Concept
Ordinary and Opinion Work Product
Work-product protection
Ordinary and Opinion Work Product
Hickman protected a lawyer’s file at two different strengths, though it never gave them names. The survivors’ signed statements stood open to a showing of need, and the plaintiff lost only because he offered none.1 The lawyer’s accounts of what witnesses told him orally fared differently: on those facts the Court did not believe any showing of necessity could justify production.2 The Court thus drew a line between factual material and work product reflecting the lawyer’s mental impressions, and it made the second kind, for practical purposes, immune from discovery.3
Rule 26(b)(3) preserved both strengths through two separate commands. The first lets a requester reach qualifying material by showing substantial need and undue hardship. The second applies even after that showing succeeds, and it directs the court to protect the mental impressions, conclusions, opinions, and legal theories of a party’s attorney or other representative. The commands produce the two categories of this section: ordinary work product, also called fact work product, and opinion work product, also called core work product. Classification matters because it sets the price of access, and the reasons for protection explain why the prices differ. A requester who proves it cannot recreate the facts no longer takes a free ride on its opponent’s investigation, so a strong enough showing can open fact material. The reasons for shielding a lawyer’s evaluation do not fade that way: candid writing is still at stake, and the requester’s own lawyer can supply an evaluation, so the case for opening it is far harder to make.
Real documents mix the two tiers, so classification works on portions rather than labels. A witness-interview memorandum may recount what the witness said and, a few lines later, record the lawyer’s judgment that the witness was evasive. A court can order the first passage produced and keep the second shielded. An answer should therefore sort the content of the document before it chooses a standard.
Ordinary Work Product
Hickman did not make a lawyer’s file untouchable, and the passage that says so defines the first tier. The Court cautioned that not all written materials an adversary’s counsel obtains or prepares with an eye toward litigation are free from discovery in every case: where relevant, non-privileged facts remain hidden in an attorney’s file and their production is essential to preparing the case, discovery may properly be had.4 The Court also put the burden on the party who would invade the lawyer’s privacy to establish adequate reasons for production.5 Material in this tier is now called ordinary work product, or fact work product: tangible work product that includes facts but no mental impressions of an attorney.6 Its protection is qualified, so an opponent can reach it only by proving need, a showing developed in the next section.7
What a document says, not who wrote it, places it in this tier. Ordinary work product consists chiefly of facts and witness statements, mostly information learned from witness interviews or other investigation.8 Hickman supplies the model. Fortenbaugh took the survivors’ signed statements “with an eye toward the anticipated litigation,” and those statements recount what the survivors saw without recording how he weighed it.9 A verbatim transcript of an interview, a photograph or diagram of an accident scene, and a client’s written chronology belong with them, because each records what happened and none records anyone’s evaluation of it.
A lawyer who discloses facts loses little. Witnesses’ names, dates, and times are available to the other side in most cases anyway, so the lawyer’s objection is chiefly to giving the opponent a free ride on the investigation, and confidentiality adds little.10 That objection has a built-in limit, because a free ride assumes the opponent could have made the trip itself. An opponent who truly cannot replicate the material borrows no one’s effort, and Hickman recognized that production might be justified where witnesses are no longer available or can be reached only with difficulty.11 A lawyer’s evaluations stand differently. A lawyer is likely to record candid views only when confident they will stay private, and the policies behind the doctrine press hardest there.12
A routine insurance claim shows the line on two documents from one file. Suppose an insurer interviews a non-party eyewitness to a crash and has the statement transcribed. The transcript was prepared in anticipation of litigation by a party’s representative, so it is protected, and the injured plaintiff can obtain it only by showing substantial need and inability to get the substantial equivalent without undue hardship. Because the plaintiff can interview the eyewitness directly, that showing would probably fail.13 Now suppose the driver’s lawyer reads the transcript and writes a memo to the file concluding that the witness appears to be lying for three stated reasons. The memo reflects the lawyer’s mental impressions and conclusions, and the showing of need that might open the transcript will not open the memo.14 The transcript reports what a witness said; the memo reports what the lawyer thinks of the witness.
A file-sorting problem tests classification before it tests need, so each item should be sorted by what it says.
Sorting a document does not turn its factual remainder into an open record. In In re Grand Jury Subpoenas, 265 F. Supp. 2d 321 (S.D.N.Y. 2003), a law firm withheld nineteen documents as work product. The government’s first answer was that the documents were not work product at all, because it sought no materials revealing the attorneys’ mental impressions. That answer misses the structure of the doctrine. Materials prepared in anticipation of litigation are discoverable only on a showing of substantial need and inability to obtain the substantial equivalent without undue hardship, and opinion work product is discoverable, if at all, only on a significantly stronger showing.15 Students often read ordinary as a synonym for unprotected. The case shows the opposite: ordinary work product stays shielded until the requester carries its burden.
Opinion Work Product
Rule 26(b)(3) adds a second command to the need-and-hardship showing: when a court orders discovery of trial-preparation materials because the requester has made that showing, it must still protect against disclosure of the mental impressions, conclusions, opinions, or legal theories of a party’s attorney or other representative concerning the litigation.17 Material that reveals that thinking is opinion work product, and it sits in the most protected category.18 Treatises describe it, also called core work product, as almost wholly immune from discovery.19 The words “other representative” matter, because an investigator, consultant, or insurance adjuster can record evaluations as readily as a lawyer can, and the Rule protects the thinking whoever writes it down. The showing that opens ordinary work product therefore does not by itself open this layer. A requester who has cleared the first tier has the facts in the file and still lacks the lawyer’s evaluation of what they mean.
The protection comes from Hickman, which treated a lawyer’s account of oral statements differently from the signed statements. As to what witnesses told Fortenbaugh orally, whether it survived as his mental impressions or as memoranda, the Court did not believe any showing of necessity could be made under the circumstances of that case to justify production.20 Two reasons support that treatment. The first is distortion. Forcing an attorney to repeat or write out what witnesses told him invites inaccuracy and untrustworthiness, and even a perfect recollection would come out in his language, permeated with his inferences, so that the first variation in the witness’s later testimony would let the opponent use the lawyer’s statement to impeach.21 The second is candor. A lawyer is likely to record frank evaluations only when confident they will stay private, so a rule opening them to the adversary would leave them unwritten and the preparation poorer for it.22 United States v. Nobles places the point at the core of the doctrine, which shelters the attorney’s mental processes and gives the lawyer a privileged area in which to analyze and prepare the client’s case.23 The adversary can usually obtain the facts from witnesses, interrogatories, and depositions, so what it would gain from the evaluation is a view of how the other side reads the case, and that is exactly what the doctrine withholds.24
Upjohn Co. v. United States set the working standard. The Government sought a company’s notes and memoranda of attorney interviews with employees. It stressed that the interviewees were scattered across the globe and that the company had forbidden its employees to answer questions it considered irrelevant, and it invoked Hickman’s remark that production might be justified where witnesses are hard to reach.25 The Court answered that the remark did not reach oral statements made by witnesses, whether preserved in the attorney’s mental impressions or in memoranda, because compelled disclosure of such memoranda would reveal the attorney’s mental processes.26 The magistrate had applied the substantial need and undue hardship standard from the first part of Rule 26(b)(3), and the Court held that standard wrong for this material: to the extent the notes did not reveal privileged communications, they revealed the attorneys’ mental processes in evaluating those communications, and such work product cannot be disclosed simply on that showing.27 The Court declined to say such material is always protected, but it held that a far stronger showing of necessity and unavailability by other means than the Government made would be needed to compel disclosure, and it remanded.28 The structure of that holding is the lesson. The Rule’s showing is a gate for facts, and the protective command for mental impressions is a second gate that the first showing does not open.
Because Upjohn left open whether the second gate ever opens, the lower courts divide, and an answer should present both positions. The Court noted that some courts treat protection of work product based on witnesses’ oral statements as absolute, while others that reject an absolute rule allow discovery only in a rare situation, and it expressly declined to decide between them.29 The first camp reads the Rule’s command as a bar. In In re Grand Jury Proceedings, 473 F.2d 840 (8th Cir. 1973), the court found a lawyer’s notes and memoranda of conversations with witnesses absolutely, not conditionally, protected, and the Fourth Circuit has said that opinion work product is never discoverable.30 The second camp reads Hickman’s reference to a rare situation and Upjohn’s call for a far stronger showing as leaving a narrow door open, and its best-known formulation allows disclosure when the mental impressions are the pivotal issue in the litigation and the need for them is compelling.31 Adlman, the Second Circuit decision applying Upjohn, declined to choose. It said it need not decide whether opinion work product is ever discoverable on a showing of necessity and unavailability by other means, and that at a minimum the Rule requires such material to be protected unless a highly persuasive showing is made.32
The contrast between Holmgren v. State Farm Mutual Automobile Insurance Co. and Adlman shows what a persuasive showing looks like. In Holmgren the plaintiff sued an insurer for settling an insurance claim in bad faith, and the insurer withheld a claims adjuster’s report evaluating its potential liability on the underlying claim. The court found the report to be opinion work product and still ordered it produced, because the adjuster’s thought processes and opinions were directly at issue, necessary to establish bad faith, and unavailable by any other means.33 In Adlman the IRS argued that a memorandum was necessary to show why a company restructured and whether a legitimate business purpose existed. The court rejected the argument, because its review showed that the memorandum recorded the legal analysis of the company’s accountants, not the motives of its executives, so it could not supply what the IRS wanted.34 In Holmgren the evaluation was the thing the claim turned on; in Adlman the thing the IRS wanted lay elsewhere, and the memorandum did not contain it. A memorandum weighing the strengths and weaknesses of prospective trial witnesses is the usual Adlman-side case, and many courts would withhold the lawyer’s impressions on any showing, because nothing in the lawsuit turns on what the lawyer thought.35
The third step decides a contested case, and the fourth is where most real documents end up. Redaction is the usual way courts honor the protection in a mixed document, because it gives the requester every fact it has proved it needs and keeps the evaluative layer out of the production. The lawyer’s evaluations of the evidence and of a witness’s credibility are the standard examples of material to strike, and a court that orders the rest produced should let the lawyer produce a redacted version.36
Some courts go further and refuse even a redacted production, on the theory that the lawyer’s choice of which verbatim statements to include itself reflects the lawyer’s mental impressions or legal theories.37 Two limits keep the category from growing beyond its reason. Protection reaches only material that actually reveals the lawyer’s thinking, and courts have said that not every item hinting at some inkling of a lawyer’s impressions counts as opinion work product. The protection also yields to the crime-fraud exception, which courts have recognized for work product as for the attorney-client privilege, even where the item is opinion work product.38 An answer that stops at substantial need and undue hardship will wrongly order production of the lawyer’s evaluation, so it must sort the content of the document first and then apply the stronger standard to the evaluative portions.