This is an interlocutory appeal. The United States and the State of Indiana (the "governments") filed separate actions under the Comprehensive Environmental Response, Compensation and Liability Act ("CERCLA"), 42 U.S.C. §§ 9601-9675, to recover costs incurred by them in connection with environmental clean-up and remediation of a site in Fort Wayne, Indiana. In addition to its federal CERCLA claim, Indiana also brought a state law action seeking similar relief. See Ind.Code § 13-25-4-8. These actions against Navistar International Transportation Corporation and the twenty-five other defendants (hereinafter referred to collectively as "Navistar") were consolidated by the district court. Navistar sought summary judgment on the ground that the governments' claims were time barred by the six-year CERCLA statute of limitations, 42 U.S.C. § 9613(g)(2)(B). The district court denied the motion for summary judgment, and Navistar sought and received certification to appeal the ruling immediately under 28 U.S.C. § 1292(b). This court accepted the certification and permitted the interlocutory appeal. For the reasons set forth in the following opinion, we reverse the district court's judgment denying Navistar's motion for summary judgment.

* BACKGROUND

The environmental site at issue in this litigation is located on the Maumee River in Fort Wayne, Indiana. The site was used as a municipal landfill from approximately 1967 to 1976 for the disposal of residential and industrial wastes, including hazardous wastes. The United States Environmental Protection Agency ("EPA") placed the site on the National Priorities List in 1986, indicating that it was considered a significant threat to the public health and the environment. See generally 42 U.S.C. § 9605(a)(8); 40 C.F.R. § 300.425. This action resulted in the initiation by the EPA of a formal administrative process regarding the site, including an evaluative process to determine what action was necessary to neutralize the threat posed by the site. See generally 40 C.F.R. §§ 300.400-440 (outlining methods and procedures for evaluating, prioritizing and remedying environmental harms). At the conclusion of this process, the EPA determined that, among other things, the landfill needed to be covered with a permanent clay cap to isolate the hazardous materials from the rest of the environment.

In February 1989, the governments filed an action against SCA Services of Indiana, Inc. ("SCA"), the owner and operator of the site at that time. Simultaneously, the parties filed a consent decree which provided that SCA would perform and pay for the necessary remedial action at the site under the governments' supervision and in accordance with the various documents produced during the formal EPA administrative process. SCA also agreed to pay a portion of the governments' costs to oversee the remedial action up to a maximum of $200,000. The consent decree was approved by the district court in July 1989.

The consent decree, in combination with several administrative documents, established a timeline for the preparation and creation of the remedial design plan for the site and for the implementation of that plan. Moreover, the decree indicates that the implementation of the remedial action was to begin after the final remedial design plans were approved by the EPA and the State. The EPA asserts that this approval had to be in writing.1 In July 1990, SCA obtained permission to begin certain activities in connection with the remedial action. These activities included hooking up utilities at the site for use in the remedial action, constructing an access road and clearing and grubbing part of the site to prepare for placement of the clay cap. The governments note that the final design of the remedial action had not been submitted at this time, and so approval to begin the implementation phase of the action had not been obtained.

On September 11, 1990, the final design documents were submitted to the EPA. On September 14, 1990, the EPA's project manager gave SCA oral approval of the final design, and on September 17, 1990, oral authorization was given to SCA to proceed with construction of the remedial action. On September 18, 1990, the first "lift" of clay to build the permanent clay cap was placed on the landfill.2 On September 20, 1990, SCA obtained written approval of the design plan and written authorization to begin construction. The construction of the remedy was completed by SCA in 1995, but the governments continued to incur oversight costs because some response activities at the site, such as groundwater monitoring, are ongoing.

Navistar and the other defendants in this action became involved in the EPA-SCA litigation in 19923 when SCA filed a third-party complaint naming them, as well as more than fifty other parties, as third-party defendants and seeking contribution for the costs of the remedial action. However, the governments did not file any claims directly against the defendants at that time; in fact, the United States attempted unsuccessfully to sever the third-party litigation from the settled action between itself and SCA. The litigation between SCA and the third-party defendants ultimately was settled through various agreements to which the governments were not a party. Because some of the agreements could have been construed to preclude any cost-recovery claims by the United States against the parties to the settlements, the governments obtained an order from the district court reserving their rights to bring separate claims against the third-party defendants to recover their oversight costs. This action was filed by the United States against Navistar on September 19, 1996, to recover its costs for the remedial action that were not covered by the SCA consent decree. The State of Indiana filed a similar action on September 20, 1996.

The complaints filed by the governments sought reimbursements for costs incurred from November 1, 1989, through April 30, 1995, and a declaration that the defendants are jointly and severally liable for future costs. The defendants moved for summary judgment on the ground that the action was brought beyond the six-year statute of limitations and thus was time barred. The statute of limitations at issue provides:

(2) Actions for recovery of costs

An initial action for recovery of the costs referred to in section 9607 of this title must be commenced--

...

(B) for a remedial action, within 6 years after initiation of physical on-site construction of the remedial action....In any such action described in this subsection, the court shall enter a declaratory judgment on liability for response costs or damages that will be binding on any subsequent action or actions to recover further response costs or damages. A subsequent action or actions under section 9607 of this title for further response costs at the vessel or facility may be maintained at any time during the response action, but must be commenced no later than 3 years after the date of completion of all response action....

42 U.S.C. § 9613(g)(2). The pertinent language of the statute is that "[a]n initial action for recovery of the costs ... must be commenced ... for a remedial action, within 6 years after initiation of physical on-site construction of the remedial action." Id. Navistar's position was that initiation of the on-site construction began no later than September 18, 1990, when the first lift of clay was placed on the landfill. Therefore, because the governments' actions were filed on September 19 and 20, 1996, they were filed after the six-year period had expired.

The district court concluded that the six-year statute of limitations was inapplicable in this case because the suit against the defendants was not an "initial action," but was instead a "subsequent action" governed by a different limitations period. The CERCLA statute provides that a "subsequent action ... for further response costs ... may be maintained at any time during the response action, but must be commenced no later than three years after the date of completion of all response action." Id. The district court determined that the case by the governments against SCA was the initial action and that this case against Navistar was a subsequent action to obtain further response costs. In the district court's view, because this action was commenced while the "response action" was ongoing, it is not time barred. Alternatively, the district court held that, even if this case is an initial action governed by the six-year statute of limitations, the action was timely filed because the activities cited by Navistar did not constitute the "initiation of physical on-site construction of the remedial action" within the meaning of the statute.4

Because the district court decided that the governments' federal claims were not time barred, it held that the secondary issue of the appropriate statute of limitations to apply to Indiana's state law-based claim was moot; Indiana had argued that a state law ten-year limitations period applied, and Navistar had argued that the federal six-year period applied. Because the court found that the actions were timely with respect to the six-year period, that dispute was rendered moot. As set forth below, we determine that the governments' federal claims are time barred, and consequently we also address whether Indiana's state law action is similarly time barred and hold that it is.

This case presents us with two questions involving the construction of the CERCLA statutes of limitations contained in 42 U.S.C. § 9613(g)(2). The first is whether the six-year statute of limitations for "initial actions" or the open-ended statute of limitations for "subsequent action[s] ... for further response costs" applies to the actions filed by the governments against Navistar in September 1996. If we were to determine that the longer period for subsequent actions applies, this appeal would be at an end; however, because we conclude that the six-year period is the correct one, we must also decide whether the governments' actions were filed more than six years following the "initiation of physical on-site construction of the remedial action." We decide that they were. Finally, we shall address the appropriate statute of limitations for Indiana's state law-based claim.

We review the district court's decision to deny Navistar's motion for summary judgment de novo, viewing the facts of record in the light most favorable to the governments. See Harden v. Raffensperger, Hughes & Co., 65 F.3d 1392, 1396 (7th Cir.1995). A party is entitled to summary judgment when the "pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law." Fed.R.Civ.P. 56(c). In this case, the facts are not in dispute; rather, the parties disagree with respect to the interpretation of the CERCLA statute of limitations and with respect to the application of that law to the facts. The district court's interpretation of the CERCLA statute is a matter of law that we shall review de novo. See United States v. Montoya, 827 F.2d 143, 146 (7th Cir.1987).

The governments urge that, in conducting that review, we construe the statute of limitations in their favor in order to avoid frustrating the salutary goals of CERCLA.5 We acknowledge this general principle, but also note that Congress had a specific purpose in mind when it added the statute of limitations to CERCLA in 1986. By implementing the statute of limitations, it expressed a determination that, in order to achieve timely clean-up of affected sites and to ensure replenishment of the fund,6 cost recovery actions must commence in a timely fashion.7 Therefore, although we shall construe ambiguities in the statute in favor of the government in an effort to avoid frustrating the beneficial purposes of CERCLA, we must recognize that Congress has determined that those beneficial purposes are serviced by the timely filing of recovery actions. Cf. Galloway v. General Motors Serv. Parts Operations, 78 F.3d 1164, 1165 (7th Cir.1996) ("[T]he Supreme Court has told us not to interpret statutes of limitations in a grudging, hostile fashion.... They serve the important purpose of encouraging the prompt filing of claims and by doing so of enhancing the likelihood of accurate determinations and removing debilitating uncertainty about legal liabilities." (internal citations omitted)).

We turn first to the question of the appropriate CERCLA statute of limitations period to apply to the actions filed by the governments against Navistar. The parties appropriately focus their initial efforts on the wording of the statute. Section 9613(g)(2) provides the limitations periods for "initial action[s]" for the recovery of costs under 42 U.S.C. § 9607; moreover, the statute specifies different limitations periods for such cost recovery actions if they relate to "removal action[s]," see § 9613(g)(2)(A), as opposed to "remedial action[s]," see § 9613(g)(2)(B). There is no dispute here that a "remedial action" is involved. Therefore, the pertinent statutory language provides that an "initial action" to recover costs must be commenced "within 6 years after initiation of physical on-site construction of the remedial action." Navistar maintains that the governments' actions against it are such "initial action[s]" subject to the six-year period.

The governments contend, however, that their actions against Navistar are "subsequent action[s]." Section 9613(g)(2) states that a "subsequent action or actions under section 9607 ... for further response costs at the vessel or facility may be maintained at any time during the response action, but must be commenced no later than 3 years after the date of completion of all response action." There is also no dispute that the "response action" at the site in this case is ongoing; consequently, if this period of limitations applies, the governments' actions were timely filed. In essence, in order to determine which period of limitations applies, we must determine whether the actions against Navistar are "subsequent" or "initial" ones.

The governments claim that the initial action for recovery of costs was the action they brought against SCA to establish primary liability for the site. That action was resolved by the consent decree, which provided for the recovery of various "response"8 costs including part of the governments' costs (up to $200,000) of overseeing the remediation process at the site. However, the consent decree did not cover all of the governments' costs; consequently, the governments brought these actions against Navistar to recover additional costs that they have incurred and to obtain a declaration that future costs also will be paid by Navistar. According to the governments, therefore, the actions against Navistar are "subsequent" ones to recover "further response costs at the ... facility" within the plain meaning of the statute.

Navistar counters that the plain language and structure of the statute actually establish certain prerequisites for a "subsequent action." Because those prerequisites are not met here, it maintains, the governments' actions only can be deemed initial ones. Navistar contends that the sentence in § 9613(g)(2) prior to the one containing the limitations period for subsequent actions establishes one such prerequisite: "In any such action described in this subsection, the court shall enter a declaratory judgment on liability for response costs or damages that will be binding on any subsequent action or actions to recover further response costs or damages." In Navistar's view, this sentence obligates a district court to enter a declaratory judgment on liability in any and all initial actions that will then govern subsequent actions. Navistar notes that no such declaratory judgment has been entered with respect to this site in general, much less with respect to itself or any of the other defendants specifically; consequently, Navistar maintains that, under the plain language of the statute, the actions against it cannot be deemed "subsequent action[s]" if there has not been first a declaratory judgment entered on an initial action.

In support of this position, Navistar cites, beyond the plain wording and structure of the statute, various authorities. The legislative history, in its view, bolsters the idea that a subsequent action can only follow an initial action concluded by a declaratory judgment.9 In addition, Navistar cites certain proposed, but ultimately not adopted, rules promulgated by the EPA to highlight that the government itself has understood the statute to require a declaratory judgment.10 Finally, Navistar asserts that the Fourth Circuit understands the CERCLA statute to require entry of a declaratory judgment in an initial action before any subsequent action may be maintained.11

The governments respond that nothing in the statute, its history, or the case law requires that all initial actions be concluded by the district court's entering a declaratory judgment with respect to liability; i.e., the statute does not make the entry of a declaratory judgment a prerequisite for a subsequent action. The governments note that, although the statute does provide for entry of a declaratory judgment for efficiency purposes, none of the authorities relied on by Navistar establishes that such a judgment is a quid pro quo or a prerequisite for the maintenance of a subsequent action.12

We agree with the governments and the district court on this issue. The structure of this subsection indicates that the intent of Congress in including the sentence was to avoid the necessity of relitigating liability questions. It provides for the declaratory judgment as a way to manage the litigation in an efficient manner. That also is what we understand the legislative history relied upon by Navistar to indicate. See H.R.Rep. No. 99-253, pt. 3, at 21 (1985), reprinted in 1986 U.S.C.C.A.N. 3038, 3044 ("In the initial cost recovery action, in order to conserve judicial time and resources, the court is to enter a declaratory judgment on liability for response costs; ..."). However, there is nothing to indicate that a declaratory judgment is the only permissible means to resolve an "action for recovery of the costs referred to in section 9607." 42 U.S.C. § 9613(g)(2). Nor is there anything to indicate that the entry of such a declaratory judgment is a prerequisite to the prosecution of a "subsequent action ... for further response costs." Id.13

Although we are unpersuaded by Navistar's first argument to establish a prerequisite for subsequent actions, we do accept its second position. Navistar also argues that the governments' actions against it cannot be "subsequent actions" because the governments have never previously asserted any claim against it. Therefore, submits Navistar, a common sense understanding of "initial" versus "subsequent" actions indicates that the governments' actions against Navistar are initial ones.

The governments respond that it is not necessary for them to have brought claims against these particular defendants in order for this action to be considered a "subsequent" one. The governments reiterate that the initial action here was against SCA; moreover, they note that Navistar and the other defendants were parties to that litigation as third-party defendants. The governments also focus on the fact that the defendants were aware of the claims against them because, in the course of the SCA litigation, the defendants blocked, for discovery purposes, the governments' attempt to sever the action against SCA from the third-party actions. Therefore, the governments urge that, although they did not assert claims directly against Navistar, it was party to the initial action and was intimately aware of the governments' future claims against it.

We cannot accept the governments' position. We believe that, from the language and structure of the statute, a "subsequent action ... for further response costs" must be one brought against the same party or parties against which an "initial action" to recover such costs has been maintained. The first time an action is brought against a given party, it cannot be anything other than an "initial action" for recovery costs. If this were not the case, as Navistar notes, the government would only need to bring an "initial action" against a single party within the six-year limitations period. Thereafter, any action brought by the government against any party in connection with that site could be deemed a "subsequent action" which, under the terms of the statute, may be brought "any time during the response action, but ... no later than 3 years after the date of completion of all response action." 42 U.S.C. § 9613(g)(2). As Navistar points out, this rule would largely write the limitations restriction for initial actions out of the statute because many response actions will be ongoing for decades. In this case, for example, the remedial action at the site is anticipated to last another thirty years. Therefore, if the governments were permitted to file "subsequent actions" against new parties as long as an "initial action" was filed within the six-year period against one party (here, SCA), the governments would have had approximately 33 more years to bring its claims against Navistar. We do not think that this understanding of the statute comports with Congress' intent in adding the statute of limitations to CERCLA--"to provide some measure of finality to affected responsible parties." H.R.Rep. No. 99-253, pt. 1, at 138 (1985), reprinted in 1986 U.S.C.C.A.N. 2835, 2920. Therefore, we hold that a "subsequent action" against a party can be brought only after an "initial action" has been brought against it. Here, the governments' actions against Navistar were the first claims brought against these defendants by the governments to recover response costs. Therefore, the appropriate limitations period in this case is the six-year period for initial actions.14C.

Having established that the six-year limitations period for initial actions, see 42 U.S.C. § 9613(g)(2)(B), applies in this case, we turn next to the question of whether the actions were timely filed. As noted above, the statute provides that an "initial action for recovery of costs ... must be commenced ... for a remedial action, within 6 years after initiation of physical on-site construction of the remedial action." Id. Therefore, the focus for this inquiry is the latter portion of that language, i.e., whether the governments filed suit within six years after the "initiation of physical on-site construction of the remedial action." The only guidance given in the statute regarding this language is that "remedial action" is defined. See supra n. 4. In part, that definition indicates that a remedial action constitutes "[t]hose actions taken consistent with [the] permanent remedy" including "such actions at the location of the release as storage, confinement, ... [and] clay cover." 42 U.S.C. § 9601(24). We recall at the outset that the governments filed suit on September 19 (United States) and September 20 (Indiana) 1996. Therefore, if the initiation of construction of the remedial action began before September 19, 1990, the governments' actions are time barred.

In Navistar's view, SCA undertook a number of activities in August and September 1990 that constitute initiation of construction of the remedial action.15 Specifically, SCA hooked up utilities at the site, set up trailers and connected them with the utilities, constructed an access road, cleared and grubbed a portion of the site to prepare it for the clay cap, compacted existing landfill at the site to prepare for the clay cap, and excavated, transported and stockpiled clay for construction of the clay cap. In addition to these activities, SCA placed the first lift of clay for the clay cap on the site on September 18, 1990. Navistar contends that these activities, particularly the placement of the clay, clearly constitute, in the terms of the statute of limitations, the "initiation of physical on-site construction of the remedial action."

In support of its position, Navistar explains that the primary aspect of the remedial action in this case was the construction of the clay cap on the site; in fact, such a "clay cover" is specifically mentioned in the definition of "remedial action." Therefore, when SCA placed the first lift of clay on the site on September 18, after it had received oral approval from the EPA to proceed, that action, according to Navistar, was the (1) physical (2) initiation (3) on the site (4) of the construction (5) of the remedial action. If this activity was not enough in itself to begin the statute of limitations running, Navistar refers us to the other activities it undertook to prepare the site for construction of the clay cap and directs our attention to various cases in which similar activities have been held to satisfy the conditions of "initiation of physical on-site construction of the remedial action."

In opposition, the governments present essentially two arguments to demonstrate why the activities undertaken by SCA prior to September 19, 1990, did not trigger the statute of limitations period. First, the governments argue that this court should establish a bright-line rule: The initiation of construction of the "remedial action" can never begin, within the meaning of 42 U.S.C. § 9613(g)(2)(B), until the EPA issues final, written approval of the remedial design for the site at issue. Because such approval was not forthcoming in this case until September 20, 1990, the governments argue that their actions were timely filed. The rationale supplied by the governments for this bright-line rule is that the term "the remedial action" in § 9613(g)(2)(B) is a term of art which refers to the specific action determined to be required as a result of the evaluative administrative process. Therefore, submit the governments, the term does not refer merely to the general "remedial action" as defined in 42 U.S.C. § 9601(24), but refers specifically to "the remedial action" determined for the site. In this way, the governments submit, Congress intended to incorporate the underlying administrative process into the statutory framework. Moreover, that process in general contemplates that construction of the remedial action will not begin until a plan for the proposed remedial action has been designed and approved. From this proposition, the governments make the deductive leap that Congress contemplated that "the remedial action" cannot begin until the EPA has issued its final written approval of the remedial design.16 The benefit of such a rule, in the governments' view, is that it will be easy to apply and will conserve judicial resources because it will limit the inquiry of the on-site actions that may have initiated construction of the remedial action to those that occur after final design approval has been issued.

Whatever might be the advantages of such an approach, we are not persuaded that Congress intended such a rule. The statute is devoid of any reference that would lead us to conclude from its plain language that Congress intended to incorporate this specific aspect of the administrative process in establishing the actions that would trigger the limitations period. On the contrary, "remedial action" is a term broadly defined by the statute--a fact of which Congress was no doubt well aware when it incorporated that term in the statute of limitations. If it had intended to require that the EPA issue its final approval of the remedial design in order for a "remedial action" to begin within the meaning of § 9613(g)(2)(B)--a contention unsupported by the definition of "remedial action"--Congress surely would have provided us with a more explicit direction to that effect. See California v. Hyampom Lumber Co., 903 F. Supp. 1389, 1393 (E.D.Cal.1995) (rejecting similar argument that a " 'remedial action' may not occur prior to the final administrative approval of the permanent remedy").17 Rather, in crafting the statute of limitations, Congress specifically made "the initiation of physical on-site construction of the remedial action" the triggering event.

The governments' fall-back position is that, even if pre-written-approval actions can trigger the running of the statute of limitations, only a narrow range of events could fall within such a definition and, here, no such event took place before September 20, 1990. The governments note that, although "remedial action" is broadly defined in the CERCLA statute, the statute of limitations provision includes a number of modifying terms which collectively define the scope of actions which can trigger the limitations period. The governments observe that the actions must be "physical," which, they submit, excludes those studies undertaken and the designs necessary for the remedial action. The actions must also be "onsite," thereby excluding work performed elsewhere, even if that work is related to the remedial action. In addition, the actions must constitute "construction," a term which the governments suggest means the creation of something new rather than a modification of something preexisting. The actions must also relate to "the remedial action," which is the permanent remedy that has been developed for the site. And, given the statutory definition of that term, the actions at issue must be "consistent with" that particular permanent remedy. Finally, the governments also suggest that the action or actions at issue must play a "critical role in [the] implementation of the permanent remedy." United States v. Akzo Nobel Coatings, Inc., 990 F. Supp. 897, 905 (E.D.Mich.1998). Viewing the actions taken by SCA in light of this understanding of the statute, the governments assert, mandates the conclusion that "initiation of physical on-site construction of the remedial action" did not occur prior to September 20, 1990.

The governments argue that the activities undertaken by SCA, such as hooking up the utilities, grubbing, etc., were necessary for the evaluation and remedial design phases of the process. Therefore, those activities were not part of the "remedial action." In addition, the governments assert that the placement of clay on the site on September 18, 1990, also was not sufficient because that effort was ineffective18 and thus did not play a "critical role" in the construction of the permanent remedy. Therefore, in the governments' view, the statute of limitations did not begin running until sometime after September 20, 1990.

We do not agree. It cannot be disputed that the placement of the clay was a physical action undertaken by SCA on the site. The dispute instead centers on whether that action constituted the initiation of construction of the remedial action. We cannot see how the action at issue could be anything other than the initiation of such construction. It is undisputed that the remedial action for the site called for the construction of a permanent clay cap. Navistar has demonstrated that SCA began building that cap on September 18, 1990. Although the cap was not completed until much later, the statute is not concerned with "completion" of the remedial action, only with its "initiation." Although it is true that the initial clay placed on the site was deemed inadequate to satisfy certain specifications required for the clay cap, we do not consider that fact relevant because the statute does not indicate that the initial construction must be successful. The action taken on September 18, 1990, was clearly the first step in constructing the permanent clay cap and therefore was the "initiation" of "construction," § 9613(g)(2)(B), and "consistent with [the] permanent remedy," § 9601(24). Therefore, the governments' actions filed on September 19 and 20, 1996, were filed more than six years after the "initiation of physical on-site construction of the remedial action" and they are time barred.19

We turn lastly to the question of whether Indiana's separate state law cause of action for recovery of response costs under Indiana Code § 13-25-4-8 is controlled by a statute of limitations other than that contained in CERCLA. The State urges that the general residual ten-year limitations period applies to such actions. Navistar, in contrast, argues that the Indiana cause of action incorporates the CERCLA statute of limitations and that Indiana's state law action thus is time barred for the same reasons that the governments' federal claims are. Whether the Indiana statute incorporates the CERCLA statute of limitations is a question that requires us to interpret state law. We must endeavor to predict how the Indiana Supreme Court would answer this question if it were presented to it. See Konradi v. United States, 919 F.2d 1207, 1213 (7th Cir.1990) ("The decision of a federal court ... in any ... case in which state law supplies the rule of decision, is an exercise in predicting how the highest court of the state would decide the case if it were presented to it.").

The Indiana statute provides that a "person that is liable under Section 107(a) of CERCLA (42 U.S.C. § 9607(a)) for ... the costs of removal or remedial action incurred by the commissioner ... is liable, in the same manner and to the same extent, to the state under this section." Ind.Code § 13-25-4-8. However, the Indiana statute does not contain any explicit provision regarding a statute of limitations for the state cause of action. Because the statute is silent on this point, argues the State, Indiana law indicates that the residual statute of limitations period applies. See Ind.Code § 34-1-2-3 ("All actions not limited by any other statute shall be brought within ten (10) years....").20

Navistar counters that the residual statute of limitations is inapplicable because the Indiana statute creating the cause of action expressly incorporates the CERCLA statute of limitations. The Indiana statute creating the cause of action expressly establishes liability under that statute "in the same manner and to the same extent" as liability exists under CERCLA. Ind.Code § 13-25-4-8. Therefore, Navistar asserts, the CERCLA statute of limitations, a provision that limits the scope of liability under CERCLA, was expressly adopted in the Indiana statute. Navistar also contends that, as a practical matter, the Indiana statute, devoid as it is of any definitional or other significant substantive provisions, is workable only to the extent that it is read to incorporate all of the important aspects of federal CERCLA law. Therefore, the CERCLA statute of limitations is the only logical choice to apply to the Indiana cause of action because it defines the appropriate limitations periods for the different kinds of actions (remedial or removal) and provides guidance with regard to the actions that trigger the periods. See 42 U.S.C. § 9613(g)(2).

We believe that Navistar is correct. The plain language of the Indiana statute indicates that the provisions defining liability under CERCLA, in all of its aspects, provides the basis for the scope of liability under the Indiana cause of action. Indiana only modified slightly that scope of liability, see § 13-25-4-8(b)-(d), in the manner that it saw fit and otherwise appears to have contemplated the complete adoption of federal CERCLA law to govern the extent of liability under its statute. We think it likely, given the wholesale adoption of federal CERCLA law necessary to effectuate the Indiana statute as it is written, that if Indiana had decided to employ a statute of limitations other than that contained in CERCLA, it would have done so explicitly.

Moreover, even if the plain language of Indiana Code § 13-25-4-8 is read not to incorporate the CERCLA statute of limitations, we think it likely that statute would be the one the Supreme Court of Indiana would adopt if it was faced with this question. It would be consistent with Indiana law for that Court to look to the federal statute of limitations governing the adopted federal cause of action in defining what limitations period is appropriate. See Kirchoff v. Selby, 686 N.E.2d 121, 128 (Ind.Ct.App.1997) ("[I]t is generally accepted in Indiana that when a state statute is modeled after a federal statute, we may look to the cases interpreting the federal statute for guidance in interpreting the Indiana statute."). Therefore, we conclude that the six-year statute of limitations under 42 U.S.C. § 9613(g)(2)(B) applies to Indiana's action under Indiana Code § 13-25-4-8 for recovery of response costs; consequently, for the same reasons we have explained that the federal causes of action were not timely filed, that action is time barred.

The governments' federal actions against Navistar in this case were initial ones governed by the six-year CERCLA statute of limitations. Because the initiation of physical on-site construction of the remedial action (here a clay cap) began more than six years before the actions were filed, they are time barred. In addition, Indiana's separate state law-based cause of action is also governed by the six-year CERCLA statute of limitations for initial actions; therefore, that claim is also time barred. Accordingly, the district court erred in denying Navistar's motion for summary judgment and that decision is reversed.

REVERSED.

EVANS, Circuit Judge, dissenting.

This is another case where Dinah Washington's1 sage advice--"What a diff'rence a day makes ... twenty-four little hours"--should be recalled. See Spears v. City of Indianapolis, 74 F.3d 153 (7th Cir.1996). Today, the court holds that this major CERCLA suit, filed on September 19 rather than September 18 of 1996, cannot proceed. Although the majority opinion presents a thorough analysis of the law governing this sticky issue, I disagree with the conclusion reached, and for that reason I dissent.

I would hold that this suit was filed within the 6-year statute of limitations provided for under CERCLA § 113(g)(2)(B). But Navistar insists that a little bit of work on the site prior to September 19, 1990, amounted to the "initiation of physical on-site construction of the remedial action." See CERCLA § 113(g)(2)(B). But "the remedial action" referred to in the statute cannot be initiated until the EPA provides written approval of the final design for that action. Here, written approval was given on September 20, 1990. Nothing that happened before that date, in my view, triggered the ticking of the statute of limitations clock.

This interpretation of the statute has an ease of application that would serve to advance the purposes of the statute of limitations, particularly in the complicated context of CERCLA suits. If actions cannot trigger the statute of limitations until formal written approval has issued, there will be no need for parties to haggle over the presence and nature of preapproval activity at a hazardous waste site. But that will be the result of today's opinion. Parties will now invest time, energy, and money in what is ultimately an issue peripheral to the real question--liability. The resulting distraction will undermine CERCLA's primary purposes: To provide a safe and efficient means of cleaning hazardous waste sites as quickly as possible, and to recover from those responsible the money spent to pay for that cleanup. See, e.g., Sidney S. Arst v. Pipefitters Welfare Educ. Fund, 25 F.3d 417, 420-21 (7th Cir.1994).

This interpretation of § 113(g)(2)(a) also has the advantage of providing the greatest possible protection to the public. Tying the statute of limitations to the date of written approval of the remedial design plan provides a concrete basis upon which federal, state, and private parties can rely in determining when the statute of limitations will bar a claim. It will also be easy for district courts to apply the rule. District judges will not have to engage in a complex weighing of factors to determine whether particular actions advanced the purpose of the remediation, and they will not have to guess what was in the minds of the actors on site when certain actions were taken. I would affirm the judgment of the district court.