ORDER
This matter is before the court on appel-lees’ petition for rehearing filed on July 11, 2002. The petition for rehearing is granted. Therefore, the court’s opinion filed June 27, 2002 is vacated and a revised opinion is attached.
OPINION
EBEL, Circuit Judge.Under Utah statutory law, the holder of a dishonored check may collect from the person who wrote the check its face amount and “a service charge that may not exceed $15.” Utah Code § 7-15-1 (1997).1 *1111The defendants in this suit attempted to collect a statutory shoplifting fee of $250 on a dishonored check. The central question presented by this case is whether the defendants are liable under the Fair Debt Collection Practices Act (FDCPA), 15 U.S.C. § 1692 et seq., which establishes civil liability for debt collectors who attempt to collect amounts not “permitted by law,” id. § 1692f(l). The district court held that the defendants’ actions were “permitted by law” because unpublished default judgments issued by state trial courts in earlier collection actions on dishonored checks had awarded shoplifting fees irrespective of the $15 maximum service charge. Johnson v. Riddle, No. 2:98CV599C, slip op. at 12 (D.Utah Dec. 20, 2000). We conclude that the district court misconstrued the term “permitted by law.” We hold that, under a correct application of that standard, Riddle’s attempt to collect a shoplifting penalty from Johnson was not permitted by law. However, because it remains necessary to determine whether Riddle may avoid liability because his error was bona fide in statutory terms, we reverse and remand for further proceedings consistent with this opinion.
I. BACKGROUND
Defendant-appellee Jesse L. Riddle is a Utah attorney whose law firm, defendant-appellee Riddle & Associates, specializes in collecting unpaid dishonored checks. In this opinion, the defendant-appellees are referred to collectively as “Riddle.” Riddle receives between 700,000 and 1.2 million unpaid dishonored checks per year, and his clients include large corporations such as K-Mart, Circle K, and 7-Eleven (the Southland Corporation). Riddle’s debt collection practice spans multiple states, including Utah.
Like many states, Utah has enacted statutory provisions that allow a debt collector to collect, not merely the amount of the debt, but also a service charge. Under Utah’s dishonored check statute, Utah Code § 7-15-1 (1997), the holder of a dishonored check is permitted to impose a “service charge that may not exceed $15,” in addition to collecting the face amount of the check. A separate Utah statute, the shoplifting statute, Utah Code § 78-11-15, governs the civil liability of shoplifters. The shoplifting statute provides:
An adult who wrongfully takes merchandise by any means, including but not limited to, concealment or attempted concealment in any manner, either on or off the premises of the merchant, with a purpose to deprive a merchant of merchandise or to avoid payment for merchandise, or both, is hable in a civil action, in addition to actual damages, for a penalty to the merchant in the amount of the retail price of the merchandise not to exceed $1,000, plus an additional penalty as determined by the court not less than $100 nor more than $500, plus court costs and reasonable attorneys’ fees.
§ 78-11-15. “Wrongful taking of merchandise,” as used in the shoplifting statute, is defined as “the taking of merchandise that has not been purchased from a merchant’s premises without the permission of the merchant or one of his employees, servants or agents.” Utah Code § 78-11-14(5).
*1112The facts relevant to this case are undisputed. Based upon his own reading of the Utah statutes, Riddle concluded that the shoplifting statute applied to persons who passed checks which later were dishonored, and thus that he could seek penalties of up to $500 against them, rather than being limited to the $15 limit on service charges set by the dishonored check statute. In 1995, Riddle filed a complaint in Utah state district court seeking a shoplifting penalty on a dishonored check. Circle K v. Coles, No. 96-0000392CV (Sept. 23, 1996). When the defendant in the case failed to appear in court, the presiding judge and Riddle spent thirty to forty minutes discussing the applicability of the shoplifting statute to a person who passed a check that was dishonored. The court issued a default judgment awarding “Statutory penalties” of $354.82. Default judgments involving dishonored checks awarding statutory penalties in excess of $15 were also issued in several other state court cases during this time. See Johnson, No. 2:98CV599C, slip op. at 10 (citing three unpublished decisions). None of these default judgments were appealed. At some later date, several Utah state trial judges began to develop misgivings about allowing shoplifting penalties in dishonored check cases, and some began limiting fee awards in default judgment cases to $15.
We turn now to the specific facts of this case. In September 1996, Brenda Johnson made a purchase at a 7-Eleven and paid with a check for $2.64. The check bounced, and the matter was referred to Riddle. Riddle sent Johnson a letter in January 1997, demanding payment of the face amount of the check plus a statutory penalty of $69. A week later Riddle sent Johnson a second letter, now demanding a penalty of $200. On August 14, 1997, Riddle filed a complaint in Utah state district court seeking a $250 penalty. Johnson was served with the complaint and summons in this suit on August 24, 1997. Johnson paid Riddle $17.64 (i.e. the face amount of the check plus the $15 service charge permitted under the dishonored checks statute), and Riddle dropped his suit against Johnson.
On August 24, 1998, Johnson filed the instant suit in federal district court for the District of Utah against Riddle, alleging that he violated the FDCPA and various state statutes by attempting to collect a shoplifting penalty in excess of the service charge permitted by the dishonored check statute. Johnson filed the suit as a class action on behalf of all persons who wrote a subsequently dishonored check from whom Riddle sought to collect a fee or penalty of more than $15.2 Riddle moved to dismiss the suit on statute of limitations grounds, and the district court denied the motion. Riddle and Johnson then filed cross-motions for summary judgment. By written order, the district court granted summary judgment in favor of Riddle on the FDCPA claim, reasoning that Riddle’s suit seeking the shoplifting penalty was “permitted by law” because of the earlier unpublished state trial court default judgments. The district court declined to exercise supplemental jurisdiction over Johnson’s remaining state law claims. Johnson timely appealed.
II. THRESHOLD ISSUES
Before reaching the substantive issues of this case, it first is necessary to consider two arguments made by Riddle as alternative grounds for affirming the district court’s summary judgnent. Riddle argues *1113that Johnson’s FDCPA claim is barred by the applicable statute of limitations and by the Rooker-Feldman doctrine. The district court was not persuaded by either argument, and neither are we.
A. Statute of limitations
The statute of limitations for FDCPA claims is found in 15 U.S.C. § 1692k(d). The heading for this subsection is “Jurisdiction.” It states, in its entirety: “An action to enforce any liability created by [the FDCPA] may be brought in any appropriate United States district court without regard to the amount in controversy, or in any other court of competent jurisdiction, within one year from the date on which the violation occurs.”
Here, Riddle sent letters demanding penalties of $69 and $200 on January 17 and January 24, 1997, respectively. Riddle filed the collection suit against Johnson in Utah state court seeking a $250 penalty on August 14, 1997. Johnson was served with the complaint and summons in the collection suit on August 24, 1997. Johnson filed the instant action against Riddle in Utah federal court on August 24, 1998. In other words, the present suit was filed on the one-year anniversary of the date on which Riddle served Johnson in the collection suit, and more than one year after Riddle sent the demand letters and filed the collection suit. Johnson does not challenge the district court’s conclusion that her personal FDCPA claims arising from Riddle’s demand letters are untimely. Johnson, No. 2:98CV599C, slip op. at 5-6. Thus, the only statute of limitations issue is whether the district court correctly held that Johnson’s FDCPA claim arising from the collection suit Riddle brought against her was timely.
Riddle offers two alternative arguments for why Johnson’s FDCPA claim is time-barred. He argues that Johnson’s FDCPA claim was untimely because “the date on which the violation occurred],” 15 U.S.C. § 1692k(d), was the date when Riddle filed the collection suit against her, not the date when Johnson was served. Alternatively, he argues that, even if the violation occurred when Johnson was served, her suit is still untimely because, by filing suit on the one-year anniversary of service, she failed to file “within one year,” id. Neither argument is persuasive.
We reject Riddle’s argument that the violation occurred upon filing rather than upon service. We hold that, where the plaintiffs FDCPA claim arises from the instigation of a debt collection suit, the plaintiff does not have a “complete and present cause of action,” Bay Area Laundry & Dry Cleaning Pension Trust Fund v. Ferbar Corp., 522 U.S. 192, 201, 118 S.Ct. 542,139 L.Ed.2d 553 (1997), and thus no violation occurs within the meaning of § 1692k(d), until the plaintiff has been served.3 If the debt collector files suit against the FDCPA plaintiff but then elects to call off the process server and abandon the collection suit before the *1114plaintiff has been served, it cannot be said that the abandoned lawsuit constitutes an “attempt to collect” on the debt within the meaning of the FDCPA, 15 U.S.C. § 1692f. Riddle concedes it is “theoretically true,” but argues it is “unlikely,” that a creditor would choose to file a suit and subsequently decide against serving the debtor. Of course, the fact that a party that has committed half an actionable wrong is likely to commit the other half cannot suffice to create a complete and present cause of action.4
Furthermore, if the limitations clock began to run with service of process rather than with filing suit, somebody in Riddle’s position could effectively block any action under the federal statute by filing suit and then delaying service. In fact, a delay of service for an entire year (so that the debtor remained unaware of the lawsuit during that time) could cause the limitations period to run out before there was any opportunity for the debtor to bring suit under the FDCPA.
We are equally unpersuaded by Riddle’s argument that Johnson’s FDCPA suit, filed on the one-year anniversary of the violation, is not “within one year” of the violation. See Malay v. Phillips, 64 F.3d 607, 608 (11th Cir.1995); Clark v. Bonded Adjustment Co., 176 F.Supp.2d 1062, 1065-68 (E.D.Wash.2001). Riddle does not dispute that, under the Federal Rules of Civil Procedure, a suit filed on the one-year anniversary of accrual of the claim is filed “within” one year, see Fed. R.Civ.P. 6(a) (“In computing any period of time prescribed or allowed ... by any applicable statute, the day of the act ... from which the designated period of time begins to run shall not be included. The last day of the period so computed shall be included .... ”), but he argues that Rule 6(a) does not apply. Riddle’s argument is that the FDCPA statute of limitations is jurisdictional and that, absent Rule 6(a), a suit filed on the one-year anniversary would fall outside the one-year statute of limitations. Thus, he argues, Rule 6(a) does not apply because the federal rules “shall not be construed to extend ... the jurisdiction of the United States district courts.” Fed.R.Civ.P. 82.
*1115We do not need to decide whether this statute of limitations is jurisdictional or not, nor do we need to decide whether Rule 6(a) is applicable. Under the plain meaning of 15 U.S.C. § 1692k(d), whether jurisdictional or not, the suit was timely because the day of the violation is not counted in calculating the running of the statute of limitations. See Sheets v. Sel-den’s Lessee, 2 Wall. 177, 69 U.S. 177, 190, 17 L.Ed. 822 (1864) (stating as a general rule of statutory construction that a cause of action measured “from ... a date named” excludes the day thus designated); Rush v. United States, 256 F.2d 862, 865 (10th Cir.1958) (treating the 6(a) Rule as “the general rule for computation of time”); Fed.R.Civ.P. 6 Advisory Committee Note to subdivision (a) (“These are amplifications along lines common in state practices .... ”); see also Fogel v. Comm’r, 203 F.2d 347, 349 (5th Cir.1953) (“In computing a period of time ... the general rule is that the ... day of the event[] is to be excluded, while the last day of the period is to be included.”); United States v. Cia Luz Stearica, 181 F.2d 695, 696 (9th Cir.1950) (stating that FRCP 6 “merely declares a rule of statutory construction having widespread judicial sanction, federal as well as state”). All of this is consistent with common usage. Whenever one speaks of something occurring “within a day,” of an event, everyone would understand that to mean the next day after the event, not solely on the same day as the event. Similarly, requiring suit “within one year” of an event means that the suit must be filed on or before the anniversary date of the event, not the day before the anniversary. Thus, applying general principles of statutory construction, the calculation of the one year limitations period begins with August 25, 1997 (the day after Riddle served Johnson in the collection action), and Johnson’s suit filed on August 24,1998 was timely.
The single case relied upon by Riddle for the proposition that suits filed on the anniversary of the violation are not filed within one year is Mattson v. U.S. W. Communications, 967 F.2d 259 (8th Cir. 1992). Mattson is dubious authority for several reasons. Mattson relied for its conclusion that claims must be filed before the anniversary on a single case. Id. at 262. That case, Rust v. Quality Car Corral, Inc., 614 F.2d 1118 (6th Cir.1980) has since been overruled — on precisely the point relied upon by Mattson — by the Sixth Circuit sitting en banc. Bartlik v. United States Dept. of Labor, 62 F.3d 163, 166 (6th Cir.1995). Apparently, every other circuit to have weighed in on the issue has rejected the position adopted by Matt-son; indeed, two circuits termed the position later adopted by Mattson “frivolous.” United Mine Workers v. Dole, 870 F.2d 662, 665 (D.C.Cir.1989); Frey, 748 F.2d at 175; see also Maahs v. United States, 840 F.2d 863, 866-67 (11th Cir.1988); Wirtz v. Peninsula Shipbuilders Ass’n, 382 F.2d 237, 239-40 (4th Cir.1967); Fogel, 203 F.2d at 349; Cia Luz Stearica, 181 F.2d at 696; cf. Union Nat’l Bank v. Lamb, 337 U.S. 38, 40-41, 69 S.Ct. 911, 93 L.Ed. 1190 (1949) (“Since [Rule 6(a) ] had the concurrence of Congress, and since no contrary policy is expressed in the statute governing this review, we think the considerations of liberality and leniency which find expression in Rule 6(a) are equally applicable to [the federal statute at issue].”).5
*1116Accordingly, we agree with the district court that Johnson’s FDCPA claim was timely filed.
B. Rooker-Feldman
Riddle’s Rooker-Feldman argument is also meritless. The Rooker-Feld-man doctrine6 “bars a party losing in state court from seeking what in substance would be appellate review of the state judgment in a United States district court, based on the losing party’s claim that the state judgment itself violates the loser’s federal rights.” Kiowa Indian Tribe v. Hoover, 150 F.3d 1163, 1169 (10th Cir. 1998) (internal quotations and alterations omitted). Thus, Rooker-Feldman bars cases in federal court that are “inextricably intertwined” with a prior state court judgment. Id. Here, Riddle argues that Johnson’s suit would in substance constitute appellate review of the state court default judgments entered in favor of Riddle’s clients prior to Riddle’s filing of suit against Johnson.
However, it is well settled that Rooker-Feldman applies only when the party against whom the doctrine is invoked was a party to the state court proceeding which the federal court is being asked in substance to review. Johnson v. Rodrigues, 226 F.3d 1103, 1109 (10th Cir.2000) (“[T]he Rooker-Feldman doctrine should not be applied against non-parties.”). Here, the state court judgments that Riddle relies upon as the triggers for Rooker-Feldman are the state court default judgments. Johnson, quite plainly, was not a party to any relevant state court default judgment, because the state collection action against her was dismissed before it proceeded to judgment.7 Therefore, Riddle’s Rooker-Feldman argument fails. Riddle acknowledges that Rooker-Feld-man does not apply against non-parties. He offers the frivolous argument that, because Johnson was a party to his lawsuit against her, therefore the prior party requirement is met. He fails to explain how it is that Johnson’s present suit against him would be in substance appellate review of his suit against her, a suit which Riddle dismissed with prejudice when she paid him $17.64.
III. PERMITTED BY LAW
We now turn to the central question in the appeal: whether Riddle is insulated from liability under the FDCPA for seeking shoplifting penalties against persons who passed dishonored checks because various Utah trial courts had issued unpublished default judgments awarding such relief. This is an issue of statutory *1117interpretation, and we turn first to the text of the FDCPA.
In passing the FDCPA, Congress found “abundant evidence of the use of abusive, deceptive, and unfair debt collection practices.” 15 U.S.C. § 1692(a); see also S. Rep. 95-882, at 2 (1977), reprinted in 1977 U.S.C.C.A.N. 1695, 1696 (“The committee has found that debt collection abuse by third party debt collectors is a widespread and serious national problem.”). The express purpose of passing the FDCPA was “to eliminate abusive debt collection practices by debt collectors, to insure that those debt collectors who refrain from using abusive debt collection practices are not competitively disadvantaged, and to promote consistent State action to protect consumers against debt collection abuses.” § 1692(e). Attorneys engaged in the collection of debts are debt collectors subject to liability under the FDCPA. Heintz v. Jenkins, 514 U.S. 291, 299, 115 S.Ct. 1489, 131 L.Ed.2d 395 (1995). Attempting to collect upon a dishonored check constitutes debt collection within the scope of the FDCPA. Snow v. Jesse L. Riddle, P.C., 143 F.3d 1350, 1353 (10th Cir.1998). Because the FDCPA, like the Truth in Lending Act (TILA), 15 U.S.C. § 1601 et seq., is a remedial statute, it should be construed liberally in favor of the consumer. See, e.g., Pfennig v. Household Credit Servs., Inc., 286 F.3d 340, 344 (6th Cir.2002) (TILA); Rossman v. Fleet Bank Nat’l Assoc., 280 F.3d 384, 390 (3d Cir.2002) (TILA); Filis v. General Motors Acceptance Corp., 160 F.3d 703, 707 (11th Cir.1998) (TILA); Plummer v. Gordon, 193 F.Supp.2d 460, 463 (D.Conn.2002) (FDCPA); Ross v. Commercial Fin. Servs., 31 F.Supp.2d 1077, 1079 (N.D.Ill. 1999) (FDCPA); Harrison v. NBD, Inc., 968 F.Supp. 837, 844 (E.D.N.Y.1997) (FDCPA).
The substantive heart of the FDCPA lies in three broad prohibitions. First, a “debt collector may not engage in any conduct the natural consequence of which is to harass, oppress, or abuse any person in connection with the collection of a debt.” § 1692d. Second, a “debt collector may not use any false, deceptive, or misleading representation or means in connection with the collection of any debt.” § 1692e. Third, a “debt collector may not use unfair or unconscionable means to collect or attempt to collect any debt.” § 1692f.8 Violation of these standards subjects debt collectors to civil liability, § 1692k, or administrative enforcement by the Federal Trade Commission, § 1692Z.
Only the third substantive standard— use of unfair or unconscionable means to collect a debt — is at issue in this case. To illustrate the meaning of this standard, the FDCPA expressly defines as a violation of that provision “[t]he collection of any amount ... unless such amount is expressly authorized by the agreement creating the debt or permitted by law.” § 1692f(l). Here, Riddle does not argue that Johnson expressly authorized collection of a shoplifting penalty against her for her dishonored cheek. Accordingly, the existence of an FDCPA violation turns upon whether Riddle’s collection suit constituted an attempt to collect an “amount ... permitted by law.”
To evaluate whether Riddle’s suit was “permitted by law,” it is necessary to determine which “law” the suit must be “permitted by.” There are several facets to this determination. At the most basic level, this is an issue of Utah law rather than *1118federal law. See, e.g., Freyermuth v. Credit Bureau Servs., 248 F.3d 767, 770 (8th Cir.2001); Pollice v. Nat’l Tax Funding, L.P., 225 F.3d 379, 407 (3d Cir.2000). Neither party disputes the district court’s conclusion that, in interpreting the phrase permitted by law, “ ‘a general statutory authorization’ satisfies the FDCPA; that is, rather than pointing to a state statute that expressly permits a particular imposition of fee, one need only identify some state statute which ‘authorizes or allows, in however general a fashion, the fees or charges in question.’ ”
The district court looked to Utah’s dishonored check statute, § 7-15-1, and Utah’s shoplifting statute, § 78-11-15, for general statutory authorization. On appeal, Riddle argues for a more lenient standard. Riddle argues that the general statute that regulates which suits a lawyer may file is Rule 11 of the Utah Rules of Civil Procedure. Under this standard, he argues, his suit was permitted by law so long as it asserted claims that were “warranted by existing law or by a nonfrivolous argument for the extension, modification, or reversal of existing law or the establishment of new law.” Utah R. Civ. P. 11.9 Riddle cites no authority, binding or otherwise, in support of his assertion that a suit is one for an “amount ... permitted by law” under the FDCPA so long as the lawyer who brought it is not subject to Rule 11 sanctions for having done so. To the contrary, every circuit court decision that has applied the “permitted by law” standard has asked simply whether state substantive law permitted the FDCPA defendant to collect the money that it demanded. E.g., Freyermuth, 248 F.3d at 770; Duffy v. Landberg, 215 F.3d 871, 873-75 (8th Cir.2000); Pollice, 225 F.3d at 407; Tuttle v. Equifax Check, 190 F.3d 9, 13-15 (2d Cir.1999).
Further, Riddle’s argument is contrary to the plain text of the statute. The statute does not ask whether Riddle’s actions were permitted by law (or at least non-sanctionable), it asks whether the amount he sought to collect was permitted by law. § 1692f(l) (“The collection of any amount ... unless such amount is ... permitted by law.”). Accordingly, we agree with the district court that the “permitted by law” inquiry demands application of state substantive law, not Rule 11.
In determining whether Riddle had attempted to collect an amount permitted by law, the district court regarded itself as bound by unpublished state trial court decisions granting default judgments for amounts in excess of the statutory fee for dishonored checks. This was error. When the federal courts are called upon to interpret state law, the federal court must look to the rulings of the highest state court, and, if no such rulings exist, must endeavor to predict how that high court would rule. See, e.g., Comm’r v. Bosch’s Estate, 387 U.S. 456, 464-66, 87 S.Ct. 1776, 18 L.Ed.2d 886 (1967); Stuart v. Colo. Interstate Gas Co., 271 F.3d 1221, 1228 (10th Cir.2001); Commerce Bank, N.A. v. Chrysler Realty Corp., 244 F.3d 777, 780 (10th Cir.2001). As the Supreme Court explained in Bosch’s Estate:
[W]hen the application of a federal statute is involved, the decision of a state trial court as to an underlying issue of state law should a fortiori not be controlling. This is but an application of the rule of Erie ... where state law as announced by the highest court of the State is to be followed. This is not a *1119diversity case but the same principle may be applied for the same reasons, viz., the underlying substantive rule involved is based on state law and the State’s highest court is the best authority on its own law. If there be no decision by that court then federal authorities must apply what they find to be the state law after giving “proper regard” to relevant rulings of other courts of the State.
387 U.S. at 465, 87 S.Ct. 1776 (citation omitted); see also DelCostello v. Int’l Bhd. of Teamsters, 462 U.S. 151, 159 n. 13, 103 S.Ct. 2281, 76 L.Ed.2d 476 (1983) (“[WJhere Congress directly or impliedly directs the courts to look to state law to fill in details of federal law, Erie will ordinarily provide the framework for doing so.”)
The district court here eschewed any effort to determine whether “the Utah Supreme Court would allow or [would] not allow use of the shoplifting statute in debt collecting actions.” Rather, the district court determined that it was sufficient to establish the use of the shoplifting penalties for dishonored checks was “permitted by law” simply to determine that that practice was permitted by several Utah district courts. In this regard, the district court erred.
We hold that an amount is “permitted by law” within the meaning of the FDCPA if state supreme court holdings establish that collection of the amount is lawful. Absent state supreme court holdings on point, we follow our familiar Erie analysis by predicting what the state supreme court would hold, or, in the appropriate case, certifying the issue to the state supreme court.10 In that analysis, lower state court decisions may be helpful, but they are not dispositive of a prediction of how the highest court in the state might rule on the issue.11
Turning to the application of .this standard to the facts of this case, we hold that the shoplifting penalty Riddle endeavored to collect was not an amount permitted by law for what was in essence merely a dishonored check claim. The Utah Supreme Court never has addressed the applicability of the shoplifting statute to writers of dishonored checks. Significantly, however, the state supreme court has held that specific statutes (such as the dishonored check statute, which very specifically and precisely applies to suits to recover on dishonored checks) trump general statutes (such as the shoplifting statute as interpreted by Riddle, which can be applied to the situation of dishonored checks only by giving it a very broad and general meaning). Carlie v. Morgan, 922 P.2d 1, 6 (Utah 1996). We find it unmistakably *1120clear from the text of Utah statutory law that shoplifting penalties are unavailable in the collection of dishonored checks. The applicable Utah statute unambiguously states that the holder of a dishonored check is permitted to collect “a service charge that may not exceed $15.” § 7-15-1 (emphasis added). The shoplifting statute applies only to “the taking of merchandise that has not been purchased from a merchant’s premises without the permission of the merchant.” § 78-11-14(5) (emphasis added). Riddle’s collection suit complaint against Johnson included a cursory allegation that Johnson took merchandise “without consent of’ the merchant.12 We do not, however, need to decide in this case whether Riddle can satisfy the “without the permission of’ requirement because, in any event, Riddle’s collection suit complaint plainly makes no allegation, and there is no evidence, that Johnson took “merchandise that ha[d] not been purchased.” To the contrary, the complaint concedes that the transaction “was facilitated by Defendant’s tender of his/her check.” Even before this court, Riddle makes no attempt to argue that Johnson had not “purchased” the merchandise. Indeed, Utah law appears to foreclose any such argument. See Utah Code 70A-2-403 (identifying as a “purchaser” one who acquires merchandise by means of a check which is later dishonored); see also id. § 70A-1-201(32) (defining “purchase” for purposes of Utah’s Uniform Commercial Code as “taking by sale ... or any other voluntary transaction creating an interest in property”). Accordingly, this transaction, like all typical transactions involving dishonored checks, does not involve “the taking of merchandise that has not been purchased from a merchant’s premises without the permission of the merchant.” § 78-11-14(5). By seeking to collect a shoplifting penalty where no shoplifting (as the term is defined by Utah statute) occurred, Riddle sought to collect an amount not permitted by law in violation of the FDCPA.
Although not dispositive, we note that our interpretation of Utah law to the effect that shoplifting penalties are not applicable to dishonored checks is consistent with the most recent decision by a Utah District Court opinion in a contested case. In Riddle v. Perry, No. 970907851, slip op. at 9-12 (Utah Dist.Ct. Aug. 22, 2001), the Utah District Court for the Third Judicial District in Salt Lake County, Utah, held under Utah law, “that the plaintiffs practice of seeking penalties under the civil shoplifting penalties from the drawers of dishonored instruments was in violation of the FDCPA.” Our ruling also is consistent with that reached in other recent federal district court decisions applying the law of Utah and other states. Scott v. Riddle, No. 2:99-CV-00042, slip op. at 2 (D.Utah Dec. 1, 2000) (upholding FDCPA claim for wrongful collection of shoplifting fees from writers of dishonored checks because the “Washington state shoplifting statute ... does not apply to a buyer who writes a check to a merchant and takes merchandise from the store with the merchant’s knowledge and permission”); Irwin v. Mascott, 112 F.Supp.2d 937, 948 (N.D.Cal.2000) (holding that collecting shoplifting charges against writers of dis*1121honored checks violates the FDCPA because “[California] Penal Code § 490.5 does not apply to a buyer who writes a check to a merchant and takes merchandise from the store with the merchant’s knowledge and permission.... The failure of the buyer’s check to clear does not make the buyer a shoplifter”); Ditty v. CheckRite, Ltd., 973 F.Supp. 1320, 1328 (D.Utah 1997) (holding that defendant violated FDCPA by attempting to collect an amount in excess of Utah’s $15 dishonored check fee, because to allow collection of a higher fee under the shoplifting statute “would undermine the effectiveness of the statute and would be inconsistent with Utah’s Legislature’s intention that dishonored checks be governed by the specific procedures outlined in the statute”).
We conclude that the Utah Supreme Court would not allow a holder of a dishonored check to collect a shoplifting penalty. Accordingly, we hold that Riddle violated the FDCPA because he attempted to collect an amount not permitted by law.
IV. BONA FIDE ERROR DEFENSE
Our conclusion that Riddle attempted to collect an amount not “permitted by law” does not end the FDCPA inquiry. Under the FDCPA, an affirmative defense exists to insulate debt collectors from liability even where they have violated the FDCPA:
A debt collector may not be held liable in any action brought under this sub-chapter if the debt collector shows by a preponderance of the evidence that the violation was not intentional and resulted from a bona fide error notwithstanding the maintenance of procedures reasonably adapted to avoid any such error.
15 U.S.C. § 1692k(c). Thus, an FDCPA defendant seeking the protection of the bona fide error defense carries the burden of proving that the violation was (1) unintentional, (2) a bona fide error, and (3) made despite “the maintenance of procedures reasonably adapted to avoid” the violation.
Because the district court concluded that Riddle’s actions were permitted by law, it did not address the applicability of the bona fide error defense here.13 Johnson argues that the bona fide error defense does not apply here because a debt collector’s mistake of law cannot be a bona fide error. We disagree for the reasons below, and accordingly we remand for the district court to consider in the first instance whether Riddle is entitled to the defense.
This circuit has never addressed whether the FDCPA bona fide error defense can apply to a mistake of law which resulted in an attempt to collect amounts not permitted by law. Two district court opinions in this circuit have addressed this issue, each concluding that the bona fide error defense is limited to clerical errors. Scott, No. 2:99 CV 00042, slip op. at 3; Martinez v. Albuquerque Collection Servs., 867 F.Supp. 1495, 1502-03 (D.N.M.1994). Outside this circuit, federal courts have split on the issue, with a majority concluding that the defense is limited to clerical errors and cannot protect mistakes of law, but a growing minority of courts reaching the contrary conclusion.14 Many other *1122courts, although not squarely addressing the clerical/legal mistake issue, have stated that the FDCPA is a strict liability statute.15
Of the cases which hold that the defense does not apply to mistakes of law, almost all dispense with the issue by citing earlier cases back to the Ninth Circuit’s decision in Baker v. G.C. Servs. Corp., 677 F.2d 775, 779 (9th Cir.1982). Baker rested its holding entirely upon the similarity of the FDCPA bona fide error defense to the “nearly identical” bona fide error defense provided in the Truth in Lending Act (TILA), 15 U.S.C. § 1640, a provision uniformly interpreted to apply only to clerical errors and not to legal errors. Baker, 677 F.2d at 779. However, the TILA analogy is faulty, as the Seventh Circuit has explained:
[The plaintiff] also analogizes the provision to a similar section in [TILA], 15 U.S.C. § 1640(c), which is limited to clerical mistakes and which does not include errors of judgment or law. But ... the TILA bona fide error provision expressly defined bona fide errors as [including, but not limited to,] “clerical, calculation, computer malfunction and programming, and printing errors, except that an error of legal judgment *1123with respect to a person’s obligations under this subchapter is not a bona fide error.” 15 U.S.C. § 1640(c). The FDCPA provision does no such thing. This, along with the statutes’ different purposes, distinguishes the two.
Jenkins v. Heintz, 124 F.3d 824, 832 n. 7 (7th Cir.1997). Unlike TILA, the plain language of the FDCPA suggests no intent to limit the bona fide error defense to clerical errors. To the contrary, § 1692k(c) refers by its terms to any “error” that is “bona fide.” We find no indication in the legislative history that Congress intended this broad language to mean anything other than what it says. Indeed, to the extent that the legislative history speaks to this issue, it suggests that the narrow reading advocated by Johnson is incorrect. In describing the bona fide error exception, the Senate Report stated: “A debt collector has no liability, however, if he violates the act in any manner, including with regard to the act’s coverage, when such violation is unintentional and occurred despite procedures designed to avoid such violations.” S.Rep. No. 95-382, at 5.
Our conclusion is bolstered by the Supreme Court’s reasoning in Heintz, 514 U.S. at 295, 115 S.Ct. 1489. Prior to Heintz, the Sixth Circuit had held (incorrectly, in light of Heintz) that lawyers áre not debt collectors for FDCPA purposes. Green v. Hocking, 9 F.3d 18, 21 (6th Cir. 1993). The Sixth Circuit’s holding was based in part on its view that any other rule “automatically would make liable any litigating lawyer who brought, and then lost, a claim against a debtor.” Heintz, 514 U.S. at 295, 115 S.Ct. 1489 (citing Green 9 F.3d at 21); see also Taylor v. Luper, Sheriff & Niedenthal Co., 74 F.Supp.2d 761, 764 (S.D.Ohio 1999) (noting that, if mistakes of law were not protected by the bona fide error defense, state ethical duty of zealous advocacy could require debt collecting lawyer to assert claims that would expose- her to FDCPA liability); Janet Flaccus, Fair Debt Collection. Practices Act: Lawyers and the Bona Fide Error Defense, 2001 Ark. L. Notes 95, 97 (2001)- (“Limiting the bona fide error defense to clerical errors would place lawyers between the proverbial rock and hard place.”). Heintz accepted for argument’s sake the Sixth Circuit’s view that any debt collection lawyer whose claim fails necessarily violates the FDCPA. Id., 115 S.Ct. 1489 However, the Court concluded that such a premise did not produce absurd results, because of the existence of the bona fide error defense rule. Id., 115 S.Ct. 1489 In other words, the Court apparently believed that the bona fide error defense would apply in at least a portion of the cases where the lawyer brought suit to collect an amount beyond that legally owed by the debtor. This reasoning at least suggests that the defense is available for mistakes of law, because presumably mistake of law. may contribute to the reasons why some of the underlying debt collection procedures are lost.
Johnson argues that the statute’s reference to “maintenance of procedures reasonably adapted to avoid any such error,” § 1692k(e), implies a congressional intent to limit the bona fide error defense to clerical errors. We acknowledge that it is more common to speak of procedures adapted to avoid clerical errors than to speak of procedures adopted to avoid mistakes of law. However, absent a clearer indication that Congress meant to limit the defense to clerical errors, we instead adhere to the unambiguous language of the statute as supported by the available legislative history. Several courts that have applied the bona fide error defense to mistakes of law also have discussed the application the “procedures” element to such claims. Jenkins, 124 F.3d at 834; Frye, *1124193 F.Supp.2d at 1088-89; Taylor; 74 F.Supp.2d at 766; Aronson, 1997 WL 1038818 at *5; of. Watkins, 57 F.Supp.2d at 1107-08. We leave for the district court to determine in the first instance what “procedures” Riddle employed and whether such procedures were “reasonably adapted to avoid” the legal error here.16 It is enough to conclude that this requirement is not incompatible with application of the bona fide error defense to mistakes of law.
Accordingly, we remand to the district court to determine whether Riddle is entitled to the bona fide error defense. As noted previously, the issue of class certification has not yet been ruled upon, so that issue remains open for further consideration upon remand.
V. STATE LAW CLAIMS
Having dismissed Johnson’s federal claim, the district court declined to exercise supplemental jurisdiction over her remaining state law claims under § 1367(c)(3). Because we conclude that dismissal of the FDCPA claim was erroneous, we reverse this decision as well.
VI. CONCLUSION
For the foregoing reasons, we REVERSE the judgment of the district court and REMAND for further proceedings consistent with this opinion.17
. In amendments effective May 5, 1997, the Utah legislature raised this limit to $20, and in amendments effective May 3, 1999, it rewrote the section to establish different limits depending upon the stage of collection. See § 7-15-1 and amendment notes (2001 Supp.). Neither party argues that the dollar limits imposed by the 1997 or 1999 amendments apply here.
The 1999 amendments also added the following sentence: "This section may not be construed to prohibit the holder of the check from seeking relief under any other applicable statute or cause of action.” § 7-15-l(8). Riddle argues that collection of a shoplifting penalty from the one who passed a dishonored check amounts to "seeking relief under any other applicable statute or cause of action." *1111Assuming for the sake of argument that this amendment is merely a clarification of prior law that Utah courts would apply retroactively, see Kilpatrick v. Wiley, Rein, & Fielding, 37 P.3d 1130, 1142 (Utah 2001), we conclude for the reasons given in part III, infra, that Utah Code § 78-11-15, the provision governing the civil liability of shoplifters, is not an “applicable statute or cause of action.” § 7-15-1(8) (emphasis added).
. In granting Riddle's motion for summary judgment, the district court did not reach the class certification issue.
. Johnson asserts an alternative reason for concluding that her claim accrued upon service rather than filing. The Tenth Circuit adheres to the traditional discovery rule in determining when a federal statute of limitations begins to run. Indus. Constructors Corp. v. United States Bureau of Reclamation, 15 F.3d 963, 969 (10th Cir.1994) (holding that federal statute of limitations does not begin to run until plaintiff knows or has reason to know of the existence and cause of his injury). However, the continuing scope of the traditional federal discovery rule, at least in the kind of claim asserted here, has been rendered uncertain by dicta in TRW Inc. v. Andrews, 534 U.S. 19, 122 S.Ct. 441, 446-47, 151 L.Ed.2d 339 (2001); see also id. at 451-53 (Scalia, J., concurring in the judgment). Given our disposition of this issue, it is unnecessary for us to consider whether a discovery rule applies to the FDCPA statute of limitations.
. We note that the same conclusion would follow under the analysis followed in at least one circuit in FDCPA cases where the alleged violation is the mailing of a demand letter. The Eighth Circuit based its holding that the FDCPA cause of action accrues upon mailing the demand letter on two considerations: the mailing of the letter is the creditor’s last opportunity to comply with the FDCPA, and the date of mailing is easy to determine, ascertainable by each party, and easily applied. Mattson v. U.S. W. Communications, Inc., 967 F.2d 259, 261 (8th Cir.1992). Here, the decision of whether to serve the debtor is the creditor's last opportunity to comply with the FDCPA. Further, as Riddle does not dispute, the date of service is easily ascertainable by each party, because Utah Rule of Civil Procedure 4(h) obliges parties to file a proof of service which includes the date of service with the court within 10 days of service.
Riddle relies upon dicta in Naas v. Stolman, 130 F.3d 892 (9th Cir.1997), for his argument that the cause of action accrues upon filing where the FDCPA violation alleged is the initiation of a collection suit. However, the choice between accrual upon filing and accrual upon service was not before the court in Naas. The Naas plaintiff did not argue that the limitations period began at filing, because if he had his claim still would have been time-barred. Id. at 893. Instead, he argued that the limitations period did not run until the collection suit judgment was upheld by the appeals court. Id. Not surprisingly, the court never discussed, and apparently never considered, the possibility that the claim accrued upon service. Indeed, it appears likely that, had the Naas court confronted the issue, it would have concluded that the cause of action accrues upon service. Naas applied Matt-son 's two-part test for determining when an FDCPA cause of action accrues. As explained above, it is clear that the Mattson test militates in favor of the conclusion that the claim accrues upon service rather than filing.
. Indeed, the Eighth Circuit's own continuing fidelity to Mattson is not free from doubt:
Rule 6(a) provides a reasonable basis for determining the appropriate ending date of the [AEDPA] grace period. See Mattson, 967 F.2d at 262 (McMillian, J., dissenting) (noting the wide acceptance of the 'modern doctrine’ under which federal statutes of limitations are calculated pursuant to Rule 6(a)); McDuffee v. United States, 769 F.2d *1116492, 494 (8th Cir. 1985) (citing with approval other courts that have applied Rule 6(a) to federal statutes of limitations).
Moore v. United States, 173 F.3d 1131, 1135 (8th Cir. 1999).
. The doctrine's name is derived from the two Supreme Court cases out of which it arose, Rooker v. Fidelity Trust Co., 263 U.S. 413, 44 S.Ct. 149, 68 L.Ed. 362 (1923), and Dist. of Columbia Court of Appeals v. Feldman, 460 U.S. 462, 103 S.Ct. 1303, 75 L.Ed.2d 206 (1983).
. Riddle argues that the prior party requirement is met because “potential class members .. . include those who had state default judg-merits entered against them.” Because the issue of class certification remains unresolved, we express no opinion on the merits of this argument or its consequences upon the scope of any class ultimately certified. Further, we note uncertainty regarding whether Riddle, as distinct from the clients he represented, was a party to each of the relevant default judgment actions.
Even if the parties were the same, Riddle would have to establish that the federal action constitutes a challenge to a relevant state court judgment. We express no opinion in this appeal as to whether that is the case.
. The FDCPA also establishes specific duties and prohibitions governing debt collectors’ communication with third parties, § 1692b, and communication with the debtor, §§ 1692c, 1692g, 1692j.
. This language is identical to that found in Rule 11 of the Federal Rules of Civil Procedure.
. The sparse case law applying the “permitted by law" provision supports our conclusion. In Watkins v. Peterson Enters., Inc., 57 F.Supp.2d 1102 (E.D.Wash.1999), for example, the defendant debt collector argued that its collection procedure was permitted by law because that procedure had been ordered by the state district courts. Unable to determine whether state law authorized the use of the challenged procedure, the district court certified the question to the Washington Supreme Court, which held that pay orders could not be used. Watkins v. Peterson Enters., Inc., 137 Wash.2d 632, 973 P.2d 1037, 1049 (1999). In light of the Washington Supreme Court’s answers to the certified questions, the Watkins court held that the use of the pay orders was not permitted by state law, even though state trial courts had permitted — indeed, required — the defendant to use them. 57 F.Supp.2d at 1106. Watkins ultimately ruled in favor of the debt collector on the basis of the bona fide error defense, id. at 1107-08, an issue we take up below in part IV.
. In this regard, we observe that trial court default judgments generally are among the least persuasive evidence available as to how a state supreme court would resolve an issue.
. In Riddle's Petition for Panel Rehearing, he explained that the basis for his allegation that Johnson took merchandise without the merchant’s consent is, "in part, ... the legal proposition that the victim of a fraud or illegal conduct does not knowingly consent to such conduct in absence of knowledge of the fraud." Riddle does not assert that the merchant was unaware that Johnson took the merchandise or that the merchant was opposed to such taking, and in fact his collection suit complaint stated that the transaction "was facilitated by Defendant's tender of his/ her check.”
. It is unclear from the record whether Riddle waived the affirmative defense of bona fide error by failing properly to raise it in district court. We assume, for purposes of this appeal only, that Riddle has not waived the defense.
. Compare Picht v. Jon R. Hawks, Ltd., 236 F.3d 446, 451 (8th Cir.2001) (stating that bona fide error defense does not apply to mistakes of law) (citing Hulshizer v. Global Credit Servs., Inc., 728 F.2d 1037, 1038 (8th Cir.1984) (per curiam)); Pipiles v. Credit Bureau of Lockport, Inc., 886 F.2d 22, 27 (2d *1122Cir.1989) (same); Baker v. G.C. Servs. Corp., 677 F.2d 775, 779 (9th Cir.1982) (mistake of law "insufficient by itself to support the bona fide error defense”); Hartman v. Meridian Fin. Servs., Inc., 191 F.Supp.2d 1031, 1045-46 (W.D.Wis.2002) (does not apply to mistakes of law and generally is limited to clerical mistakes); Arroyo v. Solomon & Solomon, P.C., No. 99-CV-8302, 2001 WL 984940, at *6 (E.D.N.Y. July 19, 2001) amended and superseded by 2001 WL 1590520 (E.D.N.Y. Nov. 16, 2001) (does not apply to mistakes of law, and collecting cases); Wilkerson v. Bowman, 200 F.R.D. 605, 608-09 (N.D.Ill.2001) (does not apply to mistaken view of the obligations imposed by the FDCPA); Edwards v. McCormick, 136 F.Supp.2d 795, 800 (S.D.Ohio 2001) (limited to clerical errors); Spencer v. Hendersen-Webb, Inc., 81 F.Supp.2d 582, 591 (D.Md.1999) (does not apply to mistakes of law and generally is limited to clerical mistakes); Booth v. Collection Experts, Inc., 969 F.Supp. 1161, 1165 (E.D.Wis.1997) (same) with Jenkins v. Heintz, 124 F.3d 824, 832 & n. 7, 833 (7th Cir.1997) (stating that bona fide error defense not limited to clerical errors and can apply to mistakes of law); Frye v. Bowman, Heintz, Boscia & Vician, 193 F.Supp.2d 1070, 1085-86 (S.D.Ind.2002) (same); Filsinger v. Upton, Cohen, & Slamow-itz, No. 99-CV-1393, 2000 WL 198223, at *2 (N.D.N.Y. Feb. 18, 2000) (not limited to clerical errors); Taylor v. Luper, Sheriff & Niedenthal Co., 74 F.Supp.2d 761, 765 (S.D.Ohio 1999) (not limited to clerical errors and can apply to mistakes of law); Watkins v. Peterson Enters., Inc., 57 F.Supp.2d 1102, 1107-08 (E.D.Wash.1999) (can apply to mistakes of law where the mistake stemmed from an official interpretation of the law); Aronson v. Commercial Fin. Servs, Inc., No. Civ.A. 96-2113, 1997 WL 1038818, at *5 (W.D.Pa. Dec.22, 1997) (not limited to clerical errors and can apply to mistakes of law). See also 17 Am.Jur.2d Consumer and Borrower Protection § 224 (2002) ("[A] mistake of law is not included in the intentional and bona fide error defense of [the FDCPA].” (citation omitted)); Janet Flaccus, Fair Debt Collection Practices Act: Lawyers and the Bona Fide Error Defense, 2001 Ark. L. Notes 95, 96 & passim (2001) (noting that cases applying the bona fide error defense to attorney mistakes of law "seem in disarray,” but arguing that results may be harmonized because courts generally have rejected the defense where the violation was reasonably clear from the controlling law and allowed the defense where violation was not at all clear).
. E.g., Russell v. Equifax A.R.S., 74 F.3d 30, 33 (2d Cir.1996); Ferguson v. Credit Mgmt. Control, Inc., 140 F.Supp.2d 1293, 1297 (M.D.Fla.2001); Skerry v. Mass. Higher Educ. Assistance Corp., 73 F.Supp.2d 47, 51 (D.Mass.1999); Raimondi v. McAllister & As-socs., Inc., 50 F.Supp.2d 825, 829 (N.D.Ill. 1999); Macarz v. Transworld Sys., Inc., 26 F.Supp.2d 368, 373 (D.Conn.1998); O’Connor v. Check Rite, Ltd., 973 F.Supp. 1010, 1020 (D.Colo.1997); Pittman v. J.J. Mac Intyre Co., 969 F.Supp. 609, 613 (D.Nev.1997); Bitah v. Global Collections Servs., Inc., 968 F.Supp. 618, 623 (D.N.M.1997).
. Of course, Riddle also must prove that the error was unintentional and bona fide.
. We deny the Appellant's Bill of Costs as untimely.