concurring in part and dissenting in part.
The Court today reverses defendant’s capital murder conviction and death sentence. In doing so, the Court determines that a defendant can be considered death-eligible only if he murdered with the intent to kill, and only if he committed the murder by his own conduct. This narrowing of the scope of the class of death-eligible murders is a significant step toward remedying the constitutional infirmities that burden the capital murder-death penalty statute. Of corollary significance, the Court, in departing as it does from the United States Supreme Court’s decision in Tison v. Arizona, 481 U.S. 137, 107 S.Ct. 1676, 95 L.Ed.2d 127 (1987), recognizes the unreliability and inadequacy of federal precedent in the development of capital-murder jurisprudence.
I subscribe to the Court’s reasoning on these essential points. Moreover, while I concur in the judgment of the Court, I would also reverse because the New Jersey capital murder-death penalty statute is unconstitutional and invalid, both as enacted and as applied. See, e.g., State v. Ramseur, 106 N.J. 123, 343 (1987) (Handler, J., dissenting). I feel constrained to maintain and repeat this position because of the evolving and unsettled nature of the law governing these unique capital-murder prosecu*145tions. See, e.g., State v. Rose, 112 N.J. 454 (1988) (Handler, J., dissenting); State v. Bey (II), 112 N.J. 123 (1988) (Handler, J., dissenting).
I write separately in this case to address two continuing troublesome issues, the status of “knowing” murder as a capital offense and the role of prosecutorial discretion in capital-murder prosecutions. The resolution of these two issues requires, I submit, the exclusion of “knowing” murder, one that lacks intent to kill, from the class of death-eligible murders; it demands also uniform, statewide standards to govern the prose-cutorial discretion in determining whether a case should properly be prosecuted as a capital offense.
I.
The Court now holds, as a matter of state constitutional law, that the intent to kill is an essential element of capital murder, distinguishing capital murder from other murders. Moreover, the imposition of a death penalty for intentional infliction of serious bodily injury that results in death is disproportionate. The majority now rules that this violates the State Constitution. Ante at 88-89. I agree.
In the context of this case, this important holding does not rest on the sufficiency or insufficiency of the evidence of intentional murder. As the Court aptly notes, the State’s evidence in this case was, arguably, sufficient to support a jury’s determination that the defendant murdered the decedent with the intent to kill. We cannot, however, have confidence in that interpretation of the jury’s verdict. The confusion in the prosecution of this case about what type of conduct could constitute purposeful or knowing murder and what might constitute aggravated manslaughter has rendered the jury’s verdict suspect. The inclusion of the serious bodily harm provision in the definition of capital murder and in the charge to the jury may have influenced the jury’s deliberations in this case because the evidence of defendant’s intentions was inconclusive *146and because this definition may have de-emphasized the jury’s consideration of a non-capital offense, aggravated manslaughter.
I also concur in the majority’s determination to find that the “own conduct” requirement must be met by showing that “the defendant actively and directly participated in the homicidal act.” Ante at 97. I believe that the interpretation clearly reflects the Legislature’s intent to narrow the class of death-eligible murderers by excluding those who would, pre-Code, have been considered accomplices. I commend this narrowing of the class of death-eligible murders and the ruling that this determination must be made by the jury at the guilt phase of its deliberations. As I have repeatedly stressed, the determination of whether the particular homicide falls within the class of death-eligible murders must be made a constituent part of the process in the determination of guilt, and it must be done apart from and before the jury is thereafter directed to consider whether the death sentence is appropriate for the particular defendant. See State v. Ramseur, supra, 106 N.J. at 387-94 (Handler, J., dissenting); State v. Bey (II), supra, 112 N.J. at 215-216 (Handler, J., dissenting).
I must, however, stress what may be more than a semantical difference; unlike the majority, I view the own-conduct requirement as a material element of capital murder rather than as a mere “triggering device.” Ante at 99 (quoting State v. Moore, 207 N.J.Super. 561, 576 (Law Div.1985)). The Court, in adopting the reasoning of Moore, determines that the commission of a homicidal act by the defendant’s own conduct is not an element of the offense of murder. There is, however, nothing in the language of N.J.S.A. 2C:ll-3 suggesting that the own-conduct requirement is anything less than a material element of the offense of capital murder. This requirement assuredly satisfies the Code’s understanding of “a material element” of a crime as a requirement that relates solely to the crime itself, N.J.S.A. 2C:l-14(i), and includes “conduct” that is a part of the “definition of the offense” and also “establishes the required *147kind of culpability,” N.J.S.A. 2C:l-14(h)(l)(a), (b). Since a defendant cannot be convicted of “capital murder” unless it is established that the homicide was by his “own conduct,” there is no escape from the conclusion that this requirement is a “material element” of the crime of capital murder. Hence, the “own-conduct” aspect of capital murder is no more, no less a material element of capital murder than the requirements involving “purpose” or “knowledge.” What follows from this is that the jury must understand clearly that the own-conduct requirement, as a material element of capital murder, is subject to the State’s continuing burden to prove each and every element of the charged crime beyond a reasonable doubt.
The narrowing of the class of death-eligible murders thus achieved by the Court is both essential and constructive. It will go far toward focusing the aim of the capital murder-death penalty statute on the truly egregious murderer, and it will significantly decrease the risk of an arbitrary and disproportionate imposition of the death penalty. Unfortunately, by failing to exclude “knowing” murder from the class of death-eligible murders, the majority fails to cohere the implications of its well-reasoned holding that the culpable state of mind of “intent to kill” is constitutionally indispensable to establish capital murder. In my opinion, the implication of the majority’s decision that the State Constitution requires a showing of “intent to kill” for capital murder is that murders committed only with knowledge that death is “practically certain” to result from the acts of the defendant cannot rise to the level of capital murder. Such murders do not, in my view, reflect the greatest level of culpability that would justify the greatest possible penal sanction, namely, death.
This constitutional perception of capital murder would be consistent with New Jersey’s prior capital-murder statute in which the death penalty was not authorized for crimes less severe than first-degree murder. See State v. Ramseur, supra, 106 N.J. at 387-90 (Handler, J., dissenting). Under the former statute, death-eligible murders were those that were *148“perpetrated by means of poison, or by lying in wait, or by any other kind of willful, deliberate and premeditated killing, or [were] committed in perpetrating or attempting to perpetrate certain felonies.” See L. 1965, c. 212, § 1; R.S. 2:138-2; L. 1917, c. 238, § 1, p. 801 (1924 Supp. § 52-107); L. 1898, c. 235, § 107, p. 824 (C.S. p. 1780, § 107). The former statute’s focus on intentional murder required that with the exception of murders during the course of certain felonies and the murder of a law enforcement officer, the State was required to prove three mental operations in order to establish first-degree murder: premeditation, deliberation, and willful execution of the plan. See State v. Anderson, 35 N.J. 472, 496-97 (1961). All other murder was presumptively second degree murder — a non-capital offense — regardless of the circumstances of the murder. Indeed, under our prior statute even most first-degree murder defendants were not subjected to a death sentence, but could under prevailing practice exercise a non vult plea. See discussion infra at 154.
By defining first-degree murder in terms of culpability as manifested through mental operations demonstrating pre-medi-tation or intent, the former statute expressed one of the goals of the new code: “greater individual justice through a closer relation between guilt and culpability ...” Knowlton, Comments Upon The New Jersey Penal Code, 32 Rutgers L.Rev. 1, 2 (1979). The relationship between culpability, guilt, and the appropriate level of punishment is squarely recognized by this Court. Indeed, the majority explicitly endorses the Supreme Court’s statement that “[d]eeply ingrained in our legal tradition is the idea that the more purposeful the conduct, the more serious is the offense, and, therefore, the more severely it ought to be punished.” Tison v. Arizona, supra, 481 U.S. at 156, 107 S.Ct. 1687, 95 L.Ed.2d at 143. Ante at 77. This Court has also stated that “[o]ur system of criminal laws is predicated usually on the imposition of punishment based on the defendant’s intent. Indeed, our Code’s ranking of crimes by degree places those crimes committed with intentional con*149duct as the highest degree of crime, for which the defendant is most severely punished.” State v. Ramseur, supra, 106 N.J. at 207-08. Ante at 77. The legislature itself effectuated this principle through adoption of the separate levels of culpability described in N.J.S.A. 2C:2-2(b). In applying this principle to criminal homicide, “[p]urposeful murder, knowing murder, aggravated manslaughter and reckless manslaughter are criminal homicides that lie on a descending scale of culpability.” State v. Grunow, 199 N.J.Super. 241, 250 (App.Div.1975), aff'd, 102 N.J. 133 (1986). These considerations all point to one basic principle: the severest sanctions should be reserved for actors exhibiting the most culpable mental states.
Although “purposely” and “knowing” are often yoked together relative to other culpability states, they nevertheless are two different measures of criminal intent. In the context of criminal homicide, “purposeful” commission of murder, N.J.S.A. 2C:2-2(b)(1), is the most culpable form, closely corresponding to the former mens rea requirement of premeditation, i.e., intent to kill. Knowing murder, however, a less culpable state than purposeful murder, contains no requirement of premeditation or willfulness and deliberation, the mental states or qualities that historically have justified the ultimate sanction of the death penalty.
The Court, in refusing to winnow out knowledge from capital murder, in effect, homogenizes distinctly different states of criminal culpability. While, in terms of culpability, the Court’s result seemingly equates only purpose and knowledge, it fails to remove recklessness from the equation. Thus, in weighing culpability, what is disquieting about the Court’s rejection of any constitutional distinction between purpose and knowledge is its failure to understand that it thereby obliterates the distinction between knowledge and recklessness.
Further, in terms of criminal culpability, the Court’s determination permits recklessness to be equated with purpose, which is the critical characteristic of the most serious form of murder. *150This is clearly what the Supreme Court did in Tison when it adopted this philosophy, viz:
[T]he reckless disregard for human life implicit in knowingly engaging in criminal activities known to carry a risk of death represents a highly culpable mental state, a mental state that may be taken into account in making a capital sentencing judgment when that conduct causes its natural, though also not inevitable, lethal result.
[ 481 U.S. at 157-158,107 S.Ct. at 1688, 95 L.Ed.2d at 144.]
Ostensibly the Court repudiates this philosophy when' it states disapprovingly that “as a matter of federal proportionality principles ... capital punishment may be imposed on one who ... can be characterized as ‘recklessly indifferent to human life.’ ” Ante at 75. Nevertheless, by including “knowing” murder in the class of capital murder, the Court does not truly separate itself from the “federal proportionality principles” that it professes to reject.
There was not any genuine disagreement with the observation made recently, that “[t]he distinction between knowledge and recklessness, between ‘practical certainty’ of a result and ‘conscious disregard’ of ‘a substantial and unjustifiable risk’ of a result, is a subtle one at best,” and that “the distinction between ‘knowing’ murder and ‘aggravated manslaughter’ turns in close cases ... on the difference between ‘practical certainty’ that one is inflicting injury with a substantial risk of death and a conscious disregard of a substantial risk of death that manifests ‘extreme indifference to the value of human life.’ ” State v. Rose, supra, 112 N.J. at 562-563 (Handler, J., dissenting). Because the distinction between knowledge and recklessness is so nebulous, I do not believe it is possible to include a non-intentional but knowing murder as capital murder without creating an intolerable risk that reckless murder will also become capital murder.
Aggravated manslaughter now exists as a lesser-included offense of knowing murder. See State v. Crisantos (Arria-gas), 102 N.J. 265 (1986). Knowing murder itself can incorporate a degree of “indifference” that can, on a given state of facts, serve to make aggravated manslaughter the functional *151equivalent of knowing murder. See State v. Palmer, 211 N.J.Super. 349, 352 (App.Div.1986) (endorsing a definition of the element of aggravated manslaughter of “circumstances manifesting extreme indifference” as conduct “practically certain to kill any one who might happen to be in the way.” (Emphasis added)). Conversely, on identical evidence, aggravated manslaughter can encompass a quality of indifference that could transform knowing murder into aggravated manslaughter. See State v. Rose, supra, 112 N.J. at 562-567 (Handler, J., dissenting).
In stressing the similarities between intentional murder and aggravated manslaughter, Justice O’Hern in his concurring opinion observes that the latter requires “[t]hat [the] actor, realizing death was rather probable, and fully aware of that, nevertheless committed the act with a state of mind that said T couldn’t care less, I don’t care at all, I don’t care in the least bit if this person dies, even though I know that my act possesses a high degree of probability of causing that person’s death.’ ” Ante at 139. The Justice goes on to say: “[t]he conduct of the actor could be as broad as the scope of human behavior” but, to constitute capital murder that conduct “would have to include” a “state of mind that is almost, but not quite, the same as purposely causing death or knowingly causing death____” (emphasis added). Ibid. The “almost, but not quite the same” test might suffice and be acceptable to distinguish murder from manslaughter when the penal consequences entail only incarceration. That test, however, cannot, in my opinion, be used to determine degrees of criminal culpability to justify death as the just penalty for a homicide.
These offenses, knowing murder and aggravated manslaughter, are couched in terms that in any given case can render the two indistinguishable. See New Jersey Penal Code, Volume II: Commentary, Final Report of the New Jersey Criminal Law Revision Commission, 1971 (quoted in Cannel, New Jersey Criminal Code Annotated (1987), at 245-46). Hence, the inclusion of knowing murder as a capital offense creates a *152system that will of necessity function arbitrarily and irrationally because it cannot reliably or consistently exclude homicides that may constitute only aggravated manslaughter and, in terms of underlying criminal culpability, are not truly different from aggravated manslaughter.
Moreover, it does not appear that the inclusion of knowing murder as a form of capital murder was itself expressly or impliedly intended by the Legislature. I have noted the omission in the enactment of the capital murder-death penalty statute of any explanation of why aggravated manslaughter is now an offense distinguishable from knowing murder. State v. Rose, supra, 112 N.J. at 563 (Handler, J., dissenting). Further, support for the conclusion that knowing murder was not intended to constitute capital murder can be derived from the statement of the sponsor of the capital murder legislation that this State’s capital murder statute is “not as broad” as statutes in some other states, in part because defendants are death-eligible only if “found guilty ... of first-degree murder, willful, premeditated murder,” a degree of murder closely corresponding only with a “purposeful” state of mind. Capital Punishment Act: Hearing on S.112 Before the Senate Judiciary Committee (1982) at 1.
In sum, as noted, the Court rejects Tison v. Arizona, supra, 481 U.S. 137, 107 S.Ct. 1676, 95 L.Ed.2d 127, ante at 75-77 in which the Supreme Court in effect ruled that it was not essential that the State show intent to kill to obtain a conviction for capital murder. Id. at 155, 107 S.Ct. at 1687, 95 L.Ed.2d at 144. Indeed, according to the Supreme Court, it was sufficient to show that a defendant participate in a murder only with the knowledge that there was a risk of death. Ibid. The Court, in accepting a death-penalty scheme that includes non-intentional but knowing murder as a capital murder, thus indirectly endorses the federal philosophy of capital-murder culpability that it purports to repudiate, a philosophy that does not blanche at» the inclusion of reckless homicide as a capital offense. In its prior decision, Enmund v. Florida, 458 U.S. 782, 102 S.Ct. 3368, 73 *153L.Ed.2d 1140 (1982), the Supreme Court insisted that capital murder required minimally an intent to kill, that this constituted the “highly culpable mental state” required to convert ordinary murder into capital murder. I believe therefore that this Court, in approving Enmund and disapproving Tison, should acknowledge the sound conceptual and policy basis for a legal understanding of capital murder that requires the intent to kill.
I conclude, consistent with the underlying premises of the majority opinion and with New Jersey’s capital murder traditions that as a matter of state constitutional doctrine “knowing” murder should not be included in the class of the death-eligible murders.
II.
A significant issue in this case, not addressed by the majority, involves the twin concerns of unfettered prosecutorial discretion and disproportionality of sentence. Most recently I commented on this in my dissenting opinion in State v. Koeda-tich, 112 N.J. 225, 265-271 (1988). These concerns are closely related to the overly broad death-eligible class of murders attributable to both the lack of a narrow definition of this class and to the problematic application of aggravating factors. The absence of uniform standards governing prosecutorial discretion heightens the uncertainty and inconsistency in the administration of the capital murder statute. Derivatively, it loosens the guidelines, complicates immeasurably the discretionary responsibilities of the jury, and inevitably compounds the risk of arbitrary and capricious death sentences.
This uncertainty and arbitrariness is exacerbated by unstructured plea-bargaining practices. Here, three young men with equal culpability participated in the murder of Paul Matusz. Two of them escaped trial and received custodial terms by pleading guilty to some of the charges brought against them. The defendant did not; because of this, he is the only one who *154may have to pay for this shared crime with his life. This demonstrates that the availability of guilty pleas on a random basis and the lack of uniform standards to control prosecutorial discretion engender an intolerable degree of sentence dispropor-tionality.
A.
What occurred in this case — the plea to lesser, non-capital offenses by co-defendants — is reminiscent of the practice found unconstitutional under our prior capital murder statute. The former New Jersey murder statute was designed so that a plea of non vult removed the death penalty from the range of sentences that could be imposed on the defendant. While this statute was in effect, however, the Supreme Court asserted the unconstitutionality of a federal statute that “needlessly encouraged” guilty pleas or jury waivers by forcing defendants to choose between pleading guilty, thus assuring nothing worse than imprisonment, or asserting their right to contest guilt or the degree of guilt, thereby incurring the possibility of the death penalty. See United States v. Jackson, 390 U.S. 570, 583, 88 S.Ct. 1209, 1217, 20 L.Ed.2d 138, 148 (1968).
We were slow to recognize the constitutional infirmity of this practice. In 1968, in State v. Forcella, 52 N.J. 263, 274-81 (subsequent history omitted), we ruled that neither the death penalty nor the non vult plea provision of our statute was unconstitutional. However, three years later in Funicello v. New Jersey, 403 U.S. 948, 91 S.Ct. 2278, 29 L.Ed.2d 859, reh’g den., 404 U.S. 876, 92 S.Ct. 31, 30 L.Ed.2d 125 (1971), the Supreme Court struck down New Jersey’s death penalty in a summary disposition that cited Jackson. This Court then declared the New Jersey death penalty unconstitutional, finding that it fell within the area that Jackson had found unconstitutional (although the non vult plea was continued after 1972, defendants were simply sentenced to life imprisonment). See State v. Funicello, 60 N.J. 60, 67-68 (per curiam), cert. den. *155sub nom. New Jersey v. Presha, 408 U.S. 942, 92 S.Ct. 2849, 33 L.Ed.2d 766 (1972).
Funicello recognized that Jackson had been narrowed by the Supreme Court’s intervening decision in Brady v. United States, 397 U.S. 742, 90 S.Ct. 1463, 25 L.Ed.2d 749 (1970). Brady held that such “dilemma” statutes did not violate a defendant’s fifth or sixth amendment rights if the non vult or guilty plea was intelligently and voluntarily made, even if the defendant faced a possible death penalty by contesting guilt. See also North Carolina v. Alford, 400 U.S. 25, 37-38, 91 S.Ct. 160, 167-68, 27 L.Ed.2d 162, 171-72 (1970) (upholding statute in which defendant , received a lesser penalty in exchange for a guilty plea). The Brady majority did not say whether death could be so significant a motivating factor that the fifth amendment was violated, but indicated that the defendant bore a very heavy burden. See Brady, 397 U.S. at 747, 90 S.Ct. at 1468, 25 L.Ed.2d at 755, approving and citing, from Jackson, supra, Laboy v. New Jersey, 266 F.Supp. 581 (D.N.J.1967) (guilty plea held voluntary despite the fact that defendant was greatly upset by prospect of the death penalty).
The federal court’s decision in Brady does not limit the protections granted by State constitutional sources. Some states have recognized that the prosecutor’s use of the power to bargain for guilty pleas in capital murder prosecutions can lead to arbitrary and unjustifiable results. See, e.g., Commonwealth v. Colon-Curz, 393 Mass. 150, 163-64, 470 N.E.2d 116, 124 (1984) (state’s newly enacted death penalty law violated the state constitution because the death penalty could only be imposed after a jury trial, thus coercing guilty pleas and impermissibly violating a defendants’ rights to demand a jury trial and against self-incrimination); State v. Frampton, 95 Wash.2d 469, 627 P.2d 922 (1981) (where, pursuant to statute, the death penalty is imposed upon conviction following a plea of not guilty, but is not imposed when there is a plea of guilty, that statute is unconstitutional); Spillers v. State, 84 Nev. 23, 436 P.2d 18 (1968) (death penalty scheme unconstitutional be*156cause it impermissibly burdened the defendant’s right to trial by jury since death penalty could be imposed as a result of jury verdict of guilty of rape).
Following the invalidation of the death penalty provisions of the former statute, this Court dealt with the issue of whether the former statute still unnecessarily coerced non vult pleas from defendants guilty only of manslaughter. In State v. Corbitt, 74 N.J. 379 (1977), aff'd, Corbitt v. New Jersey, 439 U.S. 212, 99 S.Ct. 492, 58 L.Ed.2d 466 (1978), the petitioner unsuccessfully argued, on the basis of Jackson, that any system under which the defendant received a lesser penalty in exchange for a guilty plea was unconstitutional. This Court noted that “[t]he criminal process ... is replete with situations requiring ‘the making of difficult judgments’ as to which course to follow,” Corbitt, supra, 74 N.J. at 398 (quoting McGautha v. California, 402 U.S. 183, 213, 91 S.Ct, 1454, 1470, 28 L.Ed.2d 711 (1971)), and that such dilemmas were never thought to invade constitutional rights.
Significantly, however, the Court in Corbitt was considerably less sanguine about the dilemma when one choice involved the death penalty. The Court concluded:
Finally, it is our considered view that Jackson today is authority only for a situation where a defendant faces the prospect of a possible death sentence if convicted as against the alternative of merely a prison term if he pleads guilty. We have hereinabove stressed the language of Jackson emphasising the awful pressure exerted upon a defendant whose choice of a course of action may mean his death. At the time of Jackson, the Court was undoubtedly already sensitive to the considerations which were to lead to the later broad invalidation of death penalties in Furman v. Georgia, 408 U.S. 288, 92 S.Ct. 2726, 88 L.Ed.2d 346 (1972). Cf. Whitherspoon v. Illinois, supra, 391 U.S. 510, 88 S.Ct, 1770, 20 L.Ed.2d 776 [1968],
In summary of the foregoing, we conclude: (1) the authority of Jackson is confined to the case of a choice between trial and plea of guilty when a possible consequence of the former, and only of the former, is the death penalty; (2) alternatively, if the authority of Jackson goes beyond situation (1), any encouragement or inducement of a defendant to plea non vult under our statute and thereby waive his right to trial is not the result of a needless or unnecessary procedural device but a highly useful and desirable one, and the encouragement *157is therefore not an impermissible infringement of defendant’s Fifth Amendment rights.
[ 74 N.J. at 399-400.]
Although this Court was interpreting and limiting Jackson, supra, it ignored Brady’s subsequent narrowing of Jackson in the death penalty context. On appeal, the United States Supreme Court, in a 6-8 decision, affirmed Corbitt, noting that:
[although we need not agree with the New Jersey court that the Jackson rationale is limited to those cases where a plea avoids any possibility of the death penalty ... it is a material fact that under the New Jersey law the maximum penalty for murder is life imprisonment, not death.
[Corbitt v. New Jersey, supra, at 217, 99 S.Ct. at 496, 58 L.Ed.2d. at 473.]
The undue pressure in the context of a capital case involving the death penalty that the Court acknowledged and criticized in Corbitt was present in this case. It was indeed used by the prosecutor to obtain the guilty pleas of the co-defendants. Facing the death penalty in a jury trial, they assured themselves of custodial sentences by pleading guilty. However, the vice is not that the guilty pleas of these co-defendants were unconstitutionally coerced, but rather that the sentences that have emerged from this prosecution are so unequal and disparate: the plea bargain results directly in custodial terms for the co-defendants and indirectly in the death sentence for the defendant. This practice is of deep concern because there are no standards whatsoever to guide the prosecutor, and thus the courts, in determining whether what occurred here is fair, consistent, and rational. In the absence of guided discretion, the result strongly bespeaks of intolerable disproportionality.
B.
This case is further illustrative of the propensity toward disproportionality that inheres in uncontrolled prosecutorial discretion. This concern increases in this case because of the borderline nature of the evidence supporting a capital murder prosecution. The perception of disproportionality is strengthened by the growing empirical support from statistical evidence *158that the death penalty is being administered in an arbitrary and capricious manner.
In several capital murder appeals, the Public Defender has brought to our attention statistical evidence gathered in analyzing the administration of the capital murder-death penalty statute. This study, entitled “The Reimposition of Capital Punishment in New Jersey: Homicide Cases from 1982-1986,” by Leigh Bienen, Assistant Deputy Public Defender, and Neil Alan Wiener, Senior Research Associate at the Sellin Center for Studies in Criminology and Criminal Law at the University of Pennsylvania (hereinafter Study or Report), was acknowledged by the Court in State v. Koedatich, supra, 112 N.J. at 256. The Study is instructive. Among other things, it points out a clear differential in prosecutorial practices in our various counties. According to the Public Defender’s Study, the county-to-county differences suggest that a “death-possible” case, one in which at least one statutory aggravating factor could have been alleged, was much more likely to be prosecuted as death-eligible in certain counties than in others.1 Of particular relevance to this case is the suggestion that the county-to-county disparity also relates to plea-bargaining practices.2 This correlation, if true, between pursuing prosecution aggressively while offering few plea bargains, and prosecuting relatively few death-possible homicides while settling many by means of a plea offer, suggests that homicides are being prosecuted not solely on the basis of the nature of the crime and the defendant, but also on the basis of *159political or budgetary or moral considerations that vary from county to county. See discussion, infra 166. In the absence of clear uniform standards governing capital murder prosecutions, such unchecked discretion will continue to lead to unacceptable and anomalous results.
The Public Defender’s Study is an indication that the absence of controlled prosecutorial discretion poses real, not imaginary, risks. The Study provides us at least initially with a reasonably objective framework within which to think about and confront the complex issue of uncontrolled prosecutorial discretion as an aspect of proportionality.3 The Study may have shortcomings, but the data, particularly with reference to county-to-county *160discrepancies, point to the need for controls on prosecutorial discretion. See State v. Koedatich, supra, 112 N.J. at 269-270 (Handler, J., dissenting).
The Study uses a data base of 703 homicides;4 of the 703, prosecutors served a notice of aggravating factors
in cases involving 131 defendants (18.6 percent of the 703), for 94 of whom ... the case went to trial as a capital case before a judge or jury (the factor was neither dismissed by the judge nor withdrawn by the prosecutor). Of these 94 cases, 69 (73.4 percent) resulted in a capital conviction for death eligible murder followed by a penalty phase trial before a judge or jury. Of these 69 cases, 25 (36.2 percent) resulted in the death sentence being imposed. Overall, of the 703 homicide cases, 18.6 percent [131 cases] were recommended for the death penalty by the prosecutors serving a formal notice of statutory aggravating factors (cases were designated death-eligible), 13.4 percent (94 cases) resulted in a capital trial, 9.8 percent (69 cases) went to penalty phase ... and 3.6 percent ... resulted in the death sentence being imposed.
Study, Interim Report Part I, “Characteristics of the New Jersey Homicide Cases” at 5. The Public Defender’s data also purport to show, however, that 404 of the '703 cases (57.5 percent) were death-possible because there was at least one statutory aggravating factor that could have been alleged. Only 131 of the *161404 death-possible cases (82.4 percent) were charged as capital murder prosecutions.5
Thus, according to this data, there are 278 eases in which the prosecutor had a factual basis for seeking the death penalty but declined to do so. The Study suggests that one factor that correlates with the decision to prosecute is the policy of the prosecutor of the county in which the prosecution is to occur; these policies differ from county to county. Another factor is the race of various classes of victims and defendants. See discussion infra at 168-164.
The Report does suggest that county prosecutors have adopted differing philosophies; as a result, there are no uniform prosecutorial standards. The lack of prosecutorial uniformity in the absence of uniform statewide standards should not be surprising. The issue was discussed at length in State v. Willie E. Smith, 202 N.J.Super. 578 (Law Div.1985), in which the court specifically challenged the county prosecutor to explain the basis on which discretionary decisions were made to prosecute capital cases. The prosecutor’s explanation was less than satisfactory, and serves to highlight the need for greater control over prosecutorial discretion.6
*163The concerns deriving from the statistical disparities among counties within this State is heightened in this case because defendant was prosecuted by the Atlantic County Prosecutor’s Office. The Study suggests that with the exception of Monmouth County, this Office is far more likely to serve a notice of aggravating factors in a death-possible case than any other county in the State. Close to 60 percent (59%) of death-possible cases became death-eligible in Atlantic County. Study, Interim Report Part I, at TABLE CP4. In comparison, consolidation of the relevant statistics for Essex, Hudson, Camden, Mercer, and Passaic Counties reveals that in these five counties an average of only 25 percent of death-possible cases became death-eligible. Id. This seemingly wide disparity, coupled with the arguably tentative nature of the evidence supporting the capital-murder prosecution in this case, raises serious concerns of arbitrariness and disproportionality.
C.
The borderline nature of the evidence supporting capital murder prosecution in this case, as well as the lack of standards to control the prosecutor’s discretion, is even more troubling in *164light of the fact that defendant is black and his victim white. I acknowledge that there is no actual evidence of any racial bias or invidious discrimination in this case. Nevertheless, the statistical evidence indicates that we should not disregard the potential for racial bias.
As noted in State v. Koedatich, supra, 112 N.J. at 266-267 (Handler, J., dissenting), the Supreme Court in McCleskey v. Kemp, 481 U.S. 279, 107 S.Ct. 1756, 95 L.Ed.2d 262 (1987), rejected petitioner’s claim, based on statistical studies, that the Georgia capital punishment system violates equal protection guarantees and the eighth amendment by discriminating in application according to the race of the victim.7 The Supreme Court felt that because the studies showed no discrimination in McCleskey’s individual case, this foreclosed his equal-protection claim. I strongly disagree that we should be influenced or persuaded by the federal Court’s analysis. State v. Koedatich, supra, 112 N.J. at 267 (Handler, J., dissenting). Our State Constitution, in light of our pronounced and consistent concern about the subtle evils of discrimination, demands the most scrupulous assessment of such claims. State v. Rasmeur, supra, 106 N.J. at 426 (Handler, J., dissenting).
The Public Defender’s Study indicates that the race of the victim may be of some significance in whether a case is prosecuted as a capital-murder offense, with cases involving a white victim more likely to progress to the death-eligible class than those involving black or Hispanic victims.8 The data, according to the Public Defender, also suggests that cases are designated *165death-eligible in a manner that is influenced by the race of the defendants and victims, with a black defendant/white victim homicide more likely to be designated death-eligible than a white defendant/white victim homicide.9 Indeed, the Interim Report of the Public Defender concluded its discussion of the racial statistics as follows:
At this initial point of examination of case progression probabilities, there is some indication, then, of differences in the comparative probabilities that homicide defendants will move more deeply into the New Jersey capital case processing system based on the race characteristics of the defendant and the victim.
[Study, Interim Report Part I, “Research Findings” at 10.]
These statistical differences, indicating some correlation between the race of victim and defendant and the decision by a prosecutor to serve a notice of aggravating factors in a death-possible case, are of greater concern in this case because the statistical disparity in Atlantic County is more pronounced than it is statewide. Supra at 163-64. Although the information is tentative, the study reveals that all (100%) of the death-possible cases involving a black defendant/white victim pattern became death-eligible, whereas only 25% of those involving a black defendant/black victim or 20% of those homicides involving a white defendant/white victim became death-eligible. Study, *166Interim Report Part I, at TABLE CP11B. As noted, there are difficulties with the data presented, particularly in that the critical determination of whether a ease was “death-possible” is heavily dependent on information furnished by defense counsel. Supra at 161 n. 5. Nevertheless, in light of uncontrolled prosecutorial discretion and the arguably tentative nature of the evidence supporting capital murder in the instant case, this rather emphatic difference must cause apprehension.
Most important, the question of prosecutorial discretion as it relates to the potential for racial bias cannot be utterly disregarded or discounted. Dissenting in McCleskey v. Kemp, supra, Justice Blackmun stated that “[t]he Court’s emphasis on the procedural safeguards in the system ignores the fact that there are none whatsoever during the crucial process leading up to trial,” 481 U.S. at 366, 107 S.Ct. at 1805, 95 L.Ed.2d at 326, implying that the substantial discretion afforded prosecutors during the process prior to trial might be the cause of the racial disparities shown in the statistics before the Court. As explained in Gross & Mauro, “Patterns of Death: An Analysis of Racial Disparities in Capital Sentencing and Homicide Victimization,” 37 Stan.L.Rev. 27, 106-07 (1984): “Since death penalty prosecutions require large allocations of scarce prosecu-torial resources, prosecutors must choose a small number of cases to receive this expensive treatment. In making these choices they may favor homicides that are visible and disturbing to the majority of the community, and these will tend to be white-victim homicides.” See McCleskey, supra, 481 U.S. at 294 n. 13, 107 S.Ct. at 1803 n. 13, 95 L.Ed.2d at 322-23 n. 13 (Blackmun, J., dissenting). Thus, I would urge the Court, in the context of this case, to seriously examine the problem of prosecutorial discretion as it may relate to the difficult issue of racial disparity.
D.
Prosecutorial discretion in this case is troublesome not only with respect to race, but also with respect to the apparent lack *167of proportionality evinced by the dramatically differing sentences received by Gerald and his two co-defendants. In the ordinary criminal setting, we can countenance some degree of sentencing disparity. In the context of capital sentencing, however, a system that affords prosecutors the leverage to force defendants to choose between imprisonment and a possible death sentence is no improvement over the prior unconstitutional system that forced defendants to choose between a non vult plea and a possible death sentence. See State v. Funicel-lo, supra, 60 N.J. 60 (1972).
In State v. Ramseur, I expressed the view that under state constitutional and fundamental fairness doctrines our capital murder-death penalty statute did not provide sufficient guidance to overcome the genuine risk of arbitrary and capricious-applications. 106 N.J. at 405-06. I pointed out that the global statutory definition of murder and the fusion of the guilt and sentencing determinations, with the simultaneous application of statutory aggravating factors to determine both death eligibility and the death sentence, robbed the statute of coherence and objectivity. Ibid. I emphasized that the absence of guided prosecutorial discretion and mandatory effective proportionality review exacerbated the risk of arbitrariness. Ibid. I reiterated these views in State v. Koedatich, supra, 112 N.J. at 264-271.
I continue to urge that guided prosecutorial discretion is integral to a valid death penalty scheme. As was noted by Justices Blackmun, Marshall, and Stevens in dissent in McCles-key v. Kemp, supra, “the establishment of guidelines for Assistant District Attorneys as to the appropriate basis for exercising their discretion as to the various steps in the prosecution” is needed in the interests of consistency. 481 U.S. at 365, 107 S.Ct. at 1805, 95 L.Ed.2d at 325-26 (Blackmun, J., dissenting). In the absence of a comprehensive recasting of the statute to overcome its current constitutional defects, guided prosecutorial discretion will serve at least to reduce, even if it cannot eliminate, the risk of arbitrary application and disproportionate sentencing.
*168III.
I join in the Court’s decision to reverse defendant’s conviction and death sentence. I endorse its reasoning that under our State Constitution it is impermissible to permit an offender to be convicted of capital murder and sentenced to death without the intent to kill. I therefore agree that one whose state of mind is only to cause serious bodily injury that proves fatal should not be exposed to the death sentence. I also concur in the Court’s interpretation of the “own conduct” requirement. I would, moreover, make expressly certain that under the State Constitution only offenders who intend to kill, not merely kill with knowledge of the fatal consequences of their acts, should be exposed to the death sentence.
In addition, I emphasize that the Court should find the capital murder-death penalty statute unconstitutional because of the lack of uniform standards to guide prosecutorial discretion and the failure to develop such standards in order to complement and assure proportionality of sentencing.
Accordingly, I concur in part and dissent in part.
O’HERN, GARIBALDI and HANDLER, JJ., concurring in the result.
For reversal and remandment — Chief Justice WILENTZ, and Justices CLIFFORD, HANDLER, O’HERN, POLLOCK, GARIBALDI and STEIN — 7.
For affirmance — None.
For example, "a death-possible" case was prosecuted as death-eligible in Monmouth County 70% of the time, whereas in Essex County only 23% of death-possible cases were prosecuted as death-eligible; in Hudson and Camden Counties the rates were 24% and 28%, respectively. Study, Interim Report Part I, at TABLE CP4.
Thus, according to the Study, the chances that a plea bargain would be offered in a given case were significantly higher in Hudson and Camden Counties than in Monmouth County. Only 25% of such cases were given a plea bargain in Monmouth County, whereas 74% and 63% of such cases were pled out in Camden and Hudson Counties, respectively. Id. at TABLE PT4.
The Study relies on the defense attorneys’ designation of a case as "death-possible.” According to the Preliminary Report of the Public Defender, the data base was compiled from interviews with defense counsel. The interviews were intended to "record precise information on the characteristics of the defendant and victim, the circumstances of the homicide, the legal result, and the stages of case processing.” The interviewing process is described as follows:
When a case is completed the defense attorney is interviewed in person by the field attorney using the structured interview schedule. The defense attorney is interviewed with the case file before him____ The field attorneys ask for a copy of the indictment or accusation, the judgment sheet and the verdict sheet, and they also ask for a copy of the presentence report and the initial police report____ These documents are then filed by case study number and used for verification____
In addition to recording the data points on the structured interview schedule, the field attorneys ... probe on the individual circumstances of each case____ If the case was not a capital case, but there was a factual basis for serving a notice of factors, why, in the defense attorney’s opinion, was the case not designated a capital case by the prosecutor. What were the important issues in the case. What was the nature of the evidence. How credible were the State witnesses. Were there plea offers____ How strong was the prosecution’s case, and how was the case structured prior to trial or plea. Was attorney conducted voir dire used, and was there significant pretrial publicity. What were the defense strategies. What was the prosecutor’s theory of the case. Did the defendant make a statement and was the statement admitted. What was important in this particular homicide case. Why did the case result in this particular disposition.
[Study, Preliminary Report (January 1987) at 57-59.]
The Study offers a statistical comparison of "all cases of homicide, except vehicular manslaughter, where the homicide occurred after August 6, 1982, the effective date of the reimposition of capital punishment in New Jersey.” The criteria for inclusion in the data base are "formal charge for a homicide offense by the prosecutor's office and a final disposition of that charge at the trial court level." Study, Interim Report Part I, at APPENDIX A-l — Methodology. The specific data on the case, including the assessment of whether a factual basis existed for serving a notice of aggravating factors, is assembled by interviewing "the attorney, private counsel or public defender, who represented the defendant at trial or at plea and sentencing.” Id.
A reliable indicator of "death-eligible" cases could be simply whether cases are like ones in which a notice of aggravating factors is served. The Public Defender’s Study, however, has allowed “death-possible” characteristics of a case to be defined essentially by the defense counsel in a given case, and the Study uses these “death-possible” data to make the serious charge that from 1982-1986, of 404 cases in which a factual basis existed for serving a notice of aggravating factors, prosecutors actually served such a notice on only 131 defendants. The validity of the implication of this statistic — that the odds are great that at least some cases in which the death sentence was sought and obtained are factually indistinguishable from cases in which the penalty was not even sought, and thus that the system is operating arbitrarily — is dependent on the accuracy of the data on how many death-possible cases existed.
The Smith case is highly instructive and of more than passing interest. The defendant, a juvenile charged with capital murder (the legislature's amendment proscribing execution of juvenile offenders later mooted the capital issue), see State v. Bey (I), 112 N.J. 45 (1988), challenged the Essex County Prosecutor's exercise of discretion in seeking the death penalty in his case. To support his
*162claim, Smith brought
to the court’s attention 15 other recent homicide indictments ... [E]ach of these indictments charges the defendant with purposeful or knowing murder by his or her own conduct. Each also includes a count charging that the victim of the murder was also the victim of a robbery. In each of these cases, it was factually plausible for the prosecutor to seek the death penalty. Unlike Smith, however, none of these 15 other defendants was served with notice of [aggravating factors]. Thus, the prosecutor made the decision to treat Smith’s case as a capital case, but declined to seek the death penalty against similarly situated defendants.
Smith, supra, 202 N.J.Super. at 591. Judge Cocchia acknowledged the "broad discretionary powers” with which a prosecutor is vested. Id. at 592 (citing State v. Laws, 51 N.J. 494, 510-11 (1968), cert. den., 393 U.S. 971, 89 S.Ct. 408, 21 L.Ed.2d 384; State v. Hermann, 80 N.J. 122 (1979); State v. Conyers, 58 N.J. 123, 146 (1971) (it is within the prosecutor’s discretion not to seek the death penalty)). The court also acknowledged that ”[a]s a general proposition, the conduct of a prosecutor is presumed to be valid,” State v. McCrary, 97 N.J. 132, 142 (1984), but insisted that ”[n]owhere is judicial intrusion into the realm of prosecutorial discretion more appropriately exercised or likely to be necessary than in capital criminal proceedings.” Smith, supra, 202 N.J.Super. at 592 (citing McCrary, supra, 97 N.J. at 141). The court rejected the prosecutor’s argument that "so long as there was an adequate factual basis for serving Smith with notice of the aggravating factor ... he cannot be said to have acted arbitrarily or capriciously____ This argument... confuses arbitrary and capricious prosecution with groundless prosecution." Id. 202 N.J.Super. at 593. Arbitrary, the court declared, “ 'means depending on will or discretion, that is not governed by any fixed rules or standards.'" Id. (quoting Canada Dry Ginger Ale, Inc. v. F & A Distrib. Co., 28 N.J. 444, 456 (1958)). The court concluded:
Even now, almost three years after the reenactment of the death penalty, the guidelines by which the prosecutor determines whether to proceed with a homicide case as a capital or noncapital cause remains [sic] obscure. Indeed, in oral argument ... the attorney for the State admitted that she had no notion as to how the Essex County Prosecutor's Office chooses homicide cases for capital treatment. How then is a defendant ... to meet his burden of demonstrating that the decision of the prosecutor ... is arbitrary ... where the defendant has no comprehension of the criteria and procedure by which that vital decision is made? ... The defendant ... the public ... and the court, which must be sufficiently informed about the decision-making process so as to be able to ensure that it is free of any randomness, vagary or discrimination, are all entitled to some understanding of how the prosecutor selected cases for capital treatment.
[Id. 202 N.J.Super. at 594.]
Accordingly, the court ordered "that the prosecutor submit in writing a statement setting forth the criteria and procedure by which the Essex County Prosecutor's Office selects homicide cases for capital prosecution.” Ibid. The Prosecutor’s Office responded with an unsigned, undated memorandum *163that provided none of the requested information. According to a brief filed in this case, the court responded in a letter: “The Court finds the statement submitted to be flagrantly inadequate____ The Court is displeased with what it perceives to be a cavalier response on the part of the Prosecutor’s Office— indeed, it is difficult to believe that the attorneys for the State legitimately anticipated that so cursory and deficient a statement would be acceptable to the Court." The Prosecutor’s Office followed with a more detailed description of its internal procedures in capital cases. According to the revised memorandum, the decision whether to prosecute a given homicide as a death-penalty case depends on the Assistant Prosecutor’s assessment that one or more aggravating factors can be proven.
It is clear that no explicit comparison is made among homicides; decisions are made on a case-by-case basis. According to arguments presented to this Court on behalf of the Attorney General in recent capital murder appeals, there is no commitment by that office to the need for uniform standards, nor is there any sense that non-uniformity among prosecutors constitutes any kind of a problem in the administration of the death penalty statute. This provides independent empirical support for the Public Defender’s thesis that there exists a high probability of arbitrary prosecution on a case-by-case and county-by-county basis, stemming from an absence of uniform prosecutorial standards.
The studies there indicated that a black murderer of a white person is more likely to be sentenced to death than a black murderer of a black person, and that prosecutors seek the death penalty in 70% of the cases involving black defendants and white victims and 15% of the cases involving black defendants and black victims, McCleskey v. Kemp, supra, 481 U.S. at 285, 107 S.Ct. at 1763, 95 L.Ed.2d at 275.
According to the Study, cases involving white victims comprise 40.5% of all death-possible cases but increase their percentage share substantially to 53.1% *165of the cases that are death-eligible. This increase is to be contrasted with cases involving a black victim, which comprise 44.8% of all death-possible cases, but only 38.5% of death-eligible cases. Similarly, cases involving Hispanic victims comprise 14.7% of death-possible cases, but only 8.4% of death-eligible cases. Because the odds that a case involving a white victim would advance from death-possible to death-eligible were substantially greater than in cases involving Hispanic or black victims, the Study points to some racial disparity in the plenary exercise of prosecutorial discretion over seeking the death penalty in death-eligible cases. Study, Interim Report Part I, at TABLE CP2.
Black defendant/white victim cases had the highest probability of being death-possible (.85) followed, in decreasing order, by white defendant/white victim (.62), black defendant/black victim (.52), and Hispanic defendant/Hispanic victim (.50). White victim cases, regardless of whether the defendant was black or white, exhibited the highest probability of advancing to the death eligible stage. Id. at TABLE CP3B.