Summary
holding that "only consent by both (or all) parties," that is, "the names on both sides of the 'v,' without regard to events such as service of process," will suffice to give a magistrate judge authority to enter dispositive decisions
Summary of this case from Ideavillage Prods. Corp. v. AntikerOpinion
No. 15-3254
06-16-2017
Tracey COLEMAN, Plaintiff-Appellant, v. LABOR AND INDUSTRY REVIEW COMMISSION OF the State of WISCONSIN, Defendant-Appellee.
Travis Crum, Charles A. Rothfeld, Attorneys, MAYER BROWN LLP, Washington, DC, for Plaintiff-Appellant. Anthony D. Russomanno, Attorney, OFFICE OF THE ATTORNEY GENERAL, Wisconsin Department of Justice, Madison, WI, for Defendant-Appellee.
Travis Crum, Charles A. Rothfeld, Attorneys, MAYER BROWN LLP, Washington, DC, for Plaintiff-Appellant.
Anthony D. Russomanno, Attorney, OFFICE OF THE ATTORNEY GENERAL, Wisconsin Department of Justice, Madison, WI, for Defendant-Appellee.
Before Wood, Chief Judge, and Posner and Williams, Circuit Judges.
Wood, Chief Judge.
Seeking to proceed without prepayment of fees, Tracey Coleman brought a pro se suit against the Labor and Industry Review Commission of the State of Wisconsin (Commission). At the same time as he submitted his affidavit of indigence, see 28 U.S.C. § 1915(a)(1), he filed a document indicating that he consented to have a magistrate judge decide the case. 28 U.S.C. § 636(c)(1). The magistrate judge dismissed the suit and entered a final judgment before the Commission was even served, and thus before it had any occasion either to consent to or to refuse the option of proceeding before the magistrate judge. Coleman appealed, and we recruited counsel to assist him. The question before us is whether a plaintiff's consent alone can give a magistrate judge the necessary authority to resolve a case on the basis that the complaint fails to state a claim upon which relief can be granted, in a case that otherwise requires an Article III judge. We conclude that the answer is no: only consent by both (or all) parties will suffice, and so we must remand this case for further proceedings.
I
ABM Industries hired Coleman in February 2014, and assigned him to work at Carmen High School. It fired him less than three weeks later because it believed that he had sexually harassed a school employee. Coleman responded by filing a pro se suit against ABM and Carmen in federal district court; he asserted that the real reason he was fired was racial discrimination. The suit ended more or less amicably with a stipulated dismissal on October 31. Coleman also sought relief from the state Commission's Equal Rights Division. That effort came to naught. On July 22, 2015, an administrative law judge dismissed Coleman's case because he had failed to meet certain deadlines. Coleman then turned back to federal court. He filed a second pro se suit on August 18, this time contending that the Commission had denied him due process and requesting permission to proceed in forma pauperis (IFP). See 28 U.S.C. § 1915. This was when he filed his consent to proceed before the magistrate judge.
The magistrate judge took up the matter immediately. He found that Coleman's request to proceed IFP was financially supported, but he ordered Coleman to submit an amended complaint because the initial version failed to "offer any details that could plausibly present a federal cause of action." The new complaint proved similarly deficient, and so the magistrate judge dismissed Coleman's case and entered final judgment. In so doing, he relied on the instruction in 28 U.S.C. § 1915(e)(2) that calls for "the court" to "dismiss the case at any time if the court determines that ... (B) the action ... (ii) fails to state a claim on which relief may be granted." Here's the rub: because the Commission had not yet been served, the magistrate judge was proceeding with the consent of only one of the litigants. Coleman filed a timely appeal from the final judgment, as permitted by 28 U.S.C. § 636(c)(3). Our appellate jurisdiction is secure. The important question on appeal relates to jurisdiction at the district-court level, where the case was resolved by a magistrate judge who had not received permission from "the parties" (plural) to act.
II
A
Article III vests the "judicial power" in "one Supreme Court, and in such inferior courts as the Congress may from time to time ordain and establish." U.S. CONST. art. III, § 1. It preserves the impartiality of those courts through two key safeguards: life tenure and salary protection. Wellness Int'l Network, Ltd. v. Sharif , ––– U.S. ––––, 135 S.Ct. 1932, 1938, 191 L.Ed.2d 911 (2015). The Supreme Court has insisted that core judicial functions may not be given to persons who lack these Article III protections, unless all affected parties consent. See, e.g. , Wellness , 135 S.Ct. at 1942, 1949 (Article I judge may adjudicate a case if the litigants validly consent); Stern v. Marshall , 564 U.S. 462, 482, 131 S.Ct. 2594, 180 L.Ed.2d 475 (2011) (Article III did not permit the bankruptcy court to enter final judgment on a counterclaim); Nguyen v. United States , 539 U.S. 69, 77–78, 80, 123 S.Ct. 2130, 156 L.Ed.2d 64 (2003) (panel of court of appeals judges could not include an Article I territorial judge); N. Pipeline Constr. Co. v. Marathon Pipe Line Co. , 458 U.S. 50, 76, 102 S.Ct. 2858, 73 L.Ed.2d 598 (1982) (assignment by Congress to bankruptcy judges of authority to adjudicate all civil proceedings "related to" bankruptcy violated Article III).
The case before us presents one aspect of that problem: may an Article I judge dismiss an action for failure to state a claim on which relief can be granted, if that dismissal is part of the initial screening that occurs in IFP cases and thus takes place before the defendants are served? Although the statute does not say this in so many words, we have understood section 1915(e) to "direct [ ] courts to screen all complaints filed with requests to proceed IFP." Luevano v. Wal-Mart Stores, Inc. , 722 F.3d 1014, 1022 (7th Cir. 2013). It instructs "the court" to "dismiss the case at any time" if, among other things, the action "fails to state a claim on which relief may be granted." 28 U.S.C. § 1915(e)(2)(B)(ii). The magistrate judge in our case followed that command and entered a final judgment for the as-yet-unserved defendant. We must decide whether this was permissible.
B
We begin with the pertinent language of the Magistrate Judges Act:
(c) Notwithstanding any provision of law to the contrary—
(1) Upon the consent of the parties, a full-time United States magistrate judge ... may conduct any or all proceedings in a jury or nonjury civil matter and order the entry of judgment in the case, when specially designated to exercise such jurisdiction by the district court or courts he serves.
28 U.S.C. § 636(c)(1). Valid consent is not optional; it is "essential to upholding § 636(c)'s constitutionality against arguments that it improperly vests the judicial power of the United States in non-Article III judges." Kalan v. City of St. Francis , 274 F.3d 1150, 1152 (7th Cir. 2001). But although the consent requirement is carved in stone, the identity of "the parties" whose permission section 636(c)(1) demands is less clear. Is it all of the parties? Some of the parties? Only any party who must be before the court for purposes of the ruling in question?
The circuits have come to different conclusions about this question, and unfortunately, so have we. The Fifth Circuit, in Neals v. Norwood , 59 F.3d 530, 532 (5th Cir. 1995), and our opinion in Hains v. Washington , 131 F.3d 1248, 1249 & n.2 (7th Cir. 1997) (per curiam ), take the position that an unserved defendant is not one of the "parties" for purposes of section 636(c)(1), and thus the defendant's consent is not required in order to permit a magistrate judge to issue a dispositive order when screening a complaint for IFP purposes. On the other side, the Eighth Circuit, in Henry v. Tri-Services, Inc. , 33 F.3d 931, 933 (8th Cir. 1994), and our court in Geaney v. Carlson , 776 F.2d 140, 142 (7th Cir. 1985), hold that the magistrate cannot finally dispose of the case, on screening or otherwise, without the defendant's consent.
No matter what the rules concerning consent are, there is a substantial role for the magistrate judge to play in the screening process. Under the rule of Neals and Hains , the magistrate judge enters a final judgment on the case, while under the approach of Henry and Geaney, he evaluates the complaint, takes any nondispositive actions that are appropriate (e.g. , further investigation of indigence), and recommends an action for the district court. See § 636(b)(1)(C). If the action recommended is dismissal, the district court can enter that final judgment (after considering any properly filed objections) without bothering the defendants. Under either view, to the extent that section 1915(e)(2) involves nondispositive issues, such as the truthfulness of the allegation of poverty, nothing in the Magistrate Judges' Act prevents the magistrate judge from resolving the issue. That is because a conclusion that the plaintiff is not indigent does not foreclose further legal proceedings—it just means that the plaintiff must pony up the required filing fee. See 28 U.S.C. § 636(b)(1) (authorizing magistrate judges to perform a wide variety of non-dispositive matters).
Before turning to the question whether the consent of the plaintiff alone suffices, a quick review of the screening process is in order. Until April 26, 1996, section 1915 did not cover failures to state a claim. All the statute then said, with respect to any kind of screening, was that the court "may dismiss the case if the allegation of poverty is untrue, or if satisfied that the action is frivolous or malicious." 28 U.S.C. § 1915(d) (1995). It was during the pre-1996 era that the Supreme Court decided Neitzke v. Williams , 490 U.S. 319, 320, 109 S.Ct. 1827, 104 L.Ed.2d 338 (1989), which raised the question "whether a complaint filed in forma pauperis which fails to state a claim under Federal Rule of Civil Procedure 12(b)(6) is automatically frivolous within the meaning of 28 U.S.C. § 1915(d) [1989] )." Underscoring the fact that there is some distance between the total lack of merit connoted by the word "frivolous" and a failure to state a claim, the Court concluded that the answer was no. The justices found it "evident that the failure-to-state-a-claim standard of Rule 12(b)(6) and the frivolousness standard of § 1915(d) were devised to serve distinctive goals, and that while the overlap between these two standards is considerable, it does not follow that a complaint which falls afoul of the former standard will invariably fall afoul of the latter." Id. at 326, 109 S.Ct. 1827. Nothing in Rule 12(b)(6), the Court said, "confines its sweep to claims of law which are obviously insupportable." Id. at 327, 109 S.Ct. 1827. This is because the functions of Rule 12(b)(6) and section 1915(d) are different. The latter statute permitted (then as now) judges to "pierce the veil of the complaint's factual allegations and dismiss those claims" whose factual contentions are so baseless that they are properly considered to be frivolous. Id.
The pre-1996 regime was still in place when the Supreme Court decided Denton v. Hernandez, 504 U.S. 25, 112 S.Ct. 1728, 118 L.Ed.2d 340 (1992). There the Court underscored that a dismissal for frivolousness under the IFP statute leaves the complainant free to pay the filing fee and try again. Id. at 34, 112 S.Ct. 1728. At the time, however, it was still impossible for a district court to dismiss an IFP request for failure to state a claim (or any other merits-based reason). Denton 's holding must therefore be understood as addressing only the grounds for dismissal that were possible at the time.
The number of such grounds expanded dramatically in 1996. Concerned with the ballooning number of IFP requests, particularly though not exclusively from prisoners, Congress decided to enhance the courts' screening powers. It did so in the Prison Litigation Reform Act, which added to section 1915 an expanded set of reasons that would support dismissal "at any time" (including during preliminary screening). Those reasons are as follows: (1) the allegation of poverty is untrue; (2) the action or appeal is frivolous or malicious; (3) the action fails to state a claim on which relief may be granted; or (4) the action seeks monetary relief against a defendant who is immune from such relief. 28 U.S.C. § 1915(e)(2) (2017). See Prison Litigation Reform Act of 1995, Pub. L. 104–134, 110 Stat. 1327 (1996) ; Hutchinson v. Spink , 126 F.3d 895, 899–901 (7th Cir. 1997) (discussing the 1996 amendments and anticipating some of the problems that could arise if a case is dismissed for failure to state a claim without the defendants' knowledge).
Under the current regime, at the time a person files a complaint in which she seeks permission to proceed IFP, she must also submit an affidavit setting forth all of her assets and demonstrating that she is unable to pay the required fees. The date on which that complaint is filed or properly mailed, if the plaintiff is a prisoner, is considered to be the date when the suit "commences" for purposes of Federal Rule of Civil Procedure 3. See, e.g. , Jones v. Bertrand , 171 F.3d 499, 500–02 (7th Cir. 1999) ; see generally 4 CHARLES ALAN WRIGHT, ARTHUR R. MILLER & ADAM N. STEINMAN, FEDERAL PRACTICE AND PROCEDURE § 1052 at 224–25 (4th ed. 2015). If the plaintiff ultimately is found not to be eligible for IFP status and then fails to pay the filing fee, the case is subject to dismissal on that basis, but the failure to pay does not mean that the case never started. See Farzana K. v. Ind. Dep't of Educ., 473 F.3d 703, 707 (7th Cir. 2007). C
We are now ready to turn to Coleman's case. We address three questions: whether a dismissal under section 1915(e)(2)(B)(ii) must be treated as a full-fledged dismissal on the merits; what it takes to signify consent; and whether a defendant who has not yet been served is one of the "parties" who must consent for purposes of section 636(c)(1).
The first of those questions addresses the Commission's argument that the meaning of the word "parties" in section 636(c)(1) depends in large part on whether a decision risks adversely affecting the putative litigant's rights. The Commission reasons that an unserved defendant should not be classified as one of the "parties" at the IFP screening stage, because screening does not meaningfully affect that defendant's rights. There is some force in this point, but not as much as Judge Posner's dissent and the Commission assume. Some questions asked at the screening stage do not implicate the merits (and thus the rights of both parties). But the question whether the case states a claim on which relief may be granted does.
This court reviews dismissals under 28 U.S.C. § 1915(e)(2)(B)(ii) for failure to state a claim under the same standard we use for dismissals under Federal Rule of Civil Procedure 12(b)(6). See, e.g., Arnett v. Webster , 658 F.3d 742, 751 (7th Cir. 2011) ; DeWalt v. Carter , 224 F.3d 607, 611 (7th Cir. 2000). Noting that the question is exactly the same under Rule 12(b)(6) and section 1915(e)(2)(B)(ii), we have said that "[w]e see no reason to treat dismissals under § 1915(e)(2)(B)(ii) differently" from their Rule 12(b)(6) counterparts. DeWalt , 224 F.3d at 611. This strongly indicates that the action under section 1915 is just as much an adjudication of the legal sufficiency of the complaint as a comparable decision under Rule 12(b)(6) would be.
It is telling that the courts of appeals have treated dismissals under section 1915(e)(2)(B)(ii) as final decisions for purposes of the "three-strikes" rule in section 1915(g). That part of the IFP statute says that "[i]n no event shall a prisoner bring a civil action or appeal a judgment in a civil action or proceeding under this section if the prisoner has, on 3 or more prior occasions, while incarcerated or detained in any facility, brought an action or appeal in a court of the United States that was dismissed on the grounds that it ... fails to state a claim upon which relief may be granted...." 28 U.S.C. § 1915(g).
In McLean v. United States, 566 F.3d 391, 396 (4th Cir. 2009), the Fourth Circuit had to decide "whether Congress intended an action or appeal ‘that was dismissed on the grounds that it ... fails to state a claim upon which relief may be granted’ to count as a strike under 28 U.S.C. § 1915(g) if that dismissal was specifically designated to be ‘without prejudice.’ " (ellipsis in original). It noted that "[w]hen Congress directly incorporates language with an established legal meaning into a statute, we may infer that Congress intended the language to take on its established meaning." Id. Unless the court directs otherwise, a dismissal for failure to state a claim is with prejudice and would count as a strike. The court held, however, that if the section 1915(e) dismissal was specified to be without prejudice, then it would not be an adjudication on the merits and would not count as a strike.
The Tenth Circuit came to a similar conclusion in Smith v. Veterans Admin. , 636 F.3d 1306 (10th Cir. 2011). There a district court had dismissed a prisoner's complaint on screening for failure to state a claim. Id. at 1312–13. (It reached that conclusion because it decided that the prisoner's action was barred by Heck v. Humphrey , 512 U.S. 477, 114 S.Ct. 2364, 129 L.Ed.2d 383 (1994).) The Tenth Circuit held that this dismissal counted as a strike for purposes of section 1915(g). Indeed, the Tenth Circuit would have gone beyond the Fourth by also counting as strikes dismissals without prejudice for failure to state a claim. Id. at 1313.
Further proof that dismissals for failure to state a claim under section 1915(e)(2)(B)(ii) are final decisions on the merits comes from the fact that they have res judicata effect. Thus, in Cieszkowska v. Gray Line New York , 295 F.3d 204 (2d Cir. 2002), the appellant challenged the dismissal for failure to state a claim of her IFP complaint, because it was barred by res judicata . She had filed her first case in January 2000. At that time, she sought to proceed IFP, but the district court dismissed her complaint pursuant to section 1915(e)(2)(B)(ii) for failure to state a claim under either the Fair Labor Standards Act or the Consumer Credit Protection Act. In November 2000, she returned to the district court with a new case against the same defendant; in her new action she charged unlawful discrimination on the basis of national origin. Once again she sought to proceed IFP, and once again, the district court rejected her application on the ground of failure to state a claim. This time, however, the court held that her action was barred by res judicata . The court of appeals affirmed, noting that "the factual predicates of plaintiff's allegations in the first and second complaints involve the same events concerning her employment, pay history, and termination.... Accordingly, the claims in her second in forma pauperis complaint are now barred by res judicata , and the district court properly dismissed her complaint under § 1915(e)." Id. at 206. See also Harmon v. Webster , 263 Fed.Appx. 844 (11th Cir. 2008) (nonprecedential) (holding that second complaint was barred by res judicata after first action was dismissed on IFP screening for failure to state a claim).
The fact that screening dismissals have res judicata effect undermines the argument presented in Judge Easterbrook's opinion dissenting from the decision not to hear this case en banc . He argues that the operative distinction is between persons bound by a judgment and others who are named in a complaint but not bound (for example, because they have not been served with process). That distinction would be compelling if it were not for the fact that claim preclusion attaches to a screening dismissal for failure to state a claim upon which relief can be granted. And if claim preclusion applies to a defendant who never knew about the case, it is hard to see how that person is not bound by the judgment. Claim preclusion has not followed the path taken by issue preclusion, for which the requirement of mutuality has been abandoned, see Blonder-Tongue Lab., Inc. v. Univ. of Ill. Found. , 402 U.S. 313, 91 S.Ct. 1434, 28 L.Ed.2d 788 (1971) (approving non-mutual defensive issue preclusion); Parklane Hosiery Co. v. Shore , 439 U.S. 322, 99 S.Ct. 645, 58 L.Ed.2d 552 (1979) (approving non-mutual offensive issue preclusion). In Federated Dep't Stores, Inc. v. Moitie , 452 U.S. 394, 101 S.Ct. 2424, 69 L.Ed.2d 103 (1981), the Supreme Court considered and rejected an argument to deviate on equitable grounds from the strict requirements of claim preclusion: a final judgment on the merits, between the same parties or privies, relating to the same claim. If then, as all have held, claim preclusion is proper for a screening dismissal for failure to state a claim, then both parties must be bound by that decision, not just the plaintiff. And if both parties are bound, then both must consent to the assignment of the case to the magistrate judge. This is because a dismissal at the screening stage for failure to state a claim (with the possible exception, inapplicable here, of a dismissal specified to be without prejudice) is a decision on the merits of the complaint. Not only will the plaintiff be precluded from filing the same complaint again; principles of res judicata will prevent her from filing another complaint that arises from the same underlying circumstances. We have found no instance in which a district court, acting through either an Article III judge or a magistrate judge, dismissed a complaint on screening for failure to state a claim and then permitted the same complaint to move forward upon the tender of the necessary filing fees.
The conclusion we draw is that a dismissal under section 1915(e)(2)(B)(ii) for failure to state a claim is covered by the language in the Magistrate Judges Act that prevents the district court from designating a magistrate judge to "hear and determine" a motion "to dismiss for failure to state a claim upon which relief can be granted." 28 U.S.C. § 636(b)(1)(A). That rule is not absolute, however. It yields when, as section 636(c)(1) puts it, the magistrate judge is acting with "the consent of the parties." Packed into that short phrase are two questions: what does it take to signify consent, and which parties (one, some, or all) have to consent?
Typically, both parties file written consents to the magistrate judge's jurisdiction and the case proceeds without a hitch. And as Judge Posner's dissent emphasizes, consent need not be explicit. No dispute there: consent may be implicit, as the Supreme Court recognized in Roell v. Withrow , 538 U.S. 580, 586–91, 123 S.Ct. 1696, 155 L.Ed.2d 775 (2003). But even implicit consent requires some action from the party whose consent must be found. In Roell , for example, "Roell and Garibay clearly implied their consent by their decision to appear before the Magistrate Judge, without expressing any reservation, after being notified of their right to refuse and after being told that she intended to exercise case-dispositive authority." Id. at 586, 123 S.Ct. 1696 (internal quotation marks omitted). It is one thing to infer consent from concrete actions—such as appearing and participating in a case after being told about the consequences—and quite another to infer consent from a total lack of action. This pushes us from implicit consent to assumed or hypothetical consent, and it is at most the latter that exists in our case. The Commission did not even know that it had been sued, and so it naturally took no step from which its consent to proceed before a magistrate judge for a dispositive ruling could be inferred. This is not enough to qualify as implicit consent.
With the idea of presumed consent out of the way, the question remains whether the statute permits the consent of only one party. We think not. To begin with, that is not what the statute says; it speaks instead of the consent of the "parties," plural. The Commission sees no problem in the use of the plural: at the time of screening, it notes, only one party has come before the court. It should be enough, the Commission says, to obtain only the consent of the party that is actually present. This would require us to treat the unserved defendant as a nonparty to the case for this purpose.
Such a step would ignore the general use of the term "parties" throughout the rules of civil procedure. At the time of filing any civil complaint, only one party is ever before the court. Yet no one thinks, for example, that the plaintiff alone has the right to consent to the district court's adjudication of the plaintiff's motion for summary judgment, which can be filed before the defendant is served, and indeed at "at any time until 30 days after the close of all discovery." FED. R. CIV. P.56(b) ; see 28 U.S.C. § 636(b)(1)(A), (c)(1) (prohibiting magistrate judges from deciding summary judgment motions without consent). Similarly, no one thinks that the plaintiff alone can consent to a magistrate judge's decision on a motion to certify a class action, filed before the defendant has been served and brought before the court. See FED. R. CIV. P.23(c) ; see 28 U.S.C. § 636(b)(1)(A), (c)(1).
The fact that it is the plaintiff who commences the suit does not mean that the other parties named in the complaint do not count as "parties" prior to service of process. And our case differs in one critical respect from the hypotheticals that concern Judge Easterbrook. Post, at 478–80. All of his examples involve a case that has at least one plaintiff and at least one defendant. And in all of his examples, at least one plaintiff and one defendant have consented to the authority of the magistrate judge. That is the crucial element missing here: in our case, only one side of the "v" has consented to the magistrate judge, and under the statute, that is not enough.
Nothing in the decisions in Murphy Bros., Inc. v. Michetti Pipe Stringing, Inc., 526 U.S. 344, 119 S.Ct. 1322, 143 L.Ed.2d 448 (1999), and Cambridge Holdings Grp., Inc. v. Fed. Ins. Co. , 489 F.3d 1356, 1361 (D.C. Cir. 2007), undermines that conclusion. Although both deal with the status of an unserved defendant, the questions arose in a very different setting, and we know from Devlin v. Scardelletti , 536 U.S. 1, 122 S.Ct. 2005, 153 L.Ed.2d 27 (2002), that context matters. In Scardelletti , the question was whether an unnamed member of a class certified under Rule 23 was entitled to appeal from the approval of a settlement to which he objected, when he had not first intervened at the district court level. Id. at 7, 122 S.Ct. 2005. Rather than adopting a hard-and-fast rule for party status, the Court essentially said "it depends." It catalogued various ways in which unnamed class members were "parties" (e.g. , for purposes of tolling the statute of limitations and for purposes of the amount-in-controversy rule) and ways in which they were not "parties" (e.g. , for purposes of the complete diversity requirement). It concluded with the observation that "[t]he label ‘party’ does not indicate an absolute characteristic, but rather a conclusion about the applicability of various procedural rules that may differ based on context ." Id. at 10, 122 S.Ct. 2005 (emphasis added).
The context in which the "party" question arose in Murphy Bros. and Cambridge Holdings is nothing like the one now before us. The question in Murphy Bros. dealt with the timeliness of a defendant's removal of a case from state to federal court. Section 1446(b) gives a defendant 30 days "after the receipt by the defendant, through service or otherwise, of a copy of the initial pleading...." The defendant, Michetti, had filed a contract case in an Alabama state court on January 26, 1996; three days later, before serving Murphy, Michetti faxed a courtesy copy of the complaint to Murphy. Murphy did not receive formal process, however, until February 12, 1996. On March 13, 1996—30 days after service, but 44 days after receiving the courtesy copy—Murphy removed the case to the federal court. The Supreme Court held that the removal was timely, because the 30 days began only at the time of formal service.
In reaching this decision, the Court emphasized the importance of service of process. In the absence of service (or waiver), "a court ordinarily may not exercise power over a party the complaint names as defendant." 526 U.S. at 350, 119 S.Ct. 1322. It reasoned that "one becomes a party officially, and is required to take action in that capacity, only upon service of a summons or other authority-asserting measure...." Id. Nowhere did it imply that the named defendant is not a "party" entitled to procedural protections. To the contrary, it is precisely because the unserved defendant is a party that it is entitled to the protection of formal service of process.
The interests with which the Court was concerned in Murphy Bros. related to the operation of the removal statutes. The lesson we draw is that something as important as the choice between a state court and a federal court, or between an Article I and an Article III judge, cannot be resolved against a party without bringing the party into the case through formal service of process.
There is also no conflict between our approach here and Cambridge Hol dings . That case dealt with Federal Rule of Civil Procedure 54(b) ; the question was what it means to resolve all claims of all parties. On July 12, 2004, the district court dismissed Counts 1 through 4 of Cambridge's complaint against two parties for failure to state a claim. Count 5 remained; it was a request for an injunction against a third party. Nearly a year went by before anything else happened in the case, but on June 20, 2005, the district court dismissed Count 5 for failure to prosecute and entered an order dismissing the case as a whole. Cambridge filed a notice of appeal on July 19, 2005—within 30 days of the June 20 order, but more than a year after the order dismissing Counts 1 through 4.
As the D.C. Circuit recognized, ordinarily the question whether the appeal had been filed within the required 30 days would have been an easy one to answer. Rule 54(b) says that an order disposing of less than all claims of all parties does not qualify as a final judgment unless the district court enters a Rule 54(b) certification that there is no just reason for delay and that the judgment is indeed final for the party and claim so certified. The quirk in the case once again had to do with service of process. It appeared that Count 5's third party had never been served. The court of appeals thus had to decide whether a failure to resolve a claim against an unserved party renders a judgment nonfinal for purposes of Rule 54(b), or as the court put it, "whether a defendant that has never been served is a ‘party’ for purposes of Rule 54(b)." 489 F.3d at 1360. The discussion that follows is closely tethered to Rule 54(b) and its purposes. The D.C. Circuit noted that eight circuits had concluded that the unserved party did not defeat finality for this purpose. Id. at 1360 n.1. This court had held much the same thing in Manley v. City of Chicago , 236 F.3d 392 (7th Cir. 2001), which said that "the presence of an unserved defendant does not defeat finality when an attempt by the plaintiff to serve the complaint on the unserved defendant would be untimely under Fed. R. Civ. P. 4(m) and any new complaint against the unserved defendant would be barred by the statute of limitations." Id. at 395.
We see no reason to transplant the finality rules for removal or Rule 54(b) to the setting of IFP screening at the outset of a lawsuit, particularly given the fact that in these other settings the rules have operated to protect the interests of the absentee party, not to act as if the party does not exist. These cases do not compel a finding that the unserved defendant in our case is a nonparty for purposes of 28 U.S.C. § 636(c).
Other aspects of the relevant statutes and rules reinforce our conclusion. Section 636(c)(1) of the Magistrate Judges Act twice refers to the "consent of the parties "; in no place does it call for the consent of a "party." We give statutes their plain meaning, and we see no ambiguity in the meaning of the term "parties": It refers to the names on both sides of the "v," without regard to events such as service of process. Federal Rule of Civil Procedure 4, which governs service, would make no sense if the rule were otherwise. Rule 4 begins by stating that a summons must "name the court and the parties." See FED. R. CIV. P.4(a)(1)(A). The rule refers throughout to the yet-to-be-served opponent of the plaintiff as the "defendant"—a word used only to connote a party to a lawsuit. Rule 4(i) distinguishes parties from nonparties, for purposes of actions that "challenge[ ] an order of a nonparty agency or officer of the United States." See FED. R. CIV. P.4(i)(1)(C). Rule 4(k) talks about an unserved defendant "who is a party joined under Rule 14 or 19...." FED. R. CIV. P.4(k)(1)(B).
The Commission attempts to avoid the use of the plural in the Magistrate Judges Act by looking at the Dictionary Act, which says that "[i]n determining the meaning of any Act of Congress, unless the context indicates otherwise—... words importing the plural include the singular." 1 U.S.C. § 1, ¶ 2. The last part of that excerpt might help, were it not for the words "unless the context indicates otherwise." Here the context, which includes both the remainder of section 636(c) and the constitutional rulings of the Supreme Court, does indicate otherwise. Rowland v. Cal. Men's Colony, Unit II Men's Advisory Council , 506 U.S. 194, 199, 113 S.Ct. 716, 121 L.Ed.2d 656 (1993).
The remainder of section 636 supports a reading that distinguishes between the use of the plural and the singular in this setting. When a law employs a specific term at one point, and omits it at another, we assume that "Congress intended a difference in meaning." Loughrin v. United States , –––U.S. ––––, 134 S.Ct. 2384, 2390, 189 L.Ed.2d 411 (2014). That rule applies here. Section 636(c)(1) refers to "parties," but other parts of the same statute use the singular. Thus, section 636(b)(1) says "[w]ithin fourteen days after being served with a copy [of a report and recommendation], any party may serve and file written objections...." (emphasis added). The same subsection directs the magistrate judge to mail a copy of his proposed findings and recommendations "to all parties ." (emphasis added). For purposes of appellate review, the statute allows "an aggrieved party" to appeal directly to the court of appeals. 28 U.S.C. § 636(c)(3). It also authorizes the district court, either on its own or "under extraordinary circumstances shown by any party , [to] vacate a reference to a magistrate judge." Id. § 636(c)(4) (emphasis added). The point is clear: in this particular statute, Congress used the singular when it meant one party, and it used the plural when it meant all parties.
If we had any doubt about the proper understanding of the statutory text, it would be laid to rest by the constitutional problem that would arise if we were to hold that the consent of one party alone was enough to permit an Article I judge to resolve the case on the merits. As the decisions in Wellness , Marathon , Stern , and Nguyen demonstrate, institutional concerns beyond these two parties give final decisionmaking authority only to Article III judges, unless all parties consent to an alternative. See Commodity Futures Trading Comm'n v. Schor , 478 U.S. 833, 848–51, 106 S.Ct. 3245, 92 L.Ed.2d 675 (1986) ; Brook, Weiner, Sered, Kreger & Weinberg v. Coreq, Inc. , 53 F.3d 851, 852 (7th Cir. 1995). III
This court, unfortunately, has not been consistent in its approach to the issue we have been discussing. Hains , supra , 131 F.3d 1248, concerned 28 U.S.C. § 1915A, a law under which the district courts review all complaints in civil actions in which a prisoner seeks relief from a governmental entity or officer. The Northern District of Illinois adopted a procedure under which magistrate judges conducted that screening and ruled on motions to proceed IFP, if the prisoner signed a limited consent form. Id. at 149. The prisoner complained after the magistrate judge dismissed his action without allowing him the opportunity to file objections and to obtain a de novo ruling from the district court. We rejected his challenge, reasoning that the prisoner's consent "relinquished his right to have an Article III judge rule upon his request to proceed in forma pauperis ." Id . This was true even though the defendant had not been served nor had it otherwise consented to the magistrate's authority; it was enough that the plaintiff had agreed. Id. at 149 & n.2.
Hains, however, did not cite Geaney , supra , 776 F.2d 140, in which a magistrate judge dismissed a prisoner's case under Federal Rule of Civil Procedure 41(a)(1)(i). Our decision in Geaney was unambiguous: "We hold that this court lacks jurisdiction to review the magistrate's dismissal, whether characterized as voluntary or involuntary, and we dismiss for lack of appellate jurisdiction." Id. at 141. Our reasoning was consistent with the discussion that appears above. On the assumption that the dismissal was involuntary, we held that the magistrate judge lacked authority to enter a final judgment, because the defendants never consented to this procedure under section 636(c)(1). We stated that "[a] magistrate may enter a final judgment only after a proper referral by the district court and upon consent by both parties." Id. at 142 (citing § 636(c)(1) ) (emphasis added). We justified this ruling in part by the "constitutional overtones" of the question. Id. See also Brook , 53 F.3d at 852 ("Each litigant has a right to an Article III judge, so consent must be unanimous.").
We see no principled way of reconciling Geaney and Hains , even though Hains dealt with the prisoner-specific section 1915A, whereas Coleman's case was dismissed under section 1915(e). The relevant language and structure of these two statutes are identical. See Brown v. Bargery , 207 F.3d 863, 866 n.3 (6th Cir. 2000).
As we recognized earlier, in certain situations there is some flexibility in identifying who is, and who is not, a party to a case. Unnamed class members, for instance, are not full-fledged parties, and so they need not be served with process, their citizenship does not matter for diversity purposes, and they can obtain a direct seat at the table only by intervening. A magistrate judge can therefore rule on a motion to intervene filed by an unnamed class member, both because that person is not yet a party, and such a motion in any event is nondispositive. Williams v. Gen. Elec. Capital Auto Lease, Inc. , 159 F.3d 266, 269 (7th Cir. 1998) ; People Who Care v. Rockford Bd. of Educ. , 171 F.3d 1083, 1089 (7th Cir. 1999). But this is not a class action, and it is uncontested that only Coleman consented to the magistrate judge.
In the end, Hains stands alone in this circuit. Our consistent emphasis on the importance of consent strongly supports Geaney 's analysis. See also Stevo v. Frasor , 662 F.3d 880, 883–86 (7th Cir. 2011) ("A district judge may assign a magistrate judge to hear a civil case and render final judgment, provided that all parties consent voluntarily."). We conclude that Geaney , Brook , and Stevo are more consistent with the language of the statute and better respect the constitutional line between Article III judges and other adjudicators. We thus overrule Hains. Because this resolves an internal circuit conflict and overrules one case, we have circulated this opinion to all judges in regular active service. See Cir. R. 40(e). A majority of the judges did not wish to hear this case en banc . Judge Easterbrook, joined by Judge Sykes, voted in favor of an en banc hearing. Judge Posner dissents on the merits.
As we stressed earlier, this does not mean that the work of the magistrate judge in this type of case is wasted. The role of the magistrate judges in conducting screening of prisoner and in forma pauperis actions is of great assistance to the district court. But unless all parties to the action have consented to the magistrate judge's authority to resolve the case finally, the role of the magistrate judge must parallel that of the bankruptcy judges after Stern . Rather than entering final judgments, they must "issue proposed findings of fact and conclusions of law to be reviewed de novo by the district court." Exec. Benefits Ins. Agency v. Arkison , ––– U.S. ––––, 134 S.Ct. 2165, 2168, 189 L.Ed.2d 83 (2014). We realize that this adds one extra step, but it is not a particularly burden-some one, and it does not mean that parties in the Commission's position must be served before the case can be resolved. It just means that the district judge must enter any post-screening orders that dispose of the entire case. Fears that the district court judges will drown beneath a deluge of IFP applications are overblown. The Stern dissent expressed similar fears, but this has not come to pass. Stern , 564 U.S. at 520–21, 131 S.Ct. 2594 (Breyer, J., dissenting). The magistrate judges' screening is among those "additional duties as are not inconsistent with the Constitution and laws of the United States." 28 U.S.C. § 636(b)(3).
IV
We have before us a decision from the district court that lay beyond that court's jurisdiction, because the judicial officer who entered it was not properly authorized to do so. The proper remedy, we conclude, is to vacate the judgment of the district court and remand for further proceedings. In keeping with such cases as Stern and Wellness, we note that on remand the district court will be free to take the magistrate judge's decision as a report and recommendation and proceed accordingly.
We therefore VACATE the judgment of the district court and REMAND for proceedings consistent with this opinion.
Posner, Circuit Judge, dissenting.
The plaintiff, proceeding pro se and seeking in forma pauperis status (that is, seeking permission to commence a civil suit without prepaying the court's filing fee), filed suit against a Wisconsin agency complaining of racial discrimination in employment, and consented to a magistrate judge's exercising jurisdiction over the suit. When a plaintiff who is not a prisoner (and the plaintiff in this case is not) seeks in forma pauperis status, normally the district judge or magistrate judge assigned to the case will conduct an initial screening of the complaint to determine whether the plaintiff really is too poor to pay the fees, or whether, though he is, his suit is frivolous, malicious, fails to state a claim, or seeks monetary relief against an immune defendant. 28 U.S.C. § 1915(e)(2). The plaintiff in this case did not serve the defendant before the initial screening, consistent with the guidance document of the Eastern District of Wisconsin for pro se litigants which advises that serving the defendant is the next step after screening. The defendant, whether or not aware of the suit (as it may have been despite not being served), did not express consent to a magistrate judge's exercising jurisdiction over it. But because the plaintiff had consented to a magistrate judge's handling the case, that judge screened the complaint, and concluding from the screening that it failed to state a claim entered final judgment in favor of the defendant, dismissing the suit.
The plaintiff appeals on the ground that 28 U.S.C. § 636(c)(1) requires "consent of the parties" for the entry of a final judgment by a magistrate judge, and in this case only one party—the plaintiff—consented. Agreeing with the plaintiff that "consent of the parties" means consent by both plaintiff and defendant, the majority opinion orders the judgment entered by the magistrate judge vacated on the ground that because the statute uses the singular "party" in other contexts, the plural "parties" must refer to all the parties to a case and therefore a magistrate judge can't enter a final judgment of dismissal unless the defendant consents—consents explicitly , even to a judgment in its favor. Yet in such a case, which is this case, the defendant's consent can of course be presumed, the magistrate judge having thrown out the case against it, as the judge was entitled to do because the plaintiff had consented to his adjudicating the case.
The majority opinion bases its insistence that the defendant's consent to dismissal of the case must to be effective be explicit on concern that a "constitutional problem ... would arise" if the consent of the plaintiff alone were sufficient for a magistrate judge to screen a complaint. The "constitutional problem" alleged is that if the statute required only implicit consent by the defendant this would violate the provision in Article III, section 1 of the Constitution that "the judicial Power of the United States, shall be vested in one supreme Court, and in such inferior Courts as the Congress may from time to time ordain and establish," with the "Courts" to be composed of judges who "shall hold their offices during good behaviour" and receive salaries that may not be diminished. Magistrate judges are not such judges—that is, are not Article III judges.
Yet we know that litigants may consent to adjudication by a magistrate judge and that such consent need not be express. Wellness Int'l Network Ltd. v. Sharif , ––– U.S. ––––, 135 S.Ct. 1932, 1947–48, 191 L.Ed.2d 911 (2015) ; Roell v. Withrow , 538 U.S. 580, 586–91, 123 S.Ct. 1696, 155 L.Ed.2d 775 (2003). The plaintiff in this case consented to adjudication by a magistrate judge—expressly—and the defendant consented implicitly, because the adjudication ended the litigation against the defendant, leaving no possible reason for the defendant to try to revive the litigation. Whoever heard of a defendant saying to a district judge or magistrate judge "I know you've dismissed the plaintiff's case against me, but would you please reinstate it so that I can file a statement agreeing to the dismissal?"
A bigger waste of time than sending this case back to the district court is thus hard to imagine, as well as creating tension with the Supreme Court's rulings in the Wellness and Roell cases, cited above, that a defendant's consent to a magistrate judge's adjudication of a suit against him need not be express. To repeat, the defendant's consent to the magistrate judge's exercise of jurisdiction in this case up to and concluding final judgment can be taken for granted because the defendant has no interest in having the case continue and could therefore have no quarrel with the magistrate judge's dismissing it, at least where as in this case the dismissal was a final judgment, precluding a refiling of the complaint. At this stage of the litigation, the best outcome for the defendant is for the case simply to be dismissed, and the worst outcome would be to allow the case to proceed.
Yet since the majority insists on a remand, though with no justification that I can see, we can expect that on remand the magistrate judge will reissue his order, retitled as a "proposed order," and a district judge will sign off on it and the plaintiff will again appeal—placing the case again before us with nothing for us to do but affirm. Such a waste of time can be avoided—without any negative consequences for anyone—simply by interpreting the requirement in 28 U.S.C. § 636(c)(1) of "consent of the parties" as satisfied by the plaintiff's consenting to a magistrate judge's exercising jurisdiction at the screening stage and the defendant's not objecting, in which event the defendant's consent to the procedure can be presumed at least when, as in this case, the suit is dismissed; for that outcome will delight the defendant.
If the suit is not dismissed but instead allowed to proceed, the defendant will be served and have an opportunity to defend. Cf. Hains v. Washington , 131 F.3d 1248, 1249 and n. 2 (7th Cir. 1997) (per curiam). If the defendant doesn't consent explicitly or implicitly to a magistrate judge's adjudicating the case, the "consent of the parties," explicit or implicit, to such adjudication will not have been obtained and the case will proceed in the ordinary way, with any final decision in the district court being made by a district judge (or a jury) rather than by a magistrate judge.
The majority opinion argues that our Hains decision, cited above as authority for the approach I'm advocating, can't be reconciled with an earlier decision of our court, Geaney v. Carlson , 776 F.2d 140 (7th Cir. 1985), which according to the majority opinion in this case held that a magistrate judge could enter a final judgment dismissing a case only if both parties consented. But unlike the present case, Geaney did not involve initial screening and does not say that the defendant's consent to a magistrate judge's dismissal of a suit against him during an initial screening can never be presumed—and in a case such as this the presumption is inescapable. For why, to return to what seems to me the only sensible interpretation of the consent requirement, would a defendant ever refuse to consent to the dismissal of a suit against him?
In summary we should affirm, not order a futile remand.
EASTERBROOK , Circuit Judge , with whom SYKES , Circuit Judge , joins, dissenting from the denial of rehearing en banc. I agree with the panel that Hains v. Washington , 131 F.3d 1248, 1249 & n.2 (7th Cir. 1997), conflicts with Geaney v. Carlson , 776 F.2d 140, 142 (7th Cir. 1985), and that one or the other must be overruled. The panel overrules Hains , but I think that Geaney should be the one to go. We should align ourselves with Neals v. Norwood , 59 F.3d 530, 532 (5th Cir. 1995), which holds that an un-served defendant is not among the "parties" for the purpose of 28 U.S.C. § 636(c), rather than Henry v. Tri-Services, Inc. , 33 F.3d 931, 933 (8th Cir. 1994), which holds the opposite.
As the panel sees matters, "parties" in § 636(c) means all persons named in the complaint. That's linguistically possible, though § 636(c) works just fine if "parties" means "every party bound by the judgment" rather than "every party named in the complaint." I distinguish between persons bound by a judgment (in the sense that the person can be saddled with an adverse decision) and everyone named in a complaint. Un-served defendants cannot be adversely affected by a judgment. McDonald v. Mabee , 243 U.S. 90, 92, 37 S.Ct. 343, 61 L.Ed. 608 (1917).
The panel opens with a reminder that magistrate judges lack the tenure and salary protections of Article III (op. 465–66) and later uses this as a doubt resolver (op. 473–74). Article III means that only litigants who consent to decision by an Article I officer (or for that matter a private arbitrator) can be denied the benefits of an Article III judge. But Coleman did so consent, and he is the only person adversely affected by the outcome. See Omni Capital International, Ltd. v. Rudolf Wolff & Co. , 484 U.S. 97, 104, 108 S.Ct. 404, 98 L.Ed.2d 415 (1987). There's no constitutional problem with binding consenting plaintiffs to adverse decisions by magistrate judges. Article III does not place a thumb on the scale against the possibility that "parties" in § 636(c) means "all litigants bound by an adverse judgment."
A different principle does create a helpful presumption: the principle that someone named as a defendant does not become a party until served with process. See Murphy Brothers, Inc. v. Michetti Pipe Stringing, Inc ., 526 U.S. 344, 350, 119 S.Ct. 1322, 143 L.Ed.2d 448 (1999). Murphy Brothers arose from a statute providing that "defendants" have only 30 days to remove cases filed in state court. 28 U.S.C. § 1446(b). That statute starts the time from each defendant's receipt of a copy of the complaint "through service or otherwise". One defendant got a faxed copy and did not remove within 30 days of its receipt. The Justices concluded that this did not matter because only parties have duties in the litigation, and one does not become a party until served with process. This approach—that the number of parties grows as defendants are served—is incompatible with the panel's approach to § 636(c), under which the number and identity of "parties" is fixed the moment the complaint is lodged in the district court.
By the standard of Murphy Brothers , there was only one "party" when the magistrate judge dismissed Coleman's complaint. He consented under § 636(c). It follows that the magistrate judge's decision was authorized by statute.
The panel's majority replies that Murphy Brothers interprets § 1446(b) rather than § 636(c) and therefore does not compel us to read the latter statute any particular way. That's true enough. Yet Murphy Brothers did not suggest that its understanding of the role of service in making a defendant a party depended on anything peculiar to § 1446(b). Other circuits apply Murphy Brothers beyond § 1446(b). For example, they have held that un-served defendants are not parties for purposes of Fed. R. Civ. P. 54(b), and, by extension, 28 U.S.C. § 1291. See Cambridge Holdings Group, Inc. v. Federal Insurance Co ., 489 F.3d 1356, 1361 & n.1 (D.C. Cir. 2007) (collecting cases). The panel brushes these decisions aside on the ground that they, too, do not concern § 636(c), but this treats Murphy Brothers as a sport rather than as the instantiation of a principle. Murphy Brothers reads like a principle applicable to federal civil procedure, not like dogma good for one statute only.
For the same reason I find the majority's opinion unpersuasive, I do not adopt the dissent's view. It is not sound to say that un-served defendants have consented—implicitly, constructively, or in any other way. Decision by a magistrate judge depends on consent. That consent may be inferred from action (such as litigating before a magistrate judge without protest) but must be real, and treating non-action of non-parties as consent dispenses with the need for consent of any kind. It is the sort of make-believe that gives the phrase "legal fiction" its bad odor.
Language in both the majority and the dissent suggests that this appeal is about whether magistrate judges can dismiss prisoners' complaints under § 1915(e) before defendants have been served. It is indeed about that, but the decision has broader scope. The panel holds that a magistrate judge cannot preside unless all defendants named in the complaint, plus all plaintiffs, have consented in writing.
Suits may linger in a district court with one or more defendants un-served—indeed, with one or more defendants identified only by pseudonym (for example, a police officer who is said to have used excessive force in making an arrest). Sometimes a defendant's identity is known but he can't be found to be served. For example, prisons don't want to reveal the current addresses of former guards, and defendants who reside abroad can be especially hard to serve.
The suit A v. B & C could go on for years before C is finally dismissed or discovered to be nonexistent; why can't A and B agree in the meantime that their dispute will be resolved by a magistrate judge? It would be weird if, after final decision, either A or B could have the judgment upset on appeal—despite having consented under § 636(c) —just because C was never found or served. There's no constitutional impediment to consent by A and B; both waived their right to an Article III judge. But under the majority's approach every suit with an un-served or unknown defendant must be handled by a district judge even if all of the real litigants prefer a magistrate judge. That is not sound judicial administration. Murphy Brothers offers a way to solve this problem.
The panel's majority dismisses these possibilities (op. 471) by suggesting that "parties" in § 636(c) may mean only the plaintiff plus at least one defendant. That comes out of the blue. The bulk of the majority's Part II.C is devoted to contending that "parties" means everyone named in a complaint. That is a plausible (though I think incorrect) reading. But the word "parties" cannot possibly mean "plaintiffs plus one defendant, but not other defendants." That ad hoc treatment is disconnected from the opinion's rationale.
If "parties" means "all named litigants," then problematic examples in multi-party suits are legion. (1) C moves to dismiss for lack of personal jurisdiction, and a district judge grants that motion; A and B then consent to decision by a magistrate judge. (2) C moves to dismiss under Fed. R. Civ. P. 12(c) because of a person-specific defense such as release, and a district judge grants that motion; A and B then consent to decision by a magistrate judge. (3) Same as example (2) except that C moves to dismiss for improper joinder, because A v. C has nothing to do with A's claim against B, and a district judge grants that motion; A and B then consent to decision by a magistrate judge. The panel's decision means that all of these consents by A and B are invalid, because C, having been named in the complaint, is a "party" who did not consent. Yet I cannot see why A and B should be forbidden to advance their own interests by consenting to decision by a magistrate judge, who may be able to give the case immediate attention that a district judge cannot provide.
By using these examples I do not suggest that practical concerns justify overriding statutory text. But nothing in the text of § 636(c) implies a departure from the principle of Murphy Brothers , under which "party" status depends on service. And we know from Devlin v. Scardelletti , 536 U.S. 1, 122 S.Ct. 2005, 153 L.Ed.2d 27 (2002), that people can be "parties" for some purposes while they aren't for others. So the fact that defendants are parties for the purpose of filling the blanks in a summons (which the panel discusses at op. 473) does not tell us that they are parties under § 636(c) too.
All statutory language must be read against background norms. One norm is that ambiguous statutes should not be understood to create mischief—losses to some litigants with no offsetting gains to others (or to society as a whole). Depriving litigants of their preferred forum produces losses with no offsetting gains. Murphy Brothers provides another norm. When read in light of those norms, § 636(c) requires no more than the consent of the party or parties who will be bound by an adverse judgment.