Martin, J.

The facts in this case are few and may be briefly stated. On the first day of April, 1891, the plaintiff -was standing upon a crosswalk on Main street in the city of Rochester, awaiting an opportunity to board one of the defendant’s cars which had- stopped upon the street at that place. While standing there, and just as she was about to step upon the car, a horse car of the defendant came down the street. As the team attached to the car drew near, it turned to the right and came so close to the plaintiff that she stood between the horses’ heads when they were stopped.

She testified that from fright and excitement caused by the *109approach and proximity of the team she became unconscious, and also that the result was a miscarriage and consequent illness. Medical testimony was given to the effect that the mental shock which she then received was sufficient to produce that result.

Assuming that the evidence tended to show that the defendant’s servant was negligent in the management of the car and horses, and that the plaintiff was free from contributory negligence, the single question presented is whether the plaintiff is entitled to recover for the defendant’s negligence which occasioned her fright and alarm, and resulted in the injuries already mentioned. While the authorities are not harmonious upon this question, we think the most reliable and better considered cases, as well as public policy, fully justify us in holding that the plaintiff cannot recover for injuries occasioned by fright, as there was no immediate personal injury. (Lehman v. Brooklyn City R. R. Co., 47 Hun, 355 ; Victorian Railways Commissioners v. Coultas, L. R. [13 Appeal Cases] 222; Ewing v. P., C. & St. L. Ry. Co., 147 Penn. St. 40.) The learned counsel for the respondent in his brief very properly stated that, “The consensus of opinion would seem to be that no recovery can be had for mere fright,” as will be readily seen by an examination of the following additional authorities: Haile v. Texas & Pacific R. Co. (23 Lawyers’ Rep. 774); Joch v. Dankwardt (85 Ill. 331); Canning v. Inhabitants of Williamstown (1 Cush. 451); Western Union Tel. Co. v. Wood (57 Fed. Repr. 471); Renner v. Canfield (36 Minn. 90); Allsop v. Allsop (5 Hurl. & Nor. [N. S.] 534); Johnson v. Wells, Fargo & Co. (6 Nev. 224); Wyman v. Leavitt (71 Me. 227).

If it be admitted that no recovery can be had for fright occasioned by the negligence of another, it is somewhat difficult to understand how a defendant would be liable for its consequences. Assuming that fright cannot form the basis of an action, it is obvious that no recovery can be had for injuries resulting therefrom. That the result may be nervous disease, blindness, insanity, or even a miscarriage, in *110no way changes the principle. These results merely show the degree of fright or the extent of the damages. The right of action must still depend upon the question whether a recovery may he had for fright. If it can, then an action may be maintained, however slight the injury. If not, then there can be no recovery, no matter how grave or serious the consequences. Therefore, the logical result of the respondent’s concession would seem to be, not only that no recovery can / be had for mere fright, but also that none can be had for [ injuries which are the direct consequences of it.

If the right of recovery in this class of cases should be once established, it would naturally result in a flood, of litigation in cases where the injury complained of may be easily feigned without detection, and where the damages must rest upon mere conjecture or speculation. The difficulty which often exists in cases of alleged physical injury, in determining whether they exist, and if so, whether they were caused by the negligent act of the defendant, would not only be greatly increased, but a wide field would be opened for fictitious or speculative claims. To establish such a doctrine would be contrary to principles of public policy.

Moreover, it cannot be properly said that the plaintiff’s mis- 1 carriage was the proximate result of the defendant’s negligence. Proximate damages are such as are the ordinary and natural results of the negligence charged, and those that are usual and may, therefore, be expected. It is quite obvious that the plaintiff’s injuries do not fall within the rule as to proximate damages. The injuries to the plaintiff were plainly the result of an accidental or unusual combination of circumstances, which could not have been reasonably anticipated, and over which the defendant had no control, and, hence, her damages were too remote to justify a recovery in this action.

These considerations lead to the conclusion that no recovery can be had for injuries sustained by fright occasioned by the negligence of another, where there is no immediate per-injury.

The orders of the General and Special Terms should be *111reversed, and the order of the Trial Term granting a nonsuit affirmed, with costs.

All concur, except Haight, J., not sitting, and Vann, J., not voting.

Ordered accordingly.

Wickhem, J.

In the statement of facts in both briefs it is said that deceased was looking out the window of her house watching her child cross the highway, and witnessed the negligent killing of the child by defendant. While upon a demurrer the sole question is whether the facts alleged in the complaint state a cause of action, we consider that the statement of facts concurred in by plaintiff and defendant constitutes an informal amendment to the complaint by stipulation, and will determine the questions presented as though the complaint were amended to conform to the statement of facts.

*605The question presented is whether under the Wisconsin equivalent of Lord Campbell’s Act, decedent’s husband may recover for her death under such circumstances. Under the provisions of sec. 331.03, Stats., in order that he may recover .for her death, the circumstances must have been such as to have entitled Susie Waube, had she lived, to maintain an action for her injuries. Koehler v. Waukesha Milk Co. 190 Wis. 52, 208 N. W. 901. Thus the question presented is whether the mother of a child who, although not put in peril or fear of physical impact, sustains the shock of witnessing the negligent killing of her child, may recover for physical injuries caused by such fright or shock.

The problem must be approached at the outset from the view-point of the duty of defendant and the right of plaintiff, and not from the view-point of proximate cause. The right of the mother to recover must be based, first, upon the establishment of a duty on the part of defendant so to conduct herself with respect to the child as not to subject the mother to an unreasonable risk of shock or fright, and, second, upon the recognition of a legally protected right or interest on the part of the mother to be free from shock or fright occasioned by the peril of her child. It is not enough to find a breach of duty to the child, follow the consequences of such breach as far as the law of proximate cause will permit them to go, and then sustain a recovery for the mother if a physical injury to her by reason of shock or fright is held not too remote.

Upon this point we adopt and follow the doctrine of Palsgraf v. Long Island R. R. Co. 248 N. Y. 339, 162 N. E. 99. In that case the facts are thus stated by the court:

“Plaintiff was standing on a platform of defendant’s railroad after buying a ticket to go to Rockaway Beach. A train stopped at the station, bound for another place. Two men ran forward to catch it. One of the men reached the platform of the car without mishap, though the train was al*606ready moving. The other man, carrying a package, jumped aboard the car, but seemed unsteady as if about to fall. A guard on the car, who had held the door open, reached forward to help him in, and another guard on the platform pushed him from behind. In this act, the package was dislodged, and fell upon the rails. It was a package of small size, about fifteen inches long, and was covered by a newspaper. In fact it contained fireworks, but there was nothing in its appearance to give notice of its contents. The fireworks when they fell exploded. The shock of the explosion threw down some scales at the other end of the platform, many feet away. The scales struck the plaintiff, causing injuries for which she sues.”

The court, speaking through Judge Cardozo, said:

“Negligence is not actionable unless it involves the invasion of a legally protected interest, the violation of a right. . . . The plaintiff sues in her own right for a wrong personal to her, and not as the vicarious beneficiary of a breach of duty to another. . . . The passenger far away, if the victim of a wrong at all, has a cause of action, not derivative, but original and primary. His claim to be protected against invasion of his bodily security is neither greater nor less because the act resulting in the invasion is a wrong to another far removed.”

See also 27 Illinois Law Review, pp. 774, 775, in which Dean Green, referring to the case of Hambrook v. Stokes Bros. (1925), 1 K. B. 141, hereafter to be considered, stated:

“The only problem was whether defendants owed the wife a duty with respect to possible injury to herself as a result of their negligence in imperiling the lives of her children.”

The right of a plaintiff to recover damages for nervous shock caused by negligence without actual impact has had an interesting history. See 34 Harvard Law Review, 260; 11 Cornell Law Quarterly, 513; 11 American Law Register, 141. In Victoria Railways Commissioners v. Coultas (1888), 13 A. C. 222, 226, it was held that plaintiff was not *607entitled to recover such damages. In Lehman v. Brooklyn City R. Co. (1888), 47 Hun (N. Y.), 355, 356, the same doctrine was announced. This became the prevailing doctrine in this country. Ewing v. Pittsburgh, C., C. & St. L. R. Co. 147 Pa. St. 40, 23 Atl. 340; Haile’s Curator v. Texas & Pacific R. Co. 60 Fed. 557, 9 C. C. A. 134; Mitchell v. Rochester R. Co. 151 N. Y. 107, 45 N. E. 354; Spade v. Lynn & Boston R. Co. 168 Mass. 285, 47 N. E. 88. This doctrine, however, was repudiated in a number of jurisdictions, including Wisconsin, in situations where fright without impact produced physical injuries. Alabama Fuel & Iron Co. v. Baladoni, 15 Ala. App. 316, 73 So. 205; Lindley v. Knowlton, 179 Cal. 298, 176 Pac. 440; Goddard v. Watters, 14 Ga. App. 722, 82 S. E. 304; Watson v. Dilts, 116 Iowa, 249, 89 N. W. 1068; Whitsel v. Watts, 98 Kan. 508, 159 Pac. 401; Stewart v. Arkansas Southern R. Co. 112 La. 764, 36 So. 676; Green v. Shoemaker, 111 Md. 69, 73 Atl. 688; Purcell v. St. Paul City R. Co. 48 Minn. 134, 50 N. W. 1034; Kimberly v. Howland, 143 N. C. 398, 55 S. E. 778; Salmi v. Columbia & N. R. R. Co. 75 Oreg. 200, 146 Pac. 819; Simone v. Rhode Island Co. 28 R. I. 186, 66 Atl. 202; Mack v. South Bound R. Co. 52 S. C. 323, 29 S. E. 905; Sternhagen v. Kozel, 40 S. D. 396, 167 N. W. 398; Memphis St. R. Co. v. Bernstein, 137 Tenn. 637, 194 S. W. 902; Gulf, C. & S. F. R. Co. v. Hayter, 93 Tex. 239, 54 S. W. 944; O’Meara v. Russell, 90 Wash. 557, 156 Pac. 550; Pankopf v. Hinkley, 141 Wis. 146, 123 N. W. 625; Sundquist v. Madison Rys. Co. 197 Wis. 83, 221 N. W. 392. See notes 11 A. L. R. 1119; 40 id. 983; 76 id. 681. In Pankopf v. Hinkley, supra, the conflict is noted and it is stated that the better reason favors the view that there is a cause of action for shock sustained without impact but resulting in physical injuries. It had previously been held in this state, in Summerfeld v. Western Union Tel. Co. 87 *608Wis. 1, 57 N. W. 973, and Gatzow v. Buening, 106 Wis. 1, 20, 81 N. W. 1003, that where there was no impact, and where there were no subsequent physical injuries caused by fright, no cause of action existed. The rule followed in Wisconsin appears to represent the modem tendency and to be now the rule of the New York court. Comstock v. Wilson, 257 N. Y. 231, 177 N. E. 431, 76 A. L. R. 676. In jurisdictions following the liberal rule it has been held consistently, with but two exceptions, hereafter to be noted, that in order to give rise to a right of action grounded on negligent conduct, the emotional distress or shock must be occasioned by fear of personal injury to the person sustaining the shock, and not fear of injury to his property or to the person of another. Dtilieu v. White & Sons (1901), 2 K. B. 669; Sanderson v. Northern Pac. R. Co. 88 Minn. 162, 92 N. W. 542; Mahoney v. Dankwart, 108 Iowa, 321, 79 N. W. 134; McGee v. Vanover, 148 Ky. 737, 147 S. W. 742; Bucknam v. Great Northern R. Co. 76 Minn. 373, 79 N. W. 98; Cleveland, C., C. & St. L. R. Co. v. Stewart, 24 Ind. App. 374, 381, 56 N. E. 917; Southern R. Co. v. Jackson, 146 Ga. 243, 91 S. E. 28.

Thus it may be said that the doctrine most favorable to plaintiff is not sufficiently broad to' entitle him to recover. The question presented is whether there should be an extension of the rule to cases where defendant’s conduct involves merely an unreasonable risk of causing harm to the child or spouse of the person sustaining injuries through fright or shock. There are two cases upon this point which deserve consideration in some detail. In Spearman v. McCrary, 4 Ala. App. 473, 58 So. 927, defendant’s negligent operation of an automobile caused plaintiff’s mule to take fright and run away. The mule was attached to a buggy from which plaintiff and her husband had just alighted but in which plaintiff’s two children still remained. The damages alleged were shock and fright resulting in physical injuries. The *609court there adopted the view that recovery could be had for shock or fright culminating in' physical injury although caused without physical impact. The court approaches the problem from the standpoint of proximate cause, and does not discuss the precise point involved here, although in fact recovery appears to have been given for physical injuries sustained through fright occasioned by the peril of the children rather than that of the mother. In the situation presented by that case, however, plaintiff may well have suffered fright and shock as a result of fear for her own safety.

The only case squarely dealing with this problem is Hambrook v. Stokes Bros. (1925), 1 K. B. 141. In this case a servant of defendants was in charge of a motor-truck belonging to defendants, and parked it at the top of a hill on Dover street in Folkestone, leaving it unattended, with the motor running, and without taking proper precautions to prevent it from moving. During his absence the truck started to run down the hill. The street was narrow, being not more than six feet wide in some places, and there was a curve at the lower end of it. The truck eventually came to a standstill by reason of running against the side of a house at a point below the curve. On the day in question Mrs. Hambrook, whose house was at the bottom of Dover street, accompanied her three children, a girl and two boys, part of the distance on their way to school. She walked with them to a point a little below the curve in Dover street, and then left them. Shortly afterwards she saw the truck coming rapidly around the curve in her direction. She was not in any personal danger, as the truck stopped some distance short of where she was standing, and in any case she would have had ample opportunity to step into a position of safety. She immediately became fearful for the safety of her children. A crowd collected and there were rumors of an accident. She inquired who had been injured, and a friend stated that it was a little girl with glasses. It appeared that *610her little girl wore glasses. She went to the hospital and found that her daughter had been injured. She sustained a severe shock and consequent physical injuries from which she died. The trial court- directed the jury that if the nervous shock was caused by fear of her child’s safety, as distinguished from her own, plaintiff could not recover. From a verdict in favor of defendants, plaintiff appealed. The judgment was set aside and a new trial granted for misdirection. The court reviewed the state of the English law with respect to this subject, pointing out that the Coultas Case, supra, had-been modified, and recognizing also that the Dulieu .Case, supra, had limited the recovery in fright cases to situations where the fright was occasioned by the peril of the person affrighted. Viewing the matter from the standpoint of proximate cause rather than duty, the court held that there should be no distinction between shock sustained by a mother as a result of fear for her own safety, and that sustained by reason of peril to her child. The court considered that defendant ought to have anticipated that if. the unattended truck ran down this narrow street it might terrify some woman to such an extent, through fear of some immediate bodily injury to herself, that she would receive a mental shock with resultant physical injuries, and that defendant ought also to have anticipated that such a shock might result from the peril to the child of such a woman. While the majority, mistakenly, as it seems to us, approach this problem from the standpoint of proximate cause, the dissenting opinion of Sargant, L. J., approaches it from the standpoint of duty. The dissenting opinion concedes that since it was defendant’s duty to exercise due care in the management of his vehicle so as to avoid physical injury to those on or near the highway, this duty cannot be limited to physical injuries caused by actual physical impact. The dissenting opinion, however, states that the matter is quite *611different where the shock to plaintiff is due, not to immediate fear of personal impact, but to the sight or apprehension of impact upon a third person. Reference is made to the unreported case of Smith v. Johnson & Co., referred to by Wright, J., in his judgment in Wilkinson v. Downton (1897), 2 Q. B. 57, 61. There the plaintiff became ill from the shock of seeing another killed, the death being due to the negligence of the defendant. It was held that the plaintiff could not recover, not on the ground that the harm was too remote, “but on (what is often practically equivalent) a consideration of the extent of the duty of the defendant towards the plaintiff and others on and near the highway.” The dissenting opinion then states that “it would be a considerable and unwarranted extension of the duty of owners of vehicles towards others in or near the highway, if it were held to include an obligation not to do anything to render them liable to harm through nervous shock caused by the sight or apprehension of damage to third persons.” The dissenting opinion concludes that there is no sound reason for erecting an exception in favor of the mother of a child, and points out that once the defendant’s duty is held to extend to those outside the field of physical peril, a doctrine is stated to which no rational boundaries can be erected.

The American Law Institute, in its Restatement of Torts, thus states the law:

Ҥ 313. Emotional Distress Unintended.

“If the actor unintentionally causes emotional distress to another, he is liable to the other for illness or bodily harm of which the distress is a legal cause if the actor

“(a) should have realized that his conduct involved an unreasonable risk of causing the distress, otherwise than by knowledge of the harm or peril of a third'person, and

“(b) from facts known to him should have realized that the distress, if it were caused, might result in illness or bodily harm.”

*612It is thus recognized that as the law exists, there is no liability for fright, negligently occasioned, where that fright or shock is the result of knowledge of the injury or peril of a third person. The Institute, however, adds a caveat as follows:

“The Institute expresses no opinion as to whether an actor whose conduct is negligent as involving an unreasonable risk of causing bodily harm to a child or spouse is liable for an illness or other bodily harm caused to the parent or spouse who witnesses the peril or harm of the child or spouse and thereby suffers anxiety or shock which is the legal cause of the parent’s or spouse’s illness or other bodily harm.”

Whether this caveat was added because of a difference of opinion, a dearth of authority, or a consideration of the English case is immaterial. At any rate, there is no authority except the English case, and possibly the Alabama case, for extending the duty of the defendant.as contended for by plaintiff in this case, nor do we think that sound principle calls for such an extension.

With due deference to the learned judges who concurred in the decision, we cannot • escape the conclusion that the determination in the Hombro ok Case is incorrect, both in its initial approach and in its conclusion, and that the doctrine contended for by plaintiff, and there approved, would constitute an unwarranted enlargement of the duties of users of the highway. Fundamentally, defendant’s duty was to use ordinary care to avoid physical injury to those who would be put in physical peril, as that term is commonly understood, byo conduct on his part falling short of that standard. It is one thing to say that as to those who are put in peril of physical impact, impact is immaterial if physical injury is caused by shock arising from the peril. It is the foundation of cases holding to this liberal ruling, that the person af-frighted or sustaining shock was actually put in peril of physical impact, and under these conditions it was consid*613ered immaterial that the physical impact did not materialize. It is quite another thing to say that those who are out of the field of physical danger through impact shall have a legally protected right to be free from emotional distress occasioned by the peril of others, when that distress results in physical impairment. The answer to this question cannot be reached solely by logic, nor is it clear that it can be entirely disposed of by a consideration of what the defendant ought reasonably to have anticipated as a consequence of his wrong. The answer must be reached by balancing the social interests involved in order to ascertain how far defendant’s duty and plaintiff’s right may justly and expediently be extended. It is our conclusion that they can neither justly nor expediently be extended to any recovery for physical injuries sustained by one out of the range of ordinary physical peril as a result of the shock of witnessing another’s danger. Such consequences are so unusual and extraordinary, viewed after the event, that a user of the highway may be said not to subject others to an unreasonable risk of them by the careless management of his vehicle. Furthermore, the liability imposed by such a doctrine is wholly out of proportion to the culpability of the negligent tort-feasor, would put an unreasonable burden upon users of the highway, open the way to fraudulent claims, and enter a field that has no sensible or just stopping point.

It was recognized by the court in the Hambrook Case that had the mother there merely been told of the injury to her child, instead of having been virtually a witness to the transaction, there would have been no liability. The court thus selected at least one arbitrary boundary for the extension. While there is no suggestion in the caveat attached to Restatement, Torts, §313 (b), as to what the proper rule should be, it is implied that any possible extension of the rule as stated would be limited, (1) to the relationship of *614parent and child or husband and wife, and, (2) to persons sustaining the shock as a result of witnessing the peril. It was suggested in the dissenting opinion in the Hambrook Case that if the mother may recover, why not a child whose shock was occasioned by the peril of the mother? It is not necessary to multiply these illustrations. They can be made as numerous as the varying degrees of human relationship, and they shade into each other in such a way as to leave no definite or clear-cut stopping place for the suggested doctrine, short of a recovery for every person who has sustained physical injuries as a result of shock or emotional distress by reason of seeing or hearing of the peril or injury of another. No court has gone this far, and we think no court should go this far. It is our view that fairness and justice, as well as expediency, require the defendant’s duty to be defined as heretofore stated, in accordance with the weight of liberal authority and the general statement of the rule by the American Law Institute. Human wrong-doing is seldom limited in. its injurious effects to the immediate actors in a particular event. More frequently than not, a chain of results is set up that visits evil consequences far and wide. While from the standpoint of good morals and good citizenship the wrong-doer may be said to violate a duty to those who suffer from the wrong, the law finds it necessary, for reasons heretofore considered, to attach practical and just limits to the legal consequences of the wrongful act. As was said by the New York court, speaking through Judge Cardozo, in Palsgraf v. Long Island R. R. Co., supra:

“Negligence is not a tort unless it results in the commission of a wrong, and the commission of a wrong imports the violation of a right, in this case, we are told, the right to be protected against interference with one’s bodily security. But bodily security is protected, not against all forms of interference or aggression, but only against some. One who seeks redress at law does not make out a cause of *615action by showing without more that there has been damage to his person. If the harm was not willful, he must show that the act as to him had possibilities of danger so many and apparent as to entitle him to be protected against the doing of it though the harm was unintended. Affront to personality is still the keynote of the wrong.”

The allegations in plaintiff’s complaint do not bring the interest of Susie Waube within the field of legally protected rights.

By the Court. — Order reversed, and cause remanded with directions to sustain the demurrer.

TOBRINER, J.

That the courts should allow recovery to a mother who suffers emotional trauma and physical injury from witnessing the infliction of death or injury to her child for which the tortfeasor is liable in negligence would appear to be a compelling proposition. As Prosser points out, “All ordinary human feelings are in favor of her [the mother’s] action against the. negligent defendant. If a duty to her requires that she herself be in some recognizable danger, then it has properly been said that when a child is endangered, it is not beyond contemplation that its mother will be somewhere in the vicinity, and will suffer serious shock.” (Prosser, Law of Torts (3d ed. 1964) p. 353.)

Nevertheless, past American decisions have barred the mother’s recovery. Refusing the mother the right to take her ease to the jury, these courts ground their position on an alleged absence of a required “duty” of due care of the tortfeasor to the mother. Duty, in turn, they state, must express public policy; the imposition of duty here would work disaster because it would invite fraudulent claims and it would involve the courts in the hopeless task of defining the extent of the tortfeasor’s liability. In substance, they say, definition of liability being impossible, denial of liability is the only realistic alternative.

We have concluded that neither of the feared dangers *731excuses the frustration of the natural justice upon which the mother’s claim rests. We shall point out that in the past we have rejected the argument that we should deny recovery upon a legitimate claim because other fraudulent ones may be urged. We shall further explain that the alleged inability to fix definitions for recovery on the different facts of future eases does not justify the denial of recovery on the specific facts of the instant case; in any event, proper guidelines can indicate the extent of liability for such future cases.

In the instant case plaintiff’s1 first cause of action alleged that on or about September 27, 1964, defendant drove his automobile in a southerly direction on Bluegrass Road near its intersection with Clover Lane in the County of Sacramento, and at that time plaintiff’s infant daughter, Erin Lee Dillon, lawfully crossed Bluegrass Road. The complaint further alleged that defendant’s negligent operation of his vehicle caused it to “collide with the deceased Erin Lee Dillon resulting in injuries to decedent which proximately resulted in her death.” (Complaint, p. 3.) Plaintiff, as the mother of the decedent, brought an action for compensation for the loss.

Plaintiff’s second cause of action alleged that she, Margery M. Dillon, “was in close proximity to the . . . collision and personally witnessed said collision.” She further alleged that “because of the negligence of defendants . . . and as a proximate cause [sic] thereof plaintiff . . . sustained great emotional disturbance and shock and injury to her nervous system” which caused her great physical and mental pain and suffering.

Plaintiff’s third cause of action alleged that Cheryl Dillon, another infant daughter, was “in close proximity to the . . . collision and personally witnessed said collision.” Because of the negligence, Cheryl Dillon “sustained great emotional disturbance and shock and injury to her nervous system” which caused her great physical and mental pain and suffering.

On December 22, 1965, defendant, after he had filed his answer, moved for judgment on the pleadings, contending that “No cause of action is stated in that allegation that plaintiff sustained emotional distress, fright or shock induced by apprehension of negligently caused danger or injury or the witnessing of negligently caused injury to a third person. *732 Amaya v. Home Ice, Fuel & Supply Co., 59 Cal.2d 295 [29 Cal.Rptr. 33, 379 P.2d 513] (1963). Even where a child, sister or spouse is the object of the plaintiff’s apprehension no cause of action is stated, supra, p. 303, unless the complaint alleges that the plaintiff suffered emotional distress, fright or shock as a result of fear for his own safety. Reed v. Moore, 156 Cal.App.2d 43 (1957) at page 45 [319 P.2d 80].” (Italics added.) The court granted a judgment on the pleadings against the mother’s count, the second cause of action, and denied it as to the sister’s count, the third cause of action. The court, further, dismissed the second cause of action. Margery M. Dillon, the mother, appealed from that judgment.

Thereafter, on January 26, further proceedings took place as to the third cause of action, Cheryl Dillon’s claim for emotional trauma from witnessing her sister’s death while ‘1 watching her sister lawfully cross Bluegrass Road. ’ ’

Defendant moved for summary judgment on this count. In opposition plaintiff contended that the declaration of one McKinley disclosed that Mrs. Dillon testified at her deposition that when she saw the car rolling over Erin she noted that Cheryl was on the curb, but that the deposition of Cheryl Dillon contradicts such statements. Plaintiff therefore submitted that “Since the declarations filed by defendant are contradictory and the testimony contained in the testimony of Mrs. Dillon does not establish as a matter of law that Cheryl Dillon was not in the zone of danger or had fear for her own safety, plaintiff respectfully submits that the motion must be denied. ’ ’

The court denied the motion for summary judgment on the third cause as to Cheryl on the ground that the pretrial order precluded it. The trial court apparently sustained the motion for judgment on the pleadings on the second cause as to the mother because she was not within the zone of danger and denied that motion as to the third cause involving Cheryl because of the possibility that she was within such zone of danger or feared for her own safety. Thus we have before us a case that dramatically illustrates the difference in result flowing from the alleged requirement that a plaintiff cannot recover for emotional trauma in witnessing the death of a child or sister unless she also feared for her own safety because she was actually within the zone of physical impact.

The posture of this case differs from that of Amaya v. Home Ice, Fuel & Supply Co. (1963) 59 Cal.2d 295, 298 [29 Cal.Rptr. 33, 379 P.2d 513], which involved “fright or ner*733vous shock (with consequent bodily illness) induced solely by . . . apprehension of negligently caused danger or injury to a third person” because the complaint here presents the claim of the emotionally traumatized mother, who admittedly was not within the zone of danger, as contrasted with that of the sister, who may have teen within it. The ease thus illustrates the fallacy of the rule that would deny recovery in the one situation and grant it in the other. In the first place, we can hardly justify relief to the sister for trauma which she suffered upon apprehension of the child’s death and yet deny it to the mother merely because of a happenstance that the sister was some few yards closer to the accident. The instant case exposes the hopeless artificiality of the zone-of-danger rule. In the second place, to rest upon the zone-of-danger rule when we have rejected the impact rule becomes even less defensible. We have, indeed, held that impact is not necessary for recovery (Cook v. Maier (1939) 33 Cal.App.2d 581, 584 [92 P.2d 434]). The zone-of-danger concept must, then, inevitably collapse because the only reason for the requirement of presence in that zone lies in the fact that one within it will fear the danger of impact. At the threshold, then, we point to the incongruity of the rules upon which any rejection of plaintiff’s recovery must rest.

We further note, at the outset, that defendant has interposed the defense that the contributory negligence of the mother, the sister, and the child contributed to the accident. If any such defense is sustained and defendant found not liable for the death of the child because of the contributory negligence of the mother, sister or child, we do not believe that the mother or sister should recover for the emotional trauma which they allegedly suffered. In the absence of the primary liability of the tortfeasor for the death of the child, we see no ground for an independent and secondary liability for claims for injuries by third parties. The basis for such claims must be the adjudicated liability and fault of defendant; that liability and fault must be the foundation for the tortfeasor’s duty of due care to third parties who, as a consequence of such negligence, sustain emotional trauma.

We turn then to an analysis of the concept of duty, which, as we have stated, has furnished the ground for the rejection - of such claims as the instant one. Normally the simple facts of ' plaintiff’s complaint would establish a cause of action: the complaint alleges that defendant drove his car (1) negligently, as a (2) proximate result of which plaintiff suffered *734(3) physical injury. Proof of these facts to a jury leads to recovery in damages; indeed, such a showing represents a classic example of the type of accident with which the law of negligence has been designed to deal.

The assertion that liability must nevertheless be denied because defendant bears no “duty” to plaintiff “begs the essential question—whether the plaintiff’s interests are entitled to legal protection against the defendant’s conduct. . . . It [duty] is a shorthand statement of a conclusion, rather than an aid to analysis in itself. . . . But it should be recognized that ‘duty’ is not sacrosanct in itself, but only an expression of the sum total of those considerations of policy which lead the law to say that the particular plaintiff is entitled to protection.” (Prosser, Law of Torts, supra, at pp. 332-333.)

The history of the concept of duty in itself discloses that it is not an old and deep-rooted doctrine hut a legal device of the latter half of the nineteenth century designed to curtail the feared propensities of juries toward liberal awards. “It must not be forgotten that ‘duty’ got into our law for the very purpose of combatting what was then feared to be a dangerous delusion (perhaps especially prevalent among juries imbued with popular notions of fairness untempered by paramount judicial policy), viz., that the law might countenance legal redress for all foreseeable harm.” (Fleming, An Introduction to the Law of Torts (1967) p. 47.)

Indeed, the idea of court-imposed restrictions on recovery by means of the concept of “duty” contrasted dramatically with the preceding legal system of feudal society.2 In the enclosed feudal society, the actor bore responsibility for any damage he inflicted without regard to whether he was at fault or owed a “duty” to the injured person. Thus, at that time, the defendant owed a duty to all the world to conduct himself *735without causing injury to his fellows. It may well be that the physical contraction of the feudal society imposed an imperative for maximum procurable safety and a corresponding absolute responsibility upon its members.

The Industrial Revolution, Avhich cracked the solidity of the feudal society and opened up wide and new areas of expansion, changed the legal concepts. Just as the ubav competitiveness in the economic sphere figuratively broke out of the walls of the feudal community, so it broke through the rule of strict liability. In the place of strict liability it introduced the theory that an action for negligence would lie only if the defendant breached a duty which he owed to plaintiff. As Lord Esher said in Le Lievre v. Gould (1893) 1 Q.B. 491, 497: “A man is entitled to be as negligent as he pleases towards the whole world if he owes no duty to them. ’ ’

We have pointed out that this late 19th century concept of duty, as applied to the instant situation, has led the courts to deny liability. We have noted that this negation of duty emanates from the twin fears that courts will be flooded with an onslaught of (1) fraudulent and (2) indefinable claims. We shall point out why Ave think neither fear justified.

1. This court in the past has rejected the argument that we must deny recovery upon a legitimate claim because other fraudulent ones may be urged.

The denial of “duty” in the instant situation rests upon the prime hypothesis that allowance of such an action would lead to successful assertion of fraudulent claims. (See, e.g., Waube v. Warrington (1935) 216 Wis. 603, 613 [258 N.W. 497].) The rationale apparently assumes that juries, confronted by irreconcilable expert medical testimony, Avill be unable to distinguish the deceitful from the bona fide. The argument concludes that only a per se rule denying the entire class of claims that potentially raises this administrative problem3 can avoid this danger.

In the first instance, the argument proceeds from a doubtful factual assumption. Whatever the possibilities of fraudulent claims of physical injury by disinterested spectators of an accident, a question not in issue in this case, we certainly *736cannot doubt that a mother who sees her child killed will suffer physical injury from shock. “It seems sufficiently obvious that the shock of a mother at danger or harm to her child may be both a real and a serious injury.” (Prosser, Law of Torts, supra, at p. 353.)

Over a half-century ago this court recognized the likelihood that such fright and fear would cause physical injury. In Sloane v. Southern California Ry. Co. (1896) 111 Cal. 668, 680 [44 P. 320, 32 L.R.A. 193], we affirmed a judgment for damages for a plaintiff who alleged physical injury resulting from mental suffering, saying: “It is a matter of general knowledge that an attack of sudden fright or an exposure to imminent peril has produced in individuals a complete change in their nervous system, and rendered one who was physically strong and vigorous weak and timid.” Since no one can seriously question that fear or grief for one’s child is as likely to cause physical injury as concern over one’s own well-being, rejection of the fraudulent claims contention in Sloane clearly applies here.

In the second instance, and more fundamentally, the possibility that fraudulent assertions may prompt recovery in isolated cases does not justify a wholesale rejection of the entire class of claims in which that potentiality arises. The “contention that the rule permitting the maintenance of the action would be impractical to administer ... is but an argument that the courts are incapable of performing their appointed tasks, a premise which has frequently been rejected.” (Emden v. Vitz (1948) 88 Cal.App.2d 313, 319 [198 P.2d 696].) “ [F]ear that unfounded claims may be put forward, and may result in erroneous conclusions of fact, ought not to influence us to impose legal limitations as to the nature of the facts that it is permissible to prove.” (Owens v. Liverpool Corp. (1939) 1 K.B. 394, 400.) “Certainly it is a very questionable position for a court to take, that because of the possibility of encouraging fictitious claims compensation should be denied those who have actually suffered serious injury through the negligence of another.” (Orlo v. Connecticut Co. (1941) 128 Conn. 231, 239 [21 A.2d 402], See also Goodhart, The Shock Cases and Area of Risk (1953) 16 Modern L.Rev. 14, 23; Throckmorton, Damages for Fright (1921) 34 Harv.L.Rev. 260, 276.)

On the analogous issue of whether the possibility of collusive fraud in intrafamily tort actions justified a per se rule denying recovery in all such cases, this court held that the *737interests of meritorious plaintiffs should prevail over alleged administrative difficulties. Upholding the claim of a minor child in that situation we said: “The interest of the child in freedom from personal injury caused by the tortious conduct of others is sufficient to outweigh any danger of fraud or collusion. . . . [T]he fact that there may be greater opportunity for fraud or collusion in one class of cases than another does not warrant courts of law in closing the door to all cases of that class. Courts must depend upon the efficacy of the judicial processes to ferret out the meritorious from the fraudulent in particular eases.” (Emery v. Emery (1955) 45 Cal.2d 421, 431 [289 P.2d 218]; see also Klein v. Klein, (1962) 58 Cal.2d 692, 695-696 [26 Cal.Rptr. 102, 376 P.2d 70].)

The possibility that some fraud will escape detection does not justify an abdication of the judicial responsibility to award damages for sound claims: if it is “to be conceded that our procedural system for the ascertainment of truth is inadequate to defeat fraudulent claims . . . , the result is a virtual acknowledgment that the courts are unable to render justice in respect to them.” (Chiuchiolo v. New England Wholesale Tailors (1930) 84 N.H. 329, 335 [150 A. 540].)

Indubitably juries and trial courts, constantly called upon to distinguish the frivolous from the substantial and the fraudulent from the meritorious, reach some erroneous results. But such fallibility, inherent in the judicial process, offers no reason for substituting for the case-by-case resolution of causes an artificial and indefensible barrier. Courts not only compromise their basic responsibility to decide the merits of each case individually but destroy the public’s confidence in them by using the broad broom of “administrative convenience” to sweep away a class of claims a number of which are admittedly meritorious. The mere assertion that fraud is possible, “a possibility [that] exists to some degree in all cases” (Klein v. Klein, supra, 58 Cal.2d 692, 695), does not prove a present necessity to abandon the neutral principles of foreseeability, proximate cause and consequential injury that generally govern tort law.

Indeed, we doubt that the problem of the fraudulent claim is substantially more pronounced in the ease of a mother claiming physical injury resulting froln seeing her child killed than in other areas of tort law in which the right to recover damages is well established in California. For exam-*738pie, a plaintiff claiming that fear for his own safety resulted in physical injury makes out a well recognized case for recovery.4 (Lindley v. Knowlton (1918) 179 Cal. 298 [176 P. 440] ; Webb v. Francis J. Lewald Coal Co. (1931) 214 Cal. 182 [4 P.2d 532, 77 A.L.R. 675] ; Vanoni v. Western Airlines (1967) 247 Cal.App.2d 793 [56 Cal.Rptr. 115].) Moreover, damages are allowed for “mental suffering,” a type of injury, on the whole, less amenable to objective proof than the physical injury involved here; the mental injury can be in aggravation of, or “parasitic to,” an established tort. (Sloane v. Southern California Ry. Co., supra, 111 Cal. 668; Acadia, California, Ltd. v. Herbert (1960) 54 Cal.2d 328 [5 Cal.Rptr. 686, 353 P.2d 294] ; Easton v. United Trade School Contracting Co. (1916) 173 Cal. 199 [159 P. 597, L.R.A. 1916A 394].) In fact, fear for another, even in the absence of resulting physical injury, can be part of these parasitic damages. (Acadia, California, Ltd. v. Herbert, supra, 54 Cal.2d 328, 337; Easton v. United Trade School Contracting Co., supra, 173 Cal. 199, 202.) And emotional distress, if inflicted intentionally, constitutes an independent tort. (State Rubbish Collectors Assn. v. Siliznoff (1952) 38 Cal.2d 330, 338 [240 P.2d 282].) The danger of plaintiffs’ fraudulent collection of damages for nonexistent injury is at least as great in these examples as in the instant case.

In sum, the application of tort law- can never be a matter of mathematical precision. In terms of characterizing conduct as tortious and matching a money award to the injury suffered as well as in fixing the extent of injury, the process cannot be perfect. Undoubtedly, ever since the ancient case of the tavern-keeper ’s wife who successfully avoided the hatchet cast by an *739irate customer (I de S et ux v. W de S, Y.B. 22 Edw. iii, f. 99, pl. 60 (1348)), defendants have argued that plaintiffs’ claims of injury from emotional trauma might well be fraudulent. Yet we cannot let the difficulties of adjudication frustrate the principle that there be a remedy for every substantial wrong.

2. The alleged inability to fix definitions for recovery on the different facts of future cases does not justify the denial of recovery on the specific facts of the instant case; in any event, proper guidelines can indicate the extent of liability for such future cases.

In order to limit the otherwise potentially infinite liability which would follow every negligent act, the law of torts holds defendant amenable only for injuries to others which to defendant at the time were reasonably foreseeable.

In the absence of “overriding policy considerations . . . foreseeability of risk [is] of . . . primary importance in establishing the element of duty.” (Grafton v. Mollica (1965) 231 Cal.App.2d 860, 865 [42 Cal.Rptr. 306]. See also McEvoy v. American Pool Corp. (1948) 32 Cal.2d 295 [195 P.2d 783] ; Hergenrether v. East (1964) 61 Cal.2d 440 [39 Cal.Rptr. 4, 393 P.2d 164].) As a classic opinion states: “The risk reasonably to be perceived defines the duty to be obeyed.” (Palsgraf v. Long Island R.R. Co. (1928) 248 N.Y. 339, 344 [162 N.E. 99, 59 A.L.R. 253].) Defendant owes a duty, in the sense of a potential liability for damages, only with respect to those risks or hazards whose likelihood made the conduct unreasonably dangerous, and hence negligent, in the first instance. (See Keeton, Legal Cause in the Law of Torts (1963) 18-20; Seavey, Mr. Justice Cardozo and the Law of Torts (1939) 52 Harv.L.Rev. 372; Seavey, Principles of Torts (1942) 56 Harv.L.Rev. 72.)

Harper and James state the prevailing view. The obligation turns on whether “the offending conduct foreseeably involved unreasonably great risk of harm to the interests of someone other than the actor. . . . [T]he obligation to refrain from . . . particular conduct is owed only to those who are foresee-ably endangered by the conduct and only with respect to those risks or hazards whose likelihood made the conduct unreasonably dangerous. Duty, in other words, is measured by the scope of the risk which negligent conduct foreseeably entails.” (2 Harper & James, The Law of Torts, supra, at p. 1018; fns. omitted.)

This foreseeable risk may be of two types. The first class involves actual physical impact. A second type of risk *740applies to the instant situation. “In other cases, however, plaintiff is outside the zone of physical risk (or there is no risk of physical impact at all), but bodily injury or sickness is brought on by emotional disturbance which in turn is caused by defendant’s conduct. Under general principles recovery should be had in such a ease if defendant should foresee fright or shock severe enough to cause substantial injury in a person normally constituted. Plaintiff would then be within the zone of risk in very much the same way as are plaintiffs to whom danger is extended by acts of third persons, or forces of nature, or their own responses (where these things are foreseeable).” (2 Harper & James, The Law of Torts, supra, at pp. 1035-1036; fns. omitted.)5

Since the chief element in determining whether defendant owes a duty or an obligation to plaintiff is the foreseeability of the risk, that factor will be of prime concern in every case. Because it is inherently intertwined with foreseeability such duty or obligation must necessarily be adjudicated only upon a case-by-case basis. We cannot now predetermine defendant’s obligation in every situation by a fixed category; no immutable rule can establish the extent of that obligation for every circumstance of the future. We can, however, define guidelines which will aid in the resolution of such an issue as the instant one.

We note, first, that we deal here with a case in which plaintiff suffered a shock which resulted in physical injury and we confine our ruling to that case. In determining, in such a case, whether defendant should reasonably foresee the injury to plaintiff, or, in other terminology, whether defendant owes plaintiff a duty of due care, the courts will take into account such factors as the following: (1) Whether plaintiff was located near the scene of the accident as contrasted with one who was a distance away from it. (2) Whether the shock resulted from a direct emotional impact upon plaintiff from the sensory and contemporaneous observance of the accident, *741as contrasted with learning of the accident from others after its occurrence. (3) Whether plaintiff and the victim were closely related, as contrasted with an absence of any relationship or the presence of only a distant relationship.

The evaluation of these factors will indicate the degree of the defendant’s foreseeability: obviously defendant is more likely to foresee that a mother who observes an accident affecting her child will suffer harm than to foretell that a stranger witness will do so. Similarly, the degree of foreseeability of the third person’s injury is far greater in the case of his contemporaneous observance of the accident than that in which he subsequently learns of it. The defendant is more likely to foresee that shock to the nearby, witnessing mother will cause physical harm than to anticipate that someone distant from the accident will suffer more than a temporary emotional reaction. All these elements, of course, shade into each other-; the fixing of obligation, intimately tied into the facts, depends upon each ease.

In light of these factors the court will determine whether the accident and harm was reasonably foreseeable. Such reasonable foreseeability does not turn on whether the particular plaintiff as an individual would have in actuality foreseen the exact accident and loss; it contemplates that courts, on a case-to-case basis, analyzing all the circumstances, will decide what the ordinary man under such circumstances should reasonably have foreseen. The courts thus mark out the areas of liability, excluding the remote and unexpected.

In the instant ease, the presence of all the above factors indicates that plaintiff has alleged a sufficient prima facie case. Surely the negligent driver who causes the death of a young child may reasonably expect that the mother will not be far distant and will upon witnessing the accident suffer emotional trauma. As Dean Prosser has stated: “when a child is endangered, it is not beyond contemplation that its mother will be somewhere in the vicinity, and will suffer serious shock.” (Prosser, The Law of Torts, supra, at p. 353. See also 2 Harper & James, The Law of Torts, supra, at p. 1039.)

We are not uoav called upon to decide whether, in-the absence or reduced weight of some of the above factors, we would conclude that the accident and injury were not reasonably foreseeable and that therefore defendant owed no duty of due care to plaintiff. In future cases the courts will draw lines of demarcation upon facts more subtle than the compelling ones alleged in the complaint before us.

*742The courts have in the past, in analogous situations, drawn, the limits of liability, applying general guidelines such as those above set forth to the specific facts of the cases. As examples of that process of definition we set forth the history of the “open car” cases, the rulings on recovery by persons not in privity of contract for defendant’s negligence in drafting instruments, the decisions on the intentional infliction of emotional injury, the modern English cases, and some illustrative opinions that adjudicate the specific issue before us.

The ability of courts to limit liability predicated on tests largely based upon foreseeability is well illustrated by the “open car” cases. The prototype case is the suit against the owner of a vehicle for damage caused plaintiff by a third party who can commandeer the vehicle because of the owner’s carelessness in leaving the keys inside. In Richardson v. Ham (1955) 44 Cal.2d 772 [285 P.2d 269], we posited liability on the owner of a bulldozer because of a “foreseeable risk of intermeddling” (p. 776), noting especially the great danger the bulldozer created and the special temptation it presented to third parties. Similarly, in Hergenrether v. East, supra, 61 Cal.2d 440, we upheld such liability of a truck owner on the basis of “greater potentiality of foreseeable risk” (p. 444) because of the possible danger of the vehicle, the time for which it was unattended, and the type of persons who frequent the neighborhood in which it was left.

These decisions have not led to untrammeled liability. Eather, applying the foreseeability test, the courts have held that the mere act of leaving a key in an automobile, although it may possibly raise a foreseeable risk that the car will be stolen, does not increase the risk of injury to other property and hence does not warrant liability: “[e]ven if she could have foreseen the theft, she had no reason to believe that the thief would be an incompetent driver.” (Richards v. Stanley (1954) 43 Cal.2d 60, 66 [271 P.2d 23].) In short, “each ease must be considered on its own facts to determine whether the [situation] in toto justifies the conclusion that the foreseeable risk of harm imposed is unreasonable, and that the defendant owner or one in charge of a vehicle has a duty to third persons in the class of the plaintiffs to refrain from subjecting them to such risk.” (Hergenrether v. East, supra, 61 Cal.2d 440, 445; see also England v. Mapes Produce Co. (1965) 238 Cal.App.2d 120 [47 Cal.Rptr. 506] ; Murray v. Wright (1958) 166 Cal.App.2d 589 [333 P.2d 111].)

In another category of cases, those involving the liability of a tortfeasor to a third person with whom he was not in privity *743of contract for negligent draftmanship of a legal document, we have recognized the right of the injured party to compensation and set out guidelines for the determination of future eases. In Lucas v. Hamm (1961) 56 Cal.2d 583, 588 [15 Cal. Rptr. 821, 364 P.2d 685], we applied this rule to an attorney who drew a defective will, thereby causing damage to the intended third-party beneficiary. (See also Biakanja v. Irving (1958) 49 Cal.2d 647, 650 [320 P.2d 16, 65 A.L.R.2d 1358].)

In sanctioning recovery for injury caused by intentional infliction of mental distress, this court did not defer to the argument that liability should not be imposed because of the possible future difficulty in delimiting the area of liability. Defendants urged that if recovery were to be allowed for intentional infliction of emotional distress, actions would soon be forthcoming based upon every minor personal insult or indignity. We said: “That administrative difficulties do not justify the denial of relief for serious invasions of mental and emotional tranquility is demonstrated by the cases recognizing the right of privacy." (State Rubbish Collectors Assn. v. Siliznoff, supra, 38 Cal.2d 330, 338.) Wc rejected the contention “that to allow recovery in the absence of physical injury will open the door to unfounded claims and a flood of litigation, and that the requirement that there be physical injury is necessary to insure that serious mental suffering actually occurred" (State Rubbish Collectors Assn. v. Siliznoff, supra, 38 Cal.2d 330, 338).

Indeed, the argument that “there is no point at which such actions would stop" is no more plausible today than when it was advanced in Winterbottom v. Wright (1842) 10 M. & W. 109, 111. History has exposed the fallacy of the claim that abolition of privity in enterprise liability cases would lead to “the most absurd and outrageous consequences, to which I can see no limit" (p. 114). In taking another giant step forward, in imposing product liability in tort, we were not halted by the spectre of an inability to pre-judge every future ease. The setting of boundaries upon that doctrine makes the problem of fixing lines of Emitation here appear, by comparison, almost miniscule. The widening of the area of liabiEty and the possibility of the encouragement of unfounded and undefinable claims in the products liability field was sweeping; here we deal with a comparatively isolated and unusual situation. We do not believe that the fear that we cannot successfully adjudicate future cases of this sort, pursuant to the *744suggested guidelines, should bar recovery in an otherwise meritorious cause.

The fear of an inability to fix boundaries has not impelled the courts of England to deny recovery for emotional trauma caused by witnessing the death or injury of another due to defendant’s negligence. We set forth the holdings of some English cases merely to demonstrate that courts can formulate and apply such limitations of liability.

The first and classic case, Hambrook v. Stokes Bros., supra, 1 K.B. 141, rejected the argument that recovery should be denied because of possible administrative difficulty. In Ham-brook the defendant’s servant left a truck parked at the top of a steep and narrow street with the engine running. The deceased, a pregnant woman, had walked with her children on their way to school to the point where they turned onto the street where the truck was parked. Because the driver did not take proper precautions, the truck started itself down the hill and struck one of the children. Although she herself was never in danger, the mother saw the runaway truck and feared greatly for the safety of her children. Upon inquiry she found that one of the children had been seriously injured; several months later both the mother and the foetus were dead. The trial court directed the jury that the father’s suit for loss of services could succeed only if the death were caused by the mother’s fear for her own safety, but the appellate court held that the plaintiff could recover even if the fear for the children brought about her demise.

Paced with the contention that their holding would increase the number of suits and foment possible fraudulent claims, Lord Justice Atkin quoted this passage: “ ‘I should be sorry to adopt a rule which would bar all such claims on grounds of policy alone, and in order to prevent the possible success of unrighteous or groundless actions. Such a course involves the denial of redress in meritorious cases, and it necessarily implies a certain degree of distrust, which I do not share, in the capacity of legal tribunals to get at the truth in this class of claim. ’ ” (Hambrook v. Stokes Bros., supra, quoting from Dulieu v. White and Sons [1901] 2 K.B. 669, 681, opinion by Kennedy, J.)

In a recent application of the Hambrook rule, an English court permitted recovery by a widow of a man who developed severe psychoneurotic symptoms as a result of harrowing experiences, not involving his personal safety, while serving as a rescuer at a gruesome train wreck. The court stated that *745the “ ‘test of liability for' shock is foreseeability of injury by shock. ’ ” (Chadwick v. British Railways Board [1967] 1 W.L.R. 912, 920, quoting from King v. Phillips [1953] 1 Q.B. 429, 441, opinion by Denning, L.J.)

Professor John Fleming of the School of Law, Boalt Hall, University of California, in a careful analysis of the development of English law on this subject, first explains, “It is evident, of course, that, to the extent of denying redress for certain kinds of negligently inflicted harm, the law is in effect withholding its protective mantle from corresponding human interests that may accordingly be infringed with impunity. To refuse a remedy for nervous shock is the equivalent of refusing to accede to an individual’s claim for safeguarding his emotional security. It is also the same as saying that there is no ‘duty’ owed to exercise reasonable care to avoid inflicting this type of loss or injury. Although no longer quite as fashionable in this particular context, the same idea can also, finally, be expressed by asserting that such damage is ‘too remote’ or, what amounts to the same thing, that the defendant’s negligence was not its ‘proximate cause’.” (Fleming, An Introduction to the Law of Torts (1967) p. 46.) ■

After explaining that certain English cases manipulated doctrinal approaches “to subserve ulterior purposes” in granting recovery in some situations and denying it in others, Fleming states that “a long-delayed change in attitude may perhaps be discerned in the latest decision by the Court of Appeal [Boardman v. Sanderson (1964) 1 W.L.R. 1317 (C.A.)], which sustained a father’s claim for a mental shock he suffered upon hearing the screams of his boy when the latter’s foot was negligently caught under the wheel of the defendant’s ear from which father and son had just alighted inside a service garage. Neither did the father fear for his own safety nor did he so much as even see the accident. Indeed, the claimant was not even a female—the prototype plaintiff in these cases being almost exclusively concerned with pregnancy injuries. Yet the court considered it sufficient to say that a duty was owed not only to the boy but also to his near relatives, who, to the defendant’s knowledge, were on the premises within earshot and likely to come upon the scene if any injury befell him. It remains to be seen whether this relaxation, slight as it may be, might not eventually be extended to relatives’ whose presence, though not actually known, was yet foreseeable in accordance with the prevailing *746test customarily applied to .claims for. physical injuries.” (Italics in original; fn. omitted.) (Fleming, An Introduction to the Law of Torts, supra, at p. 54.)

The English courts have likewise marked out areas of liability, excluding those injuries that are remote and unexpected. Thus a distinguished English court has held that the physical injury of a casual bystander resulting from shock or fright upon witnessing an accident would present so unusual and hence unforeseeable an event as to warrant a directed verdict for defendant. “The driver of a car or vehicle, even though careless, is entitled to assume that the ordinary frequenter of the streets has sufficient fortitude to endure such incidents as may from time to time be expected to occur in them, including the noise of a collision and the sight of injuries to others, and is not to be considered negligent towards one who does not possess the customary phlegm.” (Italics added.) (Bourhill v. Young (1943) A.C. 92, 117 (Lord Porter); see, id. at pp. 98 (Lord Thankerton), 101 (Lord Bussell), 104 (Lord MacMillan), and 107 (Lord Wright) ; King v. Phillips, supra, 1 Q.B. 429, 442.)

Thus we see no good reason why the general rules of tort law, including the concepts of negligence, proximate cause, and foreseeability, long applied to all other types of injury, should not govern the case now before us. Any questions that the cause raises “will be solved most justly by applying general principles of duty and negligence, and . . . mechanical rules of thumb which are at variance with these principles do more harm than good.” (2 Harper & James, The Law of Torts, supra, p. 1039; fn. omitted.) “The refusal to apply these general rules to actions for this particular kind of physical injury is nothing short of a denial of justice.” (Throckmorton, Damages for Fright, supra, 34 Harv.L.Rev. 260, 277; fn. omitted.)

In short, the history of the eases does not show the development of a logical rule but rather a series of changes and abandonments. Upon the argument in each situation that the courts draw a Maginot Line to withstand an onslaught of false claims, the cases have assumed a variety of postures. At first they insisted that there be no recovery for emotional trauma at all. (Amaya v. Home Ice, Fuel & Supply Co., supra, 59 Cal.2d 295, dissenting opinion by Peters, J., p. 328 fn. 9.) Retreating from this position, they gave relief for such trauma only if physical impact occurred. (Id. at p. 325 fn. 4.) They then abandoned the requirement for physical impact but *747insisted that the victim fear for her own safety (Amaya v. Home Ice, Fuel & Supply Co., supra, 59 Cal.2d 295), holding that a mother could recover for fear for her children’s safety if she simultaneously entertained a personal fear for herself. (Lindley v. Knowlton, supra, 179 Cal. 298.)6 They stated that the mother need only be in the “zone of danger” (Reed v. Moore (1957) 156 Cal.App.2d 43, 47 [319 P.2d 80]). The final anomaly would be the instant case in which the sister, who observed the accident, would be granted recovery because she was in the “zone of danger,” but the mother, not far distant, would be barred from recovery.

The successive abandonment of these positions exposes the weakness of artificial abstractions which bar recovery contrary to the general rules. As the commentators have suggested, the problem should be solved by the application of the principles of tort, not by the creation of exceptions to them. Legal history shows that artificial islands of exceptions, created from the fear that the legal process will not work, usually do not withstand the waves of reality and, in time, descend into oblivion.

We have explained that recovery here will not expose the courts to false claims or a flood of litigation. The test that we have set forth will aid in the proper resolution of future eases. Indeed, the general principles of tort law are acknowledged . to work successfully in all other cases of emotional trauma.

Yet for some artificial reason this delimitation of liability is alleged to be unworkable in the most egregious case of them all: the mother’s emotional trauma at the witnessed death of her child. If we stop at this point, however, we must necessarily question and reject not merely recovery here, but the viability of the judicial process for ascertaining liability *748for tortious conduct itself. To the extent that it is inconsistent with our ruling here, we therefore overrule Amaya v. Home Ice Fuel & Supply Co., supra, 59 Cal.2d 295.

To deny recovery would be to chain this state to an outmoded rule of the 19th century which can claim no current credence. No good reason compels our captivity to an indefensible orthodoxy.

The judgment is reversed.

Peters, J., Most, J., and Sullivan, J., concurred.

TRAYNOR, C. J.

I dissent for the reasons set forth in Amaya v. Some Ice, Fuel & Supply Co. (1963) 59 Cal.2d 295, 297-315 [29 Cal.Rptr. 33, 379 P.2d 513], In my opinion that case was correctly decided and should not be overruled.

BURKE, J.

As recently as 1963 this court, in Amaya v. Home Ice, Fuel & Supply Co., 59 Cal.2d 295 [29 Cal.Rptr. 33, 379 P.2d 513], thoroughly studied and expressly rejected the proposition (pp. 298-299) that tort liability may be predicated on fright or nervous shock (with consequent bodily illness) induced solely by the plaintiff’s apprehension of negligently caused danger or injury to a third person. As related in our Amaya opinion, plaintiff there was the mother of a 17-month-old boy who saw him struck by a truck; accordingly our ruling necessarily included all mothers of small children who observe them being injured. Yet today this court’s Amaya decision is overruled by an opinion which disdains any discussion whatever of the history and policy of pertinent law painstakingly set forth in Amaya.

Every one of the arguments advanced in today’s opinion was considered by this court and rejected, expressly or by fair implication, in Amaya. 1 Further, as Amaya points out (p. 304 of 59 Cal.2d), in every jurisdiction in this country that had ruled on the point at issue the decisions up to that time (1963) were unanimous in upholding the rule of nonliability.

So far as has been discovered, in not a single such jurisdiction has an appellate court ruled to the contrary since *749 Amaya. 2 But the majority make "no attempt in today’s opinion—as apparently they could not—to buttress their result with citations of cases based on American law, to say nothing of that of California. Instead, we are offered two English cases applying the 1925 Hambrook case (Hambrook v. Stokes Bros. [1925] 1 K.B. 141), whose ruling we expressly rejected in Amaya (pp. 303-304 [fn. 4], and 313, of 59 Cal.2d), and which, as already stated has not been followed or approved by any jurisdiction in this country.

The majority, obviously recognizing that they are now embarking upon a first excursion into the “fantastic realm of infinite liability” (Amaya, at p. 315 of 59 Cal.2d), undertake to provide so-called “guidelines” for the future. But notwithstanding the limitations which these “guidelines” purport to impose, it is only reasonable to expect pressure upon our trial courts to make their future rulings conform to the spirit of the new elasticity proclaimed by the majority.

Moreover, the majority’s “guidelines” {ante, pp. 740-741) are simply a restatement of those suggested earlier by Professor Prosser (Prosser, Torts (2d ed., 1955) p. 182) they have already been discussed and expressly rejected by this court in Amaya (pp. 312-313). Upon analysis, their seeming certainty evaporates into arbitrariness, and inexplicable distinctions appear.3 As we asked in Amaya-. What if the plaintiff was honestly mistaken in believing the third person to be in danger or to be seriously injured? What if the third person had assumed the risk involved? How “close” must the relationship be between the plaintiff and the third person? I.e., what if the third person was the plaintiff’s beloved niece or nephew, grandparent, fiancé, or lifelong friend, more dear to the plaintiff than her immediate family? Next, how “near” *750must the plaintiff have been to the scene of the accident, and how “soon” must shock have been felt? Indeed," what is the magic in the plaintiff’s being actually present? Is the shock any less real if the mother does not know of the accident until her injured child is brought into her home? On the other hand, is it any less real if the mother is physically present at the scene but is nevertheless unaware of the danger or injury to her child until after the accident has occurred? No answers to these questions are to be found in today’s majority opinion. Our trial courts, however, will not so easily escape the burden of distinguishing between litigants on the basis of such artificial and unpredictable distinctions.

Further, and again contrary to the assertions of the majority (ante, pp. 732-733), no fallacy or incongruity appears in the rule permitting recovery to one within the physical zone of danger for trauma suffered from fear of impact, but denying it to a person outside that zone. The impact feared must be to oneself, and it must be an objective fear—not merely that of- an excessively imaginative or timid plaintiff. As pointed out in the leading ease of Waube v. Warrington (1935) 216 Wis. 603, 612-613 [258 N.W. 497], “It is one thing to say that as to those who are put in peril of physical impact, impact is immaterial if physical injury is caused by shock arising from the peril. It is the foundation of cases holding to this liberal ruling, that the person affrighted or sustaining shock was actually put in peril of physical impact, and under these conditions it was considered immaterial that the physical impact did not materialize. It is quite another thing to say that those who are out of the field of physical danger through impact shall have a legally protected right to be free from emotional distress occasioned by the peril of others, when that distress results in physical impairment.” (Italics added.) Thus, California’s rule that a plaintiff’s reasonable fear for his own safety is compensable presents neither an argument for the same rule as to fear for others, nor a danger of recovery based on the plaintiff’s false claims of fear for himself.4

*751The assertion of the majority (ante, p. 735) that “The denial of ‘duty’ in the instant situation [i.e., physical impairment resulting from emotional distress occasioned by apprehension of the peril of others] rests upon the prime hypothesis that allowance of such an action would lead to successful assertion of fraudulent claims,’’ (italics added) is controverted by the very ease cited in support. (Waube v. Warrington, supra, 216 Wis. 603, 613.) Instead of reliance on any such “prime hypothesis,” the Wisconsin court had this to say in Waube: “The answer to this question cannot be reached solely by logic, nor is it clear that it can be entirely disposed of by a consideration of what the defendant ought reasonably to have anticipated as a consequence of his wrong. The answer must be reached by balancing the social interests involved in order to ascertain how far defendant’s duty and plaintiff’s right may justly and expediently be extended. It is our conclusion that they can neither justly nor expediently be extended to any recovery for physical injuries sustained by one out of the range of ordinary physical peril as a result of the shock of witnessing another’s danger. Such consequences are so unusual and extraordinary, viewed after the event, that a user of the highway may be said not to subject others to an unreasonable risk of them by the careless management of his vehicle. Furthermore, the liability imposed by such a doctrine is wholly out of proportion to the culpability of the negligent tort-feasor, would put an unreasonable burden upon users of the highway, open the way to fraudulent claims, and enter a field that has no sensible or just stopping point.”

As this court declared in Amaya (p. 315 of 59 Cal.2d), there is good sense in the conclusion of the court in Waube that ‘1 the liability imposed by such a doctrine is wholly out of proportion to the culpability of the negligent tort-feasor”; further, to permit recovery by every person who might adversely feel some lingering effect of the defendant’s conduct would throw us into “the fantastic realm of infinite liability.” Yet the majority opinion in the present case simply omits to either mention or discuss the injustice to California defendants flowing from such a disproportionate extension of their liability—an injustice which plainly constituted a “prime hypothesis” for rejection of the liability sought to *752be imposed by the plaintiffs in Waube and in Amaya. (See also Jelley v. LaFlame (N.H. 1968) supra, 238 A.2d 728, 730, citing with approval and following this ground of decision expressed in Watibe and in Amaya.)

Additionally, the majority fail to explain their bare assertion {ante, p. 733) that contributory negligence of Erin will defeat any recovery by plaintiff mother and sister.5 The familiar and heretofore unquestioned principle is that the relationships of parent and child or of husband and wife in themselves furnish no basis for imputation of contributory negligence. (Within, Summary of Cal. Law (1960) Torts, § 341, p. 1542; Rest.2d Torts, § 488.) Is this principle now abrogated in California? If so, it is a ruling extending far beyond the confines of the particular issue now before us, and reaches potentially every negligence action in which the plaintiffs are members of the same family.

It appears to me that in the light of today’s majority opinion the matter at issue should be commended to the attention of the Legislature of this state. Five years have elapsed since our Amaya decision, during which that body has not undertaken to change the law we there declared. We may presume, therefore, that the limitations upon liability there affirmed comport with legislative views. But if all alleged California tortfeasors, including motorists, home and other property owners, and governmental entities, are now to be faced with the concept of potentially infinite liability beyond any rational relationship to their culpability, then surely the point has been reached at which the Legislature should reconsider the entire subject and allow all interests affected to be heard.

I would affirm the judgment.

McComb, J., concurred.

Opinion

EAGLESON, J.

The narrow issue presented by the parties in this case is whether the Court of Appeal correctly held that a mother who did not *647witness an accident in which an automobile struck and injured her child may recover damages from the negligent driver for the emotional distress she suffered when she arrived at the accident scene. The more important question this issue poses for the court, however, is whether the “guidelines” enunciated by this court in Dillon v. Legg (1968) 68 Cal.2d 728 [69 Cal.Rptr. 72, 441 P.2d 912, 29 A.L.R.3d 1316] are adequate, or if they should be refined to create greater certainty in this area of the law.

Although terms of convenience identify the cause of action here as one for negligent infliction of emotional distress (NIED) and the plaintiff as a “bystander”1 rather than a “direct victim,” the common law tort giving rise to plaintiff’s claim is negligence. (Dillon v. Legg, supra, 68 Cal.2d 728, 730; Amaya v. Home Ice, Fuel & Supply Co. (1963) 59 Cal.2d 295, 314-315 [29 Cal.Rptr. 33, 379 P.2d 513].) It is in that context that we consider the appropriate application of the concept of “duty” in an area that has long divided this court—recognition of the right of persons, whose only injury is emotional distress, to recover damages when that distress is caused by knowledge of the injury to a third person caused by the defendant’s negligence. Although we again find ourselves divided, we shall resolve some of the uncertainty over the parameters of the NIED action, uncertainty that has troubled lower courts, litigants, and, of course, insurers.

Upon doing so, we shall conclude that the societal benefits of certainty in the law, as well as traditional concepts of tort law, dictate limitation of bystander recovery of damages for emotional distress. In the absence of physical injury or impact to the plaintiff himself, damages for emotional distress should be recoverable only if the plaintiff: (1) is closely related to the injury victim, (2) is present at the scene of the injury-producing event at the time it occurs and is then aware that it is causing injury to the victim and, (3) as a result suffers emotional distress beyond that which would be anticipated in a disinterested witness.

I

Background

On December 8, 1980, John Thing, a minor, was injured when struck by an automobile operated by defendant James V. La Chusa. His mother, plaintiff Maria Thing, was nearby, but neither saw nor heard the accident. She became aware of the injury to her son when told by a daughter that John had been struck by a car. She rushed to the scene where she saw her bloody and unconscious child, who she believed was dead, lying in the *648roadway. Maria sued defendants, alleging that she suffered great emotional disturbance, shock, and injury to her nervous system as a result of these events, and that the injury to John and emotional distress she suffered were proximately caused by defendants’ negligence.

The trial court granted defendants’ motion for summary judgment, ruling that, as a matter of law, Maria could not establish a claim for negligent infliction of emotional distress because she did not contemporaneously and sensorily perceive the accident. Although prior decisions applying the guidelines suggested by this court in Dillon v. Legg, supra, 68 Cal.2d 728, compelled the ruling of the trial court, the Court of Appeal reversed the judgment dismissing Maria’s claim after considering the decision of this court in Ochoa v. Superior Court (1985) 39 Cal.3d 159 [216 Cal.Rptr. 661, 703 P.2d 1]. The Court of Appeal reasoned that while Maria’s argument, premised on Molien v. Kaiser Foundation Hospitals (1980) 27 Cal.3d 916 [167 Cal.Rptr. 831, 616 P.2d 813, 16 A.L.R.4th 518], that she was a direct victim of La Chusa’s negligence, did not afford a basis for recovery, contemporaneous awareness of a sudden occurrence causing injury to her child was not a prerequisite to recovery under Dillon.

We granted review to consider whether Ochoa supports the holding of the Court of Appeal. We here also further define and circumscribe the circumstances in which the right to such recovery exists. To do so it is once again necessary to return to basic principles of tort law.

II

Emotional Distress as a Compensable Item of Damage in Intentional Torts

Although the theory of recovery in issue here is the tort of “negligence,” recognition of emotional distress as a distinct item of damage for which recovery may be had even absent physical injury or impact is not limited to negligence actions. Indeed, recovery for intentional conduct that invades the individual’s right to peace of mind was recognized long before such recovery was permitted in negligence actions. It is useful, therefore, to place emotional distress as a basis for a negligence action in perspective by briefly reviewing the development of common law recognition of a protectible interest in individual peace of mind—i.e., the right to be free from socially unacceptable conduct that seriously affects another’s peace of mind.

The range of mental or emotional injury subsumed within the rubric “emotional distress” and for which damages are presently recoverable “includes fright, nervousness, grief, anxiety, worry, mortification, shock, *649humiliation and indignity, as well as physical pain.” (Deevy v. Tassi (1942) 21 Cal.2d 109, 120 [130 P.2d 389].)

Express or implicit recognition that peace of mind warrants legal protection is found in recovery for emotional distress as an aggravation of damages sought under intentional tort theories. Initially, emotional distress was recognized simply as an item of damages in those actions. With few exceptions, causing mental distress did not itself create a right of action, and where mental distress alone exists the common law rarely permitted recovery of damages. (5 Witkin, Summary of Cal. Law (9th ed. 1988) Torts, § 402, p. 483.)

Possibly the first exception to that limitation existed in recovery for assault. Assault is a tort which today recognizes the right of the individual to peace of mind, to live without fear of personal harm. “A civil action for assault is based upon an invasion of the right of a person to live without being put in fear of personal harm.” (Lowry v. Standard Oil Co. (1944) 63 Cal.App.2d 1, 7 [146 P.2d 57].) It has been noted, however, that actions based on intentional conduct were originally authorized not in recognition of or to redress a right to mental tranquility, but to afford an alternate dispute resolution mechanism. Legal action was preferable to redress on the field of honor. (See Pearson, Liability to Bystanders for Negligently Inflicted Emotional Harm—A Comment on the Nature of Arbitrary Rules (1982) 34 U.Fla.L.Rev. 477, 486.)

Emotional distress is also an accepted item of damage that may be recovered in actions for abuse of process (Spellens v. Spellens (1957) 49 Cal.2d 210, 233 [317 P.2d 613]); false imprisonment (Gill v. Epstein (1965) 62 Cal.2d 611, 618 [44 Cal.Rptr. 45, 401 P.2d 397]); libel (Scott v. Times-Mirror Co. (1919) 181 Cal. 345, 365 [184 P. 672, 12 A.L.R. 1007]); and invasion of privacy. (Melvin v. Reid (1931) 112 Cal.App. 285 [297 P. 91].) Here, too, recovery has not been limited to circumstances in which the mental distress is an aggravation of a physical injury or impact.

Recognition of emotional distress as a compensable injury when caused by an intentional tort carried with it a judgment that the defendant’s conduct was sufficiently outrageous or unacceptable that an award of damages was justified to punish the tortfeasor and deter such conduct by others. This development led in turn to a focus on the nature of the defendant’s conduct, rather than on identifying a traditional tort to justify recovery for infliction *650of emotional distress, and culminated in recognition of the tort now known as intentional infliction of emotional distress.2

With recognition of intentional infliction of emotional distress as a discrete tort cause of action, this court accepted both freedom from emotional distress as an interest worthy of protection in its own right, and the proposition that it is possible to quantify and compensate for the invasion of that interest through an award of monetary damages even when the severity of the emotional distress is not manifested in physical symptoms. “If a cause of action is otherwise established, it is settled that damages may be given for mental suffering naturally ensuing from the acts complained of [citations], and in the case of many torts, such as assault, battery, false imprisonment, and defamation, mental suffering will frequently constitute the principal element of damages. [Citation.] In cases where mental suffering constitutes a major element of damages it is anomalous to deny recovery because the defendant’s intentional misconduct fell short of producing some physical injury.” (State Rubbish etc. Assn. v. Siliznoff (1952) 38 Cal.2d 330, 338 [240 P.2d 282].)

In Siliznojf, the court rejected arguments that permitting recovery for emotional distress without proof of physical injury would invite fraudulent claims and create difficulties in proof that serious mental distress resulted from the tortious conduct. The court reasoned that the defendant’s conduct *651often afforded greater proof of a serious invasion of the victim’s mental tranquility than did the presence or absence of resulting physical symptoms. (Ibid.)

Ill

Limitations in Negligence Actions

A parallel line of negligence cases permitting recovery of damages for emotional distress had developed in California at the time Siliznoff, supra, 38 Cal.2d 330, was decided. Initially, however, in negligence cases the right to recover for emotional distress had been limited to circumstances in which the victim was himself injured and emotional distress was a “parasitic” item of damages, or if a plaintiff who had been in the “zone of danger” did not suffer injury from impact, but did suffer physical injury as a result of the emotional trauma. (See Webb v. Francis J. Lewald Coal Co. (1931) 214 Cal. 182, 184 [4 P.2d 532, 77 A.L.R. 675]; Lindley v. Knowlton (1918) 179 Cal. 298, 301-302 [176 P. 440].)

Where the conduct was negligent, emotional distress caused solely by fear for a third person’s safety or apprehension of injury to the third person, was first recognized as an injury for which damages could be sought in Dillon v. Legg, supra, 68 Cal.2d 728.

But shortly before Dillon, in Amaya v. Home Ice, Fuel & Supply Co., supra, 59 Cal. 2d 295, the court had declined the opportunity to broaden the right to recover for emotional distress. Amaya, after confirming that the “impact rule” making a contemporaneous physical impact a prerequisite to recovery for negligently induced fright or shock was not applicable in California, held damages could not be recovered by persons outside the zone of danger created by the defendant’s negligence even when that shock was reflected in physiological symptoms. The court quoted with approval the statement of the general rule of nonliability for nervous shock induced by fear for a third party applied by the Court of Appeal in Reed v. Moore (1957) 156 Cal.App.2d 43 [319 P.2d 80]: “ ‘As a general rule, no recovery is permitted for a mental or emotional disturbance, or for a bodily injury or illness resulting therefrom, in the absence of a contemporaneous bodily contact or independent cause of action, or an element of wilfulness, wantonness, or maliciousness, in cases in which there is no injury other than one to a third person, even though recovery would have been permitted had the wrong been directed against the plaintiff. The rule is frequently applied to mental or emotional disturbances caused by another’s danger, or sympathy for another’s suffering. It has been regarded as applicable to a mental or emotional disturbance resulting from an injury not only to a stranger, but *652also to a relative of the plaintiff, such as a child, sister, father, or spouse.’ ” (59 Cal.2d at pp. 302-303.)

The court explained the restriction on the right to recover damages for emotional distress in negligence actions on the ground that the defendant had not breached a legal duty to the plaintiff. The court concluded that existence of a duty could not be defined, or left to the jury to find, on the basis of whether the injury was foreseeable. Rather the existence and scope of the defendant’s duty in this context was one for the court.

Several factors led to that conclusion. First was the observation that there are circumstances in which although a foreseeable risk exists, there is no duty to avoid creation of that risk. Another was the then prevalent view of other courts and commentators that the type of harm—fright or nervous shock with consequent bodily illness induced solely by apprehension of danger to another person—was not reasonably foreseeable. Ultimately, however, the court weighed the interest of the plaintiff in freedom from invasion of mental tranquility against the costs involved in recognizing a duty and concluded that factors militating against recognition of a legal duty to the third party plaintiff predominated.

First among these policy considerations was efficient administration of justice. The court’s concern here was the possibility of fraud and the difficulty in resolving disputes among witnesses over the extent and severity of the injury where negligent conduct rather than intentional conduct allegedly produced the emotional distress. A second important administrative factor was concern that it would be impossible to limit the circumstances in which liability would exist for emotional distress caused by apprehension of danger or injury not to the plaintiff but to a third person. The court concluded, also, that socioeconomic and moral factors mandate that there be some limit to the liability of the negligent actor. (Amaya, supra, 59 Cal.2d at pp. 312-313.)

The concept of relative fault weighed equally with the administrative concerns. “As long as our system of compensation is based on the concept of fault, we must also weigh ‘the moral blame attached to the defendant’s conduct.’ [Citation omitted.] Here is felt the difference between the social importance of conduct that negligently causes harm and conduct that is intended to do so. It is often said that in the latter case the defendant will be held liable for a broader range of consequences because, as the consequences are intended, they are the more ‘foreseeable.’ But in many intentional tort cases the defendant has been held liable under this reasoning for consequences far beyond those which he actually intended. [Citation.] It follows that, once more, ‘foreseeability’ is not the real answer. Rather, the increased *653liability imposed on an intentional wrongdoer appears to reflect the psychological fact that solicitude for the interests of the actor weighs less in the balance as his moral guilt increases and the social utility of his conduct diminishes.” (59 Cal.2d at p. 315.)

The Amaya view was short lived, however. Only five years later, the decision was overruled in Dillon v. Legg, supra, 68 Cal.2d 728. In the ensuing 20 years, like the pebble cast into the pond, Dillon's progeny have created ever widening circles of liability. Post -Dillon decisions have now permitted plaintiffs who suffer emotional distress, but no resultant physical injury, and who were not at the scene of and thus did not witness the event that injured another, to recover damages on grounds that a duty was owed to them solely because it was foreseeable that they would suffer that distress on learning of injury to a close relative.

In Dillon itself, the issue was limited. The mother and sister of a deceased infant each sought damages for “great emotional disturbance and shock and injury to her nervous system” which had caused them great mental pain and suffering. Allegedly these injuries were caused by witnessing the defendant’s negligently operated vehicle collide with and roll over the infant as she lawfully crossed a street. The mother was not herself endangered by the defendant’s conduct. The sister may have been. The trial court had therefore granted the defendant’s motion for judgment on the pleadings as to the mother, but had denied it with respect to the sister of the decedent. Faced with the incongruous result demanded by the “zone of danger” rule which denied recovery for emotional distress and consequent physical injury unless the plaintiff himself had been threatened with injury, the court overruled Amaya.

Reexamining the concept of “duty” as applicable to the Dillon facts, the court now rejected the argument that the possibility of fraudulent claims justified denial of recovery, at least insofar as a mother who sees her child killed is concerned, as “no one can seriously question that fear or grief for one’s child is as likely to cause physical injury as concern over one’s own well-being.” (Dillon v. Legg, supra, 68 Cal.2d 728, 736.) The court held instead that the right to recover should be determined by application of “the neutral principles of foreseeability, proximate cause and consequential injury that generally govern tort law.” (Id., at p. 737.)

The difficulty in defining the limits on recovery anticipated by the Amaya court was rejected as a basis for denying recovery, but the court did recognize that “to limit the otherwise potentially infinite liability which would follow every negligent act, the law of torts holds defendant amenable only for injuries to others which to defendant at the time were reasonably *654foreseeable.” (Dillon, supra, 68 Cal.2d at p. 739.) Thus, while the court indicated that foreseeability of the injury was to be the primary consideration in finding duty, it simultaneously recognized that policy considerations mandated that infinite liability be avoided by restrictions that would somehow narrow the class of potential plaintiffs. But the test limiting liability was itself amorphous.3

In adopting foreseeability of the injury as the basis of a negligent actor’s duty, the Dillon court identified the risks that could give rise to that duty as both physical impact and emotional disturbance brought on by the conduct. Having done so, the Dillon court conceded: “We cannot now predetermine defendant’s obligation in every situation by a fixed category; no immutable rule can establish the extent of that obligation for every circumstance of the future.” (68 Cal.2d at p. 740.) In an effort to give some initial definition to this newly approved expansion of the cause of action for NIED the court enunciated “guidelines” that suggested a limitation on the action to circumstances like those in the case before it.

“We note, first, that we deal here with a case in which plaintiff suffered a shock which resulted in physical injury and we confine our ruling to that case. In determining, in such a case, whether defendant should reasonably foresee the injury to plaintiff [mother], or in other terminology, whether defendant owes plaintiff a duty of due care, the courts will take into account such factors as the following: (1) Whether plaintiff was located near the scene of the accident as contrasted with one who was a distance away from *655it. (2) Whether the shock resulted from a direct emotional impact upon plaintiff from the sensory and contemporaneous observance of the accident, as contrasted with learning of the accident from others after its occurrence. (3) Whether plaintiff and the victim were closely related, as contrasted with an absence of any relationship or the presence of only a distant relationship.

“The evaluation of these factors will indicate the degree of the defendant’s foreseeability; obviously defendant is more likely to foresee that a mother who observes an accident affecting her child will suffer harm than to foretell that a stranger witness will do so. Similarly, the degree of foreseeability of the third person’s injury is far greater in the case of his contemporaneous observance of the accident than that in which he subsequently learns of it. The defendant is more likely to foresee that shock to the nearby, witnessing mother will cause physical harm than to anticipate that someone distant from the accident will suffer more than a temporary emotional reaction. All of these elements, of course, shade into each other; the fixing of the obligation, intimately tied into the facts, depends upon each case.

“In light of these factors the court will determine whether the accident and harm was reasonably foreseeable. Such reasonable foreseeability does not turn on whether the particular [defendant] as an individual would have in actuality foreseen the exact accident and loss; it contemplates that courts, on a case-to-case basis, analyzing all the circumstances, will decide what the ordinary man under such circumstances should reasonably have foreseen. The courts thus mark out the areas of liability, excluding the remote and unexpected.” (Dillon, supra, 68 Cal.2d at p. 741. Italics added.)

The Dillon court anticipated and accepted uncertainty in the short term in application of its holding, but was confident that the boundaries of this NIED action could be drawn in future cases. In sum, as former Justice Potter Stewart once suggested with reference to that undefinable category of materials that are obscene, the Dillon court was satisfied that trial and appellate courts would be able to determine the existence of a duty because the court would know it when it saw it.4 Underscoring the questionable validity of that assumption, however, was the obvious and unaddressed problem that the injured party, the negligent tortfeasor, their insurers, and their attorneys had no means short of suit by which to determine if a duty such as to impose liability for damages would be found in cases other than those that were “on all fours” with Dillon. Thus, the only thing that was foreseeable from the Dillon decision was the uncertainty that continues to this time as to the parameters of the third party NIED action.

*656IV

Post-Dillon Extension

The expectation of the Dillon majority that the parameters of the tort would be further defined in future cases has not been fulfilled. Instead, subsequent decisions of the Courts of Appeal and this court, have created more uncertainty. And, just as the “zone of danger” limitation was abandoned in Dillon as an arbitrary restriction on recovery, the Dillon guidelines have been relaxed on grounds that they, too, created arbitrary limitations on recovery. Little consideration has been given in post -Dillon decisions to the importance of avoiding the limitless exposure to liability that the pure foreseeability test of “duty” would create and towards which these decisions have moved.

Several post-Dillon decisions of this court are particularly noteworthy in this expansive progression. In the first, Krouse v. Graham (1977) 19 Cal.3d 59 [137 Cal.Rptr. 863, 562 P.2d 1022], this court held that the NIED plaintiff need not “visually” perceive the third party injury to satisfy the Dillon guideline suggesting that the plaintiff suffer shock from “ ‘the sensory and contemporaneous observance of the accident, . . .’ ” It was sufficient that the plaintiff knew the position of his wife just outside the automobile in which he was seated the instant before she was struck by defendant’s automobile which he had seen and realized was going to strike her. He was, therefore, a “percipient witness to the impact causing [her] injuries.” (19 Cal.3d 59, 76.)

We also find in Krouse, supra, 19 Cal.3d 59, the roots of the uncertainty reflected by the instant case over whether the plaintiff must perceive the injury causing incident at all or may recover for emotional distress suffered on viewing its “immediate consequences” even though not present at the scene when it occurred. Krouse created uncertainty as to the meaning and importance of the plaintiff’s status as a “percipient witness” by approving the conclusion of the Court of Appeal in Archibald v. Braverman (1969) 275 Cal.App.2d 253 [79 Cal.Rptr. 723], that visual perception of the accident was not required without commenting on the context in which the Archibald court made its ruling. That decision had allowed recovery by a mother who “did not actually witness the tort but viewed the child’s injuries within moments after the occurrence of the injury-producing event.” (275 Cal.App.2d 253, 255.) Thus, it appeared that this court agreed that persons who were not present at the accident scene could recover damages for the emotional distress they later suffered when told by others of the injury to their loved one or when they later came to the scene.

*657Krouse, supra, 19 Cal.3d 59, was followed by Justus v. Atchison (1977) 19 Cal.3d 564 [139 Cal.Rptr. 97, 565 P.2d 122], in which the court identified the issue as whether the plaintiff fathers’ causes of action for the emotional impact of observing the stillborn birth of their children satisfied the Dillon guideline of shock resulting from a direct emotional impact from the sensory and contemporaneous observance of an “accident.” The court did not decide this question, ruling instead that the plaintiffs could not recover because neither had learned of the death of the fetus until informed by a doctor. Thus the “disabling shock” occurred only upon being informed by another of the injury. (19 Cal.3d 564, 585.) By implication, however, it seemed that the injury producing event need not be a sudden occurrence or accident.

However, the court in Hoyem v. Manhattan Beach City Sch. Dist. (1978) 22 Cal.3d 508, 523 [150 Cal.Rptr. 1, 585 P.2d 851], reaffirmed the requirement that the shock or emotional distress necessary to a cause of action for NIED under Dillon must “result from a ‘direct emotional impact’ on the plaintiff caused by ‘sensory and contemporaneous observance of the accident.’ ”

But in Nazaroff v. Superior Court (1978) 80 Cal.App.3d 553 [145 Cal.Rptr. 657], after considering this court’s post -Dillon decisions and those of the Courts of Appeal, including Archibald v. Braverman, supra, 275 Cal.App.2d 253, the concept of contemporaneous observance was broadened. The court held that a mother who, while searching for her missing three-year-old child, heard a neighbor scream his name, realized the child must have fallen into the neighbor’s pool, and saw the child being pulled from the pool and given cardiopulmonary resuscitation as she ran up, could state a cause of action under Dillon. The court reasoned that it was not necessary that the plaintiff perceive the accident if she suffered physical harm that resulted from “the direct emotional impact from the contemporaneous observation of the immediate consequences of the defendants’ negligent act, which was the proximate cause of the injury and death of her son.” (Nazaroff, supra, 80 Cal.App.3d 553, 566.) The court explained that the plaintiff may have mentally reconstructed the accident and the child might still have been experiencing the injuries caused by the defendants’ negligence when his mother first observed him.

With the Nazaroff decision, supra, 80 Cal.App.3d 553, therefore, the Dillon guideline of sensory perception of an accident or injury was no longer considered a prerequisite to recovery. It was still necessary, however, that the plaintiff who had witnessed the immediate consequences of an injury producing incident suffer emotional distress sufficient to result in physical injury.

*658Both the physical harm and accident or sudden occurrence elements were eliminated, however, in Molien v. Kaiser Foundation Hospitals, supra, 27 Cal.3d 916, at least as to those plaintiffs who could claim to be “direct victims” of the defendant’s negligence. The court held in Molien that a defendant hospital and doctor owed a duty directly to the husband of a patient who had been diagnosed erroneously as having syphilis, and had been told to so advise the husband in order that he could receive testing and, if necessary, treatment.

In finding the existence of a duty to the husband of the patient, the court reasoned that the risk of harm to the husband was reasonably foreseeable, and that the tortious conduct was directed to him as well as the patient. (Molien v. Kaiser Foundation Hospitals, supra, 27 Cal.3d 916, 922.) The status of the plaintiff mother in Dillon was distinguished as she suffered her injury solely as a “percipient witness” to the infliction of injury on another. She was therefore a “bystander” rather than a “direct victim.”

The court did not further explain this distinction, or its relevance to whether the plaintiff should be allowed to recover damages for emotional distress. Both decisions had looked to the relationships of the parties to find foreseeability of the injury and thus a “duty to the plaintiff.” The basis for finding a duty to the mother in Dillon was the foreseeability of her emotional distress to “the negligent driver who causes the death of a young child [and] may reasonably expect that the mother will not be far distant and will upon witnessing the accident suffer emotional trauma.” (Dillon v. Legg, supra, 68 Cal.2d 728, 741.) In Molien, “[t]he risk of harm to plaintiff was reasonably foreseeable to defendants. It is easily predictable that an erroneous diagnosis of syphilis and its probable source would produce marital discord and resultant emotional distress to a married patient’s spouse; [the physician’s] advice to Mrs. Molien to have her husband examined for the disease confirms that plaintiff [husband] was a foreseeable victim of the negligent diagnosis. . . . [H] We thus agree with plaintiff that the alleged tortious conduct of defendant was directed to him as well as to his wife. Because the risk of harm to him was reasonably foreseeable we hold . . . that under these circumstances defendants owed plaintiff" a duty to exercise due care in diagnosing the physical condition of his wife.” (27 Cal.3d 916, 923.)

Molien neither established criteria for characterizing a plaintiff as a “direct” victim, nor explained the justification for permitting “direct” victims to recover when “bystander” plaintiffs could not. The immediate effect of the decision, however, was to permit some persons who had no prior relationship with the defendant that gave rise to a duty, who did not suffer physical injury as a result of emotional distress, who did not observe the *659negligent conduct, and who had not been at or near the scene of the negligent act to recover for emotional distress on a pure foreseeability-of-the injury basis. The limitations on recovery for emotional distress that had been suggested in the Dillon “guidelines” were not applicable to “direct” victims of a defendant’s negligence.

The subtleties in the distinction between the right to recover as a “bystander” and as a “direct victim” created what one Court of Appeal has described as an “amorphous nether realm" (Newton v. Kaiser Foundation Hospitals (1986) 184 Cal.App.3d 386, 391 [228 Cal.Rptr. 890]; Andalon v. Superior Court (1984) 162 Cal.App.3d 600, 610 [208 Cal.Rptr. 899]), and have contributed in some measure to the present difficulty in defining the scope of an NIED action.5 In Andalon v. Superior Court, supra, 162 Cal.App.3d 600, the court found that a physician’s duty arose out of contract, after it had abandoned the effort to resolve the “direct” or “bystander” dilemma: “The problem which arises from this cryptic explanation is: how are we to distinguish between ‘direct victim’ cases and ‘bystander’ cases? An impression is given that the foreseeability of the particular injury to the husband alone explains the result. The inference suggested is that a ‘direct victim’ is a person whose emotional distress is a reasonably foreseeable consequence of the conduct of the defendant. This does not provide criteria which delimit what counts as reasonable foreseeability. It leads into the quagmire of novel claims which the Supreme Court foresaw as an unacceptable consequence of a ‘pure’ foreseeability analysis . . . .” (162 Cal.App.3d at p. 609.)

“[F]oreseeability,” the court noted later in Newton v. Kaiser Foundations Hospitals, supra, 184 Cal.App.3d 386, 391, “is endless because foreseeability, like light, travels indefinitely in a vacuum.” Molien, supra, 27 Cal.3d 916, thus, left to future cases the “unenviable tasks of distinguishing bystander from direct victim cases and establishing limits for the latter . . . with a ‘foreseeable’ diversity of results.” (Newton, supra, 184 Cal.App.3d at p. 390.)

The Dillon-Molien waters, described by the Court of Appeal in this case as “murky,” were further muddied when the right to recover in the negligence action was expanded to include “bystander” plaintiffs who suffered emotional distress, but no contemporaneous or consequential physical injury. We had reasoned in Molien that the requirement of physical injury was no more necessary to ensure that a claim of serious emotional distress is *660genuine in NIED actions than it was in actions for intentional infliction of emotional distress. Rather the question was one of proof to be left to the jury. (Molien v. Kaiser Foundation Hospitals, supra, 27 Cal.3d 916, 930.)

Cases subsequent to Molien assumed that the reasoning which led the court to abandon the physical-injury requirement extended to “bystander” NIED actions. Physical manifestation of the serious nature of the mental distress suffered by the plaintiff was no longer an element of the cause of action. (See Ochoa v. Superior Court, supra, 39 Cal.3d 159; Hedlund v. Superior Court (1983) 34 Cal.3d 695, 706, fn. 8 [194 Cal.Rptr. 805, 669 P.2d 41, 41 A.L.R.4th 1063].)

Ochoa v. Superior Court, supra, 39 Cal.3d 159, 172, partially explained and limited “direct victim” recovery under Molien, supra, 27 Cal.3d 916, to situations in which the defendant’s negligence is “by its very nature directed at” the plaintiff. However, Ochoa also indicated that the dimensions of the NIED tort might be expanded further for “bystander” plaintiffs. Ochoa confirmed that recovery was permitted even though the injury producing event was not sudden or accidental, and even though its negligent cause was not immediately apparent. The court observed that the factors set forth in Dillon had been offered only as guidelines, and suggested that none was essential to recovery for NIED. Foreseeability that the injury would cause emotional distress was the proper inquiry. (Ochoa, supra, at p. 170.)

That dictum in Ochoa was broader than the issue presented in Ochoa, however. The plaintiff mother had observed the effects of the defendants’ negligent failure to diagnose and properly treat the illness of her teenage son. Her observation of his pain and suffering, and his deteriorating condition, as the defendants failed to either properly care for him or accede to her entreaty that she be permitted to obtain care for him, was the cause of the emotional distress for which she sought to recover. The allegations of the complaint satisfied only two of the Dillon factors—she was at the scene of the negligent injury producing conduct and was closely related to the person whose physical injury caused her distress. Defendants’ negligence in failing to give proper medical treatment, however, was not a sudden accidental occurrence and thus the second Dillon factor was not met: “Whether the shock resulted from a direct emotional impact upon plaintiff from the sensory and contemporaneous observance of the accident, . . .” (Dillon v. Legg, supra, 68 Cal.2d 728, 740-741.)

This court, after reviewing several decisions of the Courts of Appeal which had limited recovery for NIED to percipient witnesses of a “sudden occurrence,” held that this requirement was an unwarranted restriction on the cause of action authorized in Dillon. “Such a restriction arbitrarily *661limits liability when there is a high degree of foreseeability of shock to the plaintiff and the shock flows from an abnormal event, and, as such, unduly frustrates the goal of compensation—the very purpose which the cause of action was meant to further.” (Ochoa v. Superior Court, supra, 39 Cal.3d 159, 168.)

Ochoa also held that the NIED plaintiif need not be aware that the conduct was “tortious.” Reasoning that such a requirement leads to anomalous results, the court held that “when there is observation of the defendant’s conduct and the child’s injury and contemporaneous awareness the defendant’s conduct or lack thereof is causing harm to the child, recovery is permitted.” (Ochoa v. Superior Court, supra, 39 Cal.3d 159, 170.) Thus, the plaintiff in that case did not have to know that the defendants had negligently misdiagnosed her son. It was enough that she knew that they were refusing or neglecting to give him additional treatment and this was the cause of the additional injury he was suffering.

In sum, however, as to “bystander” NIED actions, Ochoa held only that recovery would be permitted if the plaintiff" observes both the defendant’s conduct and the resultant injury, and is aware at that time that the conduct is causing the injury.6 The Court of Appeal erred in concluding that Ochoa, supra, 39 Cal.3d 159, held that these NIED plaintiffs need not witness the defendant’s conduct.

V

Clarification of the Right to Recover for NIED

Not surprisingly, this “case-to-case” or ad hoc approach to development of the law that misled the Court of Appeal in this case has not only produced inconsistent rulings in the lower courts, but has provoked considerable critical comment by scholars who attempt to reconcile the cases. (See Rabin, Tort Recovery for Negligently Inflicted Economic Loss: A Reassessment (1985) 37 Stan.L.Rev. 1513, 1524-1526, hereafter Rabin; Diamond, Dillon v. Legg Revisited: Toward a Unified Theory of Compensating Bystanders and Relatives for Intangible Injuries (1984) 35 Hastings L.J. 477, *662hereafter Diamond; Bell, The Bell Tolls: Toward Full Tort Recovery for Psychic Injury (1984) 36 U.Fla.L.Rev. 333; Pearson, Liability for Negligently Inflicted Psychic Harm: A Response to Professor Bell (1984) 36 U.Fla.L.Rev. 413; Nolan & Ursin, Negligent Infliction of Emotional Distress: Coherence Emerging from Chaos (1982) 33 Hastings L.J. 583, 620; Miller, The Scope of Liability for Negligent Infliction of Emotional Distress: Making “The Punishment Fit the Crime” (1979) 1 Hawaii L.Rev. 1, hereafter Miller; Pearson, Liability to Bystanders for Negligently Inflicted Emotional Harm—A Comment on the Nature of Arbitrary Rules (1982) 34 U.FIa.L.Rev. 477; Comment, Dillon Revisited: Toward a Better Paradigm for Bystander Cases (1982) 43 Ohio St. L.J. 931, 948; Note, Limiting Liability for the Negligent Infliction of Emotional Distress: The “Bystander Recovery” Cases (1981) 54 So.Cal.L.Rev. 847; Note, Molien v. Kaiser Foundation Hospitals: California’s New Tort of Negligent Infliction of Serious Emotional Distress (1982) 18 Cal. Western L.Rev. 101; Note, Negligent Infliction of Emotional Distress: Reconciling the Bystander and Direct Victim Causes of Action (1983) 18 U.S.F. L.Rev. 145.)

Proposals to eliminate the arbitrary results of the proliferating, inconsistent and often conflicting Dillon progeny include the suggestion that recovery be allowed in any case in which recovery for physical injury is permitted. (See Ochoa v. Superior Court, supra, 39 Cal.3d 159, 178 [conc. opn. of Grodin, J.].) Another would limit recovery to the close-relatives class contemplated by Dillon, but allow recovery whenever mental distress to the plaintiff was foreseeable. (Id. at p. 196, conc. & dis. opn. of Bird, C. J.) At the other extreme, respondent here and amicus curiae Association for California Tort Reform argue, in essence, that the Dillon “guidelines” should be recognized as substantive limitations or elements of the tort.

In his thoughtful article documenting the conflicting and sometimes arbitrary results of attempts by lower courts to apply Dillon and Molien,7 Professor Diamond analyzes the “flaws” in the Dillon analysis which he believes have contributed to the problem. He concludes that the Dillon-based cause of action identifies a “duty” on the basis of the purely fortuitous circumstances in which an injury occurs. As a result, when recovery for emotional distress alone is permitted under the Dillon guidelines, “foreseeability” is not a realistic indicator of potential liability and does not afford a rational limitation on recovery. Nor, Diamond suggests, do the Dillon guidelines, particularly under the expanded right to recovery created by Molien (and, we note, Ochoa), provide such limitation. In his view, only one of the Dillon guidelines is even relevant to foreseeability—the relationship of the plaintiff to the person suffering physical injury—because it is foresee*663able that the emotional distress suffered by a close relative on witnessing that injury will be greater than that of a stranger. (Diamond, supra, 35 Hastings L.J., at pp. 487-489.)

Diamond argues that the fact that it is foreseeable that a close relative will suffer psychological trauma at witnessing the injury does not adequately limit liability for damages for such intangible losses. When recovery for emotional distress was permitted only as an item of “parasitic” damage suffered by a plaintiff who had been physically injured, the defendant’s exposure was limited by more predictable factors. Foreseeability that various activities may cause physical injury limits the potential universe of persons who may be harmed. This, in turn, by limiting the negligent actor’s exposure makes it possible to protect potential victims and the defendant through insurance or other risk-spreading mechanisms whose cost is more closely related to the risk, and ensures that this exposure bears a more rational relationship to the defendant’s culpability. (Diamond, supra, 35 Hastings L.J., at pp. 490-493.)

Another scholar suggests that any foreseeable plaintiff be permitted to recover for NIED and loss of filial consortium, but only for economic loss, thereby reconciling the divergent paths and limitations on recovery for noneconomic damages in these related torts with that permitted in actions for wrongful life. (See Miller, supra, 1 Hawaii L.Rev., at pp. 39-41.) Diamond agrees that this solution would permit recovery by all foreseeable plaintiffs in “intangible” tort cases, while restricting recovery to “economically acceptable” limits. (Diamond, supra, 35 Hastings L.J., at p. 480.)

In the NIED context, however, permitting recovery of economic damages by all foreseeable plaintiffs would expose defendants to risks no less arbitrary and unacceptable than those presently existing. While the recovery by individual victims might be less, the number of potential plaintiffs traumatized by reason of defendant’s negligent conduct toward another, would turn on fortuitous circumstances wholly unrelated to the culpability of the defendant.

Our own prior decisions identify factors that will appropriately circumscribe the right to damages, but do not deny recovery to plaintiffs whose emotional injury is real even if not accompanied by out-of-pocket expense. Notwithstanding the broad language in some of those decisions, it is clear that foreseeability of the injury alone is not a useful “guideline” or a meaningful restriction on the scope of the NIED action. The Dillon experience confirms, as one commentator observed, that “[fjoreseeability proves too much. . . . Although it may set tolerable limits for most types of physical harm, it provides virtually no limit on liability for nonphysical harm.” *664(Rabin, supra, 37 Stan. L. Rev. at p. 1526.) It is apparent that reliance on foreseeability of injury alone in finding a duty, and thus a right to recover, is not adequate when the damages sought are for an intangible injury, In order to avoid limitless liability out of all proportion to the degree of a defendant’s negligence, and against which it is impossible to insure without imposing unacceptable costs on those among whom the risk is spread, the right to recover for negligently caused emotional distress must be limited.

We acknowledged and addressed one aspect of this problem in Elden v. Sheldon (1988) 46 Cal.3d 267 [250 Cal.Rptr. 254, 758 P.2d 582], holding that cohabitation, without formal marriage, did not constitute the close relationship contemplated by the Dillon guidelines and that foreseeability of injury alone does not justify imposition of liability for negligently caused emotional distress. In so doing, we again recognized that policy considerations justify restrictions on recovery for emotional distress notwithstanding the sometimes arbitrary result, and that the court has an obligation to establish those restrictions. Elden confirmed that those policy considerations include both the burden on the courts in applying vaguely defined criteria and the importance of limiting the scope of liability for negligence. If the consequences of a negligent act are not limited an intolerable burden is placed on society. A “bright line in this area of the law is essential.” (46 Cal.3d 267, 277.)

The issue resolved in Elden was too narrow to create that “bright line” for all NIED actions. This case, however, presents a broader question and thus affords the court a better opportunity to meet its obligation to create a clear rule under which liability may be determined. In so doing we balance the impact of arbitrary lines which deny recovery to some victims whose injury is very real against that of imposing liability out of proportion to culpability for negligent acts. We also weigh in the balance the importance to the administration of justice of clear guidelines under which litigants and trial courts may resolve disputes. Thus, as we did in Elden, supra, 46 Cal.3d 267, we return to the concerns which prompted the Amaya court, supra, 59 Cal.2d 295, to deny recovery for negligent infliction of emotional distress.

Among the concerns of the Amaya court was the social cost of imposing liability on a negligent tortfeasor for all foreseeable emotional distress suffered by relatives who witnessed the injury. The court again faced this problem in Borer v. American Airlines, Inc. (1977) 19 Cal.3d 441 [138 Cal.Rptr. 302, 563 P.2d 858], in which the court was asked to recognize a child’s right to recover for the loss of a parent’s consortium, an action that, like NIED, seeks monetary damages for mental or emotional loss. Refusing to permit such “filial” consortium actions, the court concluded that the cause of action for loss of consortium must be narrowly circumscribed. *665“Loss of consortium is an intangible injury for which money damages do not afford an accurate measure or suitable recompense; recognition of a right to recover for such losses . . . may substantially increase the number of claims asserted in ordinary accident cases, the expense of settling or resolving such claims, and the ultimate liability of the defendants." (Id. at p. 444.) The decision, we explained, was one of policy. We reasoned that we could not “ignore the social burden of providing damages . . . merely because the money to pay such awards comes initially from the ‘negligent’ defendant or his insurer. Realistically the burden . . . must be borne by the public generally in increased insurance premiums or, otherwise, in the enhanced danger that accrues from the greater number of people who may choose to go without any insurance. We must also take into account the cost of administration of a system to determine and pay [the] awards; . . .” (Id. at p. 447.)

While we emphasized in Borer, supra, 19 Cal.3d 441, that our refusal to extend the right to recover damages in consortium cases did not signal a refusal to allow damages for intangible losses in other contexts, the policy bases for the decision are relevant to defining the NIED cause of action. Crucial to the Borer decision were the intangible nature of the loss, the inadequacy of monetary damages to make whole the loss, the difficulty in measuring the damage, and the societal cost of attempting to compensate the plaintiff. Multiplication of the defendant’s liability was an additional concern. The number of family members who might seek damages on the basis of a single incident could unreasonably enlarge the defendant’s burden. We rejected a suggestion that principles enunciated in Dillon mandated recognition of the children’s cause of action, noting what was then the Dillon limitation—that the Dillon plaintiff have suffered physical injury— which limited the class of potential plaintiffs. (19 Cal.3d at p. 450.)8

The court again recognized the need to limit recovery of monetary damages for intangible loss in Turpin v. Sortini (1982) 31 Cal.3d 220, 237 [182 Cal.Rptr. 337, 643 P.2d 954]. There, in an action for “wrongful life," the court limited damages to economic loss and observed that “a monetary award of general damages . . . cannot in any meaningful sense compensate the plaintiff.”

*666Ochoa v. Superior Court, supra, 39 Cal.3d 159, 165, footnote 6,9 offers additional guidance, justifying what we acknowledge must be arbitrary lines to similarly limit the class of potential plaintiffs if emotional injury absent physical harm is to continue to be a recoverable item of damages in a negligence action. The impact of personally observing the injury-producing event in most, although concededly not all, cases distinguishes the plaintiff’s resultant emotional distress from the emotion felt when one learns of the injury or death of a loved one from another, or observes pain and suffering but not the traumatic cause of the injury. Greater certainty and a more reasonable limit on the exposure to liability for negligent conduct is possible by limiting the right to recover for negligently caused emotional distress to plaintiffs who personally and contemporaneously perceive the injury-producing event and its traumatic consequences.

Similar reasoning justifies limiting recovery to persons closely related by blood or marriage since, in common experience, it is more likely that they will suffer a greater degree of emotional distress than a disinterested witness to negligently caused pain and suffering or death. Such limitations are indisputably arbitrary since it is foreseeable that in some cases unrelated persons have a relationship to the victim or are so affected by the traumatic event that they suffer equivalent emotional distress. As we have observed, however, drawing arbitrary lines is unavoidable if we are to limit liability and establish meaningful rules for application by litigants and lower courts.

No policy supports extension of the right to recover for NIED to a larger class of plaintiffs. Emotional distress is an intangible condition experienced by most persons, even absent negligence, at some time during their lives. Close relatives suffer serious, even debilitating, emotional reactions to the injury, death, serious illness, and evident suffering of loved ones. These reactions occur regardless of the cause of the loved one’s illness, injury, or death. That relatives will have severe emotional distress is an unavoidable aspect of the “human condition.” The emotional distress for which monetary damages may be recovered, however, ought not to be that form of acute emotional distress or the transient emotional reaction to the *667occasional gruesome or horrible incident to which every person may potentially be exposed in an industrial and sometimes violent society. Regardless of the depth of feeling or the resultant physical or mental illness that results from witnessing violent events, persons unrelated to those injured or killed may not now recover for such emotional upheaval even if negligently caused. Close relatives who witness the accidental injury or death of a loved one and suffer emotional trauma may not recover when the loved one’s conduct was the cause of that emotional trauma. The overwhelming majority of “emotional distress” which we endure, therefore, is not compensable.

Unlike an award of damages for intentionally caused emotional distress which is punitive, the award for NIED simply reflects society’s belief that a negligent actor bears some responsibility for the effect of his conduct on persons other than those who suffer physical injury. In identifying those persons and the circumstances in which the defendant will be held to redress the injury, it is appropriate to restrict recovery to those persons who will suffer an emotional impact beyond the impact that can be anticipated whenever one learns that a relative is injured, or dies, or the emotion felt by a “disinterested” witness. The class of potential plaintiffs should be limited to those who because of their relationship suffer the greatest emotional distress. When the right to recover is limited in this manner, the liability bears a reasonable relationship to the culpability of the negligent defendant.

The elements which justify and simultaneously limit an award of damages for emotional distress caused by awareness of the negligent infliction of injury to a close relative are those noted in Ochoa—the traumatic emotional effect on the plaintiff who contemporaneously observes both the event or conduct that causes serious injury to a close relative and the injury itself. Even if it is “foreseeable” that persons other than closely related percipient witnesses may suffer emotional distress, this fact does not justify the imposition of what threatens to become unlimited liability for emotional distress on a defendant whose conduct is simply negligent. Nor does such abstract “foreseeability” warrant continued reliance on the assumption that the limits of liability will become any clearer if lower courts are permitted to continue approaching the issue on a “case-to-case” basis some 20 years after Dillon.

(P) We conclude, therefore, that a plaintiff may recover damages for emotional distress caused by observing the negligently inflicted injury of a third person if, but only if, said plaintiff: (1) is closely related to the injury *668victim;10 (2) is present at the scene of the injury-producing event at the time it occurs and is then aware that it is causing injury to the victim;11 and (3) as a result suffers serious emotional distress—a reaction beyond that which would be anticipated in a disinterested witness and which is not an abnormal response to the circumstances.12 These factors were present in Ochoa and each of this court’s prior decisions upholding recovery for NIED.

The dictum in Ochoa suggesting that the factors noted in the Dillon guidelines are not essential in determining whether a plaintiff is a foreseeable victim of defendant’s negligence should not be relied on. The merely negligent actor does not owe a duty the law will recognize to make monetary amends to all persons who may have suffered emotional distress on viewing or learning about the injurious consequences of his conduct. To the extent they are inconsistent with this conclusion, Nazaroff v. Superior Court, supra, 80 Cal.App.3d 553, and Archibald v. Braverman, supra, 275 Cal.App.2d 253, are disapproved. Experience has shown that, contrary to the expectation of the Dillon majority, and with apology to Bernard Witkin, there are clear judicial days on which a court can foresee forever and thus determine liability but none on which that foresight alone provides a socially and judicially acceptable limit on recovery of damages for that injury.

*669VI

Disposition

The undisputed facts establish that plaintiff was not present at the scene of the accident in which her son was injured. She did not observe defendant’s conduct and was not aware that her son was being injured. She could not, therefore, establish a right to recover for the emotional distress she suffered when she subsequently learned of the accident and observed its consequences. The order granting summary judgment was proper.

The judgment of the Court of Appeal is reversed.

Each party shall bear its own costs on appeal.

Lucas, C. J., Panelli, J., and Arguelles, J.,* concurred.

KAUFMAN, J., Concurring.

We granted review in this case because of the obvious and continuing difficulties that have plagued trial courts and litigants in the area of negligent infliction of emotional distress. Of course, any meaningful review of the issue necessarily entails reappraising, in the light of 20 years of experience, our landmark holding in Dillon v. Legg (1968) 68 Cal.2d 728 [69 Cal.Rptr. 72, 441 P.2d 912, 29 A.L.R.3d 1316], that a plaintiff may recover for the emotional distress induced by the apprehension of negligently caused injury to a third person. Two such “reappraisals” have now been suggested.

The majority opinion by Justice Eagleson proposes to convert Dillon's flexible “guidelines” 1 for determining whether the risk of emotional injury was foreseeable or within the defendant’s duty of care, into strict “elements” necessary to recovery. While conceding that such a doctrinaire approach will necessarily lead to “arbitrary” results, Justice Eagleson nevertheless concludes that “[g]reater certainty and a more reasonable limit on the exposure to liability for negligent conduct” require strict limitations. (Maj. opn., p. 666.)

Justice Broussard, in dissent, opposes the effort to rigidify the Dillon guidelines. He urges, instead, that the court remain faithful to the guidelines *670as originally conceived—as specific but “flexible” limitations on liability— and adhere to Dillon's original reliance on “foreseeability as a general limit on tort liability.” (Dis. opn. of Broussard, J., p. 685.) Justice Broussard denies that Dillon has failed to afford adequate guidance to the lower courts or to confine liability within reasonable limits. On the contrary, the Dillon approach, in the dissent’s view, has provided—and continues to provide—a workable and “principled basis for determining liability. ...” (Id. at p. 689, italics added.)2

With all due respect, I do not believe that either the majority opinion or the dissent has articulated a genuinely “principled” rule of law. On the one hand, experience has shown that rigid doctrinal limitations on bystander liability, such as that suggested by Justice Eagleson, result inevitably in disparate treatment of plaintiffs in substantially the same position. To be sure, the majority freely—one might say almost cheerfully —acknowledges that its position is arbitrary; yet nowhere does it consider the cost of such institutionalized caprice, not only to the individuals involved, but to the integrity of the judiciary as a whole.

On the other hand, two decades of adjudication under the inexact guidelines created by Dillon and touted by the dissent, has, if anything, created a body of case law marked by even greater confusion and inconsistency of result. (See Nolan and Ursin, Negligent Infliction of Emotional Distress: Coherence Emerging from Chaos (1982) 33 Hastings L.J. 583, 620 [“Arbitrary and unseemly results in emotional distress cases have been the consequence of this experimentation.”].)

The situation, therefore, calls for a wholesale reappraisal of the wisdom of permitting recovery for emotional distress resulting from injury to others.

A. Background

The history of negligent infliction of emotional distress is a chronicle of “false starts.” (Nolan and Ursin, supra, 33 Hastings L.J. at p. 604.) Initially, the courts were reluctant to allow any recovery for intangible harms such as fright or emotional distress resulting from negligent conduct. (Prosser and Keeton, The Law of Torts (5th ed. 1984) § 54, p. 360.) Later, plaintiffs were permitted to recover for the emotional distress occasioned by the fear for their own safety, but only when accompanied by physical impact. (Prosser *671and Keeton, supra, § 54, pp. 363-364; Comment Dillon Revisited: Toward A Better Paradigm for Bystander Cases (1982) 43 Ohio St. L.J. 932, 933-934.)

The “impact” rule was soon recognized as inherently arbitrary, however, because of the obvious discontinuity between the rule’s scope and its underlying purpose. As one commentator wryly observed, “[A] near miss may be as frightening as a direct hit.” (Pearson, Liability to Bystanders for Negligently Inflicted Emotional Harm—A Comment on the Nature of Arbitrary Rules (1982) 34 U.Fla.L.Rev. 477, 488.)

Thus, the “impact” rule was eventually replaced by a rule allowing recovery for the emotional distress resulting from threats to the plaintiff’s safety, regardless of physical impact, if the plaintiff was within the “zone of physical impact or danger.” (See, e.g., Battalla v. State (1961) 10 N.Y.2d 257 [219 N.Y.S.2d 34, 176 N.E.2d 729]; Waube v. Warrington (1935) 216 Wis. 603 [258 N.W. 497]); Note, Limiting Liability for the Negligent Infliction of Emotional Distress: The “Bystander Recovery” Cases (1981) 54 So.Cal.L.Rev. 847, 849-850.) The seminal California decision rejecting the impact test in favor of the zone-of-danger rule is Amaya v. Home, Ice, Fuel & Supply Co., supra, 59 Cal.2d 295.

In reaffirming the zone-of-danger rule, the Amaya court rejected the argument that liability should be extended in favor of one who claims injury through fright or shock induced by negligent conduct directed not to oneself but to a third person. While acknowledging that emotional trauma induced by the apprehension of danger to a loved one was eminently “foreseeable,” the Amaya court nevertheless refused to recognize that freedom from such distress constituted a legally protected interest. A number of policy reasons were cited by the Amaya court in support of its decision, including: the risk of fraudulent claims based upon such intangible “psychic” injuries; the threat of unlimited liability and the attendant economic strain on the insurance system; the disproportionality between the culpability of the negligent tortfeasor and the liability imposed; and the impossibility of formulating a “sensible or just stopping point.” (59 Cal.2d at pp. 310-315.)

Only five years later, however, in Dillon v. Legg, supra, 68 Cal.2d 728, the court reversed field, rejecting Amaya’s zone-of-danger test as “hopelessly artificial” and recognizing for the first time in American jurisprudence the concept of “bystander” liability. In Dillon, the plaintiff’s daughter was struck and killed by a car negligently driven by the defendant. Plaintiff, who observed the accident from an unspecified location, sought damages for the emotional distress resulting from the incident; plaintiff’s other daughter, who was standing on the curb, also claimed emotional distress. The trial court sustained a motion for summary judgment as to the plaintiff, the *672decedent’s mother, on the ground that the pleadings failed to establish that she was within the zone of danger and feared for her own safety, but denied a similar motion as to the daughter because of the possibility that she was within such zone and feared for her own safety. (Id. at pp. 731-732.)

The Dillon court reversed, explaining its reasoning as follows: “[W]e can hardly justify relief to the sister for trauma which she suffered upon apprehension of the child’s death and yet deny it to the mother merely because of a happenstance that the sister was some few yards closer to the accident. The instant case exposes the hopeless artificiality of the zone-of-danger rule.” (68 Cal.2d at p. 733.) And so the Dillon court recognized the mother’s right to recover for emotional distress caused by the apprehension of danger to her daughter; delineated its now-famous guidelines for determining foreseeability; and left to “future cases” the task of delimiting “bystander” liability “upon facts more subtle than the compelling ones alleged in the complaint before us.” (Id. at p. 741.)

Of course, it was immediately apparent that the Dillon court’s holding was not really based on any inherent flaw or “artificiality” in the zone-of-danger rule. As the Dillon dissent pointed out, that rule was designed to compensate for the emotional distress occasioned by the fear of physical impact solely “to oneself.” (68 Cal.2d at p. 750, dis. opn. of Burke, J.) A rule which permitted such recovery to one within the zone of physical impact or danger was not arbitrary at all. (Ibid.) It became arbitrary only after the court recognized fear for the safety of others as a legally protected interest. (See Pearson, supra, 34 U.Fla.L.Rev. at pp. 490-491.) Indeed, as discussed below, the zone-of-danger rule remains a viable measure of emotional distress in many states where the bystander-recovery rule has been rejected as inherently arbitrary.

B. Dillon Rejected as Hopelessly Arbitrary

In light of the foregoing, it is perhaps not surprising to find that during the last 20 years Dillon has played to decidedly mixed reviews among our sister states. While many have adopted the Dillon approach or some variation thereon, many others have explicitly rejected “bystander” recovery and adopted in its place the “zone-of-danger” rule articulated in the very decision which Dillon overruled, Amaya v. Home Ice Fuel & Supply Co., supra, 59 Cal.2d 295. (See, e.g., Rickey v. Chicago Transit Authority (1983) 98 Ill.2d 546 [457 N.E.2d 1, 5]; Towns v. Anderson (1978) 195 Colo. 517 [579 P.2d 1163] and James v. Harris (Colo.App. 1986) 729 P.2d 986; Guilmette v. Alexander (1969) 128 Vt. 116 [259 A.2d 12] and Vaillancourt v. Medical Ctr. Hosp. of Vt. (1980) 139 Vt. 138 [425 A.2d 92]; Whetham v. Bismarck Hospital (N.D. 1972) 197 N.W.2d 678; Stadler v. Cross (Minn. 1980) 295 *673N.W.2d 552 and Leaon v. Washington County (Minn. 1986) 397 N.W.2d 867; Grimsby v. Samson (1975) 85 Wn.2d 52 [530 P.2d 291, 77 A.L.R.2d 436].) Indeed, despite the wide range of opinion on the subject of negligent infliction of emotional distress, several recent decisions have noted that the majority of jurisdictions where the question has been considered have rejected the Dillon approach in favor of the zone-of-danger rule. (See Rickey v. Chicago Transit Authority, supra, 475 N.E.2d at p. 5; Gillman v. Burlington Northern R. Co. (N.D.Ill. 1987) 673 F.Supp. 913, 917, fn. 1.)

While the courts rejecting bystander liability have cited a number of reasons, one argument in particular has been considered dispositive: Dillon's confident prediction that future courts would be able to fix just and sensible boundaries on bystander liability has been found to be wholly illusory—both in theory and in practice.3 (See, e.g., Stadler v. Cross, supra, 295 N.W.2d at p. 555 [“No arguments have been presented that persuade us that the problems we see in limiting liability once it is extended beyond the zone of danger of physical impact can be justly overcome.”]; Grimsby v. Samson, supra, 530 P.2d at p. 294 [“(T)here appears to be no rational way to restrict the scope of liability even as attempted by Dillon's three limiting standards.”]; James v. Harris, supra, 729 P.2d at p. 988 [“ ‘Assuming that there are cogent reasons for extending liability in favor of victims of shock resulting from injury to others, there appears to be no rational way to limit the scope of liability.’ ”].)

As the New York Court of Appeals, writing one year after Dillon, presciently observed: “Every parent who loses a child or whose child . . . suffers an injury is likely to sustain grievous psychological trauma. . . . Any rule based ... on eyewitnessing the accident could stand only until the first case comes along in which the parent is in the immediate vicinity but did not see the accident.” (Tobin v. Grossman, supra, 249 N.E.2d at p. 423.) Of course, that case “came along” in California only one short year after Dillon. In Archibald v. Braverman (1969) 275 Cal.App.2d 253 [79 Cal.Rptr. 723], the court allowed recovery by a mother who “did not actually witness the tort but viewed the child’s injuries within moments after the occurrence of the injury-producing event.” (Id. at p. 255.) Similar cases have followed. In Nazaroff v. Superior Court (1978) 80 Cal.App.3d 553 [145 Cal.Rptr. 657], for example, the court reversed the entry of summary judgment in favor of the defendant, holding that a mother who *674had witnessed her infant son being pulled from a pool minutes after drowning could state a claim for negligent infliction of emotional distress. In the very case now under review, the Court of Appeal reversed the entry of summary judgment in favor of defendants, the owners and driver of a car which struck the child of the plaintiff", Maria Thing, where the evidence showed that Maria did not see the impact but rushed to the scene in time to see her son lying in the street, bleeding, and rode in the ambulance that took him to the hospital for treatment of severe injuries.

Of course, not all interpretations of Dillon have been so liberal. In Arauz v. Gerhardt (1977) 68 Cal.App.3d 937 [137 Cal.Rptr. 619], for example, the court denied recovery to a mother who arrived at the scene of a car accident involving her child within five minutes of the event; the court noted that the plaintiff" was “not at the scene of the accident at the time of the impact and [was] not near enough to the scene to have any sensory perception of the impact. . . .” (Id. at p. 949.) In Hathaway v. Superior Court (1980) 112 Cal.App.3d 728 [169 Cal.Rptr. 435], the court denied recovery to the parents of a child who was electrocuted while playing in a relative’s yard as the parents relaxed inside. Though the plaintiff’s arrived less than one minute after the accident and saw the child lying in a puddle, gagging and spitting up, the court held that Dillon precluded recovery because the infant “was no longer gripping the water cooler and receiving the electrical charge.” (Id. at p. 736.)

Twenty-five years ago, this court posed a series of rhetorical questions concerning the guidelines later adopted in Dillon: “[H]ow soon is ‘fairly contemporaneous?’ What is the magic in the plaintiff’s being ‘present’? Is the shock any less immediate if the mother does not know of the accident until the injured child is brought home? And what if the plaintiff is present at the scene but is nevertheless unaware of the danger or injury to the third person until shortly after the accident has occurred . . . ?” (Amaya v. Home Ice, Fuel & Supply Co., supra, 59 Cal.2d at p. 313.) As the foregoing sampling of Dillon’s progeny vividly demonstrates, we are no closer to answers today than we were then. The questions, however, are no longer hypothetical—they are real: Is there any rational basis to infer that Mrs. Arauz was any less traumatized than Mrs. Dillon because she saw her bloody infant five minutes after it was struck by defendant’s car? Was the Hathaways’ suffering mitigated by the fact that they witnessed their child literally in death’s throes, but failed to witness the precipitating event? Could it be argued that the emotional distress is even more traumatic, more foreseeable, for parents such as the Hathaways who fail to witness the accident and later blame themselves for allowing it to occur?

Clearly, to apply the Dillon guidelines strictly and deny recovery for emotional distress because the plaintiff was not a contemporaneous eyewit*675ness of the accident but viewed the immediate consequences, ill serves the policy of compensating foreseeable victims of emotional trauma. Yet once it is admitted that temporal and spatial limitations bear no rational relationship to the likelihood of psychic injury, it becomes impossible to define, as the Amaya court well understood, any “sensible or just stopping point.” (59 Cal.2d at p. 311.) By what humane and principled standard might a court decide, as a matter of law, that witnessing the bloody and chaotic aftermath of an accident involving a loved one is compensable if viewed within 1 minute of impact but noncompensable after 15? or 30? Is the shock of standing by while others undertake frantic efforts to save the life of one’s child any less real or foreseeable when it occurs in an ambulance or emergency room rather than at the “scene”?

Obviously, a “flexible” construction of the Dillon guidelines cannot, ultimately, avoid drawing arbitrary and irrational distinctions any more than a strict construction. Justice Burke was right when he observed of the Dillon guidelines, “Upon analysis, their seeming certainty evaporates into arbitrariness, and inexplicable distinctions appear.” (Dillon v. Legg, supra, 68 Cal.2d at p. 749, dis. opn. of Burke, J.)

C. Dillon’s Arbitrary Approach Should Be Overturned

Of course, it could be argued that recovery—not rationality—is the essential thing; that ultimately justice is better served by arbitrarily denying recovery to some, than by absolutely denying recovery to all. I find this argument to be unpersuasive, however, for two reasons.

First, the cost of the institutionalized caprice which Dillon has wrought should not be underestimated. The foremost duty of the courts in a free society is the principled declaration of public norms. The legitimacy, prestige and effectiveness of the judiciary—the “least dangerous branch”—ultimately depend on public confidence in our unwavering commitment to this ideal. Any breakdown in principled decisionmaking, any rule for which no principled basis can be found and clearly articulated, subverts and discredits the institution as a whole.

It is not always easy, of course, to accommodate the desire for individual justice with the need for reasoned, well-grounded, general principles. We sacrifice the latter for the sake of the former, however, only at our peril. For the “power-base” of the courts, as noted above, is rather fragile; it consists of the perception of our role in the structure of American government as the voice of reason, and the faith that the laws we make today, we ourselves will be bound by tomorrow. Any “rule”—such as Dillon’s—which permits and even encourages judgments based not on universal standards but individual *676expediency, erodes the public trust which we serve, and on which we ultimately depend.

There is a second reason, apart from the inherently corrosive effect of arbitrary rules, that points to the conclusion that “bystander” liability should not be retained. The interest in freedom from emotional distress caused by negligent injury to a third party is simply not, in my view, an interest which the law can or should protect. It is not that the interest is less than compelling. The suffering of a parent from the death or injury of a child is terribly poignant, and has always been so. (See Tobin v. Grossman, supra, 249 N.E.2d at p. 422.) It is the very universality of such injury, however, which renders it inherently unsuitable to legal protection. The observation of the New York Court of Appeals in this regard is particularly pertinent: “While it may seem that there should be a remedy for every wrong, this is an ideal limited perforce by the realities of this world. . . . The risks of indirect harm from the loss or injury of loved ones is pervasive and inevitably realized at one time or another. Only a very small part of that risk is brought about by the culpable acts of others. This is the risk of living and bearing children. It is enough that the law establishes liability in favor of those directly or intentionally harmed.” (Id. at p. 424.)

A final argument against overruling Dillon is, of course, the simple fact that it has been the law for 20 years. Stare decisis should not be lightly dismissed in any thoughtful reconsideration of the law. History and experience, however, are the final judge of whether a decision was right or wrong, whether it should be retained, modified or abandoned. In this case, history and experience have shown, as the Amaya court accurately predicted, that the quest for sensible and just limits on bystander liability is “an inherently fruitless one.” (59 Cal.2d at p. 313.)

Adherence to precedent cannot justify the perpetuation of a policy ill-conceived in theory and unfair in practice. As Justice Harlan aptly observed: “[A] judicious reconsideration of precedent cannot be as threatening to public faith in the judiciary as continued adherence to a rule unjustified in reason, which produces different results for breaches of duty in situations that cannot be differentiated in policy. . . .” (Moragne v. States Marine Lines, Inc. (1970) 398 U.S. 375, 405 [26 L.Ed.2d 339, 359, 90 S.Ct. 1772].)

For the foregoing reasons, therefore, I would overrule Dillon v. Legg, supra, 68 Cal.2d 728, and reinstate Amaya v. Home Ice, Fuel & Supply Co., supra, 59 Cal.2d 295 as the law of this state. Since the plaintiff was indisputably not within the zone of danger and could not assert a claim for emotional distress as the result of fear for her own safety, she could not *677establish a right to recover. Accordingly, I concur in the majority’s conclusion that the order granting summary judgment in this case was proper.

MOSK, J., Dissenting.

I am in general agreement with the dissent of Justice Broussard. I write separately to point out my other areas of disagreement with the majority opinion.

Although the majority do not forthrightly overrule Dillon v. Legg (1968) 68 Cal.2d 728 [69 Cal.Rptr. 72, 441 P.2d 912, 29 A.L.R.3d 1316], and replace it with Amaya v. Home Ice, Fuel & Supply Co. (1963) 59 Cal.2d 295 [29 Cal.Rptr. 33, 379 P.2d 513], their preference is evident. Some historical background may be of passing interest.

When Amaya came to this court, Justice Tobriner was required to recuse himself because he had written the opinion in that case for the Court of Appeal. Three active justices and a retired justice sitting pro tern, reached a contrary conclusion, with three active justices dissenting on the grounds urged below by Justice Tobriner. Thus when Dillon v. Legg arrived at the court several years later, it was inevitable that Justice Tobriner would write the opinion, this time for a majority of the full court.

My point is that had there been a full complement of justices on this court at the time, Amaya would have mirrored the rule that ultimately prevailed in Dillon v. Legg. It can be said, of course, that a court is a court, regardless of its permanent or temporary composition. I agree as a general proposition. However, the foregoing history is an explanation for the fortuitous and short life of Amaya.

Beginning with State Rubbish etc. Assn. v. Siliznoff (1952) 38 Cal.2d 330 [240 P.2d 282], a unanimous opinion written by Justice Tray nor 37 years ago, the majority recite a monotonous inventory of cases with which they find fault.1 For the past three decades apparently all the courts in tort cases have been out of step except the current majority.

Krouse v. Graham (1977) 19 Cal.3d 59 [137 Cal.Rptr. 863, 562 P.2d 1022], is high on their list as the source of the “roots of the uncertainty reflected by the instant case.” Yet the facts in Krouse caused little controversy. Justice Richardson declared for a court unanimous on the point: “the Dillon requirement of ‘sensory and contemporaneous observance of the accident’ does not require a visual perception of the impact causing the death or injury. In the matter before us, although [the husband] did not see *678[his wife] struck by defendant’s automobile, he fully perceived the fact that she had been so struck, for he knew her position an instant before the impact, observed defendant’s vehicle approach her at a high speed on a collision course, and realized that defendant’s car must have struck her. Clearly, under such circumstances [the husband] must be deemed a percipient witness to the impact causing [the wife’s] catastrophic injuries.” (Id. at p. 76.)

As Justice Richardson perceptively noted, Dillon called for sensory observance of the accident. That means all of the senses, not merely visual perception. The husband in Krouse was at the scene, heard the impact, and was himself seriously injured in the accident. How that case can be said to cause the “roots of uncertainty,” complained of by the majority, is difficult to comprehend.

In Archibald v. Braverman (1969) 275 Cal.App.2d 253, 256 [79 Cal.Rptr. 723], the Court of Appeal held that “the shock sustained by the mother herein was ‘contemporaneous’ with the explosion so as to satisfy the ‘observance’ factor.” Perhaps my colleagues would have interpreted the facts differently, but in view of the prevailing factual determination, I cannot perceive any rationale for the criticism of the majority.

The next criticism of the majority is directed to Justus v. Atchison (1977) 19 Cal.3d 564 [139 Cal.Rptr. 97, 565 P.2d 122]. Theirs is a strange comment, for in Justus the court found no liability. While the plaintiff was in the delivery room, he was shielded from observing the death of the fetus and did not learn the fact until told later by a physician. As we said, “he had been admitted to the theater but the drama was being played on a different stage.” (Id. at p. 584.)

Nazaroff v. Superior Court (1978) 80 Cal.App.3d 553, 566 [145 Cal.Rptr. 657], a Court of Appeal case, merely followed the instruction in BAJI No. 12.83 (6th ed. 1977), and concluded there were triable issues of fact as to whether the “physical harm to the mother resulted from an emotional shock proximately caused by the direct emotional impact from the contemporaneous observation of the immediate consequences of the defendants’ negligent act, which was the proximate cause of the injury and death of her son.” (Italics added.)

Some big guns are leveled at Molien v. Kaiser Foundation Hospitals (1980) 27 Cal.3d 916 [167 Cal.Rptr. 831, 616 P.2d 813, 16 A.L.R.4th 518], but they miss the target because of misunderstanding the relevant facts in Molien. After giving the wife a physical examination defendant incorrectly and negligently diagnosed her as having syphilis. As a consequence she was *679directed by the physician to require her husband to undergo physical tests to ascertain if he was the source of her purported infection. Those negligent acts by the physician as to both husband and wife were clearly physical in nature.

Why the majority seem to have some trouble with Molien limiting recovery to the “direct victim” of the doctor’s negligence is perplexing. To assure that recovery could not be claimed by an indirect victim or a mere casual bystander, the Molien court very strictly limited recovery to only a person who was directly and physically affected by the defendant’s negligence. In that instance it could be only the wife who was physically examined and the husband who was instructed to be physically examined. The court made it abundantly clear that “the alleged tortious conduct of defendant was directed to [the husband] as well as to his wife.” (27 Cal.3d at p. 923.) Thus both were direct victims.

That the majority fail to understand Molien is underscored by their comment that the parties did not suffer physical injury. The brutal fact is that the defendant committed malpractice in negligently diagnosing the wife in a physical examination as having a serious venereal disease and in directing the husband to undergo a physical examination. I cannot be as casual as the majority in concluding there is no physical aspect to negligent examinations for syphilis. Not physical? Do they believe such examinations are conducted over the telephone?

While the majority choose some partial quotations from Andalon v. Superior Court (1984) 162 Cal.App.3d 600 [208 Cal.Rptr. 899], that suggest the court had difficulty with the Molien rule, the fact is that the court properly analyzed the rule and found the plaintiff there to have a Molien cause of action. As the court pointed out, the end and aim of “the transaction with Mrs. Molien directly implicated the interests of Mr. Molien. The injury was not merely derivative of an injury to his spouse, i.e., indirect. Both Mr. and Mrs. Molien’s interests in harmonious relations with their spouses were impinged by the inherent influence of Kaiser’s misbehavior.” (Id. at p. 610, italics in original.)

The majority comment on Newton v. Kaiser Foundation Hospitals (1986) 184 Cal.App.3d 386 [228 Cal.Rptr. 890], is curious. That court clearly understood that “the doctor-patient relationship with the wife directly implicated the interest of the husband: in Molien, misdiagnosis of a venereal disease directly implicated the sexual component of the marriage in which Mr. Molien obviously had an interest . . . .” (Id. at pp. 391-392.) The Newton court properly observed that the task of distinguishing bystanders *680from direct victims is the task of lower courts. It found a cause of action under Molien, supra, 27 Cal.3d 916, had been stated.

The majority next offer lengthy criticism of Ochoa v. Superior Court (1985) 39 Cal.3d 159 [216 Cal.Rptr. 661, 703 P.2d 1], Ochoa understood Molien clearly: “In Molien defendant’s misdiagnosis was, by its very nature directed at both the wife and the husband.” (Id. at p. 172.) However, the Ochoa court found in that case “the duty was owed to [the child’s mother] as a percipient witness, not as a direct victim of negligence.” (Id. at p. 173.) Such a holding was consistent with Dillon’s requirements, supra, 68 Cal.2d 728.

Finally, the majority recite a card catalogue list of law review articles and broadly label them “critical” of the foregoing opinions. A careful reading of many of the articles fails to confirm that characterization.

For example, the note in (1983) 18 U.S.F. L.Rev. 145, pointed out that some commentators completely misunderstood the Molien case. Yet the answer was simple: “It is clear that a plaintiff should be classified as a direct victim rather than a bystander if he or she was the one toward whom the negligent conduct was directed.” (Id. at p. 166.)

And the note in (1981) 33 Hastings L.J. 291, 312, declared that “The Molien rule involves nothing more than a return to basic tort principles.”

Nolan and Ursin in (1982) 33 Hastings L.J. 583, appropriately titled their article Negligent Infliction of Emotional Distress: Coherence Emerging from Chaos. They wrote: “Molien convincingly demonstrates the superiority of seriousness, in comparison with the requirement of physical injury, as a safeguard against fraudulent claims in emotional distress cases.” (Id. at p. 610.)

The note in (1982) 18 Cal. Western L.Rev. 101, is hardly critical of the opinions of this court with which the majority now find fault. Said the author: “The court in Molien emphasized the fact that the plaintiff, Mr. Molien, was ‘himself a direct victim of the admittedly negligent act’—that is, Mr. Molien was required to undergo blood tests to determine whether he had contracted syphilis. The court contrasted that fact pattern with the ‘bystander scenario’ which applied to the Dillon facts.” (Id. at p. 112, fns. omitted.)

Some of the articles in the majority’s card index list not only support Dillon, supra, 68 Cal.2d 728, but recommend an expansion. For example, the test proposed in (1981) 54 So.Cal.L.Rev. 847, 867, is to allow “recovery *681for those types of emotional distress for which a reasonable person would be emotionally unprepared.”

The note, Dillon Revisited (1982) 43 Ohio St. L.J. 931, 940, contains a thoughtful discussion of psychological distress: “When first announced the requirement of physical manifestation was a proper limitation. It verified the severity of the emotional distress when medical science had not yet achieved sufficient sophistication to accomplish the task. With a better understanding of emotional responses, this need to show physical manifestation should diminish. The secondary responses basically are capable of objective verification. Thus, consistent with Molien, the issue should be one of proof that utilizes objective criteria.

“Once the court acknowledges that the issue is one of proof of emotional distress, it then must determine the type or degree of emotional distress that the law will recognize. Molien held that ‘some guarantee of genuineness’ of the evidence given should be required to sustain a claim. The court, however, was addressing only the issue of the evidence necessary to avoid a demurrer. Courts must set a standard that allows the fact finder to distinguish between compensable and noncompensable emotional distress.

“The Restatement (Second) of Torts [§ 46, com. j] suggests a possible starting point to begin this analysis. The following comment attempts to define actionable emotional distress:

“ ‘It is only when it is extreme that liability arises. Complete emotional tranquility is seldom attainable in this world, and some degree of transient and trivial emotional distress is part of the price of living among people. The law intervenes only when the distress is so severe that no reasonable man could be expected to endure it.’ ” (Italics in original, fns. omitted.)

In short, the majority’s broad generalization that commentators have been critical of the tort opinions of this court is unsupported when one goes beyond a superficial glance at the cited articles. For the most part, the authors understandably purported to distill the rules that should be applied to future fact situations. Some of their conclusions were sound, others dubious. That, of course, is not inconsistent with the function of law reviews: to stimulate legal thought.

I disagree with the majority opinion not merely for its conclusion—although I concur with Justice Broussard in that respect—but with its wholesale criticism of past opinions of this court and of the Courts of Appeal, some prevailing for three decades. Such callous disregard for the doctrine of stare decisis does not constructively serve the judicial process, nor does it *682contribute to the guidance of the bench and bar. As Justice Cardozo wrote in The Nature of the Judicial Process (1921) page 34, “Adherence to precedent must ... be the rule rather than the exception if litigants are to have faith in the even-handed administration of justice in the courts.”

BROUSSARD, J.

I dissent.

“[T]he problem [of negligent infliction of emotional distress] should be solved by the application of the principles of tort, not by the creation of exceptions to them. Legal history shows that artificial islands of exceptions, created from the fear that the legal process will not work, usually do not withstand the waves of reality and, in time, descend into oblivion.” (Dillon v. Legg (1968) 68 Cal.2d 728, 747 [69 Cal.Rptr. 72, 441 P.2d 912, 29 A.L.R.3d 1316] [Tobriner, J., admonishing this court not to do what the majority do today].)

The majority grope for a “bright line” rule for negligent infliction of emotional distress actions, only to grasp an admittedly arbitrary line which will deny recovery to victims whose injuries from the negligent acts of others are very real. In so doing, the majority reveal a myopic reading of Dillon v. Legg, supra, 68 Cal.2d 728. They impose a strict requirement that plaintiff be present at the scene of the injury-producing event at the time it occurs and is aware that it is causing injury to the victim. This strict requirement rigidities what Dillon forcefully told us should be a flexible rule, and will lead to arbitrary results. I would follow the mandate of Dillon and maintain that forseeability and duty determine liability, with a view toward a policy favoring reasonable limitations on liability. There is no reason why these general rules of tort law should not apply to negligent infliction of emotional distress actions.

I.

We held in Dillon that a mother who witnesses the negligent infliction of death or injury on her child may recover for the resulting emotional distress even though the mother does not fear imminent physical harm. We recognized that the primary consideration in finding liability was foreseeability (Dillon, supra, 68 Cal.2d at pp. 730-740) and rejected the “hopeless artificiality” of the zone-of-danger rule. (Id. at p. 733.)

The majority themselves note that “foreseeability of the injury [is] the basis of a negligent actor’s duty,” (maj. opn. at p. 654) and quote from Dillon that this issue must necessarily be adjudicated only upon “a case-by-case basis.” (Maj. opn. at p. 655, quoting Dillon, supra, 68 Cal.2d at p. 741.) “ ‘We cannot now predetermine defendant’s obligation in every situation by *683a fixed category; no immutable rule can establish the extent of that obligation for every circumstance of the future.’ ” (Maj. opn. at p. 654, quoting Dillon, supra, 68 Cal.2d at p. 740, italics added.)

Though Dillon made foreseeability its lodestar, it provided three factors for courts to consider in determining whether a negligent infliction of emotional distress cause of action was stated in a particular case: “(1) Whether plaintiff was located near the scene of the accident as contrasted with one who was a distance away from it. (2) Whether the shock resulted from a direct emotional impact upon plaintiff from the sensory and contemporaneous observance of the accident, as contrasted with learning of the accident from others after its occurrence. (3) Whether plaintiff and the victim were closely related, as contrasted with an absence of any relationship or the presence of only a distant relationship.” (Dillon, supra, 68 Cal.2d at pp. 740-741.)

Dillon denounced “artificial abstractions which bar recovery contrary to the general rules” of tort law, and emphasized that “ ‘mechanical rules of thumb which are at variance with these principles do more harm than good.’ [Citation.]” (68 Cal.2d at pp. 746-747.) However, some courts have rigidly and mechanically applied the Dillon guidelines—many times at the cost of injustice to a victim of a tortfeasor’s negligent act. Professor Diamond, although quoted extensively by the majority, warned against a strict application of the Dillon guidelines: “[C]ourts have applied the Dillon guidelines mechanically, viewing them as strict preconditions to recovery. This mechanical application has led to the erection of arbitrary limitations on recovery bearing little relation to the principles of foreseeability espoused so forcefully in Dillon. While in some instances mental distress is compensated, other equally foreseeable mental injuries are not. The result is feast or famine for the plaintiff depending on the fortuities of time, location, or characterization of the plaintiff as ‘direct’ or ‘indirect.’ ” (Diamond, Dillon v. Legg Revisited: Toward a Unified Theory of Compensating Bystanders and Relatives for Intangible Injuries (1984) 35 Hastings L.J. 477, 477-478.)

The majority ignore the fundamental mandate of Dillon to consider foreseeability and duty in finding liability. Their only justification for this and a strict rule that will limit liability at the cost of arbitrary results is an amorphous “policy” one. They ironically use the term “amorphous” to describe the concepts of foreseeability and duty and state that “[i]n the present context, however, we are concerned not with whether an injury is ‘foreseeable’ as a result of the negligent conduct. . . . [T]he court’s role in deciding whether a ‘duty’ to these persons should be recognized does not depend solely on the ‘foreseeability’ of the emotional distress, but on these policy considerations.” (Maj. opn. at p. 654, fn. 3, italics added.) “[I]t is clear that *684foreseeability of the injury alone is not a useful ‘guideline’ or a meaningful restriction on the scope of the NIED action. . . . [fl] [Pjolicy considerations justify restrictions on recovery for emotional distress notwithstanding the sometimes arbitrary result, and that the court has an obligation to establish those restrictions.” (Maj. opn. at pp. 663-664, italics added.) The majority admit their “policy” reasons are only a balance of “arbitrary lines which deny recovery to some victims whose injury is very real against that of imposing liability out of proportion to culpability for negligent acts,” with a view to the “administration of justice of clear guidelines.” (Id. at p. 664.)

For these reasons, the majority impose the strict requirement that plaintiff be present at the scene of the injury producing event at the time it occurs and is then aware that it is causing injury to the victim.1 They freely admit to “drawing arbitrary lines” but complain that it is “unavoidable if we are to limit liability and establish meaningful rules for application by litigants and lower courts.” (Maj. opn. at p. 666.) Thus what in Dillon were guidelines to assist courts in assessing liability become a tripartite test, which includes the above-mentioned strict and arbitrary requirement, and displaces the consideration of foreseeability.

Under the majority’s strict requirement, a mother who arrives moments after an accident caused by another’s negligence will not be permitted recovery. No matter that the mother would see her six-year-old son immediately after he was electrocuted, lying in a puddle of water in a dying state, gagging and choking in his own vomit, as in Hathaway v. Superior Court (1980) 112 Cal.App.3d 728 [169 Cal.Rptr. 435], No matter that the mother would be following her daughters’ car and would come upon the wreckage of the car “before the dust had settled” to find the mangled bodies of her daughters, who were dead or dying, as in Parsons v. Superior Court (1978) 81 Cal.App.3d 506, 509 [146 Cal.Rptr. 495],

The answer to the question of how this court should limit liability does not lie in the majority’s rigid application of Dillon and the toleration of arbitrary results that will flow therefrom. As the Wyoming Supreme Court suggested, in Gates v. Richardson (Wyo. 1986) 719 P.2d 193, the nature of the shock to be compensated requires a realistic approach to the contempo*685raneous-perception factor: “It is more than the shock one suffers when he learns of the death or injury of a child, sibling or parent over the phone, from a witness, or at the hospital. It is more than bad news. The kind of shock the tort requires is the result of the immediate aftermath of an accident. It may be the crushed body, the bleeding, the cries of pain, and in some cases, the dying words which are really a continuation of the event. The immediate aftermath may be more shocking than the actual impact. . . .” (Gates, supra, at p. 199; see also Dziokonski v. Babineau (1978) 375 Mass. 555 [380 N.E.2d 1295, 1302]; Comment, Dillon Revisited: Toward a Better Paradigm for Bystander Cases (1982) 43 Ohio St. L.J. 931, 948.) The court there held that a close relative generally should be permitted recovery if she “observed the serious bodily harm or death shortly after its occurrence but without material change in the condition and location of the victim.” (Gates, supra, 719 P.2d at p. 199; see also Portee v. Jaffee (1980) 84 N.J. 88 [417 A.2d 521] [recovery allowed to mother who arrived minutes after her son became wedged in an elevator shaft and was present and listened to his moans and cries during the futile four-and-one-half-hour struggle to free him before he died].)

The majority’s strict requirement does not simply comprise a “bright line” rule that rationally limits liability. It is arbitrary and will lead to unjust results. Dillon condemned the “hopeless artificiality” that the majority propounds, and decried the “artificial abstractions which bar recovery contrary to the general rules” of tort law. The requirement is exactly the “mechanical rules of thumb” that Dillon explicitly admonished us not to create. We should follow Dillon and its progeny and maintain the rational and traditional rule that reasonable foreseeability is the basis for determining liability. (68 Cal.2d at p. 740.) “In order to limit the otherwise potentially infinite liability which would follow every negligent act, the law of torts holds defendant amenable only for injuries to others which to defendant at the time were reasonably forseeable.” (Id. at p. 739, italics added.) As Dillon instructed, there is “no good reason why the general rules of tort law, including the concepts of negligence, proximate cause, and foreseeability, long applied to all other types of injury, should not govern the case now before us.” (Id. at p. 746.) Dillon's test of reasonable foreseeability “facilitates rational risk spreading and correlates liability with the risks that the defendant should expect.” (Diamond, supra, 35 Hastings L.J. at p. 500.)

II.

Of course I share the majority’s policy concern that tortfeasors not face unlimited liability for their negligent acts. As stated above, the Dillon court recognized foreseeability as a general limit on tort liability. The court stated that the purpose of the three guidelines was actually to limit a defendant’s *686liability to “injuries to others which to defendant at the time were reasonably foreseeable.” (Dillon, supra, 68 Cal.2d at p. 739; see Rabin, Tort Recovery for Negligently Inflicted Economic Loss: A Reassessment (1985) 37 Stan.L.Rev. 1513, 1524-1526.)

Although I disagree with the majority’s method of placing undue and what appears to be almost total reliance on a policy rationale, the Dillon guidelines also attempt to implement public policy in favor of reasonable limitations on liability. (See Ochoa v. Superior Court (1985) 39 Cal.3d 159, 180-181 [216 Cal.Rptr. 661, 703 P.2d 1] (Grodin, J., conc.); see also Pearson, Liability to Bystanders for Negligently Inflicted Emotional Harm—A Comment on the Nature of Arbitrary Rules (1982) 34 U.Fla.L.Rev. 477, 505-506.) The Dillon court did not intend the guidelines to be exclusive (68 Cal.2d at p. 740), and it specifically reserved the question of “whether, in the absence or reduced weight of some of the [three listed] factors, we would conclude that the accident and injury were not reasonably foreseeable and that therefore defendant owed no duty of care to plaintiff.” (Id. at p. 741).

To determine whether defendants owed plaintiff a duty of care in this case, I think it is fruitful to reexamine the second Dillon guideline in light of Rowland v. Christian (1968) 69 Cal.2d 108 [70 Cal.Rptr. 97, 443 P.2d 561, 32 A.L.R.3d 496], our leading case defining a defendant’s duty of care. In Rowland, decided just two months after Dillon, we held that in the absence of a statutory exception to the legislative mandate that all persons are liable for injuries caused by failure to exercise due care (Civ. Code, § 1714, subd. (a)), “no such exception should be made unless clearly supported by public policy. [Citations.] [1|] A departure from this fundamental principle involves the balancing of a number of considerations; the major ones are the foreseeability of harm to the plaintiff, the degree of certainty that the plaintiff suffered injury, the closeness of the connection between the defendant’s conduct and the injury suffered, the moral blame attached to the defendant’s conduct, the policy of preventing future harm, the extent of the burden to the defendant and consequences to the community of imposing a duty to exercise care with resulting liability for breach, and the availability, cost, and prevalence of insurance for the risk involved.” (Rowland, supra, 69 Cal.2d at pp. 112-113; see also Sun'n Sand, Inc. v. United California Bank (1978) 21 Cal.3d 671, 695 [148 Cal.Rptr. 329, 582 P.2d 920]; Tarasoff v. Regents of University of California (1976) 17 Cal.3d 425, 434-435 [131 Cal.Rptr. 14, 551 P.2d 334, 83 A.L.R.3d 1166]; Weirum v. RKO General, Inc. (1975) 15 Cal.3d 40, 46 [123 Cal.Rptr. 468, 539 P.2d 36].)

While our cases defining a bystander’s cause of action for negligent infliction of emotional distress consistently emphasize the first of the Row*687land factors—foreseeability of harm to plaintiff—discussion of the others has been limited. The second, fourth, and fifth factors may be disposed of quickly: certainty of injury is usually a jury question, particularly since we no longer require physical manifestations of mental distress (Hedlund v. Superior Court (1983) 34 Cal.3d 695, 706, fn. 8 [194 Cal.Rptr. 805, 669 P.2d 41, 41 A.L.R.4th 1063]); moral blame almost always militates in favor of recovery; and the policy of preventing future harm favors the plaintiff, but only slightly since, in most cases, any Dillon claim is simply added to the primary victim’s complaint. The third, sixth and seventh factors, however, merit more discussion.

This court has emphasized the importance of the third Rowland factor— nexus between defendant’s conduct and the risk of injury—in establishing limitations on recovery. In J’Aire Corp. v. Gregory (1979) 24 Cal.3d 799, 808 [157 Cal.Rptr. 407, 598 P.2d 60], we stated that case law “placefs] a limit on recovery by focusing judicial attention on the foreseeability of the injury and the nexus between the defendant’s conduct and the plaintiff’s injury.” There, we limited recovery for the tort of negligent interference with economic advantage “to instances where the risk of harm is foreseeable and is closely connected with the defendant’s conduct, where damages are not wholly speculative and the injury is not part of the plaintiff’s ordinary business risk.” (Ibid., italics added.)

The sixth and seventh Rowland factors—the burden on the defendant and the community, and the cost and availability of insurance—also merit further evaluation. Amici curiae2 contend that recovery in this case would mark an unwarranted expansion of Dillon, resulting in a new category of plaintiffs, fewer settlements, higher administrative costs and premiums, delays in payment, increased litigation, and higher awards. Amici insist that Dillon’s second guideline should be applied strictly, as a prerequisite for recovery. Apparently they have convinced the majority since the majority approvingly quote the reasoning in Borer v. American Airlines, Inc. (1977) 19 Cal.3d 441, 447 [138 Cal.Rptr. 302, 563 P.2d 858]: “We reasoned that we could not ‘ignore the social burden of providing damages . . . merely because the money to pay such awards comes initially from the “negligent” defendant or his insurer. Realistically the burden . . . must be borne by the public generally in increased insurance premiums or, otherwise, in the enhanced danger that accrues from the greater number of people who may choose to go without any insurance. We must also take into account the cost of administration of a system to determine and pay [the] awards; . . .’” (Maj. opn. at p. 665.)

*688The authorities upon which amici rely do not persuade me that Dillon has significantly contributed to any substantial increase in litigation and insurance premiums. Nor do I find any indication that other jurisdictions are retreating from Dillon (see 4 Speiser et al., The American Law of Torts (1987) §§ 16:25-16:26, pp. 1119-1126, and cases cited; Prosser & Keaton, Torts (5th ed. 1984) ch. 9, p. 366 et seq., and cases cited). As the Dillon court responded to the contention that otherwise meritorious claims should be barred out of fear of increases in the number of suits and of fraudulent claims: “ * “[We] should be sorry to adopt a rule which would bar all such claims on grounds of policy alone, and in order to prevent the possible success of unrighteous or groundless actions. Such a course involves the denial of redress in meritorious cases, and it necessarily implies a certain degree of distrust, which [we] do not share, in the capacity of legal tribunals to get at the truth in this class of claim.”’” (Dillon, supra, 68 Cal.2d at p. 744, quoting Hambrook v. Stokes Bros. (1925) 1 K.B. 141, quoting Dulieu v. White and Sons (1901) 2 K.B. 669, 681, opn. by Kennedy, J.; see also Molien v. Kaiser Foundation Hospitals (1980) 27 Cal.3d 916, 928-929 [167 Cal.Rptr. 831, 616 P.2d 813, 16 A.L.R.4th 518].)

I also do not believe courts lack the means to prevent unmeritorious cases from going to trial. As the case at bar demonstrates, trial courts are well aware of their duty to determine before trial whether the defendant could have owed the plaintiíf a duty of care under the facts. (See BAJI No. 12.84 (1986).)

III.

As for the instant case, I agree with Justice Benke of the Court of Appeal who in her concurring opinion indicated that the plaintiif mother’s claim should be permitted because “[u]nder current law, it cannot be said [plaintiff] is to be denied recovery because she did not see, hear or otherwise perceive the actual accident or injury. Case law permits the cause of action if she witnessed the immediate consequences of the event itself. It is a question of fact whether she did, and hence I would reverse the judgment below.” (Italics added.) Justice Benke declined to take a more liberal standard that “would extend liability far beyond existing limitations.” Her view is consistent both with Rowland, supra, 69 Cal.2d 108, and with the balance struck in Dillon, supra, 68 Cal.2d 728, between compensating legitimate claims and limiting liability.

I would conclude that the competing considerations cited in Rowland, (see ante, p. 686) as applied to this particular Dillon cause of action by the mother who sensorily perceived the immediate consequences of an accident involving her son, does not justify a departure from this state’s fundamental *689principle that a person is liable for injuries caused by his failure to exercise reasonable care under the circumstances. (Civ. Code, § 1714.) The majority’s strict requirement that plaintiff be present at the scene of the injury-producing event at the time it occurs and is then aware that it is causing injury to the victim will only bring about arbitrary results that will frustrate justice for victims of the negligent acts of others. We should apply the concepts of foreseeability and duty to negligent infliction of emotional distress actions, with a view toward a policy favoring reasonable restrictions on liability. This is a principled basis for determining liability and would also conform this area with other areas of negligence law.

The majority charge that, as former Supreme Court Justice Potter Stewart once said about obscenity, the Dillon guidelines mistakenly assumed that a court would know “duty” when it saw it. But more appropriate to the majority’s opinion and the arbitrary and unjust results it will soon engender is what Potter Stewart said, very simply, upon being appointed to the United States Supreme Court: “Fairness is what justice really is.” (The Young Justice (Oct. 20, 1958) Time, at p. 24.)

BREYER, Circuit Judge.

In December 1979 the ship Donau Maru spilled fuel oil into Boston Harbor. The spill prevented a different ship, the Tamara, from docking at a nearby berth. The Tamara had to discharge her cargo at another pier. In doing so, she incurred significant extra labor, fuel, transport and docking costs. The Tamara, her owners, and her charterers sued the Donau Maru and her owners in admiralty. Insofar as is here relevant, they claimed negligence and sought recovery of the extra expenses as damages. The district court denied recovery on the basis of the pleadings, citing as *51authority Petition of Kinsman Transit Co., 388 F.2d 821 (2d Cir.1968) (“Kinsman IT’). The plaintiffs have appealed. We believe the district court was correct, and we affirm its judgment, for three related sets of reasons.

1. Plaintiffs-appellants seek recovery' for a financial injury caused by defendants’ negligence. We assume that the injury was foreseeable. Nonetheless controlling case law denies that a plaintiff can recover damages for negligently caused financial harm, even when foreseeable, except in special circumstances. There is present here neither the most common such special circumstance — physical injury to the plaintiffs or to their property — nor any other special feature that would permit recovery. See pp. 55-56 infra.

The leading “pure financial injury” case is Robins Dry Dock & Repair Co. v. Flint, 275 U.S. 303, 48 S.Ct. 134, 72 L.Ed. 290 (1927) (Holmes, J.). Flint had chartered a ship, agreeing with its owners to have the ship docked for repairs every few months. During that time Flint would neither use the ship nor pay rent. The dry dock negligently damaged the ship’s propeller, delaying repairs, and causing Flint to lose the use of the ship for two weeks. The Court held that the ship’s owners might sue for negligent damage to the ship, but the charterer, suffering no physical injury to himself or to his property, could not do so. Justice Holmes wrote,

as a general rule, at least, a tort to the person or property of one man does not make the tortfeasor liable to another merely because the injured person was under a contract with that other unknown to the doer of the wrong. See Savings Bank v. Ward, 100 U.S. [(10 Otto) ] 195 [25 L.Ed. 621], The law does not spread its protection so far. A good statement, applicable here, will be found in Elliot Steam Tug Co., Ltd. v. The Shipping Controller, [1922] 1 K.B. 127, 139, 140. Byrd v. English, 117 Ga. 192 [43 S.E. 419]. The Federal No. 2, 21 F.(2d) 313.

Robins Dry Dock & Repair Co. v. Flint, 275 U.S. at 309, 48 S.Ct. at 135. The facts of Robins are strikingly similar to the present case. Just as the damaged propeller prevented Flint from using the ship, so the oil spill prevented the appellants from using the dock. The defendants in both eases were negligent. The injury in both cases (despite Holmes’ use of the word “unknown”) seems likely foreseeable. The harm in both cases was purely financial, without accompanying physical harm to person or property.

While we see three possible grounds of distinction, we do not believe them controlling. First, Flint alleged a specific contract with the shipowner, while the appellants here do not allege a contract with the dock. The authority that Justice Holmes says contains a “good statement” of the legal principle does not, however, turn so much on the existence of a formal contract as on the existence of limitations upon tort recovery for financial injury, see Elliot Steam Tug Co., Ltd. v. The Shipping Controller, supra; Byrd v. English, 117 Ga. 192, 43 S.E. 419 (1902). Moreover, the present appellants must have had a “right” to use the dock; and interference with that “right” caused the loss. It is difficult in this instance to see why the technical legal label applied to that right should make a legal difference.

Second, Flint apparently sued for lost profits. Appellants here sue for expense. Typically, an extra expense is more easily proved than a lost profit. Again, however, Justice Holmes points to authority that includes both added expenses and lost profits, see Savings Bank v. Ward, 100 U.S. (10 Otto) 195, 25 L.Ed. 621 (1879); The Federal No. 2, 21 F.2d 313 (2d Cir.1927). Other cases, decided both before and after Robins, for the most part make no such distinctions. E.g., Hercules Carriers, Inc. v. Florida, 720 F.2d 1201 (11th Cir.1983), aff'd by an equally divided court, 728 F.2d 1359 (1984) (en banc); Marine Navigation Sulphur Carriers, Inc. v. Lone Star Industries, Inc., 638 F.2d 700 (4th Cir.1981); Ultramares Corp. v. Touche, 255 N.Y. 170, 174 N.E. 441 (1931); Con *52 necticut Mutual Life Insurance Co. v. New York & New Haven Railroad, 25 Conn. 265 (1856); Caltex Oil v. The Dredge “Willemstad”, 11 A.L.R. 227 (Austl.H.C. 1976); Chargeurs Reunis Compagnie Francaise de Navigation a Vapeur v. English & American Shipping Co., 9 Lloyd’s List, L.R. 464 (1921); Cattle v. Stockton Waterworks Co., L.R. 10 Q.B. 453 (1875).

Third, one might claim that Robins is wrong or out of date and, therefore, that the inferior courts are free to “limit” it through a narrow reading. For reasons set out below, however, we think the principles underlying Robins remain legally sound insofar as they place plaintiffs like those before us “outside the scope” of those to whom defendant owes a legal duty of care. Cf. Sinram v. Pennsylvania Railroad, 61 F.2d 767 (2d Cir.1932) (Hand, J.); Palsgraf v. Long Island Railroad, 248 N.Y. 339, 162 N.E. 99 (1928) (Cardozo, J.).

In Kinsman II, supra, another leading case, the Second Circuit more recently came to the same result as the Supreme Court in Robins Dry Dock. In Kinsman II, defendant’s ship broke loose from her moorings, crashed into another ship, then into a bridge, and, subsequently, with the help of ice floes, created a barrier that prevented other ships from moving upstream to unload cargo. The Second Circuit held that the financial injuries suffered by these other ships — extra unloading expenses — were too “remote” to warrant recovery. The court analogized the careening ship to a negligent driver who crashes into another car in a tunnel. Such a driver, though negligent, is not thought liable for all the inevitable (and foreseeable) financial losses resulting from the delays that he has caused. We can find in the case before us no relevant distinction from Kinsman II.

Appellants argue that Kinsman II raises a factual issue of “foreseeability.” We read Kinsman II, however, not as saying that the injury, as a matter of fact, was unforeseeable but, rather, as drawing a legal line, based on considerations of policy, cf. Sin-ram, supra, that forbids compensation for certain types of foreseeable, negligently caused, financial injury. The details of the Kinsman II accident were unusual, but the precise details of many, or most, accidents cannot be foreseen in advance. Rather, foreseeability is a matter of a class, or type, of harm. And, in terms of perfectly traditional, reasonably specific, commonsense classes, the Kinsman II blockade, delay, and extra cost were foreseeable. Still more so are the extra costs involved in the analogous Kinsman II example, the extra trucking costs arising from tunnel accident delays. Viewing the legal implications of Kinsman II in this way, we cannot distinguish a barrier created by an oil spill from a barrier created by a careening ship, each of which increases unloading costs by requiring other ships to go elsewhere. It is still more difficult for us to distinguish this case from delays created by, say, tunnel accidents, which are likely to mean extra cost for truckers, shippers and merchants, all of which are foreseeable.

A third major, and recent, decision is that of the Fifth Circuit in Louisiana ex rel. Guste v. M/V Testbank, 752 F.2d 1019 (5th Cir.1985). The majority in that case sets forth a reasonably clear rule, which says that one who suffers only financial loss, unaccompanied by physical injury, cannot recover damages from a negligent defendant, whether or not the financial loss is foreseeable. The holding is consistent with the way in which most commentators have characterized pre-existing case authority. See James, Limitations on Liability for Economic Loss Caused by Negligence: A Pragmatic Appraisal, 25 Vand.L.Rev. 43 (1972) (“[A] plaintiff may not recover for his economic loss resulting from bodily harm to another or from physical damage to property in which he has no proprietary interest.”); Atiyah, Negligence and Economic Loss, 83 L.Q.Rev. 248, 249 (1967) (no “remedy for purely financial loss in the law of negligence to a person who had suffered such loss as a result of damage to, or the destruction of, a chattel in which he had no proprietary or possessory interest”). The holding also invalidates the authority of *53several Fifth Circuit cases on which appellants here rely, and which in any event expressed a minority view. See Micmar Motorship Corp. v. Cabaneli Naviera, S.A., 477 F.Supp. 45 (E.D.La.1979); vacated and remanded, 620 F.2d 298 (5th Cir. 1980); In re Lyra Shipping Co., 860 F.Supp. 1188 (E.D.La.1973); In re China Union Lines, Ltd, 285 F.Supp. 426 (S.D. Tex.1967). We need not embrace the whole of Testbank’s rule in order to recognize that it constitutes additional, strong case law against allowing appellants to recover here.

We note that the courts in the cases we have cited have not always used the same legal terminology to describe their conclusions. One might, for example, use Judge Hand’s language in Sinram, and say that plaintiffs like those before us are persons to whom appellee owes no “duty of care.” Alternatively, one could use the slightly more obscure “proximate cause” terminology, and say that plaintiffs’ injuries are too “remote.” One could also appeal to historic legal terminology, and describe plaintiffs as suffering damnum absque injuria. Regardless of descriptive terminology, the holdings of these major cases are the same. They refuse to hold a defendant liable for negligently caused financial harm without accompanying physical injury or other special circumstances, see pp. 55-56, infra, none of which is present here. E.g., Cattle v. Stockton Waterworks Co., supra (no recovery for builder’s contractual losses caused by tunnel obstruction); Weller & Co. v. Foot & Mouth Disease Research Institute, 1 Q.B. 569 (1966) (no recovery by cattle auctioneers for losses caused by virus escaped from research institute); Robins Dry Dock & Repair Co. v. Flint, supra; Petition of Kinsman Transit Co., supra.

2. Before affirming the district court on the basis of existing precedent, we have asked ourselves whether that precedent remains good law. After all, courts have sometimes departed from past legal precedent where changing circumstances viewed in light of underlying legal policy deprived that precedent of sound support. See, e.g., Glanzer v. Shepard, 233 N.Y. 236, 135 N.E. 275 (1922) (Cardozo, J.); MacPherson v. Buick Motor Co., 217 N.Y. 382, 111 N.E. 1050 (1916) (Cardozo, J.). Here, however, precedent seems, at least in general, to rest on a firm policy foundation. The same judges who removed other recovery limitations left this one firmly in place, compare Glanzer v. Shepard, supra, and MacPher-son v. Buick Motor Co., supra, with Ul-tramares Corp. v. Touche, supra (Cardozo, J.). Much written commentary, which for a time in the 1940’s attacked the limitation, see W. Prosser, Handbook on the Law of Torts 993 (1st ed. 1941), Carpenter, Interference with Contract Relations, 41 Harv.L.Rev. 728 (1928), has more recently supported it, while offering a variety of refinements. See Rizzo, A Theory of Economic Loss in the Law of Torts, 11 J.Leg. Stud. 281 (1982) (advocating recovery where its value outweighs “channeling” costs and litigation costs); Bishop, Economic Loss in Tort, 2 Oxford J.Leg.Stud. 1 (1982); Atiyah, supra (advocating minor changes); James, supra. But see Coval, Smith & Rush, “Out of the Maze": Towards a “Clear Understanding” of the Test for Remoteness of Damages in Negligence, 61 Can.B.Rev. 559 (1983) (advocating recovery when damage falls within a class of foreseeable damages); Seidelson, Some Reflections on Proximate Cause, 19 Duq. L.Rev. 1 (1980) (advocating test based on inquiry into relationship between defendant’s conduct and plaintiff’s commercial activities). Indeed, foreign civil law systems, which do not distinguish between financial and physical harm, nonetheless seem to have devised other rules that lead to similar results. See Marshall, Inability for Pure Economic Loss Negligently Caused — French and English Law Compared, 24 Int’l & Comp.Q. 748 (1975).

The cases and commentaries, in making a plausible argument that existing precedent rests on sound considerations of policy, also reveal that these considerations are highly general and abstract. Judges lack the empirical information that would allow measurement of their force or magnitude; *54and, in particular, judges cannot apply these considerations on a case by case basis.

We have concluded that we could not find for appellants here without ignoring these policy considerations, or at a minimum, applying them case by case, a practice that we believe would be unwise. A brief description of the kinds of policy considerations typically advanced as supporting existing law (perhaps with a few modifications) will show why these considerations have led us to conclude that we must adhere to prior precedent.

First, cases and commentators point to pragmatic or practical administrative considerations which, when taken together, offer support for a rule limiting recovery for negligently caused pure financial harm. The number of persons suffering foreseeable financial harm in a typical accident is likely to be far greater than those who suffer traditional (recoverable) physical harm. The typical downtown auto accident, that harms a few persons physically and physically damages the property of several others, may well cause financial harm (e.g., through delay) to a vast number of potential plaintiffs. The less usual, negligently caused, oil spill foreseeably harms not only ships, docks, piers, beaches, wildlife, and the like, that are covered with oil, but also harms blockaded ships, marina merchants, suppliers of those firms, the employees of marina businesses and suppliers, the suppliers’ suppliers, and so forth. To use the notion of “foreseeability” that courts use in physical injury cases to separate the financially injured allowed to sue from the financially injured not allowed to sue would draw vast numbers of injured persons within the class of potential plaintiffs in even the most simple accident eases (unless it leads courts, unwarrantedly, to narrow the scope of “foreseeability” as applied to persons suffering physical harm). That possibility — a large number of different plaintiffs each with somewhat different claims — in turn threatens to raise significantly the cost of even relatively simple tort actions. See Rizzo, supra. Yet the tort action is already a very expensive administrative device for compensating victims of accidents. Indeed, the legal time, the legal resources, the delay appurtenant to the tort action apparently mean that on average the victim recovers only between 28 and 44 cents of every dollar paid by actual or potential defendants, while victims who insure themselves directly recover at least between 55 and 66 cents of each premium dollar earned by insurance companies and between 85 and 90 cents of every dollar actually paid out to investigate and satisfy claims. See O’Connell, An Alternative to Abandoning Tort Liability: Elective No-Fault Insurance for Many Kinds of Injuries, 60 Minn.L.Rev. 503-12 (1976); Best’s Aggregates and Averages: Property-Casualty 4-5 (1984). See also P. Munch, Costs and Benefits of the Tort System if Viewed as a Compensation system (Rand 1977); O’Connell, Offers That Can’t Be Refused: Foreclosure of Personal Injury Claims by Defendants’ Prompt Tender of Claimants’ Net Economic Losses, 77 Nw.U.L.Rev. 589, 593-96 (1982); James, supra at 52 & n. 40. The added cost of the increased complexity, while unknowable with precision, seems likely significant.

At the same time many of the “financially injured” will find it easier than the “physically injured” to arrange for cheaper, alternative compensation. The typical “financial” plaintiff is likely to be a business firm that, in any event, buys insurance, and which may well be able to arrange for “first party” loss compensation for foreseeable financial harm. Other such victims will be able to sue under tort principles, for they will suffer at least some physical harm to their property. Still others may have contracts with, or be able to contract with, persons who can themselves recover from the negligent defendant. A shipowner, for example, might contract with a dock owner for “inaccessibility” compensation; and the dock owner (whose pier is physically covered with oil) might recover this compensation as part of its tort damages. See Rizzo, supra at 293. Of course, such a tort suit, embodying a “contract-defined” injury, may *55still raise difficult foreseeability questions, cf. Hadley v. Baxendale, 9 Exch. 341 (1854). But the bringing of one suit, instead of several, still makes the litigation as a whole a less costly compensation device. See Rizzo, supra; Atiyah, supra. Finally, some of the “financially injured” will have suffered harm that is, in any event, noncompensable because it is not sufficiently distinguishable from minor harms typical of ordinary living. Cf. Fletcher, Fairness and Utility in Tort Theory, 85 Harv.L.Rev. 536, 543 (1972). The law does not compensate, for example, the cost of unused baseball tickets or flowers needed for apology regardless of the cause of the delay that foreseeably led to the added expense. Insofar as these considerations, taken as a whole, support recovery limitations, they reflect a fear of creating victim compensation costs that, from an administrative point of view, are unnecessarily high. See Stevenson v. East Ohio Gas Co., 73 N.E.2d 200, 202 (Ohio 1946).

A second set of considerations focuses on the “disproportionality” between liability and fault. Those who argue “disproportion-ality” are not reiterating the discredited nineteenth century view that tort liability would destroy industry, investment, or capitalism. See F. Wharton, A Suggestion as to Causation 11 (1874); Horwitz, The Doctrine of Objective Causation, in The Politics of Law: A Progressive Critique 201 (D. Kairys, ed. 1982). Rather, they recognize that tort liability provides a powerful set of economic incentives and disincentives to engage in economic activity or to make it safer, see generally G. Calabresi, The Costs of Accidents (1970). And, liability for pure financial harm, insofar as it proved vast, cumulative and inherently unknowable in amount, could create incentives that are perverse.

Might not unbounded liability for foreseeable financial damage, for example, make auto insurance premiums too expensive for the average driver? Is such a result desirable? After all, the high premiums would reflect not only the costs of the harm inflicted; they would also reflect administrative costs of law suits, jury verdicts in uncertain amounts, some percentage of unbounded or inflated economic claims, and lessened incentive for financial victims to avoid harm or to mitigate damage. Given the existing liability for physical injury (and for accompanying financial injury), can one say that still higher premiums are needed to make the public realize that driving is socially expensive or to provide greater incentive to drive safely (an incentive that risk spreading through insurance dilutes in any event, see Shavell, On Liability and Insurance, 13 Bell J. of Econ. 120 [1982])?

These considerations, of administrability and disproportionality, offer plausible, though highly abstract, “policy” support for the reluctance of the courts to impose tort liability for purely financial harm. While they seem unlikely to apply with equal strength to every sort of “financial harm” claim, their abstraction and generality, along with the comparative inaccessibility of the empirical information needed to confirm or to invalidate them, mean that courts cannot weigh or apply them case by case. What, for example, in cases like this one, are the added administrative costs involved in allowing all persons suffering pure financial harm to sue the shipowner instead of “channeling” suits (perhaps via contract) through traditionally injured plaintiffs? Is there a problem of “dispro-portionality”? How far, for example, would additional, unbounded, pure financial loss liability for negligently caused oil spills, when added to the already large potential traditional liability, affect the type of insurance carried, the incentive to mitigate losses, the incentive to transport oil safely, the likelihood that shippers will use pipelines and domestic wells instead of ships and foreign wells, and the consequences of these and other related changes? We do not know the answers to these questions, nor can judges readily answer them in particular cases.

It does not surprise us then that, under these circumstances, courts have neither enforced one clear rule nor considered the *56matter case by case. Cf. Michelman, Property, Utility, and Fairness: Comments on the Ethical Foundations of “Just Compensation”Law, 80 Harv.L.Rev. 1165, 1249-53 (1967). Rather, they have spoken of a general principle against liability for negligently caused financial harm, while creating many exceptions. See, e.g., 1) Newlin v. New England Telephone & Telegraph Co., 316 Mass. 234, 54 N.E.2d 929 (1944) (accompanying physical harm); 2) Lumley v. Gye, 2 El. & Bl. 216, 118 Eng.Rep. 749 (1853); Beekman v. Marsters, 195 Mass. 205, 80 N.E. 817 (1907) (intentionally caused harm); 3) Dalton v. Meister, 52 Wis.2d 173, 188 N.W.2d 494 (defamation), cert. denied, 405 U.S. 934, 92 S.Ct. 947, 30 L.Ed.2d 810 (1971); Systems Operations, Inc. v. Scientific Games Development Corp., 555 F.2d 1131 (3d Cir. 1977) (injurious falsehood); 4) Hitaffer v. Argonne Co., 183 F.2d 811 (D.C.Cir.) (loss of consortium), cert. denied, 340 U.S. 852, 71 S.Ct. 80, 95 L.Ed. 624 (1950); 5) Chicago, Duluth & Georgia Bay Transit Co. v. Moore, 259 Fed. 490 (6th Cir.) (medical costs of injured plaintiff paid by a different family member), cert. denied, 251 U.S. 553, 40 S.Ct. 118, 64 L.Ed. 411 (1919); 6) Hedley Byrne Co. Ltd. v. Heller & Partners Ltd., A.C. 465 (1964) (negligent misstatements about financial matters); 7) Jones v. Waterman S.S. Corp., 155 F.2d 992 (3d Cir.1946) (master-servant); 8) Western Union Telegraph Co. v. Mathis, 215 Ala. 282, 110 So. 399 (1926) (telegraph-addressee); 9) Union Oil Co. v. Oppen, 501 F.2d 558 (9th Cir.1974) (commercial fishermen as special “favorites of admiralty”). These exceptions seem designed to pick out broad categories of cases where the “administrative” and “disproportionality” problems intuitively seem insignificant or where some strong countervailing consideration militates in favor of liability. Thus an award of financial damages to one also caused physical harm does not threaten proliferation of law suits, for the plaintiff could sue anyway (for physical damages). Financial harm awards to family members carry with them an obvious self-limiting principle (as perhaps does awarding such damages to fishermen, as “favorites” of admiralty). Awarding damages for financial harm caused by negligent misrepresentation is special in that, without such liability, tort law would not exert significant financial pressure to avoid negligence; a negligent accountant lacks physically harmed victims as potential plaintiffs. See Atiyah, supra at 264.

We need not explore the exceptions in detail. Rather, we here simply point to the existence of plausible reasons underlying the judicial hesitance to award damages in a case like this one, and the need to consider exceptions by class rather than case by case. The existence of these factors, together with our comparative inability to evaluate their empirical significance, cautions us against departing from prior law.

3. We note that several dissenting Fifth Circuit judges in Testbank, supra, have advocated abandonment of traditional tort rules in this area and the adoption of a new rule that might allow recovery in this case. They would adopt a principle used to identify the class of private plaintiffs permitted to sue in “public nuisance” cases; and they would allow that class of persons to sue negligent defendants to recover foreseeable financial harm. To be more specific, they would allow “particularly damaged” plaintiffs to sue for financial harms suffered “beyond the general economic dislocation.” “Particular damages” would have to be

different in kind and degree from those suffered by the general public____ [Such damage typically arises] when the plaintiff is prevented from performing a specific contract, or is put to additional expense in performing it.... although plaintiffs who are delayed by a public nuisance cannot recover money for the delay itself ..., they can recover for actual additional expenses, such as extra fuel, additional crew expenses, and greater demurrage charges.

Louisiana ex rel. Guste v. M/V Testbank, 752 F.2d at 1047-49 (Wisdom, J., dissenting).

We do not believe we can adopt this general principle and apply it here for the *57following reasons. First, if meant literally, it would amount to a near reversal of the general judicial principle that (with exceptions) forbids recoveries for negligently caused purely financial losses. Traditionally, the “public nuisance” suit involved an intentional tort — say, a deliberate blocking of a highway. The permissible plaintiffs included all those whose damage was different from that of the ordinary traveler. This class consisted of all those suffering definite financial harm, namely virtually all business travelers. See W. Prosser & W. Keeton, Handbook on the Law of Torts § 90 (1984). It is one thing to allow these persons to sue in the narrow, and relatively unusual, instance of an intentionally caused nuisance; it is quite another to allow them to sue whenever they are negligently, and foreseeably, injured. To do so — to depart from Robins, Kinsman II and others— would simply create the problems discussed in Part 2 above.

Second, the Testbank minority could escape the general problems discussed in Part 2 only by narrowing the class of plaintiffs — by eliminating some who suffer the type of financial harm that would qualify them as plaintiffs in public nuisance cases. The Fifth Circuit dissenters may want to do this, for they say they would allow delayed ships, bait shops, tackle shops, dry docks, repair services, boat charterers, to sue for financial harm caused by a negligent oil spill. But, they would not allow seafood restaurants or, presumably, grocery stores or owners of other businesses in the area to sue though they may have suffered equivalent, and equally foreseeable, harm. We fear, however, the ad hoe quality of the examples. We recognize the difficulty of avoiding a measure of judicial fiat in the tort area. Cf. Sinram, supra. And, we understand the dissenters’ efforts to broaden liability in oil spill cases, while maintaining workable, administrable limits. But we do not see in the dissent a principle that would do so — that would broaden liability somewhat without running squarely into the practical problems outlined in Part 2. At best, the dissenters seem to have created a principle that would have to be applied case by case — yet this individualized type of application raises the difficulties we have previously discussed.

Third, we do not see how to apply the dissenters’ principle outside the area of oil spills. Their reasoning and examples, applied to tunnel accidents, would allow truckers who ordinarily use the tunnel, firms whose employees use it regularly to commute, and other businesses, to recover for foreseeable financial losses from negligent drivers. While it may make sense to allow such persons to recover from one who intentionally builds a barricade in the tunnel, to allow such suits against a negligent driver raises the problems of Part 2 in full force. We do not believe it practical for courts to distinguish between, say, oil spill accidents and tunnel accidents, depending solely on the industrial context of the accident.

We conclude that we should follow existing precedent that requires us, as a matter of law, to deny recovery. That precedent is reasonably consistent. It is supported by plausible considerations of tort policy. Appellants have failed to bring themselves within any recognized class or category in which financial damages are already allowed, and appellants have failed to provide convincing reasons for the creation of any new exception or class that would work to their legal benefit.

For these reasons, the judgment of the district court is

Affirmed.

Opinion of the Court by Cuéllar, J.

*634**883*394This case concerns a massive, months-long leak from a natural gas storage facility located just outside Los Angeles. According to the allegations before us, the accident severely harmed the economy of a nearby suburb. We must decide if local businesses - none of which allege they suffered personal injury or property damage - may recover in negligence for income *635lost because of the leak. Our decision turns on whether the entity that allegedly caused the leak had a tort duty to guard against what we and other courts have termed "purely economic losses."

The businesses argue that they deserve compensation for such losses, that the entity responsible must bear the full costs of its alleged negligence so tort law can play its essential role of forcing people and organizations to take sufficient account of the risks they generate, and that courts can sensibly apportion liability under these circumstances within meaningful limits. Tort law indeed lies in the heartland of our common law system. It serves society's interest in allocating risks and costs to those who can better prevent them, and it provides aggrieved parties with just compensation. But a proper assessment of competing considerations in light of our precedent suggests, *395and the extent of consensus across other jurisdictions confirms, that claims for purely economic losses suffered from mere proximity to an industrial accident create intractable line-drawing problems for courts. So the claims before us are best not treated as compensable in negligence.

We therefore affirm the judgment of the Court of Appeal.

I.

Because this case comes to us at the demurrer stage, we take as true all properly pleaded material facts - but not conclusions of fact or law. ( Centinela Freeman Emergency Medical Associates v. Health Net of California, Inc. (2016) 1 Cal.5th 994, 1010, 209 Cal.Rptr.3d 280, 382 P.3d 1116 ( Centinela ).)

A.

Near the northwestern corner of Los Angeles lies Porter Ranch, a residential neighborhood home to some 30,000 people. Southern California Gas Company (SoCalGas) stores vast amounts of natural gas in an underground facility in the hills surrounding the community. Known today as the "Aliso Facility," that subterranean storage site was once an oil reservoir. It was repurposed about 40 years ago for its present use. SoCalGas supplies over 21 million people with natural gas from its four storage facilities, but the Aliso Facility is the company's largest. It holds up to 80 billion cubic feet of natural gas, which SoCalGas pumps underground at high pressure into more than 100 "injection wells." Because natural gas is odorless, SoCalGas adds a nausea-causing chemical to the gas so that people notice when a leak happens.

In October 2015, a leak happened - and people noticed. An uncontrolled flow of natural gas from the Aliso Facility coated nearby neighborhoods in an oily mist. At its peak, the leak released some 55 tons of natural gas every hour. Porter Ranch residents reported unpleasant odors, headaches, dizziness, and respiratory problems. In addition to those symptoms, students at local schools complained of nosebleeds and vomiting.

That November, the Los Angeles County health department directed SoCalGas to establish a relocation program available to Porter Ranch residents who lived within a five-mile radius of the leak site. The Department of Conservation's Division of Oil, Gas, and Geothermal Resources required SoCalGas to provide real-time data about the leak, **884as well as a timeline for stopping it. A month later, with the flow of gas slowing but still significant, the Los Angeles County Board of Education decided to relocate students and *396staff from two Porter Ranch schools for the duration of the academic year. And a month after that, SoCalGas expanded its relocation program, *636citing complaints of poor air quality from people living outside the initial five-mile boundary. About 15,000 people were relocated in total, scattering to locations dozens - and in some cases hundreds - of miles away.

SoCalGas finally got the leak under control in February 2016 - four months after detecting it. All told, about 100,000 tons of natural gas escaped the Aliso Facility, releasing enough greenhouse gases into the atmosphere to erase several years' worth of efforts to combat climate change in California.

B.

Plaintiffs are Porter Ranch area businesses seeking to represent a class of "[a]ll persons and entities conducting business within five miles of the [Aliso] Facility from October 23, 2015 to [the] present."1 They allege that SoCalGas's negligence caused the leak. The resulting relocation of many Porter Ranch residents devastated the local economy: by depriving local businesses of customers, the environmental disaster cost local businesses considerable earnings.

That harm, Plaintiffs maintain, is ongoing. Sales at businesses of all stripes declined sharply, and in many cases, stayed down. Enrollment at a local martial arts center, Plaintiff King Taekwondo, nosedived during the leak and has not recovered. The same was true of a neighborhood day care, Plaintiff Polonsky Family Day Care. Restaurants, gas stations, and pharmacies were affected, too. So were beauty salons, doctor's offices, party suppliers, and a photography store.

With the en masse relocation of Porter Ranch residents and the diminution in property values caused by the leak, home mortgage lenders and home improvement businesses suffered economically as well. Plaintiff First American Realty saw clients get cold feet, loans fall out of escrow, and sales tumble. A local contractor's business dropped by 25 percent, as customers moved away or decided against home improvements for the time being. "Since the onset of the gas leak," in other words, business operations throughout Porter Ranch "have either halted or slowed substantially" - and Plaintiffs "have not yet recovered from the blow to their bottom lines."

Yet no named plaintiff in this action alleges personal injury or property damage. Accordingly, Plaintiffs acknowledge they are suing SoCalGas to recover solely for the income they lost because of the leak.

*397C.

SoCalGas demurred, arguing that Plaintiffs' negligence claims failed as a matter of law because Plaintiffs were seeking to recover for purely economic losses. Overruling the demurrer, the trial court explained that companies "must face the full cost of accidents" they create, or else "they will underinvest in precautions." The trial court acknowledged that economic losses not flowing from conventional injury to person or property, such as physical damage, are ordinarily not recoverable in tort - and that the Court of Appeal had so held in Adams v. Southern Pacific Transportation Co. (1975) 50 Cal.App.3d 37, 123 Cal.Rptr. 216 ( Adams ) on facts with some similarities to those here.2 But *637the trial court questioned the wisdom of that rule and reasoned that Adams was no longer good law after our later decision in J'Aire Corp. v. Gregory (1979) 24 Cal.3d 799, 157 Cal.Rptr. 407, 598 P.2d 60 ( J'Aire ).

After SoCalGas petitioned for a writ of mandate, the Court of Appeal granted the petition and reversed the trial court. ( Southern California Gas Leak Cases (2017) 18 Cal.App.5th 581, 583-584, 227 Cal.Rptr.3d 117.) The Court of Appeal explained that, under California law, it is "not presumed"

**885that a defendant owes a duty of care to guard against economic losses unaccompanied by injury to person or property. ( Id. at p. 591, 227 Cal.Rptr.3d 117.) And without a "special relationship" between the plaintiff and the defendant stemming in this context from a "transaction," the Court of Appeal reasoned, California law did not permit recovery for the purely economic losses sought by Plaintiffs in this case. ( Id. at p. 591, 227 Cal.Rptr.3d 117.) The Court of Appeal also took the view that our decision in J'Aire had not rejected Adams in its entirety but instead disapproved Adams only "insofar as [it] held a plaintiff can never recover purely economic losses based on a defendant's negligent conduct." ( Id. at p. 592, 227 Cal.Rptr.3d 117, italics added.) Because Plaintiffs disclaimed any desire to further amend their complaint, the Court of Appeal directed the trial court to sustain SoCalGas's demurrer without leave to amend. ( Id . at p. 595, 227 Cal.Rptr.3d 117.)

II.

Recovery in a negligence action depends as a threshold matter on whether the defendant had " 'a duty to use due care toward an interest of [the plaintiff's] that enjoys legal protection against unintentional invasion.' " ( Centinela , supra , 1 Cal.5th at p. 1012, 209 Cal.Rptr.3d 280, 382 P.3d 1116, quoting *398Bily v. Arthur Young & Co. (1992) 3 Cal.4th 370, 397, 11 Cal.Rptr.2d 51, 834 P.2d 745 ( Bily ).) We review de novo whether this " 'essential prerequisite' " to recovery is satisfied.3 ( Centinela , at pp. 1010, 1012, 209 Cal.Rptr.3d 280, 382 P.3d 1116.)

The issue here is whether SoCalGas - separate from other legal and practical reasons it had to prevent injury of any kind to the public - had a tort duty to guard against negligently causing what we and others have called "purely economic loss[es]." ( Centinela , supra , 1 Cal.5th at p. 1013, 209 Cal.Rptr.3d 280, 382 P.3d 1116 ; see also 532 Madison Avenue Gourmet Foods, Inc. v. Finlandia Center, Inc. (N.Y. 2001) 96 N.Y.2d 280, 727 N.Y.S.2d 49, 750 N.E.2d 1097, 1102 ( 532 Madison ).) We use that term as a shorthand for "pecuniary or commercial loss that does not arise from actionable physical, emotional or reputational injury to persons or physical injury to property." (Dobbs, An Introduction to Non-Statutory Economic Loss Claims (2006) 48 Ariz. L.Rev. 713 (Dobbs).) And although SoCalGas of course had a tort duty to guard against the latter kinds of injury, we conclude it had no tort duty to guard against purely economic losses.

A.

In California, the "general rule" is that people owe a duty of care to avoid causing harm to others and that they are *638thus usually liable for injuries their negligence inflicts. ( Cabral v. Ralphs Grocery Co. (2011) 51 Cal.4th 764, 771, 122 Cal.Rptr.3d 313, 248 P.3d 1170 ( Cabral ).) Under Civil Code section 1714, subdivision (a), "[e]veryone is responsible ... for an injury occasioned to another by his or her want of ordinary care or skill in the management of his or her property or person, except so far as the latter has, willfully or by want of ordinary care, brought the injury upon himself or herself." So at least in cases involving traditionally compensable forms of injury - like physical harm to person or property - we presume the defendant owed the plaintiff a duty of care and then ask whether the circumstances "justify a departure" from that usual presumption. ( Cabral , at p. 771, 122 Cal.Rptr.3d 313, 248 P.3d 1170.) In Rowland v. Christian (1968) 69 Cal.2d 108, 70 Cal.Rptr. 97, 443 P.2d 561 ( Rowland ), we identified several factors that, among others, may bear on that question: (1) "the foreseeability of harm to the plaintiff," (2) "the degree of certainty that the plaintiff suffered injury," (3) "the closeness of the connection between the defendant's conduct and the injury suffered," (4) "the moral blame attached to the defendant's conduct," (5) "the policy of preventing future harm," (6) "the extent of the burden to the defendant and consequences to the community of imposing a duty to exercise care with resulting liability for breach," and (7) "the availability, cost, and prevalence of insurance for the *399risk involved." ( **886Id. at p. 113, 70 Cal.Rptr. 97, 443 P.2d 561.) At core, though, the inquiry hinges not on mere rote application of these separate so-called Rowland factors, but instead on a comprehensive look at the " 'the sum total' " of the policy considerations at play in the context before us. ( Parsons v. Crown Disposal Co. (1997) 15 Cal.4th 456, 472, 63 Cal.Rptr.2d 291, 936 P.2d 70 ( Parsons ), quoting Ballard v. Uribe (1986) 41 Cal.3d 564, 572, fn. 6, 224 Cal.Rptr. 664, 715 P.2d 624 ; see also T.H. v. Novartis Pharmaceuticals Corp. (2017) 4 Cal.5th 145, 164, 226 Cal.Rptr.3d 336, 407 P.3d 18.)

What Civil Code section 1714 does not do is impose a presumptive duty of care to guard against any conceivable harm that a negligent act might cause. No one doubts, for example, that a child suffers gravely when an accident permanently disables her parent. But in Borer v. American Airlines, Inc. (1977) 19 Cal.3d 441, 138 Cal.Rptr. 302, 563 P.2d 858 ( Borer ), we nevertheless treated the prospect of a child recovering for loss of consortium in precisely that circumstance as "a wholly new cause of action," rather than a presumptively viable one. ( Id. at p. 447, 138 Cal.Rptr. 302, 563 P.2d 858.) And we refused to recognize such a novel - though quite sympathetic - claim for emotional harm largely because that claim, unlike a loss of consortium claim brought by a spouse, threatened indeterminate and disproportionate liability. ( Id. at pp. 448-449, 453, 138 Cal.Rptr. 302, 563 P.2d 858 ; see also Thing v. La Chusa (1989) 48 Cal.3d 644, 666-668, 257 Cal.Rptr. 865, 771 P.2d 814 ( Thing ) [strictly cabining recovery for negligent infliction of emotional distress to ensure meaningful limits on liability].) Disabled parents, after all, have parents of their own, along with "brothers, sisters, cousins, inlaws, friends, colleagues, and other acquaintances who," in some way, may "be deprived of [their] companionship." ( Borer , at p. 446, 138 Cal.Rptr. 302, 563 P.2d 858.) In declining to impose a tort duty to guard against such harms, we noted in Borer the "overwhelming approval" our conclusion enjoyed in other jurisdictions and rejected the argument that our analysis was somehow "inconsistent with the principles of tort law" established in our own. ( Id. at pp. 449-450, 138 Cal.Rptr. 302, 563 P.2d 858 ; see also *639Thing , at p. 668, fn. 11, 257 Cal.Rptr. 865, 771 P.2d 814.)

Plaintiffs do cite several cases where we presumed the defendant owed the plaintiff a duty of care and then asked whether the circumstances warranted a departure from that baseline presumption. But unlike Borer and Thing , every one of those cases involved a traditionally compensable form of harm: personal injury. (See Vasilenko v. Grace Family Church (2017) 3 Cal.5th 1077, 1082, 224 Cal.Rptr.3d 846, 404 P.3d 1196 [plaintiff was struck by a car when crossing a public street shortly after leaving defendant's parking lot]; Kesner v. Superior Court (2016) 1 Cal.5th 1132, 1140-1141, 210 Cal.Rptr.3d 283, 384 P.3d 283 ( Kesner ) [employee's household members were exposed to asbestos, causing personal injury and death]; Cabral , supra , 51 Cal.4th at p. 769, 122 Cal.Rptr.3d 313, 248 P.3d 1170 [plaintiff's husband died after colliding with a truck owned by defendant]; John B. v. Superior Court (2006) 38 Cal.4th 1177, 1181-1183, 45 Cal.Rptr.3d 316, 137 P.3d 153 [defendant infected plaintiff with HIV];

*400Parsons , supra , 15 Cal.4th at p. 460, 63 Cal.Rptr.2d 291, 936 P.2d 70 [defendant's truck startled plaintiff's horse, causing the plaintiff to fall to the ground].) And in Rowland itself, a faulty faucet in the defendant's home mangled tendons and nerves in the plaintiff's hand. ( Rowland , supra , 69 Cal.2d at p. 110, 70 Cal.Rptr. 97, 443 P.2d 561.) So yes, we have frequently begun our analysis by presuming a duty of care. But we have not universally done so.

A case in point is liability in negligence for purely economic losses, which is "the exception, not the rule" under our precedents. ( Quelimane Co. v. Stewart Title Guaranty Co. (1998) 19 Cal.4th 26, 58, 77 Cal.Rptr.2d 709, 960 P.2d 513 ( Quelimane ).) And that holds true even though Civil Code section 1714 does not, by its terms, "distinguish among injuries to one's person, one's property or one's financial interests." ( J'Aire , supra , 24 Cal.3d at p. 806, 157 Cal.Rptr. 407, 598 P.2d 60 ; see Centinela , supra , 1 Cal.5th at p. 1013, 209 Cal.Rptr.3d 280, 382 P.3d 1116 ; Dobbs, supra , 48 Ariz. L.Rev. at p. 713 [explaining that "[n]egligently inflicted economic loss that results from some other kind of injury may be recoverable, but recovery for stand-alone economic loss is frequently rejected"].)

**887The primary exception to the general rule of no-recovery for negligently inflicted purely economic losses is where the plaintiff and the defendant have a "special relationship." ( J'Aire , supra , 24 Cal.3d at p. 804, 157 Cal.Rptr. 407, 598 P.2d 60.) What we mean by special relationship is that the plaintiff was an intended beneficiary of a particular transaction but was harmed by the defendant's negligence in carrying it out. Take, for example, Biakanja v. Irving (1958) 49 Cal.2d 647, 320 P.2d 16 ( Biakanja ). There, we held that the intended beneficiary of a will could recover for assets she would have received if the notary had not been negligent in preparing the document. ( Id. at pp. 650-651, 320 P.2d 16.) A special relationship existed between the intended beneficiary and the notary in Biakanja , we emphasized, because "the 'end and aim' of the transaction" between the nonparty decedent and the notary was to ensure that the decedent's estate passed to the intended beneficiary. ( Id. at p. 650, 320 P.2d 16.)

For similar reasons, in J'Aire we held that a special relationship existed between a restaurant operator and a contractor hired by a third-party property owner to renovate the space rented by the restaurant operator. ( J'Aire , supra , 24 Cal.3d at pp. 804-805, 157 Cal.Rptr. 407, 598 P.2d 60.) So when the contractor negligently failed to complete the construction work on time, the restaurant operator could recover *640purely economic losses it suffered as a result.4 ( J'Aire , at pp. 804-805, 157 Cal.Rptr. 407, 598 P.2d 60.) *401Discerning whether there is a special relationship justifying liability of this sort can nonetheless be a subtle enterprise. In both Biakanja and J'Aire we emphasized that our duty determination rested not just on (i) "the extent to which the transaction was intended to affect the plaintiff," but also on a subset of the Rowland factors relevant to the circumstances before us in those cases: (ii) "the foreseeability of harm to the plaintiff," (iii) "the degree of certainty that the plaintiff suffered injury," (iv) "the closeness of the connection between the defendant's conduct and the injury suffered," (v) "the moral blame attached to the defendant's conduct," and (vi) "the policy of preventing future harm." ( J'Aire , supra , 24 Cal.3d at p. 804, 157 Cal.Rptr. 407, 598 P.2d 60, citing Biakanja , supra , 49 Cal.2d at p. 650, 320 P.2d 16.)

Our subsequent decision in Bily , however, underscored for negligence cases involving purely economic losses what is true of all negligence cases. Deciding whether to impose a duty of care turns on a careful consideration of the " 'the sum total' " of the policy considerations at play, not a mere tallying of some finite, one-size-fits-all set of factors. ( Bily , supra , 3 Cal.4th at p. 397, 11 Cal.Rptr.2d 51, 834 P.2d 745, quoting Dillon v. Legg (1968) 68 Cal.2d 728, 734, 69 Cal.Rptr. 72, 441 P.2d 912 ( Dillon ).) In Bily , investors in a failed company sued the company's auditor for financial losses they allegedly suffered due to the auditor's negligent preparation of a public report on the company's financial well-being. (See Bily at pp. 376-379, 11 Cal.Rptr.2d 51, 834 P.2d 745.) We rejected those claims. (See id. at p. 376, 11 Cal.Rptr.2d 51, 834 P.2d 745.) Despite acknowledging that financial losses to investors from negligently prepared audit reports are "certainly" foreseeable,5 we held that an auditor "owes no general duty of care regarding the conduct of an audit to persons other than the client." ( Bily , at pp. 376, 398, 11 Cal.Rptr.2d 51, 834 P.2d 745.)

In requiring more than mere foreseeability for imposing a duty of care in Bily , we appreciated the need to safeguard the efficacy of tort law by setting meaningful limits on liability. ( Bily , supra , 3 Cal.4th at pp. 398-399, 11 Cal.Rptr.2d 51, 834 P.2d 745.) Citing decisions from our court limiting recovery for emotional harms based on similar concerns, **888we explained that although foreseeability " 'may set tolerable limits for most types of physical harm, it provides virtually no limit on liability for nonphysical harm.' " ( Id. at p. 398, 11 Cal.Rptr.2d 51, 834 P.2d 745, quoting Thing , supra , 48 Cal.3d at p. 663, 257 Cal.Rptr. 865, 771 P.2d 814.) After all, on " 'clear judicial days' " courts " 'can foresee forever.' " ( Bily , at p. 399, 11 Cal.Rptr.2d 51, 834 P.2d 745, quoting Thing , at p. 668, 257 Cal.Rptr. 865, 771 P.2d 814.) So although exposure to liability often provides an important incentive for parties to internalize the social costs of their actions, we were *641concerned that allowing the countless people who rely on public audit reports to recover "pure economic loss suffered" due to a shoddy audit would "raise[ ] the spectre of vast numbers of *402suits and limitless financial exposure." ( Bily , at p. 400, 11 Cal.Rptr.2d 51, 834 P.2d 745.) The resulting universe of potential claims would not only raise difficult line-drawing questions for courts, it might deter socially beneficial behavior. ( Id. at pp. 400, 404, 11 Cal.Rptr.2d 51, 834 P.2d 745.) That result was unacceptable. ( Id. at p. 406, 11 Cal.Rptr.2d 51, 834 P.2d 745.) We therefore limited auditor liability to claims for negligent misrepresentation brought by plaintiffs who - like the plaintiffs in Biakanja and J'Aire -were "specifically intended beneficiaries" of the defendant's conduct. ( Bily , at pp. 406-407, 11 Cal.Rptr.2d 51, 834 P.2d 745.)

To be sure, several additional considerations cut further in favor of strictly circumscribing recovery in Bily . In the audit context, "[t]he client typically prepares [the] financial statements" on which the auditor relies in preparing a report - and that report "is not a simple statement of verifiable fact" but instead "a professional opinion based on numerous and complex factors." ( Bily , supra , 3 Cal.4th at pp. 399-400, 11 Cal.Rptr.2d 51, 834 P.2d 745.) The plaintiffs in Bily were also particularly "sophisticated" and had "efficient means of self-protection," such as diversifying their investment portfolios or conducting their own due diligence. ( Id. at p. 406, 11 Cal.Rptr.2d 51, 834 P.2d 745.) More fundamentally, purely economic losses flowing from a financial transaction gone awry - which were at issue in Biakanja , J'Aire , Bily , and our other negligence cases to date about purely economic losses6 - "are primarily the domain of contract and warranty law or the law of fraud, rather than of negligence." ( Aas v. Superior Court (2000) 24 Cal.4th 627, 636, 101 Cal.Rptr.2d 718, 12 P.3d 1125 ( Aas ), superseded by statute on other grounds as stated in Rosen v. State Farm General Ins. Co. (2003) 30 Cal.4th 1070, 1079-1080, 135 Cal.Rptr.2d 361, 70 P.3d 351 ( Rosen ).)

We nonetheless acknowledged in Bily the "need to limit liability for [purely] economic loss[es]" even in the absence of those additional considerations. ( Bily , supra , 3 Cal.4th at p. 400, fn. 11, 11 Cal.Rptr.2d 51, 834 P.2d 745.) In doing so, we pointed to a hypothetical scenario similar in many ways to the case now before us. We considered a situation where "a defendant negligently causes an automobile accident that blocks a major traffic artery such as a bridge or tunnel." ( Ibid. ; see also Kinsman Transit Co. (2d Cir. 1968) 388 F.2d 821, 825, fn. 8 [using a similar illustration]; Rabin, Tort Recovery for Economic Loss: A Reassessment (1985) 37 Stan. L.Rev. 1513, 1536-38 [same].) That defendant would of course be liable for "personal injuries and property damage suffered in such an accident." ( Bily , at pp. 400-401, fn. 11, 11 Cal.Rptr.2d 51, 834 P.2d 745.) But would "any court," we continued, "allow recovery by the myriad [other] third parties who might claim [purely] economic losses because the bridge or *403tunnel" was blocked? ( Id. at p. 401, fn. 11, 11 Cal.Rptr.2d 51, 834 P.2d 745.) Based on concerns about limitless liability and unending *642litigation, as well as on long-standing legal consensus, we considered that prospect "doubtful." (Ibid .) **889B.

What we recognized in Bily fits with numerous decisions from other jurisdictions - as well as the Restatement of Torts. That consensus cuts sharply against imposing a duty of care to avoid causing purely economic losses in negligence cases like this one: where purely economic losses flow not from a financial transaction meant to benefit the plaintiff (and which is later botched by the defendant), but instead from an industrial accident caused by the defendant (and which happens to occur near the plaintiff).

1.

Concerned about line-drawing problems and potentially overwhelming liability, courts across the country have rejected recovery for purely economic losses stemming from man-made calamity. Take the New York Court of Appeals' decision in 532 Madison . There, part of a 39-story office tower collapsed, shutting down more than a dozen bustling blocks of midtown Manhattan for several weeks. (See 532 Madison , supra , 727 N.Y.S.2d 49, 750 N.E.2d at p. 1099.) Among the plaintiffs in 532 Madison were local businesses who alleged that would-be customers "were unable to gain access to their stores" due to the disaster, forcing the plaintiffs to shut down for an extended period of time. ( Id. , 727 N.Y.S.2d 49, 750 N.E.2d at pp. 1099-1100.) They sued on behalf of themselves and "all other business entities" operating within the affected city blocks. ( Ibid. )

The plaintiffs in 532 Madison sought compensation for the income they lost from the tower collapse. The New York Court of Appeals responded by declining to hold "that a landowner owes a duty to protect an entire urban neighborhood against purely economic losses." ( 532 Madison , supra , 727 N.Y.S.2d 49, 750 N.E.2d. at pp. 1102.) It instead "limit[ed] the scope of defendants' duty to those who ha[d], as a result of th[e] [collapse], suffered personal injury or property damage." ( Id. 727 N.Y.S.2d 49, 750 N.E.2d at p. 1103.) The court explained that this limitation provided "a principled basis for reasonably apportioning liability" that was necessary to prevent potentially crushing liability to "an indeterminate group in the affected areas" who could prove "financial losses directly traceable to the" collapse. ( Ibid. ) Adopting that rule, the court added, was what "historically courts ha[d] done" with similar negligence claims. ( Ibid. )

Indeed: the Illinois Supreme Court, for example, reached the same result for similar reasons in litigation flowing from a flood caused by human error that inundated downtown Chicago in 1992. (See *404In re Chicago Flood Litigation (1997) 176 Ill.2d 179, 223 Ill.Dec. 532, 680 N.E.2d 265, 268, 276.) Consider also the West Virginia Supreme Court of Appeals' decision in Aikens v. Debow (2000) 208 W.Va. 486, 541 S.E.2d 576, the Iowa Supreme Court's decision in Nebraska Innkeepers, Inc. v. Pittsburgh-Des Moines Corp. (Iowa 1984) 345 N.W.2d 124, the Massachusetts Supreme Judicial Court's decision in Stop & Shop Companies, Inc. v. Fisher (1983) 387 Mass. 889, 444 N.E.2d 368, and the Seventh Circuit's decision applying Wisconsin law in Leadfree Enterprises, Inc. v. U.S. Steel Corp. (7th Cir. 1983) 711 F.2d 805. Those cases all concerned bridge accidents similar to the hypothetical we discussed in Bily - and those cases all rejected attempts by affected businesses to recover in negligence for purely economic losses resulting from those accidents. (See Aikens , at pp. 579, 589, ; Nebraska Innkeepers , at pp. 125, 128 ; *643Stop & Shop , at pp. 371-373 ; Leadfree Enterprises , at pp. 806, 809 ; see also American Petroleum and Transport, Inc. v. City of New York (2d Cir. 2013) 737 F.3d 185, 187, 196-197 [rejecting under federal maritime law recovery for purely economic losses stemming from a drawbridge malfunction].) Among their concerns were the endless "ripple effects of a negligence claim based upon pure economic loss." ( Aikens , at p. 591, ; see also Dundee Cement Co. v. Chemical Laboratories, Inc. (7th Cir. 1983) 712 F.2d 1166, 1172 [opining that allowing recovery in negligence for purely economic losses may unleash "multiversant possibilities" for litigation that "are staggering to the imagination"].)

Similar rationales buttressed the Court of Appeals for the District of Columbia's decision in Aguilar v. RP MRP Washington Harbour, LLC (D.C. 2014) 98 A.3d 979 ( Aguilar ) and the Connecticut Supreme Court's decision in **890Lawrence v. O & G Industries, Inc. (2015) 319 Conn. 641, 126 A.3d 569. Like the California Court of Appeal's decision in Adams , those cases rejected claims for lost wages brought by employees whose workplaces were forced to close by a man-made disaster - a flood in Aguilar and an explosion in Lawrence . (See Aguilar , at pp. 981, 983 ; Lawrence , at pp. 571, 585,.) In fact, more than a half century ago we ourselves approved a decision from an intermediate appellate court in Ohio that arrived at the same conclusion on very similar facts - another explosion causing the closure of a nearby workplace. (See Fifield Manor v. Finston (1960) 54 Cal.2d 632, 636, 7 Cal.Rptr. 377, 354 P.2d 1073, citing Stevenson v. East Ohio Gas Co. (Ohio Ct.App. 1946) 73 N.E.2d 200, 201-204.)

Federal courts sitting in admiralty have dealt with industrial accidents perhaps most like the one before us: maritime spills of oil and other pollutants. Leaving aside one narrow exception not applicable here, they too have refused to impose a duty of care to guard against purely economic losses. To wit: in State of Louisiana ex rel. Guste v. M/V TESTBANK (5th Cir. 1985) 752 F.2d 1019 ( Testbank ), two ships collided in the Mississippi River *405Gulf. ( Id. at p. 1020.) Some 12 tons of a toxic chemical called pentachlorophenol rushed into the water and caused the suspension of fishing, shrimping, and other maritime activities across four hundred square miles of marsh and waterways. ( Ibid. ) Among the plaintiffs were businesses like boat rental operators, seafood restaurants, and tackle and bait shops. ( Id. at pp. 1020-1021.) They sued to recover "for economic loss unaccompanied by physical damage" that the spill had inflicted. ( Id. at p. 1021.)

The Fifth Circuit rejected those claims. (See Testbank , supra , 752 F.2d at pp. 1028-1029.) The court echoed concerns about "wave upon wave of successive economic consequences" and stressed that "[t]hose who would delete the requirement of physical damage have no rule or principle to substitute," save perhaps letting the trier of fact determine case-by-case, whim-by-whim which claims for purely economic losses warrant recovery. ( Id. at p. 1028.) The Fifth Circuit further explained that "to the extent that economic analysis" mattered, it favored rejecting recovery for purely economic losses. ( Id. at p. 1029.) That was because defendants in industrial accident cases - despite their frequently deep pockets - will have more difficulty obtaining third-party insurance coverage against purely economic losses than will individual plaintiffs seeking comparable first-party insurance. (See ibid. ) Defendants' potential liability for purely economic losses in such cases is massive and indeterminate. ( Ibid. ) So insurance companies cannot feasibly offer them comprehensive coverage - or even fix a sensible premium based on actuarial measurement.

*644( Ibid. ) Plaintiffs' "own potential losses," by contrast, "are finite and readily discernible." ( Ibid. ) They can therefore obtain insurance to cover them - perhaps relatively cheaply. ( Ibid. ; see also Posner, Common-Law Economic Torts: An Economic and Legal Analysis (2006) 48 Ariz. L.Rev. 735, 737-738.)

Faced with an oil spill diverting a container ship at substantial cost, the First Circuit in Barber Lines A/S v. M/V Donau Maru (1st Cir. 1985) 764 F.2d 50 denied recovery in negligence for those purely economic losses. The First Circuit's analysis in many ways mirrored the Fifth Circuit's reasoning in Testbank. (See Barber Lines , at pp. 50-52.) Through the pen of then-Judge Breyer, the First Circuit explained that the "number of persons suffering foreseeable financial harm in a typical accident" - like a car crash - "is likely to be far greater than those who suffer traditional (recoverable) physical harm." ( Id. at p. 54.) And when it comes to industrial accidents, that proliferation of potential liability for purely economic losses is even more dramatic. (See ibid. ) An oil spill, for instance, "foreseeably harms" not just those whose property is "covered with oil," but also "blockaded ships, marina merchants, suppliers of those firms, the employees of marina businesses and suppliers, the suppliers' suppliers, and so forth." ( Ibid. ) That indeterminate liability, the First Circuit continued, made third-party insurance coverage against purely economic losses less feasible than first-party insurance. ( Ibid. )

*406It also risked over-deterring socially productive activities. ( Id. at p. 55.) And unable to "distinguish between, say, oil spill accidents and tunnel accidents," the First Circuit rejected the idea of adopting different duty rules depending **891on the particular "industrial context" at issue. ( Id. at p. 57.)

The narrow exception mentioned earlier, to which we now turn, does not help Plaintiffs. Applying maritime law and California law alike in Union Oil Co. v. Oppen (9th Cir. 1974) 501 F.2d 558, the Ninth Circuit held that commercial fishermen could recover in negligence for the "diminution of aquatic life" caused by an oil spill. ( Id. at pp. 563, 570.) But that was because theirs was "a pecuniary loss of a particular and special nature" grounded in the time-worn principle that "seamen are the favorites of admiralty." ( Id. at pp. 567, 570 ; see also Curd v. Mosaic Fertilizer, LLC (Fla. 2010) 39 So.3d 1216, 1228.) Recovery in Union Oil was therefore tightly circumscribed: it was "limited to the class of commercial fishermen" whose livelihoods depend on the flourishing of aquatic life in the commons of the sea and thus did not include, for example, recreational fisherman whose " 'Sunday piscatorial pleasure' " depended on angling in the same waters. ( Union Oil , at p. 570, quoting Oppen v. Aetna Ins. Co. (9th Cir. 1973) 485 F.2d 252, 260.) Indeed, the Ninth Circuit further cautioned that its narrow holding based on unique features of the maritime context did "not open the door to claims" from others "whose economic or personal affairs were discommoded by the oil spill." ( Union Oil , at p. 570.) Not "every decline in the general commercial activity of every business" nearby, the court reasoned, was "a legally cognizable injury for which the defendants may be responsible." ( Ibid. ) So in Union Oil - as in every case discussed so far - recovery in negligence for purely economic losses was the exception, not the rule.

Against all these decisions, only the New Jersey Supreme Court's opinion in People Express Airlines, Inc. v. Consolidated Rail Corp. (1985) 100 N.J. 246, 495 A.2d 107 ( People Express ) cuts definitively the other way. In People Express , a railroad fire forced a nearby terminal at Newark International Airport to shut down for *645twelve hours - a terminal housing the plaintiff's business. (See id. at p. 108.) The plaintiff brought a negligence claim for income lost as a result - a claim the New Jersey Supreme Court permitted to proceed. ( Id. at pp. 108, 116.) The court imposed a tort duty to guard against purely economic losses where there is "an identifiable class with respect to whom [the] defendant knows or has reason to know are likely to suffer such damages from its conduct." ( Id. at p. 116.) The court stressed "that an identifiable class of plaintiffs is not simply a foreseeable class of plaintiffs" - such as happenstance bystanders - but instead a class that is "particularly foreseeable in terms of the type of persons or entities comprising the class, the certainty or predictability of their presence, the approximate numbers of those in the class, as well as the type of economic expectations disrupted." ( Ibid. ) *407Yet decades after the demise of the airline that gave the case its name, People Express remains "a lonely outpost." (Rabin, Respecting Boundaries and the Economic Loss Rule in Tort (2006) 48 Ariz. L.Rev. 857, 858.) Its relatively ad hoc standard, embodied in a fact-intensive " 'particular foreseeability' " test, has been avoided by other courts with - as one scholar put it - "a striking degree of unanimity." (Ibid. ; see also 532 Madison , supra , 727 N.Y.S.2d 49, 750 N.E.2d at p. 1103 [declining to follow People Express ]; Aguilar , supra , 98 A.3d at p. 984 [same].)7

2.

Little wonder the Restatement of Torts takes the dominant view. Although acknowledging that "[d]uties to avoid the unintentional infliction of economic loss" exist in certain recognized circumstances, the latest Restatement provides that there is "no general duty to avoid the unintentional infliction of economic loss on another." (Rest.3d, Torts, Liability for Economic Harm (Tent. Draft. No. 1, Apr. 4, 2012) § 1 (Restatement T.D. 1).)

**892In justifying that position, the Restatement echoes widespread judicial concern that purely economic losses "proliferate more easily than losses of other kinds" and "are not self-limiting" in the same way. (Restatement T.D. 1, § 1, com. c.) Those characteristics, the Restatement explains, threaten "liabilities that are indeterminate and out of proportion to [a defendant's] culpability," and with them "exaggerated pressure to avoid an activity altogether." (Restatement T.D. 1, § 1, com. c.) For centuries, in fact, similar concerns have justified strict limits on private recovery for a public nuisance. (See 4 Blackstone, Commentaries 167 [noting that a public nuisance is usually not privately actionable because "it would be unreasonable to multiply suits by giving every man a separate right of action"]; accord Rest.3d Torts, Liability for Economic Harm (Tent. Draft. No. 2, Apr. 7, 2014) § 8, com. c. (Restatement T.D. 2); Civ. Code, § 3493 [originally enacted in 1872].)

Only when the foregoing considerations are "weak or absent" - such as in Biakanja and J'Aire , but not in Bily - does a duty to guard against purely economic losses exist under the Restatement approach to negligence claims. (See Restatement T.D. 1, supra , § 1, com. d; see also *646Restatement T.D. 2, supra , § 7, com. a [using 532 Madison 's facts and the *408court's holding as an illustration of the Restatement view].) But in this case, as in the mine run of man-made disaster cases, those rationales apply with full force.

C.

The allegations before us underscore the ineluctable difficulty associated with imposing a duty to guard against purely economic losses in negligence cases like this one. It may be possible to quantify the profits any one business lost because of an industrial accident, but imposing such a duty would nevertheless create line-drawing problems across - quite literally - space and time.8 So although our duty determination must ultimately "occur[ ] at a higher level of generality" than would a jury's analysis of fact-intensive issues like breach and causation ( Kesner , supra , 1 Cal.5th at p. 1144, 210 Cal.Rptr.3d 283, 384 P.3d 283 ), we examine some particulars of Plaintiffs' claims to illustrate those two sets of persistent line-drawing problems.

1.

We lack clear spatial bounds within which to cabin claims like those asserted here.

This case does not involve a so-called special relationship under our precedents. Plaintiffs concede - as they must - that their only relevant ties to SoCalGas are having the misfortune of operating near the Aliso Facility. Accordingly, they propose to limit the class they seek to represent based on geographic proximity alone. Putative class members here are businesses operating "in the area within five miles" of the leak, a space which Plaintiffs characterize as "the precise area from which residents were evacuated."

What is far from clear is why the five-mile line means anything. Others beyond that boundary were also affected. We discern no compelling basis for us to let a business operating 4.9 miles away recover its lost profits but deny such recovery to another business operating 5.1 miles away. Nor is it clear what we should do about a third business operating 6 miles away whose balance sheet was hit just as hard by the leak and ensuing evacuation - or perhaps a fourth business operating 10 miles away, whose income depends on supplying Porter Ranch businesses or offering services to its residents. Similar questions arise regarding employees of businesses operating within *409the five-mile mark but who live outside it - or even well outside it. (This is Los Angeles we're talking about.) They might have lost wages during a temporary business slowdown - or even lost their jobs if their employers were forced to cut back permanently. Those employees might not be included in Plaintiffs' proposed class, but their losses are foreseeable, too. They could come to court **893next in lawsuits of their own. And if we were to permit recovery for purely economic losses in this case, we don't see how we could justify denying it in that one.

Most of the foregoing difficulties emerge even when an evacuation zone is set in stone. But here the lines drawn were traced in sand. Plaintiffs' own complaint acknowledges that, a few weeks after the leak was detected, the evacuation zone was extended beyond the initial five-mile mark. Why businesses operating outside the original *647boundary but inside the new one should not get to recover their equally real and foreseeable financial losses we do not know.

Using the boundary of an evacuation zone as a liability line might not just lack predictability. In certain circumstances, it could also inject a dangerous incentive into disaster response efforts. Consider how a company taking after Justice Oliver Wendell Holmes's infamous "bad man" - that is, a company that "cares nothing for an ethical rule" and thus cares "only for the material consequences" of its actions - might respond to an evacuation zone rule. (Holmes, The Path of the Law (1897) 10 Harv. L.Rev. 457, 459, 461 ; see also Exxon Shipping Co. v. Baker (2008) 554 U.S. 471, 501-502, 128 S.Ct. 2605, 171 L.Ed.2d 570 [looking to "Justice Holmes's 'bad man' " in a tort case brought under federal maritime law].) If companies face liability in negligence only for traditionally compensable harms, their financial incentive with respect to evacuations points in one direction: caution. To minimize the risk of, and their liability for, harm to people and property, companies under that legal regime may indeed seek (or at least not try to avoid) large evacuation zones. But imposing liability for purely economic losses - bounded only by the size of the evacuation zone - would blunt that otherwise sharp financial incentive for caution. The larger the evacuation zone, the larger a company's potential liability for purely economic losses. So under that rule, a company taking after Justice Holmes's bad man would face a newly vexing cost-benefit analysis: will an evacuation prevent enough physical damage to offset the purely economic losses it is sure to cause? The calculus of a ruthlessly self-interested company would tend to prioritize maximizing its own economic return, not minimizing the risk of harm to people and property. And where it expects an evacuation to harm its bottom line, our proverbial "bad company" might take steps to confine or prevent one.

Such steps might include, most obviously, overt pressure on public officials to roll back or eliminate a proposed evacuation. But that's not the only possibility.

*410Public officials must often rely on company information to know what scale of risk the community faces. Case in point: during the very disaster at issue here, authorities allegedly demanded from SoCalGas real-time data about the leak - and a timeline for ending it. So public officials might simply be kept in the dark. That's bad enough when, as here, the public health concerns are things like nausea and nosebleeds. But it would be much worse when, on different facts, the stakes are life and death.

Nor is it always simple to decide what counts as an evacuation, or to resolve claims for purely economic losses where the disaster in question never triggered an evacuation. Some evacuations are mandatory, others are voluntary. And sometimes public officials issue public safety warnings without telling people to leave the area. An evacuation zone rule would require a coherent way to decide which sorts of government action count and which ones don't. We do not see one. What is more, the utility of an evacuation zone rule depends on there being at least some sort of evacuation. So adopting an evacuation zone rule would be of no help in cases where nothing remotely approaching an evacuation happens, but the economic effects are nevertheless severe. (Consider, for instance, an oil spill at sea that leaves dry land mostly untouched.) Faced with all this potential for negative consequences and doctrinal confusion, "we would be acting rashly to adopt a rule treating" evacuation zones as talismanic. ( Intel Corp. v. Hamidi (2003) 30 Cal.4th 1342, 1363, 1 Cal.Rptr.3d 32, 71 P.3d 296 [declining to extend trespass *648liability into cyberspace based on similar doctrinal and practical concerns].)

Without adopting a (not so) bright-line evacuation zone rule, the alternative is applying a fact-intensive, case-by-case standard à la People Express . But we have already experimented **894with an analogous approach regarding recovery for negligent infliction of emotional distress. It did not go well. In Thing , we lamented the "arbitrary results" and the "inconsistent and often conflicting" body of law that approach produced. ( Thing , supra , 48 Cal.3d at p. 662, 257 Cal.Rptr. 865, 771 P.2d 814.) Which is why we retreated from an ad hoc standard and imposed instead a hard-and-fast rule. ( Id. at pp. 667-668, 257 Cal.Rptr. 865, 771 P.2d 814.)

We have not forgotten that experience. Today, we are confronted with hundreds of claims brought by hundreds of businesses stemming from one industrial accident - and that's just the artificially limited class Plaintiffs seek to represent, not the full universe of potential claimants whose pocketbooks were adversely (and foreseeably) affected by the leak. We see no workable way to limit geographically who may recover purely economic losses. Without one, the dangers of indeterminate liability, over-deterrence, and endless litigation are at their apex.

*4112.

Nor do we see a viable way to limit temporally what purely economic losses could be recovered here.

Plaintiffs allege that they "have been and continue to be heavily impacted by the gas leak." (Italics added.) That is possible because Plaintiffs complain not of being forced to shut down during the disaster - no named plaintiff squarely alleges that - but of losing customers due to the exodus of neighborhood residents. And even though the leak is over, they allege that, for as long as the Aliso Facility remains in use, "business will never return to Porter Ranch as usual." (Italics added.) So Plaintiffs are, in effect, seeking pro rata recovery for the past, present, and future economic toll the leak allegedly had, has, and will have on Porter Ranch. These are claims without end.

True: we could conceivably cabin recovery for purely economic losses to those suffered during the disaster alone. Or we could allow recovery only for such losses suffered during a business closure, not merely for systemic hits to economic demand. Yet upon closer inspection, the alluring simplicity of both approaches quickly proves to be a mirage.

The "during the disaster" option would require a way of determining precisely what the words "during" and "disaster" mean in a given case. That will not always be easy. Even assuming the beginning and end of most disasters can be easily fixed by the closing of a wayward valve or its equivalent, distinctions between one disaster (say, a leak of flammable fluid) and another (a fire) can be unstable. Moreover, disasters like the gas leak at issue here happen over an extended period of time, but other industrial accidents (like tower collapses or railroad explosions) happen in an instant. So for the latter sort of disaster, we might have to use the duration of any subsequent evacuation (if there is one) to time-bound the ensuing claims for purely economic losses. But doing that would inject into disaster response efforts the very same dangerous incentives and other problems discussed above.

The "business closure" option, for its part, would likely prove self-defeating. Requiring affected businesses to close as a prerequisite for recovery in negligence would lock them into a dilemma: shut down and lose any income you might have *649earned - or stay open and lose any tort claim you might have brought. Difficult though the choice could be for some, many businesses might rationally decide they are better off shutting down. Plaintiff Mediterranean Bistro, for example, would presumably be reluctant to keep its 80-seat restaurant open to serve a handful of customers if doing so meant forfeiting a potentially valuable tort claim. Encouraging businesses to close *412could thus catalyze the very economic stagnation we want to minimize. Better instead to encourage businesses to continue their economic activity where they can.

D.

None of this is to say that denying recovery for those who did not suffer injury to person or property is a perfect solution in negligence cases like this one. Far from it. It is only the least-worst rule out there.

Like other courts, we acknowledge that denying recovery for purely economic losses under circumstances like these has "the vice of creating results in cases at its edge that **895are ... 'unjust' or 'unfair' " - or even "seemingly perverse." ( Testbank , supra , 752 F.2d at p. 1029 ; see also 532 Madison , supra , 727 N.Y.S.2d 49, 750 N.E.2d at p. 1103 [acknowledging that this rule is "to an extent arbitrary because ... invariably it cuts off liability to persons who foreseeably might be plaintiffs"].) The courthouse doors are open for people who experience slight physical injury - yet closed to others who suffer devastating purely economic losses. That line may appear arbitrary in some sense. Yet so are the alternatives we have considered and rejected - and those alternatives, as we've explained, have further flaws of their own.

At any rate, "drawing arbitrary lines is unavoidable if we are to limit liability and establish meaningful rules for application by litigants and lower courts." ( Thing , supra , 48 Cal.3d at p. 666, 257 Cal.Rptr. 865, 771 P.2d 814.) And as we have explained, the ripple effects of industrial catastrophe on this scale in an interconnected economy defy judicial creation of more finely tuned rules. Hence the admittedly imperfect legal regime that governs in most jurisdictions - and that we now confirm governs in ours.

The Legislature, however, may be able to improve that regime in ways that would be exceptionally difficult, if not impossible, for us. To name one example: after we rebuffed homeowners' efforts to recover for purely economic losses stemming from construction defects in Aas , the Legislature responded to popular calls for a more forgiving rule in that context. (See Rosen , supra , 30 Cal.4th at p. 1079, 135 Cal.Rptr.2d 361, 70 P.3d 351, citing Aas , supra , 24 Cal.4th at p. 646, 101 Cal.Rptr.2d 718, 12 P.3d 1125.) It enacted a detailed statutory mechanism specifically designed for homeowners seeking redress against negligent builders. ( Rosen , at p. 1079, 135 Cal.Rptr.2d 361, 70 P.3d 351, citing Civ. Code, § 895 et seq. ) To name another: in view of "the economic and social disruptions arising out of the Lake Davis Northern Pike Eradication Project," the Legislature set up a special process for people to recover for, among other things, purely economic losses suffered due to that environmental protection effort. ( Gov. Code, §§ 998, 998.2.)

*413With the economic consequences in this case allegedly so severe, and the number of people affected allegedly so large, the Legislature could be spurred yet again to act. To be sure, purely economic losses caused by a natural gas leak may present their own set of challenges. But so too, we can only presume, of those caused by an oil spill. And in that context the Legislature has already interceded. It enacted legislation permitting those "who derive[ ]

*650at least 25 percent" of their income from activities that utilize "natural resources" to recover - without regard to fault - for "[l]oss of profits or impairment of earning capacity due to the injury, destruction, or loss of ... natural resources" from a spill.9 ( Gov. Code, § 8670.56.5, subd. (h)(6).) Perhaps there's a basis for further industry-specific legislative or regulatory action. And through the democratic process, the Legislature can bring to bear a mix of expertise while considering competing concerns to craft a solution in tune with public demands.

A partial solution leveraging the insurance market may also prove feasible, at least for some businesses. Although many business interruption insurance policies presently available might not cover the purely economic losses alleged here (see Buxbaum v. Aetna Life & Casualty Co. (2002) 103 Cal.App.4th 434, 448-449, 126 Cal.Rptr.2d 682 ), private insurance companies could conceivably see a profit-making opportunity in today's decision. Now certain that a lawsuit seeking purely economic losses of this sort will not succeed, businesses operating near a natural gas storage facility - or a dam, shipping lane, oil well, and so forth - may be more inclined to buy insurance covering profits they stand to lose if disaster strikes. (See, e.g., Testbank , supra , 752 F.2d at p. 1029 [observing that a local business's "own potential losses" in the event of an industrial accident "are finite and readily discernible," which may enable them to obtain insurance "at a relatively low cost"].) If so, private insurance companies might expand their policy offerings accordingly.

Finally, we recognize Plaintiffs' concern that SoCalGas's alleged negligent behavior will go insufficiently deterred if we deny **896recovery here. But SoCalGas is not getting off scot-free. At oral argument, the company represented that some 50,000 claimants have alleged in other litigation that they suffered property damage caused by the leak - several hundred of whom are local businesses. It further informed us, and we have no reason to doubt, that the company has spent some $450 million on remedial measures and agreed to pay another $120 million as part of a settlement with local authorities. SoCalGas, operating in a heavily regulated domain, also remains under investigation - and may face further consequences in the future. *414III.

Risks from industrial accidents raise grave concerns for society, and we have no doubt the accident precipitating this case caused significant hardships. To compensate those harmed and to deter those who do the harming, our society assigns tort law a pivotal role. But that does not mean society's interests are best served by extending its scope indefinitely. Meaningful limits on tort liability, along with the incentives they set, are crucial to the functioning of our economy and of our courts. Where such limits leave gaps in our social fabric, tort does not stand alone: insurance also compensates, regulation also deters. And where gaps persist, the Legislature can act.

The better part of a century has passed since then-Judge Cardozo warned that permitting recovery in negligence for purely economic losses can threaten indeterminacy-cubed: "liability in an indeterminate amount for an indeterminate time to an indeterminate class." ( *651Ultramares Corp. v. Touche (1931) 255 N.Y. 170, 174 N.E. 441, 444.) Courts across the country have since heeded that warning, by and large denying recovery in negligence cases like this one even though purely economic losses inflict real pain. That prevailing rule of no recovery is, like society itself, imperfect. Yet nearly everyone follows a rule that few (if any) entirely like. California does, too. So we affirm the Court of Appeal's judgment.

We Concur:

CANTIL-SAKAUYE, C. J.

CHIN, J.

CORRIGAN, J.

LIU, J.

KRUGER, J.

GROBAN, J.