Wilson v. Sellers, 584 U.S. ___ (2018)

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

_________________

No. 16–6855

_________________

MARION WILSON, PETITIONER v. ERIC SELLERS, WARDEN

on writ of certiorari to the united states court of appeals for the eleventh circuit

[April 17, 2018]

Justice Breyer delivered the opinion of the Court.

The Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA) requires a prisoner who challenges (in a federal habeas court) a matter “adjudicated on the merits in State court” to show that the relevant state-court “decision” (1) “was contrary to, or involved an unreasonable application of, clearly established Federal law,” or (2) “was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U. S. C. §2254(d). Deciding whether a state court’s decision “involved” an unreasonable application of federal law or “was based on” an unreasonable determination of fact requires the federal habeas court to “train its attention on the particular reasons—both legal and fac- tual—why state courts rejected a state prisoner’s federal claims,” Hittson v. Chatman, 576 U. S. ___, ___ (2015) (Ginsburg, J., concurring in denial of certiorari) (slip op., at 1), and to give appropriate deference to that decision, Harrington v. Richter, 562 U. S. 86, 101–102 (2011).

This is a straightforward inquiry when the last state court to decide a prisoner’s federal claim explains its decision on the merits in a reasoned opinion. In that case, a federal habeas court simply reviews the specific reasons given by the state court and defers to those reasons if they are reasonable. We have affirmed this approach time and again. See, e.g., Porter v. McCollum, 558 U. S. 30, 39–44 (2009) (per curiam); Rompilla v. Beard, 545 U. S. 374, 388–392 (2005); Wiggins v. Smith, 539 U. S. 510, 523–538 (2003).

The issue before us, however, is more difficult. It concerns how a federal habeas court is to find the state court’s reasons when the relevant state-court decision on the merits, say, a state supreme court decision, does not come accompanied with those reasons. For instance, the decision may consist of a one-word order, such as “affirmed” or “denied.” What then is the federal habeas court to do? We hold that the federal court should “look through” the unexplained decision to the last related state-court decision that does provide a relevant rationale. It should then presume that the unexplained decision adopted the same reasoning. But the State may rebut the presumption by showing that the unexplained affirmance relied or most likely did rely on different grounds than the lower state court’s decision, such as alternative grounds for affirmance that were briefed or argued to the state supreme court or obvious in the record it reviewed.

I

In 1997 a Georgia jury convicted petitioner, Marion Wilson, of murder and related crimes. After a sentencing hearing, the jury sentenced Wilson to death. In 1999 the Georgia Supreme Court affirmed Wilson’s conviction and sentence, Wilson v. State, 271 Ga. 811, 525 S. E. 2d 339 (1999), and this Court denied his petition for certiorari, Wilson v. Georgia, 531 U. S. 838 (2000).

Wilson then filed a petition for habeas corpus in a state court, the Superior Court for Butts County. Among other things, he claimed that his counsel was “ineffective” during his sentencing, in violation of the Sixth Amendment. See Strickland v. Washington, 466 U. S. 668, 687 (1984) (setting forth “two components” of an ineffective-assistance-of-counsel claim: “that counsel’s performance was deficient” and “that the deficient performance prejudiced the defense”). Wilson identified new evidence that he argued trial counsel should have introduced at sentencing, namely, testimony from various witnesses about Wilson’s childhood and the impairment of the frontal lobe of Wilson’s brain.

After a hearing, the state habeas court denied the petition in relevant part because it thought Wilson’s evidence did not show that counsel was “deficient,” and, in any event, counsel’s failure to find and present the new evidence that Wilson offered had not prejudiced Wilson. Wilson v. Terry, No. 2001–v–38 (Super. Ct. Butts Cty., Ga., Dec. 1, 2008), App. 60–61. In the court’s view, that was because the new evidence was “inadmissible on evidentiary grounds,” was “cumulative of other testimony,” or “otherwise would not have, in reasonable probability, changed the outcome of the trial.” Id., at 61. Wilson applied to the Georgia Supreme Court for a certificate of probable cause to appeal the state habeas court’s decision. But the Georgia Supreme Court denied the application without any explanatory opinion. Wilson v. Terry, No. 2001–v–38 (May 3, 2010), App. 87, cert. denied, 562 U. S. 1093 (2010).

Wilson subsequently filed a petition for habeas corpus in the United States District Court for the Middle District of Georgia. He made what was essentially the same “ineffective assistance” claim. After a hearing, the District Court denied Wilson’s petition. Wilson v. Humphrey, No. 5:10–cv–489 (Dec. 19, 2013), App. 88–89. The court assumed that Wilson’s counsel had indeed been “deficient” in failing adequately to investigate Wilson’s background and physical condition for mitigation evidence and to present what he likely would have found at the sentencing hearing. Id., at 144. But, the court nonetheless deferred to the state habeas court’s conclusion that these deficiencies did not “prejudice” Wilson, primarily because the testimony of many witnesses was “cumulative,” and because the evidence of physical impairments did not include any physical examination or other support that would have shown the state-court determination was “unreasonable.” Id., at 187; see Richter, 562 U. S., at 111–112.

Wilson appealed to the Court of Appeals for the Eleventh Circuit. Wilson v. Warden, 774 F. 3d 671 (2014). The panel first held that the District Court had used the wrong method for determining the reasoning of the relevant state court, namely, that of the Georgia Supreme Court (the final and highest state court to decide the merits of Wilson’s claims). Id., at 678. That state-court decision, the panel conceded, was made without an opinion. But, the federal court was wrong to “look through” that decision and assume that it rested on the grounds given in the lower court’s decision. Instead of “looking through” the decision to the state habeas court’s opinion, the federal court should have asked what arguments “could have supported” the Georgia Supreme Court’s refusal to grant permission to appeal. The panel proceeded to identify a number of bases that it believed reasonably could have supported the decision. Id., at 678–681.

The Eleventh Circuit then granted Wilson rehearing en banc so that it could consider the matter of methodology. Wilson v. Warden, 834 F. 3d 1227 (2016). Ultimately six judges (a majority) agreed with the panel and held that its “could have supported” approach was correct. Id., at 1235. Five dissenting judges believed that the District Court should have used the methodology it did use, namely, the “look through” approach. Id., at 1242–1247, 1247–1269. Wilson then sought certiorari here. Because the Eleventh Circuit’s opinion creates a split among the Circuits, we granted the petition. Compare id., at 1285 (applying “could have supported” approach), with Grueninger v. Director, Va. Dept. of Corrections, 813 F. 3d 517, 525–526 (CA4 2016) (applying “look through” presumption post-Richter), and Cannedy v. Adams, 706 F. 3d 1148, 1156–1159 (CA9 2013) (same); see also Clements v. Clarke, 592 F. 3d 45, 52 (CA1 2010) (applying “look through” presumption pre-Richter); Bond v. Beard, 539 F. 3d 256, 289–290 (CA3 2008) (same); Mark v. Ault, 498 F. 3d 775, 782–783 (CA8 2007) (same); Joseph v. Coyle, 469 F. 3d 441, 450 (CA6 2006) (same).

II

We conclude that federal habeas law employs a “look through” presumption. That conclusion has parallels in this Court’s precedent. In Ylst v. Nunnemaker, a defendant, convicted in a California state court of murder, appealed his conviction to the state appeals court where he raised a constitutional claim based on Miranda v. Arizona, 384 U. S. 436 (1966). 501 U. S. 797, 799–800 (1991). The appeals court rejected that claim, writing that “ ‘an objection based upon a Miranda violation cannot be raised for the first time on appeal.’ ” Id., at 799. The defendant then similarly challenged his conviction in the California Supreme Court and on collateral review in several state courts (including once again the California Supreme Court). In each of these latter instances the state court denied the defendant relief (or review). In each instance the court did so without an opinion or other explanation. Id., at 799–800.

Subsequently, the defendant asked a federal habeas court to review his constitutional claim. Id., at 800. The higher state courts had given no reason for their decision. And this Court ultimately had to decide how the federal court was to find the state court’s reasoning in those circumstances. Should it have “looked through” the unreasoned decisions to the state procedural ground articulated in the appeals court or should it have used a different method?

In answering that question Justice Scalia wrote the following for the Court:

“The problem we face arises, of course, because many formulary orders are not meant to convey anything as to the reason for the decision. Attributing a reason is therefore both difficult and artificial. We think that the attribution necessary for federal habeas purposes can be facilitated, and sound results more often assured, by applying the following presumption: Where there has been one reasoned state judgment rejecting a federal claim, later unexplained orders upholding that judgment or rejecting the same claim rest upon the same ground. If an earlier opinion ‘fairly appear[s] to rest primarily upon federal law,’ we will presume that no procedural default has been invoked by a subsequent unexplained order that leaves the judgment or its consequences in place. Similarly where, as here, the last reasoned opinion on the claim explicitly imposes a procedural default, we will presume that a later decision rejecting the claim did not silently disregard that bar and consider the merits.” Id., at 803 (citation omitted).

Since Ylst, every Circuit to have considered the matter has applied this presumption, often called the “look through” presumption, but for the Eleventh Circuit—even where the state courts did not apply a procedural bar to review. See supra, at 4–5. And most Federal Circuits applied it prior to Ylst. See Ylst, supra, at 803 (citing Prihoda v. McCaughtry, 910 F. 2d 1379, 1383 (CA7 1990); Harmon v. Barton, 894 F. 2d 1268, 1272 (CA11 1990); Evans v. Thompson, 881 F. 2d 117, 123, n. 2 (CA4 1989); Ellis v. Lynaugh, 873 F. 2d 830, 838 (CA5 1989)).

That is not surprising in light of the fact that the “look through” presumption is often realistic, for state higher courts often (but certainly not always, see Redmon v. Johnson, 2018 WL 415714 (Ga., Jan. 16, 2018)) write “denied” or “affirmed” or “dismissed” when they have examined the lower court’s reasoning and found nothing significant with which they disagree.

Moreover, a “look through” presumption is often (but not always) more efficiently applied than a contrary approach—an approach, for example, that would require a federal habeas court to imagine what might have been the state court’s supportive reasoning. The latter task may prove particularly difficult where the issue involves state law, such as state procedural rules that may constrain the scope of a reviewing court’s summary decision, a matter in which a federal judge often lacks comparative expertise. See Ylst, supra, at 805.

The State points to a later case, Harrington v. Richter, 562 U. S. 86 (2011), which, it says, controls here instead of Ylst. In its view, Ylst should apply, at most, to cases in which the federal habeas court is trying to determine whether a state-court decision without opinion rested on a state procedural ground (for example, a procedural default) or whether the state court has reached the merits of a federal issue. In support, it notes that Richter held that the state-court decisions to which AEDPA refers include summary dispositions, i.e., decisions without opinion. Richter added that “determining whether a state court’s decision resulted from an unreasonable legal or factual conclusion does not require that there be an opinion from the state court explaining the state court’s reasoning.” 562 U. S., at 98.

Richter then said that, where “a state court’s decision is unaccompanied by an explanation, the habeas petitioner’s burden still must be met by showing there was no reasonable basis for the state court to deny relief.” Ibid. And the Court concluded that, when “a federal claim has been presented to a state court and the state court has denied relief, it may be presumed that the state court adjudicated the claim on the merits in the absence of any indication or state-law procedural principles to the contrary.” Id., at 99.

In our view, however, Richter does not control here. For one thing, Richter did not directly concern the issue before us—whether to “look through” the silent state higher court opinion to the reasoned opinion of a lower court in order to determine the reasons for the higher court’s decision. Indeed, it could not have considered that matter, for in Richter, there was no lower court opinion to look to. That is because the convicted defendant sought to raise his federal constitutional claim for the first time in the California Supreme Court (via a direct petition for habeas corpus, as California law permits). Id., at 96.

For another thing, Richter does not say the reasoning of Ylst does not apply in the context of an unexplained decision on the merits. To the contrary, the Court noted that it was setting forth a presumption, which “may be overcome when there is reason to think some other explanation for the state court’s decision is more likely.” Richter, supra, at 99–100. And it referred in support to Ylst, 501 U. S., at 803.

Further, we have “looked through” to lower court decisions in cases involving the merits. See, e.g., Premo v. Moore, 562 U. S. 115, 123–133 (2011); Sears v. Upton, 561 U. S. 945, 951–956 (2010) (per curiam). Indeed, we de- cided one of those cases, Premo, on the same day we decided Richter. And in our opinion in Richter we referred to Premo. 562 U. S., at 91. Had we intended Richter’s “could have supported” framework to apply even where there is a reasoned decision by a lower state court, our opinion in Premo would have looked very different. We did not even cite the reviewing state court’s summary affirmance. Instead, we focused exclusively on the actual reasons given by the lower state court, and we deferred to those reasons under AEDPA. 562 U. S., at 132 (“The state postconviction court’s decision involved no unreasonable application of Supreme Court precedent”).

III

The State’s further arguments do not convince us. The State points out that there could be many cases in which a “look through” presumption does not accurately identify the grounds for the higher court’s decision. And we agree. We also agree that it is more likely that a state supreme court’s single word “affirm” rests upon alternative grounds where the lower state court decision is unreasonable than, e.g., where the lower court rested on a state-law proce- dural ground, as in Ylst. But that is why we have set forth a presumption and not an absolute rule. And the unreasonableness of the lower court’s decision itself provides some evidence that makes it less likely the state supreme court adopted the same reasoning. Thus, additional evidence that might not be sufficient to rebut the presumption in a case like Ylst would allow a federal court to conclude that counsel has rebutted the presumption in a case like this one. For instance, a federal habeas court may conclude that counsel has rebutted the presumption on the basis of convincing alternative arguments for affirmance made to the State’s highest court or equivalent evidence presented in its briefing to the federal court similarly establishing that the State’s highest court relied on a different ground than the lower state court, such as the existence of a valid ground for affirmance that is obvious from the state-court record. The dissent argues that the Georgia Supreme Court’s recent decision in Redmon v. Johnson rebuts the presumption in Georgia because that court indicated its summary decisions should not be read to adopt the lower court’s reasoning. Post, at 6–8, 10–11 (opinion of Gorsuch, J.). This misses the point. A presumption that can be rebutted by evidence of, for instance, an alternative ground that was argued or that is clear in the record was the likely basis for the decision is in accord with full and proper respect for state courts, like those in Georgia, which have well-established systems and procedures in place in order to ensure proper consideration to the arguments and contention in the many cases they must process to determine whether relief should be granted when a criminal conviction or its ensuing sentence is challenged.

The State also points out that we do not necessarily presume that a silent opinion of a federal court of appeals adopts the reasoning of the court below. The dissent similarly invokes these “traditional rules of appellate practice.” See post, at 5–6, 10. But neither the State nor the dissent provides examples of similar context. Were we to adopt a “look through” approach in respect to silent federal appeals court decisions as a general matter in other contexts, we would risk judges and lawyers reading those decisions as creating, through silence, a precedent that could be read as binding throughout the circuit—just what a silent decision may be thought not to do. Here, however, we “look through” the silent decision for a spe- cific and narrow purpose—to identify the grounds for the higher court’s decision, as AEDPA directs us to do. See supra, at 1–2. We see no reason why the federal court’s interpretation of the state court’s silence should be taken as binding precedent outside this context, for example, as a statewide binding interpretation of state law.

Further, the State argues that the “look through” approach shows disrespect for the States. See Brief for Respondent 39 (“Wilson’s approach to summary decisions reflects an utter lack of faith in the ability of the highest state courts to adjudicate constitutional rights”). We do not believe this is so. Rather the presumption seeks to replicate the grounds for the higher state court’s decision. Where there are convincing grounds to believe the silent court had a different basis for its decision than the analysis followed by the previous court, the federal habeas court is free, as we have said, to find to the contrary. In our view, this approach is more likely to respect what the state court actually did, and easier to apply in practice, than to ask the federal court to substitute for silence the federal court’s thought as to more supportive reasoning.

Finally, the State argues that the “look through” approach will lead state courts to believe they must write full opinions where, given the workload, they would have preferred to have decided summarily. Though the matter is empirical, given the narrowness of the context, we do not believe that they will feel compelled to do so—at least not to any significant degree. The State offers no such evidence in the many Circuits that have applied Ylst outside the procedural context. See supra, at 5.

For these reasons, we reverse the Eleventh Circuit’s judgment and remand the case for further proceedings consistent with this opinion.

It is so ordered.

SUPREME COURT OF THE UNITED STATES

_________________

No. 16–6855

_________________

MARION WILSON, PETITIONER v. ERIC SELLERS, WARDEN

on writ of certiorari to the united states court of appeals for the eleventh circuit

[April 17, 2018]

Justice Gorsuch, with whom Justice Thomas and Justice Alito join, dissenting.

After a state supreme court issues a summary order sustaining a criminal conviction, should a federal habeas court reviewing that decision presume it rests only on the reasons found in a lower state court opinion? The answer is no. The statute governing federal habeas review permits no such “look through” presumption. Nor do traditional principles of appellate review. In fact, we demand the opposite presumption for our work—telling readers that we independently review each case and that our summary affirmances may be read only as signaling agreement with a lower court’s judgment and not necessarily its reasons. Because I can discern no good reason to treat the work of our state court colleagues with less respect than we demand for our own, I would reject petitioner’s presumption and must respectfully dissent.

Even so, some good news can be found here. While the Court agrees to adopt a “look through” presumption, it does so only after making major modifications to petitioner’s proposal. The Court tells us that the presumption should count for little in cases “where the lower state court decision is unreasonable” because it is not “likely” a state supreme court would adopt unreasonable reasoning. Ante, at 9. In cases like that too, the Court explains, federal courts remain free to sustain state court convictions whenever reasonable “ground[s] for affirmance [are] obvious from the state-court record” or appear in the parties’ submissions in state court or the federal habeas proceeding. Ibid. Exactly right, and exactly what the law has always demanded. So while the Court takes us on a journey through novel presumptions and rebuttals, it happily returns us in the end very nearly to the place where we began and belonged all along.

*

To see the problem with petitioner’s presumption, start with the statute. The Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA) governs federal review of state criminal convictions. It says a federal court may not grant habeas relief overturning a state court conviction “with respect to any claim that was adjudicated on the merits in State court proceedings” unless (among other things) the petitioner can show that the state court proceedings “resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law.” 28 U. S. C. §2254(d)(1). As the text and our precedent make clear, a federal habeas court must focus its review on the final state court decision on the merits, not any preceding decision by an inferior state court. See Greene v. Fisher, 565 U. S. 34, 40 (2011). Nor does it matter whether the final state court decision comes with a full opinion or in a summary order: the same deference is due all final state court decisions. Harrington v. Richter, 562 U. S. 86, 98 (2011); Cullen v. Pinholster, 563 U. S. 170, 187 (2011).

The upshot of these directions is clear. Even when the final state court decision “is unaccompanied by an explanation, the habeas petitioner’s burden still must be met by showing there was no reasonable basis for the state court to deny relief.” Richter, 562 U. S., at 98 (emphasis added). And before a federal court can disregard a final summary state court decision, it “must determine what arguments or theories . . . could have supporte[d] the state court’s decision; and then it must ask whether it is possible fairminded jurists could disagree that those arguments or theories are inconsistent with the holding in a prior decision of this Court.” Id., at 102 (emphasis added). Far from suggesting federal courts should presume a state supreme court summary order rests on views expressed in a lower court’s opinion, then, AEDPA and our precedents require more nearly the opposite presumption: federal courts must presume the order rests on any reasonable basis the law and facts allow.

If this standard seems hard for a habeas petitioner to overcome, “that is because it was meant to be.” Ibid. In AEDPA, Congress rejected the notion that federal habeas review should be “a substitute for ordinary error correction.” Id., at 102–103. Instead, AEDPA “reflects the view that habeas corpus is a ‘guard against extreme malfunctions in the state criminal justice systems.’ ” Id., at 102 (emphasis added). “The reasons for this approach are familiar. ‘Federal habeas review of state convictions frustrates both the States’ sovereign power to punish offenders and their good-faith attempts to honor constitutional rights.’ It ‘disturbs the State’s significant interest in repose for concluded litigation, denies society the right to punish some admitted offenders, and intrudes on state sovereignty to a degree matched by few exercises of federal judicial authority.’ ” Id., at 103 (citations omitted).

Petitioner and the Court today labor to distinguish these authorities, but I don’t see how they might succeed. They point to the fact that in Richter no state court had issued a reasoned order, while here a lower state court did. See Brief for Petitioner 28–30; ante, at 8. But on what account of AEDPA or Richter does that factual distinction make a legal difference? Both the statute and our precedent explain that federal habeas review looks to the final state court decision, not any decision preceding it. Both instruct that to dislodge the final state court decision a petitioner must prove it involved an unreasonable application of federal law. And to carry that burden in the face of a final state court summary decision, Richter teaches that the petitioner must show no lawful basis could have reasonably supported it. To observe that some final state court summary decisions are preceded by lower court reasoned opinions bears no more relevance to the AEDPA analysis than to say that some final state court summary decisions are issued on Mondays.[1]

Unable to distinguish Richter, petitioner seeks to confine it by caricature. Because that case requires a federal court to “imagine” its own arguments for denying habeas relief and engage in “decision-making-by-hypothetical,” he argues it should be limited to its facts. Brief for Petitioner 28–30, 33; Reply Brief 9. But the Court today does not adopt petitioner’s characterization, and for good reason: Richter requires no such thing. In our adversarial system a federal court generally isn’t required to imagine or hypothesize arguments that neither the parties before it nor any lower court has presented. To determine if a reason-able basis “could have supported” a summary denial of habeas relief under Richter, a federal court must look to the state lower court opinion (if there is one), any argument presented by the parties in the state proceedings, and any argument presented in the federal habeas proceeding. Of course, a federal court sometimes may con-sider on its own motion alternative bases for denying habeas relief apparent in the law and the record, but it does not generally bear an obligation to do so. See Wood v. Mil-yard, 566 U. S. 463, 471–473 (2012) (discussing Day v. McDonough, 547 U. S. 198 (2006), and Granberry v. Greer, 481 U. S. 129 (1987)).

Nor is that the end of the problems with petitioner’s “look through” presumption. It also defies traditional rules of appellate practice that informed Congress’s work when it adopted AEDPA and that should inform our work today. McQuiggin v. Perkins, 569 U. S. 383, 398, n. 3 (2013). Appellate courts usually have an independent duty to review the facts and law in the cases that come to them. Often they see errors in lower court opinions. But often, too, they may affirm on alternative bases either argued by the parties or (sometimes) apparent to them on the face of the record. See, e.g., SEC v. Chenery Corp., 318 U. S. 80, 88 (1943) (noting “the settled rule that, in reviewing the decision of a lower court, it must be affirmed if the result is correct ‘although the lower court relied upon a wrong ground or gave a wrong reason’ ”); Wood, supra, at 473. And a busy appellate court sometimes may not see the profit in devoting its limited resources to explaining the error and the alternative basis for affirming when the outcome is sure to remain the same, so it issues a summary affirmance instead. To reflect these realities, this Court has traditionally warned readers against presuming our summary affirmance orders rest on reasons articulated in lower court opinions. Comptroller of Treasury of Md. v. Wynne, 575 U. S. ___, ___–___ (2015) (slip op., at 16–17) (“ ‘[A] summary affirmance is an affirmance of the judgment only,’ and ‘the rationale of the affirmance may not be gleaned solely from the opinion below’ ”); Mandel v. Bradley, 432 U. S. 173, 176 (1977) (per curiam). The courts of appeals have issued similar warnings for similar reasons about their own summary orders. See, e.g., Rates Technology, Inc. v. Mediatrix Telecom, Inc., 688 F. 3d 742, 750 (CA Fed. 2012); DeShong v. Seaboard Coast Line R. Co., 737 F. 2d 1520, 1523 (CA11 1984). And respect for this traditional principle of appellate practice surely weighs against presuming a state court’s summary disposition rests solely on a lower court’s opinion. On what account could we reasonably demand more respect for our summary decisions than we are willing to extend to those of our state court colleagues?

Petitioner and the Court offer only this tepid reply. They suggest that their “look through” presumption seeks to reflect “realistic[ally]” the basis on which the state summary decision rests. See Brief for Petitioner 44; ante, at 7. But to the extent this is a claim that their presumption comports realistically with longstanding traditions of appellate practice, it is wrong for the reasons just laid out. In fact, applying traditional understandings of appellate practice, this Court has refused to presume that state appellate courts even read lower court opinions rather than just the briefs before them. See Baldwin v. Reese, 541 U. S. 27, 31 (2004). And surely it is a mystery how the Court might today presume state supreme courts rely on that which it traditionally presumes they do not read.

If the argument here is instead an empirical claim that the “look through” presumption comports realistically with what happened in this case and others like it, it is wrong too. Petitioner was convicted in Georgia. And during the pendency of this case in our Court, the Georgia Supreme Court issued an order confirming that lower courts in that State may not “presum[e] that when this Court summarily denies an application to appeal an order denying habeas corpus relief, we necessarily agree with everything said in that order.” Redmon v. Johnson, 809 S. E. 2d 468, 472 (Ga. 2018). The court explained that it has long followed just this rule for all the reasons you’d expect. It independently reviews the facts and law in each habeas case. If it finds something it thinks might amount to a consequential error, the court sets the case for argument and usually prepares a full opinion. But “[o]n many occasions,” the court finds only “inconsequential errors.” Id., at 471.[2] And in these cases the court normally issues a summary affirmance because the costs associated with full treatment of the appeal outweigh the benefits of correcting what is at most harmless error, especially given the court’s heavy caseload and the need to attend to more consequential matters.[3] Petitioner’s presumption thus does not seek to reflect reality; it seeks to deny it.

The presumption is especially unrealistic in another way. The Court and petitioner presume that a summary order by a state supreme court adopts all the specific reasons expressed by a lower state court. In doing so, they disregard a far more realistic possibility: that the state supreme court might have relied only on the same grounds for the denial of relief as did the lower court without necessarily adopting all its reasoning. Here, the lower state court denied petitioner’s Strickland claim on the grounds that counsel’s performance was not deficient and peti-tioner suffered no prejudice. And it gave several reasons for its conclusions: for example, the evidence petitioner sought to admit “would have been inadmissible on evidentiary grounds, cumulative of other testimony, or otherwise would not have, in reasonable probability, changed the outcome of the trial.” App. 61. In summarily denying relief, the state supreme court might have reached the same conclusions (no deficient performance and no prejudice) without resting on the exact same reasons.

While the “look through” presumption cannot be squared with AEDPA’s text, traditional rules, or Georgia’s actual practice, petitioner and the Court contend it is at least consistent with Ylst v. Nunnemaker, 501 U. S. 797 (1991). See Brief for Petitioner 38; ante, at 5–8. But it is not. In habeas review of state court convictions, federal courts may only review questions of federal law. So if a state court decision rejecting a petitioner’s federal law claim rests on a state procedural defect (say the petitioner filed too late under state rules), federal courts generally have no authority to reach the federal claim. Ylst simply teaches that, if a lower state court opinion expressly relied on an independent and adequate state ground, we should presume a later state appellate court summary disposition invoked it too. See 501 U. S., at 801, 803. The decision thus seeks to protect state court decisions from displacement and reaches a result consistent with the traditional rule that a summary order invokes all fairly presented bases for affirmance.

Neither can Ylst be reimagined today as meaning anything more. The case came years before AEDPA’s new standards for habeas review and can offer nothing useful about them. The work of interpreting AEDPA’s demands was left instead to Richter. And, as we’ve seen, Richter forecloses petitioner’s presumption. Of course, and as petitioner stresses, Richter didn’t overrule Ylst. But that’s for the simple reason that Ylst continues to do important, if limited, work in the disposition of procedural default claims because “AEDPA did not change the application of pre-AEDPA procedural default principles.” B. Means, Federal Habeas Manual §9B:3 (2017).

Uncomfortable questions follow too from any effort to reimagine Ylst. If we were to take Ylst as suggesting that summary decisions presumptively rely only on the reasons found in lower court opinions, wouldn’t we have to overrule our many precedents like Wynne and Mandel that explicitly reject any such presumption? Wouldn’t circuit courts have to discard their own similar precedents? See supra, at 5–6. Consistency would seem to demand no less.

The only answer petitioner and the Court offer is no answer at all. Consistency, they suggest, is overrated. Everywhere else in the law we should retain the usual rule that a summary affirmance can’t be read as presumptively resting on the lower court’s reasons. They encourage us to use Ylst only as a tool for making a special exception for AEDPA cases: here and here alone should we adopt petitioner’s “look through” presumption. Brief for Petitioner 18, 20; ante, at 10 (stating that “we ‘look through’ the silent decision for a specific and narrow purpose” under AEDPA). But just stating this good-for-habeas-only rule should be enough to reject it. Summary orders that happen to arise in state habeas cases should receive no less respect than those that arise anywhere else in the law. If anything, they should receive more respect, because federal habeas review of state court decisions “ ‘intrudes on state sovereignty to a degree matched by few exercises of federal judicial authority.’ ” Richter, 562 U. S., at 103.

*

Petitioner’s novel presumption not only lacks any provenance in the law, it promises nothing for its trouble. Consider the most obvious question it invites, one suggested by the facts of our own case: what happens when a state supreme court issues an order explaining that its summary affirmances do not necessarily adopt the reasons in lower court opinions? Should that be enough to rebut the “look through” presumption? After defending the presumption, even the dissent in the Eleventh Circuit decision under review recognized that a disclaimer along these lines should suffice to rebut it. See Wilson v. Warden, 834 F. 3d 1227, 1263 (2016) (en banc) (opinion of J. Pryor, J.) (“The Georgia Supreme Court could simply issue a one-line order denying an application for a certificate of probable cause that indicates agreement with the result the superior court reached but not the lower court’s reasons for rejecting the petitioner’s claim”). And, of course, the Georgia Supreme Court has recently responded to the dissent’s invitation by issuing just such a disclaimer. So in the end petitioner’s presumption seems likely to accomplish nothing for him and only needless work for others—inducing more state supreme courts to churn out more orders restating the obvious fact that their summary dispositions don’t necessarily rest on the reasons given by lower courts. Along the way, too, it seems federal courts will have their hands full. For while the Eleventh Circuit dissent had no difficulty acknowledging that an order like Georgia’s suffices to overcome petitioner’s presumption, the Court today refuses to supply the same obvious answer.

Consider, too, the questions that would follow in the unlikely event a general order like the one from the Georgia Supreme Court wasn’t considered enough to overcome petitioner’s presumption. Quickly federal courts would be forced to decide: does the “look through” presumption survive even when a state supreme court includes language in every summary order explaining that its decision does not necessarily adopt the reasoning below? What if the state supreme court says something slightly different but to the same effect, declaring in each case that it has independently considered the relevant law and evidence before denying relief? And if we start dictating what state court disclaimers should look like and where they should appear, what exactly is left of Congress’s direction that our review is intended to guard only against “ ‘extreme malfunctions’ ” in state criminal justice systems? Richter, supra, at 102. Wouldn’t we be slipping into the business of “tell[ing] state courts how they must write their opinions,” something this Court has long said federal habeas courts “have no power” to do? Coleman v. Thompson, 501 U. S. 722, 739 (1991).

Apart from whether a (general or case-specific) order from a state supreme court suffices to overcome petitioner’s presumption, there’s the question what else might. Say a lower state court opinion includes an error but the legal briefs or other submissions presented to the state supreme court supply sound alternative bases for affirmance. In those circumstances, should a federal habeas court really presume that the state supreme court chose to repeat the lower court’s mistake rather than rely on the solid grounds argued to it by the parties? What if a sound alternative basis for affirmance is presented for the first time in the parties’ federal habeas submissions: are we to presume that the state supreme court was somehow less able to identify a reasonable basis for affirmance than federal habeas counsel?

Here at least the Court does offer an answer. Petitioner insists that federal courts should presume that state supreme court summary orders rest on unreasonable lower state court opinions even in the face of reasonable alternative arguments presented to the state supreme court or in federal habeas proceedings. But seeming to recognize the unreasonableness of this request, the Court opts to reshape radically petitioner’s proposed presumption before adopting it. First, the Court states that “it is more likely that a state supreme court’s single word ‘affirm’ rests upon alternative grounds where the lower state court decision is unreasonable.” Ante, at 9. Then, the Court proceeds to explain that “a federal habeas court may conclude that counsel has rebutted the presumption on the basis of convincing alternative arguments for affirmance made to the State’s highest court or equivalent evidence presented in its briefing to the federal court similarly establishing that the State’s highest court relied on a different ground than the lower state court, such as the existence of a valid ground for affirmance that is obvious from the state-court record.” Ibid.

The Court’s reshaping of petitioner’s presumption reveals just how futile this whole business really is. If, as the Court holds, the “look through” presumption can be rebutted “where the lower state court decision is unreasonable,” ibid., it’s hard to see what good it does. Peti-tioner sought to assign unreasonable lower court opinions to final state court summary decisions. To hear now that essentially only reasonable (and so sustainable) lower state court opinions are presumptively adopted by final state court summary decisions will surely leave him sour on this journey and federal habeas courts scratching their heads about the point of it all. And if, as the Court also tells us, a federal habeas court can always deny relief on a basis that is apparent from the record or on the basis of alternative arguments presented by the parties in state or federal proceedings, then the “look through” presumption truly means nothing and we are back where we started. With the Court’s revisions to petitioner’s presumption, a federal habeas court is neither obliged to look through exclusively to the reasons given by a lower state court, nor required to presume that a summary order adopts those reasons.

All this is welcome news of a sort. The Court may promise us a future of foraging through presumptions and rebuttals. But at least at the end of it we rest knowing that what was true before remains true today: a federal habeas court should look at all the arguments presented in state and federal court and examine the state court record. And a federal habeas court should sustain a state court summary decision denying relief if those materials reveal a basis to do so reasonably consistent with this Court’s holdings. Exactly what a federal court applying the statute and Richter has had to do all along. See supra, at 2–5. And exactly what the Eleventh Circuit correctly held it had to do in this case.

*

Today, petitioner invites us to adopt a novel presumption that AEDPA, traditional principles of appellate review, and Georgia practice all preclude. It’s an invitation that requires us to treat the work of state court colleagues with disrespect we would not tolerate for our own. And all to what end? None at all, it turns out. As modified by the Court, petitioner’s presumption nearly drops us back where we began, with only trouble to show for the effort. Respectfully, I would decline the invitation to this circuitous journey and just affirm.