DURHAM, Justice:
Plaintiffs sought compensatory damages for the wrongful death of their child, as well as for emotional distress to both plaintiffs and physical injury to plaintiff Ray Johnson. They also sought punitive damages against defendants. The punitive damages claim against defendant Newspaper Agency Corporation (NAC) was based on vicarious liability for Rogers’ conduct as well as on NAC’s own conduct. The trial court granted defendants’ motions for summary judgment as to the claims for punitive damages, but denied defendants’ motion on the issue of damages for Ray Johnson’s emotional distress. Plaintiffs filed [773]*773this interlocutory appeal, and defendants cross-appealed. We affirm the trial court’s ruling as to the damages for emotional distress, but reverse on the issues relating to punitive damages.
On December 10, 1984, NAC moved for partial summary judgment on the issue of punitive damages. On March 22, 1985, NAG filed a motion seeking dismissal or partial summary judgment as to the claims for emotional distress. On March 25,1985, the trial court issued its memorandum decision granting NAC’s motion for summary judgment as to punitive damages, relying on this Court’s decision in McFarland v. Skaggs Companies, Inc., 678 P.2d 298 (Utah 1984). Finding that there was no evidence of actual malice on the part of NAC, the court granted partial summary judgment against plaintiffs on their claims for punitive damages against NAC. The court also ruled that Utah does not recognize vicarious liability for punitive damages. On the basis of that ruling, defendant Rogers moved for partial summary judgment as to his liability for both punitive damages and emotional distress. Although granting partial summary judgment with respect to Rogers’ liability for punitive damages, the court denied summary judgment on the father’s emotional distress claim, holding that a parent who is in the “zone of danger” may recover for the trauma associated with seeing a child injured. The claim for emotional distress on the part of plaintiff Frances Johnson was denied, but that ruling is not appealed here.
I. Facts
At approximately 10:00 p.m. on April 16, 1982, plaintiff Ray Johnson and his eight-year-old son David were waiting for a “walk” signal before crossing a street in downtown Salt Lake City. A truck crossed the intersection and jumped the curb, killing David and injuring Ray. The truck was owned by NAC and operated by Donald Rogers. It is admitted that Rogers was driving under the influence of alcohol and that he negligently caused the injuries.
Rogers began working for NAC in May, 1980. Rogers’ license had previously been revoked in Oregon after a conviction for driving under the influence; NAC failed to discover this fact. Rogers and other NAC employees sometimes reported to work intoxicated. Although NAC had written rules forbidding driving while intoxicated, it apparently did not enforce these rules. Rogers had been a heavy drinker for approximately six months to a year prior to the accident. Depositions of NAC employees indicate that the use of alcohol and marijuana was widespread and that no effort was made to curtail such use. These depositions also indicate that NAC vehicles were sometimes returned with beer cans in them, and on one occasion, an NAC supervisor who observed drivers smoking marijuana told the drivers to “do it on the road.” Moreover, many deponents contended that these abuses were widely known and that NAC management either knew of these practices or could easily have found out about them.
II. Standard for the Imposition of Punitive Damages
The trial court held in this case that “evil intent,” “actual malice,” or “malice in fact” is required for the imposition of punitive damages. That holding misconstrues our case law; it was made in reliance on language in McFarland v. Skaggs Companies, Inc., 678 P.2d 298 (Utah 1984), in which this Court, after extensive consideration of specific policy concerns, established an “actual malice” standard for the imposition of punitive damages in false imprisonment cases. The opinion itself, as well as the law review article quoted extensively therein, focuses solely on the “ancient tort of false imprisonment [which creates] liability for wrongfully restraining another’s freedom of movement” and concludes that an actual malice standard is necessary to balance the competing interests in shoplifting cases.
The very real problem of shoplifting pits two important considerations against each other — the right of the merchant to protect his inventory and the right of the citizen to be free from unwarranted detention and accusation....
[774]*774[B]y sanctioning unrestricted punitive damages for a good faith mistake, the Terry [v. Z.C.M.I., 605 P.2d 314 (Utah 1979) ] court tipped the balance too far in favor of the patron and against the mer-chant_ [The actual malice rule] protects the interests of both merchant and patron without opening the door to unwarranted punitive damage recoveries.
McFarland v. Skaggs Companies, Inc., 678 P.2d at 304 (quoting Note, False Imprisonment — Punitive Damages May Be Awarded As A Matter of Law, 1980 Utah L.Rev. 694, 699-700 (1980)).
McFarland departed from the Terry “malice in law” standard only for false imprisonment cases and only because of carefully detailed policy reasons related to that tort. Since McFarland, this Court has explicitly articulated a broader standard for the imposition of punitive damages in at least two cases: Atkin Wright & Miles v. Mountain States Telephone, 709 P.2d 330 (Utah 1985) (intentional interference with prospective economic relations), and Synergetics v. Marathon Ranching Co., 701 P.2d 1106 (Utah 1985) (fraud, misrepresentation, and deceit). We stated: “Punitive damages, among other things, punish conduct which manifests a knowing or reckless indifference toward, and disregard of, the rights of others.” Synergetics, 701 P.2d at 1112-13 (citations omitted). “Before punitive damages may be awarded, the plaintiff must prove conduct that is willful and malicious ... or that manifests a knowing and reckless disregard toward the rights of others.” Atkin Wright & Miles, 709 P.2d at 337 (citations omitted). In both of those cases, we cited with approval identical or similar language from cases decided prior to McFarland: Behrens v. Raleigh Hills Hospital, Inc., 675 P.2d 1179 (Utah 1983) (cited in Synergetics); Branch v. Western Petroleum, Inc., 657 P.2d 267 (Utah 1982); First Security Bank v. J.B.J. Feedyards, Inc., 653 P.2d 591 (Utah 1982) (cited in Synergetics and Atkin Wright & Miles)] Terry v. Z.C.M.I., 605 P.2d 314 (Utah 1979) (cited in Atkin Wright & Miles). The citation of the Terry standard in Atkin Wright & Miles is significant because it conclusively demonstrates that McFarland overruled the use of the Terry standard only for false imprisonment cases, contrary to the argument made in this case by the defendants.
The standard for punitive damages in non-false imprisonment cases is thus clear: they may be imposed for conduct that is willful and malicious or that manifests a knowing and reckless indifference and disregard toward the rights of others. Defendants argue that mere driving under the influence of alcohol is insufficient to support a finding of knowing and reckless indifference and disregard for the rights and safety of others, and Rogers further argues that punitive damages are inappropriate here because he was convicted and sentenced for criminal violations stemming from the same acts.
The overwhelming majority of jurisdictions which have considered the issue have ruled that punitive damages are available in drunk driving eases.1
One who willfully consumes alcoholic beverages to the point of intoxication, [775]*775knowing that he thereafter must operate a motor vehicle, thereby combining sharply impaired physical and mental faculties with a vehicle capable of great force and speed, reasonably may be held to exhibit a conscious disregard for the safety of others.
[[Image here]]
The allowance of punitive damages in such cases may well be appropriate because of another reason, namely to deter similar future conduct_ [T]he applicable principle was well expressed in a recent Oregon case....
[T]he fact of common knowledge that the drinking driver is the cause of so many of the more serious automobile accidents is strong evidence in itself to support the need for all possible means of deterring persons from driving automobiles after drinking, including exposure to awards of punitive damages in the event of accidents.
Taylor v. Superior Court of Los Angeles County, 24 Cal.3d 890, 897, 598 P.2d 854, 857, 157 Cal.Rptr. 693, 697 (1979) (quoting Harrell v. Ames, 265 Or. 183, 190, 508 P.2d 211, 214-15 (1973)).
Assessing punitive damages in cases such as this is not inappropriate or inconsistent with the deterrence function of damages. See Peterson v. Superior Court of Ventura County, 31 Cal.3d 147, 154-62, 642 P.2d 1305, 1308-13, 181 Cal.Rptr. 784, 787-92 (1982); Soria v. Sierra Pac. Airlines, 111 Idaho 594, 610, 726 P.2d 706, 722 (1986); Sebastian v. Wood, 246 Iowa 94, 100, 66 N.W.2d 841, 844 (1954); Dorn v. Wilmarth, 254 Or. 236, 239-40, 458 P.2d 942, 944 (1969); Ellis, Fairness and Efficiency in the Law of Punitive Damages, 56 S.Cal.L.Rev. 1 (1982); Schwartz, Deterrence and Punishment in the Common Law of Punitive Damages: A Comment, 56 S.Cal.L.Rev. 133 (1982). Given their deterrence value, punitive damages are not prohibited where a tort-feasor has been convicted and sentenced for criminal violations stemming from his tortious conduct. In Huff v. Chrismon, 68 N.C.App. 525, 315 S.E.2d 711, review denied, 311 N.C. 756, 321 S.E.2d 134 (1984), the North Carolina Court of Appeals stated that the assessment of punitive damages in civil cases, separate from any criminal penalties imposed, was “consistent with the trend to maximize punishment and deterrence of impaired drivers_” 68 N.C.App. at 532, 315 S.E.2d at 715. On the specific question of deterrence, the Oregon Supreme Court went on to observe:
It may be debatable whether either awards of punitive damages or the imposition of criminal penalties will effectively deter persons from driving after drinking. However, in the absence of a showing of substantial evidence to the contrary, we are not prepared to hold that law enforcement officials and courts, who have a heavy responsibility in this area, are wrong in their present apparent assumption that both criminal penalties and awards of punitive damages may have at least some deterrent effect in dealing with this serious problem. We are also not aware of any good reasons why punitive damages should not have as much deterrent effect upon this type of wanton and reckless conduct as upon other types of conduct in which awards of punitive damages are traditionally approved by the courts.
Harrell v. Ames, 265 Or. at 190-91, 508 P.2d at 215 (citations omitted; emphasis added).
In this state, the Utah Court of Appeals has held that punitive damages are available in drunk driving cases. In Biswell v. Duncan, 742 P.2d 80 (Utah Ct.App.1987), the court determined that punitive damages may be awarded upon proof of either actual or legal malice and that the imposition of punitive damages against drunk drivers is consistent with Utah’s public policy. Id. at 84. The court held that in order to recover, a plaintiff must show that the defendant acted with actual malice or a reckless disregard of the rights of others and that his drunken driving contributed to the accident. Id.
We agree with the majority of other jurisdictions and with the Utah Court of Appeals. We see no reason to exclude drunk driving from the categories of outrageous [776]*776conduct, either willful or knowingly reckless, which are eligible for the imposition of punitive damages. We emphasize that this is not a “per se” rule: the mere allegation or fact of having caused an accident or injury after drinking and driving will not support an award of punitive damages. The standard is fact-specific and requires proof of conduct which is knowingly reckless and exhibits a high degree of disregard for the safety of others. To that extent, defendants are correct when they argue that mere allegations of driving after drinking could not support a punitive damages award.
In this case, however, we have allegations, evidence, and admissions of far more outrageous behavior. Rogers was employed as a truck driver. During the period of time immediately before reporting to work on the evening of the accident, he had consumed approximately seven mixed drinks containing vodka and had “chug-a-lugged” a 27-ounce drink containing two mini-bottles of tequila. His blood alcohol content after the accident was .18 percent. He admitted that he had been a “heavy” or “problem” drinker for six months to a year prior to the accident and that he had been convicted of driving under the influence in Oregon. At 10:00 p.m., Rogers drove his truck across an intersection in downtown Salt Lake City, up over the curb and onto the sidewalk, hitting and throwing several concrete pillars and the eight-year-old victim who had been standing on the corner.
The foregoing facts, if proved to a jury, would certainly be sufficient to support a finding of knowing and reckless disregard for the safety of others. Defendants suggest that because Rogers was probably too drunk to know or control what he was doing, he cannot be found to have been “knowingly” reckless. It would be perfectly permissible for a jury, however, to find that the element of deliberation and knowledge in his behavior comes from his consumption of large amounts of alcohol just before reporting to work and his decision to get into his truck and drive it on the public streets. It would be anomalous if a sober driver who drives a vehicle onto a sidewalk with people on it could be held liable for punitive damages (which would certainly be possible in the appropriate factual context), while one who has deliberately rendered himself incapable of driving safely and then drives could not. On the other hand, mere drinking and driving, absent evidence that the manner of doing so reflects conscious disregard for the safety of others, would not qualify for punitive damages. Cf. Baker v. Marcus, 201 Va. 905, 114 S.E.2d 617 (1960). As with punitive damages in all personal injury cases, it is the extreme, outrageous, and shocking behavior that justifies their imposition in drunk driving cases. The behavior alleged in this case is sufficiently extreme, outrageous, and shocking to permit the issue to go to the jury. Therefore, we hold that the trial court erred in granting summary judgment on the availability of punitive damages and reverse on that issue.
III. Vicarious Liability of Employer for Punitive Damages
The issue of when an employer can be held vicariously liable for punitive damages because of the acts of nonmanagerial employees is one of first impression in Utah. There are currently at least four approaches to such liability in other jurisdictions. The first is a straight-forward vicarious liability rule that permits recovery against an employer whenever the employer is liable for the same conduct in compensatory damages.2
A second, more conservative approach has been adopted by a number of other jurisdictions. See J. Ghiardi & J. Kircher, Punitive Damages Law and Practice, ch. 24, at 4-5 (1987), and cases cited therein. This standard appears in the Restatement [777]*777(Second) of Torts § 909 and the Restatement (Second) of Agency § 217C:
Punitive damages can properly be awarded against a master or other principal because of an act by an agent if, but only if,
(a) the principal or a managerial agent authorized the doing and the manner of the act, or
(b) the agent was unfit and the principal or a managerial agent was reckless in employing or retaining him, or
(c) the agent was employed in a managerial capacity and was acting in the scope of employment, or
(d) the principal or a managerial agent of the principal ratified or approved the act.
The third standard takes an intermediate approach between pure vicarious liability and the Restatement standard, which is not “true” vicarious liability because it is predicated on acts of the principal. This standard has been adopted by the Supreme Court of Florida and the United States Fifth Circuit Court of Appeals and is characterized by the parties as a “some fault” rule:
Before an employer may be held vicariously liable for punitive damages under the doctrine of respondeat superior, there must be some fault on his part. Although the misconduct of the employee, upon which the vicarious liability of the employer for punitive damages is based, must be willful and wanton, it is not necessary that the fault of the employer, independent of his employee’s conduct, also be willful and wanton. It is sufficient that the plaintiff allege and prove some fault on the part of the employer which foreseeably contributed to the plaintiffs injury to make him vicariously liable for punitive damages.
Mercury Motors Express, Inc. v. Smith, 393 So.2d 545, 549 (Fla.1981); see also Dorsey v. Honda Motor Co., 670 F.2d 21 (5th Cir.1982).
Finally, there are four states which apparently prohibit “vicarious” punitive damages altogether. See Briner v. Hyslop, 337 N.W.2d 858, 864-65 (Iowa 1983), and cases cited therein.
The courts of other jurisdictions are nearly evenly divided over the proper basis for “vicarious” punitive damage liability. See J. Ghiardi & J. Kirchner, Punitive Damages Law and Practice, ch. 24, at 36-39 (1987) (twenty-one states follow traditional respondeat superior rules, and nineteen follow the “complicity rule” embodied in the Restatement).
There are serious arguments in support of the position that “vicarious” punitive damages ought to be abandoned entirely as inefficient and unfair. One recent treatment of the subject reached the following conclusion:
In summary, vicarious punitive damage liability cannot be justified as deserved punishment. Indeed, it is usually conceded to be unfair. Deterrence alone is thus relied on to justify it. The foregoing analysis indicates, however, that efficient levels of deterrence are unlikely to be promoted by vicarious punitive damage liability, even assuming that the criteria for assessing punitive damages can be made certain and predictable. When that assumption is dropped, it becomes apparent that the combination of vicarious liability with the prevailing uncertain criteria for determining liability and magnitude serves to diminish rather than to increase aggregate welfare.
Ellis, Fairness and Efficiency in the Law of Punitive Damages, 56 S.Cal.L.Rev. 1, 71 (1982).
In an article in the same symposium, another scholar suggests that closer attention should be paid to the differential analysis required by punishment versus deterrence as goals and disagrees with Professor Ellis on vicarious punitive damages:
Consider now the problem of vicarious liability for punitive damages. As a general matter, our existing criminal law regards vicarious liability as an impermissible basis for punishment (except, perhaps, for a limited range of minor penalty regulatory offenses): one man cannot be judged morally guilty on account of another man’s crime. By con[778]*778trast, our civil law of torts warmly embraces vicarious liability, apparently because the employer is in the best position to control the behavior of his own employees. If punishment is the chief purpose of punitive damages, then the criminal law model should prevail and vicarious liability should be rejected. But if deterrence is the principle purpose, the tort law model seems controlling, and vicarious liability can be endorsed.
Schwartz, Deterrence and Punishment in The Common Law of Punitive Damages: A Comment, 56 S.Cal.L.Rev. 133, 136 (1982).
The defendants in this case have not, however, argued the legitimacy of vicarious punitive damages; they have addressed only the standard for imposition. We are therefore not inclined to address the fundamental question at this time, but we are sufficiently concerned about the rationale supporting the doctrine to opt for the more conservative standard articulated by the Restatement rule. The “complicity rule” (a term apparently coined by Morris, Punitive Damages in Personal Injury Cases, 21 Ohio St.L.J. 216, 221 (1969)), limits vicarious punitive damages to those situations where wrongful acts were committed or specifically authorized by a managerial agent or were committed by an unfit employee who was recklessly employed or retained. Since the rule requires some wrongful action on the part of the employer, it addresses at least some of the punishment and deterrence goals accepted thus far in the law of this state.
The issue thus becomes whether there exist triable issues of fact regarding NAC’s liability for punitive damages under the Restatement standard. In addition to the facts described earlier in this opinion, the following facts involving NAC were the subject of testimony or other evidence during discovery: although NAC purportedly ran periodic driver’s license checks on its drivers, it failed to learn of Rogers’ previous conviction and also failed to learn that two other drivers drove on suspended licenses while in NAC employ; Rogers and other employees testified that he made no effort to hide his drinking habits and that they were apparent in the workplace; Rogers frequently came to work intoxicated, openly consumed alcohol while he was at work, and periodically took alcohol with him in company vehicles when he made deliveries; alcohol and drug use were relatively common during working hours and in company vehicles and that NAC supervisors not only were aware of such use, but were also participants on occasion; alcohol and drug use at work was common, open, and the subject of regular “office gossip”; there was virtually no supervision of employees on the night shift, and no vehicle check-out system; and NAC policies against drug and alcohol use were not enforced.
I note, of course, that many or all of the above facts may be disputed by NAC, but on a motion for summary judgment, we should determine whether the prevailing party is entitled to judgment as a matter of law on a view of the proffered facts most favorable to the losing party. Based on the brief summary detailed above, I conclude that plaintiffs are entitled to a jury verdict on the question of whether NAC “authorized the doing and the manner of the act” of taking out company vehicles on company business while in a state of extreme intoxication or whether “the agent was unfit and the principal or a managerial agent was reckless in employing or retaining him.” Restatement (Second) of Torts § 909(a), (b) (1965). Therefore, I would reverse the summary judgment on the issue of punitive damages against NAC and remand for trial.
IV. Negligent Infliction of Emotional Distress
Also before us on appeal is the question of whether a cause of action for negligently inflicted emotional distress exists in Utah and, if so, under what circumstances it exists. Defendants rely on Reiser v. Lohner, 641 P.2d 93 (Utah 1982), and Samms v. Eccles, 11 Utah 2d 289, 358 P.2d 344 (1961). In Reiser, the Court rejected without analysis a claim for emotional distress to the parents of a child damaged by allegedly negligent medical treatment. [779]*779The opinion says, “[I]t is well established in Utah that a cause of action for emotional distress may not be based upon mere negligence.” The two cases relied on for that principle are Samms v. Eccles and Jeppsen v. Jensen, 47 Utah 536, 155 P. 429 (1916).
In Samms v. Eccles, addressing a claim for emotional distress resulting from “indecent proposals,” the Court held:
Our study of the authorities, and of the arguments advanced, convinces us that, conceding such a cause of action may not be based upon mere negligence, the best considered view recognizes an action for severe emotional distress, though not accompanied by bodily impact or physical injury, where the defendant intentionally engaged in some conduct toward the plaintiff, (a) with the purpose of inflicting emotional distress, or, (b) where any reasonable person would have known that such would result; and that his actions are of such a nature as to be considered outrageous and intolerable in that they offend against the generally accepted standards of decency and morality.
11 Utah 2d at 293, 358 P.2d at 346-47 (citations omitted).
Jeppsen v. Jensen, although containing some dicta suggesting that the Court accepted the then-prevailing majority view that no recovery could be had for negligent infliction of emotional distress, does not so hold; the Court merely determined that a complaint alleging emotional distress as a result of “willful and wanton” acts stated a cause of action.
While it is true, and that it is so is not remarkable, that there are some cases which seem to hold that in no case can a recovery be had where the injuries are caused alone from fright, yet it is also true that there are many cases which seem to hold that a recovery may be had for fright alone, although the acts complained of constituted merely negligent acts. Such cases are, however, not numerous, and the great weight of authority is reflected in [the cases to the contrary cited above].
Jeppsen v. Jensen, 47 Utah at 543, 155 P. at 431.
In view of the age of Samms v. Eccles (decided twenty-seven years ago) and Jeppsen v. Jensen (decided seventy-two years ago), a reexamination of their premises is timely. This is particularly so in view of the extensive development of the law in this area in the intervening period and the abandonment by other jurisdictions of the precedents which were persuasive to this Court in its earlier opinions. Therefore, we depart from the approach taken in Reiser v. Lohner and address the question anew: Should a cause of action exist in this jurisdiction for the negligent infliction of emotional distress, and should the cause of action extend to bystanders whose emotional injuries result from the infliction of injury to or death of third persons? We note that all parties have framed the issue in this fashion. It is unclear from the record whether Rogers’ conduct is claimed to be sufficiently “wanton” to qualify under the intentional infliction of emotional distress standard.
Virtually all jurisdictions in the United States now recognize a broad protected interest in mental tranquility, first acknowledged in Utah in Jeppsen. The negligent infliction of emotional distress as a separate tort (distinct from the “willful and wanton” infliction of emotional distress or the negligent infliction of physical injuries with concomitant emotional injuries) has evolved rapidly only since the 1960s. See W. Keeton, D. Dobbs & D. Owen, Prosser and Keeton on the Law of Torts § 54 (5th ed. 1984); Note, The Negligent Infliction of Emotional Distress: A Critical Analysis of Various Approaches to the Tort in Light of Ochoa v. Superior Court, 19 Ind. L.Rev. 809 (1986). A common fact pattern for the cause of action is that existing in this case: a bystander observes negligent injury to a victim, which causes the bystander to suffer emotional distress. The courts have developed several rules affecting recovery for the emotional distress. Currently, no jurisdiction precludes recovery under any circumstances. Recovery is based upon satisfaction of one of three [780]*780standards: the impact rule, the zone-of-danger rule, or a foreseeability standard.
The impact rule requires that a plaintiff sustain some physical impact or injury which itself causes emotional distress. It was the original and most limited approach to emotional distress claims and was responsive to courts’ early concerns about speculative damages and floods of litigation over trivial claims. Although this rule was at one time a majority position, it has fallen into disfavor in recent years for a number of reasons. Suspicion regarding the authenticity of claims of mental distress has decreased with medical advances in the field of psychiatry and psychology. The concern over case load impact has come to be seen as inadequate reason to deny legitimate claims, and those courts that have abandoned the impact rule have not in fact seen drastic increases in this type of litigation. Finally, the results of the impact rule were often arbitrary, capricious, and unfair: the existence of a physical impact or injury frequently bore no rational relationship to the existence and severity of emotional injuries. Sometimes even the slightest of physical impacts (e.g., from smoke, dust, small jolts) was a sufficient predicate for recovery for emotional injuries, whereas the absence of such trivial impact could preclude recovery for the same injuries. See Note, The Negligent Infliction of Emotional Distress: A Critical Analysis of Various Approaches to the Tort in Light of Ochoa v. Superior Court, 19 Ind.L.Rev. 809 (1986). Because of its irrationality, the majority of courts abandoned the impact rule.
The zone-of-danger rule, relied on by the trial judge in this case, was adopted by many courts as a less restrictive substitute for the impact rule. It is set forth in section 313 of the Restatement (Second) of Torts:
(1) If the actor unintentionally causes emotional distress to another, he is subject to liability to the other for resulting illness or bodily harm 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.
(2) The rule stated in Subsection (1) has no application to illness or bodily harm of another which is caused by emotional distress arising solely from harm or peril to a third person, unless the negligence of the actor has otherwise created an unreasonable risk of bodily harm to the other.
(Emphasis added.)
The zone-of-danger rule offers many of the same benefits, although more expansive in its application, as did the impact rule. It is an objective standard and serves to identify in consistent fashion those who are eligible to recover. Unfortunately, it also suffers from many of the disadvantages of the impact rule: it is a rigid and inequitable limitation on recovery for injuries which are otherwise indistinguishable from each other. The parent standing next to the child hit by a car has a cause of action; the parent standing twenty feet away does not.
California was the first jurisdiction in the United States to extend liability beyond the zone of danger. Dillon v. Legg, 68 Cal.2d 728, 441 P.2d 912, 69 Cal.Rptr. 72 (1968). In Dillon, Justice Tobriner of the California Supreme Court wrote:
[T]he 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 been within it. The case 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 [781]*781have rejected the impact rule becomes even less defensible. We have, indeed, held that impact is not necessary for recovery. 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.
68 Cal.2d at 731, 441 P.2d at 915, 69 Cal.Rptr. at 75 (citation omitted).
The court went on to observe that it had in the past “rejected the argument that we must deny recovery upon a legitimate claim because other fraudulent ones may be urged,” 68 Cal.2d at 733, 441 P.2d at 917, 69 Cal.Rptr. at 77, and that “the alleged inability to fix definitions on the different facts of future cases does not justify the denial of recovery on the specific facts of the instant case; ... proper guidelines can indicate the extent of liability for such future cases.” 68 Cal.2d at 735, 441 P.2d at 919, 69 Cal.Rptr. at 79. Focusing on the foreseeability of risk, the court enunciated what has come to be known as the “Dillon rule,” consisting of three factors which when evaluated determine the degree of foreseeability of the plaintiffs injury: (1) whether the plaintiff was located near the scene of the accident; (2) whether the emotional trauma to the plaintiff was caused by actually witnessing the accident; and (3) whether the plaintiff and the victim were closely related. In light of these factors, the court can determine whether the injury was reasonably foreseeable. 68 Cal.2d at 736, 441 P.2d at 920-21, 69 Cal.Rptr. at 80.
Hawaii has broadened the Dillon foreseeability standard, using a “pure” foreseeability analysis in bystander cases. Where serious emotional distress to a plaintiff-bystander is the reasonably foreseeable consequence of the defendant’s act, the defendant’s conduct is the proximate cause of the plaintiff’s mental injury and general tort principles are applied to impose liability. Leong v. Takasaki, 55 Haw. 398, 520 P.2d 758 (1974); Rodrigues v. State, 52 Haw. 156, 472 P.2d 509, reh’g denied, 52 Haw. 283, 472 P.2d 509 (1970). In the absence of any limiting rules, however, the circumstances permitting liability extend far beyond ordinary expectations, creating grave theoretical and policy problems.
The distinction between “direct victim” and “bystander” liability and its effect on the standard of recovery used by the courts must also be examined in order for the summary of currently accepted approaches to liability to be complete. Courts have characterized the direct or primary victim as the person to whom a duty was owed and who was directly injured by the breach of that duty. The bystander or secondary victim is anyone who was not directly injured by the defendant, but who suffered mental distress as a result of his or her association with the direct victim’s injury. Thus, in fact situations involving car/pedestrian accidents, the person who is physically injured by the defendant may bring an action for emotional trauma as a direct victim, while a bystander to the accident may recover based only on theories of secondary liability.3
The direct victim/bystander distinction is important because in some jurisdictions it controls the standard which a plaintiff must meet in order to recover. In California, the Dillon standard no longer applies in direct victim cases. Instead, the court has adopted a broader foreseeability test and has abandoned the requirement that the emotional injury be physically manifested. Molien v. Kaiser Foundation Hosp., 27 Cal.3d 916, 616 P.2d 813, 167 Cal.Rptr. 831 (1980). States that have adopted a zone-of-danger rule have, in effect, limited recovery to cases involving direct victims, disallowing recovery to bystanders. Plain[782]*782tiffs who are allowed to recover because they were present within the zone of danger are direct victims because the defendant breached the duty of care owed them. Other witnesses falling outside of the zone are denied recovery due to the lack of direct injury and breach of a duty. The distinction between direct victims and bystanders must be taken into account in fashioning and applying any standard of recovery for negligent infliction of emotional distress.
A final comment must be added to the discussion of direct victim and bystander liability. As a method by which a defendant’s liability can be limited in bystander cases, many jurisdictions require that any claimed emotional injury be manifested physically. This requirement persists in most jurisdictions as a sign of courts’ concern that emotional injury claims made by plaintiffs who have not been directly injured are difficult to validate in the absence of physical symptoms and that deletion of this requirement would result in a flood of ill-founded bystander claims. However, some courts have recently dismissed these concerns as unwarranted and now allow recovery in the absence of any physical symptoms. See, e.g., Rodrigues v. State, 52 Haw. 156, 472 P.2d 509, reh’g denied, 52 Haw. 283, 472 P.2d 509 (1970); Paugh v. Hanks, 6 Ohio St.3d 72, 451 N.E.2d 759 (1983) (physical injury is evidence of the degree of mental injury).
Our research has not disclosed, nor have the parties identified, any jurisdiction in the United States which bars all recovery for the negligent infliction of emotional distress. The policy considerations in favor of realistic limits on negligence liability have given rise in turn to the impact rule, the zone-of-danger rule, and the Dillon rule, but not to a refusal to recognize the cause of action under any circumstances. Although the search for coherent, consistent application of liability principles in this area is difficult, that difficulty is an inappropriate predicate for denial of redress to a whole class of legitimate and serious claims.
It is further to be observed that the argument against allowing such an action because groundless charges may be made is not a good reason for denying recovery. If the right to recover for injury resulting from wrongful conduct could be defeated whenever such dangers exist, many of the grievances the law deals with would be eliminated. That some claims may be spurious should not compel justice to shut their [sic] eyes to serious wrongs and let them go without being brought to account. It is the function of courts and juries to determine whether claims are valid or false. This responsibility should not be shunned merely because the task may be difficult to perform.
Samms v. Eccles, 11 Utah 2d at 293, 358 P.2d at 347.
We therefore sustain the trial judge in his determination that a cause of action for negligent infliction of emotional distress may be maintained and undertake to establish general guidelines for the availability of recovery. A discussion of the rule of law to be applied must be general because the facts of this case would satisfy any of the three major tests applied in other jurisdictions. The plaintiff here suffered an impact during the accident, receiving physical injuries to his foot. Furthermore, he was in the immediate zone of danger created by Rogers’ acts. Finally, all three of the Dillon criteria are present: he was located immediately at the scene of the accident, he saw and heard all of the events associated with the violence to the victim, and the victim was his child.
I believe that the analytic approach of the Dillon rule is fundamentally sound in its focus on foreseeability as a necessary element of duty in negligence cases, and I would treat it as a legitimate starting point in the treatment of bystander causes of action for negligent infliction of emotional distress.4 I am aware, however, that the [783]*783rule as formulated in Dillon has not served as an adequate guarantee of certainty in application even by the California courts.5
I therefore emphasize that it is impossible to articulate a mechanistic test, as opposed to identifying an analytic approach, in a case which challenges no traditional limitations. In effect, I would hold that one may recover for the negligent infliction of emotional distress when one was in the zone of danger created by the negligence and suffered a physical impact. I nevertheless express the view that the less arbitrary, more traditional tort analysis embodied in the Dillon rule is appropriate. It is true that Dillon is more flexible than the impact and zone-of-danger rules, but it is not entirely free from the criticism that it too has
become hardened and mechanical in practice because the courts attempting to apply [it] have no general policies to guide them in the difficult cases.
If there is a rule which can determine liability in a more policy-oriented and less arbitrary manner while still drawing a line short of unlimited liability, such a rule should be adopted. A clear rule needs a clear rationale.
Note, Limiting Liability for the Negligent Infliction of Emotional Distress: The “Bystander Recovery” Cases, 54 S.Cal.L.Rev. 847, 867 (1981).
The scholarly literature contains some interesting proposals for amelioration of the ambiguities unsolved by the Dillon rule. See, e.g., Note, The Negligent Infliction of Emotional Distress: A Critical Analysis of Various Approaches to the Tort in Light of Ochoa v. Superior Court, 19 Ind.L.Rev. 809 (1986) (suggesting a “flexible standard of liability” premised on a general foreseeability analysis combined with a “clear and convincing” standard of proof); Note, Negligent Infliction of Emotional Distress: Developments in the Law, 14 Balt.L.Rev. 135 (1984) (“The best approach may be to treat negligent infliction of emotional distress as any other negligence action, employing the criteria of foreseeability and proximate cause, keeping in of damages to be awarded.”); Note, Limiting Liability for the Negligent Infliction of Emotional Distress: The “Bystander Recovery” Cases, 54 S.Cal.L.Rev. 847 (1981) (proposing that bystander recovery for the negligent infliction of emotional distress be limited to “those types of emotional distress for which a reasonable person would be emotionally unprepared”).
The decisional law pertaining to negligent infliction of emotional distress, particularly bystander cases, illustrates what may happen when unnecessarily doctrinaire approaches, based on policy considerations, are substituted for traditional negligence criteria. The adoption of arbitrary standards such as the impact and zone of danger rules were well-intentioned efforts at addressing fears of unlimited liability, but many deserving plaintiffs were left without a remedy.... Courts that have recognized this proposition have relaxed rigid rules, but the foreseeability standard that has been substituted in their stead has not been completely free of arbitrary criteria. While certain factors, such as proximity to and observance of the accident, relationship to the victim, and physical symptoms, may be valuable indicia of the merits of a complaint, courts should not ap[784]*784ply them so inflexibly as to preclude an otherwise provable claim.
Note, Negligent Infliction of Emotional Distress: Developments in the Law, 14 Balt.L.Rev. at 159.
Although I believe that expanding Utah’s tort principles to allow recovery for negligent infliction of emotional distress is prudent, I also recognize that philosophical concerns about the proper limits of tort liability exist. While liability standards have been greatly expanded in the past few decades, the pendulum has recently begun to swing back, with courts and commentators now reexamining the effectiveness and cost of the current system. At least one scholar believes that the focus should be on standards that prevent accidents rather than on compensation, a goal that is better served by traditional insurance than by the courts.6 It is possible that a major redesign of current tort law will become necessary and desirable. In the meantime, I am satisfied that the limits on recovery for negligent infliction of emotional distress contained in the Dillon rule strike an appropriate balance between the need for flexibility and the need for predictability.
Affirmed in part and reversed in part. Remanded for trial.