 |
|||||
|
|
|
|
|
||
|
|
|||||
|
|
|
|
|
||
In habeas cases, the statute requires the court to "dispose of the matter as law and justice require." 28 U. S. C. §2243. Although §2255 does not contain the same language, the same requirement is implicit from its purpose as a habeas substitute. See supra, at 7.(4) For claims which are not fundamental, neither law nor justice requires the reexamination of long-final judgments.
Res judicata is the general rule of the American legal system. A final judgment on the merits conclusively settles the claims and issues determined. See Montana v. United States, 440 U. S. 147, 153 (1979). An adverse determination is binding against the nonprevailing party in any court, regardless of how erroneous it may have been. In Montana, relitigation of a federal constitutional question was precluded even in this Court and even against the United States government. Id., at 164. Exceptions to the rule require strong justification. Existing exceptions should not be extended without careful consideration of whether the extension is warranted.
A. Judgments: Void, Voidable, Erroneous and Correct.
Judgments are not cleanly divided into "valid" and "invalid" judgments. Instead, they lie along a spectrum of vulnerability. At one end of the spectrum is the judgment which is absolutely void. Normally the only defect severe enough to cause absolute voidness is a lack of jurisdiction. Imprisonment under such a judgment would itself be a crime. See Ex parte Watkins, 3 Pet. (28 U. S.) 193, 203 (1830).
At the other extreme are the vast majority of judgments. They have been validly rendered by courts of competent jurisdiction in proceedings free of reversible error. Such judgments will be affirmed if they are appealed and will be enforced in the face of any collateral attack.
The middle of the spectrum consists of judgments rendered erroneously by courts of competent jurisdiction. These judgments differ from each other in their degree of vulnerability. Some types of errors may be attacked at some stages of the process but not at others. A judgment may have been entered after the erroneous admission of an important piece of evidence. Such a judgment would be reversible on appeal. In most cases, however, it would not be vulnerable to collateral attack.(5) If no appeal is taken or if the judgment is affirmed on appeal, even erroneously, the judgment is no longer subject to question.
More fundamental defects will result in a judgment being "voidable," i.e., subject to being set aside in a collateral attack such as habeas corpus. Such a judgment is nonetheless "valid" in the sense that executive officers may properly execute it until such time as it is actually set aside. The legislative branch may attach civil disabilities, enforceable by criminal sanctions, which also apply until the judgment is actually set aside. Lewis v. United States, 445 U. S. 55, 67 (1980) (firearm possession by felon).
The status of a judgment as "voidable" is not necessarily static. Upon successful completion of habeas corpus proceedings, the judgment becomes void rather than merely voidable. See id., at 61, n. 5. On the other hand, the right to challenge a judgment may be lost by the passage of time, see 28 U. S. C. §2244(d); Rule 9(a) of the Rules Governing Section 2254 Cases in the United States District Courts, or by the failure to invoke the usual appellate remedies, see Murray v. Carrier, 477 U. S. 478, 490-491 (1986).
B. Reliability and Procedural Defaults.
Petitioner repeatedly inveighs against "unconstitutional" prior convictions, as if "constitutional" and "fundamental" were synonymous. See Brief for Petitioner 7-15. Arguably, they should be. One might well argue that the detailed code of criminal procedure judicially engrafted on to the Bill of Rights, see Friendly, Is Innocence Irrelevant? Collateral Attack on Criminal Judgments, 38 U. Chi. L. Rev. 142, 156 (1970), should never have been promulgated. It is far too late in the day, though, to pretend that all the rules now considered constitutionally based are actually essential to the fundamental fairness of the trial. See Rose v. Lundy, 455 U. S. 509, 543, n. 8 (1982) (Stevens, J., dissenting).
As Justice Stevens noted, the key to recognizing which rules really are essential to reliability of the result can be found in the jurisprudence of retroactivity. If compliance with Miranda v. Arizona, 384 U. S. 436 (1966) were really essential to the reliability of trials in which confessions are used, why were all other defendants in then-pending cases denied the benefits of the decision? See Johnson v. New Jersey, 384 U. S. 719, 721 (1966). The reason is that the Court did not believe that noncompliance with its rule cast genuine doubt on the reliability of the guilt verdict. See id., at 727-728 (distinguishing Gideon and Jackson v. Denno, 378 U. S. 368 (1964), which applied retroactively due to " 'clear danger of convicting the innocent' "). Ernesto Miranda, unlike Clarence Gideon, was convicted on retrial. See L. Baker, Miranda: Crime, Law and Politics 191-193 (1983). The defendants in the three joined cases were also reconvicted or plea bargained. See id., at 191-192, n. *. That is why "it is familiar learning that all 'claims of constitutional error are not fungible.' " Brecht v. Abrahamson, 507 U. S. 619, 640 (1993) (Stevens, J., concurring).
Of all the constitutional rules which apply to the guilt phase of trial, the one which is least needed for reliability of the verdict is the Fourth Amendment exclusionary rule. See Stone v. Powell, 428 U. S. 465, 490 (1976). The impact of this rule on reliability is actually negative. The runner-up in this category is the rule of Boykin v. Alabama, 395 U. S. 238 (1969). To gauge the impact of Boykin on reliability, one need only ask any experienced trial judge how many defendants have changed their minds upon being "Boykinized" and decided to go to trial. Most will answer "none." Cases in which compliance with Boykin makes any difference at all to the outcome are exceedingly rare. Typically, the only effect of noncompliance with Boykin is to hand the defendant a ground for challenging a plea he would have entered anyway.
Along the same lines is the rule defendant seeks to invoke in this case, that the judge taking a guilty plea advise the defendant of the nature of the change. See Brief for Petitioner 2-3. In most cases, this exercise merely repeats what defense counsel has already told the defendant. See Henderson v. Morgan, 426 U. S. 637, 647 (1976). Defendant himself illustrates the disconnection of this rule from reliability of the result by asserting, "It does not matter if a defendant probably would have pled guilty anyway." J. A. 51. Amicus submits that it matters in deciding whether to extend twice-removed collateral attacks to such claims.
The implicit theme of defendant's argument is that all constitutional claims must be considered at all stages of criminal proceedings. That is not the law. Claims drop off as the processes move on, and their impact on confidence in actual guilt is often a key factor. In answer to Judge Friendly's rhetorical question, innocence is relevant.
The exclusionary rule falls off first, because of its uniquely contra-reliability effect. See Stone v. Powell, supra, 428 U. S., at 490. Next come rules created after the date of finality of defendant's case on direct review which do not "implicat[e] the fundamental fairness and accuracy of the criminal proceeding." Saffle v. Parks, 494 U. S. 484, 495 (1990). Defendants with such claims are denied relief, and they remain in prison even though their convictions, in petitioner's terminology, are "unconstitutional." Then, strongly analogous to the present case, there are claims which could have been raised on appeal but were not. These are generally not cognizable on habeas, with an exception for actual innocence. See Schlup v. Delo, 513 U. S. 298, 326-327 (1995). Finally, there are claims in which the probability of a different result rises to the level of reasonable doubt but not to the level of substantial and injurious effect. These claims warrant reversal on direct appeal but not on habeas corpus. Brecht v. Abrahamson, supra, 507 U. S., at 638.
Collateral attack on a prior conviction, as opposed to the present conviction, is a step further removed, and it makes sense that fewer attacks be allowed at this very late stage. Again, impact on reliability should be the key. Defects in the prior proceedings have no impact whatever on the reliability of the present conviction. The specter of the imprisoned innocent man, which haunts so many habeas opinions, is absent here. The issue is only how long to imprison a criminal, not whether he is a criminal. The worst-case injustice on the defendant's side is an order of magnitude less than it is with guilt-phase claims. Cf. Schlup, supra, 513 U. S., at 325-326 (contrasting claim of actual innocence with claim of "too severe" sentence).
The question, then, is where to draw the line on cognizable claims. Given the zero impact on reliability of the present guilt judgment, and given that the defendant could have challenged his priors when they were rendered, does it make any sense to open these judgments to the same full range of claims as are cognizable in a habeas attack on the present convictions? Amicus believes that question was asked and answered in Custis v. United States, 511 U. S. 485 (1994). The answer is no.
|
|
||||
|
|
|
|
||
4. On one point we agree with petitioner. There is no basis to give federal prisoners any less review under §2255 than state prisoners receive under
§2254. See Brief for Petitioner 19. If there were to be any difference, federalism concerns would point the other direction.
5. This is true even if the ground for exclusion is the Fourth Amendment. Such a claim will not be considered on federal habeas if the state
provides an opportunity for full and fair litigation of the claim. Stone v. Powell, 428 U. S. 465, 481-482 (1976).