NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.
SUPREME COURT OF THE UNITED STATES
MICHAEL ROSS, PETITIONER v.
on writ of certiorari to the united states court of appeals for the fourth circuit
[June 6, 2016]
Justice Kagan delivered the opinion of the Court.
The Prison Litigation Reform Act of 1995 (PLRA) mandates that an inmate exhaust “such administrative remedies as are available” before bringing suit to challenge prison conditions.
42 U. S. C. §1997e(a). The court below adopted an unwritten “special circumstances” exception to that provision, permitting some prisoners to pursue litigation even when they have failed to exhaust available administrative remedies. Today, we reject that freewheeling approach to exhaustion as inconsistent with the PLRA. But we also underscore that statute’s built-in exception to the exhaustion requirement: A prisoner need not exhaust remedies if they are not “available.” The briefs and other submissions filed in this case suggest the possibility that the aggrieved inmate lacked an available administrative remedy. That issue remains open for consideration on remand, in light of the principles stated below.
Respondent Shaidon Blake is an inmate in a Maryland prison. On June 21, 2007, two guards—James Madigan and petitioner Michael Ross—undertook to move him from his regular cell to the facility’s segregation unit. According to Blake’s version of the facts, Ross handcuffed him and held him by the arm as they left the cell; Madigan followed close behind. Near the top of a flight of stairs, Madigan shoved Blake in the back. Ross told Madigan he had Blake under control, and the three continued walking. At the bottom of the stairs, Madigan pushed Blake again and then punched him four times in the face, driving his head into the wall. After a brief pause, Madigan hit Blake one last time. Ross kept hold of Blake throughout the assault. And when the blows subsided, Ross helped Madigan pin Blake to the ground until additional officersarrived.
Later that day, Blake reported the assault to a senior corrections officer. That officer thought Madigan at fault, and so referred the incident to the Maryland prison system’s Internal Investigative Unit (IIU). Under state law, the IIU has authority to investigate allegations of employee misconduct, including the use of “excessive force.” Codeof Md. Regs., tit. 12, §11.01.05(A)(3) (2006). After conducting a year-long inquiry into the beating, the IIU issued a final report condemning Madigan’s actions, while making no findings with respect to Ross. See App. 191–195. Madigan resigned to avoid being fired.
Blake subsequently sued both guards under
42 U. S. C. §1983, alleging that Madigan had used unjustifiable force and that Ross had failed to take protective action. The claim against Madigan went to a jury, which awarded Blake a judgment of $50,000. But unlike Madigan, Ross raised the PLRA’s exhaustion requirement as an affirmative defense, contending that Blake had brought suit without first following the prison’s prescribed procedures for obtaining an administrative remedy. As set out in Maryland’s Inmate Handbook, that process—called, not very fancifully, the Administrative Remedy Procedure (ARP)—begins with a formal grievance to the prison’s warden; it may also involve appeals to the Commissioner of Correction and then the Inmate Grievance Office (IGO). See Maryland Div. of Correction, Inmate Handbook 30–31 (2007). Blake acknowledged that he had not sought a remedy through the ARP—because, he thought, the IIU investigation served as a substitute for that otherwise standard process. The District Court rejected that explanation and dismissed the suit, holding that “the commencement of an internal investigation does not relieve prisoners from the [PLRA’s] exhaustion requirement.” Blake
, No. 8:09–cv–2367 (D Md., Nov. 14, 2012), App. to Pet. for Cert. 38, 2012 WL 5568940, *5.
The Court of Appeals for the Fourth Circuit reversed in a divided decision. Stating that the PLRA’s “exhaustion requirement is not absolute,” the court adopted an extra-textual exception originally formulated by the Second Circuit. 787 F. 3d 693, 698 (2015). Repeated the Court of Appeals: “[T]here are certain ‘special circumstances’ in which, though administrative remedies may have been available[,] the prisoner’s failure to comply with administrative procedural requirements may nevertheless have been justified.” Ibid.
, 380 F. 3d 670, 676 (CA2 2004)). In particular, that was true when a prisoner “reasonably”—even though mistakenly—“believed that he had sufficiently exhausted his remedies.” 787 F. 3d, at
695. And Blake, the court concluded, fit within that exception because he reasonably thought that “the IIU’s investigation removed his complaint from the typical ARP process.” Id.,
at 700. Judge Agee dissented, stating that the PLRA’s mandatory exhaustion requirement is not “amenable” to “[j]udge-made exceptions.” Id.,
at 703. This Court granted certiorari. 577 U. S. ___ (2015).
The dispute here concerns whether the PLRA’s exhaustion requirement, §1997e(a), bars Blake’s suit. Statutory text and history alike foreclose the Fourth Circuit’s adoption of a “special circumstances” exception to that mandate. But Blake’s suit may yet be viable. Under the PLRA, a prisoner need exhaust only “available” administrative remedies. And Blake’s contention that the prison’s grievance process was not in fact available to him warrants further consideration below.
Statutory interpretation, as we always say, begins with the text, see, e.g., Hardt
v. Reliance Standard Life Ins. Co.
560 U. S. 242,
—but here following that approach at once distances us from the Court of Appeals. As Blake acknowledges, that court made no attempt to ground its analysis in the PLRA’s language. See 787 F. 3d, at 697–698; Brief for Respondent 47–48, n. 20 (labeling the Court of Appeals’ rule an “extra-textual exception to the PLRA’s exhaustion requirement”). And that failure makes a difference, because the statute speaks in unambiguous terms opposite to what the Fourth Circuit said.
Section 1997e(a) provides: “No action shall be brought with respect to prison conditions under section 1983 of this title, or any other Federal law, by a prisoner confined in any jail, prison, or other correctional facility until such administrative remedies as are available are exhausted.” As we have often observed, that language is “mandatory”: An inmate “shall” bring “no action” (or said more conversationally, may not bring any action) absent exhaustion of available administrative remedies. Woodford
548 U. S. 81,
; accord, Jones
549 U. S. 199,
(“There is no question that exhaustion is mandatory under the PLRA”). As later discussed, that edict contains one significant qualifier: the remedies must indeed be “available” to the prisoner. See infra
, at 8–10. But aside from that exception, the PLRA’s text suggests no limits on an inmate’s obligation to exhaust—irrespective of any “special circumstances.”
And that mandatory language means a court may not excuse a failure to exhaust, even to take such circumstances into account. See Miller
530 U. S. 327,
(explaining that “[t]he mandatory ‘shall’ . . . normally creates an obligation impervious to judicial discretion”). No doubt, judge-made exhaustion doctrines, even if flatly stated at first, remain amenable to judge-made exceptions. See McKart
v. United States
395 U. S. 185,
(“The doctrine of exhaustion of administrative remedies . . . is, like most judicial doctrines, subject to numerous exceptions”). But a statutory exhaustion provision stands on a different footing. There, Congress sets the rules—and courts have a role in creating exceptions only if Congress wants them to. For that reason, mandatory exhaustion statutes like the PLRA establish mandatory exhaustion regimes, foreclosing judicial discretion. See, e.g., McNeil
v. United States
508 U. S. 106,
(“We are not free to rewrite the statutory text” when Congress has strictly “bar[red] claimants from bringing suit in federal court until they have exhausted their administrative remedies”). Time and again, this Court has taken such statutes at face value—refusing to add unwritten limits onto their rigorous textual requirements. See, e.g., id.,
at 111; Shalala
v. Illinois Council on Long Term Care
529 U. S. 1
–14 (2000); see also 2 R. Pierce, Administrative Law Treatise §15.3, p. 1241 (5th ed. 2010) (collecting cases).
We have taken just that approach in construing the PLRA’s exhaustion provision—rejecting every attempt to deviate (as the Fourth Circuit did here) from its textual mandate. In Booth
532 U. S. 731 (2001)
, for example, the prisoner argued that exhaustion was not necessary because he wanted a type of relief that the administrative process did not provide. But §1997e(a), we replied, made no distinctions based on the particular “forms of relief sought and offered,” and that legislative judgment must control: We would not read “exceptions into statutory exhaustion requirements where Congress has provided otherwise.” Id.
, at 741, n. 6. The next year, in Porter
534 U. S. 516,
, the Court rejected a proposal to carve out excessive-force claims (like Blake’s) from the PLRA’s exhaustion regime, viewing that approach too as inconsistent with the uncompromising statutory text. And most recently, in Woodford
, we turned aside a requested exception for constitutional claims. 548 U. S., at 91, n. 2. Our explanation was familiar: “We are interpreting and applying” not a judge-made doctrine but a “statutory requirement,” and therefore must honor Congress’s choice. Ibid.
] All those precedents rebut the Court of Appeals’ adoption of a “special circumstances” excuse for non-exhaustion.
So too, the history of the PLRA underscores the mandatory nature of its exhaustion regime. Section §1997e(a)’s precursor, enacted in the Civil Rights of Institutionalized Persons Act (CRIPA), §7,
352 (1980), was a “weak exhaustion provision.” Woodford
, 548 U. S., at 84. Under CRIPA, a court would require exhaustion only if a State provided “plain, speedy, and effective” remedies meeting federal minimum standards—and even then, only if the court believed exhaustion “appropriate and in the interests of justice.” §7(a),
352. That statutory scheme made exhaustion “in large part discretionary.” Nussle
, 534 U. S., at 523. And for that reason (among others), CRIPA proved inadequate to stem the then-rising tide of prisoner litigation. In enacting the PLRA, Congress thus substituted an “invigorated” exhaustion provision. Woodford
, 548 U. S., at 84. “[D]iffer[ing] markedly from its predecessor,” the new §1997e(a) removed the conditions that administrative remedies be “plain, speedy, and effective” and that they satisfy minimum standards. Nussle
, 534 U. S., at 524. Still more, the PLRA prevented a court from deciding that exhaustion would be unjust or inappropriate in a given case. As described earlier, see supra,
at 4–5, all inmates must now exhaust all available remedies: “Exhaustion is no longer left to the discretion of the district court.” Woodford
, 548 U. S., at 85.
The PLRA’s history (just like its text) thus refutes a “special circumstances” exception to its rule of exhaustion. That approach, if applied broadly, would resurrect CRIPA’s scheme, in which a court could look to all the particulars of a case to decide whether to excuse a failure to exhaust available remedies. But as we have observed, such wide-ranging discretion “is now a thing of the past.” Booth
, 532 U. S., at 739. And the conflict with the PLRA’s history (as again with its text) becomes scarcely less stark if the Fourth Circuit’s exception is confined, as the court may have intended, to cases in which a prisoner makes a reasonable mistake about the meaning of a prison’s grievance procedures. Understood that way, the exception reintroduces CRIPA’s requirement that the remedial process be “plain”—that is, not subject to any reasonable misunderstanding or disagreement. §7(a),
352. When Congress amends legislation, courts must “presume it intends [the change] to have real and substantial effect.” Stone
514 U. S. 386,
. The Court of Appeals instead acted as though the amendment—from a largely permissive to a mandatory exhaustion regime—had not taken place.[2
Yet our rejection of the Fourth Circuit’s “special circumstances” exception does not end this case—because the PLRA contains its own, textual exception to mandatory exhaustion. Under §1997e(a), the exhaustion requirement hinges on the “availab[ility]” of administrative remedies: An inmate, that is, must exhaust available remedies, but need not exhaust unavailable ones. And that limitation on an inmate’s duty to exhaust—although significantly different from the “special circumstances” test or the old CRIPA standard—has real content. As we explained in Booth
, the ordinary meaning of the word “available” is “ ‘capable of use for the accomplishment of a purpose,’ and that which ‘is accessible or may be obtained.’ ” 532 U. S., at 737–738 (quoting Webster’s Third New International Dictionary 150 (1993)); see also Random House Dictionary of the English Language 142 (2d ed. 1987) (“suitable or ready for use”); 1 Oxford English Dictionary 812 (2d ed. 1989) (“capable of being made use of, at one’s disposal, within one’s reach”); Black’s Law Dictionary 135 (6th ed. 1990) (“useable”; “present or ready for immediate use”). Accordingly, an inmate is required to exhaust those, but only those, grievance procedures that are “capable of use” to obtain “some relief for the action complained of.” Booth
532 U. S., at 738.
To state that standard, of course, is just to begin; courts in this and other cases must apply it to the real-world workings of prison grievance systems. Building on our own and lower courts’ decisions, we note as relevant here three kinds of circumstances in which an administrative remedy, although officially on the books, is not capable of use to obtain relief. See Tr. of Oral Arg. 27–29 (Solicitor General as amicus curiae
acknowledging these three kinds of unavailability). Given prisons’ own incentives to maintain functioning remedial processes, we expect that these circumstances will not often arise. See Woodford
, 548 U. S., at 102.
But when one (or more) does, an inmate’s duty to exhaust “available” remedies does not come into play.
First, as Booth
made clear, an administrative procedure is unavailable when (despite what regulations or guidance materials may promise) it operates as a simple dead end—with officers unable or consistently unwilling to provide any relief to aggrieved inmates. See 532 U. S., at 736, 738. Suppose, for example, that a prison handbook directs inmates to submit their grievances to a particular administrative office—but in practice that office disclaims the capacity to consider those petitions. The procedure is not then “capable of use” for the pertinent purpose. In Booth
’s words: “[S]ome redress for a wrong is presupposed by the statute’s requirement” of an “available” remedy; “where the relevant administrative procedure lacks authority to provide any relief,” the inmate has “nothing to exhaust.” Id.,
at 736, and n. 4. So too if administrative officials have apparent authority, but decline ever to exercise it. Once again: “[T]he modifier ‘available’ requires the possibility of some relief.” Id.,
at 738. When the facts on the ground demonstrate that no such potential exists, the inmate has no obligation to exhaust the remedy.
Next, an administrative scheme might be so opaque that it becomes, practically speaking, incapable of use. In this situation, some mechanism exists to provide relief, but no ordinary prisoner can discern or navigate it. As the Solicitor General put the point: When rules are “so confusing that . . . no reasonable prisoner can use them,” then “they’re no longer available.” Tr. of Oral Arg. 23. That is a significantly higher bar than CRIPA established or the Fourth Circuit suggested: The procedures need not be sufficiently “plain” as to preclude any reasonable mistake or debate with respect to their meaning. See §7(a),
352; 787 F. 3d, at 698–699; supra,
at 3, 6–8. When an administrative process is susceptible of multiple reasonable interpretations, Congress has determined that the inmate should err on the side of exhaustion. But when a remedy is, in Judge Carnes’s phrasing, essentially “unknowable”—so that no ordinary prisoner can make sense of what it demands—then it is also unavailable. See Goebert
v. Lee County
, 510 F. 3d 1312, 1323 (CA11 2007); Turner
, 541 F. 3d 1077, 1084 (CA11 2008) (“Remedies that rational inmates cannot be expected to use are not capable of accomplishing their purposes andso are not available”). Accordingly, exhaustion is not required.
And finally, the same is true when prison administrators thwart inmates from taking advantage of a grievance process through machination, misrepresentation, or intimidation. In Woodford
, we recognized that officials might devise procedural systems (including the blind alleys and quagmires just discussed) in order to “trip up all but the most skillful prisoners.” 548 U. S., at 102. And appellate courts have addressed a variety of instances in which officials misled or threatened individual inmates so as to prevent their use of otherwise proper procedures. As all those courts have recognized, such interference with an inmate’s pursuit of relief renders the administrative process unavailable.[3
] And then, once again, §1997e(a) poses no bar.
The facts of this case raise questions about whether, given these principles, Blake had an “available” administrative remedy to exhaust. As explained earlier, Ross’s exhaustion defense rests on Blake’s failure to seek relief through Maryland’s ARP process, which begins with a grievance to the warden and may continue with appeals to the Commissioner of Correction and the IGO. See supra
, at 2–3; Inmate Handbook, at 30–31. That process is the standard method for addressing inmate complaints in the State’s prisons: The Inmate Handbook provides that prisoners may use the ARP for “all types” of grievances (subject to four exceptions not relevant here), including those relating to the use of force. Id.,
at 30; see App. 312. But recall that Maryland separately maintains the IIU to look into charges of staff misconduct in prisons, and the IIU did just that here. See supra,
at 2. Blake urged in the courts below that once the IIU commences such an inquiry, a prisoner cannot
obtain relief through the standard ARP process—whatever the Handbook may say to the contrary. See 787 F. 3d, at 697; App. to Pet. for Cert. 38, 2012 WL 5568940, at *5. And in this Court, that issue has taken on new life. Both Blake and Ross (as represented by the Maryland attorney general) have lodged additional materials relating to the interaction between the IIU and the ARP. And both
sides’ submissions, although scattershot and in need of further review, lend some support to Blake’s account—while also revealing Maryland’s grievance process to have, at least at first blush, some bewildering features.
Blake’s filings include many administrative dispositions (gleaned from the records of other prisoner suits) indicating that Maryland wardens routinely dismiss ARP grievances as procedurally improper when parallel IIU investigations are pending. One warden, for example, wrote in response to a prisoner’s complaint: “Your Request for Administrative Remedy has been received and is hereby dismissed. This issue has been assigned to the Division of Correction’s Internal Investigative Unit (Case #07–35–010621I/C), and will no longer be addressed through this process.” Lodging of Respondent 1; see also, e.g., id., at 18 (“Admin. Dismiss Final: This is being investigated outside of the ARP process by I.I.U.”). In addition, Blake has submitted briefs of the Maryland attorney general (again, drawn from former prisoner suits) specifically recognizing that administrative practice. As the attorney general stated in one case: “Wilkerson filed an ARP request,” but “his complaint already was being investigated by the [IIU], superceding an ARP investigation.” Id., at 23–24; see also, e.g., id., at 5 (Bacon’s grievance “was dismissed because the issue had been assigned to [the] IIU and would no longer be addressed through the ARP process”).
And Ross’s own submissions offer some confirmation of Blake’s view. Ross does not identify a single case in which a warden considered the merits of an ARP grievance while an IIU inquiry was underway. See Tr. of Oral Arg. 6 (Maryland attorney general’s office conceding that it had found none). To the contrary, his lodging contains still further evidence that wardens consistently dismiss such complaints as misdirected. See, e.g., Lodging of Petitioner 15 (District Court noting that “Gladhill was advised that no further action would be taken through the ARP process because the matter had been referred to the [IIU]”). Indeed, Ross’ materials suggest that some wardens use a rubber stamp specially devised for that purpose; the inmate, that is, receives a reply stamped with the legend: “Dismissed for procedural reasons . . . . This issue is being investigated by IIU case number: ____. No further action shall be taken within the ARP process.” Id., at 25, 32, 38; see Tr. of Oral Arg. 8–9 (Maryland attorney general’s office conceding the stamp’s existence and use).
Complicating the picture, however, are several cases in which an inmate refused to take a warden’s jurisdictional “no” for an answer, resubmitted his grievance up the chain to the IGO, and there received a ruling on the merits, without any discussion of the ARP/IIU issue. We confess to finding these few cases perplexing in relation to normal appellate procedure. See id., at 3–10, 13–15, 18–20 (multiple Justices expressing confusion about Maryland’s procedures). If the IGO thinks the wardens wrong to dismiss complaints because of pending IIU investigations, why does it not say so and stop the practice? Conversely, if the IGO thinks the wardens right, how can it then issue merits decisions? And if that really is Maryland’s procedure—that when an IIU investigation is underway, the warden (and Commissioner of Correction) cannot consider a prisoner’s complaint, but the IGO can—why does the Inmate Handbook not spell this out? Are there, instead, other materials provided to prisoners that communicate how this seemingly unusual process works and how to navigate it so as to get a claim heard?
In light of all these lodgings and the questions they raise about Maryland’s grievance process, we remand this case for further consideration of whether Blake had “available” remedies to exhaust. The materials we have seen are not conclusive; they may not represent the complete universe of relevant documents, and few have been analyzed in the courts below. On remand, in addition to considering any other arguments still alive in this case, the court must perform a thorough review of such materials, and then address the legal issues we have highlighted concerning the availability of administrative remedies. First, did Maryland’s standard grievance procedures potentially offer relief to Blake or, alternatively, did the IIU investigation into his assault foreclose that possibility? Second, even if the former, were those procedures knowable by an ordinary prisoner in Blake’s situation, or was the system so confusing that no such inmate could make use of it? And finally, is there persuasive evidence that Maryland officials thwarted the effective invocation of the administrative process through threats, game-playing, or misrepresentations, either on a system-wide basis or in the individual case? If the court accepts Blake’s probable arguments on one or more of these scores, then itshould find (consistent this time with the PLRA) that his suit may proceed even though he did not file an ARP complaint.
Courts may not engraft an unwritten “special circumstances” exception onto the PLRA’s exhaustion requirement. The only limit to §1997e(a)’s mandate is the one baked into its text: An inmate need exhaust only such administrative remedies as are “available.” On remand, the court below must consider how that modifying term affects Blake’s case—that is, whether the remedies he failed to exhaust were “available” under the principles set out here. We therefore vacate the judgment of the Court of Appeals and remand the case for further proceedings consistent with this opinion.
It is so ordered.