In the
United States Court of Appeals
For the Seventh Circuit

No. 98-1736

John Ryan,

Petitioner-Appellant,

v.

United States of America,

Respondent-Appellee.



Appeal from the United States District Court
for the Northern District of Illinois, Eastern Division.
No. 97 C 2765--Marvin E. Aspen, Chief Judge.


Argued January 18, 2000--Decided June 5, 2000



  Before Easterbrook, Kanne, and Diane P. Wood, Circuit
Judges.

  Easterbrook, Circuit Judge. Custis v. United
States, 

, 487 (1994), considered
"whether a defendant in a federal sentencing
proceeding may collaterally attack the validity
of previous state convictions that are used to
enhance his sentence" and held that "a defendant
has no such right (with the sole exception of
convictions obtained in violation of the right to
counsel)". We must decide whether Custis means
only that the time for the attack on the state
conviction is postponed to a collateral attack on
the federal sentence. Our answer is no. A
sentence imposed following the approach of Custis
is lawful and thus not subject to collateral
attack under 28 U.S.C. sec.2255 as long as the
prior convictions remain undisturbed. Accord,
Moore v. Roberts, 

, 702-03 (5th Cir.
1996); Turner v. United States, 

, 477
(6th Cir. 1999); Charlton v. Morris, 

(8th Cir. 1995); Clawson v. United States, 

(9th Cir. 1995), reiterated by United
States v. Daniels, 

(9th Cir. 1999).
Contra, Young v. Vaughn, 

(3d Cir.
1996); United States v. Clark, 

(5th
Cir. 2000) (agreeing with Young and disapproving
Charlton, but without mentioning the circuit’s
earlier decision in Moore).

  John Ryan was sentenced to 185 months’
imprisonment as a career offender under U.S.S.G.
sec.4B1.1 following his guilty plea to multiple
drug crimes. Career-offender enhancement is
mandatory for an adult who commits a drug felony
and has at least two prior felony convictions for
drug offenses or crimes of violence. Ryan
concedes that his criminal record contains two
convictions meeting that description but insists
that one of them--a 1980 conviction in Illinois
for armed robbery--is invalid. Ryan did not
appeal that conviction or subject it to
collateral attack while he was in custody under
it. At the sentencing for his federal drug
offenses, however, he asked the district judge to
inquire into its validity. Ryan contended that
his 1980 plea had been involuntary because the
panel from which his jury would have been
selected heard the judge sentence another
defendant and make comments deploring the high
incidence of crime. Ryan’s lawyer asked the judge
to secure a new pool of jurors; when the judge
refused, Ryan pleaded guilty. The judge in the
federal case remarked that Ryan could have gone
to trial and appealed (if he had been convicted)
to present his claim of error; the federal judge
did not see any possibility that simply by
denying Ryan’s motion the state judge rendered
his plea involuntary. Ryan repeated his argument
on appeal to this court but received a different
kind of response: that Custis precludes an
indirect collateral attack on the state sentence,
and that the 1980 conviction therefore counts for
career-offender purposes whether Ryan’s plea was
voluntary or not. 1996 U.S. App. Lexis 3836 (7th
Cir. Feb. 29, 1996). Under Custis, we held, only
the lack of counsel permits such an indirect
collateral attack.

  A few days before the statutes of limitations
in 28 U.S.C. sec.sec. 2244(d) and 2255 para.6
expired, Ryan launched two collateral attacks--
one on the 1980 state conviction, the other on
the 1995 federal sentence. The challenge to the
state conviction was assigned to District Judge
Bucklo, who dismissed it with the observation
that Ryan was no longer "in custody" under the
1980 conviction and therefore could not use
sec.2254 to contest its validity. See Maleng v.
Cook, 

, 492-93 (1989). Both Judge
Bucklo and this court denied Ryan’s application
for a certificate of appealability to review that
decision. The challenge to the federal sentence
was assigned to Chief Judge Aspen, who rejected
it on the ground that Custis is as applicable to
a petition under sec.2255 as it is to sentencing
and direct appeal. 

(N.D. Ill.
1997). Ryan offered a new theory of
involuntariness: that his lawyer compelled him to
plead guilty by refusing to go forward with the
trial unless paid $2,000. If Ryan did not tell
this to the state judge when entering his plea,
it is difficult to see how the subject can be
raised 20 years later. See United States v.
Stewart, 

(7th Cir. 1999). But the
record does not contain a transcript of the plea,
so Judge Aspen assumed, as shall we, that Ryan
not only could establish that his lawyer made
this demand but also did not know that an
indigent defendant is entitled to court-appointed
counsel. Still, Judge Aspen observed, Ryan had
the assistance of counsel at the time of his
plea, and no more is required by Custis.

  On this, Ryan’s second appeal, the United
States leads off with the argument that our 1996
decision is the law of the case, which Ryan
cannot avoid just by changing his theory about
why the plea was involuntary. Even if an indirect
collateral challenge to a conviction used to
enhance a federal sentence is, like a claim of
ineffective assistance, the sort of contention
that ordinarily may be deferred until a motion
under sec.2255, the fact remains that Ryan did
object at sentencing, and on direct appeal, to
the consideration of the 1980 state conviction.
A defendant who complains on direct appeal about
the quality of his lawyer can’t try again on
collateral attack unless there has been an
intervening change of law, United States v.
Taglia, 

, 417-18 (7th Cir. 1991), and
Ryan does not make such an argument. What he does
say, however, is that our 1996 decision should be
disregarded for the same reason Ryan thinks that
Custis is irrelevant: that all Custis does (and,
by implication, all we did in 1996) is postpone
decision to a collateral attack under sec.2255.
It is not possible to disentangle the argument
based on law of the case from the arguments about
the effect of Custis, so we turn directly to that
subject.

  Custis gave several reasons why a prior
conviction is conclusive for purposes of
recidivist sentencing. First, the Armed Career
Criminal Act, 18 U.S.C. sec.924(e), the statute
involved in Custis, "focuses on the fact of the
conviction and nothing [in sec.924] suggests that
the prior final conviction may be subject to
collateral attack for potential constitutional
errors before it may be 

(emphasis in original). Lewis v. United States,

(1980), holds that a person with a
felony conviction who possesses a firearm cannot
defend by insisting that he shouldn’t have been
convicted; that he was convicted is sufficient,
the Court concluded. Custis holds that sec.924(e)
should be treated like the felon-in-possession

. Although the Court
recognized that prior decisions had permitted an
indirect collateral challenge when a prior
conviction was uncounseled, it declined to extend
these cases. 

At the conclusion of
this discussion the Court remarked:

Ease of administration also supports the
distinction. As revealed in a number of
the cases cited in this opinion, failure
to appoint counsel at all will generally
appear from the judgment roll itself, or
from an accompanying minute order. But
determination of claims of ineffective
assistance of counsel, and failure to
assure that a guilty plea was voluntary,
would require sentencing courts to rummage
through frequently nonexistent or difficult
to obtain state-court transcripts or
records that may date from another era,
and may come from any one of the 50

. Finally, the Court observed that
"principles of finality associated with habeas
corpus actions apply with at least equal force
when a defendant seeks to attack a previous
conviction used for sentencing. By challenging
the previous conviction, the defendant is asking
a district court ’to deprive [the] [state-court
judgment] of [its] normal force and effect in a
proceeding that ha[s] an independent purpose
other than to overturn the prior judgment.’
[Parke v. Raley, 

(1992)] at 30. These
principles bear extra weight in cases in which
the prior convictions, such as one challenged by
Custis, are based on guilty pleas, because when
a guilty plea is at issue, ’the concern with
finality served by the limitation on collateral
attack has special force.’ United States v.
Timmreck, 

, 784 (1979) (footnote

(full citation to
Parke added; other brackets and parentheses in
the original).

  What the Court said about sec.924(e) is equally
applicable to the career-offender guideline.
United States v. Killion, 

, 846 (7th
Cir. 1994); United States v. Arango-Montoya, 

, 1336 (7th Cir. 1995). The Sentencing
Commission instructed courts to use the fact of
prior convictions as conclusive when calculating
criminal history. U.S.S.G. sec.4A1.2 Application
Note 6. The Commission recognized, as did the
Court in Custis, that some recidivist statutes
expressly permit contest to the validity of prior
convictions. E.g., 21 U.S.C. sec.851(c)(1). But
in the absence of such an independent right of
review, the Commission stated, all outstanding
convictions must be counted.

  Ryan wants us to disregard not only the
language in Custis (and the Guidelines)
emphasizing that the fact of prior conviction is
dispositive but also the Court’s reminder that
collateral attacks on prior convictions are
incompatible with principles of finality--
especially, Custis said, when the defendant
pleaded guilty, as Ryan did. Instead Ryan plays
up the Court’s observation about ease of
administration (the language in the block quote
above). Now that sentencing has been completed,
Ryan contends, it is as easy to go back and
determine the validity of the 1980 conviction as
it is to resolve any other collateral attack.
That may be true, but it does not justify
disregarding the other strands of the Court’s
reasoning. Custis concluded that it is proper to
count the prior conviction. Ryan could have
challenged his 1980 conviction by appeal or by
collateral attack when he was still in custody.
He did not do so, the time to do it is long gone,
and "principles of finality associated with
habeas corpus actions apply with at least equal
force when a defendant seeks to attack a previous
conviction used for sentencing." 

.

  The linchpin of Ryan’s argument is a belief
that only a valid conviction can justify an
increase in one’s sentence. Not so. Ryan argues
as if he were in custody once again for the armed
robbery, but he isn’t. Recidivist sentencing is
not a second or deferred punishment for the prior
offense. It is a way to determine the appropriate
punishment for the latest crime. Witte v. United
States, 

(1995). And there is no
doubt that Ryan’s 1995 convictions for drug
offenses are valid and supply adequate grounds
for his current incarceration. (The maximum
punishment Ryan faced in 1995 was 40 years’
imprisonment. 21 U.S.C. sec.841(b)(1)(B)(vii).)
When imposing sentence judges may consider acts
that did not lead to a conviction. E.g., McMillan
v. Pennsylvania, 

(1986). The
Sentencing Guidelines contain elaborate rules
about "relevant conduct" that counts against the
defendant whether or not it has been the subject
of a conviction. Edwards v. United States, 

(1998). Indeed, a judge may take account
of prior criminal conduct even though a jury has
acquitted the defendant of charges based on the
events. United States v. Watts, 

(1997). And the Court added in Nichols v. United
States, 

(1994), that even an
uncounseled conviction may be the basis for
recidivist sentencing under the Guidelines, when
the sixth amendment did not require counsel
(because the original crime was a misdemeanor
that did not lead to incarceration).

  Many events that have been determined only by a
preponderance of the evidence, and without the
safeguards of a criminal trial, lead to higher
sentences. Cases such as Edwards, Watts, Witte,
and Nichols just instantiate this more general
proposition. Consider a few more examples. (1)
Stationhouse confessions are good grounds on
which to enhance a sentence, even though
confessions standing alone do not support
convictions. (2) A person deported after informal
proceedings faces a higher sentence for entering
the United States improperly than does a person
who has never been deported. (3) A person who
cuts someone else’s hair without a license may be
punished as a criminal, even though it is
possible to deny a license application without
proof beyond a reasonable doubt. Once we see that
acts proved beyond a reasonable doubt after a
full-dress criminal trial are not the only
acceptable grounds for enhancement, it looks very
odd to invest substantial resources determining
whether a particular conviction offered in
sentencing indeed meets the standards developed
for full-dress litigation. Why should a judgment
of conviction based on a guilty plea (that is, on
a confession in open court) be less legitimate,
as a ground for enhancement, than a stationhouse
confession? No one would suppose that a lawyer’s
demand for a fee puts a stationhouse confession
off limits for sentencing purposes; should the
confession in open court to armed robbery be the
less usable when ascertaining the appropriate
sentence for Ryan’s drug crimes? Both a
stationhouse confession to armed robbery and the
1980 guilty plea show that Ryan is incorrigible--
15 years after committing an armed robbery, and
despite serving time for that offense, Ryan is
still an active criminal. That knowledge
justifies a more severe sentence to achieve both
deterrence and incapacitation.

  Custis left open the question whether a person
who has his conviction set aside by the rendering
court--say, by a writ of error coram nobis--is
entitled to reconsideration of a federal
recidivist 

. Several
courts have held or assumed that the answer is
yes. E.g., United States v. Pettiford, 

(1st Cir. 1996); United States v. Bacon, 

, 162 n.3 (4th Cir. 1996); Turner v.
United States, 

(6th Cir. 1999). We
need not address that question, however, because
Ryan has not persuaded Illinois to annul his
conviction for armed robbery. He had ample chance
but did not pursue his avenues. Today is too
late. Even had Judge Bucklo concluded that Ryan’s
federal custody authorizes a collateral challenge
to the state conviction under sec.2254 this would
not have done Ryan any good. A person seeking
federal relief against a state conviction must
fairly present his contentions to state court.
O’Sullivan v. Boerckel, 

(1999). Ryan
never did, and Illinois would deem untimely any
effort to do so at this remove. 725 ILCS 5/122-
1(c). Like Custis itself, our decision in Tredway
v. Farley, 

(7th Cir. 1994), which
treated the state’s timeliness rules as defects
in state remedial processes that entitle a
federal court to disregard a person’s failure to
present the claims to the state, was rendered
when federal law did not impose time limits on
collateral attacks. That changed in 1996, with
the Antiterrorism and Effective Death Penalty
Act. Now federal law includes a one-year period
of limitations, for collateral attacks against
both state and federal convictions. It is no
longer possible to conclude, as Tredway did in
1994, that time limits under state law excuse
convicted persons from the duty to present their
claims to state court. See Edwards v. Carpenter,

(2000) (failure to present a
claim to a state court within the state’s
timeliness rules is a default that precludes
collateral review in federal court unless both
cause and prejudice are established). By
disdaining the Illinois courts when he had the
chance, Ryan lost any entitlement to obtain a
federal declaration that his 1980 conviction is
invalid, for he does not argue that any "cause"
(which is to say, an impediment created by the
state) excuses the delay. See also, e.g., Freeman
v. Page, 

(7th Cir. 2000).

  Defendants have ample reason to challenge their
convictions on direct appeal, or collaterally
while serving their sentences. The Sentencing
Commission’s approach, like that of the Armed
Career Criminal Act, relies on this incentive.
Serious challenges are likely to be brought, and
resolved, before the sentencing for a later
offense. Convicts who wait too long can try coram
nobis, which is available in many states. This is
what happened in United States v. Tucker, 

(1972): the defendant obtained writs of
error coram nobis, which justified resentencing
on the current conviction. See also Johnson v.
Mississippi, 

(1988). Any convicted
person has ample opportunities to obtain review.
Requiring defendants to use these opportunities,
rather than tarry and then launch indirect
collateral attacks during or after sentencing for
some other offense, has significant benefits. It
sends persons to the rendering courts, which have
the records necessary to determine whether a
conviction is valid or not. It requires them to
act promptly, while the information necessary to
determine validity is available (and while
reprosecution is possible, at least in theory, if
the conviction is flawed).

  Honoring judgments that remain outstanding after
full opportunity for direct and collateral review
does not dis-honor the constitutional claims the
defendant wishes to make. It simply establishes
rules for presenting these claims to the right
court, and in a timely fashion. Respecting
judgments is the norm in our legal culture: the
full faith and credit clause and common law
principles of res judicata combine to make
respect for judgments the rule. During this
century courts began to readjudicate issues that
were, or could have been, presented to the
rendering court. E.g., Frank v. Mangum, 

(1915); Waley v. Johnston, 

(1942); Brown v. Allen, 

(1953). This
development depends on a federal statute, not on
constitutional right. See Lindh v. Murphy, 

, 871-74 (7th Cir. 1996) (en banc),
reversed on other grounds, 

(1997).
See also Williams v. Taylor, 

(2000). No statute expressly authorizes the form
of derivative collateral attack that Ryan wants
to wage, and the changes made by the AEDPA show
that long-deferred challenges that were never
presented to the state courts are no longer
appropriate grounds of federal relief. Whether
some avenue remains open to Ryan in Illinois is
a question we pretermit, along with the question
(the one reserved at the end of Custis) whether
a writ of error coram nobis would justify relief
under sec.2255. Because Ryan’s 1980 conviction
remains on the books to this day, it is not
possible to conclude, in the language of sec.2255
para.1, that his federal sentence "was in excess
of the maximum authorized by law, or is otherwise
subject to collateral attack". The sentence is
within the statutory limit, was altogether proper
under Custis, and therefore may not now be upset.
Affirmed



  DIANE P. WOOD, Circuit Judge, dissenting. Once
again, we are called upon to unravel the
complexities of collateral attacks on federal
sentences. That task, which is hard enough when
the question is what kinds of defects in the
original proceeding are serious enough to entitle
a prisoner to have a conviction or sentence set
aside, becomes even more difficult when the
attack indirectly calls into question the results
of earlier proceedings. Our case is such a case.
We must decide here whether there is any way at
all that a federal prisoner can challenge a
federal sentence that was based in part on a
void, unconstitutional prior state conviction.
The majority concludes that there is nothing such
a prisoner can do, even in a state like Illinois
that affords absolutely no way for an individual
who has finished serving the unconstitutional
prior sentence to erase it from the books. I
disagree, and I therefore respectfully dissent.
  John Ryan pleaded guilty in 1980 to state
charges of armed robbery before the Circuit Court
of Cook County, Illinois. He was sentenced to six
years’ imprisonment for that offense. He took no
appeal, nor did he file a state post-conviction
petition; instead, he served out the sentence to
its completion. Only later did it turn out that
his guilty plea may have been procured through
the most serious form of attorney misconduct.
Ryan claims that on the day of trial, his lawyer
demanded a payment of $2,000 as the price of
going forward. Ryan didn’t have the money, and so
he pleaded guilty instead. The lawyer in question
was later disbarred by the Illinois Supreme
Court. In 1987, the court accepted a voluntary
disbarment based on six pending proceedings in
which he was charged with unethical conduct
toward clients other than Ryan. The charges
included lying to a client by telling him that
the appellate court had affirmed his conviction,
when the lawyer had never appealed the case at
all; lying to another pair of clients by telling
them he had incorporated their business when he
had not done so, and the lie eventually forced
them into bankruptcy; and using client escrow
accounts to hide personal funds from the Internal
Revenue Service. Ryan claims that the lawyer’s
effort to shake him down on the morning of the
trial was one more item on this inglorious list.
If Ryan’s allegation proves to be true, then it
would be compelling evidence that he had received
constitutionally ineffective counsel in
conjunction with his 1980 conviction.

  Years later, Ryan was foolish enough to commit
another offense, this time the federal controlled
substance crime to which he pleaded guilty on
September 20, 1993, in United States v. Ryan, No.
93-CR-419-1, N.D. Illinois. For that crime, Ryan
received a sentence of 185 months plus five
years’ supervised release. He started with an
offense level of 26 and a criminal history
category of VI. Six levels were then added
pursuant to the career offender provision,
U.S.S.G. sec. 4B1.1, based on the 1980 Illinois
conviction just described and one other
conviction that he does not challenge. Last, two
levels were deducted for acceptance of
responsibility, see sec. 3E1.1(b), giving a final
offense level of 30 and a range of 168 to 210
months. Had it not been for the 1980 Illinois
conviction, the career offender enhancement would
not have been proper; and without those extra 6
levels, the final offense level would have been
24 and the range 100 to 125 months. It is
therefore obvious that Ryan and anyone similarly
situated has a substantial stake in being able to
correct the kind of problem he had in the 1980
conviction, in some court, at some time.
  The majority holds that Ryan is out of luck,
unless the state offers a way to vacate an
unconstitutional prior conviction for which the
entire sentence has been served. It believes that
its result is compelled by Custis v. United
States, 

(1994). But the Court had no
such problem before it in Custis. We must decide
now whether the principles underlying Custis
apply to this significantly different fact
pattern.

  The first question is whether Ryan should be
proceeding under 28 U.S.C. sec. 2254, the statute
normally used to attack state convictions and
sentences, or 28 U.S.C. sec. 2255, the
counterpart statute for federal convictions and
sentences. I agree with the district court that
the correct vehicle was sec. 2255. This follows
from Maleng v. Cook, 

(1989), in
which the Court held that when a state prisoner
seeks to challenge a prior state conviction via
sec. 2254, the custody that matters is the
current one, not any custody that resulted from
the prior conviction: "[T]he habeas petitioner
[is] ’in custody’ under the conviction or
sentence under attack at the time his petition is
filed." 

(emphasis added), citing
Carafas v. LaVallee, 

, 238 (1968).
See also Crank v. Duckworth, 

, 1091
(7th Cir. 1990) ("Maleng holds that when sentence
A has expired but has been used to augment
sentence B, the prisoner is ’in custody’ only on
sentence B."). Although neither the Supreme Court
nor this court has explicitly ruled on this point
before, Maleng makes clear that the district
court did not have jurisdiction to consider the
validity of Ryan’s expired Illinois sentence
under sec. 2254. Ryan has not been "in custody"
of the State of Illinois under that sentence for
many years. Neither is he under "present
restraint" pursuant to that conviction. 

. Therefore, he does not meet the
statutory requirements to seek review of that
sentence under sec. 2254. This is the first
important contrast with the situation facing the
Court in Custis, where the Court found that the
petitioner could attack the prior state
conviction through sec. 2254 because he was still
in state custody at the time of his federal
sentencing. The Custis Court had no occasion to
consider what might happen if neither sec. 2254
nor any other remedy was still available.

  Following Maleng, we have held that when a
sentence has been enhanced based on an allegedly
unlawful or factually erroneous conviction, a
prisoner may bring a habeas corpus petition in
the court with jurisdiction to grant release from
the enhanced sentence, or reduction of that
sentence. Lowery v. Young, 

, 1312-13
(7th Cir. 1989). Lowery involved a different
permutation of today’s problem, where a state
prisoner in custody under one state sentence
claimed that the sentence had been
unconstitutionally enhanced by prior, unlawful
convictions of another state. We concluded that
the district court had jurisdiction to consider
his claims, even though the prisoner was in
custody only of the state which had imposed the
enhanced sentence. Indeed, the only court with
jurisdiction to grant his requested relief--an
earlier release from his current custodian--was
the court with jurisdiction over that 

. See also 

.

  The same logic applies here. Ryan is a federal
prisoner "in custody" under a federal sentence.
His complaint pertains to the length of that
sentence, which means that the statute he can and
must use is sec. 2255. Maleng rules out sec.
2254, and the fact that the attack relates to the
sentence rules out 28 U.S.C. sec. 2241. This
conclusion follows from a number of our prior
decisions. See, e.g., Tredway v. Farley, 

, 292 (7th Cir. 1994) ("Because a person
currently serving a sentence that was enhanced on
the basis of a prior conviction is still in
custody, he may challenge the enhancing
conviction as constitutionally invalid even
though the prior conviction’s original custodial
term has expired."), quoting Smith v. Farley, 

, 1365-66 (7th Cir. 1994); 

("Whether the federal court with
jurisdiction over the custodian holding the
prisoner on [the enhanced] sentence B may inquire
into the validity of [the prior] sentence A is a
matter of comity and the rules of preclusion, not
of ’custody’."). See also United States v.
Tucker, 

(1972) (granting relief
under sec. 2255 to petitioner whose federal
sentence was enhanced based on prior,
unconstitutional state conviction); United States
v. Clark, 

, 364 (5th Cir. 2000)
(collecting cases).

  Custis itself implies that the "in custody"
requirement of sec. 2255 is met where a federal
sentence has been enhanced by a prior state
conviction. In stating that "[i]f Custis is
successful in attacking these state sentences, he
may then apply for reopening of any federal
sentence enhanced by the state 

, the Court acknowledged that the
proper vehicle to attack the enhanced sentence is
sec. 2255. Several circuits have likewise held
that a federal prisoner may bring a sec. 2255
motion after successfully attacking a prior state
conviction. See Turner v. United States, 

, 477 (6th Cir. 1999); United States v. Bacon,


, 161-62 & n.3 (4th Cir. 1996); United
States v. Cox, 

, 339 (10th Cir. 1996);
United States v. Nichols, 

, 38 (5th
Cir. 1994). The only wrinkle presented by this
case is whether the federal prisoner is entitled
to seek correction of his federal sentence if he
has not, for whatever reason, already launched a
successful collateral attack on his state
conviction.

  The majority’s principal reason for concluding
that no such entitlement exists is the fact that
Ryan’s 1980 conviction was literally on the books
at the time of his federal sentencing. It argues
that because there was a conviction, the decision
to enhance his federal sentence under sec. 4B1.1
was automatically "correct" and the sentence is
therefore immune from challenge under sec. 2255.
Custis, however, did not take such a strict
approach. Nothing in Custis implied that the
earlier state conviction used to enhance Custis’s
sentence was legally correct. Nor did the Court
suggest that the possibility that Custis’s prior
conviction was tainted with constitutional error
was irrelevant. Instead, the Court focused on the
language of the Armed Career Criminal Act
("ACCA") to resolve the narrow question whether
that statute provided persons sentenced under it
an opportunity to attack the validity of their
prior convictions during sentencing under that
statute. The Court explained: "The [ACCA] focuses
on the fact of the conviction and nothing
suggests that the prior final conviction may be
subject to collateral attack for potential
constitutional errors before it may be counted
[for 

. The Court
came to a similar conclusion in Lewis v. United
States, 

(1980). Construing a
predecessor to the current felon-in-possession-
of-a-firearm statute, 18 U.S.C.A. sec. 1202(a)(1)
(Supp. 1982), the Court held that the statute was
not limited to felons whose convictions were
subject to possible collateral attack but rather
applied to all persons with a felony conviction
on the 

.

  What the Court focused on in both of these
cases was whether the applicable recidivist
statute provided an opportunity for persons
subject to the statute to expunge their past in
the same sentencing proceeding in which the trial
court was to determine whether the individual was
subject to additional sanctions and penalties
based on those past convictions. Some statutes
allow people with a criminal history to attack
the validity of their prior convictions before
being subjected to higher penalties, see 18
U.S.C. sec. 3575(e) (dangerous special offender),
21 U.S.C. sec. 851(c)(2) (recidivism under the
Comprehensive Drug Abuse Prevention and Control
Act of 1970), while others do not, see 18 U.S.C.
sec. 924(e) (sentencing for felons in possession
of a firearm).

  My colleagues construe the criminal history
provisions of the Sentencing Guidelines, U.S.S.G.
sec. 4A1.2, in the same manner as the Supreme
Court construed the ACCA in Custis, to mean that
convictions that have not yet been expunged or
vacated may be used to enhance a sentence without
providing the person being sentenced with an
opportunity to attack the validity of those
convictions during the original sentencing
proceedings. Given the language of the criminal
history Guideline, this much of its
interpretation makes sense. Naturally, this
language also implies (consistently with Custis)
that a conviction that has previously been
expunged or vacated may not be counted.

  The question of use in the original sentencing
proceeding is analytically different from the
question of the permissible scope of a collateral
attack. The Guidelines speak indirectly to the
question of collateral attacks, and what they say
is more open to the possibility of such an attack
on an earlier, allegedly invalid, sentence than
the majority’s opinion admits. Section 4A1.2,
Application Note 6, expressly states that, "with
respect to the current sentencing proceedings,"
the Guidelines do not "confer any right to attack
collaterally a prior conviction or sentence
beyond any such rights otherwise recognized in
law." (Emphasis added). The majority reads this
language as if the sentence ended just before the
phrase I have highlighted. (Obviously the U.S.
Sentencing Guidelines are not enacting laws for
the states, and so the possibility that a state
might provide an avenue for a challenge does not
change matters. Such an avenue would not be
something "conferred" by the Guidelines.) There
is no hint in that passage that if the conviction
is on the books at the time of sentencing, any
sentence enhancement based on it is thereafter
immune from challenge. The Sentencing Commission
was saying only that the Guidelines themselves
were not a source of law for possible collateral
attack--hardly a controversial proposition.

  The question then is what kind of collateral
attack on a prior state sentence used for
enhancing purposes under the Guidelines is
"otherwise recognized in law". In my opinion, the
answer for a federal prisoner is the procedure
established in sec. 2255. This conclusion does
not take away with the left hand what the Supreme
Court gave the district courts with the right
hand in Custis: the freedom not to explore prior
convictions. That would be true only if exactly
the same grounds, under exactly the same
procedures, applied to sec. 2255 cases as apply
to original sentencing proceedings, and that is
plainly not the case. Even if a defendant may not
collaterally attack her prior sentence in the
original sentencing proceeding (except on narrow
grounds not available to Ryan), it does not
necessarily follow that such a prior sentence is
immune from all attack, on any ground, at any
time. That is the key point at which I part
company with the majority.

  Both the Supreme Court and the Sentencing
Commission have concluded that there is much to
be lost and little to be gained by clogging up
the federal district courts with collateral
attacks on prior convictions during sentencing.
One could imagine an array of complaints ranging
from the sufficiency of indictments, to
evidentiary errors at the state proceeding, to
sentencing complaints. Such attacks would bog
down the sentencing process, slowing down the
operation of the courts and clogging their
already crowded dockets. (In fact, this was part
of the Custis Court’s rationale for allowing a
collateral attack during sentencing proceedings
for violations of the right to counsel recognized
in Gideon v. Wainwright, 

(1963):
lack of counsel is one of the few constitutional
errors which will be readily apparent and can be
discovered without poring over the record of the
earlier proceedings. 

.)
Collateral attacks, in contrast, are available
only on grounds far narrower than those that can
be raised in the original case (either at
sentencing or on appeal). Well established
principles governing sec. 2255 proceedings would
leave the door open to that far smaller universe
of complaints: those that assert substantial
federal constitutional flaws with the original
state conviction. The majority overlooks this
fact when it presumes that the two stages would
be identical.

  If sentencing hearings were the last word,
always, then there would be no occasion ever to
use sec. 2255 to attack a sentence. But that is
not the system we have. To the contrary,
countless cases hold that a petitioner can attack
a federal sentence for constitutional error. See,
e.g., United States v. Addonizio, 

,
185 (1979); Waley v. Johnston, 

, 104-
05 (1942). And a subset of those cases involve
challenges to prior sentences that were used to
enhance the federal sentence once a prior
(invalid) conviction has been expunged, reversed,
or invalidated. See, e.g., 

; 

. Applying the same rule to a previously
unexamined, and presently unexaminable, prior
sentence would be no different and no more
disruptive than holding a hearing to re-examine
the conviction for which the prisoner is serving
time--an activity that is routine for the federal
courts. Complaints about ineffective assistance
of counsel are regularly thrown over to
collateral attack proceedings, precisely because
the original trial record almost never contains
the necessary information to evaluate the claim.
See, e.g., United States v. Cooke, 

,
1299 (7th Cir. 1997). The fact that a new point
can be raised on collateral attack that was
unavailable at the original proceeding in no way
undermines the significance or importance of the
original proceeding, nor would it do so in the
present context.

  The Sentencing Commission itself was concerned
with the lawfulness of prior convictions used for
enhancement purposes. The Sentencing Commission
did not say that any conviction--constitutionally
infirm, factually erroneous, or not--can count
for sentencing purposes. Instead, it carved out
an exception for convictions vacated because of
errors of law or fact or invalidated because of
constitutional error. The majority claims that
the Commission meant to exclude only those
convictions invalidated prior to federal
sentencing; the result, according to the
majority’s view, is that if the sentencing court
took into account any conviction not yet
invalidated at the time the federal sentence was
imposed, the sentence is presumptively lawful and
immune from collateral attack under sec. 2255.

  That conclusion, however, simply restates the
question, which is what to do in cases like
Ryan’s, where collateral attack is not possible
either before the federal sentencing or at the
time of the sec. 2255 petition. In hindsight, one
assumes, Ryan probably now wishes that he had
been more litigious after he pleaded guilty to
the 1980 state charges, but he was not. It is now
too late for him to raise any challenge in the
Illinois courts to that plea. See 725 ILCS 5/122-
1 (providing petition for post-conviction relief
must be filed within three years from the date of
conviction, at the latest). The majority holds
out hope for Ryan and those in his position by
speculating that some form of post-hoc relief may
be available through a writ of error coram nobis.
I am not nearly so sanguine. The common law writ
was abolished in Illinois in 1871 by the Illinois
legislature. See Ill. Rev. Stat. Ch. 83, sec. 154
(Gross 1871); People v. Touhy, 

, 830
(Ill. 1947) (recounting history and abolition of
writ). It has since been replaced with a
statutory provision which, like the state post-
conviction relief statute, has strict deadlines
for seeking review of a sentence. In Illinois a
petition for statutory coram nobis must be filed
within two years of the original judgment of
conviction. See 735 ILCS 5/2-1401 (statutory
coram nobis). And, because imprisonment does not
constitute legal disability or duress, the two-
year filing period is not tolled during the time
of confinement. See Williams v. People, 

, 469 (Ill. 1964); Morgan v. People,

, 26 (Ill. 1959). Thus, Ryan (and
others like him whose state sentences have long
since expired) has no means through which he can
attack collaterally his state conviction and, if
successful, seek to have his federal sentence re-
opened.

  I would find under these circumstances that
sec. 2255 remains available to such a prisoner to
serve its classic function: to allow an attack on
his federal sentence on the ground that all or
part of it was tainted with constitutional error.
Not every prisoner will be able to allege the
kind of error in a prior state conviction that
would give rise to a right to a hearing in this
kind of case. Indeed, most will not, both because
the state courts do not commit those kinds of
errors on a regular basis, and because
allegations of ineffective assistance of counsel
that are sufficiently serious to require a
hearing are rare. But some will. Under the
majority’s rule, even a state prisoner who had
pleaded guilty to earlier charges after 90 days
in a torture chamber would still be required to
serve the extra time for his federal sentence. I
attach no special significance to the fact that
the state may no longer offer a remedy to correct
such a sentence, after some reasonable time
period has passed. Once the sentence has been
served, the state may take the position (as
Illinois has) that it does not wish to use its
scarce judicial resources for essentially a
record-correction function. The person with a
stake in the matter is the new federal prisoner,
like Ryan, and the system that must ensure it is
not relying on an unconstitutional foundation to
justify imprisonment is the federal system.

  The only task remaining is to apply this rule
to Ryan. His allegations, backed up as they are
by the deplorable later history of the lawyer who
was allegedly representing him in the 1980
Illinois proceeding, are more than enough to
justify a hearing on his petition under sec.
2255. I would reverse the decision below and
remand his case to the district court for further
proceedings.

 I respectfully dissent.