Norman C. Kleinberg with whom Michael E. Salzman, Michael T. Isbell and Hughes Hubbard & Reed, New York City, for appellees Collins & Aikman Corp., with whom Benjamin Acosta, Jr., San Juan, P.R., for La Cor Wicker.

Anderson, Moss, Russo & Cohen, P.A., Miami, Fla., for Otis Elevator.

Brown, Todd & Heyburn, Louisville, Ky., for Firestone Tire & Rubber Co., Dow Chemical Co., Future Foam, Inc., Sealed Air Corp., Texaco Chemical Co., PPG Industries, Inc., Olin Corp., Quantum Chemical Corp., ICI Americas Inc., Goodyear Tire & Rubber Co., EI DuPont de Nemours & Co., BASF Corp., Amoco Fabrics & Fibers Co., W.R. Grace & Co., Union Carbide Corp., Products Research & Chemical Corp., Allied-Signal, Inc., and Vasallo Industries, Inc.

Chase, Rotchford, Drukker & Bogust, Los Angeles, Cal., for Mitchell-Mann, Inc. and 2M Design Studio, Inc.

Fendig, McLemore, Taylor & Whitworth, Brunswick, Ga., for G.A. Marse Associates, Inc.

Kaye, Scholer, Fierman, Hays & Handler, Washington, D.C., for ADT Security Systems, Inc.

Kirkland & Ellis, Washington, D.C., for Ralph Wilson Plastics Co.

Margolis, Edelstein, Scherlis, Sarowitz & Kraemer, Philadelphia, Pa., for The Sheraton Corp.

Paxton & Seasongood, Cincinnati, Ohio, for Shelby-Williams, Inc.

White & Case, Washington, D.C., for Tarkett, Inc., join on brief.

Before CAMPBELL, Chief Judge, TORRUELLA and SELYA, Circuit Judges.

SELYA, Circuit Judge.

This matter arises on an infrastructure of important concerns involving the prophylaxis to be accorded to attorneys' work product and the scope of trial judges' authority to confront case management exigencies in complex multi-district litigation. The critical question is somewhat novel. We delve rather deeply into the doctrinal underpinnings of the work product rule and the emergent need for increased judicial intervention in the early stages of the adjudicative process in explaining our affirmance of the challenged district court order.

On New Year's Eve 1986, a conflagration engulfed the San Juan Dupont Plaza Hotel. The blaze resulted in ninety-six deaths, numerous personal injuries, and extensive property damage. Upwards of 2,000 persons sued. Many of the suits were brought in, or removed to, federal district courts. Under the aegis of the Judicial Panel on Multi-District Litigation, those actions were consolidated for discovery purposes in the United States District Court for the District of Puerto Rico. The litigation has attained heroic proportions: there are roughly two hundred defendants and ten times that number of plaintiffs.

Pretrial discovery has proven to be a gargantuan undertaking. More than 2,000,000 documents have been produced; countless interrogatories have been served; and depositions are proceeding daily along fourteen simultaneous tracks. The trial judge recently estimated that over 2,000 depositions would be required before discovery closed. Due to the immensity of the litigation, the district court has necessarily assumed an active managerial role. As the linchpin of that endeavor, the court entered an elaborate forty-five page case management order (CMO). We described certain facets of the CMO in a recent opinion, In re Recticel Foam Corp., 859 F.2d 1000, 1001 (1st Cir.1988), and will not repastinate that ground. It suffices for today to state that, inter alia, the CMO set out general discovery guidelines, established a phased schedule for pretrial preparation, and ordered formation of a joint discovery committee (JDC). From time to time, as appropriate, the court has supplemented the CMO with additional orders and refinements, whilst repeatedly imploring counsel "to explore novel methods of discovery that would ensure expeditious progress of the litigation." Recticel, supra, at 1001.

Not surprisingly, discovery disputes occurred with monotonous regularity. On February 24, 1988, the magistrate who the judge had appointed to oversee discovery held a hearing anent one such dispute. The defense representatives on the JDC sought to require parties taking depositions to identify, five days beforehand, the exhibits which they intended to utilize at deposition. Plaintiffs' representatives complained that such a paradigm, if sanctioned, would require disclosure of attorney work product. The magistrate turned a deaf ear to the protest and adopted the identification protocol.

Upon entry of the magistrate's order, the plaintiffs' steering committee (PSC), a coterie of lawyers representing the shared interests of all of the claimants, prosecuted an appeal to the district judge. See Fed.R.Civ.P. 72(a). The judge upheld the magistrate. In essence, the district court concluded that a document list of the type required was not attorney work product; that, even if such a list could be so categorized, it was at most qualifiedly privileged--and the privilege was overborne in this instance by the special needs of the sprawling litigation; and that, therefore, the order was not clearly erroneous or contrary to law. The judge appended to his affirmance a set of guidelines aimed at easing application of the identification protocol. The guidelines modified the order slightly by requiring all parties who wished to examine at the deposition to prepare and submit document lists. Yet the PSC's basic grievance was not mollified.

After the district court certified its order for interlocutory appeal under 28 U.S.C. Sec. 1292(b),1 the PSC requested that we hear the matter, and we agreed.

Appellant makes a well-constructed four-part argument which runs along the following lines: (1) PSC members sifted through millions of pieces of paper in order to locate and identify approximately 70,000 documents which they thought relevant to the litigation; (2) although the documents themselves are not protected work product, the identification protocol requires plaintiffs' lawyers to reveal to their opponents the mental processes, impressions, and opinions of the attorneys who culled the wheat from the chaff; (3) these mental processes, impressions, and opinions constitute "opinion" work product which--unlike its poor relation, "ordinary" work product--should enjoy absolute protection; and (4) inasmuch as preidentification of relevant documents necessarily divulges the results of the attorneys' selection process, the work product doctrine interdicts the challenged order. We address this quadripartite contention by examining, first, the source of the district court's authority to manage litigation and the etiology of the disputed order. We then proceed to discuss the general nature of the work product doctrine and to chart the terrain at which the trial court's case management power intersects with the demands of that doctrine. Finally, we apply the relevant principles to the matter at hand.

A. The Judicial Role. Traditionally, the American adjudicative process has been initiated and controlled by litigants--plaintiffs and defendants. The court has played the role of a guru, overseeing litigation, deciding questions of fact, ruling on points of law, and settling disputes among the parties. See generally A. Chayes, The Role of the Judge in Public Law Litigation, 89 Harv.L.Rev. 1281, 1285-86 (1976) (discussing the jurist's place in the "traditional" adjudicative model). Although trial judges were often described as more than mere moderators, e.g., Quercia v. United States, 289 U.S. 466, 469, 53 S. Ct. 698, 699, 77 L. Ed. 1321 (1933), such statements invariably applied to the judge's role in the conduct of the trial itself. See, e.g., United States v. Polito, 856 F.2d 414, 418 (1st Cir.1988). In the pretrial phases of a civil case, the judge did little heavy lifting; the parties (and their counsel) set their own agenda and were afforded considerable leeway. Prior to, and in the early days after, the adoption of the Federal Rules of Civil Procedure in 1939, this sort of Olympian oversight was generally thought sufficient to ensure the smooth operation of our adversarial system of justice.

The inauguration of the Civil Rules, however, heralded the dawning of a new day. Once absorbed by the bar, they transformed the very nature of litigation. Relaxation of the strictures governing joinder of claims and parties "shifted the focus of the lawsuit from legal theory to factual context...." Chayes, supra, at 1290; see also Fed.R.Civ.P. 13, 14, 20. The introduction of expansive pretrial discovery, Fed.R.Civ.P. 26-36, accelerated and enhanced this transmogrification. As lawyers became more adept at utilizing the liberalized rules, litigation became larger and more intricate: the joinder rules stimulated growth in the size and complexity of cases by the addition of parties and claims, and the discovery rules stimulated growth in the size and complexity of cases by multiplying the opportunities for pretrial work per party and per claim. Paper manufacturers sang hosannas.

The combined effect of these changes was explosive. Lawsuits became increasingly more complicated, and the preparatory phases of these ever-more-complex actions took on a nightmarish quality: discovery swelled to unwieldy proportions in case after case, and often became prohibitively expensive. Trial lawyering gradually became a global, rather than a local, art--and its practitioners grew ever more peripatetic. In this auxetic environment, the courts' traditional oversight powers were too limp a set of reins with which to attempt to control--let alone manage--the pretrial aspects of modern litigation. See, e.g., ACF Indus., Inc. v. EEOC, 439 U.S. 1081, 1086-87, 99 S. Ct. 865, 868-69, 59 L. Ed. 2d 52 (1979) (Powell, J., dissenting from denial of certiorari) (commenting upon need for tighter judicial control of pretrial phases of litigation). A herd of bulls had been set loose in the decorous confines of the judge's lobby--and the china shop would never be the same.

This colliquation--which threatened to rend the tissue and sinew of the judicial system--could not be allowed to go unchecked. Some response was inevitable; after all, unlike leopards, trial judges can change their spots. See, e.g., J. Pieras, Jr., Judicial Economy and Efficiency Through the Initial Scheduling Conference, 35 Cath.U.L.Rev. 943, 958 (1986) ("[w]e, the judges, must take control of pretrial procedure...."). The bench began to use its inherent powers2 to take a more active, hands-on approach to the management of pending litigation. And eventually, in 1983, more tools became available: the Civil Rules were amended to address the reality of modern litigation by supplementing the traditional powers of the trial judge with broad new powers designed affirmatively and specifically to allow the judge to control the pretrial phases of complex litigation.3 See Fed.R.Civ.P. 16 advisory committee note (1983 amendments "necessary to encourage pretrial management that meets the needs of modern litigation"). With these implements at hand, the trial judge was considerably better equipped to set a course plotted to meet the idiosyncratic needs of any pending piece of litigation.

But the changes, we suggest, go beyond the addition of a few important sentences to the Civil Rules. Notwithstanding that the court's armamentarium has been restocked, it is worth noting that these new weapons may operate within markedly different tolerances than the old. Because the courts' neoteric managerial powers stem from a transformed conception of adjudication where judicial initiative and governance play major roles in shaping the progress and extent of pretrial activity, limitations on those powers are not necessarily coextensive with the limits which historically were thought to pertain vis-a-vis the judiciary's oversight powers.

B. The Identification Protocol. Before we can define the actual limitations which, in this case, impinge upon the district court's adoption of the identification protocol, we must determine more specifically the source of the court's power to enter the challenged order. Neither the district judge nor the magistrate attempted to elucidate this point, so we must look to the nature of the order itself, and the circumstances of its interposition, for guidance.

By its terms, the order establishes "rules ... for prior identification and production of ... exhibits." Appendix to Pretrial Order No. 57, at 1. Its text tells us that these rules were formulated by the district court "[i]n the interest of expediting the taking of depositions and to afford the parties the opportunity adequately to prepare for discovery...." Id. The key provision of the protocol states in relevant part:

Any party wishing to use exhibits during the questioning of a deponent ... shall place a list of all exhibits which it intends to use during the taking of said deposition in the Joint Document Depository, at least five (5) working days before the commencement date for said deposition.

Id. The rules constrain all parties: plaintiffs and defendants, those who notice depositions and those who receive deposition notices. They prohibit counsel referring to unlisted documents during a deposition unless their usage "could not have been reasonably anticipated." Id. at 2.

It is readily apparent that this decree is not a discovery order of the genre to which we are accustomed. Traditionally, discovery orders resolve conflicts arising when a party, through the use of one of the enumerated discovery devices, seeks to gain information from an adverse party or a third person. A, for example, propounds an interrogatory to B; dissatisfied with the response--it seems incomplete, or evasive, or simply never materializes--A asks the judge to order that a proper response be served. In that sort of situation, litigant initiative mobilizes and drives the discovery engine. The trial court acts as an umpire: the judge resolves the immediate dispute between the parties and fashions relief, usually grounded in Rule 26's oversight powers, based on the nature and circumstances of the discovery requests. But the order here at issue is not cut from this staple cloth. It is unconventional, perhaps because "[t]he art of our necessities is strange." W. Shakespeare, King Lear (1606).

The protocol presently in dispute is not so much a discovery order as a case management order. Litigant initiative is of no moment: the identification protocol is a product of judicial action, not judicial reaction. Notwithstanding the PSC's insinuations to the contrary, it seems to us crystal clear that the district court crafted the challenged decree with care to serve the overall needs of the litigation rather than the individual interests of any particular party or group of parties. The resultant order is temporally and substantively pervasive; that is to say, it will guide the entire discovery phase of the litigation and leave its mark upon the taking of an estimated 2,000 depositions. It affects the parties' abilities effectively to depose persons thought to possess relevant information. It changes the plane of the playing field, but leaves the field level in that its impact is felt equally by plaintiffs and defendants.4

This focus on the systemic needs of the litigation, combined with the pervasive scope of the ensuing guidelines, persuades us that the order derives not from the district court's familiar Rule 26 "scope-of-discovery" power, but from the court's newly-augmented authority to control and manage the litigation and the course of discovery. See Fed.R.Civ.P. 16(e), 26(f). The question then becomes whether the work product doctrine limits a court's case management powers--and if so, to what extent. Asking such a question, we think, is materially different from asking how the work product doctrine restricts a court's traditional power to compel discovery from Litigant A and cause the fruits to be delivered to Litigant B.

C. The Work Product Doctrine. The work product doctrine, first recognized by the Supreme Court in Hickman v. Taylor, 329 U.S. 495, 67 S. Ct. 385, 91 L. Ed. 451 (1947), has most frequently acted as a "limitation on the nonevidentiary material which may be the subject of pretrial discovery...." United States v. Nobles, 422 U.S. 225, 243, 95 S. Ct. 2160, 2172, 45 L. Ed. 2d 141 (1975) (White, J., concurring). Although the Court has approached the doctrine in terms reminiscent of qualified privilege, its scope and effect outside the civil discovery context is largely undefined. 422 U.S. at 238, 95 S.Ct. at 2170. Fed.R.Civ.P. 26(b)(3), for example, partially codifies the work product doctrine as recognized in Hickman, but is narrowly drawn to confine its scope to its historical origins.5 Read literally, the rule applies only where a civil litigant seeks production of tangible information in another's possession. So, it is on its face inapplicable to the district court's order. Nonetheless, because the challenged decree governs the litigation's discovery phase and operates in a manner somewhat analogous to a stock discovery order--it forces one party involuntarily to disclose information to other parties--the spirit of Rule 26(b)(3), at least, should have pertinency. To assume that the work product doctrine does not apply at all when a court's order transcends the conventional discovery model would be to ignore the evolution of discovery.

It is, therefore, not surprising that the work product doctrine has found application beyond the prototypical civil discovery realm. The Supreme Court has recently recognized that the doctrine's underlying rationales are apposite in other, analogous, contexts. See Upjohn Co. v. United States, 449 U.S. 383, 397-402, 101 S. Ct. 677, 686-89, 66 L. Ed. 2d 584 (1980) (work product doctrine applicable to Internal Revenue Service summonses); United States v. Nobles, 422 U.S. at 236, 95 S.Ct. at 2169 (recognizing work product doctrine in criminal matters). Our adversarial system of justice cannot function properly unless an attorney is given a zone of privacy within which to prepare the client's case and plan strategy, without undue interference. See Upjohn Co. v. United States, 449 U.S. at 397-98, 101 S.Ct. at 686; United States v. Nobles, 422 U.S. at 238, 95 S. Ct. at 2170; Hickman v. Taylor, 329 U.S. at 511, 67 S.Ct. at 393; In re Grand Jury Subpoena, 622 F.2d 933, 935 (6th Cir.1980); Coastal States Gas Corp. v. Department of Energy, 617 F.2d 854, 864 (D.C.Cir.1980); see also Special Project, The Work Product Doctrine, 68 Cornell L.Rev. 760, 784-87 (1983) (explaining need for zone of privacy). To preserve this quiet and secluded corner for the use and benefit of attorneys in complex civil litigation, the work product doctrine must also act to some degree as a brake on the court's power to manage the discovery and pretrial phases of such litigation. Cf. Nobles, 422 U.S. at 238, 95 S. Ct. at 2170 (styling work product doctrine as "an intensely practical one, grounded in the realities of litigation in our adversary system"). But to posit applicability of the doctrine "to some degree" is not tantamount to plotting the intersecting lines; to say that some limitation exists is merely to refocus the question rather than to answer it. Because the nature and extent of such a limitation is not necessarily coterminous with those work product restrictions which accrue in the conventional discovery context, we must take a fresh look at the doctrinal infrastructure of the work product rule and the interests which shape it.

D. Obtaining Work Product. In Hickman v. Taylor, supra, the Supreme Court ruled that materials prepared "with an eye towards litigation" were not freely discoverable without some showing of necessity. 329 U.S. at 511, 67 S.Ct. at 394. See also Fed.R.Civ.P. 26(b)(3), quoted supra at note 5. The boundaries of the doctrine were mapped, originally, by balancing the systemic interest in providing lawyers with "a certain degree of privacy, free from unnecessary intrusion by opposing parties and their counsel," 329 U.S. at 510-11, 67 S.Ct. at 393, against the societal interest in ensuring that the parties obtain "[m]utual knowledge of all the relevant facts gathered...." Id. at 507, 67 S.Ct. at 392. See generally S. Cohn, The Work Product Doctrine: Protection, Not Privilege, 71 Geo.L.J. 917, 918-22 (1983).

Although the substantial need/undue hardship standard was soon accepted as basic to the work product calculus, the standard has never been applied across the board. The Hickman Court hinted broadly that some materials prepared in anticipation of litigation merited a greater degree of protection than might routinely be accorded to others. 329 U.S. at 512-13, 67 S.Ct. at 394. The draftsmen of Fed.R. 26(b)(3) also recognized a distinction between types of work product: even where the need/hardship hurdle has been cleared, the court must continue to afford "protect[ion] against disclosure of the mental impressions, conclusions, opinions or legal theories of an attorney...." Id. Striations of this kind have led courts to distinguish between "opinion" work product and "ordinary" work product--the former category encompassing materials that contain the mental impressions, conclusions, opinions or legal theories of an attorney, the latter category embracing the residue. See, e.g., Sporck v. Peil, 759 F.2d 312, 316 (3d Cir.), cert. denied, 474 U.S. 903, 106 S. Ct. 232, 88 L. Ed. 2d 230 (1985); In Re Murphy, 560 F.2d 326, 329 n. 1 (8th Cir.1977); Duplan Corp. v. Moulinage et Retorerie de Chavanoz, 509 F.2d 730, 732 (4th Cir.1974), cert. denied, 420 U.S. 997, 95 S. Ct. 1438, 43 L. Ed. 2d 680 (1975). The dichotomy--or some form of it--has now seemingly been recognized by the Court, see Upjohn Co. v. United States, 449 U.S. at 399-400, 101 S.Ct. at 687-688, but the exact import and dimensions of the approach remains tenebrous. Courts typically afford ordinary work product only a qualified immunity, subject to a showing of substantial need and undue hardship, while requiring a hardier showing to justify the production of opinion work product. See, e.g., In Re Grand Jury Investigation, 599 F.2d 1224, 1231 (3d Cir.1979). Indeed, some courts have seemingly concluded that the protection for certain types of opinion work product is ironclad. See, e.g., In Re Grand Jury Proceedings, 473 F.2d 840, 848 (8th Cir.1973) (lawyer's personal recollections, notes, and memoranda pertaining to witness interviews safeguarded absolutely). The Supreme Court, however, has made no commitment concerning the correct standard for revelation of opinion work product. See Upjohn Co. v. United States, 449 U.S. at 401, 101 S.Ct. at 688 (specifically reserving question).

For today we merely note--but do not address--this distinction. For the reasons mentioned below, see infra Part II(E), we are satisfied that only ordinary work product is involved in the identification protocol. The standard for disclosure, therefore, would seem at first blush to be that of balancing substantial need against undue hardship. Yet even that standard, useful as it may be in the discovery context, appears maladroit when the source of the intrusion is not a discovery order but a case management order. The evolution of the work product doctrine--which sprouted and grew in the fruited plains of pretrial discovery--explains what we see as a lacuna: the need/hardship balance, such as is precisely enunciated in Rule 26(b)(3), has traditionally been a barometer only of the relative interests of the opposing litigants. This makes abundant good sense for the resolution of discovery rhubarbs, but leaves the circle unclosed in the Rule 16 milieu. Insofar as we can tell, concerns for maximizing the efficient use of judicial and litigant resources (such as are served by broadening judicial management powers) have never heretofore been a relevant consideration in formulating the work product doctrine. Given the rigors of modern-day litigation, and the increased emphasis on case management, the omission looms as intolerable.

When case management, rather than conventional discovery, becomes the hammer which bangs against the work product anvil, logic demands that the district judge must be given greater latitude than provided by the routine striking of the need/hardship balance. Because of "the taxing demands of modern-day case management," Recticel, supra, at 1007, the requirements of the litigation and the court must, we think, be weighed in determining whether a management technique impermissibly impinges upon the protected zone of work product privacy. In this context, the vista is not exclusively head-to-head, A against B, plaintiff versus defendant; the relationship is triangular, with the court itself as a third, important, player. There is no reason, then, why the crying need for efficient use of scarce judicial resources cannot--and should not--be factored into the equation. We hold that it must.

E. Classification of the Lists. Having refashioned the geometry of the weighbeam, we turn to an evaluation of the work product interest which the PSC asks us to place on the scales. We begin with an abecedarian verity: not every item which may reveal some inkling of a lawyer's mental impressions, conclusions, opinions, or legal theories is protected as opinion work product. Were the doctrine to sweep so massively, the exception would hungrily swallow up the rule. See Sporck v. Peil, 759 F.2d at 319 (Seitz, J., dissenting) ("[e]very act by a litigant or his attorney gives rise to ... vague inferences" as to strategy and counsel's thought processes). Whatever heightened protection may be conferred upon opinion work product, that level of protection is not triggered unless disclosure creates a real, nonspeculative danger of revealing the lawyer's thoughts. See Gould, Inc. v. Mitsui Mining & Smelting Co., 825 F.2d 676, 680 (2d Cir.1987); Research Institute for Medicine and Chemistry, Inc. v. Wisconsin Alumni Research Foundation, 114 F.R.D. 672, 680 (W.D.Wis.1987). There is, moreover, a second line of demarcation. Some materials do not merit heightened protection because, despite the revelations they contain as to an attorney's thought processes, the lawyer has had no justifiable expectation that the mental impressions revealed by the materials will remain private.6

Consider, for a moment, an answer to a complaint. The drafting of such a pleading--admitting certain averments, denying the remainder; advancing selective affirmative defenses, eschewing others; asserting one counterclaim, omitting another--certainly implicates counsel's opinions, ideas, thoughts, and strategy. Filing and service of the answer discloses those mental impressions to the opposing party with some appreciable degree of clarity. Yet it would be foolhardy to urge that the contents of the answer should be enswathed in cotton batting. Because pleadings are drawn with the realization that they will be served upon the other parties to the case, it has never been seriously suggested that an answer could be hidden from view under the work product rubric. So, too, countless other legal documents generated in the ordinary course of litigation: complaints, counterclaims, third-party pleadings, requests for admissions, interrogatories, motions, supporting affidavits, and the like. Even when such documents are prepared specially, at the direction of the court, the outcome is the same.

To a certain extent, discovery responses can be viewed through the same glass. The Civil Rules illustrate this proposition by allowing litigants regularly to gain revelatory information from their opponents in certain circumstances. For example, Fed.R.Civ.P. 33(b) permits a party to serve interrogatories which require the opposing party, in framing an answer, to render an opinion or spell out a contention that relates to the application of law to fact; Fed.R.Civ.P. 36(a) validates requests for admissions that relate to opinions of divers kinds; Fed.R.Civ.P. 26(b)(4) enables a resourceful litigant to obtain important insights into opposing counsel's opinions and legal theories, providing, inter alia, for disclosure of the identity of expert witnesses and production of the "substance of the facts and opinions to which the expert is expected to testify and a summary of the grounds for each opinion." Id. The key factor distinguishing such materials from protected opinion work product is that, even absent compelled disclosure, the information will probably come to light during the course of trial, if not before. The rules do not compel the disclosure of matters which would likely remain inviolate in the bosom of the lodge; they merely accelerate disclosure of matters that would probably be revealed in due course. As to materials of the latter ilk, the attorney--though he has invested his ratiocinatory capital--has a relatively low expectation of ultimate privacy; indeed, precisely the reverse is often true. When framing, say, a response to a contention interrogatory, the lawyer--far from anticipating that the inferences gleanable from his preparation will remain the darkest of secrets--knows that the inferences will become common knowledge among those connected with the litigation. In real-world terms, the better the lawyer, the more he is likely to seek, whether or not sotto voce, to raise and advance his theory of the case through the answer.

On the other hand, efficacious operation of the judicial system has much to gain by expedition of the disclosure in such circumstances. Time and effort are conserved, and no meaningful intrusion takes place. Thus, the overall balance of equities plainly favors making what amounts to a timing adjustment, in the process treating such materials as something less than fully-protected opinion work product. Cf. Fed.R.Civ.P. 26(b)(3) advisory committee note (although party required to divulge mental impressions and conclusions, lawyer remains "entitled to keep confidential documents containing such matters prepared for internal use ") (emphasis supplied).

A reconstructed balance of this kind is equally adaptable to the looming collision between legitimate work product concerns and modern case management techniques. Courts, pursuant to the powers granted by Rule 16, routinely require parties to file pretrial memoranda or otherwise to identify witnesses, pre-mark exhibits, define contentions, and spell out their legal theories. See, e.g., Spray-Rite Service Corp. v. Monsanto Co., 684 F.2d 1226, 1245 (7th Cir.1982) (compilation and exchange of witness lists), aff'd, 465 U.S. 752, 104 S. Ct. 1464, 79 L. Ed. 2d 775 (1984); Elliot v. Louisiana Power & Light Co., 671 F.2d 865, 868-69 (5th Cir.1982) (same; exhibit lists); cf. Hernandez v. Alexander, 671 F.2d 402, 407 (10th Cir.1982) (legal theory not elucidated in agreed pretrial order not cognizable thereafter). See also Note, Pretrial Conference: A Critical Examination of Local Rules Adopted by Federal District Courts, 64 Va.L.Rev. 467, 469 (1978). Like the provisions of the Civil Rules anent discovery, such case management devices do not decide whether certain aspects of the attorney's thought processes will be disclosed, but simply determine when the disclosure will occur. In other words, they do little more than change the timing, expediting revelations which would ultimately be made during the normal course of pretrial proceedings or trial itself. Since these incursions into the attorney's mental impressions are inevitable, any violation of the zone of privacy is marginal, and the sturdier prophylaxis given to opinion work product is neither needed nor warranted.

In our view, the exhibit lists demanded by the district court's identification protocol fall well within this less-shielded category. The PSC concedes that the documents themselves are nonprivileged and that, apart from work product connotations, the lists are not otherwise eligible for special swaddling. More to the point, the challenged order does not result in the evulgation of matters which would otherwise remain perpetually hidden. When the deposition is held and examination commences, the questioner's document selection, and the stratagems it reveals, will become obvious to all. Requiring preidentification merely moves up the schedule, accelerating disclosures which would inevitably take place. Consequently, the resultant lists cannot validly aspire to the stature of opinion work product, nor can they command the correlative degree of (heightened) protection.

We recognize, of course, that the process of selecting relevant documents for use in depositions "is often more crucial than legal research." Shelton v. American Motors Corp., 805 F.2d 1323, 1329 (8th Cir.1986). But as we have already pointed out, the information spotlighted by such lists (including whatever insights may be gleaned from the choice of exhibits) will undeniably be gained during the course of the deposition. Like requiring pleadings, answers to contention interrogatories, pretrial exhibit and witness lists, and trial memoranda, the district court's identification protocol merely adjusts the timing of disclosure. The situation is not remotely analogous to the situation where a party seeks an attorney's personal notes and memoranda which contain his confidential assessments of the testimony of prospective witnesses. See Hickman, 329 U.S. at 512-13, 67 S. Ct. at 394. Such notes and memoranda are usually prepared solely for the attorney's own use--or at most, for confidential consideration between attorney, client, and their privies (e.g., junior counsel, investigators, retained experts). Under ordinary circumstances, the lawyer can expect that such materials will never be subject to his opponents' scrutiny--or at least, that he can effectively control whether or not such dissemination will occur. In contrast, no lawyer can be so sanguine as to expect that the opposition will not become privy to his choice of deposition exhibits; the exhibits are integral to the taking of the deposition and will, by definition, have to be revealed during the session.

In exhorting a contrary conclusion, appellant relies heavily on Sporck v. Peil and Shelton v. American Motors Corp., both supra. In Sporck, a divided panel of the Third Circuit concluded that, "the selection and compilation of documents by counsel ... in preparation for pretrial discovery falls within the highly-protected category of opinion work product." 759 F.2d at 316. In Shelton, the Eighth Circuit, relying to an extent on Sporck, reached a somewhat similar result. See 805 F.2d at 1329. We find the PSC's reliance on these decisions to be mislaid.

In Sporck, the majority emphasized that "[i]n selecting and ordering a few documents out of thousands, counsel could not help but reveal important aspects of his understanding of the case." 759 F.2d at 316 (quoting James Julian, Inc. v. Raytheon Co., 93 F.R.D. 138, 144 (D.Del.1982)). This reasoning, we suggest, is flawed because it assumes that the revelatory nature of the sought-after information is, in itself, sufficient to cloak the information with the heightened protection of opinion work product. That is simply not the case; much depends on whether the fruits of the screening would soon be revealed in any event. See supra. Indeed, Sporck should be distinguished because, unlike in this case, the lawyer's selection process there was never designed to see the light of day; the exhibits had been selected not for use in examination of an adverse or neutral witness, but for a markedly more private purpose--preparation of the attorney's own client. We believe the distinction is a critical one. And Sporck, which dealt with a classic one-on-one discovery dispute--plaintiff's attempt to secure the results of defense counsel's document triage from a defendant during an adversary deposition--is also distinguishable because it did not in any way concern the district court's case management powers.

The Sporck case is off the point for a third reason as well. There the majority placed great weight on the assumption that "without the protection that the work product doctrine accords his preparation, defense counsel may have forgone a sifting of the documents [or not used them to prepare the deposition witness]." 759 F.2d at 317. Even if this were so--and we think it unlikely that an attorney will forgo an integral part of his preparation due to his inability to shield that preparation as opinion work product--it has no pertinence in this case. Nothing compels a lawyer to notice a discovery deposition, or to examine at one; nothing compels a lawyer who chooses to participate to use documents in the course of his examination of the deponent. In short, each lawyer can weigh the advantages and disadvantages of disclosing the identity of the exhibits he selected, and act accordingly.

Shelton, also decided by a divided panel, is similarly unhelpful for our purposes. In that case, the deponent was defendant's counsel; her mental selective process in determining which documents to review and to "rely[ ] on ... in preparing her client's case," id. at 1329, was never meant to be placed on public display. Furthermore, Shelton involved a traditional discovery expedition by one party into the other's territory; it did not deal with the peculiar demands of case management. Lastly, the panel did not differentiate between opinion work product and ordinary work product. The majority held only that the document selection process comprised work product, 805 F.2d at 1329, a conclusion with which we agree. Insofar as Shelton can be read to sweep more broadly--and we do not recommend or suggest such a reading--it does so in reliance on Sporck, and is thus subject to the distinctions limned above.

For these reasons, we rule that compelled disclosure of document lists under the district court's identification protocol does not implicate opinion work product and thus does not constitute an impermissible per se intrusion into the lawyer's protected zone of privacy.

F. Calibrating the Scales. Notwithstanding our conclusion that the exhibit lists do not comprise opinion work product, we recognize that attorneys and their staffs sorted and segregated the documents in anticipation of litigation. We do not deny that a glimpse of the selection process's yield provides insight into opposing counsel's understanding of his case. Forced production of an exhibit list makes some incursion into the attorney's quiet and secluded corner and consequently implicates the work product rule to an extent--but treating the lists as ordinary work product fully safeguards whatever legitimate privacy concerns are at stake. Adopting this taxonomy, we find, as did the court below, that the warranted degree of protection was clearly overbalanced by the exigencies of the case.

We start this phase of our analysis with a frank acknowledgement that the district court's discretion in synthesizing the ingredients of a case management order involving ordinary work product is broad. See Recticel, supra, at 1006 (particularly in complex litigation, "[d]ecisions regarding the scope of discovery ... and the protections to be afforded parties in the discovery process, are ordinarily left to the informed judgment of the district judge"). In that context, new ingredients (the court's requirements and the likely systemic effect of the order) supplement the staples (the parties' cognizable needs and the degree of hardship visited upon them) in composition of the recipe. Moreover, the trial judge, qua case manager, is in the optimal position to mix and measure those diverse ingredients: he is the sous-chef, with all the contents of the ragout simmering before him. See id. at 1006 (district court "is in a unique position to gauge and balance the potentially conflicting interests at stake"). Whatever their appetite for error, appellate courts have a far less intimate knowledge of what transpires in the trial judge's kitchen.

The district court's discretion, to be sure, is not unlimited--but the judge abuses it only "when a relevant factor that should have been given significant weight is not considered, when an irrelevant or improper factor is considered and given significant weight, or when all proper and no improper factors are considered, but the court in weighing those factors commits a clear error of judgment." United States v. Hastings, 847 F.2d 920, 924 (1st Cir.1988) (quoting United States v. Kramer, 827 F.2d 1174, 1179 (8th Cir.1987)). Given this standard, the court below was comfortably within its rights in concocting the identification protocol.

We need not belabor the point. It is, we think, readily apparent that the scope of this massive litigation and its consequent special needs called for unorthodox measures. The district court observed that over 2,000 depositions would be completed before the end of discovery and found that:

The benefits in terms of time saved by prior identification of exhibits, in view of such a large number of depositions, are undeniable. Among other things, such a rule: (1) avoids unnecessary waste of time required for review of the exhibit during the taking of the deposition; and (2) promotes the speedy resolution of objections to proposed exhibits either through agreement between the parties or by presentation of all objections to the Magistrate at once, rather than one by one during the course of the deposition.

Pretrial Order No. 57 at 7 (footnote omitted). The district court expressly found that opposing parties could not, without undue hardship, obtain the substantial equivalent of the information provided by compliance with the identification protocol. Id. And in the court's view, the parties had manifested a substantial need for deposition exhibit lists arising out of the time and delay inherent in reviewing and raising objections to deposition exhibits during, rather than before, each deposition. Id. The order, then, had a systemically beneficial component: not only would litigants' burdens be eased, but the court's ability to administer, process, and respond adequately to the case's ebb and flow would be materially enhanced.

It would be entirely unwarranted for us to disturb this thoughtful decision. The managerial gains attributable to the identification protocol are obviously substantial. Indeed, prior identification of deposition exhibits has been specifically recommended as an appropriate means of facilitating discovery. See Manual for Complex Litigation 2d, Sec. 21.456 (1985). The Manual suggests that "[d]epositions may be significantly expedited by requiring that, unless surprise is important for impeachment or similar purposes, opposing counsel and the deponent shall be notified in advance of the documents about which the deponent will be examined." Id.

Appellant attacks the district court's order because it omits a generic "surprise" exception--yet this phalanx of the PSC's assault traverses especially rocky ground. In the first place, the order, though making no overall exception for "surprise," is susceptible to modification in specific instances, for cause shown. See Fed.R.Civ.P. 26(f) (order implementing management of discovery "may be altered or amended whenever justice so requires"); see also Fed.R.Civ.P. 16(b) (scheduling order subject to modification "upon a showing of good cause"); Fed.R.Civ.P. 16(e) (pretrial orders subject to modification). More fundamentally, any protection due the document selection process must stem from its status as work product. The PSC can cite no authority for the proposition that maintaining an element of surprise is an important object of the work product doctrine. Such a thesis, in fact, is completely antithetic to Hickman's admonition that "[m]utual knowledge of all the relevant facts gathered by both the parties is essential to proper litigation." 329 U.S. at 507, 67 S. Ct. at 392. The Court's observation that the "deposition-discovery procedure simply advances the stage at which the disclosure can be compelled from the time of trial to the period preceding it, thus reducing the possibility of surprise," id. (emphasis supplied), cuts in the same direction. The sporting theory of justice was abandoned with the advent of the Civil Rules, see, e.g., Martin v. Reynolds Metals Corp., 297 F.2d 49, 56 (9th Cir.1961), and we see no reason to resurrect it in the present setting. In our opinion, the trial judge had leeway to accept or reject a generic "surprise" exemption in his design of the challenged order.

That is not to say, however, that the element of surprise no longer deserves any role in the examination of a witness. In certain very special circumstances, as where vital credibility issues can be demonstrated to exist, an attorney should have the chance not to telegraph a likely major punch. The opportunity to hear a hostile witness's testimony before it has been tailored to suit other information can be a useful truth-finding measure, even during pretrial discovery, if, for example, the witness has a strong incentive to lie and can reasonably be suspected, in advance, of an intention to do so. But the technique is frequently overdone: if inconsistencies emerge, they are probably due more to memory lapses or inartful questioning than to studied prevarication. It is a fair assumption, we think, that the circumstances requiring a "surprise" exception will be on an order of rarity comparable to black swans, cf. Juvenal, Satires, VI, 1.165, and that the showing to obtain one will have to be powerful. Still, the rare situation will arise and, when it does, the trial judge should not hesitate to act if a proper motion is made in advance.7 That flexibility, we believe, is preserved by the instant order.

We can conclude this aspect of our inquiry with little added ado. Mindful of the enormity of the litigation, and of its complexity, the fashioning of a more serviceable judicial handle on the case seems a consummation devoutly to be wished. Thus, we deem it important that no other, less intrusive means of reaping the benefits which inhere in the identification protocol come readily to mind. (Certainly, the PSC has suggested none.) All in all, we have no reason to second-guess the trial court's reasoned conclusion that the balance of relevant equities counsels in favor of the identification protocol: as the district judge supportably found, the parties have a substantial need to obtain the materials; equivalent information is unobtainable through other means without undue hardship; and the court's ability successfully to manage the litigation will be hampered in the absence of the protocol. Appellant has pointed out no relevant factor which the district court failed to consider, nor any factor which it improperly included in the mix. The record leaves us with a firm and abiding conviction that, given the unusual dimensions of this behemoth, the disputed order makes eminently good sense. The PSC has failed to demonstrate any misuse of judicial discretion.

We need go no further. The district court had adequate power pursuant to Rules 16(e) and 26(f) to order those choosing to take part in a given discovery deposition to provide, five days prior to the deposition, a list of exhibits to be utilized during the questioning. Although the work product rule can truncate the usual sweep of a court's powers, there is no requirement that case management orders which impinge upon work product be barred absolutely.

In this instance, work product considerations do not interdict the district judge's order. The information provided to opposing parties by the order's operation is ordinary work product, not opinion work product. In the context of a complex case, the court, if the needs of the litigation and the litigants reasonably so dictate, has broad discretion to command production of materials constituting ordinary work product. Given the special requirements of this mammoth collection of consolidated suits and the particularized findings which were made below, we conclude that the district court's calibration of the scales should not be disturbed. The court had power, authority, and sound reason to impose the preidentification condition.

AFFIRMED.