OPINION OF THE COURT

MEMORANDUM.

Order affirmed. Defendant raises no error warranting a reversal of his conviction.

Chief Judge LIPPMAN (concurring).

Defendant stands convicted upon a jury verdict of murdering Catherine Woods. At trial, there was evidence that he had felt himself to be romantically involved with Ms. Woods, but that after a time Ms. Woods made it clear that she did not want their relationship, such as it was, to continue. There was also evidence that defendant expressed anguish over his rejection by Ms. Woods, both to acquaintances and in his personal journals, and that, at or around the time Ms. Woods finally turned him away, he telephoned her with extraordinary frequency.1 Ms. Woods was murdered on the evening of November 27, 2005 — a little over a month after breaking with defendant — in the East 86th Street apartment she shared with her long-time boyfriend David Haughn. Haughn testified that on returning to the apartment just before 7:00 p.m., after an absence of about an hour, he found Ms. Woods on the bedroom floor lying in a pool of blood. There was forensic evidence that Ms. Woods had been stabbed repeatedly about the neck, and that the bedroom walls were spattered with her blood. Within a group of wall stains appearing to a Crime Scene Unit Detective to have been produced by the impression of a bloodied hand — a so-called "hand transfer" print — a latent fingerprint was discovered. The People's forensic experts testified that, after enhancement and analysis, the print had been determined to match that produced by defendant's left index finger. Defendant's cell phone records list some 14 calls to Ms. Woods on the day of the homicide and there was proof that of these, several, made between 5:27 p.m. and 6:33 p.m., originated from the vicinity of the victim's East 86th residence. Earlier calls from the same cell phone traced the user's progress from the neighborhood in which defendant lived toward the murder site.

Defendant acknowledges that the trial evidence, viewed as it must be on appeal, in the light most favorable to the People, was sufficient to support the murder verdict. His claim to appellate relief is premised instead upon two alleged defects in the underlying proceedings. He argues first, that he was ineffectively represented at trial because one of his attorneys had personal interests that conflicted with her professional obligations to him, and second, that the receipt in evidence of entries from his journals, dating from between six and three years before Ms. Woods' murder, documenting never-acted-upon misogynistic thoughts about two former girlfriends, was in error, because his prior bad thoughts were not properly relevant to proving his commission of the crime charged. These arguments were rejected by the Appellate Division (85 A.D.3d 409, 410-411 [2011]) and are now before us on this further appeal, pursuant to leave granted by a Judge of this Court (19 N.Y.3d 972 [2012]). We conclude, that while defendant's contentions are substantial and possess a fair measure of merit, they do not in the end make out a right to relief from the judgment of conviction.

I.

Defendant was represented at trial by two attorneys; lead counsel Laura Miranda, Esq. was second-seated by an attorney (hereinafter "co-counsel") retained for her purported expertise in dealing with forensic evidence, the most crucial component of the People's case. Before the trial began, however, it was disclosed that co-counsel had been indicted by a New York County grand jury; she was alleged to have smuggled drugs to a client in prison. Inasmuch, then, as co-counsel faced prosecution by the Office of the New York County District Attorney, the same office that was concurrently prosecuting her client, Mr. Cortez, there was at least a potential conflict of interest; it was entirely plausible that co-counsel's natural concern over how she would be dealt with in her own case would inhibit the vigor of her opposition to her prosecutor's case against her client. A Gomberg hearing was thus held for the purpose of ascertaining on the record that defendant was knowingly electing to continue with co-counsel as his attorney notwithstanding any conflict that her prosecution might pose to her single-minded advocacy on his behalf (see People v Gomberg, 38 N.Y.2d 307 [1975]).

At the hearing, the trial court first elicited from defendant that lead counsel had spoken with him about "[co-counsel's] pending matter." The court then acknowledged "an argument" that co-counsel had a conflict of interest — "that she might, for some reason, be more interested in her own matter than [that of defendant]" — but added that she was "not quite sure [she saw] it factually, frankly." Nonetheless, the court said it was important for defendant to understand that if co-counsel was convicted, she could lose her license to practice law. The inquiry concluded with the court asking defendant to make explicit that he understood what was "going on" and that he wished to proceed with co-counsel anyway. Defendant responded, "Yes. I do understand that. And she has not compromised this case on account of her own," whereupon the trial court affirmatively ended the inquiry, indicating that it would not be necessary for defendant to go into further detail about his understanding of what was "going on." Defendant contends that his waiver of co-counsel's conflict is not adequately made out by this colloquy.

While the constitutional right of a criminal defendant to effective representation entails the right to conflict-free representation (see Wood v Georgia, 450 U.S. 261, 271 [1981]; People v Ortiz, 76 N.Y.2d 652, 655-656 [1990]), most, but not all, attorney conflicts may be waived so as to permit continued representation by the defendant's attorney of choice (see People v Carncross, 14 N.Y.3d 319, 327-330 [2010]). But waivers, particularly of fundamental constitutional entitlements, to be valid, must be demonstrably knowing, intelligent and voluntary (Edwards v Arizona, 451 U.S. 477, 482 [1981]; Johnson v Zerbst, 304 U.S. 458, 464 [1938]); there must be a record sufficient to overcome the presumption against them. Defendant's attorney conflict waiver then, may not be deemed effective unless the record unambiguously permits the inference that he knowingly, intelligently and voluntarily relinquished his constitutional right to unimpaired, i.e., conflict-free, assistance of counsel.

Notwithstanding their potentially pivotal importance, we have resisted a uniform judicial catechism for the taking of attorney conflict waivers (see People v Caban, 70 N.Y.2d 695, 697 [1987]; People v Lloyd, 51 N.Y.2d 107, 112 [1980]), preferring to allow trial judges to tailor the inquiry to the particular circumstances to which the waiver relates. In Gomberg we required only that the court be "satisfied" that the waiver was informed (38 NY2d at 313); the actual task of informing the defendant as to the conflict, we indicated, was the ethical and representational obligation of counsel (id. at 314). Indeed, Gomberg may be read to allow a trial judge, in assessing whether a purported waiver was adequately informed, to rely, nearly implicitly, upon counsel's assurance that the client has been appropriately advised of the conflict and its risks (id.). A different approach — one involving a more probing inquiry by the court — we said, risked intruding upon the attorney-client relationship (id. at 313).

Gomberg, it appears, may have overstated the extent to which a court may rely upon the assurance of a possibly conflicted attorney in judging whether a defendant's election to continue with that attorney was informed. Some four years after Gomberg, in People v Macerola (47 N.Y.2d 257 [1979]), we emphasized that the court's obligation in responding to a possible attorney conflict was independent of that of counsel, and would be met "[o]nly after sufficient admonition by the trial court of the potential pitfalls" posed by the conflict (id. at 263). And, in People v Baffi (49 N.Y.2d 820, 822 [1980]) we reiterated that

More recently, in People v Solomon (20 N.Y.3d 91 [2012]), we found the record insufficient to document a valid attorney conflict waiver where, although counsel represented that she had discussed her conflict with the defendant (id. at 94), the nature of the conflict was "not even" made a matter of record by the court (id. at 95).

Defendant and amici point out that federal cases have understood inquiry respecting a possible attorney conflict and the validity of a defendant's election to waive it to be, centrally, a judicial function (see e.g. United States v Levy, 25 F.3d 146, 158 [1994]), and that the widely employed protocol for passing upon attorney conflict waivers set forth in United States v Curcio (680 F.2d 881, 888-890 [1982]) contemplates a significantly more particular and searching judicial inquiry than that described by Gomberg — one establishing on the record that the court has

While we do not adopt or require the Curcio inquiry and do not view each of its six elements as invariably indispensable to a valid conflict waiver, the protocol appears well-designed to create a record from which the validity of a conflict waiver, or the lack thereof, may be readily discerned. It is an approach that has proved workable (see e.g. United States v Graham, 493 Fed Appx 162 [2d Cir 2012]; United States v Williams, 372 F.3d 96, 109 [2d. Cir 2004]; United States v Buissereth, 638 F.3d 114, 117 [2d Cir 2011]; United States v Basciano, 384 Fed Appx 28 [2d Cir 2010]; United States v Iorizzo, 786 F.2d 52, 59 [2d Cir 1986]) and which would be prudently followed in our criminal courts where there is doubt, as there evidently was in this case,2 as to how a conflict waiver inquiry should proceed.

We need not, in any event, apply Curcio's conditions to the letter to understand that the present colloquy between defendant and the trial court simply does not provide the necessary assurance that co-counsel's conflict and its risks were understood and freely assumed by defendant in the context of a choice essentially defined by the entitlement to conflict-free representation. Defendant was never informed by the court of that entitlement, and, although the nature of co-counsel's possible conflict should, from a lawyer's perspective, have been clear (see Thompkins v Cohen, 965 F.2d 330, 332 [7th Cir 1992, Posner, J.]; United States v Levy, 25 F3d at 156), the court said that she was not sure she saw it "factually." There was, accordingly, no judicial admonition at all as to the fairly palpable risks co-counsel's continued representation could hold for defendant. It would not be reasonable to suppose that defendant, a layperson, perceived and understood a risk that the court itself was unable clearly to discern. While the People emphasize defendant's declaration that co-counsel "has not compromised this case on account of her own," that bare, unelaborated statement3 manifestly does not demonstrate defendant's awareness that co-counsel's continued representation entailed a prospective risk that his defense to the charge of murder could be compromised by his attorney's personal vulnerability to his prosecutor.

The court's near complete reliance upon attorney Miranda to explain co-counsel's conflict, and its possible ramifications to defendant, was neither consonant with our post-Gomberg conflict waiver jurisprudence, nor prudent under the circumstances. Ms. Miranda, although not laboring under co-counsel's conflict, had relied upon co-counsel to cover critical issues for the defense, most notably those involving the prosecution's potent forensic proof. She would naturally have been reluctant to dispense with co-counsel's assistance mid-trial.4 And, having herself just been held in contempt by the court for delaying the trial, would not have been anxious to incur additional judicial displeasure by apprising her client in such a way as to make the substitution of co-counsel and additional delay, or her immediate assumption of representational responsibilities for which she was unprepared, likely. The possibility that an attorney in her position would understate the risks entailed by co-counsel's continued representation was not negligible. There is, in short, serious reason to doubt whether such advice as she may have given defendant on the subject of co-counsel's conflict was the product of independent and disinterested professional judgment.

Defendant argues that, if his attorney conflict waiver was invalid, there must be a reversal because an unwaived attorney conflict of any sort functions to deprive a defendant of the right to make an informed choice as to who will represent him and thus introduces into the basic structure of a criminal trial a flaw with incalculable, potentially highly prejudicial sequellae. We, however, have discerned a meaningful distinction between conflicts that are actual and those that only potentially impair an attorney's discharge of her professional obligations in a particular matter. If falling within the former category, an unwaived conflict requires reversal (see People v Solomon, 20 NY3d at 97), but if within the latter, relief, we have held, depends upon a showing by the defendant that the conflict "operated on" the defense (see id. at 98; and see People v Sanchez, 21 N.Y.3d 216, 223 [2013]; People v Ennis, 11 N.Y.3d 403 [2008]; People v Konstantinides, 14 N.Y.3d 1, 10 [2009]).

We have not viewed an attorney conflict as "actual" for purposes of deciding whether a defendant has received effective representation, except where the attorney may be understood to have divided and incompatible loyalties within the same matter necessarily preclusive of single-minded advocacy (see e.g. Solomon, supra; People v Prescott, 21 N.Y.3d 925 [2013]). Here, defendant's argument that co-counsel was actually conflicted, since at every turn in the trial she might naturally have been apprehensive as to whether full-throttle advocacy on defendant's behalf risked antagonizing the prosecutor upon whose favorable discretion her own pending matter would likely depend, while perhaps accurately describing the risk, is not legally viable after our decision in Konstantinides (supra; see also People v Townsley, 20 N.Y.3d 294, 299-301 [2012]). There, we held that a prosecutor's allegations respecting defense counsel's criminal involvement, even with respect to the very matters with which the representation was concerned, raised only a potential conflict (14 NY3d at 13-14; but see Konstantinides v Griffin, 2011 WL 3040383, 2011 US Dist LEXIS 80618 [ED NY, July 25, 2011, No. 10-CV-05999 (JG)]). Konstantinides' second-seated counsel was accused by the prosecutor of suborning perjury in the case on trial, but neither that allegation nor the concomitant prospect that the attorney would be called as a trial witness sufficed to relieve Konstantinides of the burden to show, as a condition of relief, that the conflict raised by the prosecutor's allegations actually operated on his defense (14 NY3d at 13-14). The present facts are significantly less compelling. Co-counsel's alleged wrongdoing, unlike that of Konstantinides' attorney, was factually unrelated to her client's case and raised no possibility of her being called to give testimony adverse to defendant's interests.

If, as our precedents dictate, defendant cannot obtain relief premised on co-counsel's unwaived conflict unless he can show that the conflict operated on the defense, it is evident that he cannot now prevail. While he faults co-counsel's conduct of those portions of the defense with which she was entrusted, pointing out, among other things, that she was on occasion unprepared and arguably blundered in cross-examining two prosecution witnesses, the record affords no basis to conclude that the claimed lapses were attributable to the alleged conflict. And, contrary to defendant's suggestion, our cases do not provide that such a connection may be presumed. Although a defendant need not show that an unwaived potential conflict was a source of specific prejudice (Sanchez, 21 NY3d at 223; Konstantinides, 14 NY3d at 12-13), we have been clear that a connection between the potential conflict and the representational deficiency must be made out (see Sanchez, supra; People v Ennis, 11 NY3d at 411). Whether it is made out, moreover, is ordinarily treated as a mixed question (Konstantinides, 14 NY3d at 10). Even if co-counsel's representation of defendant was in some important respects wanting, the record sheds no light on the fact-sensitive question of whether any such deficiencies were traceable to conflict generated reticence. That being the case defendant's conflict-based ineffective assistance claim, as it is now presented, must fail.

II.

As part of their direct case, the People sought to place in evidence, not only the entries from defendant's journals in which defendant ruminated over being spurned by Ms. Woods — entries whose admissibility is not now disputed — but other entries concerning his relationships with and rejections by two other women dating from three to six years before Ms. Woods' murder. In the latter entries, defendant expressed "pent up rage" at the former objects of his affections — whom he described as "poisonous" and "dangerous" — and berated himself for having let these purportedly unfaithful partners turn him into an object of ridicule. He portrayed himself as a "beast of burden" unable "to find retribution," i.e., "to kill." He was, he said, a "monster," obsessed with "thoughts of revenge." These diary entries included poems and drawings thematically preoccupied by revenge fantasies in which knives were the retributive instrument of choice.

As the prosecutor conceded, defendant never physically harmed his prior girlfriends. The theory upon which the diary entries about them were sought to be admitted, then, was not that the entries were evidence of prior bad acts relevant in some non-propensity-based way to the proof of the charged crime — i.e., as evidence admissible under some exception to the rule barring the inference of guilt from proof of no more than propensity, classically articulated in People v Molineux (168 N.Y. 264 [1901]) — but rather that they evidenced a simmering misogynistic rage which, over time, progressed in its intensity and expression from the realm of fantasy to enactment. The trial prosecutor said that the proffered evidence as to defendant's state of mind provided "the only context in which we can understand what happened here." Defendant objected to the receipt of the evidence on the ground that it would be understood as indicative of a misogynistic propensity and would engender speculation as to whether he acted upon some previously unsatisfied homicidal impulse. Defendant also specifically argued that the People had not made any Molineux application and that, in the absence of Molineux vetting, the ground for the receipt of the propensity evidence was not clear. The trial court, in ruling the evidence admissible, responded that she was "not sure what [the evidence] would indicate" but that "the People will be arguing that there's a reflection of a growing kind of state of mind." When counsel objected that this was just a way of describing propensity evidence, the court replied that the evidence was not evidence of "propensity to commit a crime" and that she did not see it as "fitting into Molineux."

If in fact this evidence was fairly probative of "a growing kind of state of mind," as the court put it, or, as the prosecutor said, a "progression" of increasing hostility toward women culminating in Ms. Woods' murder, it was evidence of propensity, and the Molineux doctrine, at its core, forbids an inference of guilt from evidence probative of no more than predisposition to a kind of behavior (see People v Agina, 18 N.Y.3d 600 [2012]; People v Arafet, 13 N.Y.3d 460, 464-465 [2009]; People v Alvino, 71 N.Y.2d 233, 241 [1987]). It is true that, most frequently, Molineux's analytic framework is invoked where the prosecution proposes to introduce evidence of prior bad acts in proof of the crime charged. But the concern with evidence of propensity as a basis for a finding that a defendant has committed the particular crime for which he or she is on trial, is no less acute where the evidence suggestive of personal tendency is of mere thoughts. Indeed, the inference of guilt from the latter sort of propensity evidence is particularly perilous. There is a wide gulf between thought and act, especially conduct of a murderous sort. If, in the setting of a criminal trial, it is not an acceptable inference that a person is guilty simply because he has done things similar to those charged, it cannot be any more acceptable — and logically would be less so — to suppose that a defendant has done the thing of which he is accused simply because in some temporally remote context he has had thoughts of such things.

Of course, where there is an objectively discernible connection between thought and act, evidence of thought may be highly relevant and admissible to prove intent or motive (see People v Fitzgerald, 156 N.Y. 253, 258-259 [1898]; and see e.g. People v Moore, 42 N.Y.2d 421, 428 [1977]). But where thinking is itself probative of no more than mental preoccupation — i.e., it bears no readily discernible connection to an act — it cannot be reliably indicative of much. It has long been understood that "[t]he motive attributed to the accused in any case must have some legal or logical relation to the criminal act according to known rules and principles of human conduct. If it has not such relation, or if it points in one direction as well as in the other, it cannot be considered a legitimate part of the proof" (Fitzgerald, 156 NY at 258-259). Proof of no more than unconsummated mental preoccupation is an invitation to speculation incompatible with the disciplined inferential exercise required of a juror in a criminal case. The prosecutor in his summation offered that defendant was "like a volcano where the person you are is building up, building up and building up; and the volcano seems passive, dormant; and then, all of a sudden, boom; there's a huge explosion," thus exhorting the jury, on the basis of a richly descriptive but scientifically baseless geothermal analogy, to draw a connection between thoughts and an extreme act years removed. There was no competently established psychological theory to support such a connection. While the evidence of prior thoughts was purportedly introduced to supply "context," that very evidence in the end paradoxically created a need for context that the prosecutor could not and should not himself have attempted to provide.

The problem with the evidence of defendant's temporally remote broodings was, fundamentally, that it was too attenuated from any act to be relevant, even under some exception to the Molineux prohibition, to proving defendant's commission of Ms. Woods' murder. Although the People's appellate claim is that defendant's thoughts about other women who had rejected him were probative of his motive to kill, there was no issue as to whether defendant had a motive to kill Ms. Woods, only as to whether he had actually done so, and, as noted, no connection between defendant's distant misogynistic thoughts and the charged conduct was made out. What was instead invited was the conjecture that defendant possessed a misogynistic impulse that had finally blossomed into the murder of a young woman. And, while it was not difficult to construct a superficially convincing narrative based on this propensity driven supposition, it is precisely such a carelessly constructed yet highly seductive narrative that the Molineux doctrine prudently excludes from a criminal trial.

It would not be realistic to say that the introduction and exploitation of this inflammatory evidence was benign. It would, however, be at least equally unrealistic to suppose that it was outcome determinative. The properly admitted proof of defendant's morbid preoccupation with Ms. Woods, combined with the forensic crime scene evidence linking him to her murder, was extraordinarily powerful as were the cell phone records tracing defendant's movements toward and away from the locus of the crime. We agree with the Appellate Division that the proof before the jury overwhelmingly pointed to the conclusion that defendant was Ms. Woods' assailant. It may be, as defendant now argues, that the probative value of the latent print found at the crime scene should not have been as great as it was made to seem at trial; defendant in his reply brief refers to several articles challenging the accuracy with which such prints may be attributed to a particular person. But these studies, all of which were published after defendant's trial, are not part of the trial record and cannot bear upon our assessment of the strength of the proof actually before the jury. If there is a claim that defense counsel were ineffective for failing to pursue an available and potentially decisive line of defense more aggressively challenging the attribution of the latent murder scene fingerprint to their client, that would be appropriately raised, if at all, on a motion pursuant to CPL 440.10. It is not reviewable on the present record.

ABDUS-SALAAM, J. (concurring).

I agree that the trial court's inquiry into co-counsel's conflict of interest was deficient under our existing case law, and that the trial court erred by admitting the challenged journal entries into evidence. I also agree that these errors were harmless and did not deprive defendant of a fair trial. However, I write separately from Chief Judge Lippman and my two colleagues who join his concurring opinion to express my view that resolution of this case need not rest upon our adoption of a federal "protocol" governing a trial court's inquiry into an attorney's potential conflict of interest, or the novel expansion of the Molineux doctrine proposed in the Chief Judge's opinion.

I.

The right to effective assistance of counsel "ensures not only meaningful representation but also the assistance of counsel that is `conflict-free and singlemindedly devoted to the client's best interests'" (People v Berroa, 99 N.Y.2d 134, 139 [2002], quoting People v Longtin, 92 N.Y.2d 640, 644 [1998], cert denied 526 U.S. 1114 [1999]). A trial judge faced with an attorney conflict of interest must balance this right with the defendant's concomitant right to retain counsel of his or her choice (see People v Gomberg, 38 N.Y.2d 307, 313 [1975]). Although the trial judge undoubtedly "owes a duty independent of counsel to protect the right of an accused to effective assistance of counsel" (People v McDonald, 68 N.Y.2d 1, 8 [1986] [internal quotation marks omitted]), at the same time, the trial judge must be mindful not to "arbitrarily interfere with the attorney-client relationship" (Gomberg, 38 NY2d at 313).

Once the trial judge is informed of the conflict or aware of facts indicating that conflicting interests arguably exist, he or she "must conduct a record inquiry" to determine whether the defendant is aware of the possible risks involved in the potentially conflict-ridden representation and has made a knowing and informed decision to continue with that representation in spite of the conflict (McDonald, 68 NY2d at 8; see People v Solomon, 20 N.Y.3d 91, 95 [2012]; Gomberg, 38 NY2d at 313-314; see also People v Macerola, 47 N.Y.2d 257, 263 [1979]). The court's inquiry must be "sufficiently searching to assure that [the defendant's] waiver was informed and voluntary" (People v Caban, 70 N.Y.2d 695, 696-697 [1987]), but it generally "need not be as thorough or as detailed as that required of the attorney" (People v Lloyd, 51 N.Y.2d 107, 111 [1980]). As we explained in Lloyd, "[b]ecause the exact nature of the defense and particularly defense strategy must remain off limits to the court[,] the extent of the precautions to be taken by the trial court to insure that the defendant[ ] perceive[s] the risk inherent in [the] representation must necessarily involve a measure of discretion" (51 NY2d at 112). Thus, we require only that the trial judge "make a reasonable inquiry of possible conflict" (Gomberg, 38 NY2d at 316) that "examine[s] the nature of the relationship or circumstances that are alleged to establish a conflict" (People v Ennis, 11 N.Y.3d 403, 410 [2008], cert denied 556 U.S. 1240 [2009]), and admonishes the defendant as to the "potential pitfalls" inherent in the representation (Macerola, 47 NY2d at 263).

The Chief Judge's opinion apparently takes umbrage with this Court's reluctance to prescribe a particular "format or catechism that the court must follow" when conducting a conflict inquiry (Lloyd, 51 NY2d at 112; see Caban, 70 NY2d at 697). Indeed, we have preferred to "allow trial judges to tailor the inquiry to the particular circumstances to which the [conflict] waiver relates" (Lippman, Ch. J., concurring op at 1064).

In my view, this flexibility enables the trial court to effectively fulfill its duty to conduct a conflict inquiry within the particular context of the case before it. When a conflict inquiry takes place prior to trial (as we have said it should), the court may not be "fully aware of the evidence, the nature of the defendants' case or its ramifications" (Lloyd, 51 NY2d at 111), and an overly searching judicial inquiry could inadvertently "infringe upon the defendant's right to retain and confer with counsel" of his own choice" (Gomberg, 38 NY2d at 313; see Lloyd, 51 NY2d at 111 ["to require the defendant or his attorney to disclose to the court details of the defense, defense conferences, or strategy would in itself invade the defendants' rights, including the right to counsel"]). It is also not uncommon for trial judges to encounter defendants who are inclined, when confronted with probing questions from the bench, to share more information than is necessary or advisable concerning their attorneys' representation. Given that trial judges are in the best position to evaluate these case-by-case circumstances, they should be permitted to employ common sense rather than "catechisms," and should not be constrained by a conflict inquiry that is formulaic rather than adaptive to the conditions of the specific case.

The Chief Judge's opinion endorses, over our Gomberg line of cases, the stricter inquiry "protocol" outlined in United States v Curcio (680 F.2d 881, 888-890 [1982]) and "widely employed" by the Second Circuit (Lippman, Ch. J., concurring op at 1065). Like our Gomberg inquiry, the Curcio "procedures" are intended "to permit the court to determine whether the defendant's waiver of his right to conflict-free counsel is knowing and intelligent" (United States v Rodriguez, 968 F.2d 130, 139 [2d Cir 1992]). However, as the Chief Judge notes in his concurrence, the "Curcio inquiry format" is "a significantly more particular and searching judicial inquiry than that described by Gomberg" (Lippman, Ch. J., concurring op at 1065), as it requires a trial court to tick off six specific queries before it can be satisfied that the defendant validly waived the conflict.

While I am in favor of clarifying the scope of a proper Gomberg inquiry (to the extent the need for such clarity exists), we need not resort to federal precedent to do so. By suggesting that trial judges follow the Curcio inquiry, the Chief Judge's opinion retreats from decades of New York law, which requires "no prescribed catechism that [trial courts] must follow in ascertaining a defendant's understanding of his [or her] choices" (Caban, 70 NY2d at 697 [alterations and quotation marks omitted]). But as just explained, Gomberg's flexibility has its benefits, and I believe the Chief Judge's concurrence does not adequately consider how "a significantly more particular and searching judicial inquiry" could cause a trial court to infringe upon one right of the defendant in its effort to protect another (see Gomberg, 38 NY2d at 313; Lloyd, 51 NY2d at 111-112).

Further, although the Curcio protocol has "proved workable" in federal conflict cases (Lippman, Ch. J., concurring op at 1066), it may not have the same success in New York courts. While we share the Second Circuit's concern for protecting a defendant's right to conflict-free counsel, we have taken a decidedly different approach to providing what we consider adequate constitutional safeguards of that right (compare Curcio, with e.g. Gomberg, 38 NY2d at 313; Caban, 70 NY2d at 697), and it is not clear whether the Curcio approach appropriately accounts for our concern that a trial court avoid delving too deeply into the attorney-client relationship for fear of upsetting the defendant's right to retain counsel of choice (see Gomberg, 38 NY2d at 313).

The Chief Judge's opinion suggests that the Curcio protocol will resolve existing "doubt ... as to how a conflict waiver inquiry should proceed" (Lippman, Ch. J., concurring op at 1066). While there have been some calls for further clarity regarding a trial judge's responsibilities in conducting a Gomberg inquiry (see e.g. Lloyd, 51 NY2d at 112 [Jones, H.R., J., dissenting] [noting that "the responsibility of the Trial Judge in cases involving joint representation should be made clear"]), and the trial judge in this case admitted that she "never quite know[s] what to say about [an attorney's conflict]" (see Lippman, Ch. J., concurring op at 1066 n 2), these "doubts" do not appear to be so widespread as to warrant a substantial alteration in our conflict inquiry jurisprudence. Further, any need to clarify how a trial judge should conduct a proper Gomberg inquiry can be achieved by resort to our existing precedent.

I would not forgo our Gomberg conflict waiver jurisprudence, and to the extent that the Chief Judge's opinion relies on that precedent in holding that the trial court's conflict inquiry was deficient here, I concur in its rationale. I further agree that, although the trial court did not secure a valid waiver from defendant, defendant failed to meet his burden to show that co-counsel's potential conflict operated on his defense and, thus, his ineffective assistance of counsel claim must be rejected (see People v Konstantinides, 14 N.Y.3d 1, 10 [2009]).

II.

The second issue in this case concerns the admission of defendant's journal entries about his former girlfriends — written several years prior to the murder of Ms. Woods — in which defendant expressed, among other things, extreme hostility towards the girlfriends and, arguably, women in general. The Chief Judge's opinion contends that the challenged journal entries are "evidence of propensity" that should have been subject to the rigors of "Molineux's analytic framework" (Lippman, Ch. J., concurring op at 1070). While I agree that the trial court abused its discretion as a matter of law in admitting the journal entries that, in my view, should have been excluded based upon relevance and redundancy grounds, I cannot subscribe to the unwarranted expansion of the Molineux doctrine proposed in the Chief Judge's concurrence.

Decided in 1901, People v Molineux (168 N.Y. 264 [1901]) prescribed the now-familiar rule that evidence of a defendant's uncharged crimes, prior crimes, or prior bad acts is generally inadmissible when it serves "only to show the defendant's criminal propensity" (People v Caban, 14 N.Y.3d 369, 375 [2010]; see e.g. People v Cass, 18 N.Y.3d 553, 559 [2012]). The Molineux rule, we have explained, "`is based on policy and not on logic'" (People v Arafet, 13 N.Y.3d 460, 465 [2009], quoting People v Allweiss, 48 N.Y.2d 40, 46 [1979]). Although "[i]t may be logical to conclude from a defendant's prior crimes that he [or she] is inclined to act criminally," this evidence is nonetheless "excluded for policy reasons because it may induce the jury to base a finding of guilt on collateral matters or to convict a defendant because of his past" (People v Arafet, 13 N.Y.3d 460, 465 [2009] [internal quotation marks omitted], quoting People v Alvino, 71 N.Y.2d 233, 241 [1987]; see Molineux, 168 NY at 313).

In light of these unique concerns, the admission of Molineux evidence is subjected to "the most rigid scrutiny" (id.).

The trial court must be sensitive to "the particular prejudice that may result when a jury is made aware of the fact that the defendant has previously committed crimes that are similar to the charged crime" (People v Walker, 83 N.Y.2d 455, 459 [1994] [describing these concerns in the Sandoval context]). The reason for this is obvious: "it is much easier to believe in the guilt of an accused person when it is known or suspected that he [or she] has previously committed a similar crime" (Molineux, 168 NY at 313). Thus, although the second part of the Molineux inquiry is similar to "the test by which all relevant evidence is measured," a trial judge evaluating Molineux evidence must "approach the normal balancing process with a heightened awareness of the unique kind of prejudice that extrinsic offense evidence can produce" (People v Hudy, 73 N.Y.2d 40, 69-70 [1988, Wachtler, Ch. J., dissenting], abrogated on other grounds by Carmell v Texas, 529 U.S. 513 [2000]).

The Chief Judge's opinion recognizes that "Molineux's analytic framework" is most frequently "invoked where the prosecution proposes to introduce evidence of prior bad acts in proof of the crime charged" (Lippman, Ch. J., concurring op at 1070). His opinion nonetheless proposes applying that framework here because the challenged journal entries contained defendant's prior bad thoughts that were admitted, essentially, as "evidence of [his] propensity" to act on his increasingly violent and misogynistic thoughts by murdering Ms. Woods (id.).

The application of the Molineux rule suggested in the Chief Judge's concurrence represents a novel expansion of that doctrine which, in my view, is both unnecessary and ill advised. The Molineux rule was created to address a particular prejudice inherent to a particular type of proof: evidence of a defendant's prior crimes and bad acts. While we have recognized additional "nonpropensity purposes" for which prior crime evidence may be relevant beyond those announced in Molineux (People v Morris, 21 N.Y.3d 588, 594 [2013] [noting that "(t)he Molineux categories" — (1) intent, (2) motive, (3) knowledge, (4) common scheme or plan, or (5) identity of the defendant — "are not exhaustive"]), the Chief Judge's opinion points to no case in which we applied Molineux to evaluate the admission of evidence unrelated to a defendant's prior crime or misconduct.

Here, of course, the evidence concerns neither of these things. The challenged journal entries are writings that reflect, at most, defendant's prior bad thoughts about his former girlfriends. Defendant committed no crime by writing the journal entries and the parties stipulated that defendant never harmed the women referenced therein. Thus, the evidence here is simply not Molineux evidence, which we have consistently defined as evidence of a defendant's prior crimes or bad acts.

Nor should we expand the Molineux rule to include prior bad thought evidence simply because it was admitted as propensity evidence, as the Chief Judge's opinion contends. All relevant evidence of guilt in some sense shows the defendant's criminal propensity and has the potential to prejudice the defendant. But Molineux was not meant to apply to the admission of all propensity evidence. Rather, more than a century of our Molineux jurisprudence has made clear that prior crime/bad act evidence (and this particular propensity evidence alone) raises such a uniquely high risk of undue prejudice that the trial court must err on the side of deference to the defendant when considering whether to admit the evidence. Indeed, under Molineux, the prior crime evidence is often excluded because its undue prejudice outweighs any probative value. By applying Molineux in the context of this case, the Chief Judge's opinion proposes that prior bad thought evidence be subject to the same deferential analysis when, in my view, the deference should be reserved exclusively for proof of the defendant's prior crimes or bad acts.

Considering the new ground the Chief Judge's concurrence attempts to break, it provides little guidance on how to apply the expanded Molineux doctrine in future cases involving prior bad thought evidence. Ostensibly, this broadened doctrine could apply to any case involving evidence of a defendant's bad thoughts that are not part-and-parcel of the charged crime and that bear any indicia of criminal propensity. Identifying the journal entries as prior bad thought evidence was relatively straightforward here, but the task of determining whether evidence constitutes a prior bad thought that triggers the Molineux rule could prove unwieldy in future cases. The Chief Judge's opinion has also failed to consider whether the protocols attendant to the proffer of Molineux evidence — for example, the requirement that the prosecution make a pretrial application for a Molineux or Ventimiglia hearing (52 N.Y.2d 350 [1981]), and the rule that the trial court issue limiting instructions to accompany the admission of the evidence — apply when prior bad thoughts are at issue.

Ultimately, reliance on Molineux is unnecessary because the journal entries should have been excluded based on general evidentiary principles. We have observed that relevance "is not always enough" to render evidence admissible "since `even if the evidence is proximately relevant, it may be rejected if its probative value is outweighed by the danger that its admission would prolong the trial to an unreasonable extent without any corresponding advantage; or would confuse the main issue and mislead the jury; or unfairly surprise a party; or create substantial danger of undue prejudice to one of the parties'" (People v Davis, 43 N.Y.2d 17, 27 [1977], quoting Richardson, Evidence § 147 at 117 [Prince 10th ed]; see also People v Walker, 83 N.Y.2d 455, 463 [1994] [noting that, even absent the special risk of an inference of propensity arising from the use of prior crimes evidence for impeachment in the Sandoval context, the trial court's exercise of discretion in admitting evidence must be informed by "ordinary principles of common sense and fairness"]; People v Duncan, 46 N.Y.2d 74, 80-81 [1978] [finding that prior inconsistent statements admissible for impeachment purposes may still be excluded as a matter of the trial court's discretion in order to avoid undue exploration of collateral matters]). Although a trial court enjoys broad discretion in deciding whether to admit evidence challenged as unduly cumulative and prejudicial, the court commits legal error whenever the record clearly reflects the court's complete failure to exercise its discretion in response to a defendant's focused challenge to the admissibility of the evidence (see Walker, 83 NY2d at 459 [stating that this Court will disturb a trial court's exercise of discretion in admitting evidence "only where `the trial court ha(s) either abused its discretion or exercised none at all'"], quoting People v Williams, 56 N.Y.2d 236, 238 [1982]; see also People v Petty, 7 N.Y.3d 277, 286 [2006] [observing that the decision to admit or preclude evidence on the ground that it is unduly cumulative lies within the discretion of the trial court]).

Here, the Chief Judge's opinion concludes that the journal entries were "too attenuated from any act to be relevant, even under some exception to the Molineux prohibition, to proving defendant's commission of Ms. Woods' murder" (Lippman, Ch. J., concurring op at 1071). Indeed, the contested journal entries, which were about women other than Ms. Woods and were temporally remote from her murder, neither addressed defendant's actions or attitude toward Woods nor revealed information about defendant's general state of mind that could not have easily been gleaned from the journal entries about the victim herself, which were admitted without any objection. Rather, the evidence was cumulative of admitted entries, bore little probative worth, and prejudicially suggested that defendant was generally a misogynist and a bad person.

Given that defendant drew the trial court's attention to these very issues, the trial court should have at least exercised some discretion by considering the probative value, prejudicial effect, and cumulative quality of the evidence before admitting it. Instead, the record shows that the court ended its inquiry at relevance without addressing those other important considerations. Accordingly, the court abused its discretion as a matter of law by erroneously admitting the disputed evidence. As the Chief Judge's opinion notes, such an abuse of discretion was harmless in light of the other evidence proving defendant's guilt.

Chief Judge LIPPMAN and Judges GRAFFEO, READ, SMITH, PIGOTT and ABDUS-SALAAM concur; Chief Judge LIPPMAN concurs in an opinion in which Judges GRAFFEO and SMITH concur; Judge ABDUS-SALAAM concurs in an opinion in which Judges READ and PIGOTT concur; Judge RIVERA taking no part.

Order affirmed.