dissenting.
In my judgment the court’s opinion, in adopting the so-called “seat belt defense,” confuses principles of liability and damages, represents poor judicial policy, and is unfair to the appellee plaintiffs; therefore, I dissent.
The court adopts an analytical approach based on principles of comparative negligence. The starting point for the court’s position is that the nonuse of a seat belt may be considered “fault.” Next, the court adopts a definition of “fault” found in the Uniform Comparative Fault Act. At 154, 755 P.2d at 1142. The fact that Arizona has not adopted the definition used in the uniform act seems to be deemed of little significance by the court.
Reliance on comparative negligence principles is inconsistent with and serves to confuse that concept as it has been adopted in this state. Comparative negligence in Arizona occurs only when the defense of contributory negligence or assumption of risk has been found and applied by the jury. A.R.S. § 12-2505.1 Without contributory negligence or assumption of risk there is no comparative negligence. The nonuse of a seat belt in this case and in most cases is usually not a cause of the accident, so the failure to wear a seat belt is not contributory negligence. Thus, the use of a comparative negligence analysis serves to confuse the principles applicable to the issue under Arizona law.
The observations of Judge Weis of the 3rd Circuit on the subject are instructive.
Mitigation and apportionment of damages are different concepts than contributory or comparative negligence. Contributory negligence frees the defendant from liability and all responsibility for damages. Comparative negligence appraises the factors which caused the impact, collision or similar event and uses the relative degree of fault to reduce the damages. Mitigation or apportionment of damages and avoidable consequences, on the other hand, are directed toward activity (or nonaction) having a direct bearing on the extent of injury but not on the conduct causing the litigated event. (Emphasis supplied.)
The monetary result reached by apportioning damages may be the same as in application, of comparative negligence but the factors leading there must be distinguished. By way of illustration, it may be helpful to examine the results which could occur in a routine intersection accident case in a state which applies comparative negligence and also apportionment of damages. A jury might well determine that driver A was ten percent at fault for failing to observe driver B going past a stop sign, and also that if driver A had been wearing a seat belt, his damages would have been twenty percent less than those actually sustained. Both factors would then be applied to reduce the amount recoverable. Because the same result can be reached under comparative negligence without concern over whether the plaintiff’s conduct goes to liability or damages, there is a tendency to blur the line between them.
The concepts of liability and damages are distinct and must be kept so.
Vizzini v. Ford Motor Co., 569 F.2d 754, 769-70 (3rd Cir.1977) (Weis, J., concurring and dissenting).
It is not necessary or correct to use a comparative negligence formula to arrive at the ruling by the court. If the “seat belt defense” is to be adopted, Judge Weis suggests that the doctrine of avoidable consequences is more appropriate.
*159Restatement (Second) of Torts § 433 provides for apportionment of damages when plaintiffs conduct was a factor in the harm and comment f explains that the doctrine of avoidable consequences is an application of the rule. Because avoidable consequences applies to actions taken after the impact, it has been said to be inapplicable to the seat belt defense. That distinction has not always been observed, however, and should not be controlling in this case. Simply because the correct pigeon-hole for the legal theory is not immediately apparent does not mean that critical evidence should be disregarded. Its relevance is apparent. The test should be not when the challenged activity or nonactivity took place; rather, the focus should be whether it played a part in producing the event or was totally unrelated to that event and affected only the injury. The inquiry should not be chronological, but causal. (Footnotes omitted.)
Vizzini, 569 F.2d at 770.
Using an expansion of the avoidable consequences theory makes more sense than attempting to wedge a solution under so-called comparative negligence principles. The seat belt defense deals with diminution of damages and not with the existence of a cause of action. Diminution of damages is appropriate when the injured party could have avoided some of the injury by the use of reasonable efforts. Full recovery is denied because part of the harm is due to the injured person’s lack of care, and public policy requires that persons should be discouraged from wasting their resources, both physical or economic. See Restatement (Second) of Torts § 918 comment a (1977).
PUBLIC POLICY
Whether the court’s comparative negligence principles or the theory of avoidable consequences is employed, we are ultimately faced with the vital issue in this case— Public Policy. Despite the court's claim that nonuse of a seat belt is not a question of duty (at 153, 755 P.2d 1141), this decision imposes upon all motorists and passengers of this state the duty to wear seat belts, and it fixes the penalty for nonuse as reduction of the amount of damages to be recovered by an injured motorist or passenger. It is pure sophistry to declare that the decision today does not impose a specific duty on all to make use of seat belts. The court declares this non duty as part of its “obligation to participate in the evolution of tort law.” At 156, 755 P.2d 1144.
Implicit in the court’s decision is the rationale that seat belts should be used, and it is time that all of us used them. Although we held no hearings on the subject, the technical literature and opinions of experts seems to support the utility of using seat belts. The important consideration is who should make the decision to impose the duty (or non duty)? What branch of government under our constitutional system should make such a public policy decision?
Some courts such as our colleagues in our neighboring state of New Mexico faced with a similar problem, held that the creation of a “seat belt defense” is a matter for the legislature, not for the judiciary. Thomas v. Henson, 102 N.M. 326, 695 P.2d 476 (1985). There are sound reasons for such a position, none of which have apparently been considered by this court in its rush to “participate in the evolution of tort law.”
Public policy exists only as long as there is public support. Courts decide matters, institutionally, without regard to public support. Protecting the constitutional rights of unpopular individuals or groups often involves deciding matters against public opinion, but in doing so, courts perform their well-recognized “judicial function.” Courts, however, make poor law makers. They make law on a case-by-case method usually with limited information and without a legislator’s concern for public support.
In the case at issue the new policy is imposed retroactive to November 8, 1985, the date of the appellees’ accident. Such action would, generally, be foreign to the legislative system. Absent an emergency situation, the Arizona Constitution provides *160that a legislative act does not become effective for ninety days after the close of a legislative session. Ariz. Const, art. IV, pt. 1, § 1(3). Such a delay in the effective date of legislation provides not only the opportunity for mounting a referendum, but it also affords an opportunity for the public to be made aware of the new obligations created by the legislative act. Under the court’s decision today not only is no period of time allowed for public awareness of the new obligation, but the obligation and its consequences are made retroactive to a date over two years ago. This type of judicial action represents the worst form of “law making.”
There is more involved in this matter than the development of tort law. We are urged to adopt a rule which may reduce the recovery of the plaintiff in this case, but the rule may also result in lessening the extent of personal injuries in future automobile accidents. The latter is certainly a good objective, but in this area of public safety the courts, with the ease-by-case method, are powerless to provide a comprehensive program of policy, enforcement and education. The legislative branch is the proper one to deal with this complex problem if there is to be an effective and unified solution.
RETROACTIVITY
Whatever else may be said about the court’s decision, it seems grossly unfair to apply this new rule to these plaintiffs. The rule in Nash has been the law of this jurisdiction since 1974. Without any suggestion to the contrary, the rule has been that nonuse of a seat belt was deemed irrelevant to the damage issue in an automobile personal injury case. The court’s decision today not only overrules Nash, but it does so retroactive to November 8, 1985, the date of the accident in this case. Such a dramatic change in law should at least be prospective.
SUPPLEMENTAL OPINION
FELDMAN, Vice Chief Justice.Plaintiffs have moved for reconsideration of our opinion. Rule 22, Ariz.R.Civ.App.P., 17B A.R.S. They argue that the opinion should be reversed or modified. First, they argue that the original opinion should be withdrawn in its entirety, primarily because the court incorrectly based its decision on a definition of “fault” specifically rejected by the Arizona legislature when it adopted comparative negligence in 1984. Second, if the opinion is to stand, they urge that it should only operate prospectively.
DEFINITION OF “FAULT”
Contrary to plaintiffs’ assertions, the opinion in this matter was not “based” on the definition of fault found in § 1(b) of the Uniform Comparative Fault Act (UCFA), 12 U.L.A. 39-40 (Supp.1988). We simply noted that definition as one explanation of the fault concept as it has evolved in recent decades. Moreover, we do not think that the tangled legislative history of the 1984 Uniform Contribution Among Tortfeasors Act (UCATA) (codified at A.R.S. §§ 12-2501 et seq.) reveals any hint that the Arizona legislature rejected the UCFA definition of “fault.” We therefore decline plaintiffs’ invitation to reconsider our decision based on an alleged definitional error.
RETROACTIVE VERSUS PROSPECTIVE APPLICATION
We find more merit, however, in plaintiffs’ argument that the decision should only receive prospective application. Unless otherwise specified, Arizona appellate opinions in civil cases operate both retroactively and prospectively. Chevron Chemical Co. v. Superior Court, 131 Ariz. 431, 435, 641 P.2d 1275, 1279 (1982). The rule in favor of retroactivity may be overcome and a decision given only prospective effect if three conditions are present:
1. The opinion establishes a new legal principle by overruling clear and reliable precedent or by deciding an issue whose resolution was not foreshadowed;
2. Retroactive application would adversely affect the purpose behind the new rule; and
3. Retroactive application would produce substantially inequitable results.
*161Chevron Oil Co. v. Huson, 404 U.S. 97, 92 S.Ct. 349, 30 L.Ed.2d 296 (1971); Brannigan v. Raybuck, 136 Ariz. 513, 520, 667 P.2d 213, 220 (1983). See also Lemon v. Kurtzman, 411 U.S. 192, 93 S.Ct. 1463, 36 L.Ed.2d 151 (1973).
The decision of whether an opinion will only be applied prospectively involves a balancing of these three factors. See Bernstein v. Aetna Life & Casualty, 843 F.2d 359, 362-363 (9th Cir.1988) (review of Arizona rules on retroactivity).
A. New Legal Principle
The opinion in this case applied new, statutory comparative fault doctrines to scientific and technological advances. We believed, and still believe, that the adoption of comparative fault as a linchpin of tort jurisprudence required no more than the extension of existing legal principles to impose personal responsibility for one’s use or nonuse of a safety device. We agree with the Laws’ argument that the original opinion therefore imposed no novel duty but simply allowed a jury hearing a comparative negligence case to reduce damages to reflect a plaintiff’s failure to use reasonable care for his or her own safety. Thus, as the Laws contend, there was “[no] mighty change in [the] substantive law.”
While this is true, however, the consequences of extending comparative negligence to the seat belt issue were certainly far from clear, with many comparative negligence jurisdictions adhering to the view that nonuse of a seat belt was not a factor to be considered in assessing damages. See, e.g., Clarkson v. Wright, 108 Ill.2d 129, 90 Ill.Dec. 950, 483 N.E.2d 268 (1985). Further, Nash v. Kamrath, 21 Ariz.App. 530, 521 P.2d 161 (1974), was unequivocal Arizona authority for the proposition that no legal consequences flowed from nonuse of a seat belt. Except for those with the gift of judicial prophecy, therefore, the decision of our court of appeals in this case was not clearly foreshadowed by Arizona caselaw. This factor therefore weighs against retrospective operation.
B. Adverse Effect on Purpose Behind the Rule
The present decision embodied several mutually consistent goals. On one level, the case settled a conflict of authority between the divisions of our court of appeals and determined difficult issues of statewide importance. At a deeper level, the opinipn constituted an advance in the continuing evolution of tort law, furthering concepts of personal responsibility and accountability by recognizing the obligation of each person to act reasonably in light of modern-day advances in knowledge and technology. 157 Ariz. at 156, 755 P.2d at 1144. The principles espoused in this case, as in all tort opinions, were designed and intended to encourage safe and responsible conduct. From that perspective, retroactive application might not foster the purpose of the rule we adopted. We cannot encourage greater seat belt use in the past.
On the other hand, most people were well aware of the practical need to buckle up and the dangers of not doing so. Retrospective operation emphasizes the need to modify future behavior. All of this suggests at most that the balance on this factor is even.
C. Substantially Inequitable Results
The final factor to be considered is basically a question of equity.
“[W]here a decision of this Court could produce substantial inequitable results if applied retroactively, there is ample basis in our cases for avoiding the ‘injustice or hardship’ by a holding of nonretroactivity.”
Coons v. American Honda Motor Co., 96 N.J. 419, 427, 476 A.2d 763, 768 (1984), cert. denied, 469 U.S. 1123, 105 S.Ct. 808, 83 L.Ed.2d 800 (1985), quoting Huson, 404 U.S. at 107, 92 S.Ct. at 355. There is little question that the balance of equities on this factor tips strongly against retroactive application. If applied retroactively, this opinion would result in lower damage recoveries for some accident victims because they failed to use seat belts at a time when no statute required their use and caselaw made it clear that not buckling up was a *162reasonable course of action involving no adverse legal consequences. It is only from the date of decision in this case that legal consequences attached to the conduct in question. Of course, we do not indulge in the fantasy that motorists consulted Nash before deciding whether to use seat belts. Nevertheless, opinions such as Nash and this play a substantial part in setting standards for reasonable conduct.
Weighing the foregoing factors, we conclude that the balance falls heavily on the side of limiting the decision’s effect to prospective application. We realize full well such a holding is contrary to the thrust of the initial decision. See 157 Ariz. at 157, 755 P.2d at 1145. To those unhappy with the change, there are two possible answers. The first is to say, as did Baron Bramwell over one hundred years ago, that “[t]he matter does not appear to me now as it appears to have appeared to me then.” Andrews v. Styrap, 26 L.T.R. 704, 706 (Ex.1872), quoted in Fidell, Judicial Recantation: Homage to Baron Bramwell, 71 JUDICATURE 259, 259 (1988).1 More accurately, the issue of retroactive versus prospective application was not argued to the court and not thoroughly considered in the opinion. The arguments of counsel on the motion for reconsideration have given us the opportunity to thoroughly consider the problem, and we are now persuaded that our former views require modification. We conclude, therefore, that the dissent was correct on this point. There was a “dramatic change” in the law and it would be “grossly unfair” to apply it retroactively. 157 Ariz. at 160, 755 P.2d at 1148 (Holohan, J., dissenting).
We come, then, to the difficult question of whether the benefits of the seat belt defense holding should apply to the Laws, who are defendants in this action. Ordinarily, even when a decision has solely prospective effect, “the fruits of the struggle are awarded the successful litigant.” Coons, 96 N.J. at 434, 476 A.2d at 772. However, each of the factors militating against general retroactive application militates just as strongly against retroactive application against the Harders. An equally pertinent consideration in this as in almost all seat belt cases is the fact that the “fault” of the victims had no relationship to the occurrence of the accident. Because the prophylactic effect of the new seat belt rule is precluded in this case, there is no equitable reason to reduce the Harders’ damages without prior notice of the need to use a seat belt.
On balance, we believe that the best and most equitable rule is simply to make our original decision effective and applicable to all accidente occurring after the date on which the mandate in this case is filed.2 See Hawkins v. Allstate Insurance Co., 152 Ariz. 490, 505, 733 P.2d 1073, 1088 (choosing the date of mandate as appropriate point to impose new burden of proof for punitive damages), cert. denied, — U.S. -, 108 S.Ct. 212, 98 L.Ed.2d 177 (1987).
Consistent with the views expressed in this supplemental opinion, we reinstate the trial court’s protective order. The Hard-ers’ nonuse of seat belts is inadmissible at the trial of their action to prove the cause or aggravation of their injuries.
GORDON, C.J., and CAMERON, J., concur.. § 12-2505. Comparative negligence; definition
A. The defense of contributory negligence or of assumption of risk is in all cases a question of fact and shall at all times be left to the jury. If the jury applies either defense, the claimant’s action is not barred, but the full damages shall be reduced in proportion to the relative degree of the claimant’s fault which is a proximate cause of the injury or death, if any. There is no right to comparative negligence in favor of any claimant who has intentionally, wilfully or wantonly caused or contributed to the injury or wrongful death.
B. In this section, "claimant’s fault” includes the fault imputed- or attributed to a claimant by operation of law, if any.
. We might also emulate Justice Story, who explained a reversal of position by stating that "[m]y own error, however, can furnish no ground for its being adopted by this Court." United States v. Gooding, 25 U.S. (12 Wheat.) 460, 478, 6 L.Ed. 693 (1827), quoted in Fidell, supra, at 259.
. The mandate in this case was filed on July 7, 1988.