holding that “[r]oughly speaking, § 2254 [is] the exclusive vehicle for prisoners in custody pursuant to a state court judgment who wish to challenge anything affecting that custody, because it makes clear that bringing an action under § 2241 will not permit the prisoner to evade the requirements of § 2254”
How later courts described this case
- holding that “[r]oughly speaking, § 2254 [is] the exclusive vehicle for prisoners in custody pursuant to a state court judgment who wish to challenge anything affecting that custody, because it makes clear that bringing an action under § 2241 will not permit the prisoner to evade the requirements of § 2254”
- holding that "[rjoughly speaking, § 2254[is] the exclusive vehicle for prisoners in custody pursuant to a state court judgment who wish to challenge anything affecting that custody, because it makes clear that bringing an action under § 2241 will not permit the prisoner to evade the requirements of § 2254”
- holding that "if a case is properly filed as an action under 28 U.S.C. §§ 2241 , 2254, or 2255, it is not a 'civil action' to which the PLRA applies."
- holding that state prisoners challenging prison disciplinary proceedings need not obtain a certificate of appealability because the challenged detention does not “arise[ ] out of process issued by a State court.”
Written by the judges who cited it.
The opinion
EASTERBROOK, Circuit Judge,
with whom POSNER, Chief Judge, and MANION, Circuit Judge, join,
dissenting from the denial of rehearing en banc.
The panel announces two important decisions: that collateral attacks on prison discipline are not “civil actions” for the purpose of the Prison Litigation Reform Act (Part iii); and that state prisoners who want to appeal adverse decisions in cases about good-time credits do not need certificates of appealability, despite 28 U.S.C. § 2253 (c), part of the Antiterrorism and Effective Death Penalty Act (Part iv). We should consider these issues en banc, for several reasons.
• The panel has established ground rules for a substantial body of prisoners’ suits, easily more than a hundred annually in this circuit.
• The issues are important qualitatively as well as quantitatively. When must prisoners pay to litigate? May decisions rejecting frivolous collateral attacks be appealed as of right?
• Part in of the panel’s opinion overrules Part iii of Newlin v. Helman, 123 F.3d 429, 437-38 (7th Cir.1997), while Part iv creates a conflict among the circuits.
• The panel’s opinion is internally contradictory. (i) Part ii.b holds that collateral attacks about good-time credits concern “a person in custody pursuant to the judgment of a State court” (§ 2254(a)), yet Part rv holds that these same collateral attacks do not concern a “proceeding in which the detention complained of arises out of process issued by a State court” (§ 2253(c)(1)(A)). (ii) Part iii is justified in large measure by the desirability of eliminating a conflict among the circuits, yet Part iv creates a new conflict, (iii) The panel asserts in Part iii that it is too difficult to distinguish between collateral attacks on convictions and collateral attacks on prison discipline when assessing fees, yet Part iv draws exactly this line for purposes of certificates of appealability. (iv) Part iv invokes a plain-meaning approach to interpretation, while Part in disdains the statutory text.
• Both of the panel’s principal holdings are substantively questionable, for reasons that I now set out.
Part III holds that the plRa’s fee-collection mechanism (indeed, the whole puia) does not apply to any application for collateral relief. The critical text is 28 U.S.C. § 1915 (b)(1), which says that “if a prisoner brings a civil action or files an appeal in forma pauperis” then part of the filing fee must be prepaid, and the rest must be collected over time from prison trust accounts. Other features of the PLRA also apply only to “civil actions.” Is a petition for a writ of habeas corpus a “civil action”? We have been told authoritatively that the answer is “yes.” Browder v. Director, Department of Corrections, 434 U.S. 257, 269 , 98 S.Ct. 556 , 54 L.Ed.2d 521 (1978); Unit *641 ed States v. Morgan, 346 U.S. 502 , 505 & n. 4, 74 S.Ct. 247 , 98 L.Ed. 248 (1954). Is an appeal from the denial of a petition for collateral relief an “appeal in forma pau-peris”? Surely yes, when the prisoner seeks to proceed in forma pauperis. What could an appeal in forma pauperis be, other than “an appeal in forma pauperis”?
Collateral attacks are civil actions. Browder holds this (the Court wrote that “[i]t is well settled that habeas corpus is a civil proceeding”), and if that were not enough 28 U.S.C. § 1914 (a) demonstrates it: “The clerk of each district court shall require the parties instituting any civil action, suit or proceeding in such court, whether by original process, removal or otherwise, to pay a filing fee of $150, except that on application for a writ of habe-as corpus the filing fee shall be $5.” Prison discipline and ensuing collateral attacks are not criminal prosecutions, so they must be “civil actions.” See Wolff v. McDonnell, 418 U.S. 539, 556 , 94 S.Ct. 2963 , 41 L.Ed.2d 935 (1974) (remarking, in a case involving good-time credits, that “[plrison disciplinary proceedings are not part of a criminal prosecution”); Baxter v. Palmigiano, 425 U.S. 308 , 96 S.Ct. 1551 , 47 L.Ed.2d 810 (1976). If we defíne the phrase “civil action” in § 1915(b) the way Congress did in § 1914(a), a statute treating collateral attacks as a subset of all civil actions, then application of the plra follows directly.
Lawsuits arising out of prison discipline are a principal target of the PLRA. Should it make any difference for this purpose whether a warden revokes 30 days of good-time credits or puts the prisoner in segregation for six months (which also may prevent the prisoner from earning new good-time credits)? Not under the language of § 1915(b), and not for functional purposes either, yet under Part m of the panel’s opinion the warden’s choice of sanction determines whether the prisoner must prepay partial filing fees and whether prior frivolous suits require prepayment of the full filing fees, as § 1915(g) requires. (Because § 1915(g) treats any frivolous proceeding as a “strike,” I take it that even under the panel’s approach three frivolous collateral attacks would require prepayment of the filing fees in any future “civil action or appeal [of] a judgment in a civil action”.)
Part in of the panel’s opinion does not take the language of § 1915(b) seriously; it mentions the phrase “civil action” only in passing and the phrase “appeal in forma pauperis” not at all. The panel conceives the issue as whether a collateral attack arising out of the deprivation of good-time credits differs, for purposes of § 2254, from a collateral attack arising out of a conviction or sentence. Relying on cases such as Preiser and Balisok, the panel answers “no.” I agree with that answer— but it is an answer to an irrelevant question, for we must decipher the meaning of “civil action” in § 1915(b) rather than the nature of actions under § 2254. Nothing in Preiser or Balisok concerns the meaning of “civil action ... in forma pauperis” or “appeal in forma pauperis” in § 1915(b); we have to define these ourselves rather than attribute a definition to cases concerning other issues.
To understand why Part m of Newlin held that collateral attacks on prison discipline are “civil actions” under § 1915(b), one must begin with Martin v. United States, 96 F.3d 853 (7th Cir.1996). Martin allowed that petitions for habeas corpus “are technically civil proceedings and so come within the literal scope of the Act.” 96 F.3d at 855 . But Martin also observed that changes made by the aedpa contemporaneously with the PLRA subject applications for writs of habeas corpus to a special regimen. For example, under the PLRA three frivolous suits block further civil filings and appeals in forma pauperis, § 1915(g), while the aedpa limits second or successive filings via § 2244(b). As the panel wrote in Martin, 96 F.3d at 856 , “by drastically curtailing the filing of second or successive applications of [sic] habeas corpus, the antiterrorism law addressed in the *642 context of habeas corpus the same concern with groundless litigation that informs the Prison Litigation Reform Act and tailored its response to that context.” Because the principal office of § 2254 and § 2255 is “to upend a criminal judgment” ( 96 F.3d at 855 ), Martin concluded that collateral attacks on criminal convictions and sentences should be grouped with criminal rather than civil proceedings for the purpose of § 1915(b) and (g). In other words, Martin holds that collateral attacks should be treated the same way as the judgment being contested.
But prison discipline is not a criminal proceeding; Wolff and Baxter hold that it is civil in nature, and that norms of the criminal process (such as the right to counsel) do not apply. If under Martin a collateral attack is treated (so far as the plRa is concerned) as a continuation of the decision being challenged, then a collateral attack on prison discipline is civil rather than criminal. Part in of Newlin so holds, which means that petitions for habeas corpus arising from prison discipline are “civil actions” for purposes of § 1915(b), just as they are for purposes of § 1914(a), Fed. R.App. P. 4(a), and many other statutes and rules.
Part m of the panel’s opinion does not come to grips with the language of the plea, the reasons Martin read that language non-literally, and the reasons New-lin gave for declining to extend the non-literal reading to other uses of habeas corpus. Nonetheless, I am content with the outcome of Part iii (rather, would be content if the panel applied the same approach to Part iv). That other circuits have followed Martin to the limit is important — we should get rid of procedural conflicts to the extent we can do so with intellectual honesty, see Lee v. Clinton, 209 F.3d 1025 (7th Cir.2000); United States v. Hill, 48 F.3d 228, 231-32 (7th Cir.1995) — and the line between Martin and Newlin can be elusive. Handling claims and collections under the pira has not been easy, and stubborn adherence to a complicating factor that has not won support elsewhere has little to commend it. Except for the fact that Part iv of the panel’s opinion reintroduces this very complication, creating a new conflict among the circuits in the process!
Part IV holds that state prisoners who want to appeal adverse decisions in cases about good-time credits do not need certificates of appealability. Only prisoners seeking to appeal from the rejection of challenges to their convictions or sentences need certificates of appealability, the panel concludes. This distinction between challenges to convictions and challenges to prison discipline is exactly the line drawn (for purposes of § 1915) by Newlin , and overruled by Part iii of the panel’s opinion as both unprincipled (given Preiser and Balisok) and too difficult to implement. If this line is incompatible with Higher Authority and causes administrative headaches, and therefore is a Bad Thing in Part iii, it is still a Bad Thing when we arrive at Part iv.
When statutes leave no leeway, we must tolerate Bad Things. This is the claim of Part iv: that the statute leaves no room for maneuver. A state prisoner needs a certificate of appealability only when appealing from “the final order in a habeas corpus proceeding in which the detention complained of arises out of process issued by a State court”. 28 U.S.C. § 2253 (c)(1)(A). See also Fed. R.App. P. 22(b)(1). A prisoner who contests the deprivation of good-time credits is complaining about an administrative rather than a judicial decision. The panel remarks (216 F.3d at 637): “In light of the statutory language, we do not see how we can construe the words ‘process issued by a State court’ to mean ‘process not issued by a State court, but instead the outcome of an internal prison disciplinary proceeding.’ ” Claiming to act under the compulsion of plain statutory language, the panel disagrees with other circuits, which have held that state prisoners need certificates of appealability to obtain review of any deci *643 sion under § 2254. See, e.g., Montez v. McKinna, 208 F.3d 862, 866-69 (10th Cir.2000); Hallmark v. Johnson, 118 F.3d 1073, 1076 (5th Cir.1997). Section 2253(c)(1)(A) descends from § 2253 ¶3 (1994 ed.), which required prisoners to obtain “certificates of probable cause” to appeal, and courts considering the question uniformly have held or assumed that prisoners who sought restoration of good-time credits needed certificates of probable cause. See, e.g., Crowell v. Walsh, 151 F.3d 1050 (D.C.Cir.1998); Lemieux v. Kerby, 931 F.2d 1391 (10th Cir.1991). All of the Justices who wrote or joined opinions in Davis v. Jacobs, 454 U.S. 911 , 102 S.Ct. 417 , 70 L.Ed.2d 226 (1981), assumed that this is so. And the D.C. Circuit, at least, has linked the definition of “civil action” under § 1915(b) to the need for a certificate of appealability (or probable cause) under § 2253. It held in Crowell and Blair-Bey v. Quick, 151 F.3d 1036 (D.C.Cir.1998), that when application of § 1915(b) is excused in a collateral attack, application of § 2253(c) is essential. Our panel approves Blair-Bey (216 F.3d at 628) yet ignores Crowell, though these companion opinions represent two sides of the same coin.
The panel does not cite, and I could not find, any appellate decision holding that a state prisoner does not need a certificate of appealability (or did not need a certificate of probable cause) to appeal from the denial of a request for post-conviction collateral relief. The panel’s assertion, 216 F.3d at 638, that “it is now well established that the CA requirement does not apply to appeals in § 2241 cases” is not correct. The cases cited for this proposition, Bush v. Pitzer, 133 F.3d 455, 456 (7th Cir.1997), and Murphy v. United States, 199 F.3d 599 , 601 n. 2 (2d Cir.1999), address collateral attacks by federal prisoners. Section 2253(c)(1)(A) is limited to state prisoners’ collateral attacks; it is § 2253(c)(1)(B), not § 2253(c)(1)(A), that controls federal prisoners’ need for certificates of appealability, and § 2253(c)(1)(B) covers only “the final order in a proceeding under section 2255.” After the panel’s opinion, the seventh circuit becomes a minority of one in the interpretation of § 2253(c)(1)(A).
If textualist interpretive methods are essential, as Part iv proclaims, then what happened in Part hi? Petitions for habeas corpus “are technically civil proceedings and so come within the literal scope of the Act.” Martin, 96 F.3d at 855 , echoed at 216 F.3d at 636. A textual approach, consistently applied, would require us to overrule Martin , not Newlin , and to include all collateral attacks within the plea’s scope. Why give a pragmatic reading to the plRA in Part in, then switch to textualism in Part iv? Not for pragmatic reasons; the approaches yield diametrically opposed answers to a functionally identical question.
Part iv is at war with Part ii.b of the panel’s opinion as well as with Part in. In Part ii.b the panel holds that a petition for a writ of habeas corpus filed by a prisoner who wants good-time credits restored is covered by § 2254 because it is a proceeding “in behalf of a person in custody pursuant to the judgment of a State court”. 28 U.S.C. § 2254 (a). The panel holds that one may be “in custody pursuant to the judgment” without needing to attack that judgment. 216 F.3d at 636. Just so for § 2253(c). “[T]he detention complained of arises out of process issued by a State court” (emphasis added) even if the prisoner does not challenge the state court’s process. If challenges to good-time credits proceed under § 2254 because the words “in custody pursuant to the judgment of a State court” refer to the genesis of the custody rather than the claim made in the collateral attack, then the words “detention complained of arises out of process issued by a State court” also must refer to the genesis of the custody rather than the claim made in the collateral attack. If Part ii.b is right, then Part iv is wrong.
Part ilb has it right. Finfrock’s detention “arises out of process issued by a State court”. He has been convicted; his *644 conviction and sentence are essential to his custody. Revocation of good-time credits does not extend a prisoner’s sentence; instead this decision (like the denial or revocation of parole, or demotion to a lower credit-earning class) requires the prisoner to serve more of the original sentence. But the detention has been authorized by (that is, arises out of) the conviction and sentence (that is, process issued by a state court). Section 2253(c)(1)(A) deals with detention (= custody) that depends on a state-court order; it is not limited to attacks on that order. A prisoner who protests the revocation of good-time credits is appealing “the final order in a habeas corpus proceeding in which the detention complained of arises out of process issued by a State court” and therefore needs a certificate of appealability. The reading offered here is the one the tenth circuit adopted in Montez , deeming it an inevitable rendition of the statute’s language. Part iv makes a plain-language claim the other way. My colleagues and the tenth circuit thus agree on one thing: that § 2253(c)(1)(A) is clear. They just don’t agree on what it means, a sign that the language may not be so plain after all.
The difference between § 2254(a) (“application ... in behalf of a person in custody pursuant to the judgment of a State court”) and § 2253(c)(1)(A) (“proceeding in which the detention complained of arises out of process issued by a State court”) is not a linguistic quirk or oversight. Prisoners may seek writs of habeas corpus before conviction — perhaps to test the validity of pretrial custody (the Great Writ, dealing with excessive detention by executive officials), perhaps to test extradition to another jurisdiction for trial. A state prisoner who applies for such a writ and does not get it needs a certificate of appealability, because he is in “detention [that] arises out of process issued by a State court”. But such a prisoner has not been convicted and therefore need not satisfy § 2254 or any other restriction on actions by persons “in custody pursuant to the judgment of a State court”. The distinction is both historical and sensible. Until the latter half of the twentieth century no state prisoner could obtain collateral review following conviction by a court of competent jurisdiction, and federal law still makes it difficult to wage a collateral attack on the final judgment of a state court. But § 2253(c)(1)(A) is more general, applying to all whose custody can be traced to state judicial process, such as an arrest warrant or indictment. The panel supposes that § 2254(a) is broader than § 2253(c)(1)(A), but this is backward. Section 2253(c)(1)(A) applies before trial, in extradition cases, and after judgment too, including claims of convicts required to serve more of their sentences than they think they should.
Any reading of § 2253(c) other than the one urged here and adopted in Montez produces an anomaly: a prisoner who challenges 100% of his prison time (by attacking a conviction or sentence) needs a certificate of appealability and is likely to be turned away without an appellate decision on the merits, for only a fraction of appellants make the necessary “substantial showing of the denial of a constitutional right” (§ 2253(c)(2)). See Slack v. McDaniel, — U.S. -, 120 S.Ct. 1595 , 146 L.Ed.2d 542 (2000). But a prisoner who challenges only 1% or 5% of his prison time is guaranteed an appellate decision on the merits. Whatever sense may lie behind this eludes me.
There isn’t any uniform, or uniformly happy, solution to coverage issues under the aedpa. Federal prisoners’ collateral attacks on deprivations of good-time credits arise under § 2241, so they do not need certificates of appealability. But differences in the statute’s treatment of state and federal prisoners do not justify departing from § 2253(c)(1)(A), or for that matter from § 1915(b), when we must decide exactly how state prisoners’ petitions must be handled. These complex statutes govern hundreds of cases annually. It is worth the full court’s time to ensure that *645 the issues I have discussed are decided correctly.