(dissenting)—I agree with the majority that the jury instruction at issue here is unconstitutional. I also agree, although reluctantly, that the holding in Mullaney v. Wilbur, 421 U.S. 684, 44 L. Ed. 2d 508, 95 S. Ct. 1881 (1975), is retroactively applicable to this case. I do not agree, however, that the personal restraint petition in this case should be granted, because the record clearly shows that the error in giving the jury instruction here was harmless. For that reason, I respectfully dissent from the majority's opinion.
Mullaney held that Maine's presumption of implied malice, which required a defendant to prove he acted in the heat of passion on sudden provocation, violated the due process clause of the federal constitution because it removed from the prosecution the burden of proving each element of the crime beyond a reasonable doubt. Mullaney v. Wilbur, supra. See In re Winship, 397 U.S. 358, 25 L. Ed. 2d 368, 90 S. Ct. 1068 (1970). Mullaney, however, does not render unconstitutional all presumptions in the criminal law.
It permits the use of an otherwise valid presumption, so long as the ultimate burden of persuasion as to the issue remains with the prosecution and the defendant is at most only required to produce some evidence, from whatever source, in opposition to it.
State v. Roberts, 88 Wn.2d 337, 341, 562 P.2d 1259 (1977). The jury instruction in this case is unconstitutional under Mullaney because it was not expressly limited to placing an initial burden of producing evidence upon the defendant. *293See State v. Roberts, supra at 342.5
In Hankerson v. North Carolina, 432 U.S. 233, 53 L. Ed. 2d 306, 97 S. Ct. 2339 (1977), the Supreme Court held the Mullaney rule was retroactive.6 In view of the majority of the United States Supreme Court, a new constitutional doctrine held to be retroactive is equally retroactive in collateral proceedings and direct appeals. Williams v. United States, 401 U.S. 646, 28 L. Ed. 2d 388, 91 S. Ct. 1148 (1971). Therefore, the Mullaney rule is retroactively applicable to personal restraint petition cases such as that with which we are here concerned.
Although I concede the force of Supreme Court precedent, I deplore this result. I would generally limit retroactive application of new constitutional doctrines to direct appeals only. In the words of Justice Powell, such a limitation
*294is closer to the ideal of principled, evenhanded judicial review than is the traditional retroactivity doctrine. At the same time it is more attuned to the historical limitations on habeas corpus . . . [the federal equivalent of our personal restraint petition], and to the importance of finality in a rational system of justice.
(Citation omitted.) Hankerson v. North Carolina, supra at 248 (Powell, J., concurring). I agree with Justice Harlan that it is "a matter of fundamental import that there be a visible end to the litigable aspect of the criminal process." Mackey v. United States, 401 U.S. 667, 690, 28 L. Ed. 2d 404, 91 S. Ct. 1160 (1971) (Harlan, J., dissenting).
The position I have taken here was strongly endorsed by our Supreme Court in In re Myers, 91 Wn.2d 120, 587 P.2d 532 (1978). The Myers court stated at pages 123-24:
Assuming the challenged instructions in the present case are defective under Roberts and Mullaney, the question is whether this new rule should be applied retroactively to this collateral review proceeding. The importance of finality in a rational system of justice leads us to conclude that it should not.
The Myers court then related important policy reasons for this conclusion:
When a trial has been held, a conviction affirmed on appeal, and a period of time thereafter elapses, we believe the negative effect on the administration of justice outweighs the interest of a defendant in having his guilt redetermined in accordance with subsequent decisions of this court or the United States Supreme Court.
Evidence necessary for retrial may no longer be extant. Witnesses may have moved or died and memories certainly will have dimmed. Further investigation, difficult at best, may well be impossible. Such factors could, and in this case probably would, make effective retrial of Myers impossible. If another trial is to be conducted in this case with the consequent burdens on the judicial and prosecutorial resources it undoubtedly would be less reliable than the first.
In re Myers, supra at 124-25. Because of the Hankerson opinion, however, the Myers court was able to deny the *295personal restraint petition only on the ground that Myers had failed to object at trial to the jury instruction at issue.
Comparison of Myers with the instant case points out the inadequacy of the United States Supreme Court's standard for retroactive application of a new constitutional doctrine in collateral review proceedings. Myers was tried in 1959. At his trial, a burden-shifting instruction was given, the validity of which was so well settled that it was not determined to be invalid until the Mullaney case in 1975. Myers' personal restraint petition was denied in 1978 solely because his attorney had not had the prescience to anticipate by 16 years a Supreme Court opinion that would declare that instruction unconstitutional. Gilbert, on the other hand, had the good fortune to be represented by an attorney who anticipated the Mullaney decision by 6 years and did except to that instruction. As a result, Gilbert's petition will be granted. This is a very slight distinction upon which to base the granting or denial of a personal restraint petition, especially because the policy reasons against granting the petitions apply with equal force to both cases.
The majority, however, need not have reached this unfortunate result. In State v. Rogers, 83 Wn.2d 553, 520 P.2d 159, cert. denied, 419 U.S. 1053, 42 L. Ed. 2d 650, 95 S. Ct. 633 (1974), where the defendant specifically objected to the same instruction as the offending instruction in the instant case, the court said at page 557:
An accused cannot avail himself of error as a ground for reversal where the error has not been prejudicial to him. State v. White, 72 Wn.2d 524, 433 P.2d 682 (1967). Error cannot be regarded as harmful so as to require reversal unless, within reasonable probabilities, had the error not occurred, the result might have been materially more favorable to the one complaining of it. . . . We do not believe the average juror would have found the prosecutor's case significantly less persuasive had the jury not been instructed on the statutory presumption of the defendant's specific intent to kill. The error was harmless.
*296(Citation omitted.) I have viewed again the trial record in the instant case. The record shows that, as in Rogers, the error in giving the jury instruction was harmless.
In the instant case, the evidence of guilt was so overwhelming the defendant assigned no error to the sufficiency of the evidence to support the conviction in his direct appeal. There was no claim of self-defense, or any defense, to defendant's criminal act. The facts as a jury could have found them are these: The petitioner Gilbert, and others, including one John LaBeur, contacted one Edward Earl Butler for the purpose of collecting a $20 debt LaBeur claimed Butler owed him. Also, Gilbert intended to buy some marijuana. After arriving at Butler's residence, they went to the door. One of the occupants answered and attempted to get Gilbert and LaBeur to leave. They obtained entrance but were told that no marijuana was available for about 3 hours. Gilbert and LaBeur returned to their car, and, as it was driving away, Gilbert told the driver to stop because he wanted to go back and talk to Butler. LaBeur followed Gilbert to the house and a further conversation occurred between Gilbert and Butler. Gilbert stared at Butler for a few seconds, placed his hand inside his belt, drew out a gun, and shot Butler in the head, from which wound Butler soon thereafter expired. After returning to the car, Gilbert made the statement that he had shot Butler and that he thought he had killed him.
I can say with confidence that the giving of the instruction complained of did not make the State's case significantly less persuasive than it would have been had the jury not been so instructed. To require the State to retry Gilbert within 60 days for a murder that occurred more than 10 *297years ago, or set him free and dismiss the charge with prejudice, places too great a burden on the State. I therefore dissent from the conclusions reached by the majority.
The majority obscures the distinction between the holdings in Mullaney and State v. Rogers, 83 Wn.2d 553, 520 P.2d 159, cert. denied, 419 U.S. 1053, 42 L. Ed. 2d 650, 95 S. Ct. 633 (1974). Rogers predated Mullaney. Rogers held that the application of the presumption contained in RCW 9.41.030 was unconstitutional under the facts of that case. The court made a similar holding in a companion case to Rogers, State v. Odom, 83 Wn.2d 541, 520 P.2d 152 (1974). The Rogers and Odom decisions were based upon the court's findings that the State did not show beyond a reasonable doubt that the presumed fact, specific intent to commit murder, followed the proven fact, being armed with an unlicensed weapon. The Rogers-Odom cases and Mullaney therefore present two distinct due process tests under which an evidentiary presumption is to be scrutinized. State v. Roberts, supra at 340-41.
The usual test for retroactive application of a new constitutional doctrine is whether the doctrine's "major purpose ... is to overcome an aspect of the criminal trial that substantially impairs its truth-finding function and so raises serious questions about the accuracy of guilty verdicts in past trials ..." Hankerson v. North Carolina, supra at 241; Williams v. United States, 401 U.S. 646, 653, 28 L. Ed. 2d 388, 91 S. Ct. 1148 (1971).
The majority confuses this test with one enunciated in In re Farney, 91 Wn.2d 72, 583 P.2d 1210 (1978), and based on Robinson v. Neil, 409 U.S. 505, 35 L. Ed. 2d 29, 93 S. Ct. 876 (1973). Robinson made clear that its holding was limited to new constitutional doctrines that are unrelated to the truth-finding process of a trial and that are, therefore, not susceptible to analysis under the test employed in Williams and Hankerson. Consequently, the Robinson and Farney holdings are not applicable to this case.