BAXTER, J., Concurring and Dissenting.

Had the trial court in this case given an instruction that a fetus need not be viable under Penal Code section 187, subdivision (a) (hereafter section 187(a)), or had the law in California been settled that, for purposes of section 187(a), a viable fetus meant a fetus with a "probability" or a "reasonable likelihood" of survival outside the womb, then I would not hesitate in joining the lead opinion to reverse defendant's conviction. As it stands, however, neither is the case. (1e) Therefore, while I concur in the holding that viability of a fetus is not required under section 187(a), I must dissent from the decision to reverse.

As the lead opinion points out, the trial court below instructed the jury that to find defendant guilty of fetal murder under section 187(a), it must find that the fetus was viable. Inasmuch as our decision today holds that fetal viability is not a requirement of section 187(a), the instruction given inured to defendant's benefit.

The trial court further instructed: "A fetus is viable when it has achieved the capability for independent existence; that is, when it is possible for it to survive the trauma of birth, although with artificial medical aid." (Italics added.) Although fetal viability is not even a requirement under section 187(a), the lead opinion finds that the court prejudicially erred in defining viability in terms of "possible" survival, and overturns defendant's conviction.

This approach makes no sense. In finding that this definitional instruction was in error, the lead opinion is purporting to decide an issue that was unsettled both at the time defendant acted and at the time of his trial. (See post, pp. 819-821.) In effect, the lead opinion wanders into a wonderland to decide what the law might be had it not been for today's holding rejecting the viability limitation.

Rather than venture into a fictitious world, I believe we should focus on whether defendant's due process rights were violated by the giving of an instruction that defined viability in terms of possible survival. Since the unsettled status of previous case law precludes any determination whether the trial court's instructions constituted error, we should instead look to whether the trial court's reliance on the challenged definition was an unforeseeable judicial enlargement of section 187(a). (See Bouie v. City of Columbia (1964) 378 U.S. 347, 353 [12 L.Ed.2d 894, 899-900, 84 S.Ct. 1697].) I conclude it was not.

There are at least three reasons why the trial court's definition of viability did not amount to an unforeseeable judicial enlargement of the fetal murder statute. First, although at the time defendant acted some Courts of Appeal had inferred a viability limitation to section 187(a), none had directly addressed whether a viable fetus means a fetus with a probability or a reasonable likelihood of survival outside the womb, as opposed to one with only a possibility of survival (i.e., People v. Smith ( K.A.) (1976) 59 Cal.App.3d 751 [ 129 Cal.Rptr. 498] [hereafter K.A. Smith]; People v. Apodaca (1978) 76 Cal.App.3d 479 [ 142 Cal.Rptr. 830] [hereafter Apodaca]; People v. Smith ( R.P.) (1987) 188 Cal.App.3d 1495 [ 234 Cal.Rptr. 142] [hereafter R.P. Smith]; People v. Henderson (1990) 225 Cal.App.3d 1129 [ 275 Cal.Rptr. 837] [hereafter Henderson]). Second, two courts had expressed the view that a viable fetus means one that has a possibility of survival (see Apodaca, supra, 76 Cal.App.3d at p. 489; Henderson, supra, 225 Cal.App.3d at pp. 1157-1158). Third, this court indicated in 1989 — two years before the conduct in this case was undertaken — that we would consider the propriety of an instruction implying that a fetus is deemed viable when it is possible for it to survive if and when we were faced with a case involving a fetus with less than a 50 percent likelihood of survival (see People v. Hamilton (1989) 48 Cal.3d 1142, 1171-1172 [ 259 Cal.Rptr. 701, 774 P.2d 730] [hereafter Hamilton]). Thus, the definition of viability was an open question in California.

As indicated above, two Courts of Appeal, like the trial court below, posited that a viable fetus means a fetus that has a possibility of survival (see Apodaca, supra, 76 Cal.App.3d at p. 489 [Fifth App. Dist.]; Henderson, supra, 225 Cal.App. 3d at pp. 1157-1158 [Fifth App. Dist., Div. Three], citing Apodaca), as contrasted to a fetus that has a probability or a reasonable likelihood of survival.

It is likely that the court in Apodaca, supra, 76 Cal.App.3d 479, spoke in terms of possible survival because Roe v. Wade (1973) 410 U.S. 113 [35 L.Ed.2d 147, 93 S.Ct. 705] referred to viability as the time when a fetus is "potentially able" to live outside the womb, albeit with artificial aid, and has "capability of meaningful life." (See 410 U.S. at pp. 160, 163 [35 L.Ed.2d at pp. 181-182].) However, after Apodaca, supra, the United States Supreme Court, purporting to reiterate its previous view, stated that viability, or capability of meaningful life, was reached when in the judgment of the pregnant woman's attending physician, "there is a reasonable likelihood of the fetus' [ sic] sustained survival outside the womb, with or without artificial support." ( Colautti v. Franklin (1979) 439 U.S. 379, 388 [58 L.Ed.2d 596, 605, 99 S.Ct. 675], italics added [hereafter Colautti].) Subsequently, Henderson, supra, 225 Cal.App.3d 1129, held that section 187(a) was not unconstitutionally vague and therefore did not violate due process principles. In recognizing that the concept of viability was well established in the context of section 187, Henderson cited to the Apodaca definition of viability. ( Henderson, supra, 225 Cal.App.3d at pp. 1157-1158.) Although Henderson did not specifically address the reasonable likelihood language referred to in Colautti, it did conclude that Colautti had no application to the construction of section 187(a) (225 Cal.App.3d at p. 1158.)

Taken as a whole, these decisions indicate that, even though a fetus must have a reasonable likelihood of sustained survival outside the womb for purposes of various abortion issues, a fetus must simply have a possibility of survival for purposes of California's fetal murder statute. Apodaca and Henderson thus put defendant on sufficient notice that even with an implied viability limitation, some California courts were of the view that section 187(a) proscribed the unlawful killing of fetuses having only a possibility of survival. Defendant could therefore reasonably anticipate that his conduct would be proscribed. (See Rose v. Locke (1975) 423 U.S. 48, 50 [46 L.Ed.2d 185, 188, 96 S.Ct. 243].)

Our statements in Hamilton, supra, 48 Cal.3d 1142, provide yet another reason for finding against defendant on the due process issue. In Hamilton, a defendant contended that the trial court gave a jury instruction that improperly implied that a fetus is viable if capable of being born alive, even if it could not have survived for a sustained period outside the mother's womb. In fact, it appears that the allegedly erroneous instruction in Hamilton was based in part on the "possibility" language of Apodaca, supra, 76 Cal.App.3d at page 489, and that the defendant had specifically argued that the instruction was contrary to Colautti, supra, 439 U.S. at page 388 [58 L.Ed.2d at p. 605]. ( Hamilton, supra, 48 Cal.3d at p. 1171.) Although we expressly declined to address the defendant's claim on the merits because there was uncontroverted evidence that the subject fetus had attained viability under any test, we noted: "The 1970 amendment extending murder to fetuses (§ 187, subd. (a); Stats. 1970, ch. 1311, § 1, p. 2440) contained no viability proviso. As amended, section 187, subdivision (a) simply defines murder as 'the unlawful killing of a human being, or a fetus, with malice aforethought.' The Courts of Appeal have inferred a viability limitation in light of the subsequent abortion cases, which first recognized a woman's constitutional right to terminate her pregnancy before the fetus becomes viable. (See, e.g., Smith, supra, 59 Cal.App.3d at pp. 756-757.)" ( Id., at pp. 1171-1172, fn. 18.) Hamilton thus highlighted the fact that we did not accept or approve of the implied viability limitation, and signaled our intent to address both this issue and the possibility/probability issue if and when presented with the type of facts involved here.

Affirmance of defendant's conviction is fully consistent with our decision in People v. King (1993) 5 Cal.4th 59 [ 19 Cal.Rptr.2d 233, 851 P.2d 27], which, as the majority points out, overruled In re Culbreth (1976) 17 Cal.3d 330 [ 130 Cal.Rptr. 719, 551 P.2d 23] (hereafter Culbreth). In People v. King, supra, 5 Cal.4th 59, the Attorney General argued that our overruling of Culbreth should apply to the defendant because the plain words of the subject statute, together with explicit criticisms of Culbreth by various Courts of Appeal, gave fair warning that we might reconsider our stated position. (5 Cal.4th at pp. 79-81.) We rejected the argument, observing: "The mere possibility that this court might reconsider its own precedent is not the equivalent of actually overruling it." ( Id., at p. 80.)

The situation before us is clearly distinguishable. The challenged jury instruction defining viability in terms of "possible" survival is grounded in the language of two appellate decisions. Prior to the time defendant acted, no appellate court in California had ever held that for purposes of section 187(a), viability means a probability or reasonable likelihood of survival. Finally, Hamilton, supra, 48 Cal.3d at pages 1171-1172, gave ample warning to defendant that we would consider the possibility/probability issue in a case presenting the appropriate facts.

Under these circumstances, defendant cannot credibly claim that the jury instructions given in this case represented an unforeseeable judicial enlargement of section 187(a). Accordingly, defendant's due process rights were not violated by their use.

I would affirm the conviction.

George, J., concurred.

MOSK, J.

I dissent. I believe the Legislature intended the term "fetus" in its 1970 amendment to Penal Code section 187 to mean a viable fetus. I rest this belief on a number of grounds.

I

First, "'The fundamental purpose of statutory construction is to ascertain the intent of the lawmakers so as to effectuate the purpose of the law.'" ( People v. Thomas (1992) 4 Cal.4th 206, 210 [ 14 Cal.Rptr.2d 174, 841 P.2d 159].) Among the relevant evidence of that intent are "Both the legislative history of the statute and the wider historical circumstances of its enactment . . . ." ( Dyna-Med, Inc. v. Fair Employment Housing Com. (1987) 43 Cal.3d 1379, 1387 [ 241 Cal.Rptr. 67, 743 P.2d 1323].)

Here the legislative history of the 1970 amendment to Penal Code section 187 (all unlabeled statutory references are to this code) was unusual, even dramatic. On Friday, June 12, 1970, we filed our decision in Keeler v. Superior Court (1970) 2 Cal.3d 619 [ 87 Cal.Rptr. 481, 470 P.2d 617, 40 A.L.R.3d 420] (hereafter Keeler), holding that the Legislature did not intend that a viable fetus be deemed a "human being" within the meaning of section 187. On the very next working day — Monday, June 15 — the decision was publicly denounced by the majority floor leader of the Assembly. (See Comment, Is the Intentional Killing of an Unborn Child Homicide? (1971) 2 Pacific L.J. 170, 172 fn. 13 [hereafter Comment, Killing of an Unborn Child].) Although the legislative session was more than halfway over and the deadline for introducing new bills had long since passed, the majority leader nevertheless took immediate action to permit the Legislature to overrule Keeler: a fellow legislator relinquished sponsorship of a pending bill on a wholly different subject — Assembly Bill No. 816 — and on June 24 the majority leader "amended" that bill by deleting its original text in its entirety and replacing it with his new version of section 187. In quick succession the bill was then amended twice more in the Assembly (July 10 17), passed and sent to the Senate (July 27), amended in the Senate (Aug. 7), and passed and returned to the Assembly (Aug. 20), where the Senate amendments were concurred in the next day (Aug. 21). In short, the bill was passed in both houses, after multiple amendments and appropriate committee hearings in each house, in the space of barely eight weeks. And the effort had been successful: despite the evolution of its wording, the bill's author "was convinced that it would accomplish his original intent, that is, to make Robert Keeler's actions susceptible to a charge of murder." (Comment, Killing of an Unborn Child, supra, 2 Pacific L.J. at p. 175, fn. omitted.)

To determine the Legislature's intent, therefore, we need to understand precisely what were (1) "Robert Keeler's actions" and (2) this court's response to them. It is black letter law that the holding of a case is determined "'by taking account (a) of the facts treated by the judge as material, and (b) his decision as based on them.'" ( Achen v. Pepsi-Cola Bottling Co. (1951) 105 Cal.App.2d 113, 124 [ 233 P.2d 74].)

The material facts of Keeler, supra, 2 Cal.3d 619, were essentially undisputed. When Robert and Teresa Keeler were granted an interlocutory decree of divorce on September 27, 1968, Teresa was already pregnant by one Ernest Vogt. She began living with Vogt in a different city and concealed the fact from Robert. Under the decree of divorce Robert had custody of their two daughters, but Teresa had the right to take the girls on alternate weekends. Five months later, on February 23, 1969, Robert encountered Teresa driving on a rural road after delivering the girls to their home. He said to her, "I hear you're pregnant. If you are you had better stay away from the girls and from here." ( Id. at p. 623.) When Teresa got out of her car Robert looked at her abdomen and became "extremely upset." Saying, "You sure are. I'm going to stomp it out of you," he struck her in the face and pushed his knee into her abdomen. When doctors subsequently performed a Caesarean section on Teresa the fetus was delivered stillborn. The pathologist was of the opinion that the fetus's fatal injury — a skull fracture — could have been caused by a blow to Teresa's abdomen.

At the time of delivery the fetus weighed five pounds and was eighteen inches in length. In light of these facts Teresa's obstetrician estimated the fetus was 35 weeks old. An attending pediatrician likewise estimated the age of the fetus as between 34 1/2 and 36 weeks. The expert testimony thus "concluded 'with reasonable medical certainty' that the fetus had developed to the stage of viability, i.e., that in the event of premature birth on the date in question it would have had a 75 percent to 96 percent chance of survival." ( Keeler, supra, 2 Cal.3d at p. 624.)

The majority opinion in Keeler, which I authored for the court, demonstrates the materiality of the fact that the fetus was viable by repeatedly incorporating that fact into its legal analyses and conclusions. At the outset we stated that the issue before the court was "whether an unborn but viable fetus is a 'human being' within the meaning of [§ 187]," and concluded that "the Legislature did not intend such a meaning" because the common law had always required a live birth to support a charge of murder. (2 Cal.3d at p. 623, italics added.)

Thereafter the People sharply focused our attention on the importance of viability. The People contended that the common law requirement of live birth was no longer an appropriate test of what is a human being because in modern medicine a fetus born after 28 weeks or more of gestation "has an excellent chance of survival, i.e., is 'viable'"; that an "unborn but viable" fetus is now fully capable of independent life; and that one who unlawfully terminates "such a life" should be prosecuted under section 187. (2 Cal.3d at p. 631.) In response, we did not deny that such fetuses were now viable, but held on two grounds that Robert Keeler could not be prosecuted for murder by reason of his alleged act of killing "an unborn — even though viable — fetus." ( Ibid.)

The first ground was our reluctance to rewrite the murder statute to extend it to this new class of victims. We recognized (2 Cal.3d at p. 632) that the killing of an "unborn but viable" fetus may be deemed an offense as grave as common law murder, but reasoned that to thus extend liability for murder was a matter solely for the Legislature: "For a court to simply declare, by judicial fiat, that the time has now come to prosecute under section 187 one who kills an unborn but viable fetus would indeed be to rewrite the statute under the guise of construing it." ( Id. at p. 633, italics added.)

The second ground was our perception of a due process bar to the charge because Robert Keeler had no notice of such a judicial enlargement of the crime of murder. We found no California case giving him notice that the killing of an "unborn but viable" fetus was covered by section 187. (2 Cal.3d at p. 636.) We distinguished People v. Chavez (1947) 77 Cal.App.2d 621 [ 176 P.2d 92], on the ground that it did not hold that "a fetus, however viable," that was not in the process of being born was a human being under section 187. (2 Cal.3d at p. 637.) And we also distinguished civil law rules protecting the property interests of an unborn child provided it is subsequently born alive, explaining why such rules did not give Robert Keeler notice that the killing of an "unborn but viable" fetus would now be murder. ( Id. at p. 639.)

The dissenting opinion of Acting Chief Justice Burke in Keeler (2 Cal.3d at p. 639) likewise tied its analysis and conclusions closely to the fact of viability. Thus it began by stressing the evidence that the fetus in question "had reached the 35th week of development, had a 96 percent chance of survival, and was 'definitely' alive and viable" at the time of death. ( Id. at pp. 639-640.) The dissent then marshaled a number of arguments in support of its conclusion that such a viable fetus was a human being under section 187.

First, the dissent argued there is common law precedent to support the view that "a viable fetus such as a Baby Girl Vogt [i.e., Teresa Keeler's baby]" is a "human being" under the California homicide statutes. (2 Cal.3d at p. 640.) The dissent noted that the common law severely punished abortion after "quickening," i.e., the time when the fetus is first felt to move in the womb, and reasoned that "we cannot assume that the Legislature intended a person such as defendant, charged with the malicious slaying of a fully viable child, to suffer only the mild penalties imposed upon common abortionists who, ordinarily, procure only the miscarriage of a nonviable fetus or embryo." ( Id. at p. 641, italics added.) The dissent then disclaimed any suggestion, however, that "human being" in section 187 should be defined in terms of the outdated concept of quickening: "The analogous concept of viability is clearly more satisfactory, for it has a well defined and medically determinable meaning denoting the ability of the fetus to live or survive apart from its mother." (2 Cal.3d at p. 641, fn. omitted.)

Next the dissent argued that its view would not judicially create a new offense because the Legislature intended that the term "human being" in section 187 be construed as an evolving concept to be defined by the courts according to contemporary conditions. Noting that under those conditions the Keeler fetus would have had a 98.8 percent chance of survival, the dissent urged that the term "human being" in section 187 should include "the fully viable fetus." (2 Cal.3d at p. 643.)

The dissent also rejected the majority's due process ground, asking rhetorically, "Can defendant really claim surprise that a 5-pound, 18-inch, 34-week-old, living, viable child is considered to be a human being?" (2 Cal. 3d at p. 644, italics added.) In view of the decision in People v. Chavez, supra, 77 Cal.App.2d 621, the dissent found that "defendant had sufficient notice that the words 'human being' could include a viable fetus." (2 Cal.3d at p. 645, italics added.)

In sum, the dissent concluded, "There is no good reason why a fully viable fetus should not be considered a 'human being' under [§ 187]." (2 Cal.3d at p. 645, italics added.)

It was thus obvious to any reader of Keeler in 1970, as it is obvious today, that the case held that the Legislature did not intend that Robert Keeler's act of maliciously killing an "unborn but viable" fetus be prosecuted as murder under section 187. That was the holding that motivated the Legislature to act with such unusual speed upon publication of the Keeler opinions, and it was also the holding that Assembly Bill No. 816 was intended to, and did, overrule: "[Section] 187 was amended in 1970 to nullify the Keeler holding." (1 Witkin Epstein, Cal. Criminal Law (2d ed. 1988) Crimes Against the Person, § 450, p. 507.) It follows that by enacting the 1970 amendment to section 187 the Legislature extended the crime of murder, as Keeler refused to do, to include the malicious killing of a viable fetus. To read that amendment as further extending murder to include the killing of even a nonviable fetus, as the lead opinion does now, is to ignore the facts and holding of Keeler and the direct legislative response they so plainly triggered.

II

Another rule of statutory construction supports this conclusion.

The first case to construe the 1970 amendment to section 187 was People v. Smith ( Karl Andrew) (1976) 59 Cal.App.3d 751 [ 129 Cal.Rptr. 498] (hereafter K.A. Smith). There, as in Keeler, supra, 2 Cal.3d 619, the defendant assaulted a pregnant woman with the specific intent of terminating her pregnancy, and caused her to miscarry; the difference was that the woman was only 12 to 15 weeks pregnant at the time. "It was stipulated that the product of conception was nonviable." (59 Cal.App.3d at p. 754.) The trial court ruled that only a viable fetus could be an object of murder under the 1970 amendment, and therefore dismissed a murder charge. On the People's appeal the Court of Appeal affirmed the ruling, and we denied review. The Court of Appeal devoted its entire opinion to the issue; although its rationale differed from my reasoning herein, the Court of Appeal squarely held that "we construe the word 'fetus' in section 187 to refer to a viable unborn child." ( Id. at p. 759.) This construction was incorporated into the approved jury instruction defining fetal murder, which declares that the killing of "a viable human fetus" is an element of the crime and defines such viability. (CALJIC No. 8.10 (5th ed. 1988).)

All subsequent cases involving fetal murder in this state have likewise held or assumed that viability is an element of the crime. ( People v. Hamilton (1989) 48 Cal.3d 1142, 1171-1173 [ 259 Cal.Rptr. 701, 774 P.2d 730]; People v. Henderson (1990) 225 Cal.App.3d 1129, 1157 [ 275 Cal.Rptr. 837] ["The 'fetus' described in the statute means only a 'viable' fetus."]; People v. Smith ( Robert Porter) (1987) 188 Cal.App.3d 1495, 1512-1516 [ 234 Cal.Rptr. 142]; People v. Apodaca (1978) 76 Cal.App.3d 479, 487-490 [ 142 Cal.Rptr. 830].)

The lead opinion reviews the cases just cited, but overlooks People v. Bunyard (1988) 45 Cal.3d 1189 [ 249 Cal.Rptr. 71, 756 P.2d 795]. There the defendant was convicted of double murder when he hired another to kill his wife, knowing she was pregnant. As in Keeler, supra, 2 Cal.3d 619, the fetus was undoubtedly viable at the time — indeed, it was full term; there was thus no issue as to whether viability was an element of fetal murder under section 187. Yet the court plainly implied that it was an element of that crime: also as in Keeler, our opinion in Bunyard repeatedly incorporated the fact of the fetus's viability into its legal analyses and conclusions. For example, in rejecting a contention that the multiple-murder special-circumstance statute (§ 190.2, subd. (a)(3)) is ambiguous because it does not expressly refer to the murder of "a fetus" as does section 187, we said that when the two statutes are read together it is clear that the special circumstance applies to the killing, even by a single act, "of a pregnant woman and her viable fetus." (45 Cal.3d at pp. 1237-1238, italics added, fn. omitted; see also id. at p. 1239.) We next rejected a contention that to make the defendant death-eligible because of the murder of a fetus would violate the state cruel or unusual punishment clause. (Cal. Const., art. I, § 17.) We explained that "The fact that the victim murdered is an unborn child does not render defendant less culpable, or the crime less severe, in light of the Legislature's determination that viable fetuses receive the same protection under the murder statute as persons." (45 Cal.3d at p. 1240, italics added.) We then rejected an Eighth Amendment claim on the ground, inter alia, that the defendant failed to show arbitrariness in charging practices or sentencing decisions "with respect to the imposition of the death penalty for the murder of a pregnant woman and her viable fetus." ( Ibid., italics added.) Finally, we found unpersuasive the defendant's argument that California is one of only three states in which an accompanying feticide can raise a noncapital murder of a pregnant woman to capital status. We explained that the argument fails to consider the distinction that, unlike section 187, the laws of most other jurisdictions do not "make the unlawful killing of a viable fetus a homicide, or a murder, at all." (45 Cal.3d at p. 1241, italics added.)

The lead opinion correctly asserts that until the case at bar "every prior decision that had addressed the viability issue had determined that viability of the fetus was prerequisite to a murder conviction under section 187, subdivision (a)" (lead opn., ante, p. 802). The lead opinion invokes this fact to support its conclusion that to apply its new and contrary construction of section 187 to this defendant would violate his due process and ex post facto rights. But the fact also leads logically to another and equally compelling conclusion. Since the decision in K.A. Smith, supra, 59 Cal.App.3d 751, holding that viability is an element of fetal murder under section 187, the Legislature has met 18 times but has taken no step to overrule that holding by what would be a simple amendment to the statute.

The Legislature well knows how to delete statutory elements added by judicial construction, particularly on the topic of murder. For example, section 187 defines murder as a killing with malice, and section 188 provides statutory definitions of both express and implied malice. In People v. Conley (1966) 64 Cal.2d 310, 322 [ 49 Cal.Rptr. 815, 411 P.2d 911] (hereafter Conley), this court declared that "An awareness of the obligation to act within the general body of laws regulating society . . . is included in the statutory definition" of both kinds of malice. This construction of section 188 was consistently followed by our courts (see, e.g., People v. Sedeno (1974) 10 Cal.3d 703, 723 [ 112 Cal.Rptr. 1, 518 P.2d 913]; People v. Poddar (1974) 10 Cal.3d 750, 758 [ 111 Cal.Rptr. 910, 518 P.2d 342]; People v. Carpenter (1979) 99 Cal.App.3d 527, 533-534 [ 160 Cal.Rptr. 386]; People v. Fusselman (1975) 46 Cal.App.3d 289, 303 [ 120 Cal.Rptr. 282]), and was incorporated into the approved jury instructions (CALJIC No. 8.11 (4th ed. 1979)). In 1981, however, the Legislature expressly repudiated the Conley construction by adding a sentence to section 188 declaring that "An awareness of the obligation to act within the general body of laws regulating society is not included within the definition of malice." (Stats. 1981, ch. 404, § 6, p. 1593, italics added; see People v. Saille (1991) 54 Cal.3d 1103, 1113-1114 [ 2 Cal.Rptr.2d 364, 820 P.2d 588].)

I recognize that "Legislative silence after a court has construed a statute gives rise at most to an arguable inference of acquiescence or passive approval" ( People v. Daniels (1969) 71 Cal.2d 1119, 1127 [ 80 Cal.Rptr. 897, 459 P.2d 225, 43 A.L.R.3d 677]), and that the presumption of such acquiescence "is not conclusive in determining legislative intent" ( Harris v. Capital Growth Investors XIV (1991) 52 Cal.3d 1142, 1156 [ 278 Cal.Rptr. 614, 805 P.2d 873]). But in appropriate circumstances it may at least be persuasive. For example, in the recent case of Freedom Newspapers, Inc. v. Orange County Employees Retirement System (1993) 6 Cal.4th 821, 831 [ 25 Cal.Rptr.2d 148, 863 P.2d 218], in determining the intent of the Legislature when it enacted a statute in 1968 we gave weight to the fact that "the Legislature has allowed the Court of Appeal's opinion in [a single 1978 decision construing the statute] to govern [the application of the statute] for the past 14 years." Here viability as an element of fetal murder has been recognized in multiple published opinions and in approved jury instructions, and the Legislature has acquiesced in that construction even longer than 14 years.

More important, the statutory language in issue here — the 1970 amendment to section 187 extending the crime of murder to the killing of "a fetus" — was itself enacted in direct and vigorous response to a judicial opinion ( Keeler) with which the Legislature disagreed. If the Legislature had also disagreed a few years later with subsequent judicial opinions ( K.A. Smith and its progeny) limiting the statutory prohibition against killing "a fetus" to the killing of a viable fetus, surely it would have spoken again, and equally vigorously. To this day, however, the Legislature has remained silent and taken no remedial action. In these circumstances its acquiescence is persuasive evidence of its intent.

III

The lead opinion is curiously proportioned. It devotes over nine-tenths of its discussion of the main issue — i.e., whether viability is an element of fetal murder — to the relatively easy task of refuting the rationale of K.A. Smith, supra, 59 Cal.App.3d 751. (Lead opn., ante, pp. 803-810; see fn. 1, ante.) But the lead opinion spends only a few lines on the much harder task of determining the Legislature's intent when it amended section 187 in 1970. Indeed, the lead opinion essentially gives up on that task, saying the Legislature had no intent at all with regard to the meaning of the key word of the amendment: "The legislative history of the amendment suggests the term 'fetus' was deliberately left undefined after the Legislature debated whether to limit the scope of statutory application to a viable fetus." (Lead opn., ante, p. 803.) This statement implies that "the Legislature" as a whole conducted a full-scale "debate" on the question whether the amendment should be limited to viable fetuses, and that it was precisely because of that debate that the Legislature "deliberately" declined to adopt that limitation. This is a gross exaggeration of the legislative history. As its sole support the lead opinion cites Comment, Killing An Unborn Child, supra, 2 Pacific L.J. 170, 174. But that source says only that when the bill that became the 1970 amendment was heard in a committee of one of the houses, one of the committee members opposing it argued as one of his grounds of opposition that it should be expressly limited to viable fetuses. ( Id. at pp. 173-174.) This is a far cry from a full-scale debate and decision on the matter by the Legislature. Even a formal statement by the author of a bill after it is passed is entitled to weight only "when it is a reiteration of legislative discussion and events leading to adoption of proposed amendments rather than merely an expression of personal opinion." ( California Teachers Assn. v. San Diego Community College Dist. (1981) 28 Cal.3d 692, 700 [ 170 Cal.Rptr. 817, 621 P.2d 856].) When, as here, it is merely the expression of a personal position — a fortiori when voiced merely by an opponent of the bill in a committee hearing — it "is not a proper subject for consideration in determining the Legislature's intent" ( id. at p. 701). This is primarily because such expressions of opinion "do not necessarily reflect the views of other members [of the Legislature] who voted for [the bill]." ( Freedom Newspapers, Inc. v. Orange County Employees Retirement System, supra, 6 Cal.4th 821, 831.) The lead opinion's reasoning violates these settled rules of construction.

The lead opinion next asserts that "The Legislature was clearly aware that it could have limited the term 'fetus' to 'viable fetus,' for it specifically rejected a proposed amendment that required the fetus be at least 20 weeks in gestation before the statute would apply." (Lead opn., ante, p. 803.) Here the lead opinion does not simply exaggerate the legislative history, it completely misreads it. It is true that in its original version the bill that became the 1970 amendment — Assembly Bill No. 816 as amended in the Assembly on June 24, 1970 — provided that "As used in this section, 'human being' includes a fetus which has advanced to or beyond the 20th week of uterogestation." It is also true that the quoted language was deleted from the bill when it was again amended in the Assembly 23 days later, on July 17, 1970.

Contrary to the lead opinion's assertion, however, the quoted language had nothing to do with viability and everything to do with abortion and "quickening." It had nothing to do with viability because of the undisputed medical fact that fetuses are not viable at 20 weeks. Rather, the 20-week cutoff date was inserted into the first version of Assembly Bill No. 816 for the wholly different purpose of avoiding a conflict with the Therapeutic Abortion Act, adopted only three years earlier. (Health Saf. Code, § 25950 et seq.) One provision of that act declared that "In no event shall the termination [of a pregnancy] be approved after the 20th week of pregnancy." (Health Saf. Code, § 25953.) By incorporating that same cutoff date into the proposed amendment to section 187, Assembly Bill No. 816 simply sought to make it clear that the fetal murder statute would not apply to therapeutic abortions. This purpose was made even clearer when the same amendment (July 17, 1970) that deleted the 20-week proviso from Assembly Bill No. 816 substituted therefor a proviso having the same effect but expressly referring to the Therapeutic Abortion Act, declaring that the fetal murder statute did not apply to any act that "complied with the Therapeutic Abortion Act, Chapter 11 (commencing with Section 25950) of Division 20 of the Health and Safety Code." (Now § 187, subd. (b)(1).)

Having erroneously concluded that the Legislature had no intent with respect to the meaning of the key word "fetus" in the 1970 amendment to section 187, the lead opinion proceeds to legislate on the subject by supplying the assertedly missing definition: "a fetus," says the lead opinion, "is defined as 'the unborn offspring in the postembryonic period, after major structures have been outlined.' (Sloane-Dorland Ann. Medical-Legal Dict. (1987) p. 281.) This period occurs in humans 'seven or eight weeks after fertilization' ( ibid.). . . ." (Lead opn., ante, p. 810, italics added.) The lead opinion repeats its new definition in concluding that the malicious killing of a fetus is murder under section 187 as long as the state can show the fetus has progressed "beyond the embryonic stage of seven to eight weeks." (Lead opn., ante, p. 815.) For a number of reasons it is highly unlikely that such was the Legislature's intent.

As a preliminary matter, it is obvious that the Legislature would not have defined the fetal stage, as does the lead opinion, as beginning "seven or eight" weeks after conception. Perhaps doctors may speak so imprecisely, but legislators may not: the bench and bar need to know whether liability for fetal murder under section 187 attaches at seven weeks after conception or at eight weeks after conception; the choice cannot be left to the whim of the prosecuting authorities, especially when, as here, it may literally make the difference between life and death. We cannot imagine the Legislature prescribing, for example, that "Children under the age of 13 or 14" are incapable of committing crime (see Pen. Code, § 26, par. One), or — even more relevant here — "the death penalty shall not be imposed upon any person who is under the age of 17 or 18 at the time of the commission of the crime" (see id., § 190.5, subd. (a)), or — to go from the sublime to the ridiculous — "no person shall drive a vehicle upon a highway at a speed greater than 54 or 55 miles per hour" (see Veh. Code, § 22349). The lead opinion's careless statement that liability for fetal murder attaches "seven or eight" weeks after conception is no less improper.

Yet that is the least of the problems with the lead opinion's new definition of "fetus" in section 187. Because liability after seven weeks necessarily includes liability after eight weeks, we may fairly assume that prosecutors faced with the lead opinion's imprecise definition will opt for the more inclusive figure and charge murder when the fetal death occurs at seven weeks. Do my colleagues have any idea what a seven-week-old product of conception looks like?

To begin with, it is tiny. At seven weeks its "crown-rump length" — the only dimension that can be accurately measured — is approximately 17 millimeters, or slightly over half an inch. (Arey, Developmental Anatomy (1965) p. 104.) It weighs approximately three grams, or about one-tenth of an ounce. (Eichler, Atlas of Comparative Embryology (1978) p. 186.) In more familiar terms, it is roughly the size and weight of a peanut.

If this tiny creature is examined under a magnifying glass, moreover, its appearance remains less than human. Its bulbous head takes up almost half of its body and is bent sharply downward; its eye sockets are widely spaced; its pug-like nostrils open forward; its paddle-like hands and feet are still webbed; and it retains a vestigial tail. As Professor Cynthia R. Daniels aptly put it in her recent book, At Women's Expense, supra, page 20, "In fact, at eight weeks most features of the fetus are so ill defined that it would be difficult for an uninformed observer to recognize it, viewed in its entirety, as distinctly human." And as concluded in the Comment relied on by the lead opinion, "A being so alien to what we know to be human beings seems hardly worth being made the subject of murder" (Comment, Killing of an Unborn Child, supra, 2 Pacific L.J. at p. 185).

The contrast between such a tiny, alien creature and the fully formed "5-pound, 18-inch, 34-week-old, living, viable child" in Keeler (2 Cal.3d at p. 644 (dis. opn. of Burke, Acting C.J.)) is too obvious to be ignored. I can believe that by enacting the 1970 amendment the Legislature intended to make it murder to kill a fully viable fetus like Teresa Keeler's baby. But I cannot believe the Legislature intended to make it murder — indeed, capital murder — to cause the death of an object the size of a peanut.

IV

It is even more unlikely that the Legislature intended many of the consequences of the lead opinion's new definition of "fetus" in section 187. That definition should therefore be reconsidered in light of the settled rule that "When uncertainty arises in a question of statutory interpretation, consideration must be given to the consequences that will flow from a particular interpretation. [Citation.] In this regard, it is presumed the Legislature intended reasonable results consistent with its expressed purpose, not absurd consequences. [Citations.]" ( Harris v. Capital Growth Investors XIV, supra, 52 Cal.3d 1142, 1165-1166.)

First, under the lead opinion's definition a person may be subject to a conviction of capital murder for causing the death of an object that was literally invisible to everyone, and hence that the person had no reason to know even existed. A woman whose reproductive system contains an immature fetus a fraction of an inch long and weighing a fraction of an ounce does not, of course, appear pregnant. In fact, if she is one of many women with some irregularity in her menstrual cycle, she herself may not know she is pregnant: "quickening" does not occur until two or three months later. ( Keeler, supra, 2 Cal.3d at p. 625, fn. 5.) Unless such a woman knows she is pregnant and has disclosed that fact to the defendant, the defendant has no way of knowing she is carrying a fetus.

Nor is this problem limited to fetuses that are "seven or eight" weeks old. Although the length of time that a woman can be pregnant without her condition's becoming noticeable varies according to such factors as her height and weight, the size of her fetus, and even the style of her clothing, the case at bar demonstrates that it can extend well into her pregnancy. Here Flores testified that in her opinion her pregnancy "showed" on the date of the shooting, March 1, 1991; but defendant testified to the contrary, and there was persuasive evidence to support him. It was undisputed that Flores was only an inch over five feet in height, yet on February 5, 1991, at her last visit to her obstetrician, she weighed 191 pounds. While she was not weighed again on the date of the shooting, at that stage of her pregnancy she would have gained still more weight in the intervening three and one-half weeks. On this record Thomas Moore, M.D., an experienced perinatologist, testified that in his opinion it is "not likely" that on the date of the shooting a woman of Flores's stature would have showed her pregnancy when clothed and standing upright.

The jury evidently agreed with Dr. Moore. In both count 2 (attempted premeditated murder) and count 3 (robbery) the information alleged the five-year enhancement provided in section 12022.9. The elements of that enhancement are that during the commission of a felony a defendant who (1) "knows or reasonably should know that the victim is pregnant" (2) intentionally inflicts injury on her (3) that results in termination of her pregnancy. The instructions emphasized that to find the allegation true the jury must find "the necessary mental state . . . that defendant knew or reasonably should have known that the victim was pregnant." The jury returned verdicts finding the allegation untrue on both counts. Because the second and third elements of the allegation were undisputed and found true by the jury in other verdicts, it follows that the only ground on which the jury could have found the allegation untrue was that defendant neither knew nor reasonably should have known that Flores was pregnant.

Yet the expert testimony agreed that Flores was between 23 and 25 weeks — approximately 6 months — pregnant on the date of the shooting. This is the very threshold of viability: an expert witness reported on a recent study showing that at 23 weeks the survival rate of the fetus is approximately 7 percent, at 24 weeks 35 percent, and at 25 weeks 47 percent. The case at bar thus demonstrates how long the risk of liability for fetal murder may run under the lead opinion's view before the actor either knows or has reason to know that the victim of the offense even exists. I cannot believe the Legislature intended such an enlargement of liability for the crime of capital murder.

Nor is the problem solved by the fact that section 187 imposes such liability only if the defendant kills the fetus "with malice aforethought." Despite the plain wording of the statute, the People may obtain a fetal murder conviction in many cases without proving malice at all: I refer, of course, to cases brought on a theory of felony murder. I turn to that theory.

Although no one would realize it from reading the lead opinion, the case at bar is truly novel: it is the first reported case in California in which a person has been convicted of fetal murder without knowing or having reason to know of the existence of the victim. In all but one of the prior cases the defendant knew very well that the woman he assaulted was pregnant; he intended at least to kill the fetus, and thus manifested express malice toward the fetus within the meaning of section 187. (See People v. Saille, supra, 54 Cal.3d 1103, 1114 [express malice is "an intent unlawfully to kill"].)

Two of these cases closely resembled the facts of Keeler, supra, 2 Cal.3d 619: in each, the defendant knew that his wife or former wife was pregnant, and assaulted her for the sole purpose of putting an end to the pregnancy; although he did not intend to kill the woman, he plainly intended to kill the fetus. ( People v. Apodaca, supra, 76 Cal.App.3d 479, 483-484; K.A. Smith, supra, 59 Cal.App.3d 751, 753-754.)

Three other cases presented a related scenario in which the husband, or another acting on his behalf, knew the wife was pregnant and killed her in order to be rid of both her and the fetus in one stroke, thus manifesting express malice towards both victims. ( People v. Hamilton, supra, 48 Cal.3d 1142, 1151-1153; People v. Bunyard, supra, 45 Cal.3d 1189, 1200-1201; People v. Smith ( Robert Porter), supra, 188 Cal.App.3d 1495, 1499-1504.) In the last cited case, for example, Smith killed Annie, the wife of his lifelong friend, Skaggs, knowing she was six to seven months pregnant. In finding the requisite intent to kill, the court reasoned: "the very purpose of the murder . . . was to free Skaggs of the encumbrance of a wife and a child so that Skaggs could pursue a ministerial career. . . . The very motivation for the murder was the pregnancy itself. This was not a murder of an expectant mother the murderer did not know and could not recognize as pregnant. If the jury convicted Smith of first degree murder of both Annie and of baby Skaggs, as it did, Smith necessarily had the intent to kill both mother and child." (188 Cal.App.3d at p. 1511, second italics added.)

I have no difficulty believing that by enacting the 1970 amendment the Legislature intended to punish such intentional killing as fetal murder. But that is not this case: here defendant did not intend to kill the Flores fetus, for the simple reason that, as the jury found, he neither knew nor had reason to know it existed. Indeed, he did not even intend to kill Flores herself: although he was charged in count 2 with the deliberate and premeditated attempted murder of Flores, the jury acquitted him of that charge, convicting him instead of the lesser included offense of assault with a firearm. (§ 245, subd. (a)(2).)

Lacking proof of both malice and premeditation, the prosecutor sought a murder conviction on a theory that did not require him to prove defendant intended to kill the invisible fetus — the theory of felony murder. That theory had been invoked in the only prior fetal murder case that did not involve an intentional killing of the fetus, People v. Henderson, supra, 225 Cal.App.3d 1129. There the defendant Philip Henderson robbed and killed Ray and Angie Boggs; Angie was pregnant and her death caused the death of her fetus. The defendant was convicted of robbery and three counts of murder, and the murder convictions were upheld on a robbery felony-murder theory. But the case is distinguishable in two respects.

First, the defendant must have been well aware that Angie was pregnant. At the time of her death her fetus was 30 weeks — 7 1/2 months — old: it was therefore both viable and visible. And, contrary to the case at bar, Angie was not a stranger whom the defendant had encountered only moments before the murders; rather, the defendant had moved in with Angie and her husband and had lived with them in their apartment for some six weeks before robbing and killing them. Second, although the jury convicted the defendant of two counts of first degree murder in the deaths of Ray and Angie, they convicted him of only second degree murder in the death of Angie's fetus. Yet under the felony-murder rule the murder of the fetus was murder in the first degree as a matter of law no less than the murders of its parents (§ 189), and the jury were doubtless so instructed. By declining to follow that law and returning instead a verdict of murder in the second degree, the jury apparently exercised their power of nullification. (See generally, People v. Dillon (1983) 34 Cal.3d 441, 490-493 [ 194 Cal.Rptr. 390, 668 P.2d 697] (conc. opn. of Kaus, J.).) In so doing, the jury were "obviously looking for a way to avoid the harsh consequences of the felony-murder rule." ( Id. at p. 490.)

But if the result was harsh in People v. Henderson, supra, 225 Cal.App.3d 1129, it will be even harsher under the lead opinion's new definition of "fetus" in section 187. Under that definition the felony-murder rule will make a capital offense out of the death of even a nonviable and invisible fetus that the actor neither knew nor had reason to know existed. As will appear, this is such a case.

Here the prosecutor left no doubt that his theory was felony murder: as he explained to the jury in closing argument, "In some homicide cases there are many different theories of murder that may be involved, depending on the facts in the case. In this particular case there is only one murder theory, and that is what we call felony murder. . . ."

It has been clear at least since People v. Dillon, supra, 34 Cal.3d 441, 475, that "as a matter of law malice [including intent to kill] is not an element of felony murder." Again the prosecutor made sure the jury understood the implications of that rule: he stressed in argument that "the question then becomes, did the killing occur during the commission or attempted commission of a robbery. The first thing I want you to note in doing this is when we are talking murder in this particular setting, it does not require an intent. . . . It doesn't matter if that death was intentional, unintentional or accidental."

The instructions confirmed the prosecutor's advice. The court first told the jury that to convict defendant of murder in this case the prosecution need prove only that "1. A viable human fetus . . . was killed. [¶] 2. The killing was unlawful. [¶] 3. The killing occurred during the commission or attempted commission of a robbery." There was no need to prove intent to kill or any other kind of malice aforethought. Nor was it necessary that the fetus be the object of the underlying felony. (See, e.g., People v. Welch (1972) 8 Cal.3d 106, 118-119 [ 104 Cal.Rptr. 217, 501 P.2d 225].)

On the record in this case an affirmative finding of the third element — i.e., that the killing of the fetus occurred during the commission of a robbery or attempted robbery — was foreordained, because defendant admitted in his direct testimony that he intended to rob Flores and his counsel conceded the fact in closing argument. But the finding not only made the killing felony murder, it had two additional consequences. As its second consequence, it fixed the degree of the murder at the first degree (§ 189) even though, as explained above, the jury found that defendant did not intend to kill anyone: thus the court further instructed the jury, "The unlawful killing of a viable human fetus, . . . whether intentional, unintentional or accidental, which occurs during the commission or attempted commission of a robbery, is murder in the first degree when the perpetrator had the specific intent to commit such robbery." (Italics added.) And as its third consequence, the same finding also constituted a "special circumstance" increasing the punishment from a term of 25 years to life with possibility of parole (§ 190, subd. (a)) to a sentence of death or life imprisonment without possibility of parole (§ 190.2, subd. (a)).

Defendant was sentenced to life imprisonment without possibility of parole. He was 20 years old at the time of the events. He had no prior felony convictions, and only two minor juvenile offenses which the prosecutor himself characterized as an "insignificant criminal history." Viewing the statutory penalty of life imprisonment without possibility of parole in the light of these circumstances, the prosecutor conceded, "It seems in a sense like a waste of a young life and to some degree it must be considered that." The court agreed that defendant "has an insignificant criminal background," and echoed the prosecutor's view of the sentence by saying, "one of the emotional issues that underlies the sentence here is that it is so sad to see a young life thrown away. And that's what's happened here with Mr. Davis."

This draconian result was possible, of course, only because of the draconian nature of the felony-murder rule itself. As we observed in People v. Dillon, supra, 34 Cal.3d at pages 476-477, footnote omitted, "the two kinds of first degree murder in this state differ in a fundamental respect: in the case of deliberate and premeditated murder with malice aforethought, the defendant's state of mind with respect to the homicide is all-important and must be proved beyond a reasonable doubt; in the case of first degree felony murder it is entirely irrelevant and need not be proved at all. From this profound legal difference flows an equally significant factual distinction, to wit, that first degree felony murder encompasses a far wider range of individual culpability than deliberate and premeditated murder. It includes not only the latter, but also a variety of unintended homicides resulting from reckless behavior, or ordinary negligence, or pure accident; it embraces both calculated conduct and acts committed in panic or rage, or under the dominion of mental illness, drugs, or alcohol; and it condemns alike consequences that are highly probable, conceivably possible, or wholly unforeseeable."

Under the lead opinion's new definition of "fetus" another category will be added to this litany: first degree felony murder will be extended to include any death, in the commission of a listed felony, of a nonviable and invisible fetus that the actor neither knew nor had reason to know existed. And that liability will not be limited to cases in which, as here, the defendant intentionally assaults the pregnant woman with a deadly weapon. It will extend, for example, to the following scenario: an unarmed 18-year-old with no criminal record enters a store during business hours, intending to shoplift a can of spray paint; when a security guard accosts him, his nerve fails and he bolts for the door; in his haste he accidentally knocks a woman shopper to the floor; unknown to anyone the woman is 7 weeks' pregnant, and the trauma of the fall causes her to miscarry.

Before today's decision, such a youth would be guilty at most of second degree burglary (§§ 459, 460; see, e.g., People v. Corral (1943) 60 Cal.App.2d 66, 70-71 [ 140 P.2d 172], and cases cited) and would be punishable by either county jail time of up to one year or a prison term of sixteen months or two or three years (§§ 18, 461); because the fetus was not viable, its death would not have been murder on any theory. After today's decision, however, this teenager could also be found guilty of first degree murder of the fetus on a burglary felony-murder theory (cf. People v. Earl (1973) 29 Cal.App.3d 894, 898-900 [ 105 Cal.Rptr. 831]); in that event his punishment would at least be imprisonment for 25 years to life (§§ 190, subd. (a)), and he would be subject to a burglary special-circumstance allegation which, if found true, would require the punishment of either death or life imprisonment without possibility of parole (§ 190.2, subd. (a)(17)(vii)).

I cannot believe the Legislature intended the 1970 amendment to accomplish so absurd a result — especially when that amendment affected a statute that requires a finding of "malice aforethought" to support a conviction (§ 187) and had nothing to do with the statute that defines felony murder (§ 189). Yet this is where the lead opinion's definition of "fetus" inexorably takes us.

In addition, by thus expanding the scope of the felony-murder rule the lead opinion violates another settled policy of this court: almost three decades ago, in the leading case of People v. Washington (1965) 62 Cal.2d 777, 783 [ 44 Cal.Rptr. 442, 402 P.2d 130], Chief Justice Traynor stressed that "The felony-murder rule has been criticized on the grounds that in almost all cases in which it is applied it is unnecessary and that it erodes the relation between criminal liability and moral culpability. [Citations.] Although it is the law in this state (Pen. Code, § 189), it should not be extended beyond any rational function that it is designed to serve." (Italics added, fn. omitted.) We have respected and reiterated his admonition until this day (see, e.g., People v. Dillon, supra, 34 Cal.3d at pp. 462-463, citing numerous cases), most recently in People v. Patterson (1989) 49 Cal.3d 615, 621-622 [ 262 Cal.Rptr. 195, 778 P.2d 549]. As the foregoing scenario demonstrates, however, to apply the felony-murder rule to the unintentional death of a nonviable and invisible fetus that the actor neither knew nor had reason to know existed will clearly extend the rule beyond any rational function it is designed to serve.

The cases also illustrate that irrational punishments may provoke one of two undesirable responses. First, as apparently occurred in People v. Henderson, supra, 225 Cal.App.3d 1129, the jury may be driven to exercise its power of nullification. Second, as we did in People v. Dillon, supra, 34 Cal.3d at pages 477-489, a court may be compelled to hold that on the facts of the case a first degree murder conviction is a violation of the state cruel or unusual punishment clause. I cannot believe the Legislature intended to encourage either response by the 1970 amendment to section 187.

Next, the lead opinion's new definition of "fetus" will raise difficult questions of causation in prosecutions for murder of a nonviable fetus. Although a conviction on a felony-murder theory does not require proof of a "strict causal relation" between the felony and the killing (see, e.g., People v. Welch, supra, 8 Cal.3d 106, 118), some act of the defendant during the commission of the felony must at least have been a "substantial factor" contributing to the death (e.g., People v. Stamp (1969) 2 Cal.App.3d 203, 210 [ 82 Cal.Rptr. 598]), just as is required for convictions on a theory of express or implied malice (e.g., People v. Caldwell (1984) 36 Cal.3d 210, 220 [ 203 Cal.Rptr. 433, 681 P.2d 274]). Here the problem is medical reality: in lowering to seven weeks the age of a fetus whose death can trigger a murder prosecution, the lead opinion overlooks the fact that the more immature the fetus, the more likely it is to die in any event by spontaneous abortion. Spontaneous abortions are caused by a variety of factors, including genetic or developmental defects in the fetus, uterine abnormalities, maternal trauma, illness, or substance abuse, toxins in the fetal or maternal environment, etc. And they are much more common than is generally realized, particularly in the early stages of pregnancy: "The incidence of spontaneous abortion is generally believed to be 15-20% of all pregnancies. Substantial numbers of [spontaneous] abortions, however, are unreported or are very early and subclinical; some have estimated the true incidence to be as high as 50-78% [of all pregnancies]." (Rogge, Spontaneous Abortion, in Manual of Obstetrics (Niswander edit. 1987) p. 213, fn. omitted.)

Accordingly, the mere fact that a fetus aborts at some time after the woman carrying it is intentionally or unintentionally struck by another does not necessarily mean the act was a "substantial factor" in causing the fetal death. In cases in which the fetus was nonviable and immature the prosecutor could well have difficulty in proving causation, or, conversely, the jury could be misled into convicting the actor of fetal murder when no such murder in fact occurred. Again I cannot believe the Legislature intended either result by the 1970 amendment to section 187.

V

Finally, the lead opinion's construction of the 1970 amendment will make our murder law unique in the nation in its severity: it appears that in no other state is it a capital offense to cause the death of a nonviable and invisible fetus that the actor neither knew nor had reason to know existed.

To begin with, in the majority of states the killing of a fetus is not a homicide in any degree: "The majority of jurisdictions which have confronted the issue has followed Keeler, 2 Cal.3d 619, . . . in holding the term 'fetus' does not fall within the definition of a human being under criminal statutes unless the term is so defined by the legislature." ( State v. Trudell (1988) 243 Kan. 29 [ 755 P.2d 511, 515].) In those jurisdictions a live birth remains a prerequisite to a conviction of homicide. (See Meadows v. State (1987) 291 Ark. 105 [ 722 S.W.2d 584, 585-586], citing numerous cases; for later cases, see, e.g., Billingsley v. State (1987) 183 Ga. App. 850 [ 360 S.E.2d 451, 452]; State v. Keller (La. Ct. App. 1991) 592 So.2d 1365, 1366; State v. Beale (1989) 324 N.C. 87 [ 376 S.E.2d 1, 2-4]; State v. Evans (Tenn. Crim. App. 1987) 745 S.W.2d 880, 881-884.)

There are, of course, jurisdictions that have enacted statutes criminalizing the killing of a fetus. The lead opinion cites seven such jurisdictions. (Lead opn., ante, p. 808, fn. 1.) My research has turned up at least 22, with 2 additional jurisdictions so holding as a matter of common law. For convenience I have grouped these jurisdictions into three distinct categories.

First, in at least 13 jurisdictions the killing of a fetus is not criminal unless the fetus is viable or has reached a gestational age significantly more advanced than the "seven or eight weeks" prescribed by the lead opinion. Thus in Iowa any person who intentionally kills a fetus "after the end of the second trimester of the pregnancy" commits the crime of "feticide." (Iowa Code Ann. § 707.7.) In New York, homicide includes the killing of "a person or an unborn child with which a female has been pregnant for more than twenty-four weeks" (N.Y. Penal Law § 125.00). In England, where the common law rule that fetal murder is not a crime first arose, a statute now provides that any person who intentionally causes a fetal death commits the crime of "child destruction" provided the fetus was "capable of being born alive," and the fact that the woman had been "pregnant for a period of twenty-eight weeks or more" is prima facie proof that the fetus was viable. (Infant Life (Preservation) Act, 1929, 19 20 Geo. 5, ch. 34, § 1.)

In South Carolina the state's highest court has adopted, pursuant to its common law power to declare substantive criminal law, the rule that the intentional killing of a fetus is murder if the fetus was viable. ( State v. Horne (1984) 282 S.C. 444 [ 319 S.E.2d 703, 704].) The highest court in Massachusetts has not held in two cases in which the fetus was in fact viable, but has not yet taken a position on whether viability is a legal prerequisite to such liability. ( Commonwealth v. Lawrence (1989) 404 Mass. 378 [ 536 N.E.2d 571, 575, fn. 6] [murder]; Commonwealth v. Cass (1984) 392 Mass. 799 [ 467 N.E.2d 1324, 1329, fn. 8] [vehicular homicide].)

Eight states have enacted statutes criminalizing the killing of "an unborn quick child." In three of those states the act is criminal if the unborn child is killed by "any" injury to the mother. (Nev. Rev. Stat. § 200.210; Okla. Stat. Ann. tit. 21, § 713; Wash. Rev. Code Ann. § 9A.32.060.) In the remaining five states, however, the act is criminal only if the unborn child is killed "by any injury to the mother of such child which would be murder if it resulted in the death of such mother." (Fla. Stat. Ann. § 782.09; Ga. Code Ann. § 16-5-80; Mich. Comp. Laws Ann. § 750.322; Miss. Code Ann. § 97-3-37; R.I. Gen. Laws Ann. § 11-23-5.) Yet in seven of those states — all except Georgia — the resulting fetal homicide is punished only as manslaughter, while in Georgia it is deemed "feticide." As noted, in all eight states there is no crime unless the unborn child was "quick" when it was killed. In all but one of those states the statute does not define "quick"; as explained hereinabove, "quickening" is traditionally said to occur between the 16th and 20th weeks of pregnancy. And in the one state that does define the term, the legislature has given "quick" the same meaning as "viable": "For the purposes of this section 'quick child' shall mean an unborn child . . . who is so far developed and matured as to be capable of surviving the trauma of birth with the aid of usual medical care and facilities available in this state." (R.I. Gen. Laws Ann. § 11-23-5, 3d par.)

The second category of jurisdictions is composed of those in which the legislature has expressly declared that the killing of a product of conception is criminal regardless of its gestational age; contrary to our section 187, therefore, these statutes purport to apply to both viable and nonviable fetuses and to embryos — even to zygotes. There are at least six states in this category. In three the crime is not murder but a lesser offense. Thus Arizona's statute is modeled on those of the eight states discussed above that criminalize the killing of an unborn child by means of an injury to its mother; unlike those statutes the Arizona measure applies not just to a "quick" unborn child but to an unborn child "at any stage of its development"; like those statutes, however, the Arizona offense is deemed manslaughter. (Ariz. Rev. Stat. Ann. § 13-1103.) Arkansas and New Mexico have essentially identical statutes making it an offense to inflict an injury on a pregnant woman, in the commission of a felony, that causes miscarriage or stillbirth; each defines miscarriage as "the interruption of the normal development of the fetus" and stillbirth as fetal death in utero "irrespective of the duration of pregnancy." (Ark. Code Ann. § 5-13-201; N.M. Stat. Ann. § 30-3-7.) In Arkansas the offense is deemed first degree battery and is punishable by imprisonment for not less than five nor more than twenty years (Ark. Code Ann. § 5-4-401); in New Mexico the offense is called "injury to [a] pregnant woman" and is punishable by imprisonment for three years with a possible fine of $5,000 (N.M. Stat. Ann. § 31-18-5).

Three other states have enacted special statutes making the killing of a fetus homicide under various circumstances, but none punishes fetal murder as severely as our section 187. Thus in Illinois the offense is called "intentional homicide of an unborn child," and the statute defines an unborn child as "any individual of the human species from fertilization until birth." (Ill. Comp. Stat. Ann. ch. 720, § 5/9-1.2, subd. (b).) But the statute further provides that the act is not homicide unless the actor actually "knew that the woman was pregnant." ( Id., subd. (a)(3).) And the punishment for this crime "shall be the same as for first degree murder, except that the death penalty may not be imposed." (Ill. Comp. Stat. Ann. ch. 720, § 5/9-1.2, subd. (d), italics added.)

In Minnesota and North Dakota the offense is called "murder of an unborn child," and the latter is defined as the "conceived, but not yet born" offspring of a human being. (Minn. Stat. Ann. §§ 609.266, 609.2661; N.D. Cent. Code §§ 12.1-17.1-01, 12.1-17.1-02.) In Minnesota the offense is murder of an unborn child in the first degree if the actor kills the unborn child either with premeditation and "with intent to effect the death of the unborn child or another," or while committing certain serious listed felonies; in either event, however, the maximum punishment is life imprisonment. (Minn. Stat. Ann. § 609.2661.) In North Dakota the offense is the most severely punished if the unborn child is killed with express or implied malice or in the commission of certain serious listed felonies; yet in either event the maximum penalty is life imprisonment (N.D. Cent. Code § 12.1-32-01).

The third and last category of jurisdictions is composed of those in which the statute neither prescribes a minimum gestational age for a conviction of fetal murder nor expressly declares that it applies regardless of gestational age; rather, the statute is facially silent on the matter. California is such a jurisdiction, and there are at least five others. Of these Utah is the only state that, like California, criminalizes the killing of a fetus under its general murder statutes: it defines homicide as the killing of "another human being, including an unborn child." (Utah Code Ann. § 76-5-201, subd. (1)(a).) But unlike California, in Utah the crime is a capital offense only if the actor caused the death of the unborn child "intentionally or knowingly," even in a felony-murder case. ( Id., § 76-5-202, subd. (1)(d).) If, as in the case at bar, the death occurred in the commission of a listed felony but the actor did not kill "intentionally or knowingly," the crime is noncapital murder ( id., § 76-5-203) punishable by imprisonment for not less than five years ( id., § 76-3-203).

In the remaining four states the offense is given special treatment and is punished much less severely than in California. In two of these states the offense is deemed "feticide." Thus in Indiana one who "knowingly or intentionally" terminates a pregnancy commits feticide (Ind. Code Ann. § 35-42-1-6), punishable by imprisonment for four years with a possible fine of not more than $10,000 ( id., § 35-50-2-6). In Louisiana one who kills an unborn child intentionally or in the commission of a listed felony commits first degree feticide punishable by imprisonment for not more than 15 years. (La. Rev. Stat. Ann. § 14:32.6.) In South Dakota one who "intentionally kills a human fetus by causing an injury to its mother" commits a felony (S.D. Codified Laws § 22-17-6) punishable by imprisonment for 10 years with a possible fine of $10,000 ( id., § 22-6-1). And in New Hampshire one who "Purposely or knowingly causes injury to another resulting in miscarriage or stillbirth" commits first degree assault (N.H. Rev. Stat. Ann. § 631:1, subd. I(c)) punishable by imprisonment for not more than 15 years with a possible fine not to exceed $4,000 ( id., § 651:2).

Robert Keeler's act of assaulting his estranged wife for the express purpose of terminating her pregnancy by knowingly and intentionally killing her fully viable fetus would have been a crime in all the jurisdictions discussed above that have abrogated the common law rule. It would certainly be a crime in California today. But I cannot believe that in amending section 187 to make that act a crime the Legislature also intended to make California the only state in the Union in which it is a capital offense to cause the death of a nonviable and invisible fetus that the actor neither knew nor had reason to know existed. Yet this, again, is where the lead opinion's construction of the 1970 amendment inexorably takes us. I dissent from that construction.

VI

I also dissent from the lead opinion's disposition of this appeal. I agree with the lead opinion that the instruction defining viability in terms of mere "possible" survival was prejudicial error. I would nevertheless reverse the judgment of the Court of Appeal because it remands the murder count for a new trial, and would direct the Court of Appeal to order the trial court to enter a judgment of acquittal on that count. (§§ 1260, 1262; see, e.g., People v. Carlson (1974) 37 Cal.App.3d 349, 359 [ 112 Cal.Rptr. 321].) On the facts of this case a retrial of the murder count will be a futile exercise and hence a total waste of court time, prosecutorial resources, and taxpayers' money. In such a retrial, as the lead opinion correctly holds, the prosecution will be required to prove beyond a reasonable doubt that the Flores fetus was viable in the sense that its survival was not merely possible but probable, i.e., that it had better than 50 percent chance of surviving premature birth and the neonatal period (28 days after birth). But as the lead opinion correctly observes, "in this case none of the medical experts who testified at defendant's trial believed that the fetus had a 'probable' chance of survival." (Lead opn., ante, p. 814, italics added.) And the prosecution's failure to prove such viability was not through want of trying. The bulk of the prosecution's case was devoted to presenting the testimony of medical experts on the issue of viability. Among other experts the prosecution called three eminent physicians — a pediatric pathologist, a neonatologist, and a perinatologist — who were all highly experienced in determining fetal viability, and not one could testify that the Flores fetus would "probably" have survived premature birth and the neonatal period. The unanimity of the witnesses, moreover, mirrored the uniform conclusions of the medical literature, which they discussed in detail on the stand.

The prosecution thus took its best shot at proving viability, and on the facts of this case was wholly unable to do so. In a retrial the medical facts will not change: a nonviable fetus will remain nonviable. Given this reality, it is highly unlikely that the prosecution will be able to make a better case for viability, let alone prove that element beyond a reasonable doubt. In these circumstances no purpose will be served by compelling the physician witnesses to repeat their lengthy testimony or by putting the survivors — Mrs. Flores and her family — through the ordeal of once again reliving the events in public. Rather, we should exercise our discretion to end this litigation so that the defendant may serve the multiple and consecutive terms of imprisonment to which he was sentenced for the crimes he actually committed.