Invoking 28 U.S.C. § 2255 (2000), petitioner-appellant George Sepulveda, a federal prisoner, mounted a collateral attack on his conviction and sentence for witness intimidation. The district court repulsed the attack. The petitioner's ensuing appeal presents two questions of first impression within this circuit. The first asks whether the rule announced in Apprendi v. New Jersey, 530 U.S. 466, 120 S. Ct. 2348, 147 L. Ed. 2d 435 (2000), applies retroactively to cases on collateral review. The second, relevant only if Apprendi lacks such retroactivity, asks whether the enactment of the Antiterrorism and Effective Death Penalty Act (AEDPA), Pub.L. No. 104-132, 110 Stat. 1214 (1996), rendered the doctrine of Teague v. Lane, 489 U.S. 288, 109 S. Ct. 1060, 103 L. Ed. 2d 334 (1989), inapplicable to initial petitions for post-conviction relief under section 2255. Concluding, as we do, that Apprendi has no retroactive effect and that Teague is not a dead letter in the AEDPA context, we affirm the denial of the petitioner's habeas corpus application.

The petitioner, known in some circles as "King Paradise," was convicted of multiple crimes arising out of his activities as the leader of the Providence chapter of the Almighty Latin King Nation. We affirmed those convictions in United States v. Lara, 181 F.3d 183 (1st Cir.1999). Here, we reconstruct only the factual terrain pertinent to the issues before us, referring those who seek a more detailed topographic model to our earlier opinion. See id. at 190-91.

A grand jury empaneled in the District of Rhode Island returned a thirteen-count indictment against the petitioner and several others. After a lengthy trial, a petit jury convicted the petitioner of racketeering, conspiracy to commit racketeering, murder in aid of racketeering, witness intimidation, and possessing a firearm as a convicted felon. See 18 U.S.C. §§ 1962(c), 1962(d), 1959(a), 1512(b)(3), 922(g)(1). The district court imposed three concurrent life sentences, a concurrent twenty-year incarcerative term, and a concurrent ten-year incarcerative term. We affirmed the convictions and sentences, Lara, 181 F.3d at 206, and the Supreme Court eschewed further review, 528 U.S. 1127, 120 S. Ct. 960, 145 L. Ed. 2d 833 (2000).

A federal criminal conviction becomes final when the Supreme Court denies certiorari. See Clay v. United States, 537 U.S. 522, 123 S. Ct. 1072, 1076, 155 L. Ed. 2d 88 (2003). Approximately five months after the denial of certiorari in the petitioner's case, the Court decided Apprendi. The central holding of Apprendi is that "[o]ther than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt." 530 U.S. at 490, 120 S. Ct. 2348. This clarion call arguably conflicted with the procedure employed by the district court in sentencing the petitioner on the witness intimidation count. We explain briefly.

The witness intimidation count arose from an incident wherein the petitioner ordered the "termination" of Manuel Pacheco, a state prisoner who was assisting the authorities in their investigation of the Latin Kings.1 Lara, 181 F.3d at 200. The petitioner's order passed through various channels to henchmen (incarcerated Latin King members) within the penitentiary in which Pacheco was housed. Id. At the first opportunity, these acolytes beat Pacheco savagely. Id. at 200-01.

A defendant convicted of witness intimidation typically faces a maximum prison term of ten years. See 18 U.S.C. § 1512(b). At sentencing, however, the government asked the lower court to enhance the petitioner's sentence pursuant to the Criminal Street Gangs Act, which authorizes a ten-year bump in the sentence of an individual convicted of certain offenses — including witness intimidation — if it is determined that the individual committed the offense "intend[ing] to promote or further the felonious activities of [a] criminal street gang or maintain or increase his or her position in the gang." Id. § 521(d). The sentencing court found that the petitioner's conviction fit within these confines and, accordingly, imposed a twenty-year sentence (double the maximum otherwise authorized under the statute of conviction).

On December 29, 2000, the petitioner filed a federal habeas application — technically, an application to vacate, set aside, or correct his sentence — under section 2255. See Ellis v. United States, 313 F.3d 636, 641 (1st Cir.2002) (observing that, as to federal prisoners, section 2255 is in essence "a surrogate for the historic writ of habeas corpus"). In it, he argued that he had been sentenced to a term of imprisonment above the default statutory maximum for the crime of conviction based on a judge's factual finding that the circumstances attendant to his commission of that crime warranted the enhancement, and that the resultant sentence violated the Apprendi rule. The district court denied the application, holding that Apprendi could not be applied retroactively to cases on collateral review.

The petitioner moved for a certificate of appealability (COA). See 28 U.S.C. § 2253. We granted the request with respect to the effect, if any, that the Apprendi decision might have on the petitioner's conviction for witness intimidation. This appeal followed.

It is beyond cavil that the petitioner's sentence for witness intimidation exceeds the default statutory maximum for that crime, and that this overage rests on the sentencing court's finding that the petitioner had acted with the intention of furthering the felonious activities of a criminal street gang and/or maintaining his leadership position in such a gang. Thus, the petitioner has made at least a prima facie showing of a violation of the Apprendi rule.2 But the petitioner's conviction had already become final before Apprendi was handed down, and the threshold question is whether the new rule applies to his case. See Derman v. United States, 298 F.3d 34, 39 (1st Cir.2002). The petitioner advances two theories in support of the retroactive application of the Apprendi rule. We address these theories sequentially.

The Supreme Court's decision in Teague v. Lane constitutes a general bar to the retroactive application of newly announced rules of criminal procedure. Tyler v. Cain, 533 U.S. 656, 665, 121 S. Ct. 2478, 150 L. Ed. 2d 632 (2001) (citing Teague, 489 U.S. at 311-13, 109 S. Ct. 1060). The Teague bar admits of two exceptions. The first allows retroactive application of new rules that either (a) prohibit criminal punishment for certain types of primary conduct, or (b) forbid the imposition of certain categories of punishment for particular classes of defendants. O'Dell v. Netherland, 521 U.S. 151, 157, 117 S. Ct. 1969, 138 L. Ed. 2d 351 (1997). This exception is patently inapposite here: Apprendi neither places any particular type of conduct beyond the reach of the criminal law nor pretermits any particular type of punishment for a specific class of defendants. Accord McCoy v. United States, 266 F.3d 1245, 1256 (11th Cir.2001). Thus, the petitioner's challenge necessarily stands or falls on the strength of the second exception.

The second Teague exception allows retroactive application of "watershed rules of criminal procedure implicating the fundamental fairness and accuracy of the criminal proceeding." Graham v. Collins, 506 U.S. 461, 478, 113 S. Ct. 892, 122 L. Ed. 2d 260 (1993) (internal quotation marks omitted). For this exception to flourish, the new rule must pass two tests. First, "[i]nfringement of the rule must seriously diminish the likelihood of obtaining an accurate conviction." Tyler, 533 U.S. at 665, 121 S. Ct. 2478 (citations and internal quotation marks omitted). Second, the new rule must itself alter the accepted understanding of the bedrock procedural elements essential to the integrity and fairness of a criminal proceeding. Id.

Nothing in the Apprendi decision indicates to us that infringements of its rule will seriously diminish the accuracy of convictions (which, by definition, must take place before any such infringement occurs). The case before us provides a practical illustration of why this is so. The petitioner was tried before a jury and convicted of witness intimidation upon proof of his guilt beyond a reasonable doubt. There is no indication that, in the course of his trial, he received less than the process that was due. The procedural error to which the petitioner adverts may raise questions as to the length of his sentence, but inaccuracies of this nature, occurring after a defendant has been duly convicted, are matters of degree and do not trump what the Justices have termed "the general rule of nonretroactivity." Tyler, 533 U.S. at 665, 121 S. Ct. 2478.

We add, moreover, that the length of the petitioner's sentence was not plucked out of thin air, but, rather, was determined by a federal judge based upon discrete findings of fact established by a fair preponderance of the evidence. We agree with the Seventh Circuit that findings by federal judges, though now rendered insufficient in certain instances by Apprendi, nonetheless "are adequate to make reliable decisions about punishment." Curtis v. United States, 294 F.3d 841, 844 (7th Cir.2002). After all, even in the post-Apprendi era, findings of fact made by the sentencing judge, under a preponderance standard, remain an important part of the sentencing regimen. See, e.g., United States v. Caba, 241 F.3d 98, 101 (1st Cir.2001) ("[A]fter Apprendi, ... sentencing factors that boost a defendant's sentence but do not trip a new statutory maximum remain grist for the district judge's mill under a preponderance-of-the-evidence standard."); see also United States v. Robinson, 241 F.3d 115, 122 (1st Cir.2001) (rejecting the argument that when facts found by the judge trigger or increase a mandatory minimum sentence, an Apprendi violation occurs). Consequently, Apprendi's new rule not only fails to impugn the accuracy of convictions that became final beforehand but also falls short of rendering sentences imposed under the pre-Apprendi regime seriously inaccurate. See Goode v. United States, 305 F.3d 378, 385 (6th Cir.2002); cf. United States v. Sanchez-Cervantes, 282 F.3d 664, 669 (9th Cir.2002) ("We do not believe that requiring the jury to make drug quantity determinations beyond a reasonable doubt will greatly affect the accuracy of convictions.").

By the same token, we do not believe that the Apprendi rule can be characterized as a watershed rule of criminal procedure. Without in any way denigrating either the importance or the impact of Apprendi, that decision cannot plausibly be said to have altered the commonly accepted understanding of the bedrock procedural elements of our criminal justice system. Accord United States v. Sanders, 247 F.3d 139, 150 (4th Cir.2001) (rejecting the notion that, pre-Apprendi, "the country's criminal justice system malfunctioned... fundamentally"). Watershed rules affecting bedrock procedural elements are few and far between. The quintessential example of such a rule — the only example specifically mentioned by the Supreme Court — is the landmark decision in Gideon v. Wainwright, 372 U.S. 335, 83 S. Ct. 792, 9 L. Ed. 2d 799 (1963). The "sweeping rule of Gideon ... established an affirmative right to counsel in all felony cases." O'Dell, 521 U.S. at 167, 117 S. Ct. 1969. It protected each accused, "though he be not guilty, [from] the danger of conviction because he does not know how to establish his innocence." Gideon, 372 U.S. at 345, 83 S. Ct. 792 (citation omitted). This pronouncement — that representation by counsel is fundamental to a fair trial — reshaped the legal landscape and dramatically revised the common understanding of what the Due Process Clause demands in a criminal case.

Other concrete examples of watershed rules are hen's-teeth rare. See Graham, 506 U.S. at 478, 113 S. Ct. 892 ("[W]e operate from the premise that such procedures would be so central to an accurate determination of innocence or guilt [that] it [is] unlikely that many such components of basic due process have yet to emerge."). One reason for this phenomenon is that the Supreme Court has exhibited reluctance to showcase prototypes of rules that might enjoy this venerated status. That is understandable because the Teague exception "is clearly meant to apply only to a small core of rules requiring observance of those procedures that are implicit in the concept of ordered liberty." O'Dell, 521 U.S. at 157, 117 S. Ct. 1969.

Apprendi is not within this small core of cases. The rule that it announces merely assures a previously convicted defendant that increased punishment, over and above the default statutory maximum, can only be imposed if the factual predicate for the increase — other than a prior criminal conviction — is confirmed by a jury to a higher quantum of proof. See Apprendi, 530 U.S. at 484, 120 S. Ct. 2348 ("If a defendant faces punishment beyond that provided by statute when an offense is committed under certain circumstances but not others,... it necessarily follows that the defendant should not ... be deprived of protections that have, until that point, unquestionably attached.") (emphasis supplied). Although the Apprendi rule is important as a means of clarifying the proper factfinding roles of judge and jury, it affords an innocent defendant no additional shield from wrongful conviction. Refined to bare essence, the rule "merely limits the potential penalty to be imposed on [an undoubtedly] guilty defendant" to that designated by statute. Goode, 305 F.3d at 385.

In sum, "a decision ... by a judge (on the preponderance standard) rather than a jury (on the reasonable-doubt standard) is not the sort of error that necessarily undermines the fairness ... of judicial proceedings." Curtis, 294 F.3d at 843. Applying Apprendi's procedural mandate retroactively would create an unacceptably high risk that those found guilty of criminal conduct might escape condign sentences. Hence, we join every court of appeals that thus far has decided the question and hold that the Apprendi rule fails to qualify as a watershed rule within the meaning of the second Teague exception. See Coleman v. United States, 329 F.3d 77, 81 (2d Cir.2003); Goode, 305 F.3d at 382-85; United States v. Brown, 305 F.3d 304, 307-10 (5th Cir.2002) (per curiam); Curtis, 294 F.3d at 843-44; United States v. Mora, 293 F.3d 1213, 1218-19 (10th Cir.2002); Sanchez-Cervantes, 282 F.3d at 667; McCoy, 266 F.3d at 1257-58; United States v. Moss, 252 F.3d 993, 997 (8th Cir.2001); Sanders, 247 F.3d at 151.

In an effort to turn aside the combined force of reasoning and precedent, the petitioner offers several diversions. None is persuasive, but three of his sallies merit brief comment.

In the first place, the petitioner emphasizes that no less an authority than Justice O'Connor has characterized the majority opinion in Apprendi as one that announces "a watershed change in constitutional law." Apprendi, 530 U.S. at 524, 120 S. Ct. 2348 (O'Connor, J. dissenting). The petitioner's reliance on this statement demonstrates the verity that a single word often has the capacity to convey multiple shades of meaning. See Hanover Ins. Co. v. United States, 880 F.2d 1503, 1504 (1st Cir.1989) (explaining that "words can be like chameleons, which reflect the color of their environment") (citation and internal quotation marks omitted). Justice O'Connor warned that "the most significant impact of the Court's decision" in Apprendi was that it "threaten[ed] to unleash a flood of petitions by convicted felons seeking to invalidate their sentences." 530 U.S. at 550-51, 120 S. Ct. 2348. Her concern was "a practical one," prompting her to chastise the majority for "not say[ing] whether [determinate-sentencing] schemes are constitutional." Id. at 550, 120 S. Ct. 2348. She did not advert to Teague, and her use of the word "watershed," taken in this context, cannot plausibly be interpreted as a definitive conclusion that she — or anyone else — would hold the Apprendi rule to be a watershed rule for Teague purposes.

Next, the petitioner posits that the Apprendi rule is not strictly a rule of criminal procedure, but, rather, possesses a substantive component, requiring those facts that warrant sentence enhancement to be considered elements of an aggravated crime. This formulation contains more cry than wool.

The Apprendi decision is about criminal procedure, pure and simple. Accord Curtis, 294 F.3d at 843. Long before the Court decided Apprendi, basic principles of criminal and constitutional jurisprudence protected defendants from being convicted on less than proof beyond a reasonable doubt of each and every element of a charged crime. See In re Winship, 397 U.S. 358, 364, 90 S. Ct. 1068, 25 L. Ed. 2d 368 (1970) (holding that "the Due Process Clause protects the accused against conviction except upon proof beyond a reasonable doubt of every fact necessary to constitute the [charged] crime"). Notwithstanding the prevalence of this principle, courts routinely approved the practice of enhancing sentences based on a judge's factual determinations. See, e.g., United States v. Thomas, 204 F.3d 381, 383 (2d Cir.2000); United States v. Grimaldo, 214 F.3d 967, 974-75 (8th Cir.2000); United States v. Lindia, 82 F.3d 1154, 1160-61 (1st Cir.1996). The Apprendi rule limited that widespread practice.3 The fact that the Apprendi rule implements certain underlying constitutional protections makes it no less a rule of criminal procedure. See Sanders, 247 F.3d at 151. The Supreme Court has instructed us that the "distinction between substance and procedure is an important one in the habeas context," Bousley v. United States, 523 U.S. 614, 620, 118 S. Ct. 1604, 140 L. Ed. 2d 828 (1998), and we cannot disregard that distinction in applying the Teague screen.

In the last analysis, "[t]he Teague doctrine is founded on the notion that one of the principal functions of habeas corpus is to assure that no man has been incarcerated under a procedure which creates an impermissibly large risk that the innocent will be convicted." Id. (internal quotation marks and alterations omitted). Pre-Apprendi practice, though incorrect, created no such risk. While the Apprendi rule may improve the correlation between crime and punishment in future prosecutions, it does not render the correlation for past convictions completely (or even seriously) unreliable. Thus, the procedure required by Apprendi has the capacity to improve substance — but it is not substance in and of itself.

The petitioner's third line of attack attempts to compare Apprendi to the Supreme Court decisions in Bailey v. United States, 516 U.S. 137, 116 S. Ct. 501, 133 L. Ed. 2d 472 (1995), and Richardson v. United States, 526 U.S. 813, 119 S. Ct. 1707, 143 L. Ed. 2d 985 (1999). Based on these comparisons, the petitioner asserts that the rationale for applying the Bailey and Richardson doctrines retroactively supports giving Apprendi retroactive effect. We reject this attempted analogy for the most basic of reasons: the cases are not fair congeners.

Bailey involved a determination of what conduct Congress criminalized in enacting 18 U.S.C. § 924(c)(1). The Bailey Court determined that "possession" was not "use" and vacated the defendant's conviction. 516 U.S. at 143, 116 S. Ct. 501. Because Bailey involved substance, not procedure, the decision provides no guidance for a Teague analysis. See Bousley, 523 U.S. at 620, 118 S. Ct. 1604 ("[B]ecause Teague by its terms applies only to procedural rules, we think it is inapplicable to the situation in which this Court decides the meaning of a criminal statute enacted by Congress.").

Richardson is not quite so far afield. That case involved both the interpretation of a criminal statute and the procedural safeguards attendant to a defendant's right to have a jury make factual findings leading to conviction thereunder. Richardson, 526 U.S. at 816, 119 S. Ct. 1707. Although the latter half of this hybrid bears some similarities to Apprendi, the first half strays from Apprendi in so material a way as to undermine the attempted analogy.

The core holding in Richardson required jury unanimity as to each "violation" in the "series of violations" needed to convict under the continuing criminal enterprise (CCE) statute, 21 U.S.C. § 848. See Richardson, 526 U.S. at 818-19, 119 S. Ct. 1707. Jury discordance as to the identity of the individual violations that comprised the necessary series would threaten the accuracy of any conviction for violating the CCE statute, a factor that makes retroactive application of Richardson more appropriate under a Teague analysis.4 Apprendi errors pose no such threat. Thus, we find the petitioner's comparison of Apprendi to Richardson to be like comparing a plum to a pomegranate.

That ends this aspect of the matter. We hold, without serious question, that Apprendi prescribes a new rule of criminal procedure, and that Teague does not permit inferior federal courts to apply the Apprendi rule retroactively to cases on collateral review.

This conclusion does not mark the end of our odyssey. The petitioner contends in the alternative that a Teague analysis is unnecessary. This contention rests on the premise that the AEDPA evinces Congress's intent to displace the Teague doctrine in the context of initial section 2255 petitions filed within one year of the date upon which a conviction becomes final.

The petitioner's premise derives, in the first instance, from an AEDPA provision governing, inter alia, federal prisoners' initial applications for habeas relief under section 2255. Section 2255 consists of a narrative series of unnumbered paragraphs, but for ease in reference we cite to the paragraphs by artificially imposed symbol and number. The provision on which the petitioner's premise depends reads in pertinent part:

A 1-year period of limitation shall apply to a motion under [section 2255]. The limitation period shall run from the latest of —

(1) the date on which the judgment of conviction becomes final;

(2) the date on which the impediment to making a motion created by the governmental action in violation of the Constitution or laws of the United States is removed, if the movant was prevented from making a motion by such governmental action;

(3) the date on which the right asserted was initially recognized by the Supreme Court, if that right has been newly recognized by the Supreme Court and made retroactively applicable to cases on collateral review; or

(4) the date on which the facts supporting the claim or claims presented could have been discovered through the exercise of due diligence.

28 U.S.C. § 2255 ¶ 6. The petitioner argues that ¶ 6(1) allows a federal habeas petitioner to file an initial application for post-conviction relief based on a new rule of criminal procedure regardless of when the new rule was announced as long as the application is filed within one year of the date on which the judgment of conviction becomes final. In his view, this interpretation — which eliminates the need for any retroactivity analysis — is legitimized by contrasting the differing language of ¶ 6(1) with that of ¶ 6(3), which embodies a specific standard for when new rules may be given retroactive effect in connection with petitions for section 2255 relief filed more than one year after a final judgment of conviction. This is an imaginative exercise in legal legerdemain, but not a convincing one.

In discerning the meaning of the words that Congress wrote, an inquiring court invariably must start with the statutory text. United States v. Charles George Trucking Co., 823 F.2d 685, 688 (1st Cir.1987). In parsing that text, "we will not depart from, or otherwise embellish, the language of a statute absent either undeniable textual ambiguity or some other extraordinary consideration, such as the prospect of yielding a patently absurd result." Pritzker v. Yari, 42 F.3d 53, 67-68 (1st Cir.1994) (citations omitted). The commonsense corollary of these principles is that statutory language can only be ascribed meaning if taken in context. United States v. Ahlers, 305 F.3d 54, 57-58 (1st Cir.2002). Part of that context is the presumption that, in the ordinary case, "Congress expects its statutes to be read in conformity with [the Supreme] Court's precedents." Clay, 123 S.Ct. at 1076.

Viewed through this prism, we find nothing in ¶ 6 sufficient to overturn the general bar against retroactive application of new rules of criminal procedure set forth in Teague. In terms, ¶ 6 deals strictly and solely with the temporal limitations for filing section 2255 petitions. See Brown, 305 F.3d at 307. The prefatory language and structure of ¶ 6 reveal an obvious legislative intent to place a one-year limitation period on federal habeas claims that arise under different circumstances. See Derman, 298 F.3d at 40 ("In [the AEDPA], Congress for the first time established time limits applicable to the filing of habeas petitions: a prisoner (state or federal) has one year from [a certain] date ... within which to seek federal habeas relief.").

The succeeding subsections of ¶ 6 implement and embellish this intent. The baseline rule is contained in ¶ 6(1). This subsection unambiguously obligates a federal prisoner to bring all habeas claims that are extant at the time of final judgment within one year. See 28 U.S.C. § 2255 ¶ 1 (authorizing a federal prisoner to seek review of a sentence that "was imposed in violation [of the law] ..., or is otherwise subject to collateral attack") (emphasis supplied). If sentence was imposed consistent with then-prevailing law and that sentence is not otherwise subject to collateral attack, the prisoner has no claim to bring under section 2255.

Notwithstanding this baseline rule, ¶ 6(3) holds out the possibility that a change in the prevailing law may make a petitioner's conviction susceptible to collateral review in the future. Thus, if the Supreme Court later announces a new rule that arguably has retroactive application to the conviction, the date of that decision marks the accrual of a new habeas claim. Under that paradigm, ¶ 6(3) provides a further one-year limitation period within which the petitioner may bring his newborn claim.5 Analytically, then, the different sub-parts that constitute ¶ 6 deal with different starting points for the running of the limitation period. There is no substantive language in ¶ 6 that realistically can be viewed as displacing the wonted application of the Teague doctrine to federal habeas claims. See Goode, 305 F.3d at 384; Brown, 305 F.3d at 307; Mora, 293 F.3d at 1218.

The petitioner has another string to his statutory construction bow. He adverts to a variety of other provisions within the realm of federal habeas law and notes that they are worded differently than 28 U.S.C. § 2255 ¶ 6. See 28 U.S.C. § 2254(d) (imposing restrictions on state prisoners' habeas claims); id. § 2255 ¶ 8 (imposing restrictions on federal prisoners' second or successive habeas petitions). Invoking the hoary tenet that "where Congress includes particular language in one section of a statute but omits it in another ... it is generally presumed that Congress acted intentionally," Duncan v. Walker, 533 U.S. 167, 173, 121 S. Ct. 2120, 150 L. Ed. 2d 251 (2001), he asserts that we should interpret those provisions that do not contain a specific directive about how to apply new Supreme Court rules retroactively — such as ¶ 6(1) — as negating any such requirement (and, to that extent, rendering Teague obsolete). This argument lacks force.

As the petitioner notes, the language of these provisions differs from the language of 28 U.S.C. § 2255 ¶ 6(3). But the provisions that the petitioner cites impose different substantive rules on the retroactivity analysis to be employed. For that reason, the canon of construction on which the petitioner relies furnishes no sound basis for concluding that because ¶ 6(1) is silent as to retroactivity, anything goes.6 The logical reading — and the one that we endorse — is that retroactivity is not mentioned in ¶ 6(1) because the impact of new rules announced after final judgment is covered in ¶ 6(3). We explain briefly.

Every provision of federal habeas law is subject to Teague's doctrinal reach. See Clay, 123 S.Ct. at 1076. The changed phraseology of section 2254(d) and 2255 ¶ 8 evinces a congressional intent to increase the burden of certain habeas petitioners above and beyond the baseline demonstration that a case "has been ... made retroactively applicable to cases on collateral review." 28 U.S.C. § 2255 ¶ 6(3). Accordingly, 28 U.S.C. § 2255 ¶ 8 restricts second or successive habeas petitions challenging federal convictions to those new rules of constitutional law explicitly made retroactive by the Supreme Court itself to cases on collateral review. See Tyler, 533 U.S. at 662, 121 S. Ct. 2478 (holding that the identically worded requirement contained in 28 U.S.C. § 2244(b)(2)(A) is "satisfied only if [the Supreme] Court has held that the new rule is retroactively applicable to cases on collateral review"). The restrictive language of ¶ 8 plainly reflects congressional intent to limit the reach of Teague and to exact a more stringent retroactivity requirement for second or successive habeas petitions. Id.

Like 28 U.S.C. § 2255 ¶ 6, 28 U.S.C. § 2254(d) applies to initial applications for post-conviction relief. Unlike section 2255, however, section 2254 relates to state prisoners. It requires such petitioners to bear a different burden; they must demonstrate that the state court's adjudication of the claim "resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States." 28 U.S.C. § 2254(d)(1). That standard reflects congressional concern that, in the interests of comity, federal courts should defer to reasonable state-court judgments anent state prisoners — even when those judgments are arguably erroneous. See Williams v. Taylor, 529 U.S. 362, 385-86, 120 S. Ct. 1495, 146 L. Ed. 2d 389 (2000); Sanna v. Dipaolo, 265 F.3d 1, 7-9 (1st Cir.2001).

Thus, by imposing these increased burdens on differently situated habeas petitioners, sections 2254(d) and 2255 ¶ 8 function as two of the AEDPA's gatekeeping provisions. Mora, 293 F.3d at 1218. 28 U.S.C. § 2255 ¶ 6 serves a different purpose. That provision, in its entirety, deals exclusively with temporal limitations for the filing of initial petitions by federal prisoners. In terms, ¶ 6(3) does no more than designate a specific limitation period for filing motions related to new rules or rights that pass through the Teague screen. Those that fail under Teague by definition fail to "appl[y] to cases on collateral review." 28 U.S.C. § 2255 ¶ 6(3).

That resolves the question regarding Teague's continuing relevance. The short of it is that Teague remains alive and well, notwithstanding Congress's enactment of the AEDPA.

We need go no further.7 To recapitulate, we hold (1) that the Apprendi rule has no retroactive application to cases in which the judgment of conviction became final before Apprendi was decided; and (2) that the passage of the AEDPA does not affect the wonted application of Teague v. Lane to initial petitions for habeas relief filed under 28 U.S.C. § 2255.

Affirmed.