concurring.
The narrow question in this case concerns only one of thirty-nine interrogatories which defendants and their counsel refused to answer. As there was persistence in refusal after the court ordered them to answer it, counsel and clients were committed to jail by the district court until they should purge themselves of contempt.
The interrogatory asked whether statements were taken from the crews of the tugs involved in the accident, or of any other vessel, and demanded “Attach hereto exact copies of all such statements if in writing, and if oral, set forth in detail the exact provisions of any such oral statements or reports.” The question is simply whether such a demand is authorized by the rules relating to various aspects of “discovery.”
The primary effect of the practice advocated here would be on the legal profession itself. But it too often is over*515looked that the lawyer and the law office are indispensable parts of our administration of justice. Law-abiding people can go nowhere else to learn the ever changing and constantly multiplying rules by which they must behave and to obtain redress for their wrongs. The welfare and tone of the legal profession is therefore of prime consequence to society, which would feel the consequences of such a practice as petitioner urges secondarily but certainly.
“Discovery” is one of the working tools of the legal profession. It traces back to the equity bill of discovery in English Chancery practice and seems to have had a forerunner in Continental practice. See Ragland, Discovery Before Trial (1932) 13-16. Since 1848 when the draftsmen of New York’s Code of Procedure recognized the importance of a better system of discovery, the impetus to extend and expand discovery, as well as the opposition to it, has come from within the Bar itself. It happens in this case that it is the plaintiff’s attorney who demands such unprecedented latitude of discovery and, strangely enough, amicus briefs in his support have been filed by several labor unions representing plaintiffs as a class. It is the history of the movement for broader discovery, however, that in actual experience the chief opposition to its extension has come from lawyers who specialize in representing plaintiffs, because defendants have made liberal use of it to force plaintiffs to disclose their cases in advance. See Report of the Commission on the Administration of Justice in New York State (1934) 330-31; Rag-land, Discovery Before Trial (1932) 35-36. Discovery is a two-edged sword and we cannot decide this problem on any doctrine of extending help to one class of litigants.
It seems clear and long has been recognized that discovery should provide a party access to anything that is evidence in his case. Cf. Report of Commission on the Administration of Justice in New York State (1934) 41-42. *516It seems equally clear that discovery should not nullify the privilege of confidential communication between attorney and client. But those principles give us no real assistance here because what is being sought is neither evidence nor is it a privileged communication between attorney and client.
To consider first the most extreme aspect of the requirement in litigation here, we find it calls upon counsel, if he has had any conversations with any of the crews of the vessels in question or of any other, to “set forth in detail the exact provision of any such oral statements or reports.” Thus the demand is not for the production of a transcript in existence but calls for the creation of a written statement not in being. But the statement by counsel of what a witness told him is not evidence when written. Plaintiff could not introduce it to prove his case. What, then, is the purpose sought to be served by demanding this of adverse counsel?
Counsel for the petitioner candidly said on argument that he wanted this information to help prepare himself to examine witnesses, to make sure he overlooked nothing. He bases his claim to it in his brief on the view that the Rules were to do away with the old situation where a law suit developed into “a battle of wits between counsel.” But a common law trial is and always should be an adversary proceeding. Discovery was hardly intended to enable a learned profession to perform its functions either without wits or on wits borrowed from the adversary.
The real purpose and the probable effect of the practice ordered by the district court would be to put trials on a level even lower than a “battle of wits.” I can conceive of no practice more demoralizing to the Bar than to require a lawyer to write out and deliver to his adversary an account of what witnesses have told him. Even if his recollection were perfect, the statement would be his lan*517guage, permeated with his inferences. Every one who has tried it knows that it is almost impossible so fairly to record the expressions and emphasis of a witness that when he testifies in the environment of the court and under the influence of the leading question there will not be departures in some respects. Whenever the testimony of the witness would differ from the “exact” statement the lawyer had delivered, the lawyer’s statement would be whipped out to impeach the witness. Counsel producing his adversary’s “inexact” statement could lose nothing by saying, “Here is a contradiction, gentlemen of the jury. I do not know whether it is my adversary or his witness who is not telling the truth, but one is not.” Of course, if this practice were adopted, that scene would be repeated over and over again. The lawyer who delivers such statements often would find himself branded a deceiver afraid to take the stand to support his own version of the witness’s conversation with him, or else he will have to go on the stand to defend his own credibility — perhaps against that of his chief witness, or possibly even his client.
Every lawyer dislikes to take the witness stand and will do so only for grave reasons. This is partly because it is not his role; he is almost invariably a poor witness. But he steps out of professional character to do it. He regrets it; the profession discourages it. But the practice advocated here is one which would force him to be a witness, not as to what he has seen or done but as to other witnesses’ stories, and not because he wants to do so but in self-defense.
And what is the lawyer to do who has interviewed one whom he believes to be a biased, lying or hostile witness to get his unfavorable statements and know what to meet? He must record and deliver such statements even though he would not vouch for the credibility of the witness by calling him. Perhaps the other side would not want to *518call him either, but the attorney is open to the charge of suppressing evidence at the trial if he fails to call such a hostile witness even though he never regarded him as reliable or truthful.
Having been supplied the names of the witnesses, petitioner’s lawyer gives no reason why he cannot interview them himself. If an employee-witness refuses to tell his story, he, too, may be examined under the Rules. He may be compelled on discovery, as fully as on the trial, to disclose his version of the facts. But that is his own disclosure — it can be used to impeach him if he contradicts it and such a deposition is not useful to promote an unseemly disagreement between the witness and the counsel in the case.
It is true that the literal language of the Rules would admit of an interpretation that would sustain the district court’s order. So the literal language of the Act of Congress which makes “any writing or record . . . made as a memorandum or record of any . . . occurrence, or event” admissible as evidence, would have allowed the railroad company to put its engineer’s accident statements in evidence. Cf. Palmer v. Hoffman, 318 U. S. 109, 111. But all such procedural measures have a background of custom and practice which was assumed by those who wrote and should be by those who apply them. We reviewed the background of the Act and the consequences on the trial of negligence cases of allowing railroads and others to put in their statements and thus to shield the crew from cross-examination. We said, “Such a major change which opens wide the door to avoidance of cross-examination should not be left to implication.” 318 U. S. at 114. We pointed out that there, as here, the “several hundred years of history behind the Act . . . indicate the nature of the reforms which it was designed to effect.” *519318 U. S. at 115. We refused to apply it beyond that point. We should follow the same course of reasoning here. Certainly nothing in the tradition or practice of discovery up to the time of these Rules would have suggested that they would authorize such a practice as here proposed.
The question remains as to signed statements or those written by witnesses. Such statements are not evidence for the defendant. Palmer v. Hoffman, 318 U. S. 109. Nor should I think they ordinarily could be evidence for the plaintiff. But such a statement might be useful for impeachment of the witness who signed it, if he is called and if he departs from the statement. There might be circumstances, too, where impossibility or difficulty of access to the witness or his refusal to respond to requests for information or other facts would show that the interests of justice require that such statements be made available. Production of such statements are governed by Rule 34 and on “showing good cause therefor” the court may order their inspection, copying or photographing. No such application has here been made; the demand is made on the basis of right, not on showing of cause.
I agree to the affirmance of the judgment of the Circuit Court of Appeals which reversed the district court.
Mr. Justice Frankfurter joins in this opinion.