We interpret in this appeal several provisions of the qui tam section of the False Claims Act, 31 U.S.C. Sec. 3729 et seq.

Appellant David Siller, the brother and employee of a former distributor of Becton Dickinson & Company products, brought a civil qui tam action under 31 U.S.C. Sec. 3730 against Becton Dickinson & Company, alleging that the company had a practice of overcharging the government. The government ultimately elected to intervene in and proceed with the case pursuant to 31 U.S.C. Sec. 3730(b)(2)-(b)(4). On Becton Dickinson & Company's motion, the district court dismissed the government as a party plaintiff, holding that the government's failure to comply with section 3730(b)(4)'s timely intervention requirement barred the government from proceeding. The district court also dismissed Siller's action under section 3730(e)(4), holding that it was "based upon" publicly disclosed allegations. The government and Siller appeal the district court's order dismissing their complaints. For the reasons that follow, we vacate the district court's dismissal order and remand for further proceedings with the government and Siller reinstated as plaintiffs. 813 F. Supp. 410 (1993).

From January 1986 until the filing of this lawsuit in January 1991, David Siller was employed at various times, and in various capacities by Scientific Supply, Inc. (SSI), a San Antonio, Texas, distributor of health care products whose president was Siller's brother, Ruben Siller. SSI was an authorized distributor of medical device products manufactured by Becton Dickinson & Company (BD) until BD canceled its distributorship agreement in 1987.

In 1989, SSI filed suit against BD in Texas state court, asserting various causes of action arising from BD's allegedly wrongful termination of its distributorship agreement. The thrust of SSI's complaint was that BD canceled SSI's distributorship because it feared that SSI, which was seeking to sell BD products to the federal government at prices below those quoted by BD itself, would disclose that BD was overcharging the government. See J.A. at 91-99.1 BD ultimately settled with SSI in September 1989. As part of the settlement agreement, Ruben Siller agreed to keep the existence and terms of the settlement confidential. David Siller, however, Ruben's brother and employee, was not similarly bound by the settlement agreement.

David Siller filed the instant qui tam suit against BD in January 1991. According to Siller, he originally learned that BD overcharged the government through his employment with SSI, not as a result of SSI's suit against BD. In fact, Siller asserts, he obtained this knowledge before the SLS and SSI complaints were filed and had not read those complaints until BD filed its motion to dismiss in January 1993. Siller contends that he learned about the False Claims Act (FCA) and its qui tam provisions in the spring of 1990, after his brother's company's suit against BD was settled, and that he subsequently conducted his own investigation which uncovered evidence revealing how BD overcharged the government and attempted to conceal those overcharges. Siller then retained counsel and, on December 27, 1990, as required by section 3730(b)(2), voluntarily disclosed the evidence he had garnered regarding BD's overcharging practices to the Office of Inspector General for the Department of Veterans' Affairs, the agency BD allegedly overcharged. Siller filed his lawsuit against BD on January 4, 1991. Pursuant to the FCA's requirements, see id., Siller's complaint remained under seal until the government decided whether to proceed with the case.

The government's decision was 21 months in coming. Between Siller's filing of the lawsuit and the government's ultimate decision to intervene, the government moved for, and received, eight extensions of time in which to consider whether to intervene. See 31 U.S.C. Sec. 3730(b)(3). On two occasions, the government failed to comply even with the court's extended deadline.

The government first missed its March 4, 1992, deadline. Although the Assistant United States Attorney (AUSA) investigating the matter claims to have "made sure" the government's motion for an extension of time was placed in the district court's night filing box on March 4, 1992, Supp.J.A. at 36-37, the motion apparently was never processed by the clerk's office, and there is no record of the motion ever having been filed. When, three weeks later, the court notified the government that it had not received any papers before the March 4 deadline, the AUSA filed a motion for a nunc pro tunc extension of time, which Judge Legg granted on March 27, 1992.

The government subsequently failed to meet its August 3, 1992, deadline for electing whether to intervene, not filing its motion for another extension of time until August 5, 1992. Judge Legg again granted the government's motion, allowing the government until October 5, 1992, to make its decision. On October 5, the government filed a notice of election to intervene and notice of appearance, and the complaint was unsealed.

BD moved on January 19, 1993, to dismiss the case. On February 10, Judge Smalkin, to whom the case had been re-assigned, granted BD's motion. Judge Smalkin dismissed the government as a party plaintiff, holding that each of the government's two previous failures to meet its deadlines violated the FCA's mandatory timely intervention requirement, see 31 U.S.C. Sec. 3730(b)(4), divesting the court of its jurisdiction over the government's complaint. Judge Smalkin also held that Siller's action was barred under section 3730(e)(4) because it was "based upon" the prior public disclosure of allegations that BD overcharged the government which appeared in SSI's 1989 lawsuit against BD. Judge Smalkin thereafter denied the government's motion to alter or amend its judgment, and ordered the action dismissed.

We first address the government's claim that the district court erred in dismissing it as a party plaintiff. Whether the government's failure to comply with section 3730(b)(4)'s timeliness requirement jurisdictionally bars the government from proceeding with the case is a question of first impression in this, or any other, federal circuit.2 Section 3730 of the False Claims Act provides that private persons may bring civil qui tam actions for alleged violations of the FCA. 31 U.S.C. Sec. 3730(b). Under this provision, a qui tam plaintiff is instructed to serve on the government his complaint and written disclosure of material evidence. 31 U.S.C. Sec. 3730(b)(2). The government then has 60 days, during which time the relator's complaint is filed in camera and remains under seal, within which to elect whether to intervene and proceed with the case, id., although the government may, for good cause shown, move the court for extensions of time, 31 U.S.C. Sec. 3730(b)(3). Section 3730(b)(4) further requires that

[b]efore the expiration of the 60-day period or any extensions obtained under [section 3730(b)(3) ], the Government shall--

(A) proceed with the action, in which case the action shall be conducted by the Government; or

(B) notify the court that it declines to take over the action, in which case the person bringing the action shall have the right to conduct the action.

31 U.S.C. Sec. 3730(b)(4).

The district court found that the government failed to prove that it had actually filed papers with the clerk's office on or before its March 4, 1992, and August 3, 1992, deadlines. Construing section 3730(b)(4) as a mandatory jurisdictional timeliness requirement, it held that each of the government's two failures to comply with that section divested the court of jurisdiction over the government's complaint, and ordered the United States dismissed.

The government, evidently conceding that it failed to meet its March 4, 1992, and August 3, 1992, deadlines, contends that section 3730(b)(4) is not a jurisdictional requirement, and therefore that its failure to comply with that section's timely intervention directive did not bar Judge Legg from excusing the government's untimeliness and allowing it to proceed. Under an established line of authority implicitly approved by the Supreme Court in Brock v. Pierce County, 476 U.S. 253, 106 S. Ct. 1834, 90 L. Ed. 2d 248 (1986), the government argues, a government employee's failure to meet a statutory deadline does not bar further government action unless Congress has, as it has not in section 3730(b)(4), expressly specified such a consequence for the government's tardiness. BD responds that Pierce County is not controlling, and that the courts must enforce section 3730(b)(4)'s unambiguous timeliness requirement by dismissing the government for noncompliance. We agree with the government that Pierce County controls this case and, although for reasons slightly different from those advanced by the government, hold that Pierce County requires us to read section 3730(b)(4) as a non-jurisdictional statutory deadline.

The Supreme Court in Pierce County considered the question whether the government's noncompliance with the Comprehensive Employment and Training Act (CETA) requirement that the Secretary of Labor "shall" issue a final determination as to the misuse of CETA funds within 120 days after receiving a complaint alleging such misuse, 29 U.S.C. Sec. 816(b) (1976 ed., Supp. V), divested the Secretary of his power later to recover the allegedly misused funds. In ruling that it did not, the Supreme Court refused to embrace either of the polar textual arguments advanced by the parties.

On the one hand, the Court rejected the reasoning of the Court of Appeals decision it was reviewing, which reasoning was championed by the county in that case and is relied upon here by BD, that Congress' use of the word "shall" in section 816(b) itself manifests a conclusive intent that noncompliance with the statutory deadline be a jurisdictional bar to later governmental action. 476 U.S. at 262, 106 S.Ct. at 1840. Repeating its "frequently articulated ... 'great principle of public policy, applicable to all governments alike, which forbids that the public interests should be prejudiced by the negligence of the officers or agents to whose care they are confided,' " id. at 260, 106 S.Ct. at 1839 (quoting United States v. Nashville, C. & St. L. Ry. Co., 118 U.S. 120, 125, 6 S. Ct. 1006, 1008, 30 L. Ed. 81 (1886)) (other citations omitted); see Office of Personnel Management v. Richmond, 496 U.S. 414, 433, 110 S. Ct. 2465, 2476, 110 L. Ed. 2d 387 (1990) ("It ignores reality to expect that the Government will be able to 'secure perfect performance from its hundreds of thousands of employees scattered throughout the continent.' ") (quoting Hansen v. Harris, 619 F.2d 942, 954 (2d Cir.1980) (Friendly, J., dissenting), rev'd sub nom. Schweiker v. Hansen, 450 U.S. 785, 101 S. Ct. 1468, 67 L. Ed. 2d 685 (1981)), the Court admonished that where "there are less drastic remedies available for failure to meet a statutory deadline, courts should not assume that Congress intended the agency to lose its power to act." 476 U.S. at 260, 106 S. Ct. at 1839 (footnote omitted); see United States v. Montalvo-Murillo, 495 U.S. 711, 717-18, 110 S. Ct. 2072, 2077, 109 L. Ed. 2d 720 (1990) ("There is no presumption or general rule that for every duty imposed upon the court or the Government and its prosecutors there must exist some corollary punitive sanction for departures or omissions, even if negligent.") (citation omitted). Finding that a less drastic remedy was available in the form of an action to enforce CETA's 120-day deadline, 476 U.S. at 260 n. 7, 106 S. Ct. at 1839 n. 7, the Court held that "the mere use of the word 'shall' in [29 U.S.C. Sec. 816(b) ], standing alone, [was] not enough to remove the Secretary's power to act after 120 days."3 Id. at 262, 106 S. Ct. at 1840 (footnote omitted).

On the other hand, the Court declined in Pierce County to embrace the Secretary's antipodal argument, urged here by the government, that a statutory deadline for agency action can never bar later action unless Congress has explicitly specified such a consequence in the statute itself. Id. at n. 9. The Court did recognize the established line of Court of Appeals precedent that adopted such a position, see, e.g., St. Regis Mohawk Tribe, New York v. Brock, 769 F.2d 37, 41 (2d Cir.1985), cert. denied, 476 U.S. 1140, 106 S. Ct. 2245, 90 L. Ed. 2d 692 (1986) and acknowledged that its own decisions articulating the general principle that official negligence should not be allowed to prejudice the public interest constituted the underpinnings for that precedent. Id., 476 U.S. at 259-60, 106 S. Ct. at 1838-39. However, the Court expressly refused to draw a conclusive inference from Congress' failure to specify a consequence for failing to comply with section 816(b), and instead resorted to other indicia of congressional intent to determine whether Congress meant noncompliance with section 816(b)'s deadline to bar subsequent agency action. Id. at 262 n. 9, 262-64, 106 S. Ct. at 1840, n. 9, 1840-41. Only after finding that CETA's history and purpose provided no evidence of such an intent, did the Court hold that CETA's 120-day deadline was not a jurisdictional requirement. Id. at 262-66, 106 S.Ct. at 1840-42.

Thus, the only principle, unorthodox as it may be, that emerges from Pierce County is that, where a statutory deadline requiring that the government "shall" take certain action within a particular time frame fails to specify the consequences of the government's failure to comply with that deadline, courts should not assume from the statute's mandatory language itself that a jurisdictional requirement was intended, if a remedy for the government's noncompliance less drastic than dismissal is available. Id. at 262, 106 S. Ct. at 1840. Rather, in such a context, they should examine the "normal indicia of congressional intent," id. at n. 9, to determine whether Congress meant the provision to be jurisdictional. Id. at 262, 106 S. Ct. at 1840.

Applying Pierce County in the case before us requires a holding that section 3730(b)(4) is not a jurisdictional requirement. As in Pierce County, the only textual evidence that Congress intended section 3730(b)(4) to be jurisdictional is the statutory requirement that the Secretary "shall" elect whether to intervene in a timely fashion. Compare, e.g., 31 U.S.C. Sec. 3730(e)(1)-(e)(4) ("No court shall have jurisdiction over an action ..."); 28 U.S.C. Sec. 2415(a) (West Supp.1993) ("every action for money damages brought by the United States ... which is founded upon any contract ... shall be barred unless the complaint is filed within six years ...") (emphasis added); 28 U.S.C. Sec. 2415(b) (West Supp.1993) ("every action for money damages brought by the United States ... which is founded upon a tort shall be barred unless the complaint is filed within three years ...") (emphasis added). Also as in Pierce County, remedies less drastic than dismissal are available for the government's failure to meet section 3730(b)(4)'s deadline. In Pierce County, the less severe remedy was allowing adversely affected complainants to bring an action in district court to enforce CETA's 120-day deadline. See 476 U.S. at 260 n. 7, 106 S. Ct. at 1839 n. 7. Here, to the extent there is any cognizable injury at all resulting from the government's untimely decision to intervene, the less extreme remedies of an order compelling the government to comply with section 3730(b)(4)'s timeliness requirement, or even a ruling that the government's failure to comply with section 3730(b)(4) amounts to an election under section 3730(b)(4)(B) not to take over the action,4 are available. Given the availability of these remedies, Pierce County compels us to hold that the mere use of the word "shall" in section 3730(b)(4) is not enough to render that provision a jurisdictional requirement.

Having found nothing in the statutory text sufficient under Pierce County to manifest an intent that section 3730(b)(4) serve as a jurisdictional requirement, we turn, as Pierce County instructs us to do, to other indicia of congressional intent.5 Contrary to BD's assertions as to the statute's history and purpose, we find nothing there to suggest that Congress intended section 3730(b)(4) to be jurisdictional in character. BD directs us in particular to a passage in the Senate Committee's Report to the effect that extensions of the time in which the government must elect whether to intervene, see 31 U.S.C. Sec. 3730(b)(3), should only be granted upon a showing of good cause, so as to prevent unnecessary government delay in deciding whether to proceed with the action. See S.Rep. No. 345, 99th Cong., 2d Sess. 24-25 (1986), U.S.Code Cong. & Admin.News 1986, 5266, 5289, 5290. Although it is true, as this passage bears out, that Congress was concerned that section 3730(b)(3) not be a vehicle for unnecessary governmental delay, this passage simply does not address the particular question of what remedy should be afforded where the government does delay. Indeed, if any of the concerns that prompted enactment of section 3730(b)(3) are significant to the question we address here, it is not that the government should be prevented from abusing the extension privilege, but rather that, because the intervention decision may at times require substantial, time-consuming investigation, the government should not be bound in all cases to a 60-day period in deciding whether to intervene in qui tam actions. That Congress had this latter concern suggests, if anything, that it may not have intended section 3730(b)(4) as a jurisdictional requirement. Cf. Pierce County, supra, at 261, 106 S.Ct. at 1840 (fact that determination as to alleged misuse of funds claim is "more substantial task than filing a complaint" suggests that Congress did not intend dismissal for failure to timely make determination); see also supra note 3.

BD's efforts to distinguish Pierce County are unpersuasive. BD contends primarily that the rule against construing statutory deadlines as jurisdictional requirements is controlling only where, as in Pierce County, the deadline is applicable to the government acting in an administrative capacity, and that it has no application where, as here, the deadline is imposed upon the government as a party to litigation. Neither principle nor precedent, however, support a distinction of Pierce County on this ground. As a matter of principle, BD has failed to articulate and we fail to see why the broadly stated policy of prohibiting an official's negligence from prejudicing the public interest would apply with any less force where the government acts in a litigating capacity than where it acts in an administrative capacity.

Moreover, the cases relied upon in Pierce County confirm that the Pierce County reasoning and holding are equally applicable to the government as litigant. The cases from which the Court derived the "great principle of public policy" protecting the public interest from official negligence, see 476 U.S. at 260, 106 S. Ct. at 1839, invoked that principle to justify holdings that laches and statutes of limitations--the archetypical deadlines governing litigants--ordinarily will not run against the government ("quod nullum tempus occurrit regi "). See Guaranty Trust Co. of New York v. United States, 304 U.S. 126, 132-33, 58 S. Ct. 785, 788-89, 82 L. Ed. 1224 (1938) (rule that statutes of limitations will not run against the government derives its "continuing vitality" from underlying public policy of preserving public rights from negligence of public officers); Stanley v. Schwalby, 147 U.S. 508, 514-15, 13 S. Ct. 418, 420-21, 37 L. Ed. 259 (1893); Nashville, C. & St. L.Ry. Co., 118 U.S. at 125, 6 S.Ct. at 1008; see also Block v. North Dakota, 461 U.S. 273, 294, 103 S. Ct. 1811, 1823, 75 L. Ed. 2d 840 (1983) (O'Connor, J., dissenting) (reciting history of and rationale for principle "nullum tempus occurrit regi "). Indeed, Marshall v. Local Union 1374, Int'l Ass'n of Machinists and Aerospace Workers, AFL-CIO, 558 F.2d 1354 (9th Cir.1977), also cited by the Court in Pierce County, 476 U.S. at 259 n. 6, 106 S. Ct. at 1839 n. 6, as a case whose underpinnings were found in the "great principle of public policy" therein discussed, id. at 260, 106 S.Ct. at 1839, is almost directly on point with the instant case. In Local Union 1374, the Ninth Circuit held that the Secretary of Labor's failure to comply with the requirement that he "shall" bring suit within 60 days of receiving a complaint, 29 U.S.C. Sec. 482(b), a directive even more akin to a jurisdictional requirement than section 3730(b)(4)'s injunction that the government "shall" elect whether or not to intervene within 60 days, does not bar a later suit. Thus, the very cases from which the Pierce County Court derived the general principle of protecting the public interest from the consequences of official negligence, as well as authority cited by the Court as based on that principle, address themselves to the government in its litigating capacity.

Nor can we accept BD's contention that the entire Pierce County line of precedent is inapposite because its underlying concern for protecting the public interest is inapplicable where, as here, a private litigant would still be positioned to represent the public's interest after the government was dismissed. Even assuming that the relator will be able to proceed with an action where the government has been dismissed, which is not necessarily the case here and often may not be, this argument rests on the flawed assumption that private relators are equally effective in prosecuting false claims on the government as is the government itself. Although the FCA encourages qui tam relators to bring to the government's attention instances of false claims, 31 U.S.C. Secs. 3730(b)(1)-(b)(2), Congress has, not surprisingly, evidenced its contrary judgment that the public is best served by placing with the government the "primary responsibility for prosecuting [false claims] action[s]," 31 U.S.C. Sec. 3730(c)(1) (emphasis added); see generally 31 U.S.C. Sec. 3730(c). We therefore have little doubt that, at least in Congress' estimation, reading section 3730(b)(4) as a jurisdictional bar on the government would prejudice the public interest, and accordingly reject BD's argument that the Pierce County rationale is inapplicable in circumstances such as these.6 Although one can reasonably doubt the wisdom of a rule shielding public officials from the consequences of their negligence, the Supreme Court has long adhered to such a rule, and we believe it is as applicable here as in other contexts.

Concluding, on the authority of Pierce County, that use of the word "shall" is insufficient evidence of a congressional intention that section 3730(b)(4) serve as a jurisdictional requirement upon the government in qui tam actions, and finding nothing in the FCA's history or purpose suggesting that Congress did intend a jurisdictional requirement, we are constrained to hold that the government's failure to comply with section 3730(b)(4) does not bar its further prosecution of a qui tam action.7 Because we hold that section 3730(b)(4) is not a jurisdictional requirement, we reverse that portion of the district court's order dismissing the government as a party plaintiff and dismissing the government's complaint.

We next address Siller's contention that the district court erred in holding that his action was "based upon" publicly disclosed allegations, and thus barred under section 3730(e)(4).

By 1986, when section 3730(e)(4) was enacted, Congress had come to the conclusion that "fraud against the Government was apparently so rampant and difficult to identify that the Government could use all the help it could get from private citizens with knowledge of fraud," United States ex rel. LaValley v. First Nat'l Bank of Boston, 707 F. Supp. 1351, 1355 (D.Mass.1988) (summarizing sentiments motivating 1986 amendments to FCA); see S.Rep. No. 345, 99th Cong.2d Sess. 2-4 (1986), and consequently amended the FCA to, inter alia, "encourage more private enforcement suits." Id. at 23-24, U.S.Code Cong. & Admin.News 1986, at 5288, 5289. At the same time, however, Congress sought to prevent "parasitic" qui tam actions in which relators, rather than bringing to light independently-discovered information of fraud, simply feed off of previous disclosures of government fraud. See, e.g., United States ex rel. Stinson v. Prudential Ins. Co., 944 F.2d 1149, 1154 (3d Cir.1991) (collecting sources). Congress sought to strike this balance largely through the enactment of section 3730(e)(4), which provides in relevant part that--

(A) No court shall have jurisdiction over an action under this section based upon the public disclosure of allegations or transactions in a criminal, civil, or administrative hearing ... unless ... the person bringing the action is an original source of the information.

(B) For purposes of this paragraph, "original source" means an individual who has direct and independent knowledge of the information on which the allegations are based and has voluntarily provided the information to the Government before filing an action under this section which is based on the information.

31 U.S.C. 3730(e)(4).

Under the terms of section 3730(e)(4), Siller's action was properly dismissed only if (1) it was "based upon" the allegations against BD in the SSI suit; (2) the SSI suit's disclosure of these allegations constituted a "public disclosure" in a "civil ... hearing"; and (3) Siller was not an "original source," either because he did not have "direct and independent knowledge of the information on which the allegations were based" or had not "voluntarily provided the information to the Government" before he brought his suit. We address these issues, each of which raises unsettled questions of statutory construction, seriatim.

The district court held that Siller's action was barred under section 3730(e)(4)(A) because it was "based upon" the disclosures in the complaint filed in SSI's action against BD. Siller contends that a qui tam action is only "based upon" a public disclosure where the relator has actually derived from that disclosure the knowledge of the facts underlying his action. Under this reading of section 3730(e)(4)(A), Siller argues, his complaint, even if substantially similar to that in the SSI litigation, was not "based upon" the disclosures in the SSI complaint because he actually learned of BD's overcharging practices independently of the SSI complaint. BD responds that any qui tam complaint, such as Siller's, that "echoes" previously disclosed allegations is "based upon" those disclosures, regardless of whether those disclosures are actually the source from which the qui tam plaintiff derived his information.

We agree that Siller's reading of "based upon" as meaning "derived from" is the only fair construction of the statutory phrase. Section 3730(e)(4)(A)'s use of the phrase "based upon" is, we believe, susceptible of a straightforward textual exegesis. To "base upon" means to "use as a basis for." Webster's Third New International Dictionary 180 (1986) (definition no. 2 of verb "base"). Rather plainly, therefore, a relator's action is "based upon" a public disclosure of allegations only where the relator has actually derived from that disclosure the allegations upon which his qui tam action is based. Such an understanding of the term "based upon," apart from giving effect to the language chosen by Congress, is fully consistent with section 3730(e)(4)'s indisputed objective of preventing "parasitic" actions, see, e.g., Stinson, supra, at 1154, for it is self-evident that a suit that includes allegations that happen to be similar (even identical) to those already publicly disclosed, but were not actually derived from those public disclosures, simply is not, in any sense, parasitic.

We are aware, as BD points out, that other circuits have not embraced this interpretation of the phrase, assuming instead that an action is based upon a public disclosure of allegations if its allegations are identical or similar to those already publicly disclosed. However, only one court of appeals has actually held that a relator need not have derived his knowledge from a public disclosure in order for his action to have been based upon that disclosure, and that court summoned neither reasoning nor supportive case authority in defense of its holding. In John Doe Corp., the Second Circuit held that a qui tam plaintiff's action is "based upon" a public disclosure where the relator's allegations are "the same as those that ha[ve] been publicly disclosed.... regardless of where the relator obtained his information." 960 F.2d at 324. The court in John Doe, however, without a single word of analysis, merely cited two cases in support of its holding, United States ex rel. Dick v. Long Island Lighting Co., 912 F.2d 13, 18 (2d Cir.1990), and LaValley, 707 F. Supp. at 1367, and neither provides authority for the John Doe holding.8

In Long Island Lighting, the court held that "if the information on which a qui tam suit is based is in the public domain, and the qui tam plaintiff was not a source of that information, then the suit is barred." 912 F.2d at 18. Taken out of context, as it was in John Doe, this quotation might appear to represent a holding that a qui tam action is "based upon" a public disclosure whenever the allegations are identical or similar to allegations already publicly disclosed. However, the Long Island Lighting court itself never addressed the particular question of when a qui tam action is "based upon" a public disclosure. There was in that case "no dispute that [the qui tam plaintiffs'] suit [was] based upon publicly disclosed 'allegations or transactions' "; the only issue before the court was whether the relators were "original sources" under section 3730(e)(4). 912 F.2d at 16.

LaValley, the other case upon which the John Doe court relied, also does not support John Doe 's holding. There, the district court interpreted "based upon" virtually identically to the way we interpret the phrase, holding that, although the factual basis underlying the qui tam action in that case was disclosed in a bankruptcy trustee's complaint, the action was not barred because "the information and knowledge upon which [the qui tam ] action [was] based did not originate from the [b]ankruptcy [t]rustee's [c]omplaint." 707 F. Supp. at 1367 (emphasis added). In a statement relied upon by the Second Circuit, the district court went on to note that the information related by the key witness in the case, on which the action was based, could not bar the qui tam plaintiffs' suit, because that information was never "contained in a public disclosure." Id. The court in John Doe simply misunderstood this latter observation as meaning that the action was not based upon the witness' information because that information was never publicly disclosed, when in fact it meant only precisely what it said--that the information related by the key witness, on which the action was based, could not serve as a bar to the action because it was never publicly disclosed.

We recognize that other courts of appeals seem implicitly to have accepted the position, contrary to ours, that qui tam actions are "based upon" a public disclosure whenever the factual basis for the action has been disclosed into the public domain, regardless of whether the relator actually derived his knowledge of those facts from that disclosure. See United States ex rel. Springfield Terminal Ry. Co. et al. v. Quinn, 14 F.3d 645, 652-55 (D.C.Cir.1994) (Congress sought to prohibit qui tam actions when allegation of fraud or critical elements of fraudulent transaction "were in public domain."); United States ex rel. Kreindler & Kreindler v. United Technologies Corp., 985 F.2d 1148, 1158 (2d Cir.) (quoting John Doe with approval), cert. denied, --- U.S. ----, 113 S. Ct. 2962, 125 L. Ed. 2d 663 (1993); United States ex rel. Precision Co. v. Koch Industries Inc., 971 F.2d 548, 552 (10th Cir.1992) (" 'based upon' is properly understood to mean 'supported by' "), cert. denied, --- U.S. ----, 113 S. Ct. 1364, 122 L. Ed. 2d 742 (1993). None of these cases, however, actually addresses the specific question whether a qui tam relator need have derived his knowledge of the factual basis of his action from a public disclosure in order for his action to be "based upon" that disclosure. Rather, each of these cases addresses the question whether a qui tam action that is only based in part on publicly disclosed information should be deemed to be "based upon" the public disclosure under section 3730(e)(4)(A). The extent to which a qui tam action must be "based upon" a public disclosure is, of course, distinct from the analytically antecedent question of what it means for an action to be "based upon" a public disclosure. Moreover, none of these cases provides any reasoning by which to reconcile its implicit understanding of the phrase "based upon" with the phrase's plain meaning. The Precision decision, for example, baldly asserts that "[a]s a matter of common usage, the phrase 'based upon' is properly understood to mean 'supported by.' " 971 F.2d at 552. We are unfamiliar with any usage, let alone a common one or a dictionary definition, that suggests that "based upon" can mean "supported by." Preferring the plain meaning of the words enacted by Congress over our sister Circuits' as-yet unconsidered assumptions as to the meaning of those words, and over the Second Circuit's considered but unsupported interpretation, we hold that Siller's action was only "based upon" the disclosures in the SSI lawsuit if Siller actually derived his allegations against BD from the SSI complaint.

It is not clear to us that the district court found that Siller actually derived the allegations in his qui tam action from the SSI complaint. The district court found that the SSI suit publicly disclosed allegations that "derive[d] from exactly the same core of operative facts ... that form[ ] the basis for" Siller's qui tam complaint. J.A. at 14. That the facts that were disclosed in the SSI complaint and the facts that "form[ed] the basis for" Siller's action were the same does not, however, mean that Siller actually derived his allegations from the allegations in the SSI complaint. It is certainly possible that, as Siller contends, Siller actually learned of BD's alleged fraud entirely independently of the SSI suit, and derived his allegations from that independent knowledge. Therefore, because the district court made no finding on whether Siller actually derived his allegations from the SSI suit, a finding necessary to the conclusion that Siller's action was "based upon" that suit, we also vacate the portion of the district court's order dismissing Siller's action so as to enable the court to address this factual question on remand.

In doing so, however, we do not upset the district court's express finding that the same essential facts underlay both the SSI suit and Siller's qui tam action. Id. We are not persuaded by Siller's argument that the SSI complaint failed to allege a fact (i.e., BD's failure to disclose its overcharges) critical to his present action. Contrary to Siller's assertion, it appears to us that the allegation in SSI's complaint that BD charged the government more than its other customers, in violation of its contractual (and statutory, see 41 U.S.C. Sec. 253e(b)) obligation, see J.A. at 95,9 is essentially the same allegation that Siller has made against BD.

The district court also held that the disclosure of allegations in SSI's complaint that BD charged the Government more than its other customers constituted a "public disclosure of allegations" in a civil "hearing." The district court's interpretation is supported by the holding of every court of appeals to have addressed the question, that any information disclosed through civil litigation and on file with the clerk's office should be considered a public disclosure of allegations in a civil hearing for purposes of section 3730(e)(4)(A). See Quinn, 14 F.3d at 651; Kreindler, 985 F.2d at 1158; Precision, 971 F.2d at 554 n. 5; Stinson, 944 F.2d at 1154-56. Siller argues that, even if his action is deemed to be "based upon" the allegations in the SSI complaint, the action is not barred because the term "hearing" does not encompass the mere filing of a complaint. While we believe this presents a closer question of statutory construction, we affirm the district court's interpretation of this part of section 3730(e)(4)(A).10

A civil complaint is unquestionably a "public disclosure of allegations." The difficult issue is whether a disclosure in a complaint constitutes a disclosure in a "civil hearing." Because a disclosure in a complaint is a disclosure in a civil proceeding, the ultimate resolution of this issue turns on whether the term "hearing" can be understood as including a "civil proceeding." We concede, as Siller argues, that a "hearing" is ordinarily understood as an event that occurs within a civil proceeding, rather than as an entire civil proceeding itself. We do not believe, however, that the definition of the term "hearing" is so fixed that it cannot possibly refer to (or at least include) an entire civil proceeding. See United States v. Florida East Coast Ry. Co., 410 U.S. 224, 239, 93 S. Ct. 810, 818, 35 L. Ed. 2d 223 (1973) (term " 'hearing' in its legal context undoubtedly has a host of meanings" that will vary depending on the context); see also Black's Law Dictionary 721 (hearing "frequently used ... to describe whatever takes place before magistrates clothed with judicial functions and sitting without jury at any stage of the proceedings subsequent to its inception" ) (emphasis added) (1990). We are also unpersuaded by Siller's contention that, as a practical matter, Congress might have meant to include live hearings but exclude civil complaints and other court filings because disclosures in a public hearing between private parties are more likely to come to the government's attention than are disclosures in filings, and we can imagine no other reason why Congress would have intended a meaning of "hearing" that excluded civil proceedings. Given the fluidity in the meaning of the term "hearing," and the fact that we can discern no reason why Congress might have intended otherwise, we agree with our sister Circuits (albeit somewhat reluctantly) that an entire civil proceeding can constitute a "hearing" for purposes of section 3730(e)(4)(A).11 See Quinn, supra, at 651 (" 'hearing' is roughly synonymous with 'proceeding.' "); cf. Anchorage Assoc. v. Virgin Islands. Bd. of Tax Review, 922 F.2d 168, 176 (3d Cir.1990) ("hearing" under Rule 56(d) does not require oral hearing or live, formal judgment). We therefore do not disturb the district court's holding that the allegations in the SSI complaint were, for purposes of section 3730(e)(4)(A), publicly disclosed in a civil "hearing."

In addition to its holding that Siller's action was based upon previously disclosed public allegations in a civil hearing within the meaning of section 3730(e)(4)(A), the district court also held that Siller was not an "original source" within section 3730(e)(4)'s exception to its jurisdictional bar. In so holding, the court expressly adopted the standard fashioned by the Second Circuit in Long Island Lighting Co., 912 F.2d at 16; accord Wang v. FMC Corp., 975 F.2d 1412, 1418 (9th Cir.1992). That standard requires that, in order to be considered an "original source," a qui tam plaintiff must not only have direct and independent knowledge of the information on which his allegations are based, and have provided that information to the government, but also must have "been a source to the entity that publicly disclosed the allegations on which a suit is based." Long Island Lighting, 912 F.2d at 16 (emphasis added). Finding that Siller was not a source to SSI, the "entity that publicly disclosed the allegations" on which Siller's suit was found to be based, id., the district court held that Siller was not an "original source" under section 3730(e)(4).

We reject the Second Circuit's standard, and the district court's adoption of that standard, as imposing an additional, extra-textual requirement that was not intended by Congress. Reading sub-paragraph (B), as we are directed by the statute, as the definition of "original source" applicable to sub-paragraph (A), sub-paragraph (A) is properly read as providing that,

[n]o court shall have jurisdiction over an action ... based upon the public disclosure of allegations ... in a ... civil ... hearing ... unless ... the person bringing the action ... has direct and independent knowledge of the information on which the allegations are based and has voluntarily provided the information to the Government before filing an action under this section which is based on the information.

31 U.S.C. Sec. 3730(e)(4) (emphasis added). Under this reading of section 3730(e)(4), the provision unambiguously does not require, as the Second Circuit has held that it does, that a relator be a source to the original disclosing entity in order to be an "original source." Rather it requires only that the relator have direct and independent knowledge of the information underlying the allegations of a false claim and voluntarily provide the information to the government before filing his qui tam action.

The Second Circuit's attempt to portray its holding that a relator must also be a source to the disclosing entity as one derived from the provision's text is not merely unpersuasive, but implausible. The Second Circuit essentially holds that sub-paragraphs (A) and (B) each provides distinct and cumulative requirements for being an "original source." Thus, it reasons, sub-paragraph (A) requires that the relator be an original source to the entity that made the public disclosure, and sub-paragraph (B) additionally requires that the relator have direct and independent knowledge of the allegations underlying his qui tam action and that he provide his information to the government. Long Island Lighting, 912 F.2d at 16-17; see also id. at 17 ("p (4)(B) does not contain the exclusive requirements in order for one to be an 'original source' and that an additional requirement is to be found in p (4)(A).").

The Second Circuit's interpretation of section 3730(e)(4) to require that the putative plaintiff provide his information to the disclosing entity might at least be tenable were there not a definition of "original source" in sub-paragraph (B). But the term "original source" as it is used in sub-paragraph (A) is expressly defined in sub-paragraph (B). Sub-paragraph (B) necessarily, therefore, sets forth that which the Second Circuit holds it does not--"the exclusive requirements that a qui tam plaintiff must satisfy to be an 'original source.' " Id. at 17.

The Second Circuit's "close textual analysis," id. at 16, that yields this interpretation is wholly indefensible. The essence of the Second Circuit's textual reasoning is that the word "information" in sub-paragraph (A) refers to the information publicly disclosed, whereas the "information" in sub-paragraph (B) refers to the information that supplies the basis for the qui tam action itself. This "slight difference in meaning" between these two uses of the word "information," suggests the Second Circuit, "permits the interpretation that p (4)(B) does not contain the exclusive requirements in order for one to be an 'original source' and that an additional requirement is to be found in p (4)(A)." Id. Besides the fact that a "slight difference" between the meaning of "information" in the two paragraphs cannot justify a holding that sub-paragraph (B)'s definition of "original source" is not the entire definition of that term, the reasoning that underlies the Second Circuit's conclusion that the word "information" means something different in the two provisions is demonstrably incorrect.

First, it states, "if the word 'information' means the same in p (4)(A) as it does in p (4)(B), then its use in p (4)(A) would be superfluous." Id. Because the definition of "original source" in sub-paragraph (B) references the information on which the allegations are based, inclusion of the prepositional phrase "of the information" in sub-paragraph (A) is, in the most hypertechnical sense, unnecessary. However, the phrase is not superfluous in a way or to a degree that warrants ascribing different meanings to the same word used in two successive clauses. In other words, it is so unlikely that Congress would have even noticed the technical redundancy, that no significance can reasonably be inferred from inclusion of the prepositional phrase in sub-paragraph (A).

Second, the Long Island Lighting court suggests that, because the word "information" in sub-paragraph (B) is accompanied by a modifying phrase while the word "information" is without modification in sub-paragraph (A), "information" must mean something different in sub-paragraph (B) than it does in sub-paragraph (A). Id. at 16-17. Although it is true that "information" is modified in sub-paragraph (B) and is not in sub-paragraph (A), the modifying phrase used in sub-paragraph (B)--"on which the allegations are based"--itself confirms that the word "information" in sub-paragraph (B) has the identical meaning that the Second Circuit agrees the word has in sub-paragraph (A). "The allegations" referred to in the modifying phrase can only mean those allegations publicly disclosed, since those not only are the allegations referenced in the clause preceding the modifying phrase, but are the only allegations mentioned at all in section 3730(e)(4). Thus, "the information on which the allegations are based" in sub-paragraph (B) cannot possibly refer to anything but the information on which the publicly disclosed allegations are based, which is, as the Second Circuit agrees, id. at 17, the precise information referenced in sub-paragraph (A).

Finally, the Second Circuit posits, without explanation, that "the most plausible reading of p (4)(B) indicates that the 'information' referred to therein is that which supplies the basis for the qui tam action itself." Id. As set forth above, however, the fact that sub-paragraph (B) refers to "the information on which the allegations are based " confirms that the only possible reference of the word "information" in sub-paragraph (B) is to the information publicly disclosed--the exact same reference of the word in sub-paragraph (A).

We believe that, in truth, the Second Circuit's conclusion that a putative plaintiff must provide his information to the disclosing entity in order to be an original source rests not upon the statutory language, but entirely upon a reading, and misreading, of the legislative history. In fact, the Second Circuit's decision is a classic example of the use of legislative history to create an ambiguity in the statute where none exists in order to justify use of that history as dispositive evidence of congressional intent.

The Second Circuit fixes on Senator Grassley's comment that "the jurisdictional requirements of Sec. 3730(e)(4) bar[ ] a person 'who had not been an original source to the entity that disclosed the allegations ' from bringing a qui tam claim based on publicly disclosed information." Id. (quoting 132 Cong.Rec. 20,536 (1986)) (emphasis added). If the provision to which this comment was directed had been enacted as it existed at the time Senator Grassley made the comment, the comment would be some evidence of a congressional intent supporting the Second Circuit's interpretation, although even then we would not permit such a statement to override statutory language as clear as that in sub-paragraph (B). In fact, however, the version of the legislation addressed by Senator Grassley12 was changed in two significant respects.

First, Congress deleted "the media" as a party whom the original source was required to inform. Given that the media is specified in the enacted version of sub-paragraph (A) as one of the disclosing entities, Congress presumably would not have deleted the media from the "original source" definition in sub-paragraph (B) if it intended to require the plaintiff to provide his information to the disclosing entity. Second, Congress ultimately provided that an original source had to inform the government only "before filing [his qui tam ] action," 31 U.S.C. Sec. 3730(e)(4)(B), as opposed to "prior to an action filed by the Government." 132 Cong.Rec. S11,240 (daily ed. Aug. 11, 1986). It would be odd, if Congress had intended a plaintiff to be a source to the disclosing entity in order to be an "original source," for it to have allowed the plaintiff to provide his information after the public disclosure (i.e., by only requiring the provision of information before the plaintiff brings his action).13 These two changes suggest that, even assuming that Congress may at one point have intended a plaintiff to be a source to the disclosing entity to be an original source, which is anything but clear, it ultimately chose not to enact such a requirement into law.14 That the Second Circuit's textual analysis is less than persuasive was recognized by the Ninth Circuit in Wang, 975 F.2d at 1418, the only Court of Appeals to have adopted the Second Circuit's "original source" requirement. That court, instead of relying upon the Second Circuit's statutory analysis, rested its decision to impose the third original source requirement wholly on the statute's legislative history, albeit a different portion of the history from that discussed in Long Island Lighting. See Wang, 975 F.2d at 1418-19. This portion of the history of the FCA, we believe, supports, but at the very least is consistent with, the unambiguous statutory language that a plaintiff must provide his information only to the government, not to a third-party disclosing entity.

The Ninth Circuit premised its holding almost entirely on the fact that one of the principal objectives of the 1986 amendments was to ensure that, contrary to the holding of decisions such as United States ex rel. State of Wisconsin v. Dean, 729 F.2d 1100, 1104 (7th Cir.1984), a plaintiff who had himself provided the government with the information of fraud prior to bringing his action would not be barred because the government (as a result of the plaintiff's effort) possessed the information on which the action was based. 975 F.2d at 1419 (citing S.Rep. 345, 99th Cong., 2d Sess. 4 (1986)). That Congress sought to "correct" Dean provides no support whatsoever for the conclusion, see 975 F.2d at 1419, that Congress intended to require a putative plaintiff to have provided his information to the "disclosing entity" in order to be an "original source."

The jurisdictional provision in effect at the time Dean was decided did not even purport to bar qui tam suits based upon prior public disclosures; it barred only suits "based upon evidence or information in the possession of the United States ... at the time such suit was brought." 31 U.S.C. Sec. 232(C) (1982) (superseded) (emphasis added). The Seventh Circuit in Dean simply held that section 232(C) barred the qui tam suit in that case because the information was "in the possession of" the government, having been provided to it by the plaintiff. To "correct" Dean only required that Congress adopt language that would ensure that a plaintiff who had provided his information to the government would not be barred from bringing a qui tam action on the ground that the government already possessed the information. This it did in section 3730(e)(4)(B), by providing that a plaintiff who produces his independently-obtained information to the government is excepted from section 3730(e)(4)(A)'s jurisdictional bar. The Ninth Circuit's error is in its unexplained assumption that, to remedy Dean, Congress had to require a plaintiff to provide information to the disclosing entity, as opposed merely to the government. 975 F.2d at 1419.15

We are not unmindful that, in contrast to the Second and Ninth Circuits, which have largely relied on the legislative history to the 1986 amendments, some courts, out of utter frustration with the many seeming inconsistencies in the history, have sought to interpret section 3730(e)(4) instead based principally upon the overarching objectives of the legislation. For example, the Third Circuit, after noting that one can find in the legislative history of the FCA amendments "some support somewhere for almost any construction" of the statutory terms, stated that it would "look instead to what [it] [could] glean from the principal intent" of the legislation, which was "to have the qui tam suit provision operate somewhere between the almost unrestrained permissiveness" represented by those pre-1943 cases that had allowed plaintiffs to bring qui tam actions based upon allegations they had learned about by reading publicly-available criminal indictments, and those post-1943 cases that had barred even suits by plaintiffs who had initially disclosed the fraud to the government. See, e.g., Stinson, 944 F.2d at 1154. It strikes us as especially inappropriate (not to mention frighteningly treacherous) to attempt, as these courts have done, to distill from such broad, generalized objectives, the answers to the kind of specific statutory questions that we herein address; fine calibrations are just not possible through the use of such crude instruments. This is particularly so in this context, given that, although we can perhaps divine from these abstract purposes a congressional intention to balance the need to encourage qui tam actions against the need to prevent parasitic suits, we can discern virtually nothing as to precisely how Congress ultimately believed it achieved that balance. If the language of law is to have any meaning at all, then surely it must prevail over the kind of speculation that is entailed in such an enterprise as these courts have undertaken.

Accordingly, we hold that a qui tam plaintiff need not be a source to the entity that publicly disclosed the allegations on which the qui tam action is based in order to be an original source under section 3730(e)(4)(B).16 In addition to having direct and independent knowledge of the information on which the allegations in the public disclosure is based, he need only provide his information to the government before instituting his qui tam action, as the provision unambiguously states.

For the reasons discussed, we reverse that portion of the district court's order dismissing the government as a party-plaintiff and dismissing the government's complaint, and vacate that portion of its order dismissing Siller's complaint as "based upon" the allegations in the lawsuit by SSI against BD. The case is remanded for further proceedings consistent with this opinion.

REVERSED IN PART, VACATED IN PART, AND REMANDED.