Summary
holding that Rule 6(e) ’s "phrasing can, and should, accommodate rare exceptions premised on inherent judicial power" and endorsing our holding in Aisenberg
Summary of this case from Pitch v. United StatesOpinion
Nos. 04-1808, 04-1809, 04-1810.
Heard February 7, 2005.
Decided July 8, 2005.
Appeal from the United States District Court for the District of Massachusetts.
Before BOUDIN, Chief Judge, LYNCH and LIPEZ, Circuit Judges.
This is a redacted version of the original decision filed under seal on April 26, 2005. The redactions in this published version, indicated by double brackets, are intended to make public the court's legal rulings while protecting grand jury secrecy and related interests. See, e.g., In re Grand Jury Subpoena, Judith Miller, 397 F.3d 964, 1002 (D.C. Cir. 2005); In re Lindsey, 148 F.3d 1100 (D.C. Cir. 1998).
This case presents important questions involving grand jury operations and the attorney-client privilege. Because the grand jury investigation is ongoing, our discussion omits a certain amount of detail and by a separate order the decision is being sealed. However, the court anticipates that an expurgated version of the opinion or summary of the key legal rulings will be published in due course.
Background. In the course of a grand jury investigation in [], a lawyer (Lawyer I) directed his client ("Client A") to commit perjury in testifying before the grand jury, after initially advising him to tell the truth. The same lawyer has also represented another client ("Client B") who might or might not have some connection with the earlier perjury. In [], Lawyer I told Client A to recant the false testimony. Lawyer I did so after consulting with Lawyer II, who represented [] other clients (collectively "Group C") variously connected with Client B and with pertinent events.
Learning of the perjury, the government is now investigating the possible involvement of others with that perjury and with other possible crimes. The present grand jury summoned Lawyer I, and the prosecutor sought to question him about the prior perjury of Client A including the involvement of others with that perjury and its subornation. Lawyer I refused to answer a number of these questions, saying that answering them would invade Client B's attorney-client privilege and the joint-defense privilege enjoyed by the Group C clients.
That questions to Lawyer I might invade the privilege of Client B is clear but the Group C clients are not represented by Lawyer I but by Lawyer II who practices independently. However, it was Lawyer I's position, not contested by the government at this stage, that the two lawyers at the relevant times had a joint-defense arrangement which allowed them, without waiving the attorney-client privilege, to exchange confidences that their respective clients may have confided to them separately.
Eventually the government moved to compel Lawyer I to answer the disputed questions, arguing that the crime-fraud exception defeated the claim of privilege as to both Client B and the Group C clients. Upon Lawyer I's request, the court notified Client B's current counsel — unaffiliated with Lawyer I — and counsel for each of the Group C clients; [].
At the government's behest, the district court also directed Lawyer I not to reveal to anyone (other than his own counsel) the "substance of the government's motion to compel" — an order the government thought essential to protect against the risk that Lawyer I might affect or influence other potential grand jury witnesses. Client B objected that under Fed.R.Crim.P. 6(e)(2)(A), Lawyer I could not be barred from disclosing to Client B or anyone else what he (Lawyer I) learned as a grand jury witness or elsewhere.
Current counsel for two [] Group C clients also objected to the motion to compel. Lawyer II provided an affidavit describing his representation of the Group C clients; in addition to other details about his consultations with Lawyer I, he said that incident to their joint defense arrangement he had disclosed to Lawyer I confidential information received from at least one of the Group C clients.
The district court held several hearings and heard arguments as to whether or not the privilege was abrogated. The government also made two ex parte filings whose contents were not made known to opposing counsel; a redacted version of one of the filings was, however, supplied to opposing counsel. Such ex parte submissions, although surprising to those unacquainted with the practice, have precedent in certain contexts including grand jury matters and privilege claims. See In re Grand Jury Proceedings (Violette), 183 F.3d 71, 79 (1st Cir. 1999).
The court also considered the government's request that it engage in a limited type of in camera inquiry suggested as a possibility by the Supreme Court in United States v. Zolin, 491 U.S. 554, 109 S.Ct. 2619, 105 L.Ed.2d 469 (1989). In such an inquiry the judge may examine privileged material itself to determine whether it is or remains privileged. The district judge expressed reservations about his capacity, given the complexities of the background events, to conduct such a limited inquiry which would involve the judge in questioning Lawyer I and not merely reviewing documents. In the end, the government chose to rest on what it already adduced and abandoned its request for an in camera inquiry under Zolin.
Thereafter, in March 2004, the district court granted the government's motion to compel answers to the government's listed questions. Requests by Client B and one of the Group C clients to allow their lawyers to interview Lawyer I about the underlying events were denied. The objecting clients then sought access to all of the filings made so far in the district court, including the government's ex parte filings. This too was denied in October 2004 when the district court issued a memorandum explaining its decisions in more detail.
In the October 2004 memorandum, the district judge explained his reasons for concluding that the crime-fraud exception applied. The judge said he was satisfied that the government had made a " prima facie showing" not only of Lawyer I's corruption but also, [], of
a shared intent by the Clients — when making the communications the government seeks to discover — to use otherwise privileged attorney-client communications to facilitate corruption and frustration of the grand jury inquiry.
The district judge stressed that he was not finding that any of the clients had in fact corruptly communicated with counsel. He only found enough of a likelihood had been established to justify compelling Lawyer I's answers — albeit "limited only to those lines of inquiry" identified by the government.
The district court expressed discomfort with the government's bypassing of the Zolin in camera inquiry, warning that this might result in a remand. The court also explained its reasons for concluding that the court had authority to forbid Lawyer I from disclosing to anyone, including Lawyer II and to Client B and the Group C clients, what he had been asked in the grand jury or other information pertaining to "the subject matter" of the grand jury inquiry.
On this appeal, Client B and the Group C clients, as the alleged privilege holders, have attacked the prima facie standard adopted by the district court and, in the alternative, argued that even this allegedly too lenient standard has not been met. Client B also urges that the government and district court were obliged to follow the Zolin in camera procedure. Finally, the clients say that the non-disclosure order to Lawyer I was barred by Fed R.Crim. P. 6(e)(2) and that it was overly broad as well.
The customary standard of review in cases of privilege is that strictly legal questions are reviewed de novo, raw fact findings are reviewed for clear error, and the application of legal standards to specific facts with a measure of deference. See Cavallaro v. United States, 284 F.3d 236, 245 (1st Cir. 2002); United States v. Mass. Inst. of Tech., 129 F.3d 681, 683 (1st Cir. 1997). We begin with the privilege question, which involves both law and law-application issues, and then return to the Rule 6 issue which presents solely a question of law.
The decisions, e.g., Violette, sometimes say that privilege determinations are reviewed for abuse of discretion, but this is the "umbrella" usage of the phrase by which, for example, an abstract error of law — certainly not reviewed deferentially — is described as one kind of abuse of discretion. E.g., Cooter Gell v. Hartmarx Corp., 496 U.S. 384, 402, 110 S.Ct. 2447, 110 L.Ed.2d 359 (1990).
Privilege. Familiarly, the attorney-client privilege — somewhat simplified — is a privilege of a client to refuse to testify or to have his counsel testify as to confidential communications between the two made in connection with the rendering of legal representation, see Cavallaro, 284 F.3d at 245; see also Fisher v. United States, 425 U.S. 391, 403, 96 S.Ct. 1569, 48 L.Ed.2d 39 (1976); 8 Wigmore, Evidence § 2292 (McNaughton rev. 1961). It extends as well to communications made within the framework of a joint defense arrangement. In re Grand Jury Subpoena, 274 F.3d 563, 572 (1st Cir. 2001); United States v. Bay State Ambulance and Hosp. Rental Serv., Inc., 874 F.2d 20, 28 (1st Cir. 1989).
The crime-fraud exception — one of several qualifications to the attorney-client privilege — withdraws protection where the client sought or employed legal representation in order to commit or facilitate a crime or fraud. Specifically, courts have required the privilege challenger to present evidence: "(1) that the client was engag[ed] in (or was planning) criminal or fraudulent activity when the attorney-client communications took place; and (2) that the communications were intended by the client to facilitate or conceal the criminal or fraudulent activity." Violette, 183 F.3d at 75.
See generally United States v. Reeder, 170 F.3d 93, 106 (1st Cir.), cert. denied 528 U.S. 872, 120 S.Ct. 174, 145 L.Ed.2d 147 (1999); United States v. Rakes, 136 F.3d 1, 4 (1st Cir. 1998).
Like the privilege itself, the exception, where employed in a federal criminal case, is effectively a creature of federal common law. See Fed.R.Evid. 501. This means that federal judges start with a core of common precedent reflecting the privilege but also have power to refine and adjust both the substance and procedure in light of reason and experience. See Zolin, 491 U.S. at 562-63, 109 S.Ct. 2619; Violette, 183 F.3d at 77-78. The process of development is far from over.
It is often hard to determine whether the attorney-client relationship has been misused by the client for crime or fraud without seeing the document, or hearing the testimony, as to which the privilege is claimed. To overcome this problem (as well as other privilege problems) judges have sometimes been willing to review privileged materials by themselves in camera and then decide whether the other side is entitled to it. See, e.g., United States v. Jacobs, 117 F.3d 82, 87 (2d Cir. 1997). The Supreme Court has blessed this procedure in Zolin. 491 U.S. at 572, 109 S.Ct. 2619; see also United States v. Edgar, 82 F.3d 499, 509 (1st Cir.), cert. denied 519 U.S. 870, 117 S.Ct. 184, 136 L.Ed.2d 123 (1996).
Zolin also said that a "lesser evidentiary showing" is needed for in camera review than is needed finally to pierce the privilege, and that upon this showing whether to conduct in camera review is within the trial court's discretion. 491 U.S. at 572, 109 S.Ct. 2619. For such a review, all that was needed was a factual basis "to support a good-faith belief by a reasonable person" that " in camera review of the materials may reveal evidence to establish the claim that the crime-fraud exception applies." Id. (internal quotation marks omitted). "May" is a very relaxed test and, as only the judge gets this initial access, properly so.
Despite Zolin's mention of discretion, conceivably on some facts a district court might abuse its discretion by declining to conduct an in camera review prior to piercing the privilege. But a categorical rule would be unwise and Client B has not argued for one.
However, Zolin did not answer the question of what level of proof is needed to pierce the privilege itself, 491 U.S. at 563, 109 S.Ct. 2619, although it acknowledged that the statement in Clark v. United States, 289 U.S. 1, 14, 53 S.Ct. 465, 77 L.Ed. 993 (1933), that a " prima facie case" was the needed showing "has caused some confusion." Zolin, 491 U.S. at 563 n. 7, 109 S.Ct. 2619. In our own case, the district court found a prima facie showing had been established, but cautiously did not further explain what this meant. We share its caution but are obliged to attempt refinement.
"Prima facie" is among the most rubbery of all legal phrases; it usually means little more than a showing of whatever is required to permit some inferential leap sufficient to reach a particular outcome. Black's Law Dictionary 1228 (8th ed. 2004); e.g., McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973). Here, the outcome is piercing the privilege based on a certain level of proof. This circuit has previously avoided a calibration of that level, Violette, 183 F.3d at 78, as has the Supreme Court, Zolin, 491 U.S. at 563, 109 S.Ct. 2619, but it is important in this case as to one client.
As we read the consensus of precedent in the circuits, it is enough to overcome the privilege that there is a reasonable basis to believe that the lawyer's services were used by the client to foster a crime or fraud. The circuits — although divided on articulation and on some important practical details — all effectively allow piercing of the privilege on something less than a mathematical (more likely than not) probability that the client intended to use the attorney in furtherance of a crime or fraud. This is a compromise based on policy but so is the existence and measure of the privilege itself.
A "probable" or "reasonable" cause standard is used in the Second, Sixth and Ninth Circuits. Jacobs, 117 F.3d at 87; In re Grand Jury Proceedings, 87 F.3d 377, 381 (9th Cir.), cert. denied 519 U.S. 945, 117 S.Ct. 333, 136 L.Ed.2d 246 (1996); In re Antitrust Grand Jury, 805 F.2d 155, 165-66 (6th Cir. 1986). The Third Circuit requires "evidence which, if believed by the fact-finder, would be sufficient to support a finding that the elements of the crime-fraud exception were met," Haines v. Liggett Group Inc., 975 F.2d 81, 95-96 (3d Cir. 1992); accord, In re Sealed Case, 754 F.2d 395, 399 (D.C. Cir. 1985). The Third and D.C. Circuits have described their standard as similar to "probable cause." Haines, 975 F.2d at 95; In re Sealed Case, 754 F.2d at 399 n. 3.
Precisely because of the initial barrier of the privilege, it is very hard for an adversary unaided to show that the privileged communications were themselves corrupt, so the requirements for access cannot be set too high. And, if the communications were innocent, the initial look may often not damage the client. In all events, the reasonable cause standard is intended to be reasonably demanding; neither speculation nor evidence that shows only a distant likelihood of corruption is enough.
In addition, the privilege is not lost solely because the client's lawyer is corrupt — as was certainly true in this case since the lawyer suborned Client A's perjury. The crime-fraud exception requires the client's engagement in criminal or fraudulent activity and the client's intent with respect to attorney-client communications. Violette, 183 F.3d at 75, 79. Accord, In re Grand Jury Subpoenas, 144 F.3d 653, 660 (10th Cir.), cert. denied 525 U.S. 966, 119 S.Ct. 412, 142 L.Ed.2d 334 (1998); In re Sealed Case, 107 F.3d 46, 49 (D.C. Cir. 1997). If there are contrary views, we do not think that they comport with the modern view of the privilege's nature and purpose as explained in Upjohn v. United States, 449 U.S. 383, 389, 101 S.Ct. 677, 66 L.Ed.2d 584 (1981).
In re Impounded Case (Law Firm), 879 F.2d 1211, 1213-14 (3d Cir. 1989), cited by the government, says that the crime-fraud exception may apply even where "pertinent alleged criminality is solely that of the law firm," but it is unclear in the context of that case whether the court intended that this showing would forfeit the privilege of an innocent client.
An additional limitation on piercing is that it is not enough to find reasonable cause to believe that the client is guilty of crime or fraud. Forfeiture of the privilege requires the client's use or aim to use the lawyer to foster the crime or the fraud. As Rakes explained:
No general rule withdraws the privilege from communications that occur in the same time frame as criminal [conduct]. . . . [T]he attorney-client privilege is forfeited inter alia where the client sought the services of the lawyer to enable or aid the client to commit what the client knew or reasonably should have known to be a crime or fraud.
Here, we are satisfied that the reasonable cause standard is satisfied as to Client B. []
Our assessment, like that of the district judge, is preliminary, reflects no finding that Client B acted wrongfully, and is limited solely to the issue whether there is reasonable cause adequate to pierce Client B's privilege. The compelled testimony of Lawyer I (or other evidence) could turn out to vindicate Client B entirely and inculpate someone (or no one) else. The prima facie case is sometimes assumed to destroy the privilege once and for all once the opponent gains access to the material (which is one reason for the Zolin procedure). We think that there may well be room for protective remedies if the privilege is initially pierced but this turns out to have been in error.
It should be noted, finally, that the district court's initial proposal to examine Lawyer I in camera was not a finding that the government had met only the lower threshold under Zolin and therefore could only proceed to issue its motion to compel after acquiring additional inculpatory evidence obtained in camera. In its October 2004 decision, the district court made quite clear that the government had fully met its prima facie burden needed to overcome Client B's privilege, and it is this conclusion that we affirm.
The more difficult problem concerns the Group C clients. There is scant evidence that these clients' purpose in retaining Lawyer II was to use his services in furtherance of a crime or fraud. True, the district judge made a finding (quoted above) of a "shared intent" to "use otherwise privileged attorney-client communications to facilitate corruption and frustration of the grand jury inquiry." But his decision points to no facts to buttress this finding, nor — more ominously since the government certainly should know — is any such raw-fact evidence identified in the government's briefs.
It is not enough that the lawyer's work posed an obstacle to the grand jury; perfectly legitimate representation may do this. [] To pierce the privilege requires that the clients have an intent to misuse their legal representation ( e.g., by enlisting the lawyer's advice to frame their perjury) to perpetrate the crime or fraud. Violette, 183 F.3d at 75. Such evidence may exist as to Group C but it has not yet been furnished.
The government relies upon In re Sealed Case, 754 F.2d at 399-402, where a client's document destruction incident to litigation was treated as undercutting the privilege of communications between it and its lawyers in that litigation. But there, the attorneys represented the client in the very litigation for which evidence had been destroyed, and in hearings in which the client sought to prevent exposure of the destruction, in the course of which they filed false documents and verified their authenticity. Id. at 400, 402. The lawyers were thus justifiably characterized as "instrumentalities in the ongoing cover-up." Id. at 402. In this case, no such misuse of Lawyer II is shown.
Now that Client B's privilege has been pierced, some of what the government seeks can conceivably be obtained by questions directed to Lawyer I tailored to avoid disclosures that would invade Group C's privilege; arguably Lawyer I-Client B conversations can be reached regardless of what information Lawyer I passed on to the Group C clients. How far Lawyer I may be questioned about the source of the information — if it came from Lawyer II or the Group C clients — may be more debatable. Other difficult questions may also arise.
An example is the extent to which — once Client B's privilege is pierced — Lawyer I may be made to testify about whether he conveyed these clients' information to Lawyer II. See, e.g., In re Grand Jury Subpoena, 274 F.3d 563, 572 (1st Cir. 2001) ("Even when [the joint-defense privilege] applies, however, a party always remains free to disclose his own communications."); United States v. Agnello, 135 F.Supp.2d 380, 383 (E.D.N.Y. 2001) (same).
The government also remains free to ask the district court to make an in camera inquiry under Zolin directed to overcoming the Group C clients' privilege. The government abandoned this quest when it apparently judged that the district court might side with it anyway. Nothing prevents the prosecutor from pursuing this route, either before or after further questioning of Lawyer I, if the district judge is willing to follow it. It is for the district court to say whether the low Zolin standard would warrant in camera questioning of the Group C clients.
The non-disclosure order. When in December 2003 the district court gave notice of the government's motion to compel, it included a protective order directing Lawyer I and his counsel that "the substance" of the motion to compel should not be further disclosed or discussed with any third parties (including Client B, Client B's current attorney, Lawyer II or the Group C clients). In its October 2004 memorandum, the district court added that it would continue prohibiting Lawyer I "from discussing the subject matter of his grand jury [testimony] with anyone other than his own counsel."
Together, Client B and the Group C clients urge that the order is inconsistent with Rule 6(e), is in any event over-broad, and violates the First Amendment. Client B, in particular, complains that Lawyer I's counsel interpreted the order to extend beyond "matter[s] occurring before the grand jury" to preclude Lawyer I from discussing not only what he was asked but also his recollection of the underlying events in which Lawyer I was involved. These include Lawyer I's original subornation and his later turn-about advice that Client A recant.
These past events are matters assuredly of interest to Client B, given government suspicions of him. Moreover, as Lawyer I represented Client B at the time of the subornation, the latter certainly has special reason to want to know what his erstwhile lawyer recalls about those events. In the district court, Client B unsuccessfully sought to clarify and narrow the order and he now says that the district court has implicitly confirmed Lawyer I's counsel's reading. Whether or not the district court so intended, the government does not dispute that the protective order is intended to have this breadth.
The government has not contested standing and quite arguably these interests are enough to give Client B standing to challenge the existence and scope of the secrecy order. But cf. In re Swearingen Aviation Corp., 605 F.2d 125, 126-27 (4th Cir. 1979) (rejecting challenge to secrecy order by non-witnesses on grounds that secrecy was meant to benefit witnesses).
There are several layers of argument to be peeled away. The first is the clients' claim that Rule 6(e) by itself bars the imposition of any secrecy order on Lawyer I. Fed.R.Crim.P. 6(e)(2)(B) lists categories of persons ( e.g., grand jurors and court reporters) who "must not disclose a matter occurring before the grand jury," subject to certain exceptions in subsection 6(e)(3). But 6(e)(2)(B) is preceded by subsection 6(e)(2)(A) which provides: "No obligation of secrecy may be imposed on any person except in accordance with Rule 6(e)(2)(B)."
Witnesses are not listed in Rule 6(e)(2)(B), nor does Lawyer I fall into any of the listed categories of persons who "must not" disclose grand jury matters. Pertinently the original advisory committee note says that "[t]he rule does not impose any obligation of secrecy on witnesses" and, although some districts did so prior to the rule, a "seal of secrecy on witnesses seems an unnecessary hardship and may lead to injustice if a witness is not permitted to make a disclosure to counsel or to an associate." Rule 6 Advisory Committee Notes, 1944 Note to Subdivision (e).
At first blush this appears to be a strong argument for the appellants, but several courts have read the rule more narrowly. These decisions take Rule 6(e)(2)(A) to set a default rule of permitting disclosure by witnesses absent a contrary order by the court in that proceeding, but also to leave open the possibility of restrictions where they can be justified by particular and compelling circumstances. See In re Grand Jury Proceedings, 814 F.2d 61, 69 (1st Cir. 1987) (collecting cases but not deciding the issue).
E.g., In re Subpoena To Testify Before Grand Jury Directed to Custodian of Records, 864 F.2d 1559, 1563-64 (11th Cir. 1989); In re Grand Jury Subpoena Duces Tecum, 797 F.2d 676, 680 (8th Cir.), cert. dismissed 479 U.S. 1013, 107 S.Ct. 661, 93 L.Ed.2d 714 (1986). But see United States v. Radetsky, 535 F.2d 556, 569 (10th Cir.), cert. denied 429 U.S. 820, 97 S.Ct. 68, 50 L.Ed.2d 81 (1976), overruled in part on other grounds by United States v. Daily, 921 F.2d 994, 1004 n. 11 (10th Cir. 1990).
We now decide that the rule's phrasing can, and should, accommodate rare exceptions premised on inherent judicial power. Absent restriction, courts have inherent power, subject to the Constitution and federal statutes, to impose secrecy orders incident to matters occurring before them. The general power is regularly expressed in orders limiting access to discovery materials, closing sensitive proceedings, and in other contexts. Sometimes these powers are reflected in or reconfirmed by rules, e.g., Fed.R.Civ.P. 26(c), but orders of this kind predated such rules which ordinarily reflect or refine the underlying authority without displacing it.
See Chambers v. NASCO, Inc., 501 U.S. 32, 42-50, 111 S.Ct. 2123, 115 L.Ed.2d 27 (1991) (sanctions rules do not displace courts' inherent authority); United States v. Aisenberg, 358 F.3d 1327, 1347 (11th Cir.), cert. denied ___ U.S. ___, 125 S.Ct. 276, 160 L.Ed.2d 115 (2004) ("Although Rule 6(e)(3) enumerates the exceptions to the traditional rule of grand jury secrecy . . . the district courts have inherent power beyond the literal wording of Rule 6(e)(3) to disclose grand jury material. . . ."); Matter of Special March 1981 Grand Jury, 753 F.2d 575, 577 (7th Cir. 1985) (court has inherent power to cure improvident disclosure).
Those who drafted and approved Rule 6(e) seemingly intended to do no more than abolish a general practice existing in some districts of automatically silencing grand jury witnesses as to their testimony. The advisory note says that the rule "does not impose" secrecy obligations and does not say that it prohibits such obligations made by appropriate order. There is no indication in the history of an intent to cut off the courts' power to restrict disclosure based on extraordinary circumstances. Cf. Aisenberg, 358 F.3d at 1347.
In this case, we are not dealing with an ordinary grand jury witness. Lawyer I confessedly suborned perjury of a witness in a related grand jury proceeding and is connected to certain underlying events that may have comprised an effort to frustrate a grand jury investigation. Further, disclosure of what is now happening in the present grand jury could help Client B (whose own privilege has been pierced under the crime-fraud exception) and the Group C clients [] coordinate their own testimony.
Rule 6(e)(2)(A) was not directed to such a situation nor does the First Amendment bar an appropriately tailored order. The interest in the secrecy of grand jury proceedings during their pendency has been treated with great respect by the courts. The present ban is concerned with protecting an ongoing grand jury investigation from further abuse by one who has already demonstrated a capacity and intention to frustrate the investigation. The restriction as to only one witness is planted in an adequate showing of special need.
See, e.g., In re Grand Jury Subpoena, 223 F.3d 213, 218-19 (3d Cir. 2000); In re Subpoena To Testify, 864 F.2d at 1564; see also 2 Beale et al., Grand Jury Law and Practice § 9:19 (2d ed. 1997).
Our circumstances are far away from the only Supreme Court case invoked by appellants. In Butterworth v. Smith, 494 U.S. 624, 635-36, 110 S.Ct. 1376, 108 L.Ed.2d 572 (1990), the Supreme Court struck down a Florida statute precluding a state grand jury witness from ever disclosing his grand jury testimony. But the grand jury proceeding had long been completed and it was the permanency of the ban that most troubled the Supreme Court. Id. at 632, 635-36, 110 S.Ct. 1376. Lawyer I or Client B may return to the district court to request a termination of the ban whenever the need for it has disappeared.
This takes us to appellants' claim that the ban, if sustainable at all, is overly broad. The ban appears to cover two different although related subjects. One is disclosure by Lawyer I of whatever happens within the grand jury — questions he has been asked or may be asked, answers he has given or may give, documents he may provide or be shown, and the like. The other is his own independent recollection of events about which he may or may not be asked including discussions or interactions with Clients A and B.
The former are, strictly speaking, "matter[s] occurring before a grand jury" and their revelation poses the most direct threat that the government's hand may be disclosed to other witnesses, enabling them to alter their own answers or coordinate evasions. Such revelations are already prohibited for most of those connected with the grand jury, and extension of this limitation to Lawyer I, based on the special circumstances of this case, is justified.
It is more difficult to sustain a ban on Lawyer I's discussion with his own former client, or with others, about Lawyer I's independent recollection of previous events outside the grand jury even though they may be matters about which the grand jury has questioned or will question him. In particular, Client B has a clear interest in preserving the opportunity to find out what his former lawyer recollects about Client B's own involvement and the interaction. Nor does such a discussion, if honestly conducted, reveal the grand jury's plan of attack.
Of course, any discussion between Lawyer I and anyone else of interest to the grand jury presents an opportunity to conspire and tailor stories. But that opportunity exists in any event; Lawyer I has not been precluded from all contact with his own former clients or lawyers for Group C clients. Nor do we normally prevent indicted co-conspirators from discussing their cases with one another even though in some cases the prospect that they will tailor their stories is very real.
The district judge has adopted the broad construction of his order, if at all, by indirection, and has never provided an explicit grounding for extending the order beyond revelation of the proceedings before the grand jury. The original notice barred disclosure by Lawyer I only of "the substance of the government's motion to compel"; and the October memorandum ambiguously prohibited discussion of "the subject matter of [Lawyer I's] grand jury [omitted word — presumably "testimony"] with anyone."
We do not say that a witness could never be precluded from discussing independent recollections; situations involving national security are too obvious a concern to encourage general pronouncements. Instead, finding the ban unclear in scope but lacking clear justification for the broader reading, we modify the present restriction to cover only (in addition to the substance of the government's sealed motion) "matters occurring before the grand jury" and not independent recollections of the witness about external prior events as to which the witness has been or may be questioned.
There is one further caveat. It is difficult to foresee how events may develop and we are sensitive to the fact that Client B and to a lesser extent the Group C clients seemingly had a protected relationship with Lawyer I. Our modified affirmance of the order limiting disclosure by Lawyer I rests on the understanding that Lawyer I, Client B and the Group C clients remain free at any time to seek relief from the district court for good cause shown so as to permit Lawyer I to disclose some or all of his grand jury testimony to Client B, the Group C clients or both.
Accordingly, the district court's order granting the motion to compel is affirmed as to Client B's privilege but vacated as to the Group C clients' privilege (without prejudice to a further showing by the government with respect to them). The order limiting disclosure is modified, and as modified affirmed, to apply only to the proceedings before the district court on the motion to compel and to "matters occurring before the grand jury" as that phrase is customarily employed and as discussed above. No costs.
It is so ordered.