dissenting:
I respectfully dissent.
First, I believe that the district court abused its discretion in reaching the merits of Schneider’s pendent state law claims. Having determined preliminarily that the federal claims lacked merit, the court should have dismissed the state claims without prejudice pursuant to the Gibbs doctrine. See United Mine Workers v. Gibbs, 383 U.S. 715, 86 S.Ct. 1130, 16 L.Ed.2d 218 (1966).
Second, given that the district court did address the state law claims, I believe that the court erred in concluding that Schneider has failed to raise a genuine issue of material fact with respect to those claims. Under a proper understanding of the complaint, the record, and the relevant state law, Schneider has indeed presented a liti-gable controversy — both in contract and in tort. The majority reaches the opposite conclusion, in my view, only through a questionable application of evidentiary rules and local rules of practice that effectively excludes from consideration the “smoking-gun” deposition of a TRW employee.
I
According to United Mine Workers v. Gibbs, federal courts may assert federal-question jurisdiction over pendent state-law claims only if those claims satisfy a three-pronged test.1 First, “[t]he state and federal claims must derive from a common nucleus of operative fact ... such that [the plaintiff] would ordinarily be expected to try them all in one judicial proceeding.” Gibbs, 383 U.S. at 725, 86 S.Ct. at 1138. This first requirement is constitutional. If the various claims do not form one integrated “case or controversy,” then a federal court lacks article III authority to adjudicate the state claims. See id.; U.S. Const. art. III, § 2, cl. 1. Schneider’s claims, however, are all closely related. They all arise out of and concern the circumstances of her hiring and firing by TRW. Her complaint therefore satisfies this first prong.
Second, Gibbs states that “[t]he federal claim [upon which pendent jurisdiction is premised] must have substance sufficient to confer subject matter jurisdiction on the court.” Gibbs, 383 U.S. at 725, 86 S.Ct. at 1138 (citing Levering & Garrigues Co. v. Morrin, 289 U.S. 103, 53 S.Ct. 549, 77 L.Ed. 1062 (1933)). This requirement is also con*996stitutional. If the claims all arise from a common transaction or occurrence, then, “assuming substantiality of the federal issues, there is power in federal courts to hear the whole.” Id. (emphasis in original; footnote omitted). In the present case, the federal claims are arguably insubstantial, although not to the point of depriving the court of jurisdiction. See Levering & Garrigues Co., 289 U.S. at 105, 53 S.Ct. at 550 (“the federal question averred may be plainly insubstantial [if it is] obviously without merit”); see also Hagans v. Lavine, 415 U.S. 528, 539-41, 543, 94 S.Ct. 1372, 1380-81, 1382, 39 L.Ed.2d 577 (1974).
The problem in this case arises when one focuses upon the third prong of the Gibbs test. Even when a federal court has the constitutional power to hear pendent claims under the first two prongs, Gibbs holds that the court may (and should) refrain from ruling on those claims in certain circumstances. The court must weigh “considerations of judicial economy, convenience and fairness to litigants; if these are not present a federal court should hesitate to exercise jurisdiction over state claims.” Gibbs, 383 U.S. at 726, 86 S.Ct. at 1139. As the Gibbs Court explained:
Needless decisions of state law should be avoided both as a matter of comity and to promote justice between the parties, by procuring for them a surer-footed reading of applicable law. Certainly, if the federal claims are dismissed before trial, even though not insubstantial in a jurisdictional sense, the state claims should be dismissed as well. Similarly, if it appears that the state issues substantially predominate, ... the state claims may be dismissed without prejudice and left for resolution to state tribunals.
Id. at 726-27, 86 S.Ct. at 1139 (footnotes omitted).2 Pendent jurisdiction is ultimately a doctrine of discretion, and in the present ease, I am persuaded that the district court abused that discretion. See id. at 728, 86 S.Ct. at 1140.
There is no indication that the district court ever considered any of the Gibbs factors before ruling on the merits of Schneider’s state law claims. A weighing of the Gibbs factors, however, indisputably favors dismissal of those claims. First of all, fairness to the litigants clearly counsels dismissal because a surer-footed reading of state law would be available in state court. The apparent split within this panel over how to read California law attests to that. Secondly, concerns about comity and respect for the state judicial system also favor dismissal. Recent landmark developments in the state’s contract and tort law *997strongly urge federal judicial restraint. See, e.g., Foley v. Interactive Data Corp., 47 Cal.3d 654, 765 P.2d 373, 254 Cal.Rptr. 211 (1988).
Thirdly, interests of economy and convenience are not significantly advanced by federal adjudication. The federal claims were all dismissed before trial, and accumulated evidence bearing on the state claims may easily be carried “across the street” to the courtroom of a state superior court judge presumably better versed in the relevant law.
Finally, there can be no doubt that “state issues substantially predominate.” Schneider voluntarily withdrew her section 1983 claim when it became clear that there was no evidence of state action to support it, and the district court properly and swiftly rejected her two remaining federal claims — alleging discrimination on the basis of age and national origin — because Schneider “failed to specify any acts which can be regarded as discriminatory in nature.” Schneider v. TRW, Inc., No. 86-2369 at 3-4 (S.D.Cal. Aug. 3, 1989). Indeed, Schneider has abandoned all three of her federal claims on appeal, apparently conceding their insubstantiality. “When the state issues apparently predominate and all federal claims are dismissed before trial, the proper exercise of jurisdiction requires dismissal of the state claim[s].” Wren v. Sletten Constr. Co., 654 F.2d 529, 536 (9th Cir.1981) (per curiam) (citations omitted); see also Kitchens v. Bowen, 825 F.2d 1337, 1342 (9th Cir.1987), cert. denied, 485 U.S. 934, 108 S.Ct. 1109, 99 L.Ed.2d 270 (1988); Otto v. Heckler, 802 F.2d 337, 338, amending 781 F.2d 754 (9th Cir.1986); Schultz v. Sundberg, 759 F.2d 714, 718 (9th Cir.1985) (per curiam).
In response to these arguments, my colleagues appear to advance the dubious proposition that a district court’s decision to retain jurisdiction over pendent claims is essentially unreviewable so long as it does not violate article III — i.e., the first two prongs of the Gibbs test. See ante at 995 (“The district court had the power to decide the pendent claims in this case, and presumably thought it best to decide them.”) (emphasis added), 995 (“Unless the district court was without power to hear the case, which is not the situation here, we ought not surprise the district court [by reversing its decision to retain jurisdiction]”) (emphasis added). They therefore do not necessarily disagree with the results of my Gibbs analysis; they simply contend that it is not our role to engage in that analysis in the first place. I disagree. While it is clear that Gibbs commits the determination of whether to exercise pendent jurisdiction to the sound discretion of the district court, it is equally clear that we must review the district court’s determination for an abuse of that discretion. See Gibbs, 383 U.S. at 728, 86 S.Ct. at 1140; Mackey v. Pioneer Nat’l Bank, 867 F.2d 520, 523 (9th Cir.1989); Aydin Corp. v. Loral Corp., 718 F.2d 897, 903-04 (9th Cir.1983); Arizona v. Cook Paint & Varnish Co., 541 F.2d 226, 227 (9th Cir.1976) (per curiam), cert. denied, 430 U.S. 915, 97 S.Ct. 1327, 51 L.Ed.2d 593 (1977). I find such an abuse here. The district court completely failed to consider the question, and on the record before us, its tacit determination to proceed with the state claims was, in my view, clearly in error.
My colleagues also insist that, even if it would have been a preferable course of action originally, dismissal of the pendent claims at this late stage is unwarranted given the passage of time and the current consumption of judicial resources. Ante at 995. They further point out that neither party has challenged the district court’s retention of jurisdiction here, and they argue that it is inappropriate for us to raise the issue on our own initiative. Id.
These arguments are equally unpersuasive. As we recognized in Wren, “ ‘the issue whether pendent jurisdiction has been properly assumed is one which remains open throughout the litigation.’” 654 F.2d at 536 (quoting Gibbs, 383 U.S. at 727, 86 S.Ct. at 1139). Despite the majority’s exclusive focus upon the interests of judicial economy, those interests are not the only or even the most important factor for us to consider. Sunk costs can rarely justify injustice, and they do not outweigh the wrong wrought by a potentially incorrect and unfair application of state law. Moreover, jurisdictional matters and questions *998of comity require a court’s own constant vigilance, and where dismissal is clearly appropriate, a court need not wait for action by the parties.3
In short, this litigation essentially concerns contract and tort law and the inchoate principles of California employment law. It probably should not have been in federal court in the first place, and it certainly should not have stayed here so long. I would therefore vacate the district court’s entry of summary judgment on the pendent claims and remand for entry of an order dismissing those claims without prejudice.
II
Notwithstanding the Gibbs problem, and turning nonetheless to the merits of the pendent claims, I believe that Schneider has successfully presented triable issues, both in contract and in tort, when one views the evidence in the light most favorable to her case. I address the contract claim first and the tort claim thereafter.
A
1
Schneider disputes TRW’s contention and the district court’s conclusion that she was an at-will employee. Under California law, an at-will employee is terminable with or without good cause upon notice by either party. See Cal.Lab.Code § 2922 (West 1989); Foley, 47 Cal.3d 654, 665, 765 P.2d 373, 376, 254 Cal.Rptr. 211, 214. California law presumes that employment is terminable at will, but the presumption is rebut-table by evidence of an oral or written agreement to the contrary. The California Supreme Court’s seminal decision in Foley recognized two sorts of contractual promises that can trump the presumption: (1) express promises and (2) promises that are “implied in fact” by the employer’s representations or conduct.4 In short, to survive summary judgment on her breach of contract claim (and on any other claim whose success depends upon the refutation of at-will status), Schneider must establish a pri-ma facie case sufficient to allow a jury to infer the existence of a promise qualifying TRW’s right to terminate her without good cause — a promise that may be oral or written, express or implied.5 This much she has done.
In her complaint, briefs, and supporting affidavits, Schneider has maintained con*999sistently that she was not entirely confident in her ability to perform the job that TRW originally had in mind for her and that she accepted the offer upon the receipt of three assurances: (1) that the company would provide necessary training and supervision, (2) that the company would place her in an entry-level position for which someone with her background would normally be regarded as overqualified, and (3) that her poor command of the English language would not be a hindrance to her progress. At no point have the defendants purported to contradict or rebut Schneider’s claim that she received these specific assurances — assurances which, if proven to exist to the finder of fact, presumably would proscribe any decision to terminate her on the basis of inadequate training or poor language skills.6
On the other hand, the sworn statements of several TRW employees who worked with Schneider lend prima facie support to her claim that she did receive such assurances. Most notable among these statements is the deposition of Robert Mangum, who served as Schneider’s immediate supervisor during much of the period in question. Mangum, who is a codefendant in this action, has provided strong evidence in the case against the company and his superior, Don Rohner, the man whom Schneider mostly credits with having hired and fired her. In his deposition, Mangum explains:
Don Rohner wanted to talk to me concerning what to do with Maria Schneider. One of the requests that he made of me was, “Whatever,” quote, unquote, he said, “Whatever it takes, Bob, get rid of her.” And I didn’t really question him about it. I didn’t — in fact, I sort of ignored him. Within that week after that Don made it very clear to me that he wanted to get rid of Maria Schneider even though she was on a probationary status. He said, “Whatever it takes, Bob, get rid of her.” He says, “I don’t want her working here any more.[”] And I questioned him; and specifically he wanted me to not give her a chance to perform, not to fulfill our agreement in the verbal warning that we gave Maria. We contracted with her to give her 30 days to improve, to meet certain standards; and he said that, “That’s unnecessary, Bob. I want you to get rid of her.”
Deposition of Robert Mangum at 22 (Sept. 26, 1988) (emphasis added). Later in his deposition, Mangum states:
I was upset by the fact that [Rohner] was going to invalidate or void a con*1000tract we made with Maria in her verbal review to give her a chance to perform. That’s what I was upset about.... I was ... upset with the fact that we were going to get rid of someone in a somewhat unethical manner. In other words, I was told to fire her, not give her a chance to actually learn something.
[T]he personnel department ... agreed that there would be no disciplinary action taken on Maria, or in other words, Maria would not have to perform to her 30 day verbal warning, that was wiped clean.
Id. at 25, 27 (emphasis added). On the basis of this deposition alone, Schneider has presented a genuine issue of material fact regarding the company’s conduct and its promissory obligations.7
2
The court, however, excuses the district court’s failure to acknowledge this and several other inculpatory depositions by declaring that this critical evidence (a) is not part of the district court’s record, (b) is inadmissible hearsay, and (c) was not lodged with the district court in compliance with its local rules of practice. See ante at 990 n. 2. All three of these arguments are unpersuasive.
First, the district court itself has made none of these determinations, and yet all three involve matters that lie wholly within that court’s exclusive domain. An appellate court has neither the authority nor the occasion to make evidentiary rulings or to apply a district court’s local rules of practice. Nor has TRW advanced any of these arguments in this appeal. My colleagues have simply substituted their own judgment for that of the district court on matters that only relate to the district court and that, as far as the immediate record reveals, clearly did not concern the district court here. In so doing, they have subverted the time-honored principle that courts reviewing summary judgments must view the evidence and resolve any conflicts in the light most favorable to the nonmoving party.
3
Even assuming that it is appropriate for an appellate court to decide to exclude evidence on its own motion, my colleagues’ arguments are still unpersuasive. First, the depositions at issue are indeed part of the district court’s record: they are all file-stamped as having been lodged with the district court in a timely fashion; they appear as an entry on the district court’s docket sheet; and they comprise exactly fifty percent of Schneider's excerpts of record, the contents of which TRW has never challenged. It is simply indisputable that the depositions relied upon were “on file” with the district court when it heard and ruled upon TRW’s motion. See Fed.R. Civ.P. 56(c).
4
Second, there is absolutely no indication that the district court ever regarded the depositions in question to be inadmissible hearsay, and TRW has not challenged them as hearsay. The reason for this is understandable: to exclude these depositions on hearsay grounds would be (and is) simply wrong as a matter of law. Admittedly, they are out-of-court statements, and they therefore satisfy the technical definition of hearsay contained in Rule 801 of the Federal Rules of Evidence. See Fed.R.Evid. 801(c) (defining hearsay as “a statement, other than one made by the declarant while testifying at the trial or hearing, offered in evidence to prove the truth of the matter asserted.”) (emphasis added). Yet virtually all depositions are out-of-court statements; if Rule 801 governed their admissi*1001bility in civil actions, their value as a discovery tool would be largely ephemeral. Rule 801 therefore does not control the admissibility of depositions; rather, Rule 32 of the Federal Rules of Civil Procedure controls.8 That rule states in pertinent part that:
[a]t the trial or upon the hearing of a motion ..., any part or all of a deposition, so far as admissible under the rules of evidence applied as though the witness were then present and testifying, may be used against any party who was present or represented at the taking of the deposition or who had reasonable notice thereof....
Fed.R.Civ.P. 32(a); see also id. 32(b) (the absence or unavailability of the deponent is not a ground upon which to object to the admission of a deposition). The fact that the depositions were taken out of court is, therefore, not a bar to their admissibility. See id. advisory committee’s note on 1970 amendment (explaining that the italicized language quoted above “eliminates the possibility of certain technical hearsay objections which are based, not on the contents of [the] deponent’s testimony, but on his absence from court.”).
To the extent that my colleagues actually mean to suggest that certain statements within the depositions — rather than the depositions themselves — are inadmissible hearsay, their argument is equally uncom-pelling. Most of the inculpatory testimony within the depositions is not hearsay at all but represents the first-hand, personal observations of the deponents themselves. Most of that which does not (for example, Mangum’s accounts of what Rohner said) is not hearsay either but constitutes the admissions of a party-opponent. See Fed.R. Evid. 801(d)(2), 805.
Nor are any wow-hearsay challenges to the content of the depositions well taken. The depositions contain indisputably relevant and material testimony by apparently competent witnesses, and on that basis, the testimony they embody is admissible. See 6 J. Moore, W. Taggart & J. Wicker, Moore’s Federal Practice ¶ 56.11[l.-3] (2d ed. 1991). Moreover, TRW has made no objection to Schneider’s reliance on this testimony, and any objections at this stage would be untimely. See Fed.R.Civ.P. 32(d)(3); Fed.R.Evid. 103(a)(1). As an appellate court obligated to view the evidence in the light most favorable to Schneider, we are certainly in no position to raise objections of our own.
5
The court’s invocation of the district court’s local rules of practice is similarly inappropriate. The local rule in question requires that a party opposing another party’s motion file a written opposition that “contain[s] ... copies of all documentary evidence upon which the party in opposition relies.” S.D.Cal.R. 220 — 8(b)(3), quoted ante at 990 n. 2. The court implies that because Schneider did not physically attach the relevant portions of the most damning depositions to her memorandum, the testimony in those depositions was not properly before the district court. I disagree.
The Mangum deposition and the others upon which Schneider relies were lodged with the district court on April 17, 1989, only seven days after Schneider filed her original memorandum in opposition to sum*1002mary judgment and seven days before the district court’s first hearing on the motion.9 Schneider clearly intended this later filing to amend her earlier memorandum, which is replete with specific references to the depositions. To deny her reliance on this critical evidence upon the mere technicality that the two documents were filed one week apart and were not physically attached — or alternatively, on the technicality that she did not refile or re-cite the depositions when she opposed TRW’s renewal of its motion less than three months later — is harsh justice indeed. See supra note 9; see also 10A C. Wright, A. Miller & M. Kane, Federal Practice and Procedure § 2721, at 44 (1983) (court “will consider all papers of record” before granting summary judgment); Higgenbotham v. Ochsner Found. Hosp., 607 F.2d 653, 656-57 (5th Cir.1979) (“The rule [Fed.R.Civ.P. 56(c) ] does not distinguish between depositions merely filed and those singled out by counsel for special attention. Nor is this a case where the deposition that created the dispute was a needle in a paper haystack; the docket sheet shows that only five depositions were filed.”) (footnote omitted).
To bolster its invocation of the district court’s local rules, the court cites Nilsson, Robbins, Dalgam, Berliner, Carson & Wurst v. Louisiana Hydrolec, 854 F.2d 1538 (9th Cir.1988), for its holding that the district court was “under no obligation to mine the full record” for triable issues of fact. Ante at 991 n. 2. In light of the original opposing memorandum’s very specific citations to the relevant depositions, however, no such mining operation is called for in this case. Nilsson, moreover, is distinguishable. There, the non-moving party had twice failed to answer the mov-ant’s pleadings and had “repeatedly refused to conform to the Local Rules and the Federal Rules of Civil Procedure.” Nilsson, 854 F.2d at 1541. The district court had entered four orders for monetary sanctions and had imposed and lifted three default judgments, and these measures still “did not alleviate this behavior.” Id.; see id. at 1541-42. In addition, the non-moving party
w[as] put on notice by [the movant] eleven months before the summary judgment motion was ruled upon, that its memorandum in opposition to the motion for summary judgment was legally insuffi*1003cient because of the lack of evidentiary support^ and t]he trial judge, in her tentative ruling, [had] highlighted for [the non-moving party] the flaws in their op-position_ [The non-moving party], in the face of all these warnings that summary judgment would be granted unless they adequately supported their opposition to the ... motion, neglected to cite the court to a single evidentiary fact.
Id. at 1545 (emphasis added).10
To equate Schneider’s slight oversight, which was swiftly remedied, with this pattern of persistent misconduct is patently strained. Moreover, there is no evidence in this case that the district court ever gave Schneider any comparable warnings. In any event, the Nilsson court clearly based its holding first and foremost on the non-moving party’s failure to cooperate and only secondarily on the relevant local rule of practice. See id. at 1545. There can be no doubt that an obstructive failure to cooperate is costly and prejudicial to the mov-ant; there is, however, no similar sign of prejudice in the current case.
In no way do I mean to imply that the observance of local rules of practice need not be strict, and I would not suggest that one can disregard such rules with impunity. Local rules serve a necessary and important function. My complaint is not with the rules themselves but with the present application of one of those rules. The court has applied this rule unfairly and in a manner wholly inconsistent with the Federal Rules of Civil Procedure. See Fed.R. Civ.P. 83 (a district court may “make and amend rules governing its practice not inconsistent with these rules.”); see also id. 1 (purpose of federal rules is to secure outcomes that are both efficient and just); 56(c) (summary judgment not proper where “pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any,” reveal a genuine issue of material fact). Perhaps most importantly, the court has undertaken to apply this rule retroactively from the appellate stage — even though the district court’s record is devoid of any reference to its local rules and even though there is no indication that the district court would have applied this particular rule to the same end.
6
Finally, in addition to denying Schneider’s reliance on critical depositions, the court affirms the district court’s conclusion that her evidence has failed to rebut California’s at-will presumption. For this holding, the court cites Pugh v. See’s Candies, Inc., 116 Cal.App.3d 311, 171 Cal.Rptr. 917 (1981). See ante at 990. In Pugh, a case favorably cited by the Foley court, the California Court of Appeal reinstated a former employee’s suit against his former employer on grounds that the trial judge had prematurely terminated the action. The court held that the plaintiff had successfully demonstrated a prima facie case of wrongful termination by introducing evidence from which a reasonable jury could infer (a) the existence of an implied promise not to fire him without cause and (b) an unlawful breach of that promise. See Pugh, 116 Cal.App.3d at 329, 171 Cal.Rptr. at 927.
I agree that Pugh is on point. I further believe that, like Wayne Pugh, Maria Schneider has demonstrated a prima facie case from which a jury might infer both the existence and the breach of a promise not to terminate her for certain reasons. The court reaches the opposite conclusion by focusing upon four “factors” mentioned *1004by the Pugh court as indicative of whether the parties have an implied agreement modifying their ability to terminate their relationship at will.11 As the Pugh court itself explained, however, these four factors are neither dispositive nor all-inclusive. “[I]t is appropriate to consider the totality of the parties’ relationship: Agreement may be ‘shown by the acts and conduct of the parties, interpreted in the light of the subject matter and the surrounding circumstances.’” Id., 116 Cal.App.3d at 329, 171 Cal.Rptr. at 927 (quoting Marvin v. Marvin, 18 Cal.3d 660, 678 n. 16, 557 P.2d 106, 118 n. 16, 134 Cal.Rptr. 815, 827 n. 16 (1976)); see also Foley, 47 Cal.3d at 681, 765 P.2d at 388, 254 Cal.Rptr. at 226 (“the totality of the circumstances determines the nature of the contract”) (citing Pugh).
Thus, although the majority correctly points out that Schneider has admitted an awareness of “no TRW policies suggesting assurances of continued employment,” that observation does nothing to deny her contention that TRW made particular contractual assurances in her case. Similarly, the absence of an industry-wide custom of terminating employees only for good cause is relevant but not dispositive, and the fact that Schneider was only with the company for nine months does not necessarily mean that she had no reasonable expectations of job security. In fact, in light of Schneider’s claim that TRW promised to train her, the shortness of her tenure would, if anything, seem to support an inference that the company reneged on that promise. See supra note 7. Finally, despite the majority’s contention to the contrary, Schneider clearly does argue that TRW’s actions and communications conveyed contractual obligations. See Brief for Appellant at 14-17.
In short, to have granted summary judgment on Schneider’s breach of contract claim, the district court must have concluded that Schneider failed to present even a prima facie case sufficient to support the finding of a promise modifying the company’s right to fire her. I do not believe that the record supports such a conclusion. I would therefore reverse the district court’s entry of summary judgment on the breach of contract claim.
B
With respect to Schneider’s tort claims, the district court essentially held that Schneider’s status as an at-will employee precluded these theories of recovery. See Schneider, No. 86-2369 at 3. If the court is correct in concluding that Schneider was an at-will employee, then its tort analysis is also largely correct. The at-will concept would be ephemeral if employees who, as a matter of contract, are legally terminated without good cause could turn around and sue their employers in tort. The California Supreme Court recognized this fact in Foley. See Foley, 47 Cal.3d at 698 n. 39, 765 P.2d at 400 n. 39, 254 Cal.Rptr. at 238 n. 39. It therefore held that, absent a specific public policy or legislation to the contrary, an at-will employee who seeks to challenge a termination decision cannot sue his former employer for tort damages or for breach of implied covenants of good faith and fair dealing.12
Two aspects of the district court's analysis, however, call its reliance on Foley into question. First, as I have already mentioned, the evidence does not indisputably *1005establish that Schneider was an at-will employee. In my view, she has presented a prima facie case sufficient to put the question before a jury. Second, the district court has fundamentally misconstrued the nature of her tort claims. The company, she contends, was not fraudulent and deceitful when it fired her; it was fraudulent and deceitful when it hired her. In her view, TRW held out assurances to induce her to accept its offer and to move from San Francisco to San Diego without ever intending to honor those assurances. See Complaint at 111118-19; Memorandum of Points and Authorities in Opposition to Defendants’ Motion for Summary Judgment at 8. A tort claim based upon such allegations remains unaffected by the logic of Foley because it concerns what is in essence pre-contractual conduct. Even if Schneider was an at-will employee, she presumably can still sue the defendants in tort with respect to actions that preceded their employment relationship. Allowing such claims to proceed does not erode an employer’s ability to fire at will.13
Recognizing that Foley does not foreclose these claims, however, only answers half the question. One must still determine whether Schneider has introduced evidence of tortious conduct sufficient to survive summary judgment. I believe that she has, and the Mangum deposition, alone, is sufficient to present a prima facie case. First, Mangum’s statement makes it clear that TRW knew about Schneider’s inexperience and her poor command of the English language when it hired her and that the company assured her, both at the time of hiring and subsequently, that it would provide adequate training. Then, in a critical portion of the deposition, Mangum states:
Mr. Rohner believed that once he was given an approved requisition, in other words, a promise by ... upper management to alot [sic] certain dollars for a hire, he thought if he didn’t fill it within seven days he would lose his requisition. I’m serious. This is why he hired Maria. She was the first warm body that he saw that would fit the bill and he was scared to death he would lose that money, and so he was trying to build a kingdom or, I don’t know, he’s not the best manager in the world; but he definitely — I swear to God I couldn’t believe it.
Q. Did he tell you that?
A. Yes, he was scared to death he was going to lose that requisition, so he wanted to fill it. See, Don Rohner’s position was, you can always fire someone who doesn’t work out, and guess what, you still have an opening to fill. I am not making that up, that’s Don Rohner’s philosophy[.]
Deposition of Robert Mangum at 36 (Sept. 26, 1988). Again, on the basis of this deposition alone, Schneider has presented triable issues in tort sufficient to survive summary judgment.
Ill
In sum, I would affirm the district court’s judgment for the defendants on the federal claims. I would reverse its judgment for the defendants on the pendent state-law claims and would remand for entry of an order dismissing those claims without prejudice. Alternatively, I would remand for further proceedings in the district court on the merits of the pendent claims.14 Summary judgment is summary *1006justice, and the record in this case does not warrant summary justice.
. Congress has adopted the Gibbs test in recent legislation. See Judicial Improvements Act of 1990, Pub.L. No. 101-650, § 310, 104 Stat. 5089, 5113-14 (1990) (to be codified at 28 U.S.C. § 1367) (effective Dec. 1, 1990).
. My colleagues assert that the foregoing language is dictum and that the Supreme Court itself has held that this dictum " ‘does not establish a mandatory rule to be applied inflexibly in all cases.’" Ante at 993 (quoting Carnegie-Mellon Univ. v. Cohill, 484 U.S. 343, 350 n. 7, 108 S.Ct. 614, 619 n. 7, 98 L.Ed.2d 720, 730 n. 7 (1988)). There are two responses to this claim. First, I do not contend for a moment that the prudential concerns in Gibbs establish a mandatory or inflexible rule. I simply see this as "the usual case in which federal-law claims are eliminated before trial” and in which, as a result, “the balance of factors ... point[s] toward declining to exercise jurisdiction over the remaining claims.” Carnegie-Mellon, 484 U.S. at 350 n. 7, 108 S.Ct. at 619 n. 7. To have retained jurisdiction over the state claims here, the district court must have concluded that this was an unusual case, and I do not believe that such a conclusion is supportable. See Notrica v. Board of Supervisors, 925 F.2d 1211, 1213-14 (9th Cir.1991) (quoting Camegie-Mellon and affirming dismissal of pendent claims over appellant's contention that his was an unusual case requiring retention of federal jurisdiction).
Second, whether they are dicta or not, the considerations articulated in Gibbs have been applied and relied upon by courts for years. Indeed, they were recently reaffirmed as the appropriate “factors to be considered under the pendent jurisdiction doctrine” by the Supreme Court itself in Carnegie-Mellon, 484 U.S. at 350 n. 7, 108 S.Ct. at 619 n. 7, and they were essentially codified by Congress in the Judicial Improvements Act of 1990. See H.R.Rep. No. 734, 101st Cong., 2d Sess. 29, reprinted in 1990 U.S. Code Cong. & Admin.News 6802, 6875 (Act “codifies the factors recognized as relevant under current law” to the determination of whether federal courts should entertain pendent claims). Moreover, virtually all judicially announced policy considerations are dicta in a literal sense, but the taint of that word alone does not vitiate their force or immunize the district courts that ignore them. The "dicta” in Miranda v. Arizona, 384 U.S. 436, 467-73, 86 S.Ct. 1602, 1624-27, 16 L.Ed.2d 694 (1966), come to mind as one obvious example.
. Incidentally, there is little danger that a state court would dismiss Schneider’s action, if she were to refile in state court, as time-barred. California has previously held that a timely filing in federal court may equitably toll passage of the applicable limitations period. Addison v. State, 146 Cal.Rptr. 224, 578 P.2d 941, 21 Cal.3d 313 (1978); see also Notrica, 925 F.2d 1211 (no abuse of discretion to dismiss pendent claims even where evidence suggests that refiling of those claims in state court could be time-barred).
. See generally Foley, 47 Cal.3d at 671-82, 765 P.2d at 380-88, 254 Cal.Rptr. at 218-27; see also id. 47 Cal.3d at 675, 765 P.2d at 383, 254 Cal.Rptr. at 221 ("the fact that it was an implied or oral agreement is not fatal to its enforcement”).
. It is important to emphasize that the question is not whether Schneider had a contract with TRW; she clearly did. Even an at-will relationship is contractual. Rather, the question is whether she had the benefit of a promise modifying the conditions under which she might be terminated. See Foley, 47 Cal.3d at 676, 765 P.2d at 384, 254 Cal.Rptr. at 222 (“A review of other jurisdictions also reveals a strong trend in favor of recognizing implied contract terms that modify the power of an employer to discharge an employee at will.”) (citing numerous cases from other jurisdictions); Thompson v. St. Regis Paper Co., 102 Wash.2d 219, 229, 685 P.2d 1081, 1088 (1984) ("if an employer, for whatever reason, creates an atmosphere of job security and fair treatment with promises of specific treatment in specific situations and an employee is induced thereby to remain on the job ..., those promises are enforceable components of the employment relationship.”) (emphasis in original), quoted in Foley, 47 Cal.3d at 676, 765 P.2d at 384, 254 Cal.Rptr. at 222.
It is also important to emphasize that Schneider’s burden at the summary judgment stage is not to prove the existence of such a promise but only to introduce evidence that, when viewed in the light most favorable to her, suffices to render the issue a triable question of fact for the jury to decide. See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248, 106 S.Ct. 2505, 2510, 91 L.Ed.2d 202 (1986) (triable issue exists "if the evidence is such that a reasonable jury could return a verdict for the nonmoving party”); cf. Pugh v. See’s Candies, Inc., 116 Cal.App.3d 311, 329, 171 Cal.Rptr. 917, 927 (Cal.Ct.App.1981) (error to grant judgment for defendant-employer where "there were facts in evidence from which the jury could infer the exist*999ence of such an implied promise [not to terminate without good cause]"). Under California law, a prima facie case that such a promise exists shifts the burden of production to the employer to introduce evidence explaining the reason for the plaintiff’s firing. See id. The plaintiff, of course, still maintains the ultimate burden of proving wrongful termination, see id., 116 Cal.App.3d at 330, 171 Cal.Rptr. at 927, but that ultimate burden has not been implicated in this case because the district court based its holding on the threshold determination that Schneider’s evidence is insufficient to rebut the at-will presumption. In short, the majority is wrong to suggest that "Schneider bears the burden of proving that she was wrongfully discharged;” her burden in order to survive summary judgment is significantly less onerous. Ante at 990.
. The defendants have maintained in general terms that Schneider was an at-will employee. They have also argued that "Schneider was one of a number of employees laid off from the company during a reduction in [work] force in 1985” and that, although unnecessary, "good cause therefore existed for her termination.” Brief for Appellees at 1. In the course of granting summary judgment for the defendants, the district court concluded that the company ”ha[d] established that the plaintiff’s termination was occasioned by [this] reduction in work force[, which was] due to economic circumstances [within her] division." Schneider v. TRW, Inc., No. 86-2369 at 2 (emphasis added). Schneider, however, has seriously challenged the credibility of this explanation. She argues that the company posted an advertisement to fill her position on December 18, 1985, only nine days after her termination. Presumably, if her firing had been pursuant to an economic lay-off, she would have been the first person called in an effort to refill that position. This observation and the evidence discussed below suggest that the company’s explanation may be pretex-tual. In any event, in my view, the district court clearly overstepped its bounds in resolving such a disputed question of fact. See Liberty Lobby, 477 U.S. at 249, 106 S.Ct. at 2511 ("it is clear enough from our recent cases that at the summary judgment stage the judge's function is not himself to weigh the evidence and determine the truth of the matter but to determine whether there is a genuine issue for trial.”).
. Moreover, in his own deposition, Don Rohner admits that he told Schneider at the time of hiring that he personally would be training her; that he was later unable to do so because the company changed his responsibilities; and that it was clear to him that "somebody in a position of knowledge still needed to train Maria." Deposition of Don Rohner (Nov. 17, 1988). If one can reasonably conclude from this evidence that the company had a contractual obligation to train Schneider, then one might also conclude that the company reneged on that promise. Schneider was fired on December 9, 1985, only nine months after she began working at the firm and after what would seem to be a rather short trial period for a trainee.
. See Fed.R.Evid. 802 (“Hearsay is not admissible except as provided by these rules or by other rules prescribed by the Supreme Court pursuant to statutory authority or by Act of Congress.”) (emphasis added); id. advisory committee’s note (listing Federal Rule of Civil Procedure 32 as an example of the rules that trump the hearsay rule under Federal Rule of Evidence 802); 4 D. Louisell & C. Mueller, Federal Evidence § 433, at 391-92 (1980) ("Perhaps the most important provisions to which Rule 802 defers are those which regulate the use of depositions in the actions in which they were taken.... FRCP 32 governs the use of depositions in civil cases in the proceedings in which they were taken.... [I]n the event of any conflict between [the hearsay rule and Rule 32], it is FRCP 32 which controls.’’); 8 C. Wright & A. Miller, Federal Practice and Procedure § 2142, at 449 (1970 & Supp. 1991) ("Rule 32 defines the circumstances in which a deposition is admissible to any extent.”); 4A J. Moore & J. Lucas, Moore's Federal Practice ¶32.02[1], at 32-11 (2d ed. 1991) (“The Evidence Rules expand to some extent the admissibility of depositions ..., but do not limit admissibility as provided for in Rule 32.”) (emphasis added).
. Schneider filed her original memorandum in opposition to summary judgment, together with her own deposition, on April 10, 1989. She filed four other depositions on April 17, and the district court held a hearing on the motion on April 24. Thereafter, the court granted Schneider leave to amend her complaint, and Schneider filed a fourth amended complaint on May 15, 1989. The defendants then filed a “Renewed Motion for Summary Judgment,” along with a memorandum in support, on June 14, 1989. Schneider filed her renewed opposition on July 3; the defendants filed a reply on July 10; another hearing was held on July 17; and the court granted summary judgment on August 4.
Because the motion upon which the court ultimately acted was a renewed motion, it follows that any documents that were properly put before the court with respect to the original motion were still before the court when it reconsidered the issue. The defendants have erased any doubt in this regard with the following recital, which appears in the notice of their renewed motion:
The motion will be made and based upon this notice and motion, the memorandum of points and authorities and other supporting papers previously filed with this court by the defendants on January 13, 1989, the memorandum of points and authorities in opposition to the motion previously filed with this court by the plaintiff, the reply memorandum of points and authorities previously filed with this court by defendants on April 17, 1989, the memorandum of points and authorities filed contemporaneously herewith, together with all of the pleadings and records on file in this matter.
(Emphasis added). Of course, if the majority’s strict application of Rule 220-8 is to be consistent, then this court must disregard (and assume that the district court disregarded) all of the above-listed items that were not actually reproduced and resubmitted as attachments to the defendants' memorandum in support of the renewed motion. Rule 220-8(a)(l) places a requirement analogous to that of Rule 220-8(b)(3) on the party who submits a motion. See S.D. Cal.R. 220-8(a)(l). If one adopts such a strict reading, however, it becomes even more clear that the defendants do not have enough support for their motion.
. See also id. at 1541:
One week prior to th[e partial summary judgment] hearing, the trial judge advised [the non-moving party] that their opposition was inadequately supported and that ... she planned to grant the motion_ [After obtaining a continuance, the non-moving party] failed to buttress their argument against the entry of summary judgment with any eviden-tiary materials, even though they had been advised that their failure to do so had inclined the district court judge to grant the motion for partial summary judgment.
. See id., 116 Cal.App.3d at 326-27, 171 Cal. Rptr. at 925-26 (footnotes omitted):
In determining whether there exists an implied-in-fact promise for some form of continued employment courts have considered a variety of factors in addition to the existence of independent consideration. These have included, for example, [1] the personnel policies or practices of the employer, [2] the employee’s longevity of service, [3] actions or communications by the employer reflecting assurances of continued employment, and [4] the practices of the industry in which the employee is engaged.
. See generally id., 47 Cal.3d at 665-71 & 682-700, 765 P.2d at 376-80 & 389-401, 254 Cal.Rptr. at 214-18 & 227-39. Under some circumstances, California law recognizes the breach of implied covenants as a tort.
. Like the district court, the majority erroneously concludes that Schneider's implied covenant claims are barred by the logic of Foley and by Schneider’s alleged status as an at-will employee. The majority then concludes that Schneider has "abandoned” her other tort claims for fraud and deceit. I disagree. Schneider correctly argues that the district court’s judgment on these latter claims was made "dependent upon the Court’s erroneous ruling that, as a matter of law, Plaintiff was an at-will employee." Brief for Appellant at 27. If one agrees that the district court’s finding of at-will status was not justified, then one cannot fairly deem the fraud and deceit claims abandoned.
. There is one exception: I do agree with the court’s analysis and judgment with respect to *1006the emotional distress claim. I therefore would only reinstate Schneider’s claims for breach of contract, fraud, deceit, and breach of implied covenants.