with whom Circuit Judges BROWNING, ELY, HUF-STEDLER and ALFRED T. GOODWIN join, dissenting:
The question presented is whether Garner, in this criminal prosecution unrelated to the federal income tax laws, may in*241voke the protection of the Fifth Amendment privilege not to “be compelled in any criminal case to be a witness against himself . . . ” to exclude incriminating answers to questions on his tax return. The answer requires the resolution of two issues. The first, whether the disclosure requirement in the tax law, backed by statutory penalties for noncompliance, is governmental compulsion of the type proscribed by the Fifth Amendment; the second, whether by answering, Garner voluntarily relinquished his privilege. Both require careful factual analysis and scrutiny of the purposes served by Supreme Court decisions. The majority opinion avoids the analysis by purporting to reach neither issue, although, in my opinion, the conclusory rationale offered conceals an unexplained decision on one of the two, and perhaps both, grounds.
The majority agrees that the privilege applies in the context of regulatory inquiries (n. 14), but, expressly rejecting reliance on the argument that Garner might have waived his privilege by answering (p. 233), holds that the nature of the privilege precludes consideration of Garner’s constitutional claim when asserted at his criminal trial rather than on the return. The majority states:
Essentially, the question is whether the Fifth Amendment is a right which can be invoked retrospectively or whether it is a privilege which must be claimed when the incriminating information is requested. Here, we hold it to be a privilege which Garner should have asserted in his tax return, (p. 233).
The holding proceeds from the premise that the Fifth Amendment provision creates a testimonial privilege which in its historical development has come to be applied to one in Garner’s position (p. 230). The majority reasons that a privilege operates to justify a refusal to answer questions when they are first asked —not as an exclusionary mechanism to exclude answers once they have been given — and therefore Garner’s objection at trial is not properly the invocation of a privilege. The majority concludes that since the Fifth Amendment creates a privilege and not a “retrospective” right, any protection the Amendment provides is unavailable to exclude the answers compelled from Garner when introduced at trial. To the extent that the majority relies solely on this basis, its decision is completely untenable, both in theory and under the Supreme Court decisions which bind the determination in this case.
The majority draws important consequences from the technical distinction between a right and a privilege, rather than analyzing the Fifth Amendment dangers in light of the particular circumstances under which Garner answered. However, even accepting the assumption that the protection of the Fifth Amendment is limited to that afforded by a technical privilege, the majority is still wrong in the assertion that the nature of a testimonial privilege precludes Garner’s objection to the introduction of the tax return. Many testimonial privileges operate not only to excuse the person holding the privilege from answering, but also to allow him to exclude the testimony of others. For instance, the client may invoke the lawyer-client privilege to prevent the lawyer from testifying about the client’s communications, even those involving admissions of a crime. The privilege is available to exclude reliable testimony in order to advance the public policies surrounding the relationship in which the communication was made. By analogy, the policy against governmental compulsion is served by the exclusion of answers obtained in violation of the policy, and Garner’s ability to command their exclusion is the exercise of a privilege in the same sense that the client’s ability to exclude the lawyer’s recital of his prior admission is a privilege. If the majority rests its decision on the ground that the concept of a “privilege” is too narrow to encompass Garner’s objection, an analysis of common law privileges alone suggests that the majority is in error.
*242More fundamentally, the fact that the constitutional provision has been historically regarded as a “privilege” does not mean, as the majority assumes, that the clause does not create a constitutional right cognizable at Garner’s trial. In a strict Hohfeldian sense1 the freedom from testimonial compulsion created by the Amendment is a right against the court. If Garner has a privilege on his tax return, he may answer or not; and the government has no right to compel the answer. If he answers, the government is privileged to introduce his answer if he has no right which the court will recognize to prevent it. The court, however, exercises the judicial power to decide Garner’s case on the basis of the Constitution’s delegation of sovereignty, one of the express conditions of which is the Fifth Amendment. The court is under a constitutional duty not to allow Garner to be compelled to be a witness against himself because the Constitution commands forebearance for the benefit of those in his position; against the court’s duty he has a correlative right, not privilege. The scope of his right depends on the extent of the court’s duty to exclude the evidence, ordinarily entailing an analysis of the policy of the constitutional provision in light of the particular facts involved. The point is not that Garner is entitled to reversal because he has a constitutional right, but rather that the majority opinion neatly, but inaccurately, sidesteps the constitutional analysis required to decide the case.
The majority’s position has been repeatedly rejected, both explicitly and implicitly, by the Supreme Court. For instance, in Murphy v. Waterfront Commission, 378 U.S. 52, 84 S.Ct. 1594, 12 L.Ed.2d 678 (1964), Justice Goldberg states:
The constitutional privilege against self-incrimination has two primary interrelated facets: The Government may not use compulsion to elicit self-incriminating statements,' see, e. g., Counselman v. Hitchcock, 142 U.S. 547, 12 S.Ct. 195, 35 L.Ed.2d 1110 and the Government may not permit the use in a criminal trial of self-incriminating statements elicited by compulsion. See, e. g., Haynes v. Washington, 373 U.S. 503, 83 S.Ct. 1336, 10 L.Ed.2d 513. . . . Now that both governments are fully bound by the privilege, the conceptual difficulty of pinpointing the alleged violation of the privilege on “compulsion” or “use” need no longer concern us. (378 U.S. at 57 n. 6, 84 S.Ct. at 1598.)
In Lefkowitz v. Turley, 414 U.S. 70, 94 S.Ct. 316, 38 L.Ed.2d 274 (1973), Justice White, writing for a unanimous Court, states:
[A] witness protected by the privilege may rightfully refuse to answer unless and until he is protected at least against the use of his compelled answers and evidence derived therefrom in any subsequent criminal case in which he is' a defendant. Kastigar v. United States, 406 U.S. 441 [92 S.Ct. 1653, 32 L.Ed.2d 212] (1972). Absent such protection, if he is nevertheless compelled to answer, his answers are inadmissible against him in a later criminal prosecution. Bram v. United States, supra; Boyd v. United States, supra.
The coerced confession cases, from Bram v. United States, 168 U.S. 532, 18 S.Ct. 183, 42 L.Ed. 568 (1897), through Miranda and its progeny, consistently indicate that the failure of the defendant to remain silent does not end the analysis, and that the protection against self-incrimination afforded by the constitutional privilege extends to the exclusion of prior answers when they are involuntarily given. Cf. Jackson v. Denno, 378 U.S. 368, 84 S.Ct. 1774, 12 L.Ed.2d 908 (1964). The majority’s reliance on the assertion that Garner’s privilege cannot be “retrospectively” invoked because he failed to “invoke and assert the benefits of his privilege” (p. 232-233) when first asked the incriminating question is misplaced *243—the coerced confession cases do not require a ritualistic explicit invocation of the privilege, or a request to remain silent or see a lawyer, before a subsequent self-incriminating answer may be excluded at trial. Rather, the cases make clear that the privilege is applicable at trial if the prior answer was coerced. Instead of addressing the relevant inquiry of whether Garner’s answers were in fact coerced, the majority ostensibly avoids the issue by restricting the Fifth Amendment to the first of Justice Goldberg’s facets. Under the majority’s reasoning, incriminating answers extracted by governmental coercion do not fall within the Fifth Amendment if the defendant fails to invoke and assert his privilege when first asked. Clearly that is not the law.
The privilege does more than provide a defense to one who sticks to his guns and is prosecuted for his silence rather than succumbing to the governmental compulsion. In a situation analogous to the present case, the Court, in decisions relating to the wagering tax law2 found unconstitutional in Marchetti v. United States, 390 U.S. 39, 88 S.Ct. 697, 19 L.Ed.2d 889 (1968), has indicated the privilege is available to exclude incriminating answers on tax returns at trial even though the defendant had failed to invoke his privilege on the return. For instance, in Lookretis v. United States, where the defendant had given incriminating answers on the wagering tax return found violative of the privilege in Marchetti and his answers had been used against him in a subsequent prosecution (see 398 F.2d 64 (7th Cir. 1968)), the Court, apparently considering the use of the returns an equivalent violation of his privilege, vacated his conviction and remanded. Lookretis v. United States, 390 U.S. 338, 88 S.Ct. 1097, 19 L.Ed.2d 1219 (1968). In United States v. Knox, 396 U.S. 77, 90 S.Ct. 363, 24 L.Ed.2d 275 (1969), the Court, in upholding a conviction for filing a false return which under Marchetti did not have to be filed, distinguished Lookretis:
The Court of Appeals ruled that truthful disclosures made under the compulsion of § 4412 could not be introduced against their maker in a criminal proceeding. However, the Fifth Amendment was offended in Lookretis precisely because the defendant had succumbed to the statutory compulsion by furnishing the requested incriminatory information. 396 U.S. at 82, 90 S.Ct. at 366.
Adhering to that view, Mackey v. United States, 401 U.S. 667, 91 S.Ct. 1160, 28 L.Ed.2d 404 (1971), again indicated that the Court views the use of compelled answers as the equivalent of the compulsion of answers for purposes of Fifth Amendment analysis. The Court in Mackey held that its decision in Mar-chetti did not have retroactive application, but Mackey, unlike Marchetti, had filed an incriminating return without asserting a privilege. The Seventh Circuit thought Mackey’s objection might have been waived, but the Supreme Court plurality opinion, in a comment on which eight of the nine Justices agreed,3 stated:
The statutory requirement to register and file gambling tax returns was held to compel self-incrimination and the privilege became a complete defense to a criminal prosecution for failure to register and pay the related taxes. It followed that the registration and excise tax returns filed in response to the statutory command were compelled statements within the meaning of the Fifth Amendment and accordingly were inadmissible evidence as part of *244the prosecution’s case in chief. 401 U.S. at 672, 91 S.Ct. at 1163.
In contrast to the circuit court, the Supreme Court indicated that the privilege is sufficient to exclude answers at a subsequent trial if the privilege was a “defense to a criminal prosecution for failure to register,” (401 U.S. at 672, 91 S.Ct. at 1163), without analysis of whether answering waived the privilege.
The majority’s approach is irreconcilable with Mackey, particularly if one reads United States v. Sullivan, 274 U.S. 259, 47 S.Ct. 607, 71 L.Ed. 1037 (1927), as the majority purports to, for the proposition that Sullivan recognized the privilege as a defense to a prosecution for filing an incomplete return if claimed with respect to particular questions on the return. If Garner was, as the majority asserts, privileged not to answer particular questions on the return, it is only because in respect to those questions the self-reporting scheme of the tax return, backed by statutory compulsion to answer, unconstitutionally compelled him to incriminate himself. It follows that the privilege became a complete defense for failure to answer those particular questions, and under Mackey, — “It followed . . .” that answers to questions he was privileged not to answer “filed in response to the statutory command were compelled statements within the meaning of the Fifth Amendment and accordingly were inadmissible evidence as part of the prosecution’s case in chief.” 401 U.S. at 672, 91 S.Ct. at 1163. As the majority correctly points out (p. 231), the differences in the statutory schemes make Marchetti and Mackey in-apposite in determining whether there is unconstitutional compulsion in Garner’s case. However, once the majority acknowledges the existence of a privilege on the return, depending only on whether the information requested is incriminating, the distinction becomes irrelevant, unless the majority is making the unexplained, and probably inexplicable, argument that different consequences should flow from supposed qualitative differences in acknowledged constitutional violations.
The root of the majority’s dilemma is that it has mistaken a privilege for the opportunity to assert one. Garner is privileged to ignore the statutorily imposed duty to answer questions only if the court will not impose the sanctions provided by the tax laws for incomplete returns. The court will refrain from doing so only to the extent dictated by statute, or prior decision interpreting the constitutional command — in those cases the court must recognize the privilege as completely excusing the breach of the normal duty to answer. The privilege is not the possible availability of a constitutional defense if prosecuted, but rather the actual recognition that he is free not to answer. In fact, what Garner had, when filling out his tax return, was at most the knowledge that the constitutional provision might lead a court to validate the privilege as a defense, and not a privilege.
No court has ever recognized the Fifth Amendment as the source of a valid defense for failure to answer a question on a tax return. Sullivan does not. It holds only that the privilege does not excuse a failure to file a tax return. Since many of the questions were innocuous, Sullivan’s remedy of refusing to answer all of them was overbroad. It was considered an extreme and extravagant application of the privilege to allow it to defeat the social purpose in obtaining a tax return from Sullivan. The oft-repeated dicta, on which the majority relies for the proposition that Garner must assert the privilege on the return, told Sullivan that the privilege might have provided a defense for refusal to answer particular questions and that the Court might have decided the merit of his claim of privilege in respect to them. The Court has never, as it recently recognized,4 held that the privilege excuses a failure to answer particular questions on the return.
*245Either way one reads Sullivan, the majority’s holding is wrong. It is either inconsistent with Mackey, or the majority is telling Garner that the court will not decide his claim of privilege in a criminal case, the only point in time at which the constitutional provision applies on its face, because he did not .run the risk of an uncertain invocation of the privilege on his tax return.
The majority criticizes the position Garner, and the original panel opinion, take for making the Fifth Amendment into a “right which can be invoked retrospectively.” (p. 233) The comment reveals a verbal sleight-of-hand which inverts the entire history and development of the privilege. The majority’s failure to understand the significance of either the method of or reasons for the extension of the privilege to “witnesses,” which it recites (p. 230), is particularly unfortunate because an understanding of the rationale of the historical development of the privilege is necessary in deciding both issues presented by this case —the proper manner of application of the privilege in the context of regulatory inquiries, and the “voluntariness” of Garner’s answers on the tax return. Analysis of the historical development indicates that completely different issues, requiring reconciliation of different values, are involved in adjudicating a claim of privilege, depending on the nature of the governmental compulsion and the timing and form in which the claim is raised. If what is “at issue is the nature of the privilege against self-incrimination .” (p. 230), then the majority opinion is neither in accord with the policies, or development and nature, of the privilege.
The historical development and extension of the privilege reflects what the privilege is — precisely, the recognition by a court that the law, to effectuate ulterior values and policies, in the circumstances imposes no duty to answer. The privilege is a defense to a prosecution for refusing to answer because the refusal does not breach the witness’s ordinary duty to answer. The development of the privilege, or its extension to novel situations, begins initially as a recognition, in the context of a criminal prosecution, that the defendant should not be punished for failure to answer because the law does not impose the duty. Thereafter, once recognized as a defense to a prosecution, the privilege will be recognized by the court to excuse a witness from answering so long as the questions are ones which he is under no duty to answer.
As many commentators have pointed out, the courts which have radically expanded the scope of the constitutional privilege have never agreed on the values to be promoted by eliminating the duty to answer in those cases.5 Among the values most frequently advanced are the preservation of an accusatorial rather than inquisatorial method of criminal prosecution, the prevention of government immorality, akin to the first, and the preservation of privacy and individual dignity. In part, the confusion over purposes reflects their interrelationship. Concern for the accusatorial system is a concern for the preservation of individual privacy as well, reflecting the Lockean notion that government is essentially a restraint on liberty and ought to leave the individual alone. Wigmore, no lover of the privilege, thought it necessary to prevent degeneration into unjust modes of criminal prosecution:
The real objection is that any system of administration which permits the prosecution to trust habitually to compulsory self-disclosure as a source of proof must itself suffer morally thereby. The inclination develops to rely mainly upon such evidence, and to be satisfied with an incomplete investigation of other sources. The exercise *246of the power to extract answers begets a forgetfulness of the just limitations of that power. Wigmore, Evidence, § 2251 (McNaughten rev. 1961).
The concern for an “accusatorial” system reflects the necessity that the integrity of the judicial process remain high in the eyes of those subject to it, in order to preserve its effective operation. The privilege is essential to the preservation of an adversary, accusatorial system of criminal procedure because it is the cornerstone of the rest of the procedural safeguards afforded the criminal defendant — all are hollow if the defendant can be compelled to condemn himself in anticipation of trial.
The development of the privilege has historically been through recognition in novel situations that the law should impose no duty on the defendant to answer. The origin of the constitutional provision lies in the dilemma of men like More or Lilburne, who were imprisoned for refusal to subscribe to the oaths required of them. Knowing that if they took the oath and answered honestly they would have to admit holding views considered criminal under applicable statutes, they refused and were imprisoned. The constitutional privilege “against being compelled in any criminal case to be a witness against” one’s self thus originally evolved from the recognition of a defense to a prosecution or punishment for refusing the oath, or refusing to answer under oath.6 By the time of the adoption of the Fifth Amendment, the privilege had some application to pre-trial stages of criminal proceedings, and there was a common law rule excluding coerced confessions.7
The Fifth Amendment does not term itself a “privilege,” but at the time of its adoption it was so conceived. On its face, the only thing clear about the applicability of the proscription of testimonial compulsion is that the answers sought must ultimately be capable of use in a criminal case; until the late 19th century, the privilege applied only to the criminal proceeding for which the incriminating answers were sought. Justice Stewart’s comment that the “Fifth Amendment’s privilege against compulsory self-incrimination was originally meant to do no more than confer a testimonial privilege upon a witness in a judicial proceeding. . . . ”8 is accurate, with the qualification that originally and until the late 19th century, the “judicial proceeding” was the criminal trial in which the witness was defendant.
The original concept of a privilege thus limited to the defendant’s criminal trial has been abandoned. The Court has interpreted the constitutional provision to create a valid defense to a prosecution for refusal to answer questions in a variety of judicial and quasi-judicial proceedings, and, under some circumstances, for a refusal to file answers to regulatory inquiries. The landmark expansion occurred in Counselman v. Hitchcock, 142 U.S. 547, 12 S.Ct. 195, 35 L.Ed. 1110 (1892), where the Court held that the constitutional provision afforded a defense to a contempt conviction predicated on a witness’s refusal to answer questions before a grand jury. In McCarthy v. Arndstein, 266 U.S. 34, 45 S.Ct. 16, 69 L.Ed. 158 (1924), the Court held that the privilege provided a valid defense for refusal to testify in a civil, bankruptcy proceeding. Taken together, the two cases allow a witness to refuse to answer in any quasi-judicial proceeding, civil or criminal, if his answers would be incriminating in a subsequent criminal trial.
In each instance in which the Court has expanded the scope and applicability of the privilege, it has done so to prevent circumvention of its purposes by coercion *247prior to criminal proceedings, thereby making the protection afforded in the criminal trial context effective. As expressed in Counselman:
It is impossible that the meaning of the constitutional provision can only be that a person shall not be compelled to be a witness against himself in- a criminal prosecution against himself. It would doubtless cover such eases; but it is not limited to them. The object was to insure that a person should not be compelled, when acting as a witness in any investigation, to give testimony which might tend to show that he himself had committed a crime. The privilege is limited to criminal matters, but it is as broad as the mischief against which it seeks to guard. 142 U.S. at 562, 12 S.Ct. at 198.
The Court’s fear is that if the witness were not protected in his refusal to answer prior to the criminal trial, the government could effectively short-circuit the trial privilege by compelling answers in anticipation of a trial and thereafter introducing them. By compelling answers, the government could abandon independent investigation in favor of forcing the individual to identify himself as a criminal suspect, providing the first “link in the chain” of prosecution. By validating a witness’s refusal to answer outside the criminal trial context, the Court removes the incentive for employing compulsion against him, effectively relegating prosecuting officials to independent investigation, and thereby securing the privilege’s purpose.
The Court’s logic in extending the privilege beyond the trial context is not unassailable, because it assumes, perhaps un-warrantedly, that the answers compelled would be admissible at a later trial. In fact, there are almost always two ways for a court to remove the incentive for prosecutorial circumvention of a defendant’s trial privilege. The first is to recognize a privilege prior to a criminal trial to refuse to answer. The second is to require the witness to answer initially, and, upon invocation of his privilege, to exclude his compelled answers when sought to be introduced in his subsequent criminal trial, placing the burden on the government to prove beyond a reasonable doubt that the evidence it has compelled has not provided a link in the chain of evidence presented in court. The second approach closely resembles the Court’s approach to coerced confession cases, which is instructive because as in our case the initial disclosure is not sought in a judicial or administrative proceeding where the rules of evidence or privilege have an immediate impact.
Analysis of the reasons for the development of the privilege and of the alternative approaches available to a court reveals the extent of the majority’s misconception of the nature of the constitutional privilege. The “use immunity” which Garner requests is simply the second approach — a decision that faced with the constitutional imperative, the trial court must exclude compelled testimony in a criminal trial. This approach is closer to the original view of the Amendment’s protection as a privilege restricted to the criminal trial than the majority’s position. The fact that the history of the Fifth Amendment has seen the expansion of the privilege to situations where it may be claimed in anticipation of danger of use in a criminal prosecution in no way relieves the court of the responsibility of deciding whether it is applicable when claimed during the course of that trial. Even accepting the majority’s assumption that Sullivan recognized the privilege as a defense to a prosecution for failure to answer particular questions on a tax return, there is no logical basis for assuming that the Court thereby meant to eliminate the trial privilege. Precisely the opposite would be the logical conclusion — allowing the privilege to operate both before and at trial best serves the Court’s purpose in extending the privilege of preventing the circumvention of an “accusatorial system.”
More importantly, the alternative approaches available to the court provide a *248reasonable solution to the problem of reconciling the government’s need for information with the purposes of the privilege presented by “neutral,” information-gathering, regulatory inquiries, enforced by statutory compulsion. The majority, as noted earlier, indicates that it considers the privilege applicable in the context of regulatory inquiries (n.14)— in this respect we are in complete agreement. Ordinarily, coercion of incriminating answers gives rise to a claim of privilege — the result should not be altered by the fact that the coercion is created by a neutral regulatory statute. No matter how overwhelmingly non-criminal and apparently innocuous the purpose of the inquiry, self-disclosure systems contain a too obvious and substantial potential for use in criminal prosecution to ignore the Fifth Amendment dangers created. From the point of view of the individual, as witnessed by Garner’s situation, even the most innocuous questions may call for incriminating answers. Were these excluded from the scope of the privilege, the opportunity would arise for comprehensive schemes of self-reporting in non-criminal areas to become data banks containing numerous “admissions” of criminal activity, available without limitation to prosecuting authorities. The use of computers potentially capable of analyzing data and reporting criminal violations makes both real, and substantial, the danger that the unfettered exercise of governmental power may degenerate into a pattern of extracting confessions on forms for use in later prosecutions. The application of the privilege fulfills the traditional purpose of removing the incentive to fall into oppressive modes of prosecutorial conduct; to do other-wise would “. ultimately license widespread erosion of the privilege through ‘ingeniously drawn legislation.’ ” Marchetti, supra, 390 U.S. at 51, 88 S.Ct. at 704.
Because of the alternative approaches available in applying the privilege, the fact that the privilege should apply in the regulatory context does not necessarily indicate when it should apply. I do not share the majority’s certainty, in reliance on the Sullivan dictum, that the privilege is applicable when a self-incriminating question is first asked. Because Garner asserted his privilege at trial rather than as a defense for refusing to answer, it is not necessary to decide when the privilege first becomes applicable in his case; in either event he is entitled to exclude the answers at trial if they were involuntary. Nevertheless, the rationale selected is of considerable consequence, both in reconciling the government’s need for information with a potential defendant’s need for protection, and because substantially different decisions are called for, depending on when the privilege applies.
I think it proper to sustain the privilege in the courtroom to exclude compelled answers in situations where the privilege might not be a defense to a prosecution for failure to answer those same questions in a questionnaire. Particularly in the regulatory context, a different complex of values is involved in deciding the applicability of the privilege in each instance. Substantial nonpros-ecutorial social purposes are served by compelled disclosure. Analyzed in light of the attenuated possibility of prosecu-torial use of the answers, the government’s interest might reasonably preclude a court from recognizing a privilege to refuse to answer the inquiries. When the answer is introduced at trial, however, the non-criminal purpose justifying compelled responses has been served and a different decision is involved, weighing a different complex of values and reanalyzing the threat of governmental misconduct.
There may be no violation of the Fifth Amendment in initially compelling answers because even if incriminating, they are not required in a “criminal case”; the violation of the “privilege” arises only when the government seeks to introduce compelled evidence in a criminal trial. The fact that compelled self-incrimination is not per se a constitu*249tional violation 9 indicates that the Fifth Amendment language is primarily a constitutional directive to the courts on the admissibility of certain types of testimony and evidence in a criminal trial. The purpose is to prevent the exercise of governmental power which might remove the prosecutorial burden from the government and place it on defendants, but the method by which it is accomplished is simply the exclusion of evidence — answers which are or have been compelled out of the mouth of the person against whom they are sought to be used.
Thus it is important to clarify whether (a) the basis for the possible exclusion of the tax return is that statutorily compelling his self-incriminating answers violated the purposes of the Fifth Amendment and tainted the answers when Garner filed the tax return, so they were excludable by his privilege when later introduced in a criminal proceeding, or (b) that while Garner was not privileged to refuse to answer the questions, he is privileged to prevent statutorily compelled answers from introduction in a criminal trial because they then give rise to a cognizable Fifth Amendment danger. The first approach, by recognizing a privilege not to answer, which the majority not only prefers but apparently thinks required by the nature of the privilege, balances a prospective threat that the answer may in the future incriminate in a judicial proceeding against the government’s need for noncriminal information. It creates a prophylactic rule which states that the noncriminal social purpose is not as compelling as the danger of governmental circumvention of the privilege’s purposes, but does so in a context where the government’s need is certain and the danger speculative. The choice is not necessarily between serving one or the other. If one limits the privilege to the point when compelled evidence is introduced, there is no longer a socially neutral purpose being served, and the threat of a breakdown of the accusatorial system and a whipsawing of defendants is a reality.
The Supreme Court’s relatively few pronouncements on the application of the privilege in the area of regulatory self-disclosure inquiries leave open the question of which approach is appropriate. As the majority indicates (p. 232), the Court has distinquished between two different types of disclosure statutes: those designed to elicit information about specific activities from a specific group of individuals inherently suspect of criminal activities,10 and those asking questions neutral on their face, of the public at large, for valid neutral social purposes, and which require self-reporting. Despite the dictum in Sullivan, the Court has never recognized a privilege to refuse to answer questions of the second variety, whether the privilege is asserted by a total failure to file a return or claimed as to specific questions. A priori, the Court has never validated the majority’s notion that the privilege is available and must be asserted by a refusal to answer neutral questions, rather than by answering and seeking to exclude the answers if they are later offered at a criminal trial. In fact, the majority completely misconstrues the point of the distriction between the two types of inquiries, which, if anything, suggests that the course Garner took in invoking his privilege was the preferable, if not the only, method.
Sullivan held that the privilege did not justify a total failure to file a tax return, since a less extravagant method of claiming the privilege was available. In the Marchetti-Albertson11 line of cases the *250Court, reversing itself,12 recognized the privilege as a defense to a prosecution for failing to register and file returns under a variety of non-criminal statutes requiring disclosures of activities in areas “permeated with criminal statutes.” The Court distinguished Sullivan’s requirement that the return be filed by pointing out the differences in the two types of inquiries involved (390 U.S. at 48-49, 88 S.Ct. 697); thus the distinction which the majority relies upon serves to distinguish cases in which the defendant may refuse to file and those in which he may not, and is somewhat irrelevant in Garner’s case. Nevertheless, as the Marchetti line of cases are the only instance in which the Court has excused a failure to answer a regulatory inquiry, the reasons underlying the distinction are instructive in determining the issue of when the privilege applies in the regulatory context.
The original reason for the distinction was that a court trying a “failure-to-file” prosecution, in the case of questions directed at suspect groups in areas permeated with criminal statutes, could be far more certain that the reason for the refusal to file was that the answers called for were incriminating. Albertson, supra, at 79. See California v. Byers, 402 U.S. 424, 91 S.Ct. 1535, 29 L.Ed.2d 9 (1971), at 469 (Brennan, J., dissenting). The objection of the Sullivan court to extending the privilege to allow a complete failure to file a tax return was that it made the taxpayer the sole judge of the merits of his claim. Under the standard of United States v. Burr, 25 F.Cas. p. 38, No. 14,692e (1807), and Hoffman v. United States, 341 U.S. 479, 71 S.Ct. 814, 95 L.Ed. 1118 (1951), a claim of privilege must be honored when it appears that a “real and substantial” possibility exists that an honest answer would provide a link in the chain of incriminating evidence. While the witness cannot be compelled to prove the basis of his claim, since to do so would abandon it, the claim is subject to review to prevent obviously frivolous claims. Because of the impossibility of administratively reviewing a claim if no return is filed, the extension of the privilege in Sullivan would have subjected the tax system to bogus invocations of the privilege. Because of the difference in the questions, the danger of fraudulent claims was far less compelling in the Marchetti line of cases, the relevance of the distinction being to determine when a defendant could fail to file a return.
The rationale of the distinction — the Sullivan Court’s fear of making the taxpayer judge of the merits of his case— clearly cuts in favor of the course Garner took. By answering and asserting a privilege at trial, Garner subjects his claim of privilege to the acid test: not whether the answer has a substantial possibility of being incriminating, but whether it is so in fact. The majority’s suggestion that Garner’s failure to claim the privilege on the return makes him “the final arbiter of the merits of the claim. . . .” (p. 5, 1. 24-30, p. 5a, 1. 1-11 and n. 19), is completely illogical. *251Garner’s interjection of the privilege at trial, after answering, is the only situation in which he does not, to some extent, judge the merits of his claim.13 Hence, Garner would not possibly immunize himself from prosecution by “volunteering” information on a tax return. To the extent that a criminal includes information not reasonably called for by the questions, his answers are not statutorily compelled, and hence not subject to a claim of privilege when introduced at trial. As to those answers compelled by statutory penalty, the judge hearing the claim of privilege will determine whether the answers are in fact incriminating, rather than, as under the majority’s preferred approach, whether there is a substantial possibility that refused answers might incriminate.
A second reason for the distinction is that when the area of inquiry is permeated with criminal statutes, anyone filing a return, even one claiming a privilege, identifies himself as a criminal — thereby undermining the accusatorial system at the investigatory stage. Again, the thrust of the distinction is aimed at determining when a complete failure to file is justified, but suggests that when the intrusion on Fifth Amendment values is less immediate, the defendant should be relegated to an invocation of the privilege in a manner which gives greater scope to the government’s interest, such as Garner’s method.
The third purpose of the distinction is that a statute which directs inquiries to an area permeated with criminal statutes alerts the Court to the predominance of prosecutorial over neutral goals and the concealment of Fifth Amendment dangers. The Court emphasized that every question and requirement in Marchetti tended to identify the registrant as a criminal, unlike Sullivan, where some, if not all, of the questions served only neutral purposes, even when directed at criminal income. As a result, the Court gave less deference to the noncriminal goals, which it recognized in Marchetti, allowing them to be defeated to prevent governmental circumvention of the accusatorial system. In the present case, as in Byers and Sullivan, the questions do not reveal an overt attempt to whipsaw defendants. The relevance of the difference is either that regulatory statutes which are neutral and present statistically small possibilities of compelled self-incrimination do not come within the ambit of the privilege at all, or it goes to when the privilege may be claimed. Since the majority and I agree that the privilege applies in this case, presumably the majority offers the distinction for the latter reason. Again, the majority draws precisely the opposite conclusion from that warranted by the rationale for the distinction. By allowing the suspect Marchetti variety statute to be challenged by a total failure to file, and relegating challenges to neutral non-suspect Sullivan type statutes either to be made on the return, or by answering and asserting a privilege at a subsequent trial (the point has never been decided), the Court indicated that when the government is pursuing legitimate non-criminal goals, the application of the privilege should be in a manner that allows as narrow an interference as possible consistent with the preservation of constitutional protection. The path Garner used is more consistent with the Court’s rationale than the majority’s approach. The fact that answering fulfills entirely the neutral aims of the questions suggests that Garner’s approach may be ultimately determined by the Court as the proper method of reconciling the government’s neutral aims with the privilege.14
*252Nor is there anything in the plurality opinion15 in California v. Byers, supra, on which the majority relies (p. 232, n. 17, n. 18), to support the majority’s decision that Garner’s answers are admissible at trial because he did not claim a privilege on the return; if anything, the plurality opinion suggests that in the case of a non-suspect Sullivan type inquiry which poses a low probability of incrimination when assessed over the entire class at whom the questions are directed, the balance of interests is such that even a person facing a high risk of incrimination must answer initially, and assert his privilege at trial. The California Supreme Court held that absent a use immunity, Byers was privileged to refuse to stop and identify himself when involved in an accident if to do so would incriminate him. The plurality, noting the statute’s neutral purpose, need for self-reporting, direction at the general public, and low danger of self-incrimination, thought the ease controlled by Sullivan. As I read the plurality’s opinion, the question they were answering was whether Byers’ failure to stop and identify himself, the equivalent of failing to file in Sullivan, was constitutionally privileged. The California Supreme Court said it was; the plurality said that, as in Sullivan, no use immunity was required. Particularly because the plurality thought Byers' identity non-testimonial under Schmerber v. California, 384 U.S. 757, 86 S.Ct. 1826, 16 L.Ed.2d 908 (1966), its rationale indicates that the decision is limited to the precise issue of whether a failure to stop is privileged, and decides neither what may be required thereafter,16 nor whether any of the information derived after the stop may be introduced at a criminal trial. Read this way, the Byers plurality is Sullivan — it clarifies the standards by which failure to answer or file is to be judged in light of Marchetti.
The reasoning of the plurality opinion supports Garner's method of asserting his privilege; only a slight extension of the plurality’s approach would require it. The rationale of the plurality was that since the non-criminal interests in obtaining a driver’s identity outweigh the danger of incrimination drivers face as a class, the privilege cannot be asserted as a defense for failing to answer by a particular driver, even if he faces a substantial risk of incrimination. Again, the Court’s solicitude for non-criminal interests served by disclosure supports Garner’s method of claiming his privilege. Were one to disregard both the fact that the plurality considered the compelled identity non-testimonial and hence not subject to a claim of privilege, and its explicit disclaimer of having decided whether privilege was available after identity was disclosed,17 and to read the plurality’s opinion to hold that whenever the neutral interest in receiving the answers outweighs the dangers of incrimination posed to the class, the individual must answer,18 then Garner’s method of asserting his privilege is the only one available. To assess the operation of the statute from a societal overview, in terms of the interests served and the risks of incrimination posed, would determine whether the disclosure requirement was an overt attempt to con*253struct a reporting system intended to whipsaw defendants and circumvent Fifth Amendment policies. If it were not, it might be appropriate not to defeat the regulatory interest served by disclosure even if the particular case revealed a high potential for prosecutorial use of the answers.
This would not mean, as the majority apparently thinks, that the privilege is thereafter unavailable to exclude the compelled answers;19 in fact, it must be available to reconcile the requirement that regulatory inquiries be answered with the need for constitutional protection. The approach is appropriate, both in judging a complete failure to file and to preclude the privilege on the return, only if the compelled witness is allowed to assert his privilege when the social purposes which the balance is intended to accommodate are served and the government is using the information solely to convict him. A decision not to extend protection to a “witness” who anticipates that answering will incriminate him in a subsequent criminal trial balances completely different interests than those involved when the answers are introduced at trial. At trial, as a majority of the Justices in Byers indicates (402 U.S. at 437-442, 459, 91 S.Ct. 1535), the risk of incrimination weighed must be that faced by the particular individual compelled to answer. The majority, in seeking to distinguish the governmental compulsion in this case from that declared constitutionally impermissible in Marchetti and Miranda, emphasizes the fact that the group queried by income tax returns, on the whole, faces a low risk of self-incrimination (p. 232). If the majority means to imply that Garner’s claim at trial should be judged'by that standard, then the majority flies in the face of its own recognition that the privilege is personal and the fact that under Hoffman, Burr, and a welter of other cases, the privilege is available when the individual faces a substantial possibility of self-incrimination — -not on a showing of illicitly intended governmental action.20 An approach which recognizes the privilege only when a question manifestly intends to incriminate would open a broad path for disingenuous circumvention of the purposes of the privilege.
Whether the Court ultimately holds that a taxpayer has a privilege to refuse to answer particular questions on his return, or, under the logic suggested by the Byers plurality opinion, holds that a taxpayer must first answer the questions and assert the privilege at trial, Garner is entitled to have his tax return excluded.
Under the latter approach, Garner must file a complete return, and the privilege is returned to its historic setting— that of the criminal trial. To have his claim of privilege sustained and the return excluded, Garner must show that he answered pursuant to a statutory directive which contained sanctions for not answering, and that his answers were incriminating. That is precisely what he has shown in this case. 26 U.S.C. § 7203 makes it a crime to fail to file any return, pay any tax, or supply any information. Answers are not optional or voluntary — the statute prescribes up to $10,-000 fines and one year in jail for ñoñ*254compliance. 26 U.S.C. § 7206 makes it a crime for a person to make and subscribe any return “which he does not believe to be true and correct as to every material matter. . . .”26 U.S.C. § 7203 commands that those subject to tax answer all questions posed in the return. On the standard 1972 1040 Form, immediately above the signature line, one reads: “Under penalties of perjury I declare that [this return] is true, correct, and complete.” At line 44 of the form the question asks taxpayer to state “Other” income, telling him to “(state nature and source)”. The “1972 Instruction for Form 1040”, (Point 1, page 8), states “Line 44 — Other-—Use this line to report and tell the source of any income you cannot find a place for on your return or other schedules. . . . ” At page 3, column 1, the taxpayer is cautioned: “Don’t forget to show occupations in space[s] in upper right corner just below social security blocks,” and then at page 6: “you have to report all income in whatever form received.” Among the examples is included “Embezzled or other illegal income.” Thus, there is no doubt that Garner’s answers were statutorily compelled within the meaning of the Fifth Amendment; they were incriminating and should have been excluded under the second approach. The government’s interest in the coercion, when introducing the answers at trial, is solely prosecutorial; as to that purpose, the Fifth Amendment embodies the constitutional decision that governmental interests in criminal law enforcement standing alone do not justify the forced extraction and subsequent introduction of the evidence.21
On the other hand, if the privilege ultimately is available when filling out a tax return, Garner still should have been privileged to have his returns excluded at his trial; under Mackey the privilege may be asserted at either time. The majority’s opposite conclusion can be correct only if there were no constitutional violation involved in Garner’s answering the questions. Despite the majority’s assertions about the nature of the privilege, the basis of the majority’s decision appears to be that Garner’s answers were given voluntarily:21a “Garner provided the source of his income on his return and failed to invoke his privilege.” (p. 231) Since the source of Gamer’s income is information which he is required to give under the threat of § 7203, his providing it is objectively involuntary. Nevertheless, the majority relies on the premise that since Garner might have refused to answer, by responding he evidenced that the answers were not subjectively compelled by the statute — hence voluntary for constitutional purposes. Because his answers are not produced by the proscribed compulsion, the majority reasons that as to him, the statute does not violate the purpose of the Fifth Amendment, and there is no constitutional violation.22 Note that the assumption that the answers were subjectively voluntary makes totally irrelevant the majority’s argument that the nature of the privilege precludes Garner’s objection at trial — if there is no constitutional viola*255tion, there is obviously no privilege at trial.
The majority’s assumption that Garner’s answers were subjectively voluntary rests entirely on the analogy, derived from Shushan v. United States, 117 F.2d 110 (5th Cir. 1941), and Stillman v. United States, 177 F.2d 607 (9th Cir. 1949), and recited by the majority opinion (p. 231-232), to the consequences of failing to invoke a privilege in a civil trial. Just as the failure to invoke the privilege evidences that the answers given in the civil trial are voluntary and are not subject to subsequent exclusion, the analogy runs, so here Garner evidenced that he was not coerced to answer, and therefore his answers may not be excluded. The analogy is totally unsound for a variety of reasons. First, Mackey rejects the logic. Mackey was privileged not to file his wagering tax return — the Court did not treat his doing so as indicating his answers were voluntary.
Second, the majority’s analogy is unsound because the applicability of the privilege in judicial proceedings is historically more certain than in other contexts, thereby making an answer on a tax return far less persuasive evidence that the answer is given voluntarily, rather than as a result of the penalties for not answering. “Courts indulge every reasonable presumption against waiver” and “do not presume acquiescence in the loss of fundamental rights.” Johnson v. Zerbst, 304 U.S. 458, 464, 58 S.Ct. 1019, 1023, 82 L.Ed. 1461 (1938); Carnley v. Cochran, 369 U.S. 506, 516, 82 S.Ct. 884, 8 L.Ed.2d 70 (1962). In order to find that by answering Garner has relinquished the constitutional protection he would otherwise enjoy, the government should logically have to rebut the presumption that his answer was involuntary — the majority presumes precisely the opposite. The factual inference that his answers are voluntary rests on the presumption that Garner knew of the availability and scope of his privilege. There is no evidence in this record to indicate that Garner knew of the privilege, and under Zerbst’s standard for the waiver of a known right, the court is directed to indulge the reasonable presumption that he did not. The presumption is reasonable not only because there is absolutely nothing on the return itself to make him aware of the privilege, and because it is not common knowledge that one may refuse to answer questions on a tax return, but also because it is still uncertain, to lawyer and layman alike, that such a privilege indeed exists.
The analogy is further unsound because even in the civil case to which the majority analogizes the tax return, if the witness were prevented from exercising his privilege by an erroneous rejection of his claim, a threat of contempt, or some other sanction being placed on his free exercise of the privilege, then when the evidence is later introduced, he has not voluntarily waived his objection and the court must hear it. That is precisely the issue in Garner’s case which the majority has assumed away, calling his answers “voluntary.”
Garner’s failure to invoke the privilege is the product of the same coercion which produced his incriminating answers — the threat of a criminal conviction under § 7203. The majority ignores the fact that when Garner answered the questions on the tax return, the only way for him to get a determination of the merits or applicability of his claim of privilege, if rejected by the Tax Service, was by asserting the privilege as a defense in a subsequent prosecution for failure to file a complete return. In a judicial proceeding if a judge disallowed his claim of privilege, Garner could go on and testify, without sanction. Here he must choose between alternative criminal prosecutions under either § 7203 or the gambling laws.
In this situation the majority is precluded from treating Gamer’s answers as “voluntary” by Garrity v. New Jersey, 385 U.S. 493, 87 S.Ct. 616, 17 L.Ed.2d 562 (1967), and the logic of a line of *256cases,23 recently reaffirmed by a unanimous Court in Lefkowitz v. Turley, supra, which indicate that when a sanction is placed on the free exercise of the privilege, the resulting answers are not voluntary, and their subsequent use violates the Fifth Amendment. In Garrity the Court confronted a situation in which police officers who were to be questioned during an investigative hearing were threatened with removal from office if they refused to answer the questions. The officers answered; the answers were admitted over their claim of privilege at a subsequent criminal trial, and the Court reversed. “The Court . . . held that in the context of threats of removal from office the act of responding to interrogation was not voluntary and was not an effective waiver of the privilege against self-incrimination. . . .” Lefkowitz v. Turley, 414 U.S. at 80, 94 S.Ct. at 323. Garrity defined a waiver as “the free choice to admit, to deny, or to refuse to answer.” In Garrity the threat of a civil sanction (which arguably served valid non-criminal purposes) was sufficient to make the answers constitutionally involuntary. Here the threat of a criminal sanction — a conviction under § 7203 if the court rejects his claim of privilege —should likewise make the answers involuntary as a matter of law.
The analysis of the particular facts of this case which the majority foregoes emphasizes Garner’s dilemma. He has no effective procedure to invoke the privilege on the return even if he were aware he had one.24 The provisions of the Internal Revenue Code applicable to gamblers provide different rules in calculating taxable income than those available to non-gamblers. A person whose income is derived from wagering may deduct his wagering losses only to the extent of his winnings from wagering, 26 U.S.C. § 165(d). Gambling losses may not be carried back or carried over [See 5 Mertens, Law of Federal Income Taxation § 28.85], and a gambler whose losses offset his winnings must nevertheless report all winnings as gross income and losses as deductions. McClanahan v. United States, 292 F.2d 630, 631 (5th Cir. 1961), cert. denied, 368 U.S. 913, 82 S.Ct. 193, 7 L.Ed.2d 130. Thus, as a practical matter, in order for Garner to claim a privilege on the return he would have to enter a claim of privilege at the lines calling for his source of income, his profession, and throughout the return where step by step calculations are called for. He would have disclosed only his name and taxable income figure. This would subject him to prosecution for failure to file a return at all, United States v. Klee, 494 F.2d 394 (9th Cir., 1974), United States v. Porth, 426 F.2d 519 (10th Cir. 1970), and the substantial risk that a court would determine that he was not privileged to do so under Sullivan. If he only claimed a privilege on the line where his profession or source of income was requested, he would reveal by the method of his calculations that he was a gambler. Thus the “free choice” to refuse to answer required by Garrity is missing — in this case the choice is to run the risk of being convicted under § 7203 because of a judicial determination that he answered too few questions, that his privilege did not extend to certain subjects, or that his claim of privilege was invalid because under Hoffman’s standard he did not face a real possibility of incrimination.
The strongest indication that his answers are involuntary, however, is provided simply by the fact that no court has ever held there is a privilege which excuses failure to answer questions on the return. I venture that no lawyer whom Garner consulted when filling out the return could have told him with certainty that the law recognized such a privilege — only that it might. He had to run the risk of a criminal conviction to establish the existence of his privilege. Because of that threat, his failure to *257claim a privilege on the return does not indicate that his answers were not coerced by penalties under § 7203; on the contrary, as an evidentiary matter it strongly appears the answers are involuntary. Under Garrity, the majority is precluded as a matter of law from considering the answers voluntary, or from holding that the privilege is not applicable at trial to exclude them.
The majority, in support of the decision not to exclude the return, suggests that Congress should decide whether a “use immunity” is appropriate (n. 20). It seems to me a fundamental of constitutional adjudication that confronted with a legislative or prosecutorial action arguably in conflict with the Fifth Amendment, regardless of what mix Congress might choose among prose-cutorial and taxing purposes, it is the duty of the court to decide whether the evidence violates the constitutional privilege. “A witness does not need any statute to protect him from the use of self-incriminating testimony he is compelled to give over his objection. The Fifth Amendment takes care of that without a statute.” Adams v. Maryland, 347 U.S. 179, 181, 74 S.Ct. 442, 445, 98 L.Ed. 608 (1954). If, as the majority assumes, Garner was privileged not to answer the questions, then the constitutional violation involved in coercing him involuntarily to answer requires that the answers be excluded at trial. The government may not use the information because it may not initially compel it.
Moreover, regardless of when the privilege first becomes applicable, this case involves a situation in which a judicial recognition25 of “use immunity” is entirely appropriate for a “rational accommodation between the imperatives of the privilege and the legitimate demands of government to compel citizens to testify.” Kastigar v. United States, 406 U.S. 441, 446, 92 S.Ct. 1653, 1657, 32 L.Ed.2d 212 (1972). The Court has recently reaffirmed the principle that the price the government must pay to pursue important neutral goals is sufficient immunity from the prosecutorial use of required information. Lefkowitz v. Tur-ley, supra, 414 U.S. at 78, 79, 81, 84, 94 S.Ct. 316. The rejection of use immunity in Marchetti, and by Justice Harlan in Byers, cited by the majority was for reasons inapplicable to the present situation.26 Here, the imposition of a use re*258striction effectuates the paramount Congressional purpose of raising revenue by-removing the dangers which lead most criminal income to be either not reported or reported under misleading labels.27 Becadse the immunity would apply only to prosecutions unrelated to the tax laws, it would affect only a small proportion of the criminal prosecutions using tax returns. Here, unlike Marchetti, the incidental criminal purpose is severable from the revenue raising purpose in obtaining information necessary for determining tax liability, making a use restriction appropriate to prevent the unconstitutional prosecutorial use of answers while advancing the overriding Congressional purpose in passing the statute — taxation .28
The majority declares that accepting Garner’s claim of privilege would create the “disquieting” and “unpalatable” (p. ■ 232) result that every public response would be subject to a claim of privilege when introduced in a criminal trial. In fact, every response would not be subject to a claim of privilege because many governmental inquiries are not accompanied by statutory penalties for refusing to answer. As to those that are, the majority purports to recognize a privilege not to answer. Thus the sum total of answers over which the majority is dismayed are those which are given when a privilege might have been claimed, but was not. Of those, if the answer was voluntary in fact, then it may not be shielded by a claim of privilege. This leaves as the real source of the majority’s anxiety the prosecutorial windfall case where no privilege was claimed but the answer was involuntarily given. That is an inadequate constitutional basis for the majority’s decision.29
The judgment ought to be reversed.
. See generally, Corbin, Legal Analysis and Terminology, 29 Yale L.J. 163, 167-8 (1919).
. 26 U.S.C. §§ 4411-4412.
. Justices Douglas and Black thought Mackey entitled to relief, assuming a constitutional violation. 401 U.S. at 713, 91 S.Ct. 1160. Justices Brennan and Marshall did not think Marchetti was controlling, but indicated Mackey would have been entitled to use immunity if his answers had been introduced in a prosecution unrelated to the statute under which disclosure was compelled, as in our case. 401 U.S. at 710-711, 91 S.Ct. 1160. On his theory of retroactivity Justice Harlan did not reach the issue. 401 U.S. at 700, 91 S.Ct. 1160.
. California v. Byers, 402 U.S. 424, 434 n. 6, 91 S.Ct. 1535, 26 L.Ed.2d 9 (1971).
. California v. Byers, supra, at 448-451, 91 S.Ct. 1535 (Harlan, J., concurring). Friendly, The Fifth Amendment Tomorrow: The Case for Constitutional Change, 37 U.Cin.L.Rev. 671 (1968) ; McKay, Self-Incrimination and the New Privacy, 1967 S.Ct.Rev. 193, 193-98 (1967).
. See Miranda v. Arizona, 384 U.S. 436, 459, 86 S.Ct. 1602, 16 L.Ed.2d 694 (1966); Friendly, supra, at 677-79. See generally, Levy, Origins of the Fifth Amendment, The Right Against Self-Incrimination (1968).
. Friendly, supra, at 678.
. Grosso v. United States, 390 U.S. 62, 76, 88 S.Ct. 709, 718, 19 L.Ed.2d 906 (1968).
. The government may compel self-incriminating admissions from a witness so long as he is given immunity from their use against Mm.
. For instance, possessors of unlawful firearms, persons engaged in unlawful wagering activities, communists, and persons dealing in unlawful drugs.
. Marchetti v. United States, 390 U.S. 39, 88 S.Ct. 697, 19 L.Ed.2d 889 (1968) ; Grosso v. United States, 390 U.S. 62, 88 S.Ct. 709, 19 L.Ed.2d 906 (1968) ; Haynes v. United *250States, 390 U.S. 85, 88 S.Ct. 722, 19 L.Ed.2d 923 (1968) ; Leary v. United States, 395 U.S. 6, 89 S.Ct. 1532, 23 L.Ed.2d 57 (1969) ; Albertson v. Subversive Activities Control Board, 382 U.S. 70, 86 S.Ct. 194, 15 L.Ed.2d 165 (1965).
. Marchetti overruled United States v. Kah-riger, 345 U.S. 22, 73 S.Ct. 510, 97 L.Ed. 754 (1953), and Lewis v. States, 348 U.S. 419, 75 S.Ct. 415, 99 L.Ed. 475 (1955). The Court had reasoned that since the registration requirement and return dealt with prospective acts, there was no valid claim of privilege involved, because the statutory compulsion to file could be avoided entirely at the choice of the defendant, by not entering voluntarily on the conduct which made failure to file criminal. Marchetti rejected this concept of implied “waiver” (strictly speaking, it is not a waiver at all) because the filing requirement focused on past acts and because the privilege was held to apply to prospective acts. While the Court’s rejection of the “waiver” in Marchetti is not controlling, the majority opinion neither reflects the Court’s hostility to implied “waivers” nor does it give the required scrutiny of the circumstances in the present case before implying such a “waiver.” See 390 U.S. at 51, 88 S.Ct. 697.
. See 86 Harv.L.Rev. 914, 915 n. 9 (1973).
. In Marchetti Chief Justice Warren would have adopted the use immunity desired by the government in order to preserve the neutral taxing purpose of the statute, despite the predominance of impermissible criminal purposes served. He felt that Sullivan and Marchetti inquiries should not be treated differently — in both situations the offense to the purposes of the privilege results from transmittal of the information for prosecu-*252tonal use, and in both situations restriction on the use of answers adequately protects the defendant. 390 U.S. at 81, 88 S.Ct. 697. The majority in Marchetti rejected his approach because it thought the neutral and criminal goals of the statute unseverable.
. Four justices dissented, thinking Marchetti controlling. Justice Harlan concurred in the result reached by the plurality, but for sub- . stantially different reasons.
. 402 U.S. at 434 n. 6, 91 S.Ct. 1535.
. Id.
. Such a reading of Byers would effectively “overrule” the suggestion of the Sullivan dictum. The information required by particular questions on tax returns poses a low danger of incrimination when viewed as a statistical matter — probably lower than that faced by persons stopping to give their identity after car accidents — and otherwise conforms to all of the indicia of neutrality found important by the Byers plurality. The answer would therefore have to be given on the tax return.
. It is difficult for me to determine how the majority reads the Byers plurality opinion. The majority, at the same time it cites Byers for the proposition that the returns should not be excluded (p. 232), advances the incompatible argument that Garner was privileged not to respond; the whole point of the Byers plurality opinion was that Byers was not so privileged. If, as did the original dissent (p. 236), the majority reads Byers to mean that the compelled identity is admissible in a subsequent criminal trial, then the implied “waiver” position the majority advances as the ground for its decision is untenable. The identity in Byers would be admissible because neutral regulatory inquiries do not raise a valid privilege claim despite high individual risks of incrimination • — for precisely the same reason, Garner would not be privileged to refuse to answer (see n. 18, supra), and his answers would be admissible at trial.
. A “group focus” for assessing the risk of incrimination posed by an inquiry advocated by the majority was in fact rejected by a majority of the Court in Byers.
. See 85 Harv.L.Rev. 269, 270 (1971).
. The majority, in seeking to avoid the critical factual inquiry of whether Garner’s incriminating answers were “provided” of his own free will or because of compulsion to answer, contends that the question of volun-tariness here “diverts from the critical issue.” (p. 232). In fact, as in all Fifth Amendment cases, it is the critical issue. Although the character of governmental conduct may vary, a statutory penalty for not answering is a compulsion to answer (compare p. 232), and an involuntary answer coerced by that compulsion violates the privilege if it incriminates.
. I have discussed the voluntariness of Garner’s answers in terms of whether there is a violation of his privilege, rather than under the rubric of waiver. Logically, if his answers are not extracted by the operation of the statute, there is never a constitutional claim to be waived. Nevertheless, the difference is one of form only. Under the rubric of waiver, the argument is that a constitutional claim is raised by statutory coercion which is not waived by answering unless the answers are voluntary.
. See, e. g., Spevack v. Klein, 385 U.S. 511, 87 S.Ct. 625, 17 L.Ed.2d 574 (1967).
. 86 Harv.L.Rev., supra, at 916 n. 13.
. E. g., Murphy v. Waterfront Commission, 378 U.S. 52, 84 S.Ct. 1594, 12 L.Ed.2d 678 (1964).
. In Marchetti the majority rejected use immunity because the prosecutorial purposes of the wagering tax statute were indicated on the face of the statute and were unseverable from the incidental neutral taxing purpose. The majority was unwilling to substitute its own judgment for Congress’ as to the appropriate mix between criminal and taxing purposes. That reason is clearly inapplicable under the general income tax statutes — any incidental prosecutorial use of the information can appropriately be severed from the paramount purpose of raising taxes.
Justice Harlan rejected use immunity in Byers for a reason which indicates it is appropriate in this case. Under Justice Harlan’s own approach to balancing, the factor which ultimately tipped the scale in favor of finding the privilege inapplicable was that the imposition of a use immunity on socially necessary information would have the effect of defeating valid criminal law enforcement purposes which did not depend on compelled information. Use immunity would place what he considered a difficult burden on proving evidence was independently obtained, defeating prosecutions in which there was no taint. However, in Byers the police served both data-gatliering and prosecutorial functions — hence the difficulty in proving that information compelled had not been used. In the present case different branches of the executive conduct the two functions, and it is possible to keep the tax returns insulated. Under C.F.R. § 301.6103 (a)-1(f) and (h), returns can only be transmitted for inspection or use on a demand from the Justice Department, not at the Service’s initiative, except in tax prosecutions. The regulations could be amended to prevent transmittal and procedures implemented to ease the difficulty of proving freedom from taint. Because of the relative ease with which the holes in the present “confidentiality” system of § 7213 can be plugged, I doubt valid prosecutorial efforts will be undermined by a use immunity on information in tax returns.
. See Bittker, Federal Income, Estate and Gift Taxation 117 (3d ed. 1964).
. The idea of a use restriction on information required by regulatory inquiries has been almost uniformly approved by commentators. See, e. g., Friendly, supra, at 720; Mansfield, The Albertson Case: Conflict Between the Privilege Against Self-Incrimination and the Government’s Need for Information, 1966 S.Ct.Rev. 103, 159-66; McKay, supra, at 232; 86 Harv.L.Rev., supra, at 922.
. The majority’s concern
“. . . appears to mean that current technological progress enabling the Government more easily to use an individual’s compelled statement against him in a criminal prosecution should be matched by frank judicial contraction of the privilege against self-incrimination lest the Government be hindered in using modern technology further to reduce individual privacy.” 402 U.S. at 464, 91 S.Ct. at 1556 (Brennan, J., dissenting).