People v. Anderson

LUCAS, C. J.

I concur in the judgment, and in all aspects of the majority opinion with one exception. I am less certain than the majority that admission of codefendant Sheila Anders’s statements to psychiatrists constituted Bruton/Aranda error. (See Bruton v. United States (1968) 391 U.S. 123 [ 20 L.Ed.2d 476 , 88 S.Ct. 1620 ]; People v. Aranda (1965) 63 Cal.2d 518 [ 47 Cal.Rptr. 353 , 407 P.2d 265 ].) Although these statements implicated defendant, they were elicited on cross-examination of these psychiatrists for the limited purpose of impeaching their opinions supporting Sheila’s diminished capacity defense, and the jury was so advised. Under these circumstances, I doubt that any denial of defendant’s right of confrontation occurred here.

In any event, in light of the majority’s conclusion that any error in admitting the statements was harmless, I simply state my reservations to the finding of error without prolonged discussion.

BROUSSARD, J.

Today the majority of this court overrule Carlos v. Superior Court (1983) 35 Cal.3d 131 [ 197 Cal.Rptr. 79 , 672 P.2d 862 ] and People v. Turner (1984) 37 Cal.3d 302 [ 208 Cal.Rptr. 196 , 690 P.2d 669 ], *1152 and hold that under California law a person can be executed for an unintentional killing. I am compelled to dissent from that portion of the majority’s opinion.

In Carlos this court, by a six-to-one majority, held that the special circumstance of felony murder under the 1978 death penalty initiative applies only when the defendant actually intended to kill. 1 Thirteen subsequent decisions of this court have confirmed the Carlos interpretation of the 1978 law. 2 Now the majority overrule this line of precedent and declare that in California a person can be executed for an accidental or negligent killing. 3 They say that this is the clear meaning of the 1978 death penalty initiative, even though they can arrive at this meaning only by interpreting part of the 1978 law to mean the opposite of what it says. The Carlos opinion, however, demonstrates that neither the inconsistent and murky language of the initiative, the circumstances of its enactment, nor the accepted principles of statutory construction require us to conclude that the initiative permits execution of an unintentional killer. There is no good reason for this court to depart from the reasonable construction of the initiative established in Carlos and approved in numerous subsequent decisions.

Periodically, when the political winds gust in a new direction, it becomes necessary to remind all concerned of the virtues of a steady course. As lawyers and judges, we sometimes deliver our reminder in Latin: stare *1153 decisis. The reminder is particularly pertinent today, for the Attorney General has asked this court to reconsider not only Carlos v. Superior Court, supra, 35 Cal.3d 131 , and People v. Turner, supra, 37 Cal.3d 302 , but also virtually every other decision construing the 1977 or 1978 death penalty laws. As appellate counsel remarked in one case, to argue or decide an appeal under such circumstances is like standing on quicksand; if all premises are open to question, one can never reason to a conclusion. Justice Cardozo observed that “the labors of judges would be increased almost to the breaking point if every past decision could be reopened in every case, and one could not lay one’s own course of bricks on the secure foundation of the courses laid by others who have gone before him.” (Cardozo, The Nature of the Judicial Process (1921) p. 149.) 4

From the vast array of commentary on the importance of following precedent, I have selected as most pertinent to the present setting the words of Justice Ryan of the Illinois Supreme Court. In 1979 that court, in a four-to-three decision, upheld the constitutionality of the Illinois death penalty statute. (People ex rel. Carey v. Cousins (1979) 77 Ill.2d 531 [ 397 N.E.2d 809 ], cert. den. (1980) 445 U.S. 953 [ 63 L.Ed.2d 788 , 100 S.Ct. 1603 ]). Two years later, a member of the majority was replaced by a new justice who believed the statute was unconstitutional. There was now a four-to-three majority for unconstitutionality, on what was certainly a question of continuing public importance. The court, however, again upheld the statute, this time by a six-to-one vote. (People v. Lewis (1981) 88 Ill.2d 129 [430 N.E.2d. 1346].) Justice Ryan, who dissented in the earlier case, wrote: “This court, not seven individual justices, has considered the constitutionality of our death penalty statute and this court found it to be constitutional. Those of us who disagree with that conclusion voiced our dissent. Having done so, it is now our obligation to accept the law as pronounced by this court. This is not to say that the holdings of this court are cast in stone and forever unchangeable. However, nothing has changed since this court’s decision in Cousins except one member of the Court that decided Cousins has retired. As Mr. Justice Clark stated in his concurrence, the circumstances which warrant changes in the law do not include changes in personnel of the court. If the law were to change with each change in the makeup of the court, then the concept that ours is a government of law and not of men would be nothing more than a pious cliche.” (P. 1364, conc. opn. of Ryan, J.) 5

*1154 A. The alleged reason for reconsidering Carlos.

The majority open their discussion by explaining their reasons for reconsidering Carlos—a curious explanation, because it appears to concede that Carlos was correctly decided.

The majority note that Carlos was based in part on the United States Supreme Court decision in Enmund v. Florida (1982) 458 U.S. 782 [ 73 L.Ed.2d 1140 , 102 S.Ct. 3368 ]. Carlos noted that Enmund was susceptible of alternative interpretations: a narrow one which would require proof of intent to kill before an accomplice could be executed, and a broader one which would require such proof for the actual killer as well. Since Enmund himself was an accomplice, the decision did not decide the intent requirement for the actual killer. As Justice Brennan, speaking for four of the five members of the Enmund majority, remarked in Tison v. Arizona (1987) 481 U.S. _ , _ [ 95 L.Ed.2d 127, 152 , 107 S.Ct. 1676 ], the constitutionality of the death penalty for the unintentional killer was a question “reserved” by Enmund—an unsettled question which might be decided either way in a later case.

Carlos accordingly turned to the principle that “where a statute is susceptible of two constructions, by one of which grave and doubtful constitutional questions arise and by the other of which such questions are avoided, our duty is to adopt the latter.” (United States v. Delaware & Hudson Co. (1909) 213 U.S. 366, 407-408 [ 53 L.Ed. 836, 849 , 29 S.Ct. 527 ], quoted in Carlos, 35 Cal.3d at p. 148 .) Applying this principle, it became our duty to construe the 1978 death penalty law to impose an intent to kill requirement for the actual killer.

The majority, as I read their opinion, dispute none of this. They agree that Carlos’s interpretation of Enmund was a reasonable one; indeed they cite commentary, unavailable when Carlos was decided, which supports our view. Neither do they dispute the proposition that courts have a duty to interpret statutes to avoid serious constitutional questions, nor that such duty required us to interpret the 1978 law to include an intent to kill requirement for actual killers. In short, the majority appear to agree that Carlos was correctly reasoned on this fundamental point, and in view of the importance of the point, that is equivalent to conceding that Carlos was correctly decided.

*1155 In short, the majority do not say that Carlos was wrong but that it is outdated. They assert that two subsequent Supreme Court decisions (Cabana v. Bullock (1986) 474 U.S. 376 [ 88 L.Ed.2d 704 , 106 S.Ct. 689 ] and Tison v. Arizona, supra, 481 U.S. _ [ 95 L.Ed.2d 127 , 107 S.Ct. 1676 ], both five-to-four decisions, read Enmund more narrowly then we did. 6 The language the majority quote from those decisions, however, offers little support for their assertion; it consists only of insignificant dictum from Cabana v. Bullock 7 and a mistaken footnote in Tison v. Arizona. 8

In any case, the majority’s assertions fail on several grounds. In the first place, Carlos was not based solely on the fear that Enmund v. Florida, supra, 458 U.S. 782 , might invalidate the 1978 law. Our decision also considered the text of the statute, the ballot arguments, and the rule that a defendant is entitled to the benefit of any reasonable doubt concerning the interpretation of a penal law. None of these considerations are affected by the cited United States Supreme Court decisions.

Furthermore, the majority’s assertion that Cabana v. Bullock, supra, 474 U.S. 376 and Tison v. Arizona, supra, 481 U.S. _ [ 95 L.Ed.2d 127 , 107 S.Ct. 1676 ] undermine the reasoning of Carlos clearly rests on a misunderstanding. Carlos did not say that one interpretation of Enmund was right and another wrong. We said that Enmund could reasonably be interpreted to require proof of intent for the actual killer. The fact that a bare majority of the court in Tison adopted a narrower view in no way refutes that assertion. And in a still later case, Booth v. Maryland (1987) 482 U.S. _ [ 96 L.Ed.2d 440 , 107 S.Ct. 2529 ], another five-to-four majority appeared to reject much of the reasoning of Tison. 9

*1156 Finally, the majority’s confidence that the constitutional issues raised by Enmund have been conclusively settled may be unwarranted. The pattern of five-to-four decisions demonstrates that four of the eight justices currently sitting believe that the death penalty can be justified, if at all, only by considerations of deterrence and culpability, and that both deterrence and culpability turn on the intent of the defendant. Such a theory forecasts the unconstitutionality of executing any person, including an actual killer, who did not intend to kill. A constitutional theory which would receive a favorable hearing from one half of the United States Supreme Court 10 is a matter of serious concern. It was and remains our duty to construe the 1978 death penalty initiative to avoid the risk that it will be declared unconstitutional under that theory.

In short, Carlos was correctly decided in 1983, and its reasoning, unimpaired by future cases, should govern today, it is disingenous to claim that a passing remark in Cabana v. Bullock and a mistaken footnote in Tison v. Arizona justify reconsideration of that decision.

B. The meaning of the statutory language.

The majority assert that the 1978 death penalty law is “realistically” susceptible of only one meaning: that intent to kill is required for accomplices to a felony murder, but not for the actual killer. Anyone familiar with principles of statutory construction will recognize that something is wrong with this assertion—the word “realistically” does not belong. It has long been established (see Ex parte Rosenheim (1890) 83 Cal. 388, 391 [ 23 P. 372 ]) that if a penal statute is reasonably susceptible of more than one interpretation, the courts should give it that interpretation which favors the *1157 defendant. (See, e.g., People v. Garfield (1985) 40 Cal.3d 192, 200 [ 219 Cal.Rptr. 196 , 707 P.2d 258 ]; People v. Davis (1981) 29 Cal.3d 814, 828 [ 176 Cal.Rptr. 521 , 633 P.2d 186 ]; People v. King (1978) 22 Cal.3d 12, 23 [ 148 Cal.Rptr. 409 , 582 P.2d 1000 ]; Bowland v. Municipal Court (1976) 18 Cal.3d 479, 488 [ 134 Cal.Rptr. 630 , 556 P.2d 1081 ].) This rule derives from the fundamental precept that a “defendant is entitled to the benefit of every reasonable doubt, whether it arise out of a question of fact, or as to the true interpretation of words or the construction of language used in a statute.” (In re Tartar (1959) 52 Cal.2d 250, 257 [ 339 P.2d 553 ]; People v. Davis, supra, 29 Cal.3d 814, 828 ; People v. Craft (1986) 41 Cal.3d 554, 560 [ 224 Cal.Rptr. 626 , 715 P.2d 585 ].) But all of these cases refer to “reasonable” interpretation and “reasonable” doubt. The majority inexplicably substitute the word “realistic” for “reasonable” and, as if to stress that the change is deliberate, italicize the words. 11 But they do not tell us how a “realistic” interpretation differs from a “reasonable” one, or a “realistic” doubt from a “reasonable” doubt. Pending an explanation, I will assume the majority intend no change in the rules of statutory construction, and will speak of “reasonable” doubt and “reasonable” interpretations of a statute. 12

I turn then to the question whether Penal Code section 190.2 is reasonably susceptible of only a single interpretation, one which requires intent to kill for accomplices to felony murder but not for killers.

Section 190.2, subdivision (a)(17) (hereinafter paragraph (17)) is the felony-murder special circumstance. 13 It provides for a special circumstance if “the murder was committed while the defendant was engaged in or was an accomplice in the commission of, attempted commission of, or the immediate flight after committing or attempting to commit” nine listed felonies— robbery, kidnapping, rape, sodomy, lewd and lascivious conduct, oral copulation, burglary, arson, and train wrecking. It is silent as to intent to kill. As *1158 we acknowledged in Carlos , one could infer from this silence the absence of an intent requirement.

As Carlos pointed out, however, a construction of paragraph (17) without an intent requirement would have anomalous results. Because section 189, the felony-murder provision, lists five of the felonies listed in paragraph (17)—robbery, rape, lewd conduct, burglary, and arson—a defendant convicted of one of those offenses could be executed without proof of intent to kill. One convicted of one of the other felonies listed in paragraph (17)— kidnapping, forcible sodomy, forcible oral copulation, or train wrecking— could not be convicted of first degree murder, and thus could not be executed, without proof of intent. Rather than blaming this anomaly on careless draftmanship, the majority unwisely attempt to defend it. They suggest that the voters may have believed kidnapping less inherently dangerous than arson, rape, robbery, burglary, or child molesting. Paragraph (17) expressly includes kidnapping in violation of section 209—that is, kidnapping for ransom or robbery—and it seems quite unlikely that the voters thought this crime less dangerous than burglary. The majority also suggest that forcible sodomy and oral copulation occur less frequently than rape and therefore stand less in need of deterrence. Are these crimes less frequent? Is there any reason to think the drafters thought they were? That the voters thought they were? Frequency is in any case irrelevant, since the objective, I assume, is to impose the death penalty for the most atrocious murders, not the most common murders. Isn’t it more reasonable to assume that the voters believed that all felonies listed in paragraph (17) would be treated alike? 14

In any event, the majority agree with Carlos that section 190.2 requires proof of intent to kill for accomplices such as Carlos; they maintain only that it does not extend this requirement to the actual killer. They cannot derive this distinction from paragraph (17), which expressly applies to both actual killers and accomplices and says nothing about intent for either. Instead, for that, the majority turn to section 190.2, subdivision (b), which states that “[ejvery person whether or not the actual killer found guilty of intentionally aiding, abetting, counseling, commanding, inducing, soliciting, requesting, or assisting any actor in the commission of murder in the first degree shall suffer death or confinement in state prison for a term of life without the possibility of parole. ...” The majority acknowledge that “in the abstract the introductory phrase of section 190.2 [, subdivision] (b)— *1159 ‘[ejvery person whether or not the actual killer’—encompasses both the actual killer and his [accomplice].” (Maj. opn. at p. 1142.) I would go further, and maintain that such is the single plain meaning of those words. In an attempt to avoid the plain meaning of those words, the majority note that the statute goes on to speak of persons found guilty of “aiding, abetting, counseling, commanding, inducing, soliciting, requesting assisting” in the commission of a murder. They do not, however, adhere to the straightforward, literal meaning of the subdivision, and apply it to require intent to kill for every person, including actual killers, who also assist other murderers. That would be an unreasonable interpretation, for there is no conceivable reason to require proof of intent to kill for the actual killer who acts with a confederate, but not for one who acts alone.

The majority, however, assert that the only reasonable meaning of subdivision (b) is one which imports an intent requirement for accomplices but not the actual killer. They do not go through the language of the section, word by word and phrase by phrase, to show how such is the clear and necessary import of the words and phrases employed. If they did, it would become apparent that the only way to arrive at the majority’s interpretation is to turn the subdivision’s opening phrase upside down. “Every person whether or not the actual killer” must be transformed into “Every person except the actual killer.” I recognize that when a court deals with a statute as poorly drafted as this one, it will often be necessary to depart from the literal language of the statute; it may even be necessary to construe some phrase to mean the opposite of its apparent meaning. But if the court concludes that it must construe a phrase to mean the opposite of what it apparently means, it cannot justify this by the single plain-meaning rule.

Indeed, the basic problem with the majority opinion is that it simply adopts one reasonable interpretation of subdivision (b), justifying that interpretation by arguments which would be unnecessary and inappropriate if the language had a single clear meaning. It invokes, for example, a comparison of the language of the 1978 and 1977 law, the presumption that the deletion of former language is intended to change the law, and the ordinary judicial construction of words describing accomplices. It ignores, however, similar canons of construction, such as the one which requires the court to give meaning and effect to all words and phrases. (See, e.g., Clements v. T. R. Bechtel Co. (1954) 43 Cal.2d 227, 233 [ 273 P.2d 5 ]; Select Base Materials v. Board of Equal. (1959) 51 Cal.2d 640, 645 [ 335 P.2d 672 ]; Mahdavi v. Fair Employment Practice Com. (1977) 67 Cal.App.3d 326, 334 [ 136 Cal.Rptr. 421 ].) Having in this manner arrived at one reasonable interpretation of subdivision (b), the majority then announce that this is the plain meaning of the statute and the only reasonable construction.

*1160 I think it is clear that subdivision (b) has no single clear meaning on its face. The majority have advanced one reasonable interpretation of that subdivision. But theirs is not the only reasonable interpretation. As five justices of this court, including the author of the present majority opinion, recognized in 1983, one can also reasonably argue that a subdivision which purports to apply to “every person whether or not the actual killer” does in fact apply to every person whether or not the actual killer.

Thus the issue in this case, as Carlos, supra 35 Cal.3d 131 , is one of choosing between alternative reasonable interpretations. One interpretation, however, is now fortified by precedent. The interpretation of subdivision (b) to require proof of intent to kill for “all persons” including the actual killer, has been reaffirmed many times by this court (see cases cited, fn. 2, ante ). Nothing has changed which warrants yet another reconsideration of that matter.

C. The ballot presentation and arguments.

The majority refer to the 1978 ballot to support their position. They first quote the presentation by the Legislative Analyst. He said that “this proposition would specifically make persons involved in the crime other than the actual murderer subject to the death penalty or life imprisonment without possibility of parole under specified circumstances.” The majority consider this an evident reference to subdivision (b); I think it refers to the proposition as a whole. In any event, the analyst’s statement is unquestionably true; both Carlos, supra, 35 Cal.3d 131 , and the majority today interpret the statute so that an accomplice is subject to the death penalty under some circumstances. But the analyst did not tell the voters what the circumstances were, nor whether a particular circumstance—intent to kill—applied to the actual killer. 15

The majority then refer to the ballot arguments. The opponents of the initiative charged that it would impose the death penalty upon the accomplice of an unintentional killer. The proponents responded: “The opposition maintains if someone were to lend a screwdriver to his neighbor and the neighbor used it to commit a murder, the poor lender could get the death penalty, even though ‘he had No Intention that anyone be killed.’ [fl] *1161 Please turn back and read Section 6b [now § 190.2, subd. (b)]. ... It says that the person must have Intentionally aided in the commission of a murder to be subject to the death penalty under this initiative.” After reading this argument, the voters who supported the initiative could be confident that the lender of the screwdriver could not get the death penalty. But what about the “neighbor” who “accidentally killed someone” during the burglary? The argument leaves his fate obscure. If the drafters of the 1978 law believed that it, unlike the 1977 law, would leave the accidental killer subject to the death penalty, here was their opportunity to explain that change to the voters. But no explanation was offered.

As we said in Carlos , “[t]he adoption of a law to permit infliction of the death penalty upon an accidental killer would be a momentous step, raising grave moral questions.” ( 35 Cal.3d at p. 145 .) 16 It certainly would be a very significant departure from the 1977 law, perhaps the most significant change effected by the initiative. One could reasonably expect a change of this magnitude would be made clear in both legal text and ballot argument. The absence of any such language suggests that the drafters did not intend to permit execution of an unintentional killer. More importantly, it compels a conclusion that the voters in approving the initiative were unaware that it could be construed to permit such an execution.

D. Canons of intepretation of penal statutes.

As we observed earlier, “[w]hen language which is reasonably susceptible of two constructions is used in a penal law ordinarily that construction which is more favorable to the offender will be adopted. [1f] The defendant is entitled to the benefit of every reasonable doubt, whether it arises out of a question of fact, or as to the true interpretation of words or the construction of language used in a statute.” (People v. Davis, supra, 29 Cal.3d 814, 828.) 17

*1162 The majority reject the application of this principle with the claim that Carlos’s construction of the 1978 law is unreasonable. Of course, there is no reason to discuss any extrinsic aid—the language of the prior statute, the ballot arguments, constitutional considerations, etc.—unless one is choosing between reasonable alternatives. With respect to all other extrinsic aids, however, the majority argue that such aids fail to support the choice we made in Carlos , and may even support the majority’s construction. When it comes to considering the rule interpreting penal statutes in favor of the defendant, however, the majority simply argue that such an interpretation is unreasonable. I accept this as a tacit acknowledgment that if the majority’s interpretation and the Carlos interpretation were reasonable alternatives, this canon of construction would support the Carlos view.

E. Interpretation to avoid doubtful constitutionality.

The majority recognize the rule that a statute should be construed to avoid grave and doubtful constitutional questions, and that when Carlos was decided a construction of the 1978 law which would permit the execution of an unintentional killer might violate the Eighth Amendment as construed in Enmund v. Florida, supra, 458 U.S. 782 . (Maj. opn., ante, at pp. 1139-1140, 1146-1147.) The majority assume that such questions raised by Enmund have been fully and finally resolved. As we explained earlier, this assumption is itself dubious. (Ante, p. 1154.)

Carlos also noted two other constitutional problems which have not been addressed by the United States Supreme Court. First, the court in Godfrey v. Georgia (1980) 446 U.S. 420, 427 [ 64 L.Ed.2d 398, 405-406 , 100 S.Ct. 1759 ], required that a state death penalty law provide a “meaningful basis for distinguishing the few cases in which [the penalty] is imposed from the many cases in which it is not.” We questioned whether a law which included some who killed accidentally, while excluding the nonfelony but premeditated murder, would meet this standard. 18

We also noted that the statute classifies otherwise similarly situated persons—convicted murderers—and provides that a person who committed a willful and premeditated murder is not subject to the death penalty unless a special circumstance is present. An interpretation of section 190.2 that *1163 allows the death penalty for unintentional felony murderers, may violate the equal protection clause.

Under accepted constitutional theory, a classification which directly abridges a fundamental constitutional right is tested by the stringent “compelling interest test.” I can think of no more fundamental right than life itself, nor a more direct abridgement of that right than execution. I doubt, however, that the classification imposed by the California statute (as construed by the majority) could even meet the rational-relationship test employed by courts to test classifications which do not affect fundamental rights. The state has an interest in deterring unintended felony murders, but surely it is no greater than its interest in deterring premeditated murder, so it cannot justify a greater punishment. ( Enmund, supra, 458 U.S. 782 , would suggest that the state’s interest in deterring unintentional felony murder is far less than its interest in deterring premeditated murder.) And while the unintentional felony murderer bears a measure of responsibility for his victim’s death, and can be subject to retributive punishment accordingly, surely he bears no greater responsibility than the murderer who premeditates. In sum, even if one could find a rational basis for punishing unintentional felony murderers as severely as premeditating killers, there is no rationale that would justify punishing them more severely. 19

The majority’s response to these arguments is to assert that “[wjhether or not we approve of the wisdom of the statutory classification, it appears to be generally accepted. ...” But many practices are generally accepted until they are constitutionally challenged. The majority cite only one case, Gray v. Lucas (5th Cir. 1982) 677 F.2d 1086 , to support their conclusion that the classifications of the 1978 law are constitutional. If that decision is distinguishable or erroneous, then the constitutional issues raised in Carlos remain unsettled.

Gray concerned a challenge to a Mississippi law which imposed the death penalty on felony murderers but not ordinary murderers. The decision of the court upholding that law can be distinguished on several grounds: (1) The jury expressly found that Gray committed an intentional felony murder. The Gray court noted that imposing the death penalty upon unintentional felony murders “might pose a difficult question of constitutional law,” (667 F.2d at p. 1103), but found it unnecessary to reach the issue. (2) In addressing the issue whether imposition of the death penalty on intentional felony murderers violated equal protection, the court assumed that the rational-basis test applied. In California, however, precedents such as *1164 People v. Olivas (1976) 17 Cal.3d 236 [ 131 Cal.Rptr. 55 , 551 P.2d 375 ] (which held “liberty” a fundamental interest), might compel use of a more exacting standard. (3) Gray found a rational basis for the challenged classification under the hypothesis that the Mississippi Legislature may have believed felony murder and ordinary murder posed two different problems, and planned to address the felony-murder problem first. That rationale could not be applied to the California statute: the 1978 law dealt comprehensively with felony and premeditating murderers. The voters were told expressly that the initiative applied to all murderers. (4) Gray asserted that the legislature could reasonably judge that “the death penalty would be more effective in deterring felony murders since an experienced felon is more likely to assess the consequences of his acts,” or conversely that it “might not effectively deter simple murders since such persons are likely as a group to act on passion or impulse and thus be unmindful of the consequences of their crime.” ( 677 F.2d at p. 1104 .) But whatever its rationality in distinguishing between intentional felony murderers and intentional non-felony murderers under Mississippi law, that reasoning breaks down completely when applied to the unintentional felony murderer. When a person who does not intend to kill contemplates the consequences of his acts, the possibility of the death penalty will not “enter into the cold calculus that precedes the decision to act.” (Enmund, supra, 458 U.S. at p. 799 [ 73 L.Ed.2d at p. 1153 ].) No one has suggested a rational basis for believing that the death penalty is a more effective deterrent of unintentional felony murder than of premeditated murders.

In Newland v. Board of Governors (1977) 19 Cal.3d 705, 711 [ 139 Cal.Rptr. 620 , 566 P.2d 254 ], we promised that when constitutional issues are raised, we would undertake “a serious and genuine judicial inquiry into the correspondence between the classification and the legislative goals.” When the court undertakes such an inquiry, it will find the imposition of greater punishment upon unintentional felony murderers than upon premeditated murderers presents a substantial constitutional question.

F. Stare decisis.

The doctrine of stare decisis necessarily implies that the fact a current majority disagrees with a prior decision is not in itself sufficient to justify overruling it. Otherwise all prior decisions would be exposed to continuous challenge, the concept of precedent would be meaningless, and the decisions of this court will sway with the political winds. In short, it is not enough for the majority to say that Carlos, supra, 35 Cal.3d 131 , was wrongly decided; something more is needed to justify a departure from stare decisis. The majority offer nothing except to note that the matter is one of continuing concern, as if stare decisis meant only the duty to follow inconsequential *1165 precedent. Taking the opposite road, I submit three reasons why, even assuming Carlos erred in its interpretation of the 1978 law, that decision should not be overruled.

First, it is rare for this court to overrule a recent precedent construing a statute. The Legislature is competent to correct a statutory construction with which it disagrees, and does so regularly. The voters too have that power, and exercised it in 1973 to overturn a decision by this court finding the death penalty to be cruel or unusual punishment. The absence of legislative or initiative action suggests that there is no consensus among either the Legislature or the voters that an unintentional murderer should be executed, and argues against the court taking this step on its own.

Second, since Carlos was filed in December of 1983, intent-to-kill instructions have been given routinely in all capital cases. (Some trial judges, anticipating our decision, gave such instructions in pr e-Carlos cases.) We have heard no complaint that such instructions have led to confusion, inefficiency, or brought about mistaken or unjust verdicts.

Indeed, the only reason Carlos is an issue today is that this court has failed to decide a number of cases that were tried prior to December of 1983 in which an intent-to-kill instruction was not given. 20 Some of those cases, including the present one, will have to be retried anyway, 21 but in others Carlos stood as a possible barrier to the execution of that minority of murder defendants unlucky enough to have their cases still pending before this court in January of 1987. The majority tear down that barrier, heedless of the effect of their decision upon cases not yet tried.

Third, overruling Carlos now, four years after it was filed, creates arbitrary distinctions between defendants. Some of those tried before December 1983 received an intent-to-kill instruction. Of those who did not, some have had convictions reversed, and their case retried with an intent-to-kill instruction. 22 None of the above are affected by the present decision. The *1166 impact of the present decision falls on the minority of pre-1983 defendants whose appeals, for one reason or another, have not been resolved by this court four or more years after their trials. Whether a particular defendant falls within that group—and thus perhaps whether he lives or dies—is largely a matter of chance.

Finally, I am impressed by the failure of anyone to argue that it is desirable for the state to execute unintentional murderers. To the contrary, many would think such an execution disproportionate to the culpability of the defendant. Indeed at oral argument the Attorney General said that if a case of an unintentional felony murder arose, he would expect the prosecutor or the jury to use their discretion to avoid the death penalty. However, we cannot rely on prosecutors or juries to refrain from applying the 1978 death penalty law as the majority have construed it, harsh and arbitrary as that construction may be. 23 Sooner or later the present decision will result in an execution unjust to the defendant and unnecessary to protect any interest of the public. We should therefore decline to reconsider Carlos , and leave it for the People or the Legislature to act if they desire to change established precedent.

G. The penalty verdict.

The trial court’s instruction on the Governor’s commutation power—the Briggs Instruction—violated the California Constitution. (People v. Myers (1987) 43 Cal.3d 250, 270-272 [ 233 Cal.Rptr. 264 , 729 P.2d 698 ]; People v. Montiel (1985) 39 Cal.3d 910, 928 [ 218 Cal.Rptr. 572 , 705 P.2d 1248 ]; People v. Ramos (1984) 37 Cal.3d 136, 153 [ 207 Cal.Rptr. 800 , 689 P.2d 430 ].) Those decisions mandate that the imposition of a death sentence following a penalty trial in which the Briggs Instruction was given be reversed as a “miscarriage of justice” within the meaning of article VI, section 13 of that Constitution.

The Attorney General, however, asks us to overrule those decisions. The majority properly reject that plea, for principles of stare decisis require us to adhere to prior precedents establishing that the giving of the Briggs Instruction is reversible error. 24 Of course, those same principles should also *1167 command our adherence to Carlos v. Superior Court, supra, 35 Cal.3d 131 and People v. Turner, supra, 37 Cal.3d 302 .

People v. Turner, supra, applied Carlos to hold that intent to kill was an essential element of the multiple-murder special circumstance.

People v. Whitt (1984) 36 Cal.3d 724 [ 205 Cal.Rptr. 810 , 685 P.2d 1161 ]; People v. Ramos (1984) 37 Cal.3d 136 [ 207 Cal.Rptr. 800 , 689 P.2d 430 ]; People v. Turner, supra, 37 Cal.3d 302 ; People v. Armendariz (1984) 37 Cal.3d 573 [ 209 Cal.Rptr. 664 , 693 P.2d 243 ]; People v. Anderson (1985) 38 Cal.3d 58 [ 210 Cal.Rptr. 777, 694 P.2d 1149 ]; People v. Boyd (1985) 38 Cal.3d 762 [ 215 Cal.Rptr. 1 , 700 P.2d 782 ]; People v. Hayes (1985) 38 Cal.3d 780 [ 214 Cal.Rptr. 652 , 699 P.2d 1259 ]; People v. Chavez (1985) 39 Cal.3d 823 [ 218 Cal.Rptr. 49 , 705 P.2d 372 ]; People v. Guerra (1985) 40 Cal.3d 377 [ 220 Cal.Rptr. 374 , 708 P.2d 1252 ]; People v. Fuentes (1985) 40 Cal.3d 629 [ 221 Cal.Rptr. 440 , 710 P.2d 240 ]; People v. Silbertson (1985) 41 Cal.3d 296 [ 221 Cal.Rptr. 152 , 709 P.2d 1321 ]; People v. Hamilton (1985) 41 Cal.3d 408 [ 221 Cal.Rptr. 902 , 710 P.2d 981 ], revd. on other grounds, _ U.S. _ [ 92 L.Ed.2d 734 , 106 S.Ct. 3328 )]; and People v. Ratliff (1986) 41 Cal.3d 675 [ 224 Cal.Rptr. 705 , 715 P.2d 665 ].

Among the cases now subject to the death penalty are the following: (a) A burglar startles a resident, who dies of a heart attack. (Cf. People v. Stamp (1969) 2 Cal.App.3d 203 [ 82 Cal.Rptr. 598 ].)

(b) A robber inflicts only a minor injury, but the victim dies weeks later of unexpected medical complications.

(c) While defendant is on the way to committing an armed robbery, his gun fires accidentally, killing his accomplice. (Cf. People v. Johnson (1972) 28 Cal.App.3d 653 [ 104 Cal.Rptr. 807 ].)

(d) While defendant is driving the get-away car, he causes an accident, killing a bystander. (Cf. People v. Fuller (1978) 86 Cal.App.3d 618 [ 150 Cal.Rptr. 515 ].) Indeed the defendant would be subject to the death penalty even if he were driving carefully, so long as he could be said to be “the actual killer,” and even if his victim was the robber.

Imagine, for example, the chaos if some years in the future a newly constituted California Supreme Court, taking its view of stare decisis from the present decision, were to overrule the present decision and reinstate Carlos , thus invalidating most death penalty judgments during the interim.

A similar event occured in California. We upheld the constitutionality of the 1977 death penalty law in People v. Jackson (1980) 28 Cal.3d 264 [ 168 Cal.Rptr. 603 , 618 P.2d 149 ], a *1154 four-to-three decision. When Justice Manuel died, the court was deadlocked, three to three, on the constitutional issue. Justice Tobriner nevertheless voted to affirm the penalty judgment in People v. Harris (1981) 28 Cal.3d 935, 964 [ 171 Cal.Rptr. 679 , 623 P.2d 240 ], on the ground that he was bound by the prior decision.

To be consistent with their reasoning, the majority should hold their decision to have only prospective effect, or extend retroactivity limited to cases tried after the filing of Tison v. Arizona, supra, 481 U.S. _ [ 95 L.Ed.2d 127 , 107 S.Ct. 1676 ], Under their reasoning a defendant tried after Enmund v. Florida, supra, 458 U.S. 782 , but before Tison, should have received an intent-to-kill instruction.

The issue before the court in Cabana v. Bullock, supra, 474 U.S. 376 was whether the finding of intent to kill must be made by a jury or could be made by a reviewing court. In introducing that issue, the United States Supreme Court described its prior decision in Enmund . The language quoted by the majority is part of that introduction. It does not address any issue before the court in Cabana v. Bullock , and does not consider or purport to resolve any controversy concerning the meaning of Enmund .

Tison v. Arizona, supra, 481 U.S. _ , _, footnote 8 [ 95 L.Ed.2d 127, 141-142 , 107 S.Ct. 1671, 1686 ] said that the California Supreme Court “construed its capital murder statute to require a finding of intent to kill . . . only ... in light of perceived federal constitutional limitations stemming from our then recent decision in Enmund . ” The majority, however, recognized that this statement is a “mischaracterization of the basis of Carlos” (maj. opn., ante, at p. 1141) since Carlos was based upon a number of other reasons in addition to the perceived effect of Enmund .

Booth v. Maryland endorsed the following language from a California Court of Appeal opinion: “We think it obvious that a defendant’s level of culpability depends not on fortuitous *1156 circumstances such as the composition of his victim’s family, but on circumstances over which he has control. A defendant may choose, or decline, to premeditate, to act callously, to attack a vulnerable victim, to commit a crime while on probation, or to amass a record of offenses. ... In contrast, the fact that a victim’s family is irredeemably bereaved can be attributable to no act of will of the defendant other than his commission of homicide in the first place. Such bereavement is relevant to damages in a civil action, but it has no relationship to the proper purposes of sentencing in a criminal case.” (People v. Levitt (1984) 156 Cal.App.3d 500, 516-517 [ 203 Cal.Rptr. 276 ], quoted in Booth v. Maryland, supra, 482 U.S. _ , _, fn. 7 [ 96 L.Ed.2d 440, 449 ].)

If we apply this reasoning to the present setting, we would conclude that a defendant can choose to kill, but he cannot choose to kill unintentionally; the fact that his unintentional act results in a death is not attributable to an act of defendant’s will. It is, therefore, a matter which has no relationship to the proper purposes of criminal sentencing. Such reasoning, admittedly, differs from the reasoning of Tison v. Arizona, supra, 481 U.S. _ [ 95 L.Ed.2d 127 , 107 S.Ct. 1676 ]; the Booth dissenters, all of whom joined the majority in Tison, argued that Booth and Tison were inconsistent. (See Booth, 482 U.S. _ , _ [ 96 L.Ed.2d 440, 458-460 ] (Scalia, J., dis.).

At this writing, the United States Supreme Court has eight members. Justice Powell has retired, and his successor has not yet been confirmed.

One can surmise that the majority substituted “realistic” for “reasonable” because they have concluded that the Carlos interpretation of the 1978 law was “reasonable.”

The creation of a new rule of construction, under which the court cannot adopt a reasonable construction favorable to the defendant if such action is “unrealistic,” would be a revolutionary step. If the majority actually intend to take that step, one would expect them to address the matter candidly, to justify their decision, and to explain what difference it makes to the result in the present case. One would also expect some elucidation of the concept of a “realistic” interpretation; standing alone, the word carries ominous overtones of judicial submission to political realities.

Carlos noted the longstanding judicial antipathy toward the felony-murder rule, which has led courts to give it the narrowest possible application. (People v. Satchell (1971) 6 Cal.3d 28, 33-34 [ 98 Cal.Rptr. 33 , 489 P.2d 1361 , 50 A.L.R.3d 383 ].) That antipathy stems from the artificial and arbitrary character of the rule. When the felony-murder rule standing alone— without proof of malice or premeditation—serves as a basis for the death penalty, the danger of arbitrariness is increased, and the need for a limiting construction greater.

All statutory references are to the Penal Code unless otherwise noted.

The majority correctly point out that the Carlos opinion did not eliminate all anomalies. The 1978 law, as interpreted by the present majority, requires proof of intent to kill, deliberation, and premeditation for some felony murderers, but not for others. Carlos eliminated this irrational distinction with respect to the element of intent, but not for deliberation and premeditation. As between an opinion which eliminates some of the anomalies in the law, and one which retains them all, the former would seem preferable.

The majority also refer to the practice of printing the ballot pamphlet with provisions to be deleted in strikeover type and those to be added in italic type. But the voter who sought to utilize this tool would discover that the entire 1977 law was to be repealed and that all provisions of the 1978 law, including those identical to the 1977 law, were considered new provisions. Any attempt to determine the meaning of the new law by comparing the provisions in italic type with those in strikeover type would be a laborious task, and at the end the voter would only discover the uncertainties and ambiguities discussed in this case, and many others as well.

The majority disagree with Carlos’s claim that the elimination of an intent to kill requirement would be a “momentous step,” but their discussion mistakenly equates “momentous” with “unconstitutional.” (See maj. opn., ante, p. 1145, fn. 8 .) Although Carlos did discuss the constitutionality of executing an unintentional killer, that is a separate issue; the point of the quoted language from Carlos is simply that the elimination of an intent requirement is of sufficient importance, compared to the other changes effected by the 1978 law, that one would expect the ballot arguments to discuss the question.

Davis itself offers an analogy to the present case. It construed section 190.5 of the 1977 death penalty law, which provided that “the death penalty shall not be imposed upon any person who is under the age of 18 years at the time of the commission of the crime.” This section said nothing about imposing a punishment of life imprisonment without possibility of parole. One could argue that on its face the statute unambiguously permitted that punishment; Justice Richardson, who on the same ground dissented in Carlos , so argued. Yet the majority, by Justice Mosk, found the statute ambiguous, and construed it to bar such punishment.

The quoted language from Godfrey v. Georgia, supra, 446 U.S. 420 , assumes that the state standards will permit the death penalty in only a few of the many murder cases. It would be interesting to see statistics that show what percentage of murders in California fall within one or another of the 19 special circumstances. I suspect the 1978 California law has the opposite effect from that envisioned by the United States Supreme Court: that it serves to distinguish those few murder cases in which the death penalty may not be imposed from the many in which, in the discretion of the prosecutor and jury, it may be imposed.

The premeditating killer, without special circumstances, can be sentenced at most to life with possibility of parole; the murderer with felony-murder special circumstances must be sentenced at least to life without possibility of parole.

A substantial percentage of those cases were in fact decided by this court, but the decisions were vacated after a change in the membership of the court.

Because the present case must be retried, it is an inappropriate one for overruling Carlos . Since the new jury will be unfamiliar with the case, the new trial will probably include all evidence from the first trial, including that bearing on intent to kill. Carlos would add only a requirement that the new jury also determine whether defendant intended to kill; hardly a difficult determination under the facts of this case. In short, Carlos makes very little difference in this case, and defendant will have little incentive to seek review of this decision in the United States Supreme Court. But having overruled Carlos in a case where it does not matter, the majority will be able to cite this case to affirm a death penalty upon a defendant who did not intend to kill.

A few cases in which the defendant was sentenced to life imprisonment without possibility of parole were affirmed by the Court of Appeal and not heard before this court.

One could refer to this construction of the 1978 law as “draconian,” a term which refers to a severe and arbitrary law. (See Black’s Law Dict. (5th ed. 1979) p. 443.) Yet, Drakon’s laws, harsh though they may have been, distinguished between premeditated and unpremeditated homicide, and imposed the lesser punishment of exile upon those who did not premeditate. (See Hooper, Greek Realities (1967) p. 137; see generally Stroud, Drakon’s Law on Homicide (1968 U. Cal. Publications) 3 Classical Studies.)

Citing People v. Myers, supra, 43 Cal.3d 250, 272-273 , the Attorney General contends that Ramos error may be harmless and that it was so here. The majority, in response, adhere *1167 to People v Montiel, supra, 39 Cal.3d 310 , which held Ramos error reversible per se, and reject any contrary holding of Myers .

In my view, there is no need to posit a nonexistent conflict between Montiel and Myers . The simple answer to the Attorney General’s contention is that Myers does not support it; our decision in that case declared that “in view of the very serious potential for prejudice emphasized in Ramos , we strongly doubt that we could ever confidently conclude that there was no reasonable possibility that this instruction improperly tainted the jury’s decision-making process.” ( 43 Cal.3d at p. 272 .)