Kastigar v. United States/Dissent Marshall

[p467] MR. JUSTICE MARSHALL, dissenting.

Today the Court holds that the United States may compel a witness to give incriminating testimony, and subsequently prosecute him for crimes to which that testimony relates. I cannot believe the Fifth Amendment permits that result. See Piccirillo v. New York, 400 U.S. 548, 552 (1971) (BRENNAN, J., dissenting from dismissal of certiorari).

The Fifth Amendment gives a witness an absolute right to resist interrogation, if the testimony sought would tend to incriminate him. A grant of immunity [p468] may strip the witness of the right to refuse to testify, but only if it is broad enough to eliminate all possibility that the testimony will in fact operate to incriminate him. It must put him in precisely the same position, vis-à-vis the government that has compelled his testimony, as he would have been in had he remained silent in reliance on the privilege. Ullmann v. United States, 350 U.S. 422 (1956); McCarthy v. Arndstein, 266 U.S. 34 (1924); Hale v. Henkel, 201 U.S. 43 (1906); Brown v. Walker, 161 U.S. 591 (1896); Counselman v. Hitchcock, 142 U.S. 547 (1892).

The Court recognizes that an immunity statute must be tested by that standard, that the relevant inquiry is whether it "leaves the witness and the prosecutorial authorities in substantially the same position as if the witness had claimed the Fifth Amendment privilege." Ante, at 462. I assume, moreover, that in theory that test would be met by a complete ban on the use of the compelled testimony, including all derivative use, however remote and indirect. But I cannot agree that a ban on use will in practice be total, if it remains open for the government to convict the witness on the basis of evidence derived from a legitimate independent source. The Court asserts that the witness is adequately protected by a rule imposing on the government a heavy burden of proof if it would establish the independent character of evidence to be used against the witness. But in light of the inevitable uncertainties of the fact-finding process, see Speiser v. Randall, 357 U.S. 513, 525 (1958), a greater margin of protection is required in order to provide a reliable guarantee that the witness [p469] is in exactly the same position as if he had not testified. That margin can be provided only by immunity from prosecution for the offenses to which the testimony relates, i.e., transactional immunity.

I do not see how it can suffice merely to put the burden of proof on the government. First, contrary to the Court's assertion, the Court's rule does leave the witness "dependent for the preservation of his rights upon the integrity and good faith of the prosecuting authorities." Ante, at 460. For the information relevant to the question of taint is uniquely within the knowledge of the prosecuting authorities. They alone are in a position to trace the chains of information and investigation that lead to the evidence to be used in a criminal prosecution. A witness who suspects that his compelled testimony was used to develop a lead will be hard pressed indeed to ferret out the evidence necessary to prove it. And of course it is no answer to say he need not prove it, for though the Court puts the burden of proof on the government, the government will have no difficulty in meeting its burden by mere assertion if the witness produces no contrary evidence. The good faith of the prosecuting authorities is thus the sole safeguard of the witness' rights. Second, even their good faith is not a sufficient safeguard. For the paths of information through the investigative bureaucracy may well be long and winding, and even a prosecutor acting in the best of faith cannot be certain that somewhere in the depths of his investigative apparatus, often including hundreds of employees, there was not some prohibited use of the compelled testimony. Cf. Giglio v. United States, 405 U.S. 150 (1972); Santobello v. New York, 404 U.S. 257 (1971). The Court today sets out a loose net to trap tainted evidence and prevent its use against the witness, but it accepts an intolerably great risk that tainted evidence will in fact slip through that net.

[p470] In my view the Court turns reason on its head when it compares a statutory grant of immunity to the "immunity" that is inadvertently conferred by an unconstitutional interrogation. The exclusionary rule of evidence that applies in that situation has nothing whatever to do with this case. Evidence obtained through a coercive interrogation, like evidence obtained through an illegal search, is excluded at trial because the Constitution prohibits such methods of gathering evidence. The exclusionary rules provide a partial and inadequate remedy to some victims of illegal police conduct, and a similarly partial and inadequate remedy to some victims of illegal police conduct, and a similarly partial and inadequate deterrent to police officers. An immunity statute, on the other hand, is much more ambitious than any exclusionary rule. It does not merely attempt to provide a remedy for past police misconduct, which never should have occurred. An immunity statute operates in advance of the event, and it authorizes–even encourages–interrogation that would otherwise be prohibited by the Fifth Amendment. An immunity statute thus differs from an exclusionary rule of evidence in at least two critical respects.

First, because an immunity statute gives constitutional approval to the resulting interrogation, the government is under an obligation here to remove the danger of incrimination completely and absolutely, whereas in the case of the exclusionary rules it may be sufficient to shield the witness from the fruits of the illegal search or interrogation in a partial and reasonably adequate manner. For when illegal police conduct has occurred, the exclusion of evidence does not purport to purge the conduct of its unconstitutional character. The constitutional violation remains, and may provide the basis for other relief, such as a civil action for damages (see 42 U.S.C. § 1983 and Bivens v. Six Agents, 403 U.S. 388 (1971)), or a criminal prosecution of the responsible [p471] officers (see 18 U.S.C. §§ 241-242). The Constitution does not authorize police officers to coerce confessions or to invade privacy without cause, so long as no use is made of the evidence they obtain. But this Court has held that the Constitution does authorize the government to compel a witness to give potentially incriminating testimony, so long as no incriminating use is made of the resulting evidence. Before the government puts its seal of approval on such an interrogation, it must provide an absolutely reliable guarantee that it will not use the testimony in any way at all in aid of prosecution of the witness. The only way to provide that guarantee is to give the witness immunity from prosecution for crimes to which his testimony relates.

Second, because an immunity statute operates in advance of the interrogation, there is room to require a broad grant of transactional immunity without imperiling large numbers of otherwise valid convictions. An exclusionary rule comes into play after the interrogation or search has occurred; and the decision to question or to search is often made in haste, under pressure, by an officer who is not a lawyer. If an unconstitutional interrogation or search were held to create transactional immunity, that might well be regarded as an excessively high price to pay for the "constable's blunder." An immunity statute, on the other hand, creates a framework in which the prosecuting attorney can make a calm and reasoned decision whether to compel testimony and suffer the resulting ban on prosecution, or to forgo the testimony.

For both these reasons it is clear to me that an immunity statute must be tested by a standard far more demanding than that appropriate for an exclusionary rule fashioned to deal with past constitutional violations. Measured by that standard, the statute approved today by the Court fails miserably. I respectfully dissent.