—[*1]

Judgment, Supreme Court, New York County (Edward McLaughlin, J.), rendered June 19, 2002, convicting defendant, after a jury trial, of two counts each of robbery in the first and second degrees, and sentencing him to concurrent terms of 9½ years, reversed, on the law, and the matter remanded for a new trial.

Defendant challenges the court's charge on the burden of proof, which included the following instruction:

"During the course of a criminal trial in the United States, every defendant is entitled to every factual inference in his favor which can reasonably be drawn from the evidence. And where two factual inferences can be drawn from the evidence, both being of equal weight and strength, one factual inference consistent with guilt and the other factual inference consistent with innocence, any defendant is entitled to the factual inference of innocence.

"With regard to facts and with regard to the verdict, there are two different burdens of proof operating simultaneously. . . . With regard to finding of fact during the course of your deliberations, the burden of proof is simply that it is more likely than not that the fact exists, 50.1 beating 49.9."

Initially, we reject the People's contention that defendant's challenge is unpreserved. At the conclusion of the charge, defense counsel stated that he had "a problem. I understand the 50.1 and 49.9 point. I think it may be confusing to the jury. And it might apply to the evidence charge as opposed to the beyond a reasonable doubt. I think it serves to confuse them." He requested that the court reread the standard reasonable doubt charge on "to convict or not convict."

There is no question that counsel "registered" his "protest" to the court's instruction "when the court had an opportunity of effectively changing the same" (CPL 470.05 [2]). It also is plain that the court understood counsel's protest as an objection to its elaboration on the "two-inference" charge. Indeed, the court acknowledged that it is aware that the "two-inference" charge is a subject of controversy. [*2]

This Court has repeatedly expressed its disapproval of the "two-inference" charge (see e.g. People v Cruz, 172 AD2d 383, 383 [1991] [two-inference "instructions have been criticized as potentially confusing to the jury"], lv denied 78 NY2d 964 [1991]; People v Cohen, 61 AD2d 929 [1978] [use of phrase, "if the scales are even," a variation on two-inference language, "does not facilitate the resolution of the issues before the jury in a criminal case"], lv denied 44 NY2d 854 [1978]; People v Fox, 72 AD2d 146, 147 [1980] ["We take this opportunity again to repeat the admonition that it is undesirable in criminal cases to refer to . . . 'the scales are even' "], lv denied 49 NY2d 804 [1980]; People v Stinson, 186 AD2d 23, 25 [1992] ["allusions to 'equal inferences' . . . pose the same risk recognized in 'scales are even' language, namely that, if instructed that a defendant should be acquitted 'if the scales are even', a jury may draw the negative implication that, if the scales are not even, there is enough evidence to convict"; People v Allan, 192 AD2d 433, 435 [1993] [referring to "infirmity" in instructions that permit the jury to conclude that if inferences were not equal, evidence was sufficient to convict]; People v Ellis, 202 AD2d 301 [1994] ["we have previously noted that it is undesirable to use the phrase 'the scales weigh even' in the reasonable doubt charge in a criminal case"], lv denied 83 NY2d 910 [1994]; People v Durden, 211 AD2d 568, 569 [1995] [referring to "disfavored 'two-inferences' charge language"], lv denied 85 NY2d 937 [1995]).

The two-inferences language has been described as "formally correct" (Fox, supra at 147) and "technically correct" (Stinson, supra at 28 [Sullivan, J., dissenting]). And it is "obviously correct as far as it goes. But such an instruction by implication suggests that a preponderance of the evidence standard is relevant, when it is not. Moreover, the instruction does not go far enough. It instructs the jury on how to decide when the evidence of guilt or innocence is evenly balanced, but says nothing on how to decide when the inference of guilt is stronger than the inference of innocence but no[t] strong enough to be beyond a reasonable doubt" (United States v Khan, 821 F2d 90, 93 [1987]).

Since "[i]n a charge that properly instructs the jury on reasonable doubt, the two-inference language adds nothing" (id. [internal quotations marks and citation omitted]), to the extent that it does not taint the entire charge, we have held that it is not reversible error (see, Cruz; Cohen; Ellis; Durden; but see Stinson [reversing in part on ground of error in "equal inferences" charge]; Allan at 435 [reversing in part on ground of "infirmity" in two-inference instructions]). Of course, language that explicitly authorizes the improper inference, i.e., a charge that "there must be a preponderance in order to establish guilt," is error so egregious as to warrant reversal (People v Bailey, 121 AD2d 189, 191 [1986]). The lesson with respect to the disfavored if not fatal two-inference charge is that "[i]t is . . . better to reduce the risk, however remote, that a juror might draw the negative inference that if the scales are uneven, that is enough to convict" (Fox, supra at 147-148).

The lesson apparently has gone unheeded in at least one courtroom.

The instant charge began with the two-inference instructions that "have been criticized as potentially confusing to the jury" (Cruz, supra, 172 AD2d at 383) because they permit the jury to find the defendant guilty "when the inference of guilt is stronger than the inference of innocence but no[t] strong enough to be beyond a reasonable doubt" (Khan, supra, 821 F2d at 93). That is, the court instructed that "where two factual inferences can be drawn from the evidence, both being of equal weight and strength, one factual inference consistent with guilt and the other [*3]factual inference consistent with innocence: any defendant is entitled to the factual inference of innocence." The court then instructed the jury that "[w]ith regard to facts and with regard to the verdict, there are two different burdens of proof operating simultaneously"; that "[w]ith regard to finding of fact during the course of your deliberations, the burden of proof is simply that it is more likely than not that the fact exists, 50.1 beating 49.9"; and that "the People's obligation is to prove a defendant's guilt beyond a reasonable doubt."

These instructions are not merely potentially confusing. They could undermine the jury's understanding of the presumption of innocence, the burden of proof and the standard of proof beyond a reasonable doubt. Instead of reducing the risk "that a juror might draw the negative inference that if the scales are uneven, that is enough to convict" (Fox, supra at 147-148), they increase it. They risk eliminating the reasonable doubt standard from the trial. While the remainder of the court's instructions on reasonable doubt were proper, they did not remove the taint from the reasonable doubt charge as a whole.

Moreover, in its final instruction to the jury, which addressed the requirement of a unanimous verdict, the court repeated language that it had used to explain the burden of proof. "The two most important civic functions that citizens do are to vote and to serve on juries. In voting in elections, close counts, 50.1 beats 49.9 every time. . . . The same group, citizens who can't elect anybody by acclamation, who elect some people 50.1 vote to 49.9, minus one vote, have for about two hundred thirty years rendered unanimous verdicts. How does that happen[?] . . . [Y]ou can change your opinion, your vote, provided based on reason, logic, common sense, and reliance on the record of this case if somebody can cause you to change your mind. . . ."

Thus, the last instruction the jurors heard before they began deliberating was a numerical expression of the simple majority in voting that echoed the numerical expression of the preponderance standard that the court had used earlier in explaining the burden of proof. This instruction could only have reinforced the court's improper instruction on the standard of proof.

Justice Saxe's concurrence sets forth the additional reasons for the deficiency of the preponderance charge on factfinding when given without proper explanation.

We have held that, while the trial court in a criminal case should use the statutory phrase "beyond a reasonable doubt" (CPL 300.10 [2]) in "explicitly" charging that "guilt must be proved beyond a reasonable doubt," "this does not mean that elaboration and semantic variations on this theme which do not reduce the standard are error" (Fox, supra at 146-147). We added parenthetically on that occasion that "[i]n an unpredictable world, cautious Trial Judges will of course continue to try not to stray from language that has been approved by appellate courts" (id. at 147). The instant case compels us to admonish that elaboration and semantic variations on the theme of reasonable doubt that reduce the standard are error and cautious trial judges will continue to try to avoid using language that has met with disapproval by appellate courts. "[A] Judge should think long and hard about the wisdom of departing from the standard charge on such elementary matters as reasonable doubt [and] burden of proof" (People v Nunez, 182 AD2d 527, 528 [1992] [Sullivan, J., concurring], lv denied 80 NY2d 836 [1992]). Concur—Ellerin, Lerner and Gonzalez, JJ. Tom, J.P., and Saxe, J., concur in separate memoranda. Gonzalez, J., also concurs in the separate memorandum of Saxe, J.

Tom, J.P. (concurring).The court instructed the jury that if two factual inferences may be drawn, each equally consistent with guilt or nonguilt, the defendant is entitled to a factual inference of innocence. Taken with the court's extensive instructions in regard to the People's burden for conviction and the requirement of proof beyond a reasonable doubt, the court's charge, on the whole, arguably served to apprise the jury of the appropriate standard for conviction (see People v Fields, 87 NY2d 821, 823 [1995]; People v Maldonado, 220 AD2d 212, 213 [1995], lv denied 87 NY2d 904 [1995]). However, the additional instruction that the jurors should resolve factual inferences by a bare preponderance of the evidence ("50.1 to 49.9") was an unnecessary departure from the Criminal Jury Instructions that introduced the potential for confusion, thereby tending to dilute the prosecution's burden on the ultimate question of guilt.

Accordingly, the judgment of conviction should be vacated and the matter remanded to Supreme Court for a new trial.

Saxe, J. (concurring). Appellate courts reviewing jury instructions frequently take issue with portions of the charge without doing more than indicating disapproval. After all, we are required to consider errors in the context of the charge as a whole (see Cupp v Naughten, 414 US 141, 146-147 [1973]). Yet, there comes a time when repeated appellate admonitions have proven futile in preventing charge errors which have the effect of causing jury confusion. Since I believe such charge errors happened in this case, I conclude that a new trial is necessary here.

Trial judges must always be mindful that a jury is not left with a fundamental misapprehension of how to go about deciding a case. In a criminal case, nothing is more fundamental than the requirement that a defendant may only be convicted if each and every element of the crime is proved beyond a reasonable doubt. When there is a distinct possibility of confusion in this area, admonitions are not enough.

The portion of the jury instruction which is at issue here concerns the People's burden of proof. After instructing the jury on what constitutes evidence, the court went on as follows:

"[E]very defendant is entitled to every factual inference in his favor which can reasonably be drawn from the evidence. And where two factual inferences can be drawn from the evidence, both being of equal weight and strength, one factual inference consistent with guilt and the other factual inference consistent with innocence, any defendant is entitled to the factual inference of innocence.

"With regard to facts and with regard to the verdict, there are two different burdens of proof operating simultaneously. . . .

"With regard to finding of fact during the course of your deliberations, the burden of proof is simply that it is more likely than not that the fact exists, 50.1 beating 49.9."

Later in the charge, the trial court fully instructed the jury concerning the presumption of innocence and the People's burden of proving the elements of the charged crimes beyond a reasonable doubt. [*4]

At the conclusion of the court's charge, defense counsel stated while he understood that the "50.1 and 49.9" language applied to the evidence charge as opposed to the reasonable doubt charge, the jury might be confused. Counsel asked the court to reread the standard reasonable doubt charge on "to convict or not convict." The court denied counsel's request.

As my colleagues note, there is a fundamental problem with the portion of the above charge known as the "two inference" charge, in which the court instructed the jury that where two factual inferences can be drawn from the evidence, one consistent with guilt and the other consistent with innocence, the defendant is entitled to the factual inference of innocence. The problem with this charge as it was given in this case is that it fails to note that if the inference consistent with guilt is stronger than that consistent with innocence, yet not strong enough to prove the defendant's guilt beyond a reasonable doubt, acquittal is required. "This instruction erroneously allow[s] the jury to use the preponderance of the evidence standard in a criminal case" (United States v Hughes, 389 F2d 535, 537 [1968]; see also United States v Khan, 821 F2d 90, 93 [1987]; but see People v Tubens, 237 AD2d 170 [1997], lv denied 90 NY2d 898 [1997]). Its use, adding nothing and creating only the potential for misunderstanding, should be prohibited outright.

An equally disturbing related problem was created by the portion of the charge instructing that "[w]ith regard to finding of fact during the course of your deliberations, the burden of proof is simply that it is more likely than not that the fact exists, 50.1 beating 49.9." As defense counsel protested, this instruction is too confusing, and can only lead to the possibility of the jury finding elements of the crime based upon a preponderance of the evidence standard.

A degree of support for this portion of the charge is concededly found in United States v Viafara-Rodriguez (729 F2d 912 [1984]), which stated that the burden of proof beyond a reasonable doubt "does not operate upon each of the many subsidiary facts on which the prosecution may collectively rely to persuade the jury that a particular element has been established beyond a reasonable doubt" (id. at 913, citing United States v Valenti, 134 F2d 362, 364 [1943], cert denied 319 US 761 [1943]). But, despite the technical accuracy of this assertion, it serves no useful function within a jury charge, and indeed, can only serve to confuse the jury as to the degree of proof required for the different findings of fact that together lead to the verdict. As the Second Circuit remarked in Viafara-Rodriguez in elaborating on the complex range of ways in which burdens of proof operate in criminal cases, "[t]he range of such matters suggests the hazard of attempting to tell the jury anything more than it needs to know in a particular case. That the burden [of proof beyond a reasonable doubt] applies to the defendant's guilt and to every element necessary to establish guilt will normally be sufficient" (729 F2d at 913).

There may be particular circumstances in which it is necessary to explain to the jury that certain types of "subsidiary" facts need only be decided by a preponderance of the evidence: for instance, if the jury was to ask whether an ancillary witness's credibility must be established beyond a reasonable doubt before his testimony is relied upon (cf. Justice v Hoke, 45 F3d 33, 35 [1995]). However, in the absence of a particular question, or some other unique circumstance making such an instruction important, giving this "subsidiary facts" charge in the context of the [*5]general closing charge to the jury is a pernicious practice, having no positive impact and being very likely to serve only to confuse the jury.

Moreover, even if we were to approve inclusion in a jury charge of a lesser burden of proof regarding "subsidiary facts," the language employed by the trial court here did not suffice to correctly convey the proper standard. Instead, the court, without explaining the distinction between "facts" and the "verdict," or even using the term "subsidiary facts," simply instructed that "[w]ith regard to facts and with regard to the verdict, there are two different burdens of proof operating simultaneously." This cannot be permitted to stand. We cannot allow a jury charge to refer to the use of the preponderance standard without an explanation of which types of facts may be proven by a preponderance of the evidence, and which, such as the actual commission of the act constituting the crime, must be proved beyond a reasonable doubt.

For all the foregoing reasons, I conclude that defendant's conviction must be reversed and the matter remanded for a new trial.