(dissenting) — The majority rejects the tests for determining retroactivity in civil cases set forth in the Court of Appeals. Bradbury v. Aetna Cas. & Sur. Co., 19 Wn. App. 66, 573 P.2d 395 (1978). It adopts a single standard: Was there justifiable reliance by defendant on the law as it existed prior to our decision in Cammel v. State Farm Mut. Auto. Ins. Co., 86 Wn.2d 264, 543 P.2d 634 (1975)?
Defendant argues State Farm Mut. Auto. Ins. Co. v. Bafus, 77 Wn.2d 720, 725, 466 P.2d 159 (1970), indicated this court would not approve stacking. In Bafus we said:
No provisions in either contract of insurance, nor in the statute, have been pointed out which would warrant a stacking of liability, one policy upon the other, to allow duplicate or overlapping awards for the same bodily injuries.
Accordingly, we deem it the rule that, in the absence of a statute or contract to the contrary, one who has coverage under the uninsured motorist provisions of two or *513more policies of insurance in separate companies is entitled to but one recovery to the extent of his maximum coverage under the larger of them, and cannot increase his total award by stacking one liability upon the other. Maryland Cas. Co. v. Howe, 106 N.H. 422, 213 A.2d 420 (1965); Burcham v. Farmers Ins. Exch., 255 Iowa 69, 121 N.W.2d 500 (1963). Supporting this conclusion is the rationale of our decision in Miller v. Allstate Ins. Co., 66 Wn.2d 871, 405 P.2d 712 (1965).
Of this language, the Court of Appeals said, "This quotation led to the inference that our Supreme Court would not approve the 'stacking' of uninsured motorist coverage." Bradbury, at 69. Was Aetna justified in relying on this inference? The conclusion of the majority, that Aetna should have known of this court's position on the stacking issue because the issue is a matter of considerable dispute nationwide, does not answer the critical question. The decisions reached in other jurisdictions and the extent to which attorneys debate the merits of stacking are not really relevant considerations in determining whether Aetna justifiably relied on this court's language in Bafus.
I believe that reasonable persons, including attorneys and insurance companies, could have interpreted this court's language as disapproving the stacking of policies. The defendant was entitled to rely on the inference that we would disapprove of stacking regardless of how many companies issued the policies. At least four judges (the trial court and the Court of Appeals) believed stacking was not the policy in this state. The defendant was entitled to rely on a similar belief.
The majority points to the distinctions between Bafus and Cammel. These distinctions are valid and form the basis for our holding in Cammel that stacking of policies from the same company is proper. But it is unreasonable to expect persons who read Bafus to recognize, in advance, that the Cammel situation would produce a different result. We may now say that prior decisional law did not exist. But the belief it did exist was certainly justified.
*514One of the primary functions of a court of last resort is to write its opinions in such a way that they can reasonably be relied upon by attorneys and their clients in determining a future course of action. The defendant read Bafus and acted accordingly. The Cammel holding should not be applied retroactively against the defendant and others who relied on an earlier decision of this court. Otherwise, we might as well abandon any claim for the predictive quality of our decisions or any doctrine of reasonable reliance.
Hicks, J., concurs with Dolliver, J.