective juror who displays a disqualifying state of mind unless the juror gives a clear and *76unambiguous assurance that he or she can be impartial. “[T]he Trial Judge should require the prospective juror to ‘expressly state that his prior state of mind . . . will. . . not influence his verdict, and . . . that he will render an impartial verdict based solely on the evidence’ ” (People v Johnson, 94 NY2d at 613, quoting People v Blondo, 41 NY2d 483, 485 [1977]).

A general statement of impartiality that does not explicitly address the specific cause of the preexisting bias is not sufficient. Instead, “jurors must clearly express that any prior experiences or opinions that reveal the potential for bias will not prevent them from reaching an impartial verdict. If there is any doubt about a prospective juror’s impartiality, trial courts should err on the side of excusing the juror, since at worst the court will have ‘replaced one impartial juror with another’ ” (People v Arnold, 96 NY2d 358, 362 [2001]).

Defendant argues that the trial court should have dismissed prospective juror No. 23 and should have rejected the People’s cause challenge with respect to prospective juror No. 855. While recognizing that some of the initial answers given by those prospective jurors could arguably reflect on their ability to be impartial on the guilt phase, I join the Court in its conclusion that the potential lack of impartiality of these two prospective jurors affected their impartiality only as to the sentencing phase of the trial.

Anticipatory Deadlock Instruction

Where the death penalty is sought on a first degree murder indictment, if the jury finds the defendant guilty, it must then determine how the defendant’s crime should be punished. To assist the jury with its sentencing responsibilities, CPL 400.27 (10) requires the trial court to instruct the jury that it must unanimously decide whether to impose the death sentence or a sentence of life imprisonment without the possibility of parole. That provision further requires the court to instruct the jury that “in the event the jury fails to reach unanimous agreement with respect to the sentence, the court will sentence the defendant to a term of imprisonment with a minimum term of between twenty and twenty-five years and a maximum term of life.” Defendant argues that this instruction is unconstitutional because of the substantial risk that jurors who believed that a sentence of life without parole was appropriate would be coerced into voting for the death penalty in order to avoid the possibility that the defendant might someday be released from prison. *77At the outset, it is necessary to note that because of its severity and irrevocability, the penalty of death is qualitatively different than any other type of sentence, regardless of the length of the period of imprisonment (see Woodson v North Carolina, 428 US 280, 305 [1976] [plurality op]). Because of this qualitative difference, the United States Supreme Court has recognized that there is a heightened “need for reliability in the determination that death is the appropriate punishment in a specific case” (id.; see People v Harris, 98 NY2d at 497-506 [Smith, J., concurring in part and dissenting in part], citing Furman v Georgia, 408 US 238, 306 [1972] [Stewart, J., concurring]; Caldwell v Mississippi, 472 US 320, 340 [1985]; California v Ramos, 463 US 992, 998-999 [1983]). In order to ensure that “the death penalty is indeed imposed on the basis of ‘reason rather than caprice or emotion,’ ” the Court has “invalidated procedural rules that tended to diminish the reliability of the sentencing determination” (Beck v Alabama, 447 US 625, 638 [1980]; see Gardner v Florida, 430 US 349, 358 [1977]). A jury instruction that introduces a level of uncertainty and unreliability “cannot be tolerated in a capital case” (Beck at 643; see People v Harris, 177 Misc 2d 160, 162 [Sup Ct, Kings County 1998, Feldman, J.]).

New York leaves the decision of capital sentencing in the hands of the jury. Like the majority of such jurisdictions, New York provides that when a jury is unable to reach a unanimous verdict, the trial judge assumes the responsibility to impose the defendant’s sentence according to its respective statute.

New York is but one of four states that instructs its juries as to the consequences of a failure to decide between death and some determined lesser sentence.4 However, in each of the other states that requires such an instruction, the court in the event of deadlock, is required to impose a lesser penalty that the jury had been considering. Only in New York is the trial judge required to impose a sentence more lenient than the two sentencing options upon which the jury had deliberated. Such a sentencing scheme is irrational and carries a substantial risk of coercing a unanimous sentencing verdict from a jury, even under circumstances where individual jurors hold a genuine belief that their position, though different than that of their fellow jurors, is the just one. The sentencing scheme to be employed upon a jury deadlock is therefore unconstitutional.

*78While the legislative history of the New York death penalty statute offers no insight into why the only sentence available for the trial court to impose upon a jury deadlock would be a sentence more lenient than the options the jury had been considering, the legislative debates demonstrate that the members of the Legislature were aware of the risk that the jury instruction could result in a coerced verdict but consciously disregarded the risk.5 6 During the State Senate debate over the issue of the anticipatory deadlock instruction, the following exchange took place:

“senator dollinger: . . . [H]ow do you avoid the problem of a jury that is hung up on the issue of either life in prison without parole or the death penalty of putting additional pressure on the jurors, knowing that if they failed to agree they are going to face a penalty that is less than either of the two penalties that they are currently in dispute over?

“senator volker: That’s, I think, the option that they face.” (New York State Senate Debate on Senate Bill S 2850, Mar. 6, 1995, at 1912.)6 At a subsequent exchange, the following occurred:

“senator dollinger: . . . But isn’t it inherently coercive to tell them that you have to [reach a consensus on the most severe penalties]; otherwise, there is going to be another penalty imposed?

“senator volker: They don’t have to do anything. I mean what you do have to do, I think, Senator, the problem would be—my own personal feeling is I think there could be a problem if you didn’t inform the jury right up front as to what happens when they fail to make a decision” (id. at 1916).

*79Senator Volker has since introduced a bill amending the death penalty statute that, among other things, proposes that the sentence to be imposed by the trial judge upon a jury impasse with regard to sentencing would be the non-death sentence that the jury had been considering—life imprisonment without the possibility of parole (see 2001 NY Senate Bill S 5409, New York 224th Annual Legislative Session, introduced May 31, 2001).

The sentencing consequence of a deadlock, as provided by CPL 400.27 (10), creates an intolerable risk of coercion on the jury’s deliberative process. Essentially, jurors are informed at the very outset of their deliberations that if they cannot unanimously decide whether the defendant should be sentenced to death or to life imprisonment without the possibility of parole, the judge will impose a prison term of life with the possibility of parole after serving from 20 to 25 years. Such an instruction plainly goes to the heart of one of the chief concerns a deliberating jury would have about the defendant—his or her future dangerousness (see Simmons v South Carolina, 512 US 154, 162 [1994] [“a defendant’s future dangerousness bears on all sentencing determinations made in our criminal justice system”]; Jurek v Texas, 428 US 262, 275 [1976] [“any sentencing authority must predict a convicted person’s probable future conduct when it engages in the process of determining what punishment to impose”]).

Future dangerousness of a defendant is such a crucial issue for a jury deciding a defendant’s sentence because jurors do not want to be responsible for the release of a defendant they believe will continue to be a societal threat (see William J. Bowers and Benjamin D. Steiner, Death by Default: An Empirical Demonstration of False and Forced Choices in Capital Sentencing, 77 Tex L Rev 605, 701 [1999] [according to accounts, jurors often impose death, not because they deem the sentence to be retributively appropriate, but for the incapacitative purpose of removing the defendant from society]). According to an empirical study conducted by the Capital Jury Project in South Carolina, one of the most important considerations jurors make in assessing a defendant’s future dangerousness is “the probable actual duration of the defendant’s prison sentence” if the death penalty is not imposed (Eisenberg and Wells, Deadly Confusion: Juror Instructions in Capital Cases, 79 Cornell L Rev 1, 4, 7 [1993]). Jurors perceive a defendant likely to be released after a shorter prison term as more dangerous than the same defendant expected to serve a longer term. Consequently, jurors tend *80to sentence to death those defendants expected to serve a shorter term (see id.; Laurie B. Berberich, Note, Jury Instructions Regarding Deadlock in Capital Sentencing, 29 Hofstra L Rev 1301, 1324 [2001] [“The sooner (jurors) think the defendant will get out of prison, the more likely they are to vote for death”]).7

It rationally follows that a jury’s perception of a defendant’s future dangerousness is vastly diminished when the only options for his sentence are death and life imprisonment without the possibility of parole. The availability of a life sentence without the possibility of parole necessarily means that the death penalty is no longer the only way to achieve a defendant’s permanent incapacitation (see Bowers and Steiner, Death by Default, 77 Tex L Rev at 707 [life without parole and death are “essentially equivalent for incapacitative purposes”]). “Indeed, there may be no greater assurance of a defendant’s future non-dangerousness to the public than the fact that he never will be released on parole” (Simmons, 512 US at 163-164).

However, a jury instruction that the failure to reach a unanimous verdict between those two options would result in a parole-eligible life sentence interjects the issue of the defendant’s future dangerousness into the deliberative process where it otherwise would have no place. Where jurors are divided as to which incapacitative sentence to impose, they inevitably are faced with the possibility that if they do not come to a consensus, the defendant will be sentenced to a term that will possibly result in his eventual release from prison—a prospect *81that the jurors are not being asked to contemplate, and possibly may not want.8

*82Plainly, the most disturbing risk that the statute poses is that a juror who is in the minority in his or her vote for life without parole would inevitably feel pressured to vote with the majority for the death sentence in order to avoid the possibility of defendant’s eventual freedom.9 And where the jury vote is substantially in favor of the death sentence, the jurors in the minority no longer are choosing between death and life without parole; they are choosing between death and life with the possibility of parole within 20 to 25 years as a result of creating an impasse. Faced with the possibility that they will not otherwise be able to prevent the defendant’s return to society, jurors favoring life without parole may relinquish their conscientiously held beliefs and vote for the death penalty (see Berberich, Jury Instructions, 29 Hofstra L Rev at 1325 [“Common sense leads to the conclusion that a juror who would impose death over a life sentence because of his or her fear that a capital offender would be released on parole, would also impose a death sentence if he or she believed that a hung jury would lead to the same result.”]).

A unanimous death sentence, secured under these circumstances, would not be the product of reason but rather of uncertainty and coercion; it therefore cannot withstand constitutional scrutiny (see Beck v Alabama, 447 US 625 [1980] [where the death penalty was automatic upon conviction of first degree murder and the state statute forbade the jury from considering lesser included offenses, the Supreme Court found the statute unconstitutionally coercive in violation of the Eighth Amendment because rather than determining solely whether the evidence established the defendant’s guilt, the statute inevi*83tably forced jurors to take into account whether the defendant would die for his crime or simply go free]).

The People argue that the challenged jury instruction could also have the effect of pressuring jurors who favor the death sentence to vote in favor of life without parole in order to avoid the more lenient deadlock sentence. The People are certainly correct on this point. But for two important reasons, the possibility of such an outcome does not matter. First, as explained above, the finality and irrevocability of a death sentence makes that punishment qualitatively different than any other punishment that may be imposed, no matter how long the period of imprisonment (see Woodson, 428 US at 305 [plurality op]). It is the possibility of the death sentence that triggers a heightened scrutiny analysis, not the possibility of a life sentence. Surely, a statutory provision which creates a substantial risk that jurors would be coerced into sentencing a defendant to die cannot be saved by the fact that the provision also creates the risk that the jury would be coerced into sparing his life. These are not comparable outcomes. Moreover, because we can never predict in any given case how the votes of the jury would be configured, we cannot anticipate how the pressure of the deadlock sentence would be applied (see Harris, 177 Misc 2d at 164). In any event, the substantial risk of coercion posed by the deadlock sentence cannot be tolerated.

Second, regardless of whether the deadlock provision favors the defendant or the prosecution, it remains coercive. It introduces a measure of uncertainty and unreliability into the deliberative process. Thus, there is a substantial risk that the jury verdict may not reflect the true conscience of the jury. As a constitutional matter, such a result cannot be countenanced in a capital case (see Beck v Alabama, 447 US at 643).

“Any criminal defendant, and especially any capital defendant, being tried by a jury is entitled to the uncoerced verdict of that body” (Lowenfield v Phelps, 484 US 231, 241 [1988]). However, “when [capital jurors] impose a death sentence because life without parole, the sentence they deemed most appropriate, is not available, death is a forced choice” (Bowers and Steiner, Death by Default, 77 Tex L Rev at 611 n 22 [1999]). The result is an unreliable death verdict that does not accurately reflect the collective conscience of the community and seriously offends principles of due process and fundamental fairness. Thus, CPL 400.27 (10) is unconstitutional insofar as it provides that upon the failure of the sentencing jury to reach a unani*84mous verdict, the trial judge must impose a maximum sentence of life imprisonment and a minimum sentence of between 20 and 25 years.

Just as significant, the appropriate cure for this coercive statutory provision is not simply to withhold from the jury the fact that the trial judge could impose a more lenient sentence should the jury fail to reach a unanimous decision. New York has, with good reason, decided that the jury should be made aware of what would happen in the event of a deadlock. Providing such instruction is rooted in the significant interest of minimizing the risk of an arbitrary or capricious action. Indeed, without such an instruction, there is an intolerable likelihood that jurors would speculate as to what would happen to the defendant in the event of a deadlocked jury. Studies have shown that such speculation often tends to favor the imposition of the death sentence because jurors tend to underestimate the severity of the non-death sentencing alternative (see Eisenberg and Wells, Deadly Confusion, 79 Cornell L Rev at 4; Bowers and Steiner, Death by Default, 77 Tex L Rev at 705). Thus, telling jurors nothing about the sentence to be imposed upon a jury deadlock would risk the same unacceptable results that already exist with the instructions provided in the current statute.

Therefore, the only proper way to cure the infirmity of the provision is to strike the deadlock instruction itself. The only constitutionally proper sentence for a judge to impose upon the failure of the jury to decide between death and life imprisonment without parole is the lesser sentence considered by the jury—life imprisonment without parole. Such a result ensures that the jury’s verdict expresses the true conscience of the community and is not the result of uncertainty or coercion.

The Arbitrariness of New York’s Death Penalty Scheme

Prosecutorial Discretion

Defendant has also raised the issue of the arbitrariness of the death penalty statute. While in light of the disposition here, I do not come to a definitive conclusion on the issue of arbitrariness, the argument is a strong one and should be addressed. I write to highlight two aspects of this argument.

Defendant claims that New York’s death penalty scheme allows district attorneys to select capital defendants arbitrarily, inconsistently and discriminatorily, in violation of article I, § 5 (cruel and unusual punishment), § 6 (due process of law), and *85§ 11 (equal protection of the law) of the New York State Constitution and the Eighth and Fourteenth Amendments to the United States Constitution. Although raised in Harris, this Court has not previously addressed this issue (Harris, 98 NY2d at 497 n 23).

Traditionally, district attorneys have enjoyed great discretion in making charging decisions. That discretion need not be disturbed by this opinion. However, notice of intent to seek the death penalty pursuant to CPL 250.40 is not the same as a charging decision. By the time this notice is filed, the defendant has already been charged and indicted for first degree murder under Penal Law § 125.27.

CPL 250.40 does not provide any guidelines for when and if a district attorney should seek the death penalty. There is no provision under section 250.40 or any other portion of the CPL mandating that the district attorney set out the reason he or she has or has not sought death. The decision can be based on personal, irrelevant or even improper reasons without preserving a record for appeal.

The arbitrary way in which a defendant becomes exposed to the death penalty may be in violation of a defendant’s constitutional rights on both a state and federal level.

The People argue that the defendant failed to allege how the statute “as applied to him violates a constitutional norm” (People v Parker, 41 NY2d 21, 24 [1976]). The People call for a specific allegation. In their words, “[a]ppellant only insinuates that the use of prosecutorial discretion to file a notice to seek capital punishment pursuant to CPL 250.40 resulted in his arbitrary, capricious selection as a capital defendant; he does not point to evidence in the record to support an injury in fact.” Additionally, the People claim that the figures kept by the Office of Court Administration and the Capital Defender Office do not provide evidence of a per se injury.

According to the figures kept by the Office of Court Administration and the Capital Defender Office, between September 1, 1995 and December 31, 2001, grand juries indicted 376 defendants for first degree murder in New York. During this same period, district attorneys filed death notices in only 43 cases. That is only 11.4% of all first degree murder indictments, according to statistics maintained by the Capital Defender Office.

Without a reason being given by the People, the defendant is at a distinct disadvantage when it comes to making a specific *86challenge alleging improper reasoning by the district attorney. Yet the decision made by the district attorney has a direct impact on the defendant.

Under the statute as written, a defendant who is allegedly aggrieved by the People’s filing of the CPL 250.40 notice has no recourse and no remedy for improper reasons of the district attorney. Certainly no one would deny a defendant’s challenge if he or she could prove, for example, that the reasoning behind the notice was race-based. To deny a defendant’s challenge where the district attorney is allowed to conceal reasoning improperly stacks the deck against the defendant. Death is different and a defendant should be allowed to seek out and challenge the action of the district attorney. Failure to permit a challenge to the filing of the death notice stands in direct contradiction to the principles on which our legal system is founded.

The Arbitrariness of New York’s Death Penalty Statute Racial Discrimination

It is a sad and ugly badge of shame that the imposition of death in this country has been tainted with racial prejudice. It is no coincidence that when the Supreme Court in Furman v Georgia (408 US 238 [1972]) declared that the death penalty as then enacted constituted a cruel and unusual punishment in violation of the Constitution, the three defendants in the case were Black. The Court’s opinion is terse, but the opinions of the individual Justices make the case one of the longest in the United States Reports. The first of the opinions is that of Justice Douglas, who concluded that

“[t]hose who wrote the Eighth Amendment knew what price their forebears had paid for a system based, not on equal justice, but on discrimination. In those days the target was not the blacks or the poor, but the dissenters . . .

“[W]e know that the discretion of judges and juries in imposing the death penalty enables the penalty to be selectively applied feeding prejudices against the accused if he is poor and despised, and lacking political clout, or if he is a member of a suspect or unpopular minority, and saving those who by social position may be in a more protected position” (id. at 255).

*87Examining the raw numbers of those executed, Justice Marshall noted that “Negroes were executed far more often than whites in proportion to their percentage of the population.” (Id. at 364.) He then added that “[s]tudies indicate that while the higher rate of execution among Negroes is partially due to a higher rate of crime, there is evidence of racial discrimination” (id.).

Although concluding that racial discrimination in the cases before the Court was not established, Justice Stewart referred to Justices Douglas’s and Justice Marshall’s “demonstration] that, if any basis can be discerned for the selection of these few to be sentenced to die, it is the constitutionally impermissible basis of race” (id. at 310). Although Justice Brennan did not mention race, he concluded that the death penalty did not “comport[ ] with human dignity” because of, among other things, the “strong probability that it is inflicted arbitrarily” (id. at 305). Similarly, Justice White concluded that “the death penalty is exacted with great infrequency even for the most atrocious crimes and that there is no meaningful basis for distinguishing the few cases in which it is imposed from the many cases in which it is not.” (Id. at 313.)

Furman represents, to varying degrees, a certain distrust of jurors’ ability to consistently make rational and nondiscriminatory judgments as to which defendants are deserving of the ultimate punishment. Just one year before Furman, the Court had held, in McGautha v California (402 US 183, 207 [1971]), that “[i]n light of history, experience, and the present limitations of human knowledge, we find it quite impossible to say that committing to the untrammeled discretion of the jury the power to pronounce life or death in capital cases is offensive to anything in the Constitution.” To Justice Marshall, one of the dissenters in McGautha, the holding “was an open invitation to discrimination” (Furman, 408 US at 365).

In his dissenting opinion in Furman, Chief Justice Burger stressed “the assumption underlying the McGautha ruling is that juries ‘will act with due regard for the consequences of their decision’ ” (Furman, 408 US at 387-388). He added that jurors are the “conscience of the community” (id. at 388) and that “to assume from the mere fact of relative infrequency that only a random assortment of pariahs are sentenced to death, is to cast grave doubt on the basic integrity of our jury system” (id. at 388-389). Chief Justice Burger did seem to concede that the studies presented to the Court showed that jurors had *88discriminated in imposing death, but he brushed them aside by noting that they “cover[ed] periods when Negroes were systematically excluded from jury service and when racial segregation was the official policy in many States” (id. at 389 n 12).

Justice Powell, the author of another dissenting opinion, also conceded that the death penalty had been applied in a discriminatory manner in the past, but concluded that this did not justify invalidating death penalty sentences “in all cases in which sentences were handed out to members of the class discriminated against.” (Id. at 450.) Expressing a rather optimistic view, he stated that “[t]he possibility of racial bias in the trial and sentencing process has diminished in recent years” (id.).

Four years later, in Gregg v Georgia (428 US 153, 198 [1976]), the Court condoned Georgia’s revised sentencing procedures in death penalty cases because jurors were required to make specific findings “as to the circumstances of the crime or the character of the defendant.” In addition, the Georgia Supreme Court was required to compare a sentence to other similarly situated defendants to ensure that it was not disproportionate. Although the Court did not focus on race, it is fair to say that, in light of Furman, the purpose of limiting jury discretion was to ensure that it would not be exercised in a discriminatory manner.

One would expect then that after Furman and Gregg, there would be a change in the statistical landscape of those who are subjected to the death penalty and those who are not. It has not occurred. The jury discretion limits imposed by Georgia in 1972, after Furman, did not have an immediate profound effect on the collective racial consciousness of Georgia jurors. This was shown by an academic study spearheaded by Professor David C. Baldus which analyzed data on about 2,500 homicides committed in Georgia from 1973 to 1979. Taking into account 230 nonracial variables, Professor Baldus made the following findings, as noted by the Supreme Court in McCleskey v Kemp (481 US 279 [1987]):

“[T]he death penalty was assessed in 22% of the cases involving black defendants and white victims; 8% of the cases involving white defendants and white victims; 1% of the cases involving black defendants and black victims; and 3% of the cases *89involving white defendants and black victims. Similarly, Baldus found that prosecutors sought the death penalty in 70% of the cases involving black defendants and white victims; 32% of the cases involving white defendants and white victims; 15% of the cases involving black defendants and black victims; and 19% of the cases involving white defendants and black victims.” (.McCleskey at 286-287.)

The numbers seem to make clear that in a significant number of cases, the taking of a White life by a Black defendant deserved a harsher punishment, namely death, than any other type of case.

By a slim majority, the Supreme Court rejected McCleskey’s equal protection and Eighth Amendment challenges to the Georgia statute. Of course, McCleskey could not prove directly that the jurors who sentenced him to death did so because of his race. But he relied on the study to show an inference of discrimination in cases involving White victims and Black defendants. As in Furman, the argument was a salvo aimed at jury discretion, the heart of the criminal justice system. In an opinion by Justice Powell, the Supreme Court rejected the argument by finding as follows:

“Because discretion is essential to the criminal justice process, we would demand exceptionally clear proof before we would infer that the discretion has been abused. The unique nature of the decisions at issue in this case also counsels against adopting such an inference from the disparities indicated by the Baldus study. Accordingly, we hold that the Baldus study is clearly insufficient to support an inference that any of the decisionmakers in McCleskey’s case acted with discriminatory purpose” (481 US at 297).

Addressing jury and prosecutorial discretion in the context of the Eighth Amendment challenge, the Court found that:

“[w]here the discretion that is fundamental to our criminal process is involved, we decline to assume that what is unexplained is invidious. In fight of the safeguards designed to minimize racial bias in the process, the fundamental value of jury trial in our criminal justice system, and the benefits that discretion provides to criminal defendants, we hold that *90the Baldus study does not demonstrate a constitutionally significant risk of racial bias affecting the Georgia capital sentencing process.” (McCleskey at 313.)

The Court also stated that an acceptance of McCleskey’s claim would place into jeopardy cases not involving race and noncapital cases. For example, the Court reasoned, unattractive people could make an argument similar to that of McCleskey. Finally, the Court ended with the caution that McCleskey, or rather those who supported his arguments, should appeal to the legislature's.

It is noteworthy that Furman, Gregg, and McCleskey all dealt with Georgia defendants. The Court in Furman analyzed data from the entire nation, while Gregg and McCleskey were limited to Georgia defendants. While many other states do not share the same legacy of institutional racism common to Georgia (see McCleskey, 481 US at 328-330 [dissenting op of Brennan, J.]) and the rest of the South, when it comes to the application of the death penalty, Georgia and the other states with death penalty statutes are not so far apart.

That much can be gathered from a report, entitled “Death Penalty Sentencing: Research Indicates Pattern of Racial Disparities,” prepared by the United States General Accounting Office (GAO), an independent and nonpartisan agency, pursuant to the Anti-Drug Abuse Act of 1988 (Pub L 100-690).10 The report evaluated 28 carefully selected studies covering the pe*91riod of post -Furman to 1988. The report undertook this approach because of the availability of “both . . . sufficient quality and quantity [of research concerning discrimination] to warrant the evaluation synthesis approach” (at 2). The synthesis showed “a pattern of evidence indicating racial disparities in the charging, sentencing, and imposition of the death penalty.” (At 5.)

With respect to the race of the victim, the evidence was clear:

“In 82 percent of the studies, race of victim was found to influence the likelihood of being charged with capital murder or receiving the death penalty, i.e., those who murdered whites were found to be more likely to be sentenced to death than those who murdered blacks. This finding was remarkably consistent across data sets, states, data collection methods, and analytic techniques. . . .

“The evidence for the race of victim influence was stronger for the earlier stages of the judicial process (e.g., prosecutorial decision to charge defendant with a capital offense, decision to proceed to trial rather than plea bargain) than in later stages. This was because the earlier stages were comprised of larger samples allowing for more rigorous analyses. However, decisions made at every stage of the process necessarily affect an individual’s likelihood of being sentenced to death. . . .

“The analyses show that after controlling statistically for legally relevant variables and other factors thought to influence death penalty sentencing (e.g., region, jurisdiction), differences remain in the likelihood of receiving the death penalty based on race of victim” (at 5-6).

With respect to the race of the defendant, the findings were not as clear cut. The report determined that “more than half of the studies found that race of defendant influenced the likelihood of being charged with a capital crime or receiving the death penalty,” and that three fourths of these studies “found that black defendants were more likely to receive the death penalty.” (At 6.) The report cautioned, however, that the relationship between race and outcome was complex. For example, in rural areas Black defendants were more likely to receive death sentences, but in urban areas White defendants were *92more likely to receive death sentences. A few studies revealed “that the black defendant/white victim combination was the most likely to receive the death penalty.” (At 6.)

The GAO report is now 15 years old. Since its publication, a number of additional studies have been prepared. Many of these studies have concluded that race, whether of the victim or the defendant, or both, plays a role in the imposition of the death penalty.11 Others have found no statistically significant evidence that race plays a role in capital sentencing.12

*93The GAO report did not include New York. While the imposition of the death penalty in New York is older than the state itself,13 the last execution took place in 1963, several years before Furman. When the death penalty was in place, New York was often a leader in the number of executions carried out.14 Unlike Georgia, however, New York has often been a leader in enacting legislation prohibiting discrimination.15 Yet New York is no less deserving of a badge of shame for the role race has played in capital punishment. An analysis of those executed from 1888, the year capital punishment was centralized, to 1963, reveals the following:

“[Wjhile in the 1920s Blacks comprised only 1.9% of New York’s population, they represented 14% of those executed. Similarly, during the 1930s the Black population in New York accounted for only 3.28% of all New York residents, yet Black men constituted 17% of all persons executed during this time. ... By the 1960s, Black people comprised 80% of all capital defendants executed in New York while representing only 8.45% of the state’s population.”16

As to the makeup of the race of the victims, the statistics are not much different. In 90.4% of the cases resulting in executions in New York since 1890, the victims were White. On the other hand, “only 6.3% of the victims were Black and 3.3% . . . were from other racial groups.”17 Moreover, “[w]hile there have been ninety-six cases where a Black person was executed for killing a White person (14% of all cases), there has been only a single case in New York where a White person was executed for killing a Black person (0.1% of all cases).”18

*94The foregoing numbers, of course, involve a time before Fur-man. As the People point out, the current death penalty statute contains numerous safeguards against racial discrimination. During voir dire, parties may question jurors on “the possibility of racial bias” (CPL 270.16 [1]). Upon a conviction of first degree murder, but before the sentencing stage, the court must “determine whether any juror has a state of mind that is likely to preclude the juror from rendering an impartial decision based upon the evidence adduced during the proceeding” (CPL 400.27 [2]). On direct appeal, this Court must determine whether the sentence “was imposed under the influence of passion, prejudice, or any other arbitrary or legally impermissible factor including whether the imposition of the verdict or sentence was based upon the race of the defendant or a victim of the crime” (CPL 470.30 [3] [a]). Similarly, in determining whether the sentence was excessive or disproportionate in comparison to other cases, the Court must, if requested by the defendant, “review whether the sentence of death is excessive or disproportionate to the penalty imposed in similar cases by virtue of the race of the defendant or a victim of the crime for which the defendant was convicted.” (CPL 470.30 [3] [b].)

Yet, it is questionable whether racial prejudice can really be extirpated from capital proceedings. It is undeniable that racial prejudice is alive in many aspects of society, and no one expects it to die any time soon. There are numerous laws and agencies charged with stamping out discrimination from employment, housing and public accommodation. Yet, discrimination in those areas continues. The criminal justice system, of course, is supposed to be different. Prominently carved on top of the United States Supreme Court building are the powerful words “Equal Justice Under Law.” But we must acknowledge that equality under the law is an ideal the system strives for, not a reality.

Despite the well-established mass of jurisprudence, from Congress, the New York State Legislature, the United States Supreme Court, and this Court, prohibiting discrimination in the criminal justice system, as recently as 1988, a study conducted by the New York State Judicial Commission on Minorities, established by the then-Chief Judge of this Court, concluded

“that there are two justice systems at work in the courts of New York State, one for Whites, and a very different one for minorities and the poor.

*95“The system serving most minorities does not conform to our society’s notion of individualized justice, of hallowed halls, of impartial, reflective decision-making. Many minorities in our courts receive ‘basement justice’ in every sense of the phrase . . .

“[inequality, disparate treatment and injustice remain hallmarks of our state justice system.”

Moreover, a study by the New York State Division of Criminal Justice Services, entitled “Disparities in Processing Felony Arrests in New York State, 1990-1992,” concluded that similarly situated minorities and Whites were not treated equally, with the former held in jail on indictment and sentenced to incarceration more often, particularly in counties outside of New York City.

There is no doubt that conditions have improved, but try as it might, the justice system cannot immunize itself against racial prejudice. At the heart of the system are jurors with a great deal of discretion. It is probably the case that no safeguards are needed to ensure that many jurors exercise their discretion in a nondiscriminatory manner. But how many jurors who hold views determined largely by race, which can range from virulent to latent, will publicly acknowledge their views during voir dire? And can it be said that such jurors who sit in judgment of minority defendants will invariably put their views aside and decide the case on the law? Can the justice system realistically protect minority defendants against jurors who hold these views?19 In noncapital cases, the claims that sentencing determinations were race-influenced are generally not pursued or investigated. But should the danger be tolerated in cases where the penalty is death? Does the need to have the death penalty as a punishment outweigh the danger that it will be applied in a discriminatory manner? Are the efforts of the Legislature to insure that a death sentence will not be imposed because of race sufficient in light of the history of capital punishment in New York State?

Under McCleskey, the importance of jury discretion outweighs the danger that it could be used in a discriminatory manner. Al*96though the Supreme Court has said that death is different, in McCleskey it made no distinction between death and other penalties.20

Another ignored imperfection of the criminal justice system is a difference in the treatment of similarly situated defendants that results from prosecutorial discretion. Like jurors, prosecutors can use their “broad discretion” (United States v Goodwin, 457 US 368 [1982]) in a discriminatory manner. While it may not be difficult to imagine a juror whose vote for death may be tainted by views on race, it jars the mind to think that the race of a defendant or victim would influence a prosecutor’s decision to seek the death penalty. The difficulty is diminished, however, if determinative views on race, whether blatant or latent, exist in the collective consciousness of the community that elected the prosecutor. The Supreme Court has had to “repeatedly state[ ] that prosecutorial discretion cannot be exercised on the basis of race” (McCleskey, 481 US at 309 n 30).

Perhaps the most prominent example of a curb on prosecutorial discretion in order to protect minority defendants is Batson v Kentucky (476 US 79 [1986]) which held that prosecutors may not exclude jurors on account of race. Up until then, prosecutors had the unfettered discretion to peremptorily exclude jurors without having to offer any reasons. After Batson, upon a defendant’s showing of a prima facie case of discrimination, a prosecutor must offer race-neutral reasons for a peremptory challenge to a juror. If prosecutors—public officials—never used their discretion to discriminate, Batson would not be needed. The important interests served by peremptory challenges gave way to the need to ensure that prospective jurors are not excluded based on race.

While a prosecutor’s decision to seek the death penalty is limited by the nature of the crime, within the class of death-eligible homicides (first degree murders as defined in Penal Law § 125.27), prosecutors retain broad discretion. There is no safeguard to ensure that race plays no role in the decision to *97seek the death penalty. In addition, other factors, such as the availability of the resources to prosecute a death penalty case, may also play a role. When the death penalty is sought against a defendant in one county, but not against a similarly situated defendant in another county, the decision as to who gets the death penalty may be said to be arbitrary. The death penalty statute does not prohibit this kind of arbitrariness.

It must be acknowledged that so far there is no apparent indication that the race of the defendant has played a role in the determination to seek the death penalty in New York under the present statute. According to raw numbers compiled by the New York Capital Defender Office, from 1995 to 2001, district attorneys across the state filed death notices in 43 cases.21 Nineteen of those cases involved Black defendants, who were a subgroup of the larger group of about 194 Black defendants who were indicted for first degree murder. The number of White defendants against whom a death notice was filed was 18, although the larger group of eligible defendants was about 65. The race of the victims, however, is a different matter. Of the 43 cases in which prosecutors sought the death penalty, 21 (or 48.8%) of the victims were White, while only 11 (or 25.6%) were Black. Thus, of the victims of defendants charged with first degree murder whose race is known, the number of Black victims is half the number of White victims. Concededly, these are raw numbers, but they are startling nevertheless.

The numbers for geographic disparities are even more startling: “Although upstate counties experience approximately 19% of all homicides, they nonetheless account for 61% of all capital prosecutions.” (Id.)

The question in death penalty cases is whether the New York State Constitution should provide more protection against the risk of racial discrimination than McCleskey. Historically, the New York Constitution has at times provided greater guarantees for individuals than those provided by the Federal Constitution (People v Scott, 79 NY2d 474, 491 [1992]; People v Bora, 83 NY2d 531, 535 [1994]; People v Harris, 77 NY2d 434, 437-441 [1991]; People v Torres, 74 NY2d 224, 226 [1989] People v Griminger, 71 NY2d 635, 637-639 [1988]; People v P.J. Video, Inc., 68 NY2d 296, 304 n 4 [1986]). As this Court stated in People v Barber (289 NY 378, 384 [1943]):

*98“[I]n determining the scope and effect of the guarantees of fundamental rights of the individual in the Constitution of the State of New York, this court is bound to exercise its independent judgment and is not bound by a decision of the Supreme Court of the United States limiting the scope of similar guarantees in the Constitution of the United States.” (See also People v Alvarez, 70 NY2d 375, 379 [1987]; Sharrock v Dell Buick-Cadillac, 45 NY2d 152, 159 [1978].)

New York must remain committed to the ideal of equality under the law. A defendant should not have to show direct evidence of discrimination, as effectively required by McCleskey. While the administration of the death penalty in New York was tainted with racial and ethnic bias, and minorities have not always received equal justice, New York has been particularly steadfast in its condemnation of racial discrimination. Both the Federal and State Constitutions prohibit the denial of equal protection. Article I, § 11 of the New York Constitution provides as follows:

“No person shall be denied the equal protection of the laws of this state or any subdivision thereof. No person shall, because of race, color, creed or religion, be subjected to any discrimination in his or her civil rights by any other person or by any firm, corporation, or institution, or by the state or any agency or subdivision of the state.”

Under the New York State Constitution, the death penalty cannot be imposed if the risk that race played a role in its imposition is significant.

Thus, I conclude, along with Judge Ciparick, that the deadlock jury instruction required by CPL 400.27 (10) is coercive and unconstitutional. While I make no determination on the issue of the arbitrariness of the death penalty, I conclude that before a death sentence can be imposed, the issue must be addressed and resolved.

I concur with Point I of the majority’s opinion. But I believe the jury’s determination of James Cahill’s guilt of murder in the first degree, based on two statutory aggravating factors—felony murder (see Penal Law § 125.27 [1] [a] [vii]) and witness elimination (see Penal Law § 125.27 [1] [a] [v])—should be up*99held. In my view, a majority of this Court now rewrites the first-degree felony-murder statute in a manner contrary to the plain language of the provision and clear public policy enacted into law by our elected representatives. Equally disturbing, the majority concludes that the witness elimination murder charge was supported by sufficient proof, yet reverses that conviction without explaining why the jury, which had the ability to assess the credibility of witnesses, could not have found beyond a reasonable doubt that—regardless of whatever else motivated defendant—he intended to eliminate Jill Cahill as a potential witness. In addition, I cannot agree with the majority that the trial court erred in life and death qualification of the prospective jurors and fully concur with Judge Read’s jury selection analysis. I therefore dissent from so much of the majority decision that overturns defendant’s first-degree murder convictions.

I. Penal Law § 125.27 (1) (a) (vii): First-Degree Felony Murder

The majority’s holding that the evidence against defendant was legally insufficient to establish his guilt of felony murder in the first degree rests upon two separate, yet intertwined legal theories: an application of the so-called “merger” doctrine (although the majority does not use that term) and an implicit interpretation of the “in furtherance of’ requirement in the first-degree murder statute. I cannot accept either analysis. There are certain well-settled principles of statutory construction— construing laws in accord with the plain meaning of the language chosen by the Legislature, avoiding irrational results and adopting interpretations that do not undermine the constitutionality of a statute—that this Court adheres to in order to effectuate the legislative intent of a law. These rules should apply in all cases, including capital cases.

The fundamental flaw in the majority’s analysis is its disregard of these rules and of the clear and unequivocal language of the felony-murder aggravator. The majority declines to interpret the statutory provision in accord with its plain language on the basis that it “would have to stretch the statute’s meaning” (majority op at 71). I strongly disagree.

The carefully selected language of the first-degree felony-murder statute should inform the outcome in this case. Penal Law § 125.27 (1) (a) (vii) provides, in relevant part:

“[a] person is guilty of murder in the first degree when . . . [w]ith intent to cause the death of another person, he causes the death of such person or *100of a third person; and . . . the victim was killed while the defendant was in the course of committing or attempting to commit and in furtherance of robbery, burglary in the first degree or second degree, kidnapping in the first degree, arson in the first degree or second degree, rape in the first degree, sodomy in the first degree, sexual abuse in the first degree, aggravated sexual abuse in the first degree or escape in the first degree, or in the course of and furtherance of immediate flight after committing or attempting to commit any such crime or in the course of and furtherance of immediate flight after attempting to commit the crime of murder in the second degree” (emphasis added).

Our function as judges is to interpret this law. “The governing rule of statutory construction is that courts are obliged to interpret a statute to effectuate the intent of the Legislature” (People v Finnegan, 85 NY2d 53, 58 [1995]). The “ ‘clearest indicator of legislative intent’ ” is the statute itself (People v Robinson, 95 NY2d 179, 182 [2000], quoting Majewski v Broadalbin-Perth Cent. School Dist., 91 NY2d 577, 583 [1998]). If the language chosen by the State Legislature is clear and unambiguous, and “involves no absurdity or contradiction, there is no room for construction and courts have no right to add to or take away from that meaning” (Tompkins v Hunter, 149 NY 117, 123 [1896]; see People ex rel. Harris v Sullivan, 74 NY2d 305, 309 [1989]). When this doctrine is violated, a court impermissibly encroaches upon the legislative and executive domains and thereby violates the foundation of the separation of powers doctrine (see People v Finnegan, 85 NY2d at 58).

The statute encompasses first- or second-degree burglary charges, both of which may be premised upon the intent to commit any crime {see Penal Law §§ 140.25, 140.30). We have recognized that in enacting the burglary and felony-murder statutes, the Legislature “intended that the definition be ‘satisfied if the intruder’s intent, existing at the time of the unlawful entry or remaining, is to commit any crime’ ” {People v Miller, 32 NY2d 157, 161 [1973] [emphasis in original], quoting Denzer and McQuillan, Practice Commentary, McKinney’s Cons Laws of NY, Book 39, at 355). “[A]ny crime” obviously includes murder.

The majority asserts that “the Legislature did not expressly sweep within Penal Law § 125.27 all killings in which the *101murderer unlawfully entered the victim’s home” (majority op at 65). This is true in one respect—only intentional murders are included, not “all killings” (majority op at 65). But by specifying first- and second-degree burglary in the felony-murder provision, the Legislature certainly intended to include within the ambit of the statute intentional murderers who unlawfully enter another person’s dwelling with the intent to kill (see Penal Law § 125.27 [1] [a] [vii]).

This is consistent with additional language in the statute— the requirement that the murder must occur “in the course of . . . and in furtherance of’ the burglary. To satisfy these elements, the People must prove that the homicide not only occurred during “the course of’ a burglary, but that the killing was logically related to the unlawful intrusion as well. The latter requirement has long been part of New York law. As we explained in People v Wood (8 NY2d 48 [I960]), interpreting a predecessor felony-murder provision:

“a felony murder embraces not any killing incidentally coincident with the felony . . . but only those committed by one of the criminals in the attempted execution of the unlawful end. Although the homicide itself need not be within the common design . . . the act which results in death must be in furtherance of the unlawful purpose” (id. at 51 [emphasis added]).

Significantly, the. Legislature echoed the terminology we articulated in Wood when it added the phrase “in furtherance of’ to the felony-murder statute in 1965. This phraseology was then carried forward in the felony-murder provision of the 1995 death penalty act. Because the Legislature chose to use a phrase first adopted by this Court when it enacted the felony-murder provision in 1965 and then used the identical language in the 1995 capital felony-murder provision, ■ it is reasonable to conclude that the Legislature intended to incorporate the Wood rule into the death penalty statute.

Consistent with Wood, courts have interpreted the “in furtherance of’ element as requiring a logical nexus between a murder and a felony (see People v Lewis, 111 Misc 2d 682, 686 [Sup Ct, NY County 1981] [citing Wood for the proposition that “ ‘in furtherance’ places a relation requirement between the felony and the homicide . . . the nexus must be one of logic or plan”]; see also State v Young, 191 Conn 636, 642, 469 A2d *1021189, 1193 [1983] [citing Wood: “(w)e agree with the New York courts that the phrase ‘in furtherance of’ was intended to impose the requirement of a relationship between the underlying felony and the homicide beyond that of mere causation in fact, similar to the concept of proximate cause in the law of torts”]).1 Thus, by requiring that a rational relationship be established between the perpetrator’s actions and the homicide, the Legislature expressed its intent to exclude deaths that are merely coincidental to the occurrence of a felonious event or completely unrelated to a defendant’s purpose or conduct (see People v Hernandez, 82 NY2d 309, 317 [1993]).

Wood therefore established that a murder is logically related to and thereby furthers another crime when, among other possibilities, it is designed to help achieve the purpose of the other crime. In this case, the “unlawful end” (People v Wood, 8 NY2d at 51) that prompted defendant to commit the burglary was the murder of his wife. Thus, under the Wood analysis, the murder of Jill Cahill furthered the unlawful entry into her hospital room.

The majority entirely avoids addressing the “in furtherance of’ requirement and adopts a rule that is inconsistent with this statutory language. Indeed, although the majority asserts that its analysis is “more faithful to the Legislature’s language and design” and is consistent with “the statutory plan” (majority op at 72), it never actually interprets the language of the statute. Nor does it cite any legislative history in support of its conclusions (nothing in the legislative history supports either the conclusion the majority has reached or the untraditional review it engages in to get there). Instead, the majority concludes that if the purpose of a predicate felony is to commit murder, there is no “independent criminal objective” for the felony; the burglary necessarily furthers the homicide but the burglary may not be furthered by the killing. Thus, in the *103absence of an independent criminal purpose, the majority reasons that the underlying felony is an integral component of the killing and is thereby subsumed within the homicide, precluding a capital felony-murder conviction.

This is, in effect, the “merger” rule that many other states— including New York (see People v Miller, 32 NY2d 157 [1973])— have rejected.2 Clearly, “there is no logic or reason to preclude a felony murder charge from being based upon a burglary charge that, in turn, is premised upon either an intent to assault or an intent to murder” (People v Lewis, 791 P2d 1152, 1154 [Colo Ct App 1989]; see also State v Tillman, 750 P2d 546, 581 [Utah 1987] [Stewart, Assoc. Ch. J., concurring]). The California Supreme Court—the court that invented this analytical theory—has characterized as “artificial” and “anomalous” the notion that “a felon who acts with a purpose other than specifically to inflict injury upon someone—for example, with the intent to sell narcotics for financial gain . . .—is subject to greater criminal liability for an act resulting in death than a person who actually intends to injure the person of the victim” (People v Hansen, 9 Cal 4th 300, 315, 885 P2d 1022, 1030 [1994]).

The majority contends that the holding in this case does not undermine this Court’s precedent in People v Miller. If this is true, there are now two different “merger” rules in this state— one for first-degree felony murder and one for second-degree felony murder—despite the fact that the language of those statutes is identical in all relevant respects. Before today, those statutes had the same meaning. But where the first-degree murder provision currently reads “burglary in the first degree or *104second, degree,” in light of the majority’s decision, the statute will now be subject to an additional, judicially-imposed restriction—“except where the underlying intent is to kill”—a particularly strange limitation for a capital murder statute.

The majority adopts this new limitation because it concludes that the statute as written by the Legislature does “not narrow the class of those eligible for the death penalty, but. . . widen[s] it” (majority op at 65). The majority’s conclusion stems from a mischaracterization of the Supreme Court’s narrowing principle and a failure to give weight to the constitutional concept that narrowing must rationally distinguish between defendants who are and are not subject to the possibility of capital punishment.

“[A] capital sentencing scheme must ‘genuinely narrow the class of persons eligible for the death penalty and must reasonably justify the imposition of a more severe sentence on the defendant compared to others found guilty of murder’ ” (Lowenfield v Phelps, 484 US 231, 244 [1988], quoting Zant v Stephens, 462 US 862, 877 [1983]). In other words, states must make “ ‘the death penalty ... an available sentencing option—even potentially—for a smaller class of murders’ ” (Lowenfield v Phelps, 484 US at 245-246, quoting Jurek v Texas, 428 US 262, 271 [1976]) and must make “a principled distinction between those who deserve the death penalty and those who do not” (Lewis v Jeffers, 497 US 764, 776 [1990]; see Gregg v Georgia, 428 US 153, 198 [1976]).

In New York, the Legislature specified certain aggravating factors that a jury must find proven beyond a reasonable doubt before it may consider whether a sentence of death or life without parole is warranted (see Penal Law § 125.27 [1] [a] [i]-[xiii]; CPL 400.27 [11] [a]). Because the first-degree felony-murder statute applies to only intentional murders, it necessarily excludes unintentional felony murders. Thus, the statute restricts death eligibility to limited categories of homicides and identifies a subclass of felony murder for which capital punishment and life imprisonment are authorized.3

The Legislature also drew a distinction for aggravating factor purposes between burglary in the first and second degrees, and *105burglary in the third degree. Burglary in the third degree covers the unlawful entry of any building, even if it is not a dwelling, and does not require that the perpetrator be armed with a weapon (see Penal Law § 140.20). By excluding burglary in the third degree from the class of felonies underlying the capital felony-murder provision, the Legislature indicated its desire to punish more severely illegal entry into dwellings or entry while armed with a dangerous weapon. Thus, in crafting the capital statute, the Legislature fulfilled its narrowing responsibility by incorporating our State’s historic preservation of the sanctity of a home (see e.g. People v Miller, 32 NY2d at 160). Because crimes committed in dwellings can rapidly escalate to violence, the Legislature adopted a reasonable demarcation among possible burglary-murder scenarios. As we determined in People v Harris, it is rational for the Legislature to consider “the inherent potential for violence and physical injury” in making distinctions between capital-eligible and noncapital-eligible offenses (98 NY2d 452, 477 [2002]). By including first- and second-degree burglary as predicate felonies but excluding third-degree burglary, the Legislature did just that: there was a “ ‘meaningful basis for distinguishing the few cases in which (the death penalty) is imposed from the many cases in which it is not’ ” (id. at 476, quoting Gregg v Georgia, 428 US at 198).

The majority overlooks the requirement that narrowing, while necessary, must be rational. Under the majority’s interpretation, an intruder who unlawfully enters a dwelling with the intent to steal—a television, for instance—and intentionally kills an occupant in furtherance of that burglary may be indicted on a capital murder charge. In sharp contrast, an intruder who unlawfully enters a dwelling with the intent to murder—someone like defendant who obtained a toxic chemical known for its ability to kill, violated an order of protection by surreptitiously entering his wife’s hospital room, wore a disguise to avoid detection and, in a face-to-face encounter, poured cyanide down her throat, inducing a particularly painful death—confronts a parole-eligible prison sentence. It is inconceivable that the Legislature intended such an irrational delineation between capital- and noncapital-eligible defendants.

In adopting its “independent criminal objective” rule the majority addresses only felony murder based on burglary. But it is *106possible that the majority could extend its analysis to other crimes included in the first-degree felony-murder statute, such as kidnapping, arson and robbery, thereby creating additional “merger” distinctions contrary to the clear language of New York’s first-degree felony-murder statute. Did the Legislature intend that a murderer who kidnaps to extort money may face a capital charge, but a murderer who kidnaps in order to transport a victim to a secluded location for the killing falls outside the statute?4 Or that a defendant who sets fire to a house in order to kill the occupants receives a parole-eligible prison sentence, but a defendant who sets a fire in order to obtain insurance money or to drive the occupants outside and murder them in person can be indicted for first-degree murder?5

I therefore question whether the majority’s version of the merger analysis “rationally distinguish[es]” (Spaziano v Florida, 468 US 447, 460 [1984]) between death-eligible defendants on a “principled basis” (People v Harris, 98 NY2d at 476). If additional violent felonies such as kidnapping, arson and robbery were to be excised from the statute along with the class of burglaries the majority excises, few qualifying felony-murder offenses would remain. The majority’s rationale thus creates a potential conflict with our reasoning in People v Harris (98 NY2d at 476-477) where, in response to a constitutional challenge to the felony-murder provision, we held that New York’s death penalty statute was not irrationally underinclusive. There, we noted that “[t]he Legislature made the assessment that the predicate felonies for first-degree felony murder should be those that are potentially the most violent and involve a substantial risk of physical injury,” and concluded that the “decision to exclude felonies of a similar grade, but lacking the inherent potential for violence and physical injury, was . . . rational” (id. at 476, 477). In Harris, we recognized that it was the Legislature’s prerogative to determine, as a matter of public policy, that a certain crime—including, I would argue, the illegal entry of a dwelling with the intent to kill and while armed with a deadly weapon—is a “violent” offense that “involves a substantial risk of physical injury” justifying its inclusion as an aggravating factor for jury consideration (id. at 476). I would adhere to that determination.

*107Even applying the majority’s new merger rule, I believe there was sufficient proof that defendant committed felony murder in the first degree—he did, in fact, have an independent criminal objective. The majority concedes that its analysis does not apply to a defendant who has more than one purpose for committing the underlying felony (majority op at 65). Thus, a felony-murder conviction will be sustained where the purpose of the predicate felony is not exclusively to kill but also to commit some other offense, such as concealing evidence of another crime (see People v Mendoza, 24 Cal 4th 130, 183-184, 6 P3d 150, 183 [2000]).

In this case, the majority holds that the evidence was legally sufficient to prove that one of the reasons defendant killed his wife was to eliminate her as a witness to the April assault (majority op at 59)—a point with which I concur. That determination necessarily means that the evidence was sufficient to prove several, lesser included felony offenses, including tampering with a witness in the first and second degrees (Penal Law § 215.13 [1]; § 215.12 [1]), and intimidating a victim or witness in the first and second degrees (Penal Law § 215.17 [1]; § 215.16 [1]). Hence, defendant intended to commit the crime of tampering with a witness in the first degree by “intentionally caus[ing] serious physical injury to” Jill Cahill “for the purpose of . . . preventing . . . the giving of testimony in a criminal proceeding by [her]” (Penal Law § 215.13 [1]). The majority does not dispute that the People offered sufficient evidence that, at the moment defendant unlawfully entered Jill Cahill’s hospital room, he possessed the intent to commit crimes that require “a felonious intent independent of the murder” (majority op at 64). This is precisely the “double crime—murder ‘plus’ ” that the majority claims is required for first-degree felony murder (majority op at 64).

For all of these reasons, I cannot join the majority’s discussion of the first-degree felony-murder provision. The majority’s analysis is undeniably contrary to the language in the statute, produces irrational results and fails to give proper weight to the principled narrowing choices made by the Legislature. In my view, it is evident that the Legislature intended to apply the felony-murder rule to egregious cases like this. Even if I disagreed with that legislative determination, my personal belief as to whether this defendant is an appropriate candidate for first-degree murder would be irrelevant for one very basic reason—I am a judge, not a legislator.

*108II. Penal Law § 125.27 (1) (a) (v): Witness Elimination Murder

The majority sets aside defendant’s conviction of witness elimination murder because it believes that the jurors, who had the opportunity to see and hear the witnesses and assess their credibility, failed to properly evaluate the evidence presented at trial. In doing so, the majority substitutes its view of the evidence, based on the cold record, for that of the jurors who sat in judgment of defendant. Because I cannot agree that the jury failed to properly weigh the evidence in this domestic violence case, I would uphold defendant’s conviction under Penal Law § 125.27 (1) (a) (v).

According to the majority, a critical factor in assessing the weight of the evidence in this case is whether “defendant thought he could avoid an assault conviction by murdering Jill” (majority op at 62). The witness elimination statute, however, is unconcerned with the probability of a conviction. The plain language of the statute requires only proof that the murder was committed to “prevent[ ] the intended victim’s testimony” (Penal Law § 125.27 [1] [a] [v]).6 Our task is to determine whether the jury’s assessment of this statutory requirement was against the weight of the evidence.

People v Bleakley (69 NY2d 490 [1987]) provides the standard for weight of the evidence review:

“If based on all the credible evidence a different finding would not have been unreasonable, then the appellate court must, like the trier of fact below, ‘weigh the relative probative force of conflicting testimony and the relative strength of conflicting inferences that may be drawn from the testimony’ . . . . If it appears that the trier of fact has failed to give the evidence the weight it should be accorded, then the appellate court may set aside the verdict” (id. at 495).

This is not an open invitation for a court to substitute its judgment for that of the jury. Bleakley commands that “[g]reat deference” be accorded to the jury’s resolution of factual issues (id.), a principle the majority fails to even mention. Deference is an integral ingredient of the weight of the evidence analysis *109because it is the “fact-finder[ ]” that has the “opportunity to view the witnesses, hear the testimony and observe demeanor” (id.).

Thus, although an appellate court engaging in weight of the evidence review has been described as fulfilling the role of a “thirteenth juror” (see Tibbs v Florida, 451 US 31, 42 [1982]; People v Rayam, 94 NY2d 557, 560 [2000]), this shorthand is not a literal description of the court’s appropriate analytical process. To the contrary, there are “[w]ithout question . . . differences between what the jury does and what the appellate court does in weighing evidence” (People v Bleakley, 69 NY2d at 495), which is apparent given that a jury’s ability to assess credibility and weigh competing inferences is far superior to that of an appellate court working with only a “printed record” (People v Cohen, 223 NY 406, 423 [1918]). Jurors use all of their senses to evaluate testimony and evidence presented at trial—only they “are able and entitled to assess, at first hand, the credibility and reliability of the witnesses” because they have the opportunity to actually see, hear and feel the evidence in person (People v De Tore, 34 NY2d 199, 206 [1974]). A healthy respect for the jury’s unanimous assessment of the evidence is therefore a necessary component of a proper weight of the evidence review.

Applying the Bleakley standard in its entirety, I cannot concur with the majority’s conclusion that the jury verdict on the witness elimination count was against the weight of the evidence. The People presented compelling evidence that defendant intentionally killed his wife to prevent her from challenging his version of “the truth” about the circumstances leading to the assault on April 21, 1998. As the majority recognizes, the evidence overwhelmingly established that defendant intended to kill Jill when he assaulted her with a baseball bat, crushing her skull. But defendant offered a different explanation at the time of the assault and at trial, claiming that Jill attacked him first and he initially responded in self-defense.7 This is the factual *110rendition of the assault that defendant wanted the world to hear and believe.

Yet, Jill Cahill survived the first attack. Over the next six months, she fought back from the brink of death to the point that she was able to respond and speak. The testimony of medical personnel and her family members clearly relayed to the jury that Jill was beginning to communicate in October 1998. What would she have testified to had she been given the chance—that she did not attack defendant with a knife and initiate the assault? We will never know. But we do know that, once defendant learned Jill could speak—after his mother visited Jill in the hospital in October 1998—he immediately executed his plan to permanently prevent her from attempting to contradict his account of “the truth.”

In overturning the jury verdict on this count, the majority overlooks a number of significant facts and fails to identify the important connections between them that the jury may have found determinative. During the May 11, 1998 Family Court hearing, an assistant district attorney announced the People’s intention to indict defendant for the April assault and advised defendant that his children were potential witnesses. The same day, defendant began to search the Internet for Web sites relating to “cyanide.” Approximately one week later, defendant was present for another Family Court proceeding and learned that his children would undergo psychological evaluation. These impending interviews gave defendant reason to believe that his *111children would be questioned about the assault on their mother and the effect that it had on them. This had the potential to elicit additional information regarding the events surrounding the assault. After the Family Court appearance, defendant returned home and searched the Internet for information on “ordering potassium cyanide.”

The timing of pertinent events—when defendant actually ordered the cyanide, when he obtained it, and when he used the poison—forcefully establishes that defendant’s actions were inextricably related to the prosecution of the assault charges. Between the date defendant was arraigned on an indictment charging him with three counts of first-degree assault and his next court appearance, defense counsel and the prosecutor had “numerous conversations” about the case, including plea negotiations. Defendant’s decision to order cyanide on the forged letterhead of a local manufacturing company on July 8 occurred only a day before the People’s plea offer was officially placed on the record. The offer, under which defendant would have had to serve at least 10 years in prison, was not accepted by defendant, making a trial increasingly certain.

Most importantly, although defendant intercepted the shipment of cyanide on July 17 by approaching the delivery person and taking possession of the package from the supplier he had contacted, he did not use the poison to kill his wife until three months later. Why? The evidence presented to the jury pointed to a logical explanation: defendant did not execute his homicidal plan until he was able to confirm that Jill was conscious and had recovered sufficiently to be a “threat.”

Defendant’s mother supplied defendant (and the jury) with the information that provided the impetus for him to carry out Jill’s murder.8 During a visit with Jill Cahill in the hospital on October 5, Mrs. Cahill observed a “conversation” between Jill and two nurses. Significantly, Jill actually spoke to the nurses. Defendant’s mother told her son about these “physical observations” of Jill about two weeks later. Defendant’s response to his mother demonstrates that he appreciated the significance of this information and knew that Jill was regaining her ability to communicate: “ T hope that when this is all concluded that she can tell the truth about what led to the break-up’ of the[ ] marriage.” For the first time since the April assault, defendant *112became aware that Jill could speak and, therefore, posed a threat as a possible witness against him at trial. Defendant immediately put his murderous plan into action.

Within days of learning of Jill’s improved condition and ability to speak, defendant—disguised as a janitor—was found in Jill’s hospital room by a nurse’s assistant. He quickly departed without incident but returned one week later, again in disguise. This time he was able to bring his plan to fruition—he murdered Jill by poisoning her with cyanide. Defendant had successfully eliminated the primary witness to the April baseball bat attack.9 A hearing to determine whether defendant’s confessions could be used at the assault trial was only days away.

Giving due deference to the jurors’ assessment of the evidence presented, this sequence of events amounted to compelling evidence supporting the conclusion the jury unanimously reached—defendant killed Jill Cahill to prevent her from communicating with the authorities and testifying at the impending assault trial. The trigger for the murder was defendant’s awareness of the fact that Jill had regained her ability to speak, and with that ability, Jill may have refuted defendant’s account of the events on the night of the assault.

Indeed, no other rationale accounts for defendant’s timing in carrying out his plan for murder. The majority posits that defendant was motivated “to poison his wife because their marriage and family life were being destroyed, not because he wanted to kill a witness to the assault case” (majority op at 59-GO). Defendant may have been motivated to kill because of the deterioration of his family life but the proof does not support the conclusion that this was his only motivation. And what other evidence was presented to the jury to explain why defendant waited three months after he acquired cyanide to carry out his plan to poison Jill? If it was only hatred and family disruption that impelled defendant’s murderous intent, why didn’t he attempt to kill Jill in August after his parents surrendered custody of the children to Jill’s parents, further removing him from his children? In the absence of any plausible reason *113explaining why defendant waited until mid-October to attempt to administer the cyanide—notably, just a day or two after his mother informed him of Jill’s improved condition and shortly before the pretrial Huntley hearing scheduled for November 2—the jury was entitled to conclude that defendant’s motivation was to prevent Jill from communicating her version of the facts underlying the assault. Hence, the jury could reasonably infer that, whatever else may have driven defendant to take Jill’s life, he intended to eliminate Jill as a witness. Therefore, giving “[g]reat deference” to the jury’s findings (People v Bleakley, 69 NY2d at 495), I conclude that defendant’s conviction of witness elimination murder was not against the weight of the evidence.

The majority’s reasons for disregarding the jury’s resolution of these issues focus on facts that were not known to defendant and therefore shed no light on his motivation. For example, there is no basis in the record to suppose that defendant was aware about the fact the majority finds “[m]ost compelling[ ]”— that Jill had never been interviewed by the police or a prosecutor (majority op at 60)—nor was there any proof that defendant knew whether Jill had or had not retained her memory of the events surrounding the assault (majority op at 60). These facts are meaningless to the witness elimination analysis because defendant’s motives for acting could not have been influenced by that which he did not know. What is pertinent is that the jury heard evidence permitting the inference that defendant believed Jill posed a risk and he killed her to prevent her from testifying.

Similarly unavailing is the majority’s decision to discount the importance of the evidence provided by defendant’s mother. The majority claims that her “testimony does not even suggest, let alone reveal, what defendant knew about Jill’s condition and speaking ability” (majority op at 61). That assertion is not borne out in the record. Defendant’s mother testified that she told defendant about her observations of Jill’s physical condition. This testimony came moments after she described to the jury what she had observed—that Jill spoke to the nurses. It is irrelevant that “we do not know what Jill said” (majority op at 61)—what is significant is that Jill could communicate and defendant was aware of this critical fact.

The majority also places undue emphasis on its conclusion that “there is scant basis to believe that defendant thought he could avoid an assault conviction by murdering Jill” (majority *114op at 62). The witness elimination statute does not require proof that defendant intended to avoid a criminal conviction. Rather, the statute requires only that a defendant seeks to “prevent[ ] the intended victim’s testimony” (Penal Law § 125.27 [1] [a] [v]). The facts of this case highlight the difference between the plain language of the statute and the majority’s interpretation. It is true that defendant admitted that he beat Jill with a baseball bat during the April assault, although this was not a full confession since defendant maintained he did so in self-defense. The prosecution clearly had a “powerful case without Jill’s testimony” (majority op at 61)—defen-dant confessed to wounding himself and acknowledged striking Jill after she had been incapacitated, thereby severely undermining if not “foreclosing any plausible claim of self-defense” (majority op at 61). But in order to convict defendant of witness elimination murder, the jury was not required to find that defendant hoped (either reasonably or unreasonably) that by killing Jill he could avoid a criminal prosecution. The jury had only to conclude that defendant intended to prevent her testimony.

In this case, the jury could logically infer from all the evidence that defendant was driven by a desire to convince everyone, including his children, that Jill initiated the incident; that she attacked him with a knife; and that, in the course of protecting himself, defendant simply “snapped.” This is the story defendant told the police in April 1998 and this is the explanation that was ultimately presented to the jury. While this account would not avoid a criminal prosecution and might not engender jury leniency sufficient to produce an acquittal— and it did not in this case—defendant did succeed in preventing the jury from hearing the whole story. With Jill’s death, defendant knew that there would be no one left to directly dispute his version of the events on the night of the assault. Defendant committed the ultimate act of domestic violence and, by killing the prime witness, he also impeded the truth-seeking function of a trial, which is precisely why witness elimination murder is among the crimes most strongly condemned in New York State. In my view, if we faithfully apply the weight of the evidence standard of review in this case and give deference to the jury’s unanimous finding, there is no basis to set aside defendant’s conviction of witness elimination murder in the first degree (see Penal Law § 125.27 [1] [a] [v]).

III. Conclusion

The extreme disagreement that envelops the members of this Court stems from three general ideas and differing philosophies *115about how we should exercise the power that the State Constitution vests in this institution. By failing to respect the plain language of the first-degree felony-murder provision, the majority has substituted its “wisdom” and public policy choices for those of the Legislature when it enacted the statute. The majority also rejects the jury’s assessment of the probative value of the evidence and the reasonable inferences derived therefrom. Finally, the majority sets aside the conclusions reached by the Trial Judge, who had the ability to actually see and hear the prospective jurors discuss their ability to fairly consider possible sentences, reinterpreting the statements made during the life and death qualification process from the cold record. In this case, I would respect, rather than second guess, the life/death qualification assessments of the Trial Judge, the factual findings of the jury and the public policy choice made by the Legislature.

Accordingly, I would affirm both first-degree murder convictions. Due to the majority’s holding in this case, it is unnecessary for me to address the penalty phase issues raised in defendant’s brief and I therefore express no view on those issues or the propriety of the sentence imposed.

Read, J. (concurring in part and dissenting in part). In order to reverse defendant’s conviction for murder in the first degree under Penal Law § 125.27 (1) (a) (vii) (capital felony murder, here predicated on burglary in the second degree), the majority rewrites clear statutory language and disregards our 30-year-old precedent in People v Miller (32 NY2d 157 [1973]). The majority also discovers jury selection error where none reasonably exists. I respectfully dissent.

I. The Burglary Aggravating Factor

The People charged defendant with first-degree felony murder, for which the underlying felony was second-degree burglary. As relevant here, first-degree felony murder occurs when a defendant, “[w]ith intent to cause the death of another person, . . . causes the death of such person . . . ; and . . . the victim was killed while the defendant was in the course of committing . . . and in furtherance of . . . burglary in the . . . second degree” (Penal Law § 125.27 [1] [a] [vii]). In turn, and again as relevant here, “[a] person is guilty of burglary in the second degree when he knowingly enters or remains unlawfully in a building with intent to commit a crime therein, and . . . [t]he building is a dwelling” (Penal Law § 140.25 [2]).

*116The People contend that defendant committed burglary in the second degree because he “knowingly enter [ed]” or “remain[ed] unlawfully” in his wife’s “dwelling,” her hospital room (it was after visiting hours; an order of protection forbade defendant from seeing his wife, although he contests notice of it) “with intent to commit a crime therein”; namely, intentional murder. Further, “inten[ding] to cause [her] death,” and “caus[ing] [her] death,” defendant killed his wife while “in the course of committing” and “in furtherance of” the burglary when he forced potassium cyanide down her throat as she lay virtually helpless on her hospital bed. Hence, the People argue, defendant committed first-degree felony murder, a capital offense in this case. The jury agreed.

Defendant contests his conviction for first-degree felony murder on two grounds. First, he argues that the phrase “in furtherance of’ in Penal Law § 125.27 (1) (a) (vii) requires a murder to “advance” or “facilitate” the commission of the predicate felony. In the circumstances of this case, he contends, it is logically impossible for the murder to “advance” or “facilitate” the burglary because the two crimes share the identical criminal objective, intentional murder.

Second, principally relying on People v Green (27 Cal 3d 1, 609 P2d 468 [1980]), defendant asserts that absent an independent criminal objective, the burglary “merged” into the murder; therefore, the burglary cannot serve as an aggravating factor without contravening the Eighth Amendment’s mandate that any aggravating factor must genuinely narrow the class of defendants eligible for a capital sentence.

The majority skirts defendant’s “in furtherance of’ argument and, instead, grounds its opinion on the felony murder merger (or “merger”) doctrine, holding that the burglary aggravating factor may not be predicated solely on the intent to commit intentional murder. The majority reaches this conclusion principally (it would appear) because of an inability to believe that the Legislature meant what it actually said {see majority op at 65).

Further, the majority labors mightily to slip the strictures of our precedent by belittling Miller as a traditional, not a first-degree, felony murder case, which dealt with an assaultive, not *117murderous intent.1 Then the majority treats the holding in Miller as virtual dictum since the two concurring judges—although obviously unsuccessfully—promoted a narrower basis for reaching the result. Finally, the majority relies on Arkansas and Delaware cases; and, while pointing out the Legislature’s unquestioned resolve to enact a death penalty statute satisfying the requirements of the Eighth Amendment as enunciated in Gregg v Georgia (428 US 153 [1976]), extensively discusses Green, a case in which the California Supreme Court judicially narrowed that state’s circa-1970’s capital felony murder “special circumstance” to conform with Gregg.

A. Statutory Interpretation

First and foremost, this case presents us with a pure issue of statutory interpretation. As a general rule, “[t]he primary consideration of courts in interpreting a statute is to ‘ascertain and give effect to the intention of the Legislature’ ” (Riley v County of Broome, 95 NY2d 455, 463 [2000], quoting McKinney’s Cons Laws of NY, Book 1, Statutes § 92 [a], at 177). Moreover, the plain meaning of the statutory text is the best evidence of legislative intent and, in fact, the only authoritative basis for interpretation {Riley v County of Broome, 95 NY2d at 463 [“Of course, the words of the statute are the best evidence of the Legislature’s intent. As a general rule, unambiguous language of a statute is alone determinative”]).

Reiterating the statutory language relevant here, a defendant commits first-degree felony murder when “[w]ith intent to cause the death of another person, he causes the death of such person . . . ; and . . . the victim was killed while the defendant was in the course of committing . . . and in furtherance of . . . burglary in the . . . second degree” (Penal Law § 125.27 [1] [a] [vii]). Further, a defendant “is guilty of burglary in the second degree when he knowingly enters or remains unlawfully in a building with intent to commit a crime therein, and . . . [t]he building is a dwelling” (Penal Law § 140.25 [2]).

Obviously, nothing in the wording of the first-degree felony murder statute in any way limits the type of crime that a burglar must intend to commit or, put another way, excludes *118intentional murder as the sole criminal objective of a second-degree burglary. Belatedly, the second-degree burglary statute manifestly fails to limit the type of crime that a burglar must intend to commit or to exclude intentional murder from those crimes.

Moreover, we have—until today—always construed the second-degree burglary statute as broadly as its unambiguous language demonstrates the Legislature intended (see e.g. People v Mackey, 49 NY2d 274, 279 [1980] [holding that a defendant charged with burglary in the second degree is not entitled to a bill of particulars specifying which crime he intended to commit within the building since “the Revised Penal Law’s definition of burglary is satisfied if the intruder’s intent, existing at the time of the unlawful entry or remaining, is to commit any crime” (internal quotation marks omitted; emphasis added)]). We must presume that the Legislature was aware of our decisions interpreting the second-degree burglary statute as broadly as its words naturally allow (see e.g. People v Robinson, 95 NY2d 179, 184 [2000]).

Further, when “words hav[e] a precise and well settled legal meaning in the jurisprudence of the state[, they] are to be understood in such sense when used in statutes, unless a different meaning is plainly indicated” (McKinney’s Cons Laws of NY, Book 1, Statutes § 233, at 396 [emphasis added]). For the “different meaning” to which the majority subscribes to have been “plainly intended” here, the Legislature would have to have (as it easily could have) drafted the first-degree felony murder provision to exclude burglary predicated on intentional murder from its ambit. That the Legislature did not make this particular drafting choice is telling in light of the numerous ways in which the first-degree felony murder provision was, in fact, limited or, in the language of Gregg, “narrowed” by the Legislature.

That is, the Legislature created the first-degree felony murder provision by borrowing the language of the second-degree felony murder provision, with its well settled legal meanings, and then limiting or narrowing this language in several key respects. Most importantly, first-degree felony murder, unlike second-degree felony murder, requires an intentional killing. First-degree felony murder also provides for more limited acces*119serial liability;2 and includes a predicate felony not found in the second-degree felony murder provision; namely, an intentional killing while in immediate flight from attempting second-degree murder.

Significantly, the Legislature also cut back on the underlying felonies upon which first-degree (as opposed to second-degree) felony murder may be predicated, including, as is especially relevant for present purposes, the types of burglary. That is, second-degree felony murder may be predicated on any degree of “burglary,” while first-degree felony murder can rest only on first- or, as in this case, second-degree burglary. By excluding third-degree burglary (which references entry into a building), the Legislature deliberately limited first-degree felony murder to those burglaries that expressly involve invasions of a dwelling, the place where a citizen reasonably expects to be most secure and is, in fact, most vulnerable to serious injury or death when cornered by an intruder bent on murder, as happened here.

In short, the Legislature, while carefully limiting the scope of the first-degree felony murder provision with respect to the types of burglary covered, otherwise did not see fit to exclude a burglary committed with intentional murder as its criminal objective. While the majority may find this hard to accept, the choice is the Legislature’s—not ours—to make.

B. Merger and the Miller Case

The felony murder merger doctrine—a judicial gloss developed in the context of traditional (that is, unintentional) felony murder—bars felony murder when the predicate felony is subsumed in the homicide. With the advent of the revised Penal Law, we cast off merger in Miller. By holding that first-degree felony murder predicated on burglary must have a criminal objective independent of intentional murder, the majority has now revived (and, in fact, expanded) merger in our jurisprudence in ways that the Legislature could never have anticipated.

*1201. The Scope of the Merger Doctrine in New York

We developed and discussed our own distinctive brand of merger in several key cases decided in the late nineteenth and early twentieth centuries, when felony murder, even if unintentional, was punishable by death. In Buel v People (78 NY 492 [1879]), the defendant, who apparently strangled his victim while raping her, contended that the felony (the rape) merged with the homicide since violence was a requisite element of the rape; therefore, the defendant argued, the rape was not a separate and distinct offense from the killing and so the judge should not have instructed the jury on felony murder.3 We disagreed: “While force and violence constitute an important element of the crime of rape, they do not constitute the entire body of that offence. The unlawful or carnal knowledge is the essence of that crime; and without this, no matter what degree of force or violence may be employed, rape is not established” (id. at 497 [emphasis added]).

The defendant in People v Huter (184 NY 237 [1906]), after fleeing from premises in which he had committed a burglary, shot and killed a police officer who was pursuing him. When the defendant was convicted of felony murder predicated on burglary,4 we were asked to review. Concluding that the defendant had ceased to be engaged in the commission of a burglary at the time of the murder, we reversed the judgment and conviction. We then turned our attention to the “much mooted” question of whether the assault in which the defendant was engaged at the time of the killing merged into the homicide, a question that would “doubtless” come up upon retrial (id. at 242-243).

*121Citing Buel, we stated that “[i]n order ... to constitute [felony murder], . . . the violence may constitute a part of the homicide, yet the other elements constituting the felony in which [the defendant] is engaged must be so distinct from that of the homicide as not to be an ingredient of the homicide, indictable therewith or convictable thereunder” (Huter, 184 NY at 244 [emphasis added]).

In People v Wagner (245 NY 143 [1927]), the proprietor of a boarding house detected a lodger acting suspiciously. When she confronted the lodger outside the bathroom on the boarding house’s second floor, he suddenly struck her in the head with a “blackjack.” The proprietor cried out for help, and another occupant of the boarding house rushed to her assistance. The rescuer struck the defendant a blow, and the proprietor broke free and ran down the stairs and out the front door, shouting for help. She heard shots and ran back into the boarding house and upstairs, where she saw the defendant and her rescuer lying side by side on the floor. When the proprietor took off her shoe and began to beat the defendant about the head with it, he arose, struck her in the face and made his escape. Her rescuer died of gunshot wounds that the defendant had inflicted on him.

The trial court, among other things,5 charged the jury that defendant was guilty of felony murder6 if he killed the victim while engaged in a felonious assault upon either the victim or the proprietor. Citing Huter, we held that the trial court erred in charging the jury regarding the assault on the victim, which was an “ingredient” of the resulting homicide and therefore not an independent crime. The error was harmless, though, because the homicide was committed when the felonious assault on the proprietor was still in progress, and this assault was an independent felony.

Our decision in Wagner also relied on People v Patini (208 NY 176 [1913]), another case in which a Good Samaritan intervened in an assault on someone else only to lose his own life {see also People v Giblin, 115 NY 196 [1889] [wife killed when she rushed *122into store to assist her husband, who was scuffling with gunman who had shot him in altercation over payment for purchases]). Patini, which involved a felonious assault upon the brother of the victim, was grounded at least in part on our interpretation of the word “otherwise” in section 1044 (2) of the former Penal Law to “include [ ] a felony upon a person other than the one killed” (Patini, 208 NY at 178; see also People v Miles, 143 NY 383 [1894]).

In People v Moran (246 NY 100 [1927]), the defendant, after he and his companions were halted by two policemen, drew a revolver and fired two shots at one officer and a third at the other, killing them both. The trial judge charged the jury upon the single theory of felony murder, which we held to be error in light of the merger doctrine. That is, if, as was charged, the defendant shot the second officer while trying to escape after having shot the first officer, then the first felony—the assault upon the first officer—was over, while the second felony—a felonious assault upon the second officer—was not independent of the homicide, “[fit was the homicide itself’ (id. at 103).

With Chief Judge Cardozo speaking for a unanimous Court, we explained the purpose of the merger doctrine:

“Homicide is murder in the first degree when perpetrated with a deliberate and premeditated design to kill, or, without such design, while engaged in the commission of a felony. To make the quality of the intent indifferent, it is not enough to show that the homicide was felonious, or that there was a felonious assault which culminated in homicide.

Such a holding would mean that every homicide, not justifiable or excusable, would occur in the commission of a felony, with the result that intent to kill and deliberation and premeditation would never be essential. The felony that eliminates the quality of the intent must be one that is independent of the homicide and of the assault merged therein, as, e.g., robbery or larceny or burglary or rape” (id. at 102 [emphasis added and citations omitted]).

In other words, the merger doctrine prevented prosecutors from using felony murder to sidestep proof of a culpable mental state in the majority of cases.

Next, the jury in People v La Marca (3 NY2d 452 [1957]) convicted the defendant, who kidnapped a one-month-old infant later found dead in a wooded area, of kidnapping and of having *123killed the infant while engaged in the commission of the kidnapping. The defendant urged that the felony murder count should have been dismissed because the acts constituting the kidnapping merged into the homicide.

In rejecting this argument, we first noted that while other jurisdictions limited felony murder to specified felonies, our Legislature had chosen a different path, decreeing a homicide to be a capital offense if perpetrated in the commission or attempted commission of any felony. As a result, “[o]ne necessary qualification . . . engrafted” by us on New York’s felony murder rule was the merger doctrine; i.e., “that the underlying felony must be independent of the homicide” because “otherwise, every homicide, not justifiable or excusable, would occur in the commission of a felony—namely, the assault which ended in death—with the result that premeditation, deliberation and intent to kill would never have to be established” (id. at 465). We cited Huter, Wagner and Moran for this proposition.

We reasoned in La Marca that the underlying felony of kidnapping was independent of the murder. Just as in Buel, while force and violence were an important element of the kidnapping, its “essence” was something altogether separate, i.e., “[a] taking or detaining of the infant with intent to keep or conceal ... or to extort . . . money” (La Marca, 3 NY2d at 466).

To sum up, the merger doctrine developed as a limitation that we “engrafted” on felony murder in order to maintain distinctions between culpable mental states. This was necessary because the felony murder statutes at the time encompassed assault as a predicate crime, potentially converting every homicide into capital felony murder since assault is an “ingredient” of every homicide. We consistently declined to expand merger to crimes in addition to assault since the “essence” of crimes such as rape, kidnapping, robbery, larceny or burglary is independent of the homicide.7 In several cases, we sustained a conviction for felony murder notwithstanding merger when the underlying assault was committed on someone other than the murder victim. *124This result, however, was premised at least in part on language (the word “otherwise”) appearing in both section 183 of the Penal Code and its successor provision, section 1044 (2) of the former Penal Law. Thus, the status of our merger jurisprudence in 1973 when we decided Miller.

2. People v Miller

Miller barged uninvited into Fennell’s apartment to assault him with a spray can (apparently to disable him with choking gas) and a butcher knife, and wound up killing Fennell’s roommate, Aleem, who came to Fennell’s aid. In other words, the facts were essentially indistinguishable from those of our Good Samaritan cases, such as Wagner, Patini and Giblin (see also People v Luscomb, 292 NY 390 [1944]). The statute under which Miller was prosecuted was, however, quite different as was the People’s theory.

Miller was indicted for and convicted of, among other things, felony murder predicated on burglary to commit an assault. The trial court set aside this conviction on the ground that the People had not established burglary. The trial court, the majority in the Appellate Division and Miller all subscribed to the view that assault, the target crime of the burglary, was actually the underlying felony; therefore, the conviction could not stand because the felony murder statute, Penal Law § 125.25 (3), did not list assault as a predicate crime. Accordingly, we phrased the issue upon appeal as “whether a burglary based upon the crime of assault can properly serve as the predicate for a felony-murder conviction” (Miller, 32 NY2d at 158).

As we correctly saw it, Miller was urging us to “extend” the merger doctrine so that neither the assault against Fennell nor the assault against Aleem would qualify as the target crime of the burglary. This we refused to do, noting that the considerations prompting us to “announce” the merger doctrine in the first place were no longer germane because the Legislature had amended the Penal Law to remove the “fundamental defect” that merger was designed to cure (id. at 159).

Specifically, under the old felony murder statutes (section 1044 [2] of the former Penal Law and its predecessors) any felony, including assault, could be the predicate for murder. Thus, “[s]ince, a fortiori, every homicide, not excusable or justifiable, occurs during the commission of assault, every homicide would constitute a felony murder” (Miller, 32 NY2d at 160). The Legislature remedied this “fundamental defect” in the *125revised Penal Law (Penal Law § 125.25 [3]), however, by enumerating specific felonies, not including assault and “all involving violence or substantial risk of physical injury,” which were the predicate crimes for felony murder (id.). “The legislative purpose for this limitation was ‘to exclude from felony murder, cases of accidental or not reasonably foreseeable fatality occurring in an unlikely manner in the course of a non-violent felony’ ” (id. [citation omitted]).

We further described—in detail—the Legislature’s obvious motivation or purpose for including burglary in the list of enumerated predicate crimes for felony murder:

“[P]ersons within domiciles are in greater peril from those entering the domicile with criminal intent, than persons on the street who are being subjected to the same criminal intent. Thus, the burglary statutes prescribe greater punishment for a criminal act committed within the domicile than for the same act committed on the street. Where, as here, the criminal act underlying the burglary is an assault with a dangerous weapon, the likelihood that the assault will culminate in a homicide is significantly increased by the situs of the assault. When the assault takes place within the domicile, the victim may be more likely to resist the assault; the victim is also less likely to be able to avoid the consequences of the assault, since [the] paths of retreat and escape may be barred or severely restricted by furniture, walls and other obstructions incidental to buildings. Further, it is also more likely that when the assault occurs in the victim’s domicile, there will be present family or close friends who will come to the victim’s aid and be killed” (id. at 160-161).

Since the Legislature “did not exclude from the definition of burglary, a burglary based upon the intent to assault, but intended that the definition be ‘satisfied if the intruder’s intent, existing at the time of the unlawful entry or remaining, is to commit any crime’ ” (id. at 161 [citation omitted; emphasis in original]), we concluded that any burglary, even one premised on an intent to assault, could support a felony murder conviction.

Tellingly, in Miller we also expressly “reject[ed]” the “reasoning and holding” of People v Wilson (1 Cal 3d 431, 462 P2d 22 [1969]) (Miller, 32 NY2d at 161 n 3). Along with its California *126progeny, Wilson anchors the majority’s holding that capital felony murder predicated on burglary must have a criminal objective independent of intentional murder.

In Wilson, the defendant, armed with a shotgun, broke a glass window in the outside door of his estranged wife’s apartment building and entered her apartment, where she and three men were present. Initially, the defendant shot two of the men, killing one. He placed the muzzle of his shotgun against the door of the bathroom into which his wife had fled in a futile search for refuge, and shot the door off its hinges. The defendant’s wife was found dead in the bathtub, having been shot twice in the chest at close range.

The defendant was eventually found guilty of first-degree felony murder of his wife based on the crime of burglary, and second-degree felony murder of the other victim, based on assault with a deadly weapon.8 The California Supreme Court reversed both convictions with merger as the rationale.

As to the second-degree felony murder count, this result followed directly from traditional merger principles, adopted by the court in People v Ireland (70 Cal 2d 522, 539-540, 450 P2d 580, 589-590 [1969] [second-degree felony murder improperly predicated on crime of assault with a deadly weapon because this felony was an integral part of the homicide]). The court reversed the first-degree murder count because the burglary was not independent of the assault, however, and thereby expanded merger. Moreover, in support of this expansion, the court purported to follow New York merger doctrine.

Specifically, the court noted that in Ireland (70 Cal 2d at 540, 450 P2d at 591) it had overruled, in relevant part, People v Hamilton (55 Cal 2d 881, 901, 362 P2d 473, 485 [1961]) and People v Talbot (64 Cal 2d 691, 703, 414 P2d 633, 641 [1966]), both of which had “indicated that . . . the New York ‘merger’ doctrine ... is not to be applied in this jurisdiction [i.e., California] to preclude a first degree felony-murder instruction based *127upon a burglary as to which the intended felony is the homicide itself or an offense included therein” (Wilson, 1 Cal 3d at 439-440, 462 P2d at 27-28 [emphasis added]).9 This statement implies that our traditional merger doctrine in fact would have precluded felony murder based upon a burglary as to which homicide or assault was the target crime.

As previously discussed, however, we never broadened merger beyond assault proper. Indeed, Chief Judge Cardozo specifically cited burglary as an example of a felony “independent” of the homicide and therefore not merged in it for purposes of felony murder (see Moran, 246 NY at 102; see also La Marca, 3 NY2d at 466). And in Miller, of course, we declared the merger doctrine nugatory anyway in light of revised Penal Law § 125.25 (3). We also specifically (and directly contrary to Wilson) held that a burglary premised on an intent to assault would support a felony murder conviction. No wonder, then, that we made an emphatic point in Miller to disassociate ourselves from Wilson.

Of course, today the majority completely reverses field, embracing Wilson and its progeny, which unquestionably sup*128port its holding,10 and backing away furiously from Miller, which unquestionably does not. Accordingly, the majority advances a number of superficial distinctions designed to distance itself from Miller.

First, the majority contends, Miller obviously does not “govern” because there we interpreted Penal Law § 125.25 (3) whereas this case involves Penal Law § 125.27 (1) (a) (vii) (majority op at 65). But, of course, the operative language in these two provisions is identical. The Legislature concededly “borrow[ed]” (majority op at 67) from the language of Penal Law § 125.25 (3) to fashion a narrower version of it as Penal Law § 125.27 (1) (a) (vii). We must presume that the Legislature was aware of Miller when adopting the first-degree murder statute (see e.g. People v Robinson, 95 NY2d at 184). Thus, the only reasonable inference to draw is that the Legislature intended (and expected) a burglary underlying a charge of first-degree felony murder to encompass “any crime” (People v Miller, 32 NY2d at 161 [internal quotation marks omitted; emphasis in original]; see also Mackey, 49 NY2d at 279); and “any crime” manifestly includes intentional murder.

Further, the Legislature would have had no reason to expect (and indeed would have had every reason not to expect) us to find Wilson and California-style merger11 suddenly persuasive, and to revive and expand the merger doctrine so as to restrict the scope of burglary under the first-degree felony murder statute. Moreover, given the purpose of the merger doctrine—to preserve the distinctions among the culpable mental states of murder—the majority’s revival and expansion of it for first-degree felony murder, where intent must always be proven, but *129not for second-degree felony murder, which may be unintentional, makes no logical sense.12

Next, the majority opines, Miller is different because it dealt with assault, not murder, and so we would, in fact, be “extending] ” Miller if we were to decide that a burglary based upon the crime of intentional murder can properly serve as the predicate for first-degree (or, for that matter, second-degree)13 felony murder (majority op at 66 [emphasis omitted]). The dispute in Miller was, indeed, whether a burglary based upon an assault would support a conviction for felony murder; however, this was because, it was argued, the assault merged in the murder and so the assault and the murder became one and the same. We rejected merger, and concluded that the criminal objective of the predicate burglary could be any crime, which surely includes intentional murder as well as assault. I fail to see how we would be “extending]” Miller by reading the decision to mean what it manifestly says.

Finally, the majority seems to prefer the reasoning of the concurrence in Miller, and to accept defendant’s argument that the majority opinion in that case swept too broadly, or at least more broadly than was necessary to support the result (majority op at 66). To state the obvious, though, Miller’s holdings are to be found in the majority opinion, not the concurrence; the reasoning of the concurrence was not accepted by the Miller majority. Further, the concurrence’s rationale is problematic.

*130Specifically, the concurrence would have “predicated [the result] on the narrower ground that even under the old law, when the doctrine of merger was in full bloom, a conviction of felony murder was sustained where the underlying offense was assault if the assault was committed on a person other than the one killed” (Miller, 32 NY2d at 161 [Jones, J., concurring]). Wagner and the other cases cited for this unexceptionable proposition,14 however, were all decided under the “old law” (section 1044 [2] of the former Penal Law), and so their holdings were grounded in statutory language superseded in the revised Penal Law (see discussion at 120-124). Additionally, these were all assault cases, not burglary-predicated-on-assault cases, as Miller was. It bears repeating that “even . . . when the doctrine of merger was in full bloom” (Miller, 32 NY2d at 161), we consistently declined to extend merger beyond assault to other crimes (e.g., rape, kidnapping, larceny, robbery or burglary). Although force and violence were (or might be) an element of these other crimes, their “essence” (e.g., for burglary, the unlawful entering or remaining in a building) was independent of homicide.

C. Foreign Cases

Lacking support in New York statutes and precedent, the majority finally retreats to foreign cases to support its holding that capital felony murder must have a criminal objective independent of the intentional murder (majority op at 69). As the majority acknowledges, two of these sister-state cases—Williams v State (818 A2d 906 [Del 2002]) and Parker v State (292 Ark 421, 731 SW2d 756 [1987])—are nothing more than interpretations of Delaware’s and Arkansas’ rather different capital statutes. Additionally, Williams and Parker both deal with capital felony murder statutes that are closer to our traditional (second-degree) felony murder statute than to our first-degree felony murder statute.

*131Specifically, these statutes do not require an intentional killing, and instead appear designed “to reduce the disproportionate number of accidental homicides which occur during the commission of the enumerated predicate felonies” (Miller, 32 NY2d at 161). Indeed, Delaware requires a “reckless[ ]” murder (see Del Code Ann, tit 11, § 636 [a] [2]), while in Arkansas, the killing had to have been “under circumstances manifesting extreme indifference to the value of human life” (Parker, 292 Ark at 425, 731 SW2d at 758).

It is, thus, quite surprising that the majority considers Parker and Williams so convincing. After all, the majority views our own traditional felony murder jurisprudence in Miller as irrelevant mainly because that crime, like the crimes addressed in Williams and Parker, does not require an intentional murder (majority op at 65). By contrast, the Arkansas Supreme Court itself considered Miller entirely relevant, but rejected the Miller distinction between a murder committed in an occupied dwelling as opposed to out of doors. The court concluded that what might make sense in New York did not automatically make sense in Arkansas because of its different statutes (Parker, 292 Ark at 426-427, 731 SW2d at 759).

The majority cannot have it both ways. If Miller is of no value here, as the majority contends, then Williams and Parker are at least equally useless. Alternatively, if it is legitimate to rely on foreign cases dealing with statutes similar to the statute underlying Miller, then Miller must be relevant here.

Ultimately, the majority rests its conclusion principally upon the opinion of the California Supreme Court in People v Green (27 Cal 3d 1, 609 P2d 468, supra) (majority op at 70-71). Our decisions in Miller and People v Harris (98 NY2d 452 [2002]), however, undermine any reliance on Green.

In Green, the California Supreme Court justified the need for an “independent objective” by reference to Eighth Amendment narrowing concepts. In Harris, however, we expressly rejected a claim that the first-degree felony murder statute did not comport with the narrowing requirements of the Eighth Amendment (Harris, 98 NY2d at 475-477). There is no reason for us to shrink from that conclusion here.

We noted in Miller that it was “apparent that the Legislature, in including burglary as one of the enumerated felonies as a basis for felony murder, recognized that persons within domiciles are in greater peril from those entering the domicile *132with criminal intent, than persons on the street who are being subjected to the same criminal intent” (Miller, 32 NY2d at 160). The concern for preventing violence in domiciles that led the Legislature to include burglary in the second-degree felony murder statute informs the first-degree statute under consideration here.

As discussed earlier, the traditional, second-degree statute is predicated upon “burglary” of any degree (see Penal Law § 125.25 [3]). But when the Legislature “borrowed” from that statute, it slightly narrowed intentional felony murder so as to include only first-degree or second-degree burglary (see Penal Law § 125.27 [1] [a] [vii]). Significantly, only those two degrees of burglary expressly encompass crimes committed in a “dwelling” (see Penal Law § 140.25 [2]; § 140.30); third-degree burglary references only a “building” (see Penal Law § 140.20). Thus, the Legislature clearly structured the first-degree felony murder statute to deter and punish intentional killings in a “dwelling.”

The legislative goal of protecting the citizens of our state from burglars with a murderous intent nullifies any Eighth Amendment fears. As just discussed, we took note in Miller of the extremely rational legislative “recognition]” of the “greater peril” faced by persons in their homes (Miller, 32 NY2d at 160). Because of that “greater peril,” the Legislature ensured that “the burglary statutes prescribe greater punishment for a criminal act committed within the domicile than for the same act committed on the street” (Miller, 32 NY2d at 160).

The “greater punishment” for committing the “same act” in a dwelling rather than “on the street” obviates the need for the burglary here to have had an objective independent of murder. The Legislature quite rationally concluded that a burglar—like defendant—who fulfills a specific murderous intent is deserving of “greater punishment” than a defendant who commits the “same act ... on the street” (id.). Thus, providing the death penalty for the crime committed here in no way runs afoul of the Eighth Amendment (Harris, 98 NY2d at 476). Indeed, it is not merely an intent to kill, but rather burglarizing a dwelling to fulfill that murderous intent which differentiates (and correspondingly narrows) this murder from its lesser included offense of second-degree intentional murder.

*133Simply put, the Eighth Amendment concerns expressed by the California Supreme Court in Green are not present here.15 The Eighth Amendment does not require that a burglary underlying a felony murder (especially an intentional murder) be predicated upon an independent criminal objective (see e.g. State v Tillman, 750 P2d 546, 571 [Utah 1987]; Smith v State, 499 So 2d 750, 754 [Miss 1986]; see also State v Dann, 205 Ariz 557, 568 n 7, 74 P3d 231, 242 n 7 [2003] [implying that merger doctrine would not preclude capital felony murder conviction for burglary premised on intent to kill]).

II. Jury Selection

I cannot agree that the trial court erred in its rulings under CPL 270.20 (1) (f) regarding prospective jurors Nos. 23 and 855. The trial court’s rulings, viewed in light of the entire voir dire for these prospective jurors, comported with the standards for life/death qualification that we just enunciated in Harris (98 NY2d at 485-487). The majority spotlights selected snippets from a lengthy record. We were not there; the Trial Judge was. As we recognized long ago when construing section 377 of the former Code of Criminal Procedure (the statutory predecessor of CPL 270.20 [1] [f]), “it was for the court to say, from the whole examination of the juror, including his appearance and demeanor, whether he was fit and competent to perform fairly and impartially” (People v Carolin, 115 NY 658, 659 [1889]).16

*134In the end, this Trial Judge presided over eminently fair, error-free jury selection, in which he correctly applied subdivision (1) (f) in an even-handed manner. Indeed, the trial court and the parties went to extraordinary lengths to impanel a fair and impartial jury: jury selection involved over 900 prospective jurors, took more than two months, and included a comprehensive written questionnaire followed, as necessary, by extensive individual and then group voir dire.17 We should not upset this handiwork because we might have asked prospective jurors more or different or better questions during voir dire or might have otherwise created a fuller record; or because some of the prospective jurors strike us as more or less qualified than the Trial Judge reasonably assessed them to be in the immediacy of a trial setting. Jury selection is not a mechanical exercise. In a capital case, in particular, jury selection is necessarily exhaustive and emotional. We do not want to discourage trial courts from independently assessing a prospective juror’s credibility and sincerity; we do not want to force trial courts to rely on rote assurances of impartiality whether or not they “ring true.” Nor do we want to convey a sense that jury selection errors are inevitable in every capital case because of our willingness to parse the record so as to make them so.

*135A. Prospective Juror No. 23

Under the standards for life/death qualification explained in Harris, the trial court properly concluded that prospective juror No. 23 had been life qualified. Certainly, prospective juror No. 23 never expressly “rule[d] out voting against” the death penalty (majority op at 44). Indeed, prospective juror No. 23 repeatedly and consistently assured the parties and the trial court that he could consider both death and life. Responding to the prosecutor, for example, he unequivocally stated that he would “fulfill my [sentencing] responsibility[ ] fairly and impartially,” and that he could consider both sentences “[e]qually.” Prospective juror No. 23 then made similar assurances during a colloquy with defense counsel.

The majority’s conclusion that prospective juror No. 23 had not been life qualified rests entirely upon his voluntary disclosure of a personal near experience with domestic violence. By zeroing in on that lone section of the lengthy voir dire, the majority deems irrelevant prospective juror No. 23’s prior voir dire assurances that he could, indeed, consider a life verdict. Significantly, in Harris we rejected an essentially indistinguishable challenge to the life qualification of prospective juror No. 233.

Prospective juror No. 233 expressed “strong skepticism” regarding the defendant’s proposed mitigating factor of child abuse (Harris, 98 NY2d at 485). The defendant subsequently challenged prospective juror No. 233 for cause under subdivision (1) (f), and the trial court rejected the challenge (id. at 486) . In holding that the trial court ruled properly, we determined that subdivision (1) (f) required a “connection” between a proposed juror’s death penalty views “and its effect on that juror’s consideration of the appropriate sentence” (id. at 486, 487) . We noted that the “[defendant made no correlation between Juror No. 233’s views on child abuse and her views on the death penalty or her ability to exercise sentencing discretion conferred by statute” (id. at 487).

While defendant here may have established some “correlation” between prospective juror No. 23’s views and his sentencing ability, defendant never demonstrated the link that we viewed as “[c]ritical[ ]” in Harris (id. at 486). Specifically, as was the case with prospective juror No. 233 in Harris, prospective juror No. 23 “was never asked if ... [ ]he could still vote to give defendant life without parole” (id.).

*136No doubt, defendant took prospective juror No. 23 right up to the brink of that question. The voir dire ended with prospective juror No. 23’s acknowledgment that, as defendant phrased it, this “life experience ... is causing you some concern as to whether you’d be able to consider both penalties fairly”; and “[t]hat would cause you some problems with being able to consider life without parole as the appropriate penalty other than death” (emphasis added). But having “some concern” or “some problems” does not translate into an affirmative statement that prospective juror No. 23 could not “vote to give defendant life without parole” (Harris, 98 NY2d at 486). Thus, here, as in Harris, “defendant has failed to make a claim that his for-cause challenge against Juror No. 23 [ ] pursuant to CPL 270.20 (1) (f) should have been granted” (id. at 487).

The majority unfairly takes the trial court to task for supposedly failing to pose additional questions {see majority op at 46). But, as the majority recognizes, “[t]he proponent of the challenge must demonstrate through questioning the juror’s inability to fulfill his or her oath” (majority op at 46, citing Morgan, 504 US at 733). Thus, defendant bore the burden of making the necessary record to support his subdivision (1) (f) challenge. The trial court heard nothing that merited continued questioning or granting defendant’s challenge. Again, the facts of the voir dire are instructive.

The record reveals someone struggling with his emotions, informing the court and parties honestly that he had every intention of following his obligations if selected. Indeed, prospective juror No. 23 expressly stated that “it’s possible that [my life experience] could come to mind. As much as I’m sitting here and I’m trying to promise everybody that I’d be fair, I’m still human, I’m still a man, that could come to mind. That’s being honest.” Thus, the record—far from showing a juror who was not life qualified—bespeaks someone caught up in the assuredly emotional context of a capital case. In fact, the trial court specifically cited prospective juror No. 23’s candid and heartfelt disclosures as a basis for rejecting defendant’s challenge. Following group voir dire, defendant renewed his challenge. The trial court adhered to its original ruling, noting that prospective juror No. 23 had actually been crying when he related his life experience.

Prospective juror No. 23’s emotional reflections and the depth of his feelings did not constitute grounds for his exclusion. Construing subdivision (1) (f) in Harris, we concluded that “[w]e *137understand the word ‘preclude’ to relate to the ability of a prospective juror to perform his or her duty to act impartially and to follow a court’s instructions in accordance with the law and not as a measurement of the strength or sincerity of an individual’s feelings on the death penalty” (Harris, 98 NY2d at 484). As we also noted in Harris, under controlling Supreme Court precedent “ ‘[n] either nervousness, emotional involvement, nor inability to deny or confirm any effect whatsoever is equivalent to an unwillingness or an inability on the part of the jurors to follow the court’s instructions and obey their oaths, regardless of their feelings about the death penalty’ ” (id. at 483, quoting Adams v Texas, 448 US 38, 50 [1980]).

If anyone is to blame for any lack of follow-up in this case, it is defendant, not the trial court or the People. While the majority states that defendant “immediately” challenged prospective juror No. 23, that is not the case. After eliciting only that prospective juror No. 23 could have “some concernís],” defendant asked for “just a minute.” After the consequent pause in the proceedings, defendant asked no further questions; specifically, defendant did not ask prospective juror No. 23 if he could vote for life, the question that we found to be critical in Harris.

In sum, prospective juror No. 23 never expressed any “unwillingness or an inability” to obey his oath. At most, he indicated “some” concerns premised on his emotions, nothing more. The trial court properly took into account prospective juror No. 23’s “appearance and demeanor”—specifically noting that he had been crying—when rejecting defendant’s for-cause challenge under CPL 270.20 (1) (f) (Carolin, 115 NY at 659; see also Harris, 98 NY2d at 486-488). Accordingly, the trial court did not err when it denied defendant’s subdivision (1) (f) challenge to prospective juror No. 23.

B. Prospective Juror No. 855

The trial court also did not err when it granted the People’s subdivision (1) (f) challenge to prospective juror No. 855. The full voir dire and questionnaire of this juror reveal someone who was not death qualified. Furthermore, the record completely belies the majority’s assertion that the trial court somehow favored the People in this instance.

In People v Carolin (115 NY 658, supra), the trial court granted several challenges by the People to prospective jurors, each of whom had “said, in substance, that his scruples were such as would render him ‘extremely reluctant to find the de*138fendant guilty of murder in the first degree’ ” {id. at 659). Later, the jurors’ statements were “attenuated and somewhat modified by further examination” {id.). We held that the trial court properly granted the People’s challenges because—despite the statements of impartiality by the jurors—“it was for the court to say, from the whole examination of the juror, including his appearance and demeanor, whether he was fit and competent to perform fairly and impartially” {id.).

Here, by contrast, the prospective juror was more than “extremely reluctant” to impose the death penalty. Pointing to prospective juror No. 855’s questionnaire, the majority states merely that she “equivocated” in her views regarding the death penalty (majority op at 46). Indeed, at times, prospective juror No. 855 checked boxes marked “yes” to indicate that she could deliberate impartially and consider the death penalty. Those generic answers, however, must be juxtaposed against her initial, express written statement (unmentioned by the majority) that “[i]n my heart I do not believe I have or could condone death for another person for [sic] punishment.”

Several other answers and written statements (again, all unmentioned by the majority) similarly reveal a juror who appears unwilling to impose the death penalty. For instance, in the first question requiring a checked answer regarding the death penalty in this case, prospective juror No. 855 checked “no” when asked “[i]f the People prove the defendant guilty beyond a reasonable doubt of Murder in the First Degree, could you consider the death penalty as one of the available punishments.”

Then, prospective juror No. 855 indicated that she would not be able to “set aside [her] own views and conscientiously apply the law to the facts of this case in accordance with the Judge’s instructions” “[i]f the law, as presented to [her] by the Judge in this case, is contrary to [her] present personal beliefs concerning the death penalty.” Explaining this answer, prospective juror No. 855 wrote that “I am morally responsible for my own actions and decisions.”

As if these answers were not sufficient reason to disqualify prospective juror No. 855, another of her responses could only have caused the People and the trial court grave apprehension. Specifically, prospective juror No. 855 did not consider it “wrong for a juror to go through the whole trial and then during the penalty phase deliberations express for the first time that they would never or always vote for the death penalty regardless of the facts of the case.”

*139Thus, the trial court was confronted with a potential “stealth” juror; i.e., a juror who believed that her personal opposition to capital punishment trumped any obligation to comply with her oath, and also allowed her to seek to be selected so that she could ensure a non-death sentence; in short, a juror willing to sabotage the trial. This had to make the Trial Judge seriously question the credibility of those questionnaire answers in which prospective juror No. 855 indicated that she would deliberate fairly. Similarly, even if this juror later said all the “right things” during voir dire, her questionnaire responses could only cause the trial court to view her with some skepticism if not outright suspicion.

Indeed, when challenging prospective juror No. 855 the People expressly pointed out to the Trial Judge that, “[f]irst of all you have the ability to have seen her body language and her in person.” And, contrary to the majority’s view of the record, the People never asked for prospective juror No. 855’s exclusion because her death penalty views were “unknown” (majority op at 48). The People expressed concern that prospective juror No. 855 “has an inner moral compass that says she is opposed to the death penalty except ... in quote certain circumstances [w]hich she is unwilling or unable to tell us what they are.” The People then noted that “[t]he only criteria she elaborates on is multiple victims.” The only thing “unknown,” according to the People, was “[w]hat those other inner workings of her own personality are that trigger her consideration of the death penalty”; not prospective juror No. 855’s death penalty views.

The People challenged prospective juror No. 855 because of the reasonable belief that she was irrevocably opposed to capital punishment except perhaps in a case involving multiple victims. Contrary to the majority’s conclusion, the People never invited the trial court to exclude prospective juror No. 855 on the basis of her “unknown” views of the death penalty. Instead, the People focused on prospective juror No. 855’s inability to credibly indicate that she could vote for death “in this case” (Harris, 98 NY2d at 485 n ll).18

The majority also ignores and misconstrues the trial court’s reason for granting the People’s challenge. In fact, the majority *140never mentions that the trial court premised the exclusion of prospective juror No. 855 on the facts that “I have listened to [her], I have formed my opinion.” And the trial court had to be “listening]” to prospective juror No. 855 and observing her “body language” during voir dire, while inevitably mindful of her dubious questionnaire responses (see Carolin, 115 NY at 659). As we have long recognized, even if a prospective juror’s “extreme[ ] reluctan[ce]” to impose the death penalty is “attenuated and somewhat modified by further examination, it was for the court to say, from the whole examination of the juror, including his appearance and demeanor, whether he was fit and competent to perform fairly and impartially his duty as a trior of the issues which involved such result” (id.). This is exactly what the trial court did here.19

Nor did the trial court apply different standards to prospective jurors Nos. 23 and 855. In reaching a contrary conclusion, the majority does not acknowledge that the trial court in both instances took into consideration “appearance and demeanor” (Carolin, 115 NY at 659), and was in a position—as we most decidedly are not—to assess these jurors’ questionnaire and voir dire answers in that light.

The trial court took pains to note that prospective juror No. 23 had been crying while relating his life experience. The trial court concluded that prospective juror No. 23 was serious about carrying out his oath if selected. By contrast, the trial court considered prospective juror No. 855, who had indicated a willingness to deceive the court and parties regarding her true intentions, to be unworthy of belief after “listening]” to her and watching her “body language.” The trial court did not apply different standards to each party’s for-cause challenge; applying the same standard, he reasonably reached different *141conclusions about each prospective juror’s credibility and ability to deliberate impartially.

As the majority cautions, trial courts must scrupulously guard against jury selection errors, especially in a resource-intensive capital case. Of course, the trial court here properly excluded prospective juror No. 855 precisely because seating her might have led to a waste of resources. Indeed, a “stealth” juror bent on determining the outcome of a capital case forces both the parties and the judiciary to expend resources needlessly. A “stealth” juror poses an equal danger of injustice to a capital defendant, the People, and the justice system.

III. Conclusion

The majority opinion is a remarkable piece of judicial legerdemain, shot through with after-the-fact analysis. Legitimate choices made by the Legislature, by the Miller court and by the Trial Judge are vigorously second-guessed and overthrown from the safe remove of our appellate aerie. Finally, of course, the majority overturns the measured, reasoned verdict that 12 citizens of our state undoubtedly struggled to reach in what was likely the most solemn responsibility they will ever have to undertake.

For all the reasons stated, I would affirm defendant’s conviction for first-degree felony murder. For the reasons stated by Judge Graffeo, I would also affirm defendant’s conviction for first-degree witness elimination murder. I respectfully dissent from the majority’s conclusion in Part II regarding prospective jurors Nos. 23 and 855, and from the reversal of defendant’s first-degree murder convictions in Part III.

Finally, the majority’s disposition of this case renders the sentencing issues academic (see Harris, 98 NY2d at 497; see also People v Davis, 43 NY2d 17, 29 [1977], cert denied 435 US 998 [1978]). We are constitutionally and prudentially precluded from issuing advisory opinions (see Harris, 98 NY2d at 497).

Chief Judge Kaye and Judge Ciparick concur with Judge Rosenblatt; Judge G.B. Smith concurs in result in a separate opinion in which Judge Ciparick concurs insomuch thereof as addresses deadlock jury instructions; Judge Graffeo concurs in part and dissents in part in another opinion in which Judge Read concurs; Judge Read concurs in part and dissents in part in an opinion in which Judge Graffeo concurs.

Judgment modified, etc.