United States v. Scott A. Fountain

840 F.2d 509

• Court: Court of Appeals for the Seventh Circuit
• Decided: May 16, 1988
• Docket: 86-2622, 87-1465
• Status: Published

Opinion

EASTERBROOK, Circuit Judge.

On January 29, 1984, Matthew Granger slew Boyd Spikerman, a guard at a federal prison. Prison officials found Spikerman in a pool of his own blood, stabbed more than a dozen times. Repeated application of a fire extinguisher to Spikerman’s head finished the job. Granger, lounging nearby, maintained that he “heard voices and didn’t mean to do it.” He didn’t say whose voices. Prison officials soon concluded that the voices came from other inmates, and that Granger certainly meant “to do it” — for he had acquired the knife the night before and planned ahead enough to arrange for other prisoners as lookouts.

An indictment charged Granger, Scott A. Fountain, and a third prisoner with first-degree murder of a federal employee, 18 U.S. C. §§ 1111 and 1114, and conspiracy to commit murder, 18 U.S.C. § 1117. According to the indictment, Fountain supplied the knife and was a lookout. Granger and Fountain pleaded guilty on the same day in a joint proceeding under Fed.R.Crim.P. 11. We consolidated the ensuing appeals. We affirmed Granger’s conviction but vacated Fountain’s, holding that the record did not contain an adequate factual basis for his plea. United States v. Fountain, 777 F.2d 351, 356-57 (7th Cir.1985). The mandate in Fountain’s case issued on January 14,1986. The Speedy Trial Act provides that the trial must commence “within seventy days from the date the action occasioning the retrial becomes final”. 18 U.S.C. § 3161(e). The record did not accompany the mandate; it stayed here because Granger had filen a motion for extension of time to file a petition for rehearing. Before the time to seek rehearing expired, Granger filed a petition for certiorari, which automatically stayed issuance of the mandate. Fed.R.App.P. 41(b). The mandate in Granger’s case finally issued on March 28, 1986, and the record went with it. Only then did the district court recognize that Fountain was entitled to plead anew — but by then the 70 days were gone.

The court re-arraigned Fountain on April 10, 1986, and eight days later he invoked his rights under the Speedy Trial Act, see 18 U.S.C. § 3162. On June 4, 1986, the district court dismissed the 1984 indictment without prejudice. A grand jury returned a fresh indictment the same day; Fountain pleaded not guilty that afternoon. He was later tried, convicted on both counts, and sentenced to life plus 150 years. The evidence at trial showed that Fountain held Spikerman while Granger attacked him. The life sentence for first-degree murder was mandatory under § 1111. The conspiracy statute, § 1117, permitted the court to order imprisonment “for any term of years or for life.” The court selected 150 years and added that Fountain would not be eligible for parole for 50 years, invoking 18 U.S.C. § 4205(b)(1).

Fountain challenges every step of this. He insists that the 1984 indictment should have been dismissed with prejudice, that the 1986 indictment was obtained improperly, that the trial was defective, and that the sentence is illegal. We consider his contentions in that order.

I

Fountain argues with some force that the only effect of the Speedy Trial Act in his case was to postpone his trial. Once everyone realized the problem in late March or early April 1986, the court could have given Fountain a prompt trial. But Fountain’s motion to dismiss took the case off the trial calendar. The dismissal, rein-dictment, and plea in close order on June 4, 1986, simply started a new 70-day period. And as things turned out, motions challenging the procedures used to obtain the 1986 indictment postponed the trial to August 11, 1986. The trial thus occurred 209 days after our mandate issued, more time than would have elapsed if Fountain had accepted the violation of the Speedy Trial Act stoically, indeed probably more time than would have elapsed if there were no Speedy Trial Act. As Fountain sees things, the violation of the Act was at least reckless, and the “remedy” an insult. Only dismissal with prejudice would do, he contends.

One may reply that the new plea on April 10 was taken no later than if we had held the mandate until the final resolution of Granger’s case. If Granger had obtained relief from the order affirming his conviction, any further proceedings in Granger’s case should have been linked to Fountain’s. If the prosecutor had recognized the significance of’ the staggered issuance of the mandates, the district court doubtless would have granted an exclusion under 18 U.S.C. § 3161(h)(7) (exclusion of a “reasonable period ... when the defendant is joined for trial with a codefendant as to whom the time for trial has not run”). Section 3161(e), starting the 70 days on “the date the action occasioning the retrial becomes final”, creates problems when an action is “final” only in retrospect. The Solicitor General had 60 days to file a petition for a writ of certiorari to review our decision reversing Fountain’s conviction, and was entitled to ask for 30 more, see Supreme Court Rule 20.1. McDonald v. Schweiker, 726 F.2d 311 (7th Cir.1983), holds that a judgment is not “final”, starting the 30 days in which to seek attorneys’ fees under the Equal Access to Justice Act, until the time to appeal has expired or an appeal, once filed, has been withdrawn; perhaps a similar approach should be employed under § 3161(e). But the prosecutor has not made this argument, and we do not decide cases on the basis of arguments not made and steps not taken. The delay violated the Speedy Trial Act, we must assume. “What might have been” still is of interest when deciding whether the district court lacks discretion to dismiss the indictment without prejudice.

Section 3162(a)(2) allows the district court to dismiss with or without prejudice. The statute lists three things, “among others”, for the court’s consideration: “the seriousness of the offense; the facts and circumstances of the case which led to the dismissal; and the impact of a reprosecution on the administration of this chapter and on the administration of justice.” An open-ended list of this sort imbues a court with great discretion. When a court must balance incommensurables, when the factors do not have weights, it is hard to identify “error”. Such a statute calls for the exercise of sound judgment above all, and appellate review is limited to ensuring that the district court made an informed choice. United States v. Beasley, 809 F.2d 1273, 1278-79 (7th Cir.1987); Wisconsin Real Estate Investment Trust v. Weinstein, 781 F.2d 589, 597-99 (7th Cir.1986); Metlyn Realty Cory. v. Esmark, Inc., 763 F.2d 826, 831-32 (7th Cir.1985). All this means that under the Speedy Trial Act, as so often when Congress prescribes methods rather than outcomes, review on appeal is deferential. E.g., United States v. Janik, 723 F.2d 537, 546 (7th Cir.1983). See also, e.g., United States v. Salgado-Hernandez, 790 F.2d 1265, 1267 (5th Cir.1986). One court has held that a “lackadaisical” attitude by prosecutors requires dismissal with prejudice notwithstanding other considerations. United States v. Taylor, 821 F.2d 1377, 1385-86 & n. 3 (9th Cir.1987), cert. granted, — U.S. -, 108 S.Ct. 747, 98 L.Ed.2d 760 (1988). Other courts have held that district judges must consider all' of the statutory desiderata, e.g., United States v. Melguizo, 824 F.2d 370 (5th Cir.1987), a position in accord with the practice in our circuit. We therefore inquire not whether the prosecution was careless (it was), but whether the district judge made a reasoned decision in light of the statutory criteria.

The district court looked carefully at the facts of the case and the factors in the statutory list. The court observed that first-degree murder is a grave offense, which calls for dismissal without prejudice. The “penalty” imposed on the prosecutor (and the rest of society) by dismissing with prejudice rises with the seriousness of the crime. Dismissing a misdemeanor with prejudice promotes enforcement of the Speedy Trial Act at smaller cost to potential victims of crime than does dismissing a murder case. Important as it is to enforce the Speedy Trial Act, it is also important to deter murder and punish murderers. Certainly the murder was a more serious offense than the violation of the Speedy Trial Act. As for “facts and circumstances”: the violation was inadvertent, and Fountain did not assert his right to a speedy trial in time to alert the prosecutor or the court to the problem. Perhaps he thought delay advantageous. Delay injures the prosecution, after all, as memories fade. The prosecutor bears a heavy burden of persuasion, and the degradation of evidence generally cuts against the party with the burden. A defendant who waits passively while the time runs has less claim to dismissal with prejudice than does a defendant who demands, but does not receive, prompt attention. See United States v. Regilio, 669 F.2d 1169, 1172 (7th Cir.1981).

Fountain pins his hopes on a claim that the prosecutor “exploited” the delay by gathering new evidence from four witnesses. The facts do not support the contention. The FBI interviewed two of these four in January 1986, while there was still plenty of time. The other two witnesses testified that during April 1986 Fountain boasted to them of his role in killing Spik-erman. This evidence, at first glance, is a fruit of forbidden delay. Yet consider the circumstances. Fountain made his admissions to prisoners in the Dane County Jail in Madison, Wisconsin. After the murder Fountain was consigned to the Control Unit of Marion, the nation’s most secure prison. He returned to Wisconsin in April 1986 for arraignment and trial. There he met a new group to impress with his exploits. The fact that he started talking on entering a new prison is more significant than the April 1986 date; he would have had the same reason and opportunity if he had been returned in January 1986. We do not think that delay worked to Fountain’s detriment. The district court was entitled to dismiss the indictment without prejudice.

That said, we are dismayed by the way this case proceeded and encourage the United States Attorney and the district court to prevent repetitions. Fountain did not have counsel between the issuance of our mandate and his arraignment on April 10, 1986. His original lawyer had withdrawn, and new counsel were appointed for the appeal in 1985. The appellate lawyers understood their appointment as for appellate purposes only, a plausible construction given the way in which this court recruits lawyers to represent indigent defendants. We obtain great assistance from young lawyers whose education gives them more appellate than trial skills, and who are not necessarily located in the state (or members of the bar of the court) in which proceedings on remand would be conducted. We could not induce (or in good conscience recruit) such lawyers to represent indigent defendants if the appointment entailed an obligation to continue the representation in the district court. Circuit Rule 4 refers to these appointments as “appellate appointments.” A district court therefore needs—and the Western District of Wisconsin lacks—a method for providing counsel promptly after receiving the mandate in a case like Fountain’s. A lawyer might have prevented the problems Fountain encountered.

There is, moreover, no excuse for the prosecution’s neglect of this case. The explanation—that it slipped through the cracks while the record lingered in our court—suggests that the United States Attorney’s office relies on memory rather than a tickler system to ensure compliance with the Speedy Trial Act. Not having a method to log and track cases on remand from this court is a managerial decision, one predictable consequence of which is that some cases will not be tried on time. If the problem in our case recurs, it will not be so easy to chalk it up to inadvertence.

II

Fountain asked the district court to dismiss the 1986 indictment on three grounds: that the prosecutor used perjured testimony, that the prosecutor neglected to present exculpatory testimony, and that the entire presentation was hearsay.

The supposed perjury was the testimony of Jeffrey Logie, an inmate who appeared before the grand jury in 1984 and inculpated Fountain; the prosecutor gave the 1986 grand jury a copy of Logie’s testimony. Michael Stotts, another inmate, has asserted that Logie later told him that Logie’s testimony to the 1984 grand jury was false. Thus the charge of perjury. This circuit has reserved the possibility that knowing use of perjured testimony, amounting to the creation of trumped-up charges, might justify' the dismissal of an indictment. United States v. Roth, 777 F.2d 1200 (7th Cir.1985) (discussing earlier cases); United States v. Thomas, 788 F.2d 1250 (7th Cir.1986). We have not, however, ordered any indictment dismissed on this ground, and therefore have not decided whether to convert these ruminations into a holding. We do not do so here, either, because the use of Logie’s testimony is not the “knowing” use of perjury. Logie testified at the trial of Randall Dahlin — accused of being one of Granger’s accomplices — and denied recanting. Logie’s testimony at Dahlin’s trial in December 1984 is consistent with the grand jury testimony taken in 1984 and reused in 1986. Logie gave consistent testimony at Fountain’s trial in 1986, and the jury evidently believed it. We therefore do not have knowing use of perjury; we do not even have perjury; we have only a (denied) claim of recantation, a depressingly familiar one in any criminal trial involving inmate witnesses. The prosecutor is entitled to present such testimony to the grand jury for whatever it is worth.

The other challenges to the indictment are irrelevant once the petit jury convicted Fountain. We may assume that despite cases such as United States v. Calandra, 414 U.S. 338, 345, 94 S.Ct. 613, 618, 38 L.Ed.2d 561 (1974), Costello v. United States, 350 U.S. 359, 362-63, 76 S.Ct. 406, 408-09, 100 L.Ed. 397 (1956), and United States v. Murphy, 768 F.2d 1518, 1533-34 (7th Cir.1985), a district judge in an exceptional case may conclude that the prosecutor’s use of hearsay (misleading the grand jury about the quality of the evidence) and omission to present exculpatory evidence (misleading the grand jury about the weight of the evidence) requires the government to present its case anew. The function of the grand jury is to screen out cases in which there is not even probable cause to believe a person committed the offense, and it is possible to imagine manipulations of the grand jury that prevent it from fulfilling this mission. 1 But the people harmed by such misdeeds are the innocent, those exposed to the travail and expense of trial even though they were fated to prevail in the end. Once a person is convicted by the petit jury, we may be confident that a full presentation to the grand jury would have ended in indictment. The trial is a full, adversarial presentation, in which the live witnesses and exculpatory evidence missing in the grand jury come to the fore. The grand jury acts by majority vote and uses a lesser burden; it is impossible to imagine evidence sufficient to produce a conviction at trial that would not also produce an indictment. So it would be silly to reverse a conviction on the ground that the evidence before the grand jury was insufficient. We know that this defendant is not a member of the class that is harmed by sloppy or overbearing conduct before the grand jury. We know that he could be reindicted in a trice (using the record of the trial as a basis), and would be tried anew in the same fashion.

These considerations led the Supreme Court to hold in United States v. Mechanic 475 U.S. 66, 106 S.Ct. 938, 89 L.Ed.2d 50 (1986), that a violation of Fed.R.Crim.P. 6(d) was not a good reason to reverse a conviction, even on the assumption that it would have been a good reason to dismiss the indictment before trial. The Court concluded, 475 U.S. at 70, 106 S.Ct. at 941 (footnote omitted):

The Rule protects against the danger that a defendant will be required to defend against a charge for which there is no probable cause to believe him guilty_ But the petit jury’s subsequent guilty verdict means not only that there was probable cause to believe that the defendants were guilty as charged, but also that they are in fact guilty as charged beycnd a reasonable doubt. Measured by the petit jury’s verdict, then, any error in the grand jury proceeding connected with the charging de- cisión was harmless beyond a reasonable doubt.

This holds true here. If the prosecutor erred in presenting evidence to the grand jury, that error was harmless. It is inappropriate to reverse the ensuing conviction.

Fountain dismisses Mechanik as a case concerned with a “technicality”. It is fair to apply this label to Rule 6(d); the violation was that two witnesses testified simultaneously, while the Rule regards anyone other than the prosecutor, a court reporter, and a single testifying witness an unauthorized person. The Court's reasoning, however, does not depend on the nature of the rule at hand; the Court assumed that the violation was sufficiently substantial to permit dismissal of the indictment. Me-chanik proceeds by identifying the purpose of the rule (to protect the innocent from being indicted) and then says that a rule with this purpose should not be enforced by reversing a conviction obtained after trial—because we know, as surely as courts “know” anything, that the convicted defendant is not a member of the class of beneficiaries of the rule. Most important legal rules, even constitutional ones, may be enforced only by their beneficiaries. One could say, for example, that permitting anyone aggrieved by the introduction of evidence to invoke the exclusionary rule would contribute to the enforcement of the Fourth Amendment, but the Court nonetheless limits the sanction of exclusion to the person whose privacy interest the search invaded. See Rawlings v. Kentucky, 448 U.S. 98, 104-06, 100 S.Ct. 2556, 2561-62, 65 L.Ed.2d 633 (1980); United States v. Payner, 447 U.S. 727, 731-33, 100 S.Ct. 2439, 2443-45, 65 L.Ed.2d 468 (1980).

We recognize that at least one court, United States v. Taylor, 798 F.2d 1337 (10th Cir.1986), had held that Mechanik is limited to “technical violations” before the grand jury. For the reasons we have given, we respectfully disagree with Taylor and instead follow United States v. LaRouche Campaign, 829 F.2d 250 (1st Cir.1987); United States v. Dederich, 825 F.2d 1317 (9th Cir.1987); and United States v. Benjamin, 812 F.2d 548 (9th Cir.1987), which apply Mechanik to rules that are designed to prevent the indictment of innocent persons. This is, after all, the ground on which Mechanik distinguished Vasquez v. Hillery, 474 U.S. 254, 106 S.Ct. 617, 88 L.Ed.2d 598 (1986), which allowed convicted defendants to challenge racial bias in the selection of grand jurors. The rule against racial bias serves functions other than the protection of the innocent; the rules (if these are rules) against presenting too much hearsay to a grand jury, and requiring the presentation of exculpatory evidence, serve exclusively to protect the innocent.

In saying this we take no position on a question that has divided the First and Ninth Circuits: whether a defendant’s inability to obtain reversal after conviction means that he is entitled to a pretrial appeal if the district court declines to dismiss the indictment. The Ninth Circuit answered “yes” and the First Circuit “no”. Today’s case does not present an occasion for addressing that question.

Ill

Fountain presents a number of claims based on the trial and its aftermath. These require only brief discussion.

1. The indictment charged Fountain with both murder and conspiracy to commit murder. To the extent it described the facts, it implied that Granger stabbed Spik-erman while Fountain supplied the knife and acted as lookout. At trial, however, three witnesses testified that Fountain had admitted or boasted to them about restraining Spikerman while Granger flailed away. Fountain now describes this as an “amendment” of the indictment requiring its dismissal. See United States v. Miller, 471 U.S. 130, 105 S.Ct. 1811, 85 L.Ed.2d 99 (1985) (collecting' authorities).

This contention was forfeited at trial. When the witnesses testified to Fountain’s role in the attack, he did not object. Fountain did not raise this question even after trial. Had he objected, the district judge might have excluded the testimony, avoiding the error Fountain now perceives. A contention not raised in the district court may be pursued on appeal only in the event of plain error, Fed.R.Crim.P. 52(b), which this is not. The indictment charged Fountain with murder, not with the particular method of murder. Testimony at trial often supplements the sketchy terms of the indictment. This testimony did not contradict anything in the indictment (he supplied the knife and was a lookout in addition to restraining Spikerman) or surprise Fountain. He is really using this to attack the indictment by the back door, on the ground that the grand jurors had not heard from these three witnesses. They had not; but that the defendant’s hands are a darker shade of red than the grand jurors believed is not a reason to dismiss the charges.

2. Inmate-witnesses have their drawbacks—for example, they all have convictions—but they may be the only witnesses available in cases like this. Donald Bur-nim, one of the inmate-witnesses, testified that Fountain boasted about his role in Spikerman’s execution. Five months later he recanted in a statement to the FBI, adding for good measure that six other inmate-witnesses lied at the trial. He then sent a letter to another agent recanting his recantation. Two weeks later he sent a letter to defense counsel reinstating the recantation but refusing to swear to it. Some time after that he sent a letter to the United States Attorney disavowing the recantation and asserting that he recanted in the first place only because he feared retaliation by Fountain’s allies in prison—the Aryan Brotherhood, a singularly vicious prison gang. See, e.g., United States v. Abel, 469 U.S. 45, 105 S.Ct. 465, 83 L.Ed.2d 450 (1984); United States v. Clayton Fountain, 768 F.2d 790 (7th Cir.1985); Gometz v. Henman, 807 F.2d 113 (7th Cir.1986). Fountain insists that the prosecutor knowingly used perjured testimony or, at a minimum, that the conviction rests on perjury, so that there must be a new trial. See United States v. Kaufmann, 803 F.2d 289 (7th Cir.1986) (collecting authorities).

The district court conducted a hearing at which Burnim testified, as did Eddie Geouge, the prison guard who Bumim’s initial recantation accused of coercing his testimony. Geouge testified that he had not coerced Burnim and that inmates planning to testify for the prosecution were treated the same as inmates who planned to testify for Fountain. The district court found that Burnim had not perjured himself, that he recanted out of fear of the Aryan Brotherhood, and that Geouge had not applied pressure or promised favors. These findings are not clearly erroneous.

Fountain complains about the conduct of this hearing. Burnim invoked his privilege against self-incrimination in response to numerous questions, and Fountain insists that the district judge should have rejected the claim of privilege or directed the prosecutor to obtain a grant of immunity. Burnim was entitled to assert the privilege, for his waffling positions exposed him to the risk of prosecution for perjury or for making a material, false statement to a federal officer, see 18 U.S.C. § 1001. As for immunity: we have never held that the Constitution requires the government to overcome, for the defendant’s benefit, a constitutional (or any other) privilege that affects the availability of evidence. The circumstances of this case do not call for a break from that practice.

3. Gary Lumley, another inmate-witness, declined to speak with Fountain’s lawyers before trial. When approached, he said he would have to talk to his prison counselor; after the counselor talked with the prosecutor, Lumley refused to speak with the defense. Fountain says this must mean that the prosecutor told Lumley not to talk to the defense. Such an instruction would have been error. The problem lies in the inference. Other inmate-witnesses talked to Fountain’s lawyers before trial. There was obviously no campaign to insulate witnesses from probing by counsel. Inmate-witnesses may have sound reasons not to talk to the defense team. See United States ex rel. Jones v. DeRobertis, 766 F.2d 270, 274-75 (7th Cir.1985). A single witness’s refusal to meet with counsel before trial does not call for an evidentiary hearing into all of the influences that led to this decision.

4. Still a third inmate-witness, Donald Priest, had a history of mental problems. One psychiatrist diagnosed him as a paranoid schizophrenic, a person with trouble distinguishing reality from fantasy. Fountain asked the district court to require Priest to submit to a psychiatric examination before testifying. The court declined, observing that when Priest testified in Randall Dahlin’s trial, Judge Crabb found him coherent and lucid. Fountain then showed psychiatric reports concerning Priest to John Greist, a psychiatrist, and asked Dr. Greist to assess Priest’s ability to perceive correctly the events of January 1984. The district court sustained the prosecutor’s objection to Dr. Greist’s proposed testimony, ruling that he could not give an expert opinion because he had not examined Priest, had not seen him testify, and could rely only on reports that the court characterized as hearsay.

This sequence has something of a Catch-22 flavor. Having denied the request for a psychiatric examination of Priest, the court ruled that Dr. Greist was incompetent to testify because he had not examined Priest. But things are not quite so bad. Fountain could have subpoenaed the psychiatrists who had examined Priest. Dr. Greist could have watched Priest’s testimony and used that (plus the reports) as a foundation for expert analysis; he did not. The district court was entitled to be leery of both psychiatric examinations of witnesses and psychiatric testimony about witnesses, because the jury can observe for itself (as Judge Crabb did in Dahlin’s trial) the witness’s behavior. Criminal trials are complex enough without turning them into collateral investigations of the witnesses—investigations that would not only drag out trials and confuse jurors but also discourage people from serving as witnesses.

We have held repeatedly that district courts possess broad discretion in deciding whether to require witnesses to endure mental examinations. E.g., United States v. Raineri, 670 F.2d 702, 709 (7th Cir.1982); United States v. Gutman, 725 F.2d 417 (7th Cir.1984); United States v. Eschweiler, 745 F.2d 435, 438 (7th Cir.1984). The district court exercised his discretion in this case against the background of another district judge’s decision, after listening to Priest give more than 100 pages of testimony in another criminal trial, that Priest was a competent witness. The court then concluded, having heard Priest’s testimony in Fountain’s trial, that a psychiatrist’s evaluation of other psychiatrists’ reports would not assist the jury to assess Priest’s testimony; that, too, was a considered decision. True, the district court gave one incorrect reason—that the testimony would be hearsay because the underlying reports were hearsay, a conclusion inconsistent with Fed.R.Evid. 703—but the remaining reasons were sufficient to support the decision.

IV

Fountain received consecutive terms of life for murder and 150 years for conspiracy to commit murder, with 50 of these 150 to pass before becoming eligible for parole. The 50-year minimum was possible because of a conjunction of §§ 1117 and 4205(b). Section 1117 authorizes a court to order imprisonment “for any term of years or for life." Section 4205(b)(1) provides that when imposing a sentence of more than one year, a court may “designate in the sentence of imprisonment imposed a minimum term at the expiration of which the prisoner shall become eligible for parole, which term may be less but shall not be more than one-third of the maximum sentence imposed by the court”. All of this seems straightforward, an example of clear and therefore conclusive statutory language. E.g., United States v. Turkette, 452 U.S. 576, 580, 101 S.Ct. 2524, 2527, 69 L.Ed.2d 246 (1981); Swain v. Pressley, 430 U.S. 372, 377-78, 97 S.Ct. 1224, 1227-28, 51 L.Ed.2d 411 (1977).

The plain language produces a jarring result, however, when mixed with the equally plain language of some other statutes. Section 1111(b), dealing with first-degree murder, requires a court to impose a mandatory sentence of life. It does not mention terms of years, doubtless because its authors assumed that terms of years are less than life. The reference to “term of years” in § 1117 gives the district court power to impose sentences less than life, a desirable outcome for a crime less serious than murder. Yet because “any term of years” may include terms greater than life, § 1117 may be the basis of the harsher punishment. It is not obvious that “any term of years” means “any term of years less than the age of the universe” rather than “any term of years less than life”. The “any term” language complements the maximum of life and distinguishes § 1117 from § 1111(b), which sets life as the minimum. It is most unlikely that Congress wrote the “any term of years” language to allow courts to jigger parole dates. Still, Fountain does not argue that his 150-year sentence is excessive under § 1117. We need not decide whether it is.

The use of § 1117 to decree terms exceeding the defendant’s life expectation would not matter if § 4205(a) governed, for it draws the sting of 999-year sentences. Section 4205(a) says that any prisoner is eligible for parole after serving one-third of the term “or after serving ten years of a life sentence or of a sentence over thirty years”. So the life sentence and the sentence to 6.02 x 1023 years yield the same result: eligibility for parole after 10 years. 2

The trouble is that § 4205(a) applies, as its last words provide, “except to the extent otherwise provided by law.” Three courts of appeals have held that § 4205(b)(1) is such an “otherwise” provision, so that district judges may insist that parole come no sooner than one-third of the sentence — which under a statute such as § 1117 may be an eternity. United States v. O’Driscoll, 761 F.2d 589, 598 (10th Cir.1985); Rothgeb v. United States, 789 F.2d 647, 651-52 (8th Cir.1986); United States v. Gwaltney, 790 F.2d 1378, 1387-89 (9th Cir.1986). The result is that a first-degree murderer is eligible for parole after 10 years, because he receives a “mere” life sentence — while confederates with lesser responsibility may receive life in prison without possibility of parole if the judge imposes, say, a 300-year sentence under § 1117 with a 100-year minimum under § 4205(b)(1). The incongruity has led two judges to conclude that § 4205(b)(1) is not a statute that “otherwise provides” for purposes of § 4205(a). Rothgeb, 789 F.2d at 655-56 (Lay, J., dissenting); Gwaltney, 790 F.2d at 1389-94 (Norris, J., dissenting). See also Clayton Fountain, 768 F.2d at 799, stating that § 4205(b)(1) allows judges to make parole available before the time provided by § 4205(a) but not to postpone it, a dictum withdrawn on rehearing, 777 F.2d 345, 346 (7th Cir.1985), in favor of reserving the question.

Incongruity is not the only strike against using § 4205(b)(1) to defeat the maximum-minimum rule of § 4205(a). For decades “everyone knew” that § 4205(b)(1) and its predecessors could be used only to advance the date of parole eligibility. The statute was enacted in 1958. To our knowledge, not until 1984 did a court employ it to postpone eligibility for parole. The Department of Justice also “knew” that § 4205(b)(1) was a one-way street. The Bureau of Prisons, part of the Department, said in Program Statement 5050.9, at p. 2 (May 21, 1979), that it is so limited. This is not the sort of interpretation to which courts defer, because the Bureau of Prisons does not administer § 4205(b)(1). Although the Bureau computes release dates using judgments, the portion of § 4205 in question is addressed to sentencing courts, and deference runs to the decisions of those to whom statutes are addressed, those who must resolve ambiguities and exercise discretion to make the statute work. See Homemakers North Shore, Inc. v. Bowen, 832 F.2d 408, 411-12 (7th Cir.1987). But the Bureau’s interpretation shows the assumption widely shared at the time, indeed universally held until O’Dris- coll. The Department of Justice has yet to withdraw its interpretation. 3

One may say in reply that the language of § 4205(b)(1) is clear, and it is not illogical to allow a district court to defer as well as accelerate eligibility for parole. We do not have a case in which a court must bend plain language in order to avoid an absurdity. Church of the Holy Trinity v. United States, 143 U.S. 457, 459, 12 S.Ct. 511, 512, 36 L.Ed. 226 (1892). We do not even have a clear statute cancelled out by clearer legislative history. See Train v. Colorado Public Interest Research Group, Inc., 426 U.S. 1, 96 S.Ct. 1938, 48 L.Ed.2d 434 (1976); see also Arthur W. Murphy, Old Maxims Never Die: The “Plain Meaning Rule” and Statutory Interpretation in the “Modem” Federal Courts, 75 Colum.L. Rev. 1299 (1975); Note, Intent, Clear Statements, and the Common Law: Statutory Interpretation in the Supreme Court, 95 Harv.L.Rev. 892 (1982). (We return to the history of § 4205 below.) We have at most a surprising result, and judicial discomfort with a surprisingly harsh rule is not enough to permit its revision. TVA v. Hill, 437 U.S. 153, 98 S.Ct. 2279, 57 L.Ed.2d 117 (1977). It is not enough even in criminal law, as Turkette shows in taking the Racketeer Influenced and Corrupt Organizations Act for all it could be worth instead of the least it has to be worth.

When the statutes were enacted, their treatments of first-degree murderers and conspirators also were not so surprising as they now appear. The bite of any incongruity comes from the Supreme Court, not Congress. We have treated § 1111 as requiring life imprisonment, but this is not what it says. Its penalty provision reads:

Whoever is guilty of murder in the first degree, shall suffer death unless the jury qualifies its verdict by adding thereto “without capital punishment”, in which event he shall be sentenced to imprisonment for life.

The statute read this way in 1958 when Congress enacted the predecessor to § 4205(b)(1), just as it reads today. If this language were applied, first degree murderers would face a punishment more severe than 150 years in prison with parole after 50 years. Only in 1972, when Furman v. Georgia, 408 U.S. 238, 92 S.Ct. 2726, 33 L.Ed.2d 346 (1972), held unconstitutional statutes granting juries unconfined discretion to impose or withhold capital punishment, did it become possible for a conspirator’s sentence to exceed the principal’s. Neither Congress nor the Sentencing Commission has supplied a sentencing structure for capital punishment under § 1111, so the upshot is a mandatory life sentence. But this development does not retroactively change the meaning of a statute passed in 1958. 4 When the forerunner to § 4205(b)(1) was enacted, there was nothing incongruous in exposing murderers to death and their accomplices to long terms before parole eligibility.

The change in law concerning the death penalty may explain, however, the “discovery” in recent years of the potential of § 4205(b)(1). The lack of a federal death penalty, coupled with § 4205(a)’s maximum 10 years before becoming eligible for parole, means that there is no effective deterrent to murder by federal prisoners serving long sentences. Clayton Fountain describes the cases of multiple murderers in federal prison who committed their latest deeds in plain -dew and explained as they were herded back to their cells that they had to kill anyone who “disrespected” them. They knew that although a new offense would extend their parole eligibility dates by 10 years, they were not likely to get out before then anyway. The effective incremental punishment was zero. See also, e.g., United States v. House, 808 F.2d 508 (7th Cir.1986), another inmate-murderer whose only effective punishment was deprivation of potato chips and chocolate from the prison commissary until he could make restitution for the small burial costs of the victim. Small wonder that district judges have sought authority to impose real punishment when one inmate kills a guard or another inmate. The sentence in this case, effectively a “natural life” sentence, is more fitting than the one in United States v. Jackson, 835 F.2d 1195 (7th Cir.1987), which upheld a natural-life sentence for a four-time armed bank robber.

We arrive at the legislative history of § 4205(b)(1). Congress established a system of parole in 1910, 36 Stat. 819-20, and provided that any prisoner could be released after serving one-third of his term. This denied lifers the opportunity to be considered for parole. Congress extended that opportunity by the Act of January 23, 1913, 37 Stat. 650, providing that a prisoner who has observed the rules of the institution

and who, if sentenced for a definite term, has served one-third of the total of such term or terms for which he was sentenced, or, if sentenced for the term of his natural life, has served not less than fifteen years, may be released on parole as hereinafter provided.

This statute created the incongruity we now confront: a lifer became eligible for parole after serving 15 years, but a prisoner serving a sentence longer than 45 years would wait more than 15 years before becoming eligible.

The incongruity was more serious on paper than in practice, because in 1913 almost no one was sentenced to more than 45 years in prison. Few federal statutes permitted long terms of years. There were no federal drug offenses; the Lindberg Kidnapping Act and the Continuing Criminal Enterprise statute lay in the future. Capital punishment was common; anyone whose crime was serious enough to call for a long term was likely to be sentenced to death. So the criminal code carried this language through to 1951.

By 1951 there were 43 federal prisoners serving terms exceeding 45 years. H.Rep. No. 387, 82d Cong., 1st Sess. 2 (1951). The Department of Justice thought the incongruity too great to bear and recommended that the 15-year minimum be applied to prisoners serving terms of years. A letter from Deputy Attorney General Ford, attached to H.Rep. 387, states: “It seems inconsistent to deny [prisoners serving sentences longer than 45 years] the same parole privilege as is provided for prisoners serving life sentences.” Congress therefore enacted 28 U.S.C. § 4202 (1951), 65 Stat. 150 (1951), providing that any prisoner who has observed the rules of the institution “may be released on parole after serving one-third of such term or terms or after serving fifteen years of a life sen tence or of a sentence of over forty-five years.” The Senate report, in language similar to the House report, S.Rep. No. 524, 82d Cong., 1st Sess. 2 (1951), U.S.Code Cong. & Admin.Serv.1951, pp. 1676, 1677, explained that the new language

would make prisoners sentenced to imprisonment of over 45 years eligible for consideration of parole after serving 15 years to conform with the minimum period of eligibility for life prisoners. The present inflexible rule that a prisoner ... must serve one-third of his sentence to become eligible for parole seems unjust in its application to prisoners sentenced for more than 45 years because a prisoner serving a life sentence becomes eligible in 15 years. Thus, under the present law, a prisoner sentenced to a total of 60 years on a charge less severe in its nature than homicide, will have to serve 20 years before becoming eligible for parole, while a person sentenced to life for homicide becomes eligible for parole after serving 15 years. While there are only 43 Federal prisoners at present who are serving sentences in excess of 45 years, the committee Delieves the amendment desirable as removing a patent discrimination.

Between 1951 and 1958 prisoners sentenced to any term of years, no matter how long, were eligible for release on parole after 15 years. Judges had no authority to tinker with that date.

The inexorable 15-year minimum was the impetus for the change in 1958. Indeterminate sentencing was becoming common. In 1957 California created a system under which every adult sentence was zero-to-life, with the release date to be picked by the Adult Authority. 1957 Cal.Stat. c. 2256 § 58. 5 The Department of Justice submitted a bill that would move federal sentencing practices in the direction of California — not by indeterminate sentencing in every case, but by giving judges two new options for adult sentences: to make parole available from the beginning of the sentence, and to make parole available at any intermediate time up to the one-third mark. Deputy Attorney the Senate Judiciary Committee:

The present statute ... which provides that a prisoner is not eligible for parole consideration until he has served one-third of the sentence imposed, is purely arbitrary and does not take into consideration the response made by an individual prisoner to the rehabilitation programs carried on in our Federal institutions. Much bitterness is engendered in many prisoners who have otherwise admirably responded to rehabilitation programs by the knowledge that they must be incarcerated for a purely arbitrary period as is now provided by the Parole Act. The enactment of [our proposal] would have the practical effect of authorizing the imposition of indeterminate sentences by Federal courts. Indeterminate sentence laws are effective in 38 States and in the District of Columbia and the result has been uniformly beneficial.

Letter from Lawrence E. Walsh to Senator James 0. Eastland, reprinted as an appendix to S.Rep. No. 2013, 85th Cong., 2d Sess. (1958), U.S.Code Cong. & Admin.News 1958, p. 3891. This is a strong plea for eliminating minimum eligibility rules for parole. It represents a rehabilitative model that has been abandoned by Congress (the code of 1984 abolished parole for crimes committed after November 1, 1987) but in 1958 was widely shared. The Administrative Office of the United States Courts, on behalf of the Judicial Conference, supported the bill in these words (ibid., emphasis added):

This is a bill to authorize the court in sentencing a prisoner to fix an earlier date when the prisoner shall become eligible for parole_ The additional flexibility in setting parole eligibility dates [has] received substantial support from Federal judges.... Through it judges would be given wider discretion in the formulation of sentences and could share the responsibility for determining the date of parole eligibility. This would provide judges with a choice which has some features of an indeterminate sentence which now is not available to Fed eral judges, except for those offenders committed under the Youth Corrections Act.

Neither Deputy Attorney General Walsh nor the Administrative Office suggested that the pending legislation would undo the decision, made in 1951, to eliminate the difference between life sentences and those for long terms of years. The point was quite the opposite — to accelerate rather than retard eligibility for parole.

The committee reports on the legislation are in the same vein. The House deleted this portion of the Department’s legislative proposal but the Senate restored it, explaining:

The purpose of this section is to provide the court with optional procedures which will enable it to impose sentences which may be indeterminate in nature. This would permit the court, at its discretion, to share with the executive branch the responsibility for determining how long a prisoner should actually serve. The court could impose a term of imprisonment either under the existing definite sentencing system, or it could fix the maximum term of the sentence and (1) direct that the prisoner shall be eligible for parole at any time up to one-third this maximum, as now provided by law, or (2) specify that the Board of Parole shall decide when the prisoner will be considered for parole.

S.Rep. No. 2013, at 3, U.S.Code Cong. & Admin.News 1958, p. 3892 (emphasis added). The Conference Committee followed the Senate on this question, and its report is similar. H.R.Conf.Rep. No. 2579, 85th Cong., 2d Sess. (1958), U.S.Code Cong. & Admin.News 1958, p. 3905. The statute was enacted, Act of August 25, 1958, 72 Stat. 845-46 (1958).

No one raised in 1958 the interaction between the new sentencing options and the provision, added in 1951, to treat prisoners serving terms longer than 45 years in the same way as those serving life sentences. The references to the “existing definite sentencing system” and the “maximum, as now provided by law” strongly suggest, however, that Congress anticipated that the 1951 decision would be unaffected by the 1958 change. The function of the new language was to move in the direction of indeterminate sentencing — that is, earlier eligibility for parole — at the judges’ option, and not to create an option to restore the incongruity that had been eliminated seven years before.

The Parole Commission and Reorganization Act of 1976, which among many other things recodified the 1958 statute as 18 U.S.C. § 4205(b), did not alter the substance or even the language of this law. The 1976 bill was largely written by the Conference Committee. The conference report, H.R.Conf.Rep. No. 94-838, ' 94th Cong., 2d Sess. 24 (1976), U.S.Code Cong. & Admin.News 1976, pp. 335, 357, recites that “[ejxisting provisions of law are reenacted.” The only change of interest was the reduction from 15 to 10 in the number of years served before becoming eligible for parole.

The legislative history, from 1913 through 1976, is both silent and informative. Silent because, neither a Member of Congress nor a proponent of the several changes in the rules attempted to integrate the judges’ power under 18 U.S.C. § 4205(b) with the 15 (or 10) year eligibility rule under § 4205(a). Informative because the 1951 revision was explicitly justified as a rule that would eliminate the anomaly of parole on a term of years being available later than parole on a life sentence, informative because the 1958 revision was explicitly justified as giving judges flexibility to move in the direction of indeterminate sentences, making parole available even earlier. Not a voice was raised in 1951 and 1958 to suggest that judges should retain (or in 1958 regain) the authority to adhere to a strict one-third criterion for terms of years.

It is hazardous to draw inferences from silence. Perhaps the language chosen in 1958 was so “clear” that comment was unnecessary. E.g., Harrison v. PPG Industries, Inc., 446 U.S. 578, 592, 100 S.Ct. 1889, 1897, 64 L.Ed.2d 525 (1980); United States Railroad Retirement Board v. Fritz, 449 U.S. 166, 179, 101 S.Ct. 453, 461, 66 L.Ed.2d 368 (1981). But we do not find it that clear. It would be surprising if in 1958, when a vogue for indeterminate sentences was sweeping the country, Congress silently abandoned its decision of seven years previous to treat long terms of years no worse than it treated life sentences for purposes of parole. Cf. Watt v. Alaska, 451 U.S. 259, 267, 101 S.Ct. 1673, 1678, 68 L.Ed.2d 80 (1981); Upjohn Co. v. United States, 449 U.S. 383, 397-99, 101 S.Ct. 677, 686-87, 66 L.Ed.2d 584 (1981), both using the absence of discussion in Congress as the basis of an inference that Congress did not change existing law. The incongruity that confronts us today was considered and eliminated in 1951; it was not considered or consciously recreated in 1958; we therefore conclude that § 4205(b) does not authorize judges to postpone eligibility for parole beyond the time specified by § 4205(a). 6

Scott Fountain’s crime is serious, his conviction valid. The Parole Commission may well choose to detain him for the 50 years the district court thought appropriate. But § 4205 leaves to the Parole Commission the timing of his release, once he has served 10 years. The sentence on the conspiracy conviction is modified to delete the 50-year period before eligibility for parole, and the judgment, as modified, is

Affirmed.

1 . Though it is also true, as the Court stressed in Calandra and Costello, that attempts to superintend the presentation of evidence to the grand jury create the risk of enervating diversion from the main issues in the case and unwholesome delay. The Supreme Court has never held that a district court may dismiss an indictment because of the quantity or quality of evidence used and has stated repeatedly that it may not. Thus our assumption is strictly for purposes of argument.

2 . In using the present tense we have not overlooked § 218(a)(5) of Pub.L. 98-473, 98 Stat. 2027-31, which repeals § 4205 effective November 1, 1987, as part of the new system of determinate sentencing. Before passing out of existence, the Parole Commission must convert older sentences to determinate sentences, and in doing so must respect any mandatory minima it encounters. Section 235(b)(1) — (4), 98 Stat. 2032-33. The terms of § 4205 and the district court’s 50-year minimum will control the Commission’s conversion of Fountain’s sentence.

3 . The government filed a brief in this case embracing O’Driscoll, Rothgeb, and Gwaltney, despite the fact that the published position of the Department of Justice is to the contrary. Inquiry at oral argument produced the admission that the United States Attorney had not troubled to check with his superiors at the Department before confessing error on two published statements. When we asked for further briefs, the prosecutor furnished copies of the memoranda the Solicitor General filed in opposition to the petitions for writs of certiorari in O’Driscoll, cert. denied, 475 U.S. 1020, 106 S.Ct. 1207, 89 L.Ed.2d 320 (1986), and Gwaltney, cert. denied, — U.S. -, 107 S.Ct. 1337, 94 L.Ed.2d 187 (1987), asserting that these showed that the Department had changed its position. They show nothing of the sort. Each memorandum asks the Court to deny review on two grounds: that there is no conflict among the circuits, and that the problem has little prospective significance because § 4205 has been repealed. Each studiously avoids taking a position on the merits. The memorandum in Gwaltney even adds, in a footnote, that the Department has a published statement to the contrary. We accordingly ordered the United States Attorney to file a second post-argument brief. The second brief asserts that the United States Attorney obtained the verbal authorization of Assistant Attorney General Weld to repudiate the published statement. The memorandum does not give Mr. Weld’s reasons.

This is not a satisfactory way to litigate. A court ought to be able to assume that published documents of the Department of Justice state the Department’s position, unless rescinded by appropriate procedures. Phone calls from Assistant Attorney:. General are no substitute for published statements of reasons — and far from issuing such a statement, the Department has left its manual in force for three years after O'Driscoll showed that its contents are questionable. The Assistant Attorney General for the Criminal Division is not the supervisor of the Bureau of Prisons and lacks the authority to rescind or countermand its manuals. The Solicitor General’s careful treatment of this subject has not been replicated.

4 . Despite Dean Calabresi’s provocative suggestion to the contrary, see Guido Calabresi, A Common Law for the Age of Statutes (1982), Congress rather than the courts still is in charge of keeping statutes up to date. See National Broiler Marketing Ass’n v. United States, 436 U.S. 816, 827, 98 S.Ct. 2122, 2129, 56 L.Ed.2d 728 (1978).

5 . California abandoned wholly indeterminate sentencing in 1976, see 1976 Cal.Stat. c. 1139 § 281. The current statute is Cal.Penal Code §§ 3000-41.

6 . Because this holding creates a conflict among the circuits, Part IV of this opinion was circulated to all active judges. See Circuit Rule 40(f). No judge requested rehearing en banc.