In Banc

GILLETTE, J.

In these consolidated contempt cases, defendants refused to testify before a grand jury investigating a possible burglary at the General Store in Klamath Falls, claiming their right under Article I, section 12, of the Oregon Constitution not to be compelled to testify against themselves. After extensive proceedings, the trial court granted them use and derivative use immunity under ORS 136.617- 136.619 and ordered them to testify. They persisted in their refusals, claiming that the statutory immunity was insufficient to replace their constitutional right and that only transactional immunity would be sufficient to protect them. The court held that the statutory immunity was adequate and found them to be in contempt. They appeal from the resulting sentences. We reverse.

THE OREGON CONSTITUTION

In recent years, the Oregon Supreme Court has made it clear that the Oregon Constitution has a content independent of that of the federal constitution and that Oregon courts should consider state constitutional claims before examining issues of federal law. See, e.g., In re Lasswell, 296 Or. 121, 673 P.2d 855 (1983); State v. Lowry, 295 Or. 337, 667 P.2d 996 (1983); State v. Kennedy, 295 Or. 260, 666 P.2d 1316 (1983); State v. Davis, 295 Or. 227, 666 P.2d 802 (1983); Hewitt v. SAIF, 294 Or. 33, 653 P.2d 970 (1982); State v. Caraher, 293 Or. 741, 653 P.2d 942 (1982). The state nonetheless argues that, with respect to the issue now before us, we should generally construe the Oregon Constitution to the same effect as the United States Supreme Court construes similar provisions of the federal constitution. It points out that many of the guarantees in both constitutions come from the same sources and urges that the United States Supreme Court's construction of them should control the Oregon construction. The state's arguments miss the point of the Oregon Supreme Court's holdings.

While many guarantees of the state and federal constitutions have their roots in the same sources, they are embodied in different constitutions, with different ultimate interpreters, and may reflect variations in their values and purposes. They generally appeared first in state constitutions and were later added to the federal. No court is the primary interpreter of those guarantees. Under the Fourteenth Amendment, the United States Supreme Court's construction of the federal version of those guarantees is both authoritative for the federal system and a constitutional minimum which states must obey. Its decisions do not, however, decide the meaning of the Oregon Constitution. In that respect, a United States Supreme Court majority is no more binding in Oregon than is a United States Supreme Court minority, a decision of the Supreme Courts of Hawaii, California, or Georgia, or a well-reasoned law review article. Judicial opinions from other jurisdictions are helpful in interpreting the Oregon Constitution to the extent that their reasoning is persuasive and their background is applicable to Oregon. Cf. State v. Kennedy, supra (establishing a double jeopardy rule under the Oregon Constitution different from that proposed by the majority or the minority of the United States Supreme Court). This independent construction of Oregon law holds even when the Oregon law is directly derived from the federal law, which the Oregon Bill of Rights is not. See State v. Pottle, 296 Or. 274, 677 P.2d 1 (1984).

The specific right involved here, not to be a witness against one's self, was a recognized privilege at common law before American independence and first took constitutional form in state constitutions. State courts were also the first to rule on the issues we consider in this case; when the United States Supreme Court first acted it simply adopted one line of state decisions and rejected another. We see no particular reason that we should give greater weight to the views of the United States Supreme Court, which came late to this subject, rather than to the states whose constitutional provisions predated the federal and whose court decisions came earlier.

IMMUNITY LEGISLATION AND THE STATE COURTS

The common law rule, nemo tenetur seipsem accusare, (no person is bound to accuse himself) first received constitutional status in the Declaration of Rights in the Virginia Constitution of 1776, and then in several other state constitutions. In 1791, it was added to the United States Constitution as part of the Fifth Amendment and, by the beginning of the twentieth century, was part of every state constitution except those of New Jersey and Iowa, where it exists as part of the due process of law, State v. Height, 117 Iowa 650, 91 N.W. 935 (1902), or as a rule of the common law. State v. Zdanowicz, 69 NJL 619, 55 A 743 (1903). The constitutional language varies, but courts generally treat the basic principle as the same in all the states. Before Malloy v. Hogan, 378 U.S. 1, 84 S Ct 1489, 12 L Ed 2d 653 (1964), the Fifth Amendment statement of the right applied only to the federal system. Twining v. New Jersey, 211 U.S. 78, 29 S Ct 14, 53 L Ed 97 (1908). As a result, state legislatures and state courts were the first to attempt to balance the constitutional right against the need to gain accomplice or other self-incriminating testimony in enforcing criminal laws. We therefore first examine the state decisions which set the framework for evaluating immunity legislation.

The common law privilege against being required to testify or produce evidence against oneself had several exceptions; early immunity legislation drew on them. All arose in situations where it was clear that any criminal liability which might attach to the witness' answers had expired. Thus, if the witness had been pardoned, or had been tried and acquitted or convicted (and, under some statements of the exception, had served the sentence), or if the limitations period had run, the witness had to testify. These exceptions applied to the constitutional right, which was based on the common law privilege. See State v. Quarles, supra n 4; State v. Jack, 69 Kan. 387, 397, 76 P. 911 (1904), aff'd sub nom Jack v. Kansas, 199 U.S. 372, 26 S Ct 73, 50 L Ed 234 (1905); People v. Mather, 4 Wend 230 (NY 1830); LaFontaine v. Southern Underwriters Assn.,

83 N.C. 132 (1880). Legislatures seeking to overcome the problems of discovering crimes, such as gambling or bribery, in which all those involved might be subject to criminal liability and thus not available to testify, sought to provide additional exceptions to the right by immunizing witnesses who testified. The crucial problem courts faced in evaluating these exceptions was the extent to which it was necessary to protect against derivative use of the immunized testimony.

The first state immunity acts provided only use immunity. The first courts to consider them held that they provided adequate substitutes for the right, but those cases did not involve any derivative use problems. In State v. Quarles, supra n 4, the Arkansas Supreme Court held that the right was merely not to be required to produce evidence to establish one's guilt. It did not directly consider use of the testimony to search out other evidence against the witness (derivative use), but it may have suggested that such a use was too remote to consider. The California Supreme Court in Ex Parte Rowe, supra n 4, took a similar position, as did the Georgia Supreme Court in Higdon v. Heard, 14 Ga. 255 (1853). Neither court considered the derivative use problem.

When courts did face the derivative use problem, they originally held that it was not necessary to protect a witness against such use. In Wilkins v. Malone, supra n 4, the Indiana Supreme Court rejected the plaintiff's claim that he was entitled to transactional immunity before he could be required to testify concerning alleged usury in a note he sued to collect. Although his testimony might provide facts previously unknown to anyone but himself, the court said, the prosecution would still have to prove those facts by other means. The New York Court of Appeals stated this position clearly in People ex rel Hackley v. Kelly, supra n 4, 24 N.Y. at 83:

"If a man cannot give evidence upon the trial of another person without disclosing circumstances which will make his own guilt apparent or at least capable of proof, though his account of the transactions should never be used as evidence, it is the misfortune of his condition and not any want of humanity in the law."

The possibility that a witness' immunized testimony would point the prosecution to the evidence necessary for a conviction and, thus, would be as a practical matter incriminating, led several courts to reject limits on immunity. Massachusetts took the lead in 1871 in Emery's Case, supra n 4. Basing its ruling partly on the provision in the Massachusetts constitution prohibiting requiring a person to furnish evidence as well as to give incriminating testimony, the court held that use of compelled evidence to search out other evidence against a witness was unconstitutional; only transactional immunity could fully protect the witness. Virginia followed the same course in Cullen v. Commonwealth, supra n 4, also emphasizing that only complete amnesty could prevent the use of the immunized testimony to search out other evidence.

The New Hampshire Supreme Court later upheld a statute which required the witness to testify, but provided that no testimony could be used, directly or indirectly, as evidence against the witness and that the witness could not prosecuted for any offense disclosed. State v. Nowell, supra n 4. The court found this combination of use and derivative use immunity with transactional immunity to be an adequate substitute for the right. It emphasized that a valid immunity statute must relieve the witness from all liability on account of the matters the witness is compelled to disclose. "He is to be secured against all liability to future prosecution as effectually as if he were wholly innocent." 58 N.H. at 315.

By the 1880s, then, the problems in connection with immunity statutes were clear, and there were two lines of authority. Those courts that limited the right against self-incrimination to providing testimony that would be used directly against the witness at trial upheld use immunity statutes. Those that extended the right to protect the witness from having to give any information that could be used in obtaining a conviction insisted on transactional immunity. They did not, as they could have done, construe use immunity statutes to protect also against derivative use. The one court which directly considered a use and a derivative use statute simply ignored the latter aspect and upheld the law because it also provided transactional immunity. State v. Nowell, supra n 4. There is no reason to believe that those courts failed to conceive of the possibility of use and derivative use immunity; they simply rejected it. The only way they saw to prevent derivative use was to grant transactional immunity.

IMMUNITY LEGISLATION AND THE FEDERAL COURTS

Federal immunity legislation and court action paralleled that in the states. The first federal immunity statute, 11 Stat 156, ch 19, § 2 (1857), provided transactional immunity for witnesses before congressional committees. Congress soon became concerned that the statute's automatic operation allowed witnesses to acquire immunity too easily and, in 1862, it changed the statute to provide use immunity, relying on state cases holding such to be sufficient. 12 Stat 333, ch 11 (1862). See Grant, Immunity from Compulsory Self-Incrimination in a Federal System of Government, 9 Temp L.Q. 57, 62-65, 194 (1934); Comment, The Federal Witness Immunity Acts in Theory and Practice: Treading the Constitutional Tightrope, 72 Yale L.J. 1568, 1571-75 (1963). In 1868, Congress adopted a general use immunity statute, 15 Stat 37, ch 13, later codified as Rev Stat § 860. Several federal trial courts upheld use immunity under it. See, e.g., United States v. McCarthy, 18 F 87 (S.D.N.Y 1883); United States v. Williams, 28 F Cas 670 (No. 16,717) (S.D. Ohio 1872); United States v. Brown, 24 F Cas 1273 (No. 14,671) (D Or 1871); In re Phillips, 19 F Cas 506 (No. 11,097) (D Va 1869). The courts in Brown and Phillips specifically held that the Fifth Amendment did not require protection from derivative use, citing People ex rel Hackley v. Kelly, supra n 4.

Then, in cases decided in 1892 and 1895, the United States Supreme Court held that the Fifth Amendment required an immunity statute to protect against derivative use and apparently also held that only transactional immunity could provide that protection. In doing so, it adopted one of the two lines of state cases and provided an analysis of the problem which almost all courts accepted for the next 70 years. In Counselman v. Hitchcock, supra n 4, the court held the use immunity of Rev Stat § 860 to be constitutionally insufficient. It first held that the Fifth Amendment right applies to all criminal proceedings: "The object was to insure that a person should not be compelled, when acting as a witness in any investigation, to give testimony that might tend to show that he himself had committed a crime." 142 US at 562. The problem with the statute, the court found, was that it did not prohibit derivative use and, thus, was not an adequate substitute for the right:

"It could not, and would not, prevent the use of his testimony to search out other testimony to be used in evidence against him or his property, in a criminal proceeding in such court. It could not prevent the obtaining and the use of witnesses and evidence which should be attributable directly to the testimony he might give under compulsion, and on which he might be convicted, when otherwise, and if he had refused to answer, he could not possibly have been convicted." 142 US at 564.

In reaching its conclusion, the court followed the Emery-Nowell line of cases without relying, as the Emery and Nowell courts had, on variations in the constitutional language. It held that the Massachusetts and New Hampshire prohibition on being required to furnish incriminating evidence was also part of the Fifth Amendment right, despite the greater similarity between the Fifth Amendment and the New York constitutional language, and that the broader prohibition prevented any derivative use of the compelled testimony. The only way the court found to prevent derivative use was to require transactional immunity:

"We are clearly of opinion that no statute which leaves the party or witness subject to prosecution after he answers the criminating questions put to him, can have the effect of supplanting the privilege conferred by the Constitution of the United States. * * * In view of the constitutional provision, a statutory enactment, to be valid, must afford absolute immunity against future prosecution for the offense to which the question relates." 142 US at 585-86.

In response to Counselman, Congress granted transactional immunity to witnesses in Interstate Commerce Commission proceedings. In Brown v. Walker, supra n 5, the Court, in a five to four decision, upheld the statute. The majority pointed out that a pardoned witness must testify, because the criminality is gone. The immunity law was analogous to an act of general amnesty and thus resembled a pardon. Relying on state cases construing similar laws, the court held this effect to be sufficient:

"[T]he witness cannot properly be said to give evidence against himself, unless such evidence may in some proceeding be used against him, or unless he may be subject to a prosecution for the transaction concerning which he testifies." 161 US at 604.

The court did not consider the possibility of use and derivative use immunity; the statute did not require it to do so. It simply applied the implications of Counselman and held, over strong dissents, that the statute was broad enough to replace the constitutional right.

After Counselman and Brown, most of those states which had previously considered use immunity sufficient changed their positions, and states which had not previously considered the issue accepted what became the general rule that only transactional immunity is an adequate substitute for the right.

There are several reasons for the nearly universal adoption of transactional immunity. First, the previous exceptions to the common law privilege and the constitutional right all required that potential criminal liability be removed before a witness could be compelled to answer. A pardon, a previous acquittal or conviction or the expiration of the statute of limitations all meant that the witness could not possibly incriminate himself. It was natural, particularly for those courts favorably inclined to the right, to insist that the legislature give the same protection to witnesses compelled to testify under a statute as courts required the witness to have before compelling testimony without a statute. The new requirement thus fit comfortably within the old exceptions.

Second, the courts which decided the early cases, which generally accepted use immunity, did not consider the full implications of their actions. The glib dismissal of the derivative use problem in People ex rel Hackley v. Kelly, supra n 4 ("it is the misfortune of his condition and not any want of humanity in the law") forced other courts to face the problem directly, to recognize that allowing prosecutors to require use-immunized potential defendants to provide leads for cases against them did not comport with the right and to insist on statutory protection adequate to preserve the right. As Judge Cardozo later said in In the Matter of Doyle, 257 N.Y. 244, 256-57, 177 N.E. 489 (1931), "[t]he privilege is not removed if there are loopholes in the tender of immunity through which a prosecutor can cut a way to indictment and conviction."

Finally, the cases discussing immunity issues were often the leading cases on the scope of the right itself. It was in those case that courts decided that the right applies to witnesses as well as to defendants, that it applies in grand jury and legislative investigations and that it does not include the common law privilege against disclosing degrading information. Those cases are the foundation for the modern law on the right against self-incrimination. Counselman and Brown were accepted widely as being solidly grounded in 40 years of state court consideration and as an acceptable accommodation between the individual's rights and the government's need for the evidence. Even states that previously had accepted use immunity seized the opportunity to reverse themselves. For 70 years thereafter the basic issue was settled under the federal and almost all constitutions.

THE FEDERAL ABANDONMENT OF TRANSACTIONAL IMMUNITY

Brown v. Walker, supra n 5, alluded to a problem which it did not resolve: the effect of immunity granted by one sovereign on the right of another sovereign to prosecute the witness for a related crime. See Grant, Immunity From Complusory Self-Incrimination, supra. When it considered the issue, the Supreme Court took a separate sovereignties approach, requiring immunity only from prosecution by the government compelling the testimony. The witness had to answer, even if the testimony might prove a case in another jurisdiction. Knapp v. Schweitzer, 357 U.S. 371, 78 S Ct 1302, 2 L Ed 2d 1393 (1958); Feldman v. United States, 322 U.S. 487, 64 S Ct 1082, 88 L Ed 1408 (1944); United States v. Murdock, 284 U.S. 141, 52 S Ct 63, 76 L Ed 210 (1931).

Among the reasons for the court's position were considerations similar to the "silver platter doctrine," which allowed the use in federal trials of evidence that state officials had illegally seized if there was no federal participation in the seizure, and the court's holding that the Fifth Amendment right against self-incrimination did not apply to the states. When these rules disappeared in Elkins v. United States, 364 U.S. 206, 80 S Ct 1437, 4 L Ed 2d 1669 (1960), and Malloy v. Hogan, supra, respectively, the court took another look at its position. In Murphy v. Waterfront Comm. of New York, 378 U.S. 52, 84 S Ct 1594, 12 L Ed 2d 678 (1964), decided the same day as Malloy, the court overruled its previous cases and held that the Fifth Amendment requires that an immunized witness be protected against self-incrimination under the laws of other sovereigns as well as under the laws of the one compelling the testimony.

An obvious question which Murphy had to consider was whether the immunized witness was entitled to transactional immunity in the other jurisdictions. The problems of federalism implicit in allowing one sovereign to grant amnesty from prosecution by another sovereign — especially when the other sovereign was not involved in the decision to grant immunity — weighed heavily against that result, as did the court's belief that the threat of prosecution in the other jurisdiction was generally less than in the immunizing jurisdiction. The court therefore held that the Fifth Amendment was satisfied if the non-immunizing sovereign were prohibited from using the immunized testimony or its fruits in any criminal prosecution against the witness. This was the first adoption and, indeed, the first significant mention, of use and derivative use immunity by the Supreme Court.

In explaining this holding, which came at the end of an opinion focusing on other issues, the court quoted the requirement in Counselman v. Hitchcock, supra n 4, that an immunized witness be protected from use and derivative use of the immunized testimony, quoted Malloy v. Hogan, supra, that the standards to determine whether a witness' silence is justified must be the same in both state and federal proceedings and held that the constitutional rule was

"* * * that a state witness may not be compelled to give testimony which may be incriminating under federal law unless the compelled testimony and its fruits cannot be used in any manner by federal officials in connection with a criminal prosecution against him." 378 US at 79.

The court did not mention transactional immunity, let alone discuss how its holding in Murphy related to the nature of the immunity the Fifth Amendment required an immunizing state sovereign to provide. Justice White concurred, arguing, for the first time at the Supreme Court level, that transactional immunity was greater than the constitution requires from the immunizing sovereign and that use and derivative use immunity is a sufficient substitute for the right. His views were a harbinger.

The Supreme Court's decision in Murphy revived interest in the extent of immunity. The apparent inconsistency in requiring transactional immunity for the immunizing jurisdiction while accepting use and derivative use immunity in other jurisdictions led some to conclude, with Justice White, that use and derivative use immunity was acceptable in all circumstances. The Supreme Court gave conflicting signals in the following years. It finally resolved the issue, so far as the federal constitution is concerned, in 1972 in Kastigar v. United States, supra n 4, in which it held, by a 5-2 vote, that use and derivative use immunity is sufficient protection under the Fifth Amendment, whether the compelling jurisdiction is a state or the federal government.

In Kastigar, the court re-examined Counselman v. Hitchcock, supra n 4, and Brown v. Walker, supra n 5, and noted that, in both cases, the court's concern was that the statute protect a witness from all use of the compelled testimony, including using it to search out other evidence against the witness. It treated the explicit Counselman requirement of transactional immunity as dictum and did not consider the possibility that the Counselman court believed that only transactional immunity could fully protect against derivative use. Kastigar v. United States, supra, 406 US at 453-55 and n 39. It held that Murphy v. Waterfront Comm., supra, had accepted use and derivative use immunity as sufficient in all circumstances, but did not consider how the two-jurisdiction problem might have affected the decision in Murphy. 406 US at 455-59. In order to insure full protection for the witness, the court required the government to prove that all evidence it proposed to use in a subsequent prosecution be derived from a legitimate source wholly independent of the compelled testimony.

Unlike its earlier decisions, the Supreme Court's action in Kastigar stirred up disagreements rather than resolving the issue. The extensive law review commentary was generally critical. See, e.g., Comment, The Scope of Testimonial Immunity Under the Fifth Amendment: Kastigar v. United States, 6 Loy La L Rev 350 (1973); Comment, Self-Incrimination: Choosing a Constitutional Standard: Kastigar v. United States, 32 MD L Rev 289 (1972); Comment, Kastigar v. United States: the Required Scope of Immunity, 58 Va L Rev 1099 (1972); Note, Criminal Procedure — Immunity: Fifth Amendment Privilege Against Self-Incrimination Eclipsed by Use Immunity — Kastigar v. United States, 406 U.S. 441 (1972), 48 Wn. L Rev 711 (1973). State courts evaluating use and derivative use immunity under their own constitutions have split. See Attorney General v. Colleton, supra n 4 (transactional); State v. Miyasaki, 62 Haw. 269, 614 P.2d 915 (1980) (transactional); Ex Parte Shorthouse, 640 S.W.2d 924 (Tex Crim App 1982) (follows Kastigar); In re Caito, supra n 8 (follows Kastigar). The drafters of Uniform Rule of Criminal Procedure § 732 (successor to the Model State Immunity Act) retained transactional immunity despite Kastigar, giving an extensive summary of the arguments in favor of it. With this history and the competing considerations of policy in mind, we turn to a consideration of Oregon's history of immunity legislation.

OREGON IMMUNITY CONSIDERATIONS

Oregon law has included immunity provisions almost from statehood. No Oregon case has decided their constitutionality, although the Supreme Court has given indications of its probable approach. In addition, what early legislatures considered necessary to give witnesses in return for abrogating the Oregon constitutional right not to testify against oneself, Or Const, Art I, § 12, is relevant as an informed, if non-judicial, expression of Oregon constitutional values.

The earliest immunity provisions related to gambling offenses. Section 677 of the Criminal Code of 1864 (Deady 1845-1864) required all persons to testify in criminal and gambling forfeiture proceedings despite any claim of self-incrimination related to those crimes and provided use immunity for the testimony. In 1867, the Oregon Supreme Court held another section of the gambling laws void for vagueness. State v. Mann, 2 Or. 238 (1867). The 1868 legislature then re-enacted the laws with changes to answer the court's concerns. The re-enactment retained the obligation to testify but gave transactional, rather than use, immunity: "[N]o indictment or prosecution shall afterward be brought against him for the particular offense concerning which he testified as a witness." Act of October 28, 1868, § 7 (Criminal Code § 714) (Deady and Lane, 1845-1872). The legislature re-enacted that provision in 1876, and it continued in effect in that form until the revision of the Criminal Code in 1971. Hill's Ann Laws § 3534; former ORS 167.520 ( repealed by Or Laws 1971, ch 743, § 432). This change from use to transactional immunity occurred before any appellate court had invalidated use immunity, after several had approved it, and after Congress had moved from transactional to use immunity in its enactments. The continuation of transactional immunity in the 1876 re-enactment occurred after Judge Deady's decision in United States v. Brown, supra, holding that use immunity was sufficient under the United States Constitution. We infer from this history that the legislature made a conscious decision that use immunity, even though it was constitutional elsewhere, is inconsistent with the constitutional rights of Oregonians and that only transactional immunity is a full substitute for that right.

Most other Oregon immunity statutes have also provided transactional immunity, either originally or as amended. Or Laws 1915, ch 141, § 35, covering testimony in prohibition act violations, originally provided only use immunity. The next session of the legislature amended it to add transactional immunity to the use immunity; the act remained in that form until the repeal of prohibition in 1933. Or Laws 1917, ch 40, § 4. Most currently effective statutes unambiguously provide transactional immunity. ORS 136.619 is the only Oregon law providing use and derivative use immunity. As originally adopted in 1971, it provided transactional immunity; at that time the federal status of use and derivative use immunity was unclear. See Johnson, Oregon's Witness Immunity Law, 51 Or L Rev 573 (1972). In the only Oregon appellate cases involving a witness immunity statute, the Supreme Court applied the gambling transactional immunity law according to its terms, holding that the witness did not need to claim the right in order to be protected; immunity flowed automatically from the act of testifying. State v. Logsdon, 196 Or. 542, 250 P.2d 377 (1952); State v. Hennessey, 195 Or. 355, 245 P.2d 875 (1952). The court had no occasion to consider what scope of immunity was constitutionally necessary.

While there are no cases directly concerning the issue we now face, the Oregon Supreme Court has expressed itself on issues relevant to this problem in several cases concerning psychiatric interviews of defendants who raise mental defenses. Those cases strongly indicate that, were it faced with the issue, the court would require transactional immunity. Beginning with Shepard v. Bowe, 250 Or. 288, 442 P.2d 238 (1968), it has consistently held that a psychiatrist conducting an examination for the prosecution may not question a defendant concerning the events surrounding the commission of the alleged crime, even if the defendant's state of mind at that time is crucial to the mental defense. It rejected arguments that it should only prohibit the use of the statements at trial, because that rule would still allow the use of the statements to develop leads. It also rejected an argument that it should simply prohibit the psychiatrist from sharing the statements with the prosecution, thus necessarily preventing derivative use. The latter possibility, it stated, would be impossible to enforce.

"[W]e are of the opinion that under certain circumstances there is more than a remote chance that such statements would become known to others in addition to the trial court." 250 Or at 293-94.

The court spelled out its concerns and reaffirmed its position, despite doubts from some members of the court, in State ex rel Johnson v. Woodrich, 279 Or. 31, 566 P.2d 859 (1977):

"[W]e are aware that certain expert witnesses are regularly hired by the state to perform pretrial psychiatric examinations. This procedure creates relationships which, if the suggested solution [prohibiting the psychiatrist from divulging incriminating statements] was [sic] adopted, would make us apprehensive and uneasy." 279 Or at 36-37.

The court's skepticism about relying on the prosecution to protect a defendant's rights was expressed most clearly in its rejection of a proposed prohibition of only the use and derivative use of incriminating statements made during a psychiatric interview:

"The state contends that pretrial discovery will permit a criminal defendant to spot and to stop any attempt by the prosecution to make such use of the statements. We are not convinced that this would be so. The temptation on the part of prosecutors to develop their cases would be almost irresistable. It is unrealistic to give a dog a bone and to expect him not to chew on it. We are not inclined to compel criminal defendants to cooperate fully with state psychiatrists in reliance on the ability of defendants to detect every instance of improper use of the information so gained." 279 Or at 36. (Emphasis supplied.)

The Oregon Supreme Court has thus rejected the foundation of the United States Supreme Court's approval of use and derivative use immunity: its belief that such immunity will in fact adequately protect the witness' right and that it will, as a practical matter, be possible to determine whether proposed prosecution evidence is derived from impermissible use of the compelled testimony. The court has taken its position in a context where the defendant, not the prosecution, has raised the issues to which the compelled statements are directed and in which the defendant will not testify under oath and will thus not be subject to a perjury prosecution. If the Supreme Court will not accept compelled statements in those circumstances, it is unlikely that it will accept them when the state is engaged in a fishing expedition, the witness is under oath and thus is threatened with the possibility of perjury for false testimony and the witness has done nothing to open up the issues which the prosecution wishes to explore. The court thus appears to have aligned itself with the reasoning of the Emery/Counselman/Brown cases that use and derivative use protections are not an adequate substitute for the right against self-incrimination.

THE OREGON CONSTITUTION REQUIRES TRANSACTIONAL IMMUNITY

We hold that Article I, section 12, of the Oregon Constitution requires transactional immunity as a substitute for the right not to testify against oneself. ORS 136.619 does not pass this test.

Our analysis starts with the language of the United States Supreme Court:

"It is quite clear that legislation cannot abridge a constitutional privilege, and that it cannot replace or supply one, at least unless it is so broad as to have the same extent in scope and effect." Counselman v. Hitchcock, supra n 4, 142 U.S. 585. (Emphasis supplied.)

The court later modified this statement in Murphy v. Waterfront Commission, supra, to require only that the witness be left "in substantially the same position as if the witness had claimed his privilege in the absence of a state grant of immunity." 378 US at 79. (Emphasis supplied.) However correct this modification may be when the issue is whether the federal government may use evidence procured under a state grant of immunity, and however proper it may be in that context to refer to immunity as an exclusionary rule, we cannot agree with the Supreme Court in Kastigar v. United States, supra n 4, 406 US at 457, that the Murphy statement provides the rule by which we should evaluate the immunity which the sovereign compelling the testimony must provide. The citizens of Oregon are entitled, under their constitution, not merely to a "substantial" substitute for their constitutional rights, but to one which has "the same extent in scope and effect." Mere use and derivative use immunity falls short of that constitutionally-required minimum.

First of all, use and derivative use immunity does not clearly protect against non-evidentiary as well as evidentiary use of the immunized testimony. A prosecutor who is aware of a witness' immunized statements is likely to allow that knowledge to affect discretionary decisions: whether to pursue the investigation, what direction to take in doing so, whether to prosecute the witness, whether to plea bargain and, if so, what to seek in the bargaining. It is hard to see how the most conscientious prosecutor could avoid letting the knowledge that the witness had admitted the crime while immunized affect these decisions. The immunized testimony might also be helpful in explaining other information in the prosecutor's hands, pointing out the significance of items previously ignored, aiding in the preparation of cross-examination and, more generally, in deciding trial strategy. At the least, a prosecutor could use immunized testimony as a deposition, learning the witness' complete story before trial. A prosecutor may use immunized testimony in any of these ways without introducing it or its fruits into evidence, but it is obvious that none of these actions would be possible if the witness had simply asserted the right to remain silent and had not testified. The witness is thus more subject to prosecution and conviction because of the testimony. Immunity that permits non-evidentiary use is thus not "so broad as to have the same extent in scope and effect" as the right it supposedly supplants.

Non-evidentiary use of the compelled testimony is sometimes justified by an analogy to the rules excluding illegally obtained evidence and prohibiting the use of coerced confessions or their fruit. Those analogies are inappropriate. Illegal police actions occur during the course of investigations, often when there is little time for reflection, and are contrary to law. Prohibition of the use of the fruits of illegal police activity in court is necessary to vindicate the violated rights. See State v. Davis, supra, 295 Or at 231-37. It is neither necessary nor practical to hold that illegal police activities automatically immunize a defendant from all prosecution; it is enough that courts refuse to participate in the illegality. In fact, holding that immunity comes from such disapproved actions by police officers in the field would require the state to give up the right to prosecute while receiving nothing in return.

In contrast, a grant of testimonial immunity is a conscious prosecutorial choice, taken after careful consideration and with full knowledge of its results. It is not a violation of the witness' constitutional right but a determination to substitute for that right another one granted by the legislature and found by the courts to be an adequate replacement. The constitutional right given up is not to be compelled to testify against oneself. A witness granted immunity and then required to testify has not received full value for that lost right if there is any way the testimony can cause harm to the witness in that prosecution.

Because use and derivative use immunity, to be constitutional, must prohibit non-evidentiary use of the testimony, the methods of assuring against that non-evidentiary use are crucial. One federal court has held that the prosecutor's reading the immunized testimony was in itself a non-evidentiary use sufficient to require dismissal of the case. United States v. McDaniel, 482 F.2d 305 (8th Cir 1973). Other cases have taken a less strict approach, see Strachan, Self-Incrimination, supra n 20, 56 Tex L Rev at 813, but the problems involved remain serious. Kastigar v. United States, supra, 406 U.S. 460-461, requires the prosecution to bear the burden of showing that its proposed evidence is without taint; the prosecution would also have to bear the burden of showing that its case is unaffected by prohibited non-evidentiary use. Yet carrying the latter burden would require the prosecution to open up its investigatory process to thorough examination and would permit the defense to question the prosecutor concerning proposed evidence and trial tactics and their sources in order to ferret out subtle non-evidentiary uses of which even the prosecutor might not be consciously aware.

The prosecution might attempt to avoid these problems by erecting a "Chinese wall" around the immunized evidence, so that the prosecutors and investigators will remain unaware of it. This action would be easiest in the two-jurisdiction situation in which use and derivative use immunity first arose. It might also be effective in a single jurisdiction where the immunized testimony is sought for use against a different defendant for a crime only indirectly related to that for which the witness is prosecuted. Even in that situation, however, such a wall would be difficult to construct, especially considering the size of many investigating and prosecuting offices and the amount of information sharing that normally occurs within them. Too many people would have to participate, and there would be too many opportunities for breaks in the wall.

In a large conspiracy investigation or a major multiple-party crime, situations for which immunity is particularly designed, effective use of such a wall would, we believe, be impossible. There is simply no way that a prosecutor could try a case against one member of a conspiracy, for instance, without having a solid grasp of all of the evidence related to that conspiracy, and it is hard to imagine how that evidence, and the discretionary decisions involved in bringing the case to and managing it during trial, could be unaffected by the immunized testimony which his or her colleagues secured. The Supreme Court in State ex rel Johnson v. Woodrich, supra, expressed its skepticism, which we share, about the ability of a prosecutorial team to keep items of evidence from crossing such a wall. We conclude that there is a wide range of situations in which use and derivative use immunity simply cannot adequately protect the witness.

"It is unrealistic to give a dog a bone and to expect him not to chew on it." State ex rel Johnson v. Woodrich, supra, 279 Or at 36. We hold that Article I, section 12, of the Oregon Constitution forbids giving the dog the bone. Only transactional immunity is constitutional in Oregon. Because ORS 136.619 did not afford such immunity to these defendants, it is unconstitutional and their refusals to testify according to its terms were justified. Their convictions for contempt are reversed.

Reversed.