An action against a physician and a hospital whose negligence permanently disabled plaintiffs mother brings before us the issue whether a minor child may recover damages for the loss that a mother’s incapacitation means for the child. According to the complaint, the defendants’ negligent treatment caused brain damage that will require the mother to have lifelong custodial care, with the result that “plaintiff has been deprived of his mother’s society, companionship, support [and] education” as well as incurring a future obligation to support his mother. A memorandum opposing defendants’ motion to dismiss the complaint adds that when the mother was disabled she was 25 years old and the sole surviving parent of the then three-year old plaintiff.

The circuit court accepted the defendants’ position that plaintiffs claim was one for “parental consortium” unknown to Oregon law and dismissed the complaint. The Court of Appeals affirmed, three judges dissenting, 52 Or App 853, 631 P2d 1377 (1981), and having allowed review, we also affirm.

I. The basis for decision.

Novel issues of nonstatutory law, and especially tort claims, pose recurring questions of the sources and methods of law. Because other courts examining the child’s tort claim for a parent’s disablement have based divergent results on a variety of reasons, well briefed by the present parties, we review the reasons that do not as well as those that do enter into our assessment of the present state of Oregon law on this issue.

Discussion of the child’s claim often begins with a statement that such a claim was unknown at common law. This implies that to allow the claim means a change in existing law and therefore places on the plaintiff the burden to show why the law should be changed by judges rather than by legislators.1 If by “common law” one means the common law of England and those among its one-time dependencies that continued to follow the decisions of *546English appellate courts, the statement denying the child’s claim appears to be correct. See Fleming, The Law of Torts 142, 644 (5th Ed 1977). Until recently, it seemed true also in this country, at least where the issue had been litigated.2 When the Restatement of Torts was revised in 1969, the American Law Institute adopted a new section 707A, denying liability to a minor child for loss of parental support and care, with the comment that the rule was stated “with some reluctance on the part of several of the drafting group, and under compulsion of the case law.”3 This is no longer accurate; recent decisions in three states have recognized a child’s claim in cases like that before us. Moreover, other states are counted as rejecting the claim on the strength of *547decisions of intermediate courts or of federal courts which may prove not to be accurate reflections of the state’s law.

Whatever may have been assumed in the early years of American law, the decentralization of private law in our federal system precludes reference to a single American common law.4 When the Supreme Judicial Court of Massachusetts, in Ferriter v. Daniel O’Connell’s Sons, Inc., 413 NE2d 690 (Mass 1980), first allowed the children of a man disabled by another’s negligence to demand damages for mental anguish and loss of consortium and society, the court did not purport to change the common law of Massachusetts. It undertook to relate the claim to the state’s existing law without assuming that, in the absence of a prior decision, it would be more a change to allow the claim than to deny it. Like that court, we do not assume that without some radical innovation, present law precludes liability to the minor children of a person disabled by a defendant’s negligence, an assumption that calls into question the roles of court and legislature in such innovations. That question is proper, even unavoidable, when a court is asked to depart from its own well-settled prior doctrine; but when a claim is new to this court, either a decision for or one against liability may be harder to square with established law. The Court of Appeals rightly considered the *548child’s claim to be open for judicial decision, although the majority and the dissent divided over the significance of existing legislation in deciding it.

That a novel issue is open to judicial resolution says little about how to resolve it. Analysis often depends on the starting point from which one enters upon it. Two characteristics of the claim at issue here are that the injury to the plaintiff occurs as a consequence of an injury to another person, and that this consequential injury is to plaintiffs psychic interests rather than to his physical person or tangible property. If one starts from a broad premise that every person predictably injured by another’s negligence is entitled to recover appropriate damages, defendants are hard put to present a compelling obstacle to allowing such damages to a child negligently deprived of a functioning parent. Starting, on the other hand, from a body of negligence law that generally imposes liability only to the initial victim but not to others who depend on that victim, and no liability for psychic harm divorced from tangible injury, a plaintiff is hard put to show how and why such liability extends to the minor children of a disabled parent but to no other classes of plaintiffs whose comparable interests are similarly harmed.

A number of courts have undertaken to allow or deny the child’s action for loss of parental society, care, and support by assessing numerous arguments of policy and practicality adduced for and against such a cause of action. An examination of these arguments will explain why we do not follow that course.

In one of the early decisions to consider the issue, the Supreme Court of Kansas sympathetically stated the case for the child’s claim:

“It is common knowledge that a parent who suffers serious physical or mental injury is unable to give his minor children the parental care, training, love and companionship in the same degree as he might have but for the injury. Hence, it is difficult for the court, on the basis of natural justice, to reach the conclusion that this type of action will not lie. Human tendencies and sympathies suggest otherwise. Normal home life for a child consists of complex incidences in which the sums constitute a nurturing environment. When the vitally important parent-child *549relationship is impaired and the child loses the love, guidance and close companionship of a parent, the child is deprived of something that is indeed valuable and precious. No one could seriously contend otherwise.”

Hoffman v. Dautel, 189 Kan 165, 168, 368 P2d 57, 59 (1962). The court nevertheless rejected the claim because of “far-reaching results,” which it identified as the growth of a “new field of litigation” with the possibilities of multiple actions and double recovery. Id. The New Jersey Supreme Court later denied the child’s cause of action upon similar “considerations of policy.” Russell v. Salem Transportation Co., 61 NJ 502, 295 A2d 862 (1972). The court assumed that jury verdicts for an injured parent took into account the consequential injury to children and considered this preferable to the separate appraisal and cumulation of multiple awards arising out of a single incident and the segregation of damages awarded the child from the parents’ disposition of the overall family finances.

Academic commentary has been critical, finding none of the concerns stated by the Kansas court to be “persuasive.” Clark, Law of Domestic Relations 279 (1968). Dean Prosser characteristically was more outspoken:

“It is not easy to understand and appreciate this reluctance to compensate the child who has been deprived of the care, companionship and education of his mother, or for that matter his father, through the defendant’s negligence. This is surely a genuine injury, and a serious one, which has received a great deal more sympathy from the legal writers than from the judges. There is of course the same problem of preventing double compensation as in the case of the wife’s action, since the child will to some extent benefit by any sum recovered by the injured parent; but it is quite evident that this will not and cannot recompense him for all that he has lost. The obstacles in the way of satisfactory limitation of recovery are no greater than in the case of the wife.” (Footnotes omitted).

Prosser, The Law of Torts 896-97 (4th ed 1972). Two articles published in 1976 reviewed the arguments against a child’s cause of action for the intangible harm of a parent’s disablement in more detail and found them without merit. Love, supra note 2; Note, supra n. 2.

Nevertheless, the California Supreme Court in 1977 declined to recognize what it called “a new cause of *550action for loss of consortium in a parent-child relationship.” Borer v. American Airlines, 19 Cal3d 452, 563 P2d 858, 860, 138 Cal Rptr 302 (1977) (child’s action); Baxter v. Superior Court of Los Angeles County, 19 Cal 3d 461, 563 P2d 871, 138 Cal Rptr 315 (1977) (parent’s action). The court denied liability for psychic injury to children of a negligently disabled person, even when this injury is foreseeable, on grounds of “social policy.” As reasons it listed the inadequacy of money to restore or compensate for the intangible harm of losing a disabled mother’s companionship and guidance, substituting only a future benefit unrelated to that loss; the consequent difficulties of measuring damages for the child’s loss and of instructing juries so as to avoid double recovery; the multiplication of actions and of damages when there are several children; and the resulting public burden of increased insurance rates. 563 P2d at 862-63. Justice Mosk, dissenting, pointed out that three years earlier the California court had rejected each of these policy arguments in establishing a wife’s action for loss of consortium,5 except for the potential recovery by more than one child. 563 P2d at 867-68.

Ferriter v. Daniel O’Connell’s Sons, supra, reached its contrary result in Massachusetts without rehearsing the policy arguments previously debated by the California court and its predecessors.6 The following year, Michigan’s supreme court became the second to allow the child’s cause of action. Berger v. Weber, supra n. 1. The case illustrates the uncertain role of such arguments in common law decisions. The court’s opinion began by finding analogies for the child’s claim in “existing judicial and legislative policies” such as the action for loss of consortium between spouses, the right of parents to recover for the loss of companionship and society of negligently injured children, and the child’s similar recovery under the wrongful death statute and under a “dramshop act.” It then reviewed and rejected the pragmatic arguments concerning multiple *551litigation and insurance costs only as unpersuasive objections to an otherwise valid claim. 303 NW2d at 425-27. Justice Levin, writing for three dissenters, complained that the majority had not subjected the reasons given for the cause of action to the same critical scrutiny as the objections to it and argued for an “independent reexamination” and “subtle balance” of the competing interests and policy considerations involved. 303 NW2d at 430.

Most recently, the Iowa Supreme Court blended analytical and policy arguments in the third decision allowing the child’s claim, Weitl v. Moes, 311 NW2d 259 (Iowa 1981). Like Michigan, the Iowa court began with the implications of an existing parental action for loss of a child’s “consortium.”7 It then entered upon an extended assessment of the arguments against a parallel action by the child for loss of “parental consortium,” including a contention that the action would be “anomalous” in a state that denies a child’s action for the intentional alienation of a parent’s affection, as well as the conventional arguments of social policy and litigation processes previously discussed. 311 NW2d at 265-68. After examining the “drawbacks,” the court turned to “considerations weighing in favor of recognizing a child’s independent action for loss of consortium.” 311 NW2d at 268. It found these partly in general principles, such as the trend to recognize the independent legal personality and rights of minors, and partly in the factual assumption that “in any disruption of the parent-child relationship, it is probably the child who suffers most.” Finally, the Iowa court concluded that the child would have to sue separately if any damages were to be recovered for the child’s loss, since the court reversed its interpretation of a statute that previously had been held to include this loss in the recovery of the injured parent, see supra note 7. 311 NW2d at 269.

The arguments reviewed in the foregoing opinions and commentary juxtapose different kinds of reasons. Contentions about the character of plaintiffs loss and its *552compensability by money address facts or assumptions about human beings in society, in our case a society that postulates the dependence of children on their parents and that leaves other services and satisfactions to an economic market. We accept the view that a parent’s disablement is likely to mean a painful and possibly permanent psychic injury to a child, although one to be proved in the individual case, and that in principle it is no more or less compensable in money than other psychic injuries for which damages are allowed. But whatever the opposing psychological and social contentions mean for a legal obligation to compensate a child for incapacitating one of its parents, they cannot properly be “weighed” against concern about insurance rates, at least not by a court of law. A person’s liability in our law still remains the same whether or not he has liability insurance; properly, the provision and cost of such insurance varies with potential liability under the law, not the law with the cost of insurance.

Nor are such contentions about the child’s loss commensurable with concerns about the processes of litigation. No doubt there are genuine wrongs that courts are ill suited to set right, and others that do not merit the social costs of litigation. But if these costs are to be the reason for denying an otherwise meritorious cause of action, that is one judgment to be made by legislatures rather than by courts. Courts exist to serve whatever rights people have, Or Const art I, § 10;8 it is not for them to weigh or “balance” their own institutional concerns against the merits of such a right. Assuming that a child has a cognizable claim for damages at all, we would agree with the view of the Iowa court that it is more rational to identify and evaluate that claim individually, preferably in a consolidated proceeding, rather than with the New Jersey court’s view that the child’s loss will implicitly be included in a jury verdict for the injured parent’s damages.9 If existing *553procedures make it difficult to consolidate different claims for trial or to avoid overlapping recoveries for the same loss, the obvious answer is not to deny that there is a claim but to reform the procedures. Shortfalls in procedural reform do not justify shortchanging otherwise valid claims.

There is another reason not to explain the court’s understanding of the existing state of the law by the court’s views of desirable social policy. Legislators, unlike judges, may change the law at any time without any demonstration of error, illogic, or incongruity, simply upon changes in personnel and in the political agenda. That is what elections and legislative debates are for. One day’s unsuccessful proponents or opponents of a social policy may renew the campaign the next day to change the law. They should be free to debate the merits untrammeled by a court’s arguments why its view of the existing law represents the better policy. This court expresses no such view of the present issue.

We therefore lay aside the pragmatic arguments adduced for and against a child’s damage action for the disablement of a parent and turn to the question how plaintiffs claim relates to other comparable claims.

II. Analogies examined by other courts.

As we have said, two characteristics of the harm for which plaintiff would hold a negligent defendant liable are that it occurs as a consequence of an injury to another person, and that it is psychic or emotional harm divorced from any injury to plaintiffs physical person or tangible property. These characteristics generally, though not invariably, preclude recovery based on negligence. Plaintiff, however, narrows the asserted liability specifically to injury to a minor child’s interest in its closest family relationship, that with its parents.

Legal arguments for and against the existence of such a tort claim therefore have sought analogues in the treatment of other claims for damages for emotional distress in such family relationships. Typically these include *554(1) actions for loss of spousal consortium, (2) parents’ recovery of damages for negligent injuries to minor children, (3) recognition of the child’s noneconomic loss in the wrongful death of a parent, (4) actions for alienation of affections, and (5) bystanders’ actions for emotional distress from witnessing the death or injury of a close relative. Courts confronted with a child’s claim based on negligent injury to a parent have examined some or all of these analogues with divergent results. Not only do the foregoing actions differ among the states, they may reflect apparent inconsistencies within one state, and each is distinguishable in some respect from the child’s claim.

The Massachusetts court in Ferriter, the first decision to allow the child’s cause of action, drew support from the parent’s recovery of damages for emotional injury in actions for abduction or seduction of a child, in which the father’s loss of the child’s services was recognized to be a fiction. Although such an action was also available if one’s child was beaten or wounded, the court conceded that no Massachusetts case had compensated a parent’s mental suffering from physical injury to the child, so this offered little support for the court’s assertion that such a claim by a child had “analogous precedent” in such compensation for the parents’ “sentimental as well as economic injuries.” 413 NE 2d at 692-93. Moreover, the Massachusetts court faced its earlier rejection of a child’s claim for damages against a defendant who had enticed her mother to desert her and her father, which it described in Ferriter as the “disfavored action for alienation of affections” and distinguished as posing greater threats of “extortionate litigation” and of pitting family members against each other. 413 NE 2d at 694. These may be distinctions of policy, but they do not differentiate the character of the harm inflicted upon the minor child by the negligent as distinct from the intentional tort. Finally, the Ferriter court fell back on the inclusion of damages for loss of the decedent’s society in the statutory action for wrongful death, concluding: “We think it entirely appropriate to protect the child’s resonable expectation of parental society when the parent suffers negligent injury rather than death.” 413 NE 2d at 695.

The Iowa Supreme Court’s decision in favor of the claim, as already mentioned, was substantially influenced *555by its reading of an Iowa statute which provided for the recovery of “the value of services and support as spouse or parent, or both” by the directly injured person or her estate and further provided that “recovery for these elements of damage may not be had by the spouse and children, as such.” The court previously had interpreted this statute as permitting recovery for all elements of the child’s claim for injury to a mother, though procedurally incorporated in the action by the mother or her administrator. Hankins, supra n. 7. When it reexamined its interpretation in Weitl v. Moes, supra, the court cited Hankins as recognizing a statutory claim for a child’s loss of parental consortium. It therefore treated its new decision as merely moving the child’s damages from the claim on behalf of the parent into a separate action: “[T]he more drastic change in our law would result not from recognition of the claim, but from its denial.” 311 NW2d at 269.10

In the Michigan case, Berger v. Weber, supra n. 1, the majority briefly listed the action for loss of spousal consortium, the parents’ action for a fictitious loss of an injured child’s services, and inclusion of loss of companionship in damages under the wrongful death act as well as under a “dramshop act” as sufficiently demonstrating the state’s policy to recognize a child’s cause of action based on a negligent injury to a parent.11 It discounted the asserted contrary policy of a statute barring suits for alienation of affection, because this bar coexists with actions for negligent injury to one’s spouse or child. The court concluded that “the real anomaly is to allow a child’s recovery for the loss of a parent’s society and companionship when the loss-attends the parent’s death but to deny such recovery when the loss attends the parent’s injury.” 303 NW 2d at 426.

*556The Michigan dissenters dismissed the analogy to a parent’s action for injury to a child because in such an action damages are limited to “loss of services and expenses incurred.” 303 NW2d at 434, n. 38. Their main attack was on the analogy to the action for loss of consortium, which they described as a “historical curiosity” that first had been unthinkingly expanded from old intentional torts into negligence and thereafter equalized between the spouses by extending it to the wife rather than by abolishing the husband’s action.12 Apparently the dissenters were not prepared to make a similar attack on the inclusion of damages for lost society and companionship under the wrongful death act; they only warned that this analogy for a claim based on nonfatal injuries would be available to all who are eligible claimants under the wrongful death act, not only to the injured person’s children.

To turn to the major recent opinion rejecting the child’s cause of action, Borer v. American Airlines, supra, the California Supreme Court faced an additional argument for recovery by analogy to that court’s decision in Dillon v. Legg, 68 Cal 2d 728, 441 P2d 912, 69 Cal Rptr 72 (1968), which allowed a mother’s cause of action for emotional trauma in witnessing the negligently caused death of her child. The holding that such a trauma is compensable by damages is hard to square with the statement in Borer that money cannot compensate for a child’s trauma of growing up with a permanently disabled and helpless mother, which may well be longer lasting and harder to forget than the shock of witnessing an injury or even a death. The Borer court, however, explained Dillon as being limited to cases in which the bystander’s psychic trauma has somatic consequences. 563 P2d at 864. This sufficed to deny the analogy to the child’s claim in Borer, although only until another child claims that such consequences accompanied its psychic injury from a parent’s disabled condition; however, the California court later abandoned the requirement of physical harm in an action for negligent infliction of *557emotional distress. Molien v. Kaiser Foundation Hospitals, 27 Cal 3d 916, 616 P2d 813, 167 Cal Rptr 831 (1980).13

In Borer, the California court also denied any anomaly in allowing recovery for emotional loss in wrongful death actions but not when the injured parent is permanently disabled. It saw the primary purpose of the wrongful death statute not in compensating the survivors for their loss but in preserving the deterrent function of tort law when a fatally injured person’s own claim did not survive, thus “providing to tortfeasors a. substantial incentive to finish off their victims.”14 563 P2d at 865. In rejecting the analogy to the spousal action for loss of consortium, the Borer court added to its previously described reasons of “social policy” the distinction that the spousal action rests in part on impairment of plaintiffs sexual life. 563 P2d at 863.

In sum, the opinions of other courts have ascribed varying significance to the several evident analogies in deciding for or against the child’s damage claim for the negligent incapacitation of a parent. The child’s negligence action has been labeled as one for “loss of consortium,” and then allowed as involving a similar family relationship as the spousal action or denied for lacking the sexual element, or because that action is an anachronism which should not be extended. An analogy has been drawn to the parent’s *558action for loss of the child’s “services,” but the analogy is shaky when that action has not previously been allowed for solely emotional injury. Recovery for psychic trauma from witnessing serious injury to a family member is potentially a close parallel, but that action and its prerequisites are unsettled.15 Denial of an action for intentional alienation of affections has been cited as incompatible with allowing an action for negligent invasion of the same interest, or it has been distinguished as involving a dispute over the voluntary conduct of a family member.

Probably the recovery of psychic damages for a parent’s wrongful death offers the strongest parallel for a similar recovery for losing a parent by negligent disablement. That action is statutory, and the opinions have not agreed on what underlying policies a state’s wrongful death statute enacts.

In the light of these divergent analyses, we examine how the opposing contentions fit into the existing law of this state.

III. Obstacles in Oregon negligence law.

Like other common law jurisdictions, Oregon has few precedents for liability for negligent injury to solely psychic interests or for harm resulting from an injury to another person. Although defendant’s motion to dismiss characterized plaintiffs complaint as pleading a cause of action for loss of “parental consortium,” plaintiff himself describes it simply as an “action in negligence.” Viewed simply as a common law negligence case, without a special statutory basis, plaintiffs action must escape each of these two obstacles.

If there are few causes of action for psychic or emotional harm as such, the reason is not found in objections to monetary damages for harm of that nature. The reason may be found by focusing, not on the nature of the plaintiffs loss, but on the source and scope of the defendant’s liability. This court has recognized common law liability for psychic injury alone when defendant’s conduct *559was either intentional or equivalently reckless of another’s feelings in a responsible relationship,16 or when it infringed some legally protected interest apart from causing the claimed distress, even when only negligently.17 The court has found infringements of legal rights in an invasion of privacy, Hinish, supra n. 4, in the negligent removal of the remains of a deceased spouse, Hovis, supra n. 17, and in the negligent delivery of a passport that allowed plaintiffs child to be taken from this country, McEvoy, supra n. 17. But we have not yet extended liability for ordinary negligence to solely psychic or emotional injury not accompanying any actual or threatened physical harm or any injury to another legally protected interest.18

Under these principles, to use a simple illustration, a child might well have a cause of action for solely emotional distress if someone, in order to cause that distress; injured not the child’s parents but a favorite family pet. Cf. Fredeen v. Stride, supra n. 17. Arguably, also, the child has rights in the parental relationship sufficiently like those asserted in Hovis and McEvoy to support a similar recovery for a psychic injury inflicted even by negligence. The nature of the harm asserted here therefore does not alone defeat plaintiffs claim. There remains, however, the objection that the loss he asserts, though it is an injury to himself, arises solely as a consequence of an injury to another person.

We have recently reaffirmed the denial of damages to one person economically injured in consequence of a *560negligent injury to another person. Ore-Ida Foods v. Indian Head, 290 Or 909, 627 P2d 469 (1981). In his opinion for the Court, Justice Peterson reviewed, without endorsing, the various ways in which this denial of a negligence claim for consequential injuries has been explained. These include such propositions as that the defendant’s negligence, though the cause of plaintiffs loss, was not the “proximate cause,” a concept no longer employed in our cases;19 or that the consequence is “too remote,” which generally means the same thing; or that it is not foreseeable, which is a question of fact and will often be untrue; or that claims by consequentially injured plaintiffs would burden the courts, a reason we have rejected above; or that the negligent defendant’s duty to avoid unreasonable risk of harm to the person initially injured does not extend to such plaintiffs, which merely states the result.20 290 Or at 916-17. As the dissenting opinion in the Court of Appeals pointed out, this court has said that generally the scope of the duty to avoid negligent injury to another is governed by the foreseeable risk of harm, and on appropriate evidence a factfinder might well find it within a physician’s or hospital’s knowledge or reasonable expectation that a 25-year-old female patient has one or more young children who risk immediate and long-range psychic harm if she is incapacitated.21 Nonetheless, all these propositions express a rule that negligence alone, as a reason to shift the burden of a *561resulting loss, has not been deemed so grievous as to hold the negligent actor liable beyond the immediate victim’s injury to others who suffer a loss only in consequence of that injury.

The denial of recovery for a third person’s consequential economic loss in Ore-Ida Foods, supra, meets the present plaintiffs claim for his alleged loss of his mother’s support and his own future obligation to support his disabled mother. The question remains whether the law recognizes a wider range of liability for foreseeable noneconomic consequences to others beyond a physically injured person. We note that plaintiff, and those courts that have allowed claims like his, do not assert such liability for the foreseeable psychic harm to any person who is in a close emotional relationship to the injured person. They speak only of the relationship between parents and minor children. See Ferriter v. Daniel O’Connell’s Sons, supra, 413 NE 2d at 696; Weitl v. Moes, supra, 311 NW 2d at 270. But see Berger v. Weber, supra n. 1, 303 NW2d at 427 (reserving decision on whether action is limited to minor children with “severely” injured parents). This has its own difficulties in a society whose practices and common assumptions about such relationships are rapidly changing. Would the parentage that entitled a child to claim a psychic injury follow from biology, or cohabitation, or only married cohabitation, or solely from legal parentage by birth or adoption?22

In any event, to limit recovery to minor children means that the premise for the asserted cause of action is not found in general negligence theory applied to indirect but foreseeable psychic injury as such, because this can occur in many other close relationships. Rather, as shown by the cases discussed in Part II, the premise must be sought in those parts of the state’s law that specifically deal with injuries to family relationships.

*562IV. Tortious injuries to family relationships.

Several of the analogies from which other courts have drawn arguments for or against a child’s tort claim for psychic harm from the negligent disablement of a parent are themselves open issues in Oregon.

Of the custodial parent’s action for the injury of a child, ORS 30.010, this court has said that the statute preserved the parent’s common law action for such injury independent of an action by the child or the child’s estate;23 but the court has not decided whether the parent has a claim for emotional distress from a child’s nonfatal injury, perhaps because almost all the decisions involved deaths, for which recovery of such damages was excluded until 1973. The parent’s action therefore does not lend support to such a claim on the part of a child. The same applies to recovery for emotional trauma from observing a physical injury to a member of the family, another open issue in this state. Similarly, there is no Oregon precedent on a child’s action for alienation of a parent’s affection.24

Other injuries to family relationships have had recent legislative attention and perhaps might provide principles for one coherent view of the role of civil liability in protecting such relationships. This proves not to be so.

As between spouses, both the action for alienation of affection and that for loss of consortium began as actions available only to the husband. This court held that the act removing a wife’s civil disabilities allowed her to sue for alienation of her husband’s affection, but that the same act did not extend to the wife a cause of *563action for loss of the husband’s consortium.25 The court later explained that alienation of affections differed from loss of consortium in being an intentional tort and one “inflicted directly upon the wife” rather than a consequence of injury to another.26

The action for loss of consortium was soon extended to the wife by statute. Or Laws 1941, ch 228. The evident concern was with equality between the spouses, not with the principle of the tort itself. In more recent times, however, equality with respect to alienation of affections was provided not by making this claim available to wives but by abolishing it altogether, along with the tort of criminal conversation. Or Laws 1975, ch 562.27 The legislative history makes clear that despite their intentional character and the resulting emotional harm to the injured spouse, these actions for invasion of the family relationship were considered outmoded by changing views of marriage, divorce, and sexual relations, as reflected in the repeal in 1971 of criminal laws against adultery and enactment of no-fault divorce laws. No attention was given to the question whether children had or, if so, should retain an action for deprivation of a parent’s affection. Again, this history does not lend support to a negligence claim for psychic injury from loss of a parent’s society and companionship based specifically on the family relationship, as distinct from other relationships of close emotional dependence.

Most directly in point is the inclusion of the survivor’s loss of a decedent’s society and companionship in *564the statutory action for wrongful death. As stated above, this parallel also gave the majority in Borer v. American Airlines, supra, and the dissent in Berger v. Weber, supra, the most difficulty. We are not convinced by the California court’s distinction of the wrongful death action on the ground that, unlike an action based on a nonfatal injury, its main object is to assure that the tortfeasor not escape the deterrent effect of liability. More in point is the observation of the Michigan dissenters that the wrongful death law recognizes intangible losses of others besides a decedent’s children, not so much because this scope of the law threatens to extend recovery for such losses from nonfatal dis-ablements, but because it undercuts the premise that the law expresses a special concern for minor child’s psychic dependence on its parents.

In 1972, the history of Oregon’s wrongful death act and decisions under it were reviewed in Justice Tongue’s opinion for the Court in Coheen v. General Motors Corp., 263 Or 145, 502 P2d 223 (1972). See also Escobedo v. Ward, supra n. 23, 255 Or at 97-99. In summary, the original version of Lord Campbell’s Act enacted in the Deady Code of 1862 allowed the personal representative of a decedent to recover damages for the wrongful death of the decedent, limited to $5000 for the benefit of the estate. The act made no special provision for the decedent’s family or other dependents, and it was interpreted to allow no “solatium” for their grief and anguish but only for pecuniary loss to the estate.28 This was changed in 1939 to provide that the personal representative would act “for the benefit of the widow or widower and dependents,” and only if there were none of these for the benefit of the estate, but the economic premise of the recovery was not changed. Or Laws 1939, ch 466. The act again was amended in 1967 to remove the limitation on the maximum recovery (then $25,000) and to specify damages in terms of reasonable and fair compensation for the actual pecuniary loss, if any, to the spouse, dependents, or estate. Or Laws 1967, ch 544; see Goheen, supra 263 Or at 168-171.

*565Following Goheen, the 1973 legislature further amended the specification of damages allowable in an action for wrongful death, Or Laws 1973 ch 718. See Comment, Wrongful Death Actions in Oregon: New Developments, 10 Will L J 217 (1974). Two provisions of the 1973 statute are relevant to the present issue. One provision expanded the damages recoverable by the decedent’s spouse, children and parents to include damages “for loss of the society, companionship and services of the decedent.” ORS 30.020(2)(d).29 Another provided for the apportionment of these damages according to each beneficiary’s loss. ORS 30.040, ORS 30.050. Indisputably this legislation recognizes that a person’s immediate family suffers more from his or her untimely death than merely pecuniary loss. What does this recognition mean for the child’s cause of action for a parent’s nonfatal disablement?

The correct answer is by no means self-evident. The implications drawn from the historical development by the Chief Justice’s dissent and by Judge Roberts in the Court of Appeals are tenable conclusions. A majority of the Court, however, has concluded that the amendments to the wrongful death statute do not demonstrate a more general exception from the limits on negligence liability for one person’s loss from injury to another.

The legislative history of the 1973 wrongful death legislation offers some help. The amendments were opposed in committee testimony on the grounds that the emotional injury of lost companionship cannot be compensated by money and that such compensation would lead to increased litigation and higher insurance costs. The legislature’s rejection of these objections supports our rejection of *566the same arguments made by the present defendants and in Borer v. American Airlines, supra. On the other hand, the legislative history also makes clear that the object of the legislation was to correct shortcomings in the existing measure of damages for an established cause of action, not to recognize a new one. The prescription of “actual pecuniary loss,” if strictly applied, often would preclude any recovery for wrongful death, particularly when the decedent was very young or old or a disabled person. The sponsor of the 1973 amendments cited the opinion in Goheen to suggest that juries nevertheless would take nonpecuniary elements into consideration, so “why not put them in the statute and be frank with the jury.”30

The legislature proceeded to make that adjustment in the familiar action for wrongful death without further reexamining either the historical or the contemporary basis for that action. The tortfeasor’s liability to the survivors of a deceased victim was not in question. Although the emotional impact of the lost “society, companionship and services” resulting from a nonfatal disablement is similar to that now recognized in recoveries for wrongful death, and may even be more lasting and severe, it need not follow from this alone that a defendant’s liability is the same. The law may treat liability for causing death as a special case, in which compensation necessarily extends beyond the ordinary restitution to the immediately injured person. Perhaps that distinction between the often fortuitous death or survival of the injured person deserves reexamination when the focus is moved to the consequential harm to third persons.31 But the single analogy to the 1973 amendments of the *567wrongful death act does not suffice to show that among such third persons, children now have a cause of action for such harm from nonfatal injuries that deprive them of a parent’s society and companionship.

V. Equal protection.

Plaintiff also argues that the law must entitle him to damages for his injury in order to avoid denying him a privilege on equal terms with other citizens, contrary to article I, section 20 of the Oregon Constitution, or the equal protection of the laws, contrary to the federal 14th amendment.32 He attacks two distinctions, one between the child of a dead parent, who recovers for loss of the parent’s society, and the child of a disabled parent, who does not, the other between the spouse of a disabled person, who recovers for loss of consortium, and that person’s child, who does not. It may be questioned whether these are actually “classes” of people, at least as that word is used in article I, section 20.33 As between two children, either may lose a parent to disablement or to death, or first to one followed by the other, as indeed children lose parents who leave the family for other reasons. When the legislature provided for damages for the child’s loss in a wrongful death action, it did not distinguish between children ad hominem, by personal or social characteristics, as illustrated by laws discriminating against children born out of wedlock that have *568been found to deny them equal protection.34 The distinction is not among kinds of children but between the scope of defendants’ liability for causing fatal as distinct from nonfatal injuries to the immediate victims of their negligence.

The rules that allow a cause of action to an injured person’s spouse but not to a child can more plausibly be described as a distinction by personal or social characteristics. At least it could be so described if the law purported to deny otherwise identical claims to a child merely by reason of age or relationship, although the classification would not necessarily violate article I, section 20. The age of majority, of course, is a criterion for many legal privileges or immunities; indeed, it is the plaintiff who argues for making this the distinction between otherwise identical claims of children who have lost the society and companionship of a disabled parent.35 But the law is not constitutionally obliged to treat the spousal relationship and that between parent and child as identical. The notion that constitutional equality in tort law must be judged by the sole test of the plaintiffs harm was rejected when this court held that a guest in an automobile could be held to proving greater negligence by the driver than one who is not a guest, or a guest injured in a different setting.36 The analysis in that 1974 decision was based primarily on federal equal protection concepts, and no more recent decision of the United States Supreme Court has been cited to us as requiring a different result in the present case.

VI. Conclusion.

To recapitulate: Plaintiff’s complaint claims damages for injuries resulting from the incapacitation of his mother by defendants’ negligence. While some of the pleaded consequences to plaintiff might be described as economic, his emphasis in this court is on the loss of his mother’s society and companionship.

*569The injury encompassed within that pleading might well be proved to be real and severe. It is not ineligible for the recovery of compensatory pecuniary damages merely because the injury itself is to psychic and emotional rather than to physical or economic interests, for injury to such interests of personality and emotional well-being is compen-sable in various other contexts. The present holding does not judge the desirability of such compensation as a matter of policy.

Plaintiffs injury, however, does not result from the defendants’ negligent treatment of himself but as a consequence of their negligent treatment of his mother. He may be able to show that defendants reasonably should expect this consequential injury to a child if they caused the kind of injury to a young woman that the complaint alleges here. The obstacle to plaintiffs action is that ordinarily negligence as a legal source of liability gives rise only to an obligation to compensate the person immediately injured, not anyone who predictably suffers loss in consequence of that injury, unless liability for that person’s consequential loss has a legal source besides its foreseeability.

Plaintiff points to exceptions to this general rule in three actions for consequential injury to family members: The parent’s action for injury to a child, the spouse’s action for loss of consortium, and the wrongful death action as recently expanded to entitle a spouse, child, or parent to damages for loss of a decedent’s society, companionship, and services. The argument that these analogues imply a different rule for consequential damages within close family relationships has been well presented, and it has proved persuasive to some courts with perhaps distinguishable statutes and common law precedents. The foregoing review of these different actions in Oregon, however, does not convince us that they can be cumulated to support a general liability toward dependent children for negligent injury to their parents. Finally, we hold that the state and federal constitutions do not preclude different treatment of the injuries represented by these claims.

The decision of the Court of Appeals is affirmed.

dissenting.

I dissent because I believe plaintiff has stated ultimate facts sufficient to constitute a claim. ORCP 21A. I desire, however, to acknowledge the truly excellent discussion by the author of the majority opinion insofar as parts I, II and III are concerned. Indeed, I find little to dispute in the majority opinion preceding the discussion of the 1973 legislative amendments to the action for damages for death resulting from the wrongful act of another. It is with the rest of the opinion, commencing 293 Or at 565, 652 P2d at 330, that I find that my views diverge from those of the majority. In coming to my contrary conclusion, I desire to acknowledge the assistance of the author of the majority opinion in shaping my remarks. Both in oral conference and in written suggestion of form and content, he has been of invaluable aid to me in expressing the holding I would reach.

*571As noted in the majority opinion, the legislature in 1973 recognized that the loss the immediate family suffers from the death of a member is something above and beyond “mere” pecuniary loss and fashioned the measure of recovery accordingly. See also Prauss v. Adamski, 195 Or 1, 23-24, 244 P2d 598 (1952), in which this court had recognized some twenty years earlier a like concept:

“Furthermore, we are of the opinion that the words ‘pecuniary loss’, as applied to the damages suffered by beneficiaries of a deceased wife and mother, mean something more than the actual earning of money or money’s worth or contributions to the support of the beneficiaries at or before the date of the death, where there was a reasonable expectation of pecuniary benefit from the continuance of the life. It consists not only of the loss of financial assistance which the beneficiaries might reasonably be expected to have received from the deceased had her career not been shortened by the act of defendant, but also the loss of other things which have a pecuniary worth, such as the loss of a mother’s care and attention to the physical, moral, and educational welfare of her children, and a husband’s loss of her services in the household.”

The measure of damages thus legislatively recognized was apportioned according to each beneficiary’s loss. Clearly, the 1973 legislation reaffirmed and extended liability for ordinary negligence to persons whose loss is an “indirect” or “secondary” consequence of immediate injury to another. If such extended, consequential liability were deemed a departure from principle and an anachronistic anomaly in negligence law, the attention given the subject in 1973 provided an occasion to reconsider it. The legislature and citizens concerned with tort liability had the opportunity to reexamine whether recovery of damages for wrongful death was properly premised on a tortfeasor’s liability to the decedent, represented by his estate, for the destruction of the decedent’s own economic future, or whether the premise was compensation for the consequential losses of his surviving family, or perhaps both.

Neither the legislature nor witnesses questioned the propriety of consequential liability to members of the victim’s immediate family. In fact, the legislature extended this liability to include nonpecuniary, emotional losses that *572in no way could be a substitute for the victim’s own economic loss. The amendments were opposed in committee testimony on the grounds that the emotional injury of lost companionship cannot be compensated by money and that such compensation would lead to increased litigation and higher insurance costs, the very arguments made by the present defendants and accepted by the California court in Borer v. American Airlines, Inc., 19 Cal 3d 441, 563 P2d 858, 138 Cal Rptr 302 (1977). The legislature rejected these objections to liability for the nonpecuniary loss of society and companionship of a spouse, parent, or child.

Of course, this does not mean that the legislature implicitly enacted a parallel liability when nonfatal injuries to the initial victim are corresponding losses on the stated family members. It shows, however, that the usual limitation of negligence liability recognized in Ore-Ida Foods v. Indian Head, 290 Or 909, 627 P2d 469 (1981), is not an ironclad principle. It does not always stand in the way of ordinary liability for negligently inflicted injury based on its foreseeability. It has not been a principle in the narrow area of immediate family relationships. That much is true of the long existing parent’s action under ORS 30.010, as well as of the spouse’s action for loss of consortium, whatever else distinguishes them from the present claim. The legislature might have repealed the husband’s action instead of extending it to the wife in 1941, just as it might have reconsidered the individual family member’s consequential claims in 1973, if the limitation of liability to immediate victims were thought an overriding principle of negligence law. Obviously it has not been thought so. Sometimes it has yielded to compensation for consequential injury suffered by a tort victim’s spouse, parent, or child.

The present plaintiff has pleaded an ordinary negligence claim except insofar as it may lie beyond the scope of negligence liability by resulting from injury to a third person, his mother, and by claiming damages for loss of society and companionship. The majority has discussed why the first of these is not an insuperable obstacle. In retaining and expanding the claims there reviewed, the legislature did not merely retain historic anomalies. Its abolition of actions for alienation of affection shows that it will deal with perceived anachronisms in family law. Rather, the *573legislature’s choice to retain and expand compensation for those family members closest to an injured tort victim echoes implicit social assumptions and values that also form much of the common law.

The implicit assumptions reflected in these highly selective torts do not concern the justice of compensating consequential injury as such. They do not extend to all who in fact suffer genuine material or emotional injury in the loss of the immediate victim’s support or society. They do not even extend to every close relative. Rather, what the history of treatment of these torts reflects is the value that has long and continuously been attached specifically to the family relationships between married spouses1 and between parents and children. The premise of recovery in tort was not emotional dependency in fact but assumptions arising from these specific relationships, though, of course, a recovery would depend on the facts in the individual case.

The barrier that limits negligence liability to the initially injured victim and prevents recovery by others foreseeably injured as a consequence, as in Ore-Ida Foods, therefore does not govern here. Nor does the nature of the harm for which the child seeks compensation bar recovery. As the majority acknowledges, this court has allowed recovery based on ordinary negligence for various kinds of psychic or emotional harm, unaccompanied by physical injury, when the negligent defendant invaded some legally protected interest of plaintiff separate from the infliction of the psychic or emotional injury itself. As between a child and its custodial parent, the very concept of legal custody expresses the set of mutual obligations and interests imposed and protected by law.2 The child’s interest in the *574relationship of custody and its accompanying rights surely is no less than that of the parent which was held to support a negligence claim in McEvoy v. Helikson, 277 Or 781, 562 P2d 540 (1977).

It cannot be said that the consequential injury to this plaintiff was beyond the class of foreseeable injuries as a matter of law. This is not a case of negligence toward one of many unknown potential victims, such as a collision with a vehicle carrying an unknown driver or passenger, or liability for unsafe business premises, where foreseeability might be reduced to an issue of mere statistics. Here we deal only with liability in a relationship of professional responsibility for the health of a known patient, a 25 year old woman. A factfinder could well conclude that defendants should have foreseen that such a patient would be likely to have one or more minor children, if not indeed that defendants knew that she did. The issue of foreseeability is one for determination on the evidence adduced at trial. Stewart v. Jefferson Plywood Co., 255 Or 603, 469 P2d 783 (1970).

For these reasons, I would hold plaintiff has pleaded a claim under the ordinary principles of liability for injury foreseeably resulting from defendants’ negligence.

concurring.

I concur generally, but add these few words of qualification.

The majority reasons that the anticipated effect of our holding on insurance rates is immaterial to our decision and I agree. I do not wish to be understood, however, as holding that economic realities relevant to assignment of risk are never relevant in making policy judgments as discussed in Ore-Ida Foods v. Indian Head, 290 Or 909, 627 P2d 469 (1981).

Also, I am hesitant to create judicially an entitlement as unpredictable, formless and limitless as compensation of children for psychic losses due to nonfatal injury to their parents. Unlike the courts, the legislature has the power to create not only new bases for recovery but also appropriate procedures to account for all affected interests. The legislature did so with the current wrongful death statutes, for example. ORS 30.020. Thus, I do not necessarily reject the Iowa approach; rather I recognize that practical limits sometimes render judicial power an unsuitable means for the recognition of new theories of recovery.

Campbell and Carson, JJ., join in this opinion.