Nanette HISLOP, a single woman; on her own behalf and on behalf of her minor sons, Adam, Jacob and James Hislop, Plaintiff-Appellant, v. SALT RIVER PROJECT AGRICULTURAL IMPROVEMENT AND POWER DISTRICT, a political subdivision of the State of Arizona; Dycom Industries, Inc., a corporation doing business in Arizona; S.T.S., Inc., a corporation doing business in Arizona, Defendants-Appellees; Michael McLaurin and Debra McLaurin, husband and wife; on their own behalf, and on behalf of their minor daughters, Latasha, Carrie, Michelle and Sherry McLaurin, Plaintiffs-Appellants, v. Salt River Project Agricultural Improvement and Power District, a political subdivision of the State of Arizona; Dycom Industries, Inc., a corporation doing business in Arizona; S.T.S., Inc., a corporation doing business in Arizona, Defendants-Appellees, 197 Ariz. 553

OPINION SULT, Judge. ¶ 1 In this appeal we are asked to expand the category of claimants who can recover damages for emotional distress caused by witnessing harm to another. Specifically, appellants request that they be permitted to pursue a claim for the distress they experienced at witnessing the electrocution and burning of their co-worker and friend caused by the negligence of appellees. Because we conclude that current Arizona law does not require that appellants be permitted to prosecute their claim, and because we believe that sound policy reasons favor denying recovery on such a claim, we affirm the trial court’s grant of

summary judgment to appel-lees. BACKGROUND ¶ 2 Appellants Nanette Hislop and Michael McLaurin worked with Larry Matthews on a City of Phoenix sewer service crew. On August 22, 1995, Matthews was working in an open trench repairing a sewer main. Appellants were standing at the edge of the trench, just inches away from Matthews. ¶ 8 Matthews was using a jackhammer to break some concrete surrounding the sewer pipe when the jackhammer struck a high-voltage underground power line. Matthews burst into flames as electricity coursed through his body.

The fireball emitted by the electrocution shot up out of the trench some four to six feet and momentarily engulfed appellants, although they were not burned. ¶4 Appellant Hislop ran to the truck to radio for help. Appellant McLaurin seized a nearby fire extinguisher and put out the flames that were consuming Matthews. McLaurin then climbed into the trench and carried the unconscious Matthews to the surface. Matthews died three weeks later. ¶ 5 Appellants sued appellees, and among other claims, they sought recovery for negligent infliction of emotional distress for witnessing Matthews’ electrocution.

Appellants alleged that they both had been close friends of Matthews, had been in the “zone of danger” when he was electrocuted, and had suffered mental, physical, and emotional injuries as the result of witnessing the injury to Matthews. ¶ 6 Appellees filed a motion for

summary judgment, arguing that a claim for bystander emotional distress is available only to a plaintiff who has witnessed an injury to a close family member. Because appellants were not family members, appellees argued, they were not entitled to maintain such a claim. ¶7 The trial court agreed, noting that Arizona has never permitted anyone other than an actual family member to recover for bystander emotional distress.

On the question whether recovery ought to be expanded for such a claim, the trial court referred to California law and noted that recovery in that state is denied to friends, housemates, or those who have merely a meaningful relationship to the victim.

Based thereon, the trial court declined to extend recovery for bystander emotional distress to encompass close friends, noting that “the emotional trauma sustained by [appellants] was not reasonably foreseeable where the attachment to the victim derived from association as a friend, and not as a member of the victim’s family.” ¶ 8 The parties stipulated to the dismissal of the remaining negligence and negligence per se claims. Appellants timely appealed.

ISSUES ¶ 9 Appellants claim that existing Arizona law sanctions recovery for bystander emotional distress even when the bystander is not a family member of the principal victim. Appellants also implicitly argue that if existing case law does not extend liability this far, we should do so. ANALYSIS ¶ 10 Appellants first argue that in Keck v. Jackson, 122 Ariz. 114, 593 P.2d 668 (1979), our supreme court held that a person could recover for emotional distress caused by witnessing an injury to a non-family member.

In Keck, the court enumerated three factors that must be established to recover for emotional distress from witnessing harm to another: (1) “the shock or mental anguish of the plaintiff must be manifested as a physical injury”; (2) “the emotional distress must result from witnessing an injury to a person with whom the plaintiff has a close personal relationship, either by consanguinity or otherwise”; and (3) “the plaintiff/bystander must himself have been in the zone of danger so that the negligent defendant created an unreasonable risk of bodily harm to him.” Id. at 115-16, 593 P.2d at 669-70. ¶ 11 In this case, there is no dispute that appellants satisfied elements one and three.

The issue is whether their status as “coworker and close personal friend” to Matthews satisfies the Keck requirement that the claimant have a “close personal relationship, either by consanguinity or otherwise” with the victim. ¶ 12 We acknowledge that Keck’s language regarding the required relationship, particularly the “or otherwise” phrase, could arguably be interpreted to include “co-worker and friend” as a relationship with sufficient standing to permit recovery.

However, we do not believe the Keck court intended to go that far. ¶ 13 Certainly, the facts of Keck do not require such a conclusion as the ease involved a daughter who witnessed the death of her mother. Id. at 114, 593 P.2d at 668. We also note that the great weight of authority nationwide at the time Keck was decided would have denied recovery based solely on friendship or co-worker status. See William L. Prosser, Law of Torts § 54, at 334-35 (4th ed.1971); John S.

Herbrand, Annotation, Relationship Between Victim and Plaintiff-Witness as Affecting Right to Recover Damages in Negligence for Shock of Mental Anguish at Witnessing Victim’s Injury or Death, 94 A.L.R.3d 486, 1979 WL 52294 (1979). If Keck had intended such a significant expansion of bystander recovery, we believe the court would have taken pains to announce it clearly. ¶ 14 We believe that what Keck meant by “or otherwise” can be discerned from the court’s use of Leong v. Takasaki, 55 Haw. 398, 520 P.2d 758 (1974), to illustrate that phrase.

Leong sanctioned a recovery by a ten-year-old boy for mental and emotional distress as a result of witnessing the death of his step-grandmother. 520 P.2d at 760. In justifying this result, the Leong court relied not only on the fact that the step-grandmother had lived with the boy and eared for him as a natural grandmother but also on the nature of Hawaiian and Asian families, where it is not uncommon for members of three or four generations to five in one household. Id. at 766.

Keck appears to be saying, in essence, that while Arizona will not adhere strictly to a blood relationship requirement, there must still be a familial relationship, or something closely akin thereto, between the victim and the bystander to warrant the bystander’s inclusion as a recognized claimant.

Keck does not, contrary to appellant's’ contention, require recognition of their “coworker and friend” status as includable in the category of those who qualify for bystander recovery. ¶ 15 While Keck does not mandate recognition of appellants’ claim, neither does it directly preclude us from considering expanding the category of persons who may recover for bystander emotional distress. It is true that as we read Keck, the court preferred a close, familial-type relationship between victim and bystander on which to base recovery.

However, because Keck involved a parent-child relationship, the court’s discussion of this element of recovery was merely illustrative. It was not intended to foreclose consideration of other relationships and whether they could support a recovery for emotional distress. Tort law develops in an evolutionary fashion, and the “expansion of tort law claims is peculiarly within the realm of our judicial ... system.” Reben v. Ely, 146 Ariz. 309, 311, 705 P.2d 1360, 1362 (1985) (recognizing claim for filial loss of consortium).

We therefore turn to whether Arizona should sanction recovery for emotional distress for a bystander who is a “co-worker and friend.” ¶ 16 Before we address this, the principal issue, we must choose which analytical framework to employ, one based in theory or one based in policy. California, a state with an extensive jurisprudential history regarding bystander recovery for emotional distress, has elected to use negligence theory for the framework within which the specific question of who may recover should be determined. In the seminal ease of Dillon v.

Legg, 68 Cal.2d 728, 69 Cal.Rptr. 72, 441 P.2d 912 (1968), the California Supreme Court settled on “foreseeability” as the factor by which courts should judge whether a particular claimant could recover. In granting recovery to a mother who witnessed the death of her daughter, the court noted that “recovery should be had in such a case if defendant should foresee fright or shock severe enough to cause substantial injury in a person normally constituted.” Id. at 920.

To determine foreseeability, courts should take into account, inter alia, whether the “plaintiff and the victim were closely related, as contrasted with an absence of any relationship or the presence of only a distant relationship.” Id.; cf. Drew v.

Drake, 110 Cal.App.3d 555, 168 Cal.Rptr. 65, 65-66 (1980) (denying recovery for bystander distress to person who lived three years with victim as de facto spouse because relationship too remote to satisfy foreseeability test). ¶ 17 We do not find the foreseeability framework to be a particularly useful mechanism by which to ascertain and delimit a tortfeasor’s liability to a bystander for emotional distress.

If applied honestly, foreseeability would permit recovery in all situations where the ordinary person could reasonably predict that' observing injury to another would significantly distress the particular observer. Here, for example, we could not with a straight face maintain that it was not reasonably foreseeable that appellants would suffer significant emotional distress at seeing their close friend and co-worker engulfed in a ball of fire.

Alternatively, foreseeability can be skewed beyond recognition when a court, making what in effect is a policy decision, decides on the basis of a lack of foreseeability that a certain type of bystander should not be permitted to recover for emotional distress. See, e.g., Trapp v. Schuyler Construction, 149 Cal.App.3d 1140, 197 Cal.Rptr. 411, 412 (1983) (minor appellants as close friends, first cousins, and regular playmates of the minor deceased did not meet the Dillon foreseeability criteria); Kately v.

Wilkinson, 148 Cal.App.3d 576, 195 Cal.Rptr. 902, 907, (1983) (best friends akin to natural sisters insufficient to satisfy foreseeability test of Dillon). ¶ 18 Confining a liability decision to the theoretical construct of a foreseeability analysis precludes consideration of other factors that, as a matter of good policy, ought to enter into a decision concerning where the lines of liability are to be drawn. We are cautioned, correctly, that in descrying the outer limits of liability for negligence, we should not attempt to mask the policy aspects inherent in such a decision. See Rogers v.

Retrum, 170 Ariz. 399, 403, 825 P.2d 20, 24 (1991) (acknowledging policy considerations affecting determination as a matter of law that there was no negligence on the part of a school district in not having a closed campus).

A necessary corollary, we believe, is that when the liability question is one that is peculiarly susceptible to determination by reference to policy concerns, we should straightforwardly decide it on that basis and not attempt to shoehorn policy considerations into an ill-fitting theoretical construct. ¶ 19 We believe that for the,question presented here, namely what type of relationship must exist to justify recovery for bystander emotional distress, a foreseeability analysis is unworkable and cannot lead to an acceptable resolution for all circumstances.

Rather, we conclude that the determination should rest entirely on policy considerations. See Napier v. Bertram, 191 Ariz. 238, 244, ¶ 21, 954 P.2d 1389, 1395, ¶21 (1998) (concluding after review of competing policies that it would be inappropriate to recognize a negligence duty by insurance agent to non-client); cf. Bovsun v. Sanperi, 61 N.Y.2d 219, 473 N.Y.S.2d 357, 461 N.E.2d 843, 847 (1984) (“[D]elineation of limits of liability in tort actions is usually determined on the basis of consideration of public policy.”).

We turn now to a discussion of those considerations. ¶20 Militating in favor of allowing recovery is Arizona’s strong policy interest “in fully compensating injured plaintiffs to make them whole. Thus, Arizona allows unlimited recovery for actual damages, expenses for past and prospective medical care, past and prospective pain and suffering, lost earnings, and diminished earning capacity.” Bryant v. Silverman, 146 Ariz. 41, 47, 703 P.2d 1190, 1196 (1985).

Emotional distress must be included in this listing because “Arizona courts long ago abandoned a skeptical attitude toward emotional injuries and have increasingly been willing to compensate those having validity.” Barnes v. Outlaw, 192 Ariz. 283, 286, 964 P.2d 484, 487 (1998).

Moreover, the relationship of “co-worker and friend” is a valued one in our society, encompassing many of the human virtues we cherish, and is therefore worthy of consideration for protection under the law of torts when the emotional component of the relationship is negligently harmed. ¶ 21 Notwithstanding these compelling reasons, courts have been slow to extend protection beyond the ambit of the family. Cf. Consolidated Rail Corp. v.

Gottshall, 512 U.S. 532, 557, 114 S.Ct. 2396, 129 L.Ed.2d 427 (1994) (in imposing zone of danger rule for emotional distress claim under Federal Employers’ Liability Act, the court noted the common-law restrictions on such claims and commented “[w]e believe the concerns that underlie the common-law tests, and particularly the fear of unlimited liability, to be well founded”). But see Paugh v. Hanks, 6 Ohio St.3d 72, 451 N.E.2d 759, 766-67 (1983) (stating that a strict blood relationship is not required for bystander recovery).

The primary reason for hmiting the category of those who can recover for bystander distress has been expressed by a leading commentator thus: It would be an entirely unreasonable burden on all human activity if the defendant who has endangered one person were to be compelled to pay for the lacerated feelings of every other person disturbed by reason of it, including every bystander shocked at an accident, and every distant relative of the person injured, as well as all his friends. W. Page Keeton et al., Prosser and Keeton on the Law of Torts § 54, at 366 (5th ed.1984).

This worry has been echoed by many courts as they face the difficult task of drawing the line between those who may recover and those who may not. See, e.g., Elden, 250 Cal.Rptr. 254, 758 P.2d at 588 (excluding unmarried cohabitant while recognizing that such a relationship can offer as much affection, solace, and support as is found in the immediate family setting); Gates v. Richardson, 719 P.2d 193, 198-99 (Wyo.1986) (limiting bystander distress recovery to those included by the legislature in the wrongful death statute, thereby excluding “a business partner or a friend”).

A dominant concern has been the perceived need to maintain a proportionate economic relationship between liability and culpability, the failure to do which underlies much of the criticism of the foreseeability test. Richard S. Miller, The Scope of Liability for Negligent Infliction of Emotional Distress: Making “The Punishment Fit the Crime,” 1 U. Haw. L.Rev. 1, 33-36 (1979); Richard N. Pearson, Liability to Bystanders for Negligently Inflicted Emotional Harrm-A Comment on the Nature of Arbitrary Rules, 34 U. Fla.

L.Rev. 477, 512 (1982). ¶ 22 This possibility of an unreasonable burden being placed on defendants is magnified by the obvious difficulty of drawing principled limiting distinctions once bystander recovery moves beyond the category of family. If recovery is extended to a “co-worker and friend,” why not just a co-worker, or why not just a friend? The New Mexico Supreme Court’s answer was to decide that “[t]he tort of negligent infliction of emotional distress is a tort against the integrity of the family unit.” Ramirez v. Armstrong, 100 N.M. 538, 673 P.2d 822, 825 (1983), overruled on other grounds, Folz v.

State, 110 N.M. 457, 797 P.2d 246 (1990). Consequently, the court specifically limited the action to husband and wife, parent and child, grandparent and grandchild, brother and sister, and.to persons who occupy a legitimate position in loco parentis.

Ramirez, 673 P.2d at 825; see also Elden, 250 Cal.Rptr. 254, 758 P.2d at 588 (cautioning that recognizing a de facto marriage relationship opens the door to many other kinds of de facto family relationships, causing multiplication of actions and damages). ¶23 We do not need to decide here precisely where the outer limit of liability lies, whether it extends beyond the family unit or not. Because this is still a developing area of the law, see generally, J. Mark Appleberry, Negligent Infliction of Emotional Distress: A Focus on Relationships, 21 Am.

J.L. & Med. 301 (1995), we confine our holding to the specific facts presented to us, as we believe the better approach is to decide questions of qualifying relationships on a case-by-ease basis. We therefore leave for another court and another time the question whether Arizona should circumscribe recovery for bystander emotional distress as California and New Mexico have done, or whether we can, and should, recognize and protect the many quasi-family relationships that exist in today’s society. See Dunphy v.

Gregor, 136 N.J. 99, 642 A.2d 372, 380 (1994) (extending category of bystanders who can recover to unmarried cohabitant who had an “intimate familial relationship” with the decedent). ¶24 We do hold that in Arizona a coworker and friend of a negligently injured person cannot recover for the emotional distress suffered from witnessing the injury to that person.

Simply put, we conclude that the policies in favor of allowing such a recovery are insufficiently persuasive to justify placing what we perceive to be an unreasonable economic burden on negligent defendants, a perception that is shared by other courts and commentators. Requiring payment for the bystander emotional distress of a co-worker and friend would be out of proportion to the culpability inherent in conduct that is merely negligent. We therefore decline to impose such a requirement. ¶25 A brief word regarding the dissent.

Our colleague apparently believes that Keck’s use of the phrase “or otherwise” resolves the issue, concluding that our supreme court intended that the jury in each case should decide whether the particular relationship involved supports a recovery. Infra ¶¶ 30, 33. Unlike the dissent, we do not believe that the phrase conclusively settles the matter or that the Keck court intended that it should. A fair reading of the case demonstrates that the court was primarily concerned with establishing generally the limits on recovery for bystander emotional distress to which Arizona courts would adhere in the future.

When faced with such a chore, an appellate court will often choose words with the potential for a broader application than the case before it requires, not because the court presently intends the broader application, but because the court does not want to hamstring later courts who will be asked to fine-tune the seminal decision by applying it to different facts. To read Keck as broadly as does our dissenting colleague disserves this evolutionary process by foreclosing the case-by-case consideration of different relationships as each is presented.

While such a leap may sometimes be appropriate in the judicial development of tort law, it should be accompanied by more justification than the dissent offers here. ¶ 26 Our colleague also appears to conclude that we interpret Keck as definitively limiting recovery to a bystander who has a blood or familial relationship with the injured party and that we approve such a limitation. Infra ¶ 30, 33. This is not our position; rather, we reiterate that Keck neither precludes consideration of non-familial relationships nor mandates their recognition.

Whether any of the quasi-familial relationships discussed in the dissent’s cited authorities would be recognized in Arizona is a question for a court other than this one. We hold only that for policy reasons, the relationship of “co-worker and friend” will not be recognized because it is, in the words of one of the dissent’s authorities, “outside the range of circumstances within which there may be liability.” Dziok- onski v.

Babineau, 375 Mass. 555, 380 N.E.2d 1295, 1302 (1978) (in a case involving parents and an injured child, the court abandoned the “impact rule” and established a rule of recovery dependent, inter alia, on “what degree there was of familial or other relationship”). CONCLUSION ¶ 27 While we do not adopt the trial court’s foreseeability analysis, we find that it reached the correct result. We therefore affirm the grant of

summary judgment to appellees. CONCURRING: MICHAEL D. RYAN, Judge. . California has moved away from exclusive reliance on foreseeability. In Elden v. Sheldon, 46 Cal.3d 267, 250 Cal.Rptr. 254, 758 P.2d 582 (1988), a case denying recovery for bystander emotional distress to a participant in an unmarried but stable cohabitation relationship, the California Supreme Court indicated that Dillon did not foreclose considerations of policy and such considerations could "dictate a cause of action should not be sanctioned no matter how foreseeable the risk." Id. at 586; see also Thing v.

La Chusa, 48 Cal.3d 644, 257 Cal.Rptr. 865, 771 P.2d 814, 827 (1989) (affirming the role of policy considerations in defining scope of liability).