dissenting:
Res ipsa loquitur is a convenient invention of the common law; it has permitted judges and juries to make appropriate assignments of the risks of accidents where the facts do not lend themselves to the familiar contours of traditional negligence law. It has provided a flexible and practical evidentiary doctrine that has allowed fact-finders to “fill in the blanks" in cases *459where the exact causes of an accident are either unknown or unknowable. I fear, however, that my colleagues in this ease have so wrenched the res ipsa loquitur doctrine from these moorings and given it such rigid dimensions that it will no longer successfully serve those previous purposes.
The majority finds that a department store elevator door’s closing on a customer is, as a matter of law, an event that ordinarily occurs without negligence; I would hold otherwise and permit the jury to infer, from the evidence presented, that such an event ordinarily does not occur without negligence. Accordingly, I would affirm the district court’s decision to give a res ipsa loquitur instruction to the jury.
As the majority correctly states, and as the district court properly instructed the jury, “[t]he principle of res ipsa loquitur permits a jury to draw an inference of negligence upon special circumstances where direct evidence of negligence is lacking.” Maj.Op. at 455 (emphasis added) (citations omitted); see also Sweeney v. Erving, 228 U.S. 233, 240, 33 S.Ct. 416, 418, 57 L.Ed. 815 (1913). Moreover, “[although the criteria for res ipsa loquitur are generally set forth as separate requirements, they are treated by the courts as interrelated and interdependent factors.” Maj.Op. at 455.
Where the majority errs, in my view, is its assertion that “[t]he indispensable requirement that plaintiff ‘bring [the negligence] home to the defendant’ is what plaintiffs here have failed to establish, thus rendering the principle inapplicable in this case.” Id. at 456 (citation omitted). The first element of the test for res ipsa loqui-tur, however, is only that “[t]he event must be of the kind which ordinarily does not occur in the absence of someone’s negligence;” Rassoulpour v. Washington Metro. Area Transit Auth., 826 F.2d 98, 100 (D.C.Cir.1987) (emphasis added) (citation omitted). The majority fails to appreciate that there is no dispute that the accident was caused by an instrumentality in the defendant’s exclusive control and that the accident was not due to any voluntary action by the plaintiffs. It is precisely the interrelationship among the three threshold elements, not any one element, that “bring[s] the negligence home to the defendant.” I would therefore hold that plaintiffs in this case, having shown that the accident is of the type that does not ordinarily occur without negligence, that defendant had exclusive control over the instrumentality that caused the accident, and that plaintiffs did not contribute to the accident, have met the three prerequisites for permitting the jury to infer, subject to rebuttal, defendant’s negligence.
More importantly, in holding that res ipsa loquitur does not apply in this case because plaintiffs have not established that “the defendant’s negligence most probably caused the accident,” Maj.Op. at 455, the majority effectively conflates the three-part threshold test for applying the doctrine into a single requirement that the plaintiff prove by a preponderance of the evidence that the defendant was negligent. Indeed, the majority’s reasoning significantly undermines the res ipsa loquitur doctrine, because if the plaintiff could “bring the negligence home to the defendant,” i.e., prove with direct evidence that the defendant was negligent, then a res ipsa loquitur instruction would never be necessary.
The majority’s reliance on the case law is also misplaced. The doctrine of res ipsa loquitur in the District of Columbia clearly permits jurors to make a common-sense inference that defendants in this case were negligent. See, e.g., Otis Elevator Co. v. Henderson, 514 A.2d 784, 785 (D.C.1986) (affirming res ipsa loquitur instruction even though “any one or more of a number of specific mechanical defects could have led to [the cause of the accident]”); Bell v. Westinghouse Electric Corp., 483 A.2d 324, 329 (D.C.1984) (res ipsa loquitur appropriate where “direct evidence of negligence is lacking, or where there is ‘little evidence,’ or where the cause of the injury is left in doubt by the evidence or not clearly shown”) (citing Levy v. D.C. Transit System, Inc., 174 A.2d 731, 733 (D.C.1961)); McCoy v. Quadrangle Development Corp., 470 A.2d 1256, 1261 (D.C.App.1983) (for res ipsa loquitur to apply, plain*460tiff need not “eliminate with complete certainty every other possible cause or inference”); cf. Londono v. Washington Metro. Area Transit Auth., 766 F.2d 569, 571-75 (D.C.Cir.1985) {res ipsa loquitur applies where child suffered lacerations of unknown cause while riding defendant’s escalator).
The majority places great reliance on Hafferman v. Westinghouse Electric Corp., 653 F.Supp. 423 (D.D.C.1986), even though that case is not controlling here and is readily distinguishable. In Hafferman, the court found that plaintiffs had neither circumstantial evidence of defendant’s negligence nor “the slightest shred of direct evidence of negligence.” Id. at 433. The Hafferman court concluded that “[i]t is thus sheer speculation to make any conclusion as to the cause of this elevator’s failure to level properly.” Id. Unlike the plaintiffs in Hafferman, however, plaintiffs in the instant case did not rely on “sheer speculation” to allow the jury to infer defendant’s negligence. Although plaintiffs’ experts (defendant’s engineers) testified that a variety of factors could have caused the electronic safety edge to malfunction, see Trial Transcript (“Tr.”) at 16, 53, Appendix (“App.”) at 29, 66, plaintiffs also established that defendant could not identify whether or when, prior to the accident, the “wear-and-tear” parts had been properly replaced or the safety edge properly maintained, see Tr. at 17-18, 37-41, App. at 30-31, 50-54. In light of the evidence, considered as a whole, the majority’s ipse dixit that “it would be sheer speculation to conclude that the cause of the defect was due to the negligence of the defendant,” Maj.Op. at 457, is simply overstated.
If the district court decision in Haffer-man stands for the proposition assigned to it by the majority, then I think it is neither good law nor in accord with the precedents that do bind us. Cf. Prosser and Keeton on the Law of Torts § 40, at 261-62 (5th ed. 1984) (“If the defendant seeks a directed verdict in his favor, he must produce evidence which will destroy any reasonable inference of negligence, or so completely contradict it that reasonable persons could no longer accept it. * * * But if the defendant merely offers evidence of his own acts and precautions amounting to reasonable care, it is seldom that a verdict can be directed in his favor.”). We ought not set aside well-settled principles and precedents of res ipsa loquitur doctrine on such chimerical grounds.
Likewise, the majority misreads Otis Elevator Co. v. Henderson, 514 A.2d 784 (D.C.1986), in an attempt to distinguish it from the instant case. Maj.Op. at 458. The majority, conceding the factual similarities between the two cases, contends that the court in Otis Elevator “was never called upon to apply the elements of the principle at issue here to the facts presented,” and that therefore Otis Elevator is of no prece-dential value to this case. Id. at 458. I submit that, because it was beyond question that in Otis Elevator the elements of res ipsa loquitur applied to the facts, it is hardly surprising, much less a basis for distinction, that the issue was not raised in that case. In the light of everyday experience that elevator doors simply do not ordinarily close on passengers in the absence of negligence, a res ipsa loquitur instruction was proper in Otis Elevator and, by the same analysis, is appropriate in this case as well.
Finally, I emphasize that this court is in a singularly weak position to second-guess inferences properly left to the jury. When reasonable minds can differ over the inferences to be drawn from the evidence presented at trial, our system of justice entrusts the jury, which has the opportunity to observe the evidence first-hand, to resolve such disputes. In the case sub judice, because the jury may reasonably and properly have discounted the self-serving and unsubstantiated assertions of defendant’s engineers, the possibility that defendant was negligent under the proof offered was at least as great as the possibility that defendant exercised due care. I am therefore unable to conclude (as the majority does), after considering the evidence as a whole in the light most favorable to plaintiffs, as we must, see Maj.Op. at 455, that no reasonable jury could find for the plain*461tiffs. The trial judge properly denied defendant’s motions for a directed verdict, allowed the jury to infer negligence based on plaintiffs’ res ipsa loquitur theory, and denied defendant’s subsequent motion for a judgment notwithstanding the verdict. By holding otherwise, our decision today dis-serves not only the doctrine of res ipsa loquitur, but the even more important doctrine of deference to fact-finders on matters within their province. I would affirm the district court in all respects.