Fifth Amendment First Principles - The Self-Incrimination Clause

Yale Law School
Yale Law School Legal Scholarship Repository

Faculty Scholarship Series Yale Law School Faculty Scholarship
1-1-1995
Fifth Amendment First Principles: The Self-

Incrimination Clause
Akhil Reed Amar
Yale Law School
Renee Lettow Lerner

Yale Law School
Follow this and additional works at: http://digitalcommons.law.yale.edu/fss_papers

Part of the Law Commons
Recommended Citation
Amar, Akhil Reed and Lerner, Renee Lettow, "Fifth Amendment First Principles: The Self-Incrimination Clause" (1995). Faculty
Scholarship Series. Paper 993.
http://digitalcommons.law.yale.edu/fss_papers/993
This Article is brought to you for free and open access by the Yale Law School Faculty Scholarship at Yale Law School Legal Scholarship Repository. It
has been accepted for inclusion in Faculty Scholarship Series by an authorized administrator of Yale Law School Legal Scholarship Repository. For
more information, please contact [email protected].
FIFH AMENDMENT FIRST PRINCIPLES:
THE SELF-INCRIMINATION CLAUSE
Akhil Reed Amar*
and Renge B. Lettow**
TABLE OF CONTENTS
INTRODUCTION ............................................ 857
I. THE PUZZLE ........................................... 860
A. "Person"?......................................... 861
B. "Compelled"? ..................................... 865
C. "In Any Criminal Case"? ......................... 874
D. "Witness"? ........................................ 883
E. What's the Big Idea? .............................. 889
H. THE SOLUTION ........................................ 898
A . "Person" .......................................... 901
B. "Compelled" ...................................... 904
C. "In Any Criminal Case".......................... 909
D. "Witness" ......................................... 919
E. The Big Idea(s) ................................... 922
CONCLUSION .............................................. 927
INTRODUCTION
No person... shall be compelled in any criminal case to be a witness
against himself .... I
The Self-Incrimination Clause of the Fifth Amendment is an un-
solved riddle of vast proportions, a Gordian knot in the middle of
our Bill of Rights. From the beginning it lacked an easily identifi-
able rationale; in 1789, the words of the clause were more a slogan
than a clearly defined legal rule, and in the preceding four centuries
* Southmayd Professor, Yale Law School. B.A. 1980, J.D. 1984, Yale. - Ed.
** Student, Yale Law School. A.B. 1990, Princeton; M.Litt. 1993, Oxford. - Ed. This
article is independent of, but builds upon, an earlier article. See Akhil Reed Amar, Fourth
Amendment FirstPrinciples,107 HARV. L. REv. 757 (1994).
For helpful comments on early drafts, we thank Bruce Ackerman, Vik Amar, Ian Ayres,
Steven Bibas, Jon Blue, Paul Cassell, Miran Damaska, Steve Duke, George Fisher, Abe
Goldstein, Nancy J. King, Stan Krauss, John Langbein, David Luban, Bill Stuntz, Bill
Treanor, Eugene Volokh, Stephen Williams, and workshop members at the University of
California at Davis, Chicago-Kent School of Law, George Washington University, the Uni-
versity of Pennsylvania Law School, Quinnipiac College School of Law, and Yale Law
School.
1. U.S. CONST. amend. V.
Michigan Law Review [Vol. 93:857
the slogan had stood for at least four different ideas. 2 Today, things
are no better: the clause continues to confound and confuse. Be-
cause courts and commentators have been unable to deduce what
the privilege is for, they have failed to define its scope in the most
logical and sensible way. In this article we try our hand at solving
the riddle and untying the Gordian knot. We propose both a ra-
tionale for, and a definition of the proper scope of, the Self-Incrimi-
nation Clause.
The Supreme Court's interpretation of the Fifth Amendment is
currently in a jumbled transitional phase. The key question, though
rarely recognized as such, is what sort of immunity the clause re-
quires before a person may be made to tell all outside his own
"criminal case," beyond the earshot of the petit jury. Over the
years the answers have varied considerably. In its 1892 chestnut
Counselman v. Hitchcock,3 the Supreme Court required total
("transactional") immunity from prosecution whenever a person is
compelled to testify against himself outside the courtroom. But in
1972, Kastigarv. United States4 in effect overruled Counselman and
established a new, narrower rule of "use plus use-fruits immunity":
a person forced to speak pretrial may be prosecuted, but neither the
compelled statement nor any evidence it led to ("fruits") can be
introduced in the criminal trial. Kastigar provided no persuasive
basis for stopping where it did in fashioning its new rule, and the
Court is slowly - if not always consciously - chipping away at the
foundations of the new rule, emphasizing the difference between
testimony and physical evidence. In this article we argue that the
time is ripe to take a decisive step in the direction the Court has
been leaning: the Court should move beyond the way station of
Kastigarand declare that a person's (perhaps unreliable) compelled
pretrial statements can never be introduced against him in a crimi-
nal case but that reliable fruits of such statements virtually always
can be. Thus, the government should be allowed to require a sus-
pect to answer relevant questions in a civilized pretrial hearing pre-
sided over by a judge or magistrate. Under penalty of contempt, a
suspect must answer truthfully, but he will be entitled to "testimo-
nial immunity": that is, the compelled words will never be intro-
duced over the defendant's objection in a criminal trial - the
defendant will never be an involuntary "witness" against himself
2. See John H. Langbein, The HistoricalOrigins of the Privilege Against Self-Incrimina-

don at Common Law, 92 Mxci. L. Rnv. 1047, 1084-85 (1994); infra section I.E.
3. 142 U.S. 547 (1892).
4. 406 U.S. 441 (1972).
March 1995] Fifth Amendment First Principles
"in" a "criminal case" - but the fruits of these compelled pretrial

words will generally be admissible.
This clean rule of testimonial immunity would openly vindicate
the Court's recent emphasis on reliability as a preeminent criminal
procedure value, but it has many other virtues as well. For starters,
it has some important history on its side. Testimonial immunity,
supported by English precedent, was the majority rule in America
before Counselman and was explicitly endorsed by Congress at the
time it adopted the Fourteenth Amendment, which reglossed the
Bill of Rights and made its privileges and immunities applicable
against the states. More importantly - as the long history of com-
pelled self-incrimination is admittedly tangled - testimonial immu-
nity makes the best sense of the words of the Fifth Amendment and
kindred constitutional provisions, like those of the Sixth Amend-
ment. Read our way, the words of the clause fit neatly together and
tightly cohere with the rest of the Constitution. In short, our read-
ing works, textually and functionally - no small thing in so muddy
a field. Indeed, our proposal creates a rare win-win solution in
criminal procedure: more guilty persons will be brought to book,
while the plight of the truly innocent defendant will be improved.
Once we see the clear meaning of the rule against compelled
self-incrimination - that self-incriminating words compelled from
a defendant must be excluded from his criminal case - and the
best reason underlying that rule - reliability - we can also see the
independent roles that other rules must play, radiating from other
clauses of the Constitution and motivated by other rationales. As
we shall see, in both civil and criminal contexts, needlessly intrusive
questioning, fishing expeditions, and offensive impositions upon a
person's body raise obvious Fourth Amendment concerns; attempts
to probe a citizen's mind for heresy or political incorrectness or a
map of his inner life or conscience implicate the First Amendment;
unregulated and lawless police behavior must be stamped out in the
name of due process and the rule of law; and so on. By misreading
- and often overreading - the scope and rationale of the Self-
Incrimination Clause, courts and commentators have often ob-
scured the proper role of other clauses, and so another happy side
effect of our reading is that it yields a sensible division of labor
among constitutional provisions.
The very breadth of the Self-Incrimination Clause, as currently
construed, creates huge challenges, substantively and doctrinally,
for this article. Substantively, we must note that an enormous
amount of modern criminal law enforcement has been shaped by
Michigan Law Review [V'ol. 93:857
the Self-Incrimination Clause, as (mis)construed over the years. By

narrowing government's ability to tap suspects as testimonial re-
sources in civilized pretrial proceedings, the clause has driven some
interrogation underground into less-than-civilized police station
rooms and squad cars; has spurred on surprise searches, wiretaps,
and other intrusions that fall outside the Fifth Amendment; has in-
creased pressure to use "sting" operations and government infor-
mants, who often must be bribed in unappealing ways; and has
ramified in countless other directions. A different reading of the
Self-Incrimination Clause would likewise ripple out far and wide,
and a single article cannot trace all these likely ripples with preci-
sion. On the other hand, the vast substantive significance of the
Self-Incrimination Clause only deepens the puzzle of current doc-
trine and scholarship, which cannot persuasively explain what the
clause means and why.
Doctrinally, the vastness of the Self-Incrimination Clause,
sprawling across the U.S. Reports into a great many doctrinal cor-
ners and crevices, makes exposition difficult. It is hard to get one's
hands on the beast, and there is no self-evidently superior way of
organizing the multiheaded case law for exposition and analysis. In
this article we have chosen the technique - call it a gimmick, if you
like - of organizing our account around key words of the clause
itself.5 Each word can be used as a window onto a different cluster
of doctrinal difficulties. Although these clusters might at first seem
unconnected, by the end of this article we hope to pull them all
together - to show how each cluster of current problems can be
solved by the same elegant rereading of the Self-Incrimination
Clause.
In Part I of this article, we examine the global puzzle of the Self-
Incrimination Clause and the local confusion or perversion lurking
behind virtually every key word and phrase in the clause as now
construed. In Part II we elaborate our reading of the clause and
show how it clears up the local problems and solves the overall
puzzle.
I. THE PuzzLE
Under current interpretations, courts cripple innocent defend-
ants while the guilty wrap themselves in the clause and walk free.
Modem understandings of the clause deviate far from its early
5. For a similar organizational strategy, see Bruce A. Ackerman, Beyond Carolene Prod-
ucts, 98 HARV. L. RFv. 713 (1985).
American implementation, from plain meaning, and from common

sense. The catalogue of interpretive difficulties is long: the privi-
lege protects the wrong "person" - the guilty witness rather than
the innocent defendant; courts struggle with impractical definitions
of "compulsion"; what is excluded from "any criminal case" by
grants of immunity has fluctuated widely over the years; and relia-
ble physical fruit, such as a bloody knife or a dead body, is now
excluded because it is "witnessing" against oneself, while other reli-
able physical evidence (the defendant's own blood, for example) is
allowed because it is not testimonial. In short, virtually every word
and phrase - person, compelled, in any criminal case, and witness
- sits atop considerable confusion or perversion because courts do
not yet understand how the words fit together, or what big idea(s)
might underlie the clause.
A. "Person"?
Perhaps the most striking flaw in current application of the priv-

ilege is that in some important cases defendants are not helped but
harmed. And worse: the privilege seems perversely designed to aid
the guilty defendant while punishing the innocent one. Indeed, an
overbroad reading of the privilege ends up undercutting the most
basic of all criminal procedure rights - the right of an innocent
defendant to mount a truthful defense.
Begin with the following question: Does the Self-Incrimination
Clause prevent a "person"in a criminal case from being compelled
to testify against himself even when that person is not on trial but
only a witness? Today the answer is yes. As a result, the Fifth
Amendment prevents an innocent defendant from compelling self-
incriminating testimony from a guilty witness. But this invocation
of the Fifth Amendment by the witness denies the accused her ex-
plicit Sixth Amendment right "to have compulsory process for ob-
taining witnesses in [her] favor."'6 The Fifth and Sixth Amendments
seem to be at war with one another, and the Fifth Amendment
rights of a witness apparently trump the Sixth Amendment rights of
the defendant, even though the defendant is of course the "person"
whose criminal case it is.7
6. U.S. CoNsr. amend. VI (emphasis added).

7. The quandary of an innocent defendant who wishes to show that someone else com-
mitted the crime is highlighted by Professor Peter Tague. See Peter W. Tague, The Fifth
AmendmenL If an Aid to the Guilty Defendan4 an Impediment to the Innocent One 78 GEo.
L.J. 1 (1989).
Let's consider a simple situation - a single defendant truthfully

claiming that she is innocent and that another specific person com-
mitted the crime - and trace the sequence of events that would
most likely follow in a criminal trial.8 Defendant subpoenas the
guilty party to testify at her trial, in accordance with her explicit
Sixth Amendment right to "comp[el]" the production of "witnesses
in [her] favor." The guilty witness notifies defense counsel before
trial that he will refuse to testify, invoking his Fifth Amendment
privilege not to be "compelled" to be a "witness" against himself.
After defense counsel informs the judge of the witness's intent, the
judge conducts a hearing to determine whether the witness may as-
sert the privilege. At the hearing, the witness refuses to answer
questions. This refusal, combined with other evidence the defense
has discovered suggesting the witness's guilt, convinces the judge
that the witness might incriminate himself were he to testify truth-
fully. The court therefore holds that the witness may assert the
privilege.
Faced with the witness's silence, the defendant begs the prosecu-
tion to grant the witness immunity to compel him to testify.9 Note
the imbalance of power here. While the prosecution can freely
grant immunity and compel witnesses to testify, the defense has no
such power. At a minimum, one might have thought the defend-
ant's explicit Sixth Amendment right of compulsory process should
put her on a level playing field with the prosecution; if the prosecu-
tor is empowered to subpoena a reliable witness, the defendant
should be so empowered. 10 Prima facie, it might seem that the
8. This account is drawn from Peter Tague. See iL at 1-3.
9. As will be explained in more detail infra in section I.C, the type of immunity now
required is "use plus use-fruits" immunity, which prevents the prosecution from using either
the words of the testimony or any evidence (fruits) found as a result of the testimony. See
Kastigar v. United States, 406 U.S. 441, 453 (1972).
10. See Webb v. Texas, 409 U.S. 95, 96, 98 (1972); Washington v. Texas, 388 U.S. 14, 22
(1967); 388 U.S. at 24-25 (Harlan, J., concurring). New Jersey's state constitutional precursor
of the federal compulsory process clause explicitly gave defendants "the same privileges of
witnesses ... as their prosecutors are or shall be entitled to." NJ. CONST. art. XVI (1776)
(emphasis added). This provision traces back to William Penn's 1701 Pennsylvania Charter
of Privileges: "THAT all criminals shall have the same Privileges of Witnesses... as their
Prosecutors." PA. CHARTER, art. V (1701) (emphasis added). Similarly, Sir William Black-
stone's widely influential treatise defined the compulsory process principle as giving the de-
fendant "the same compulsive process to bring in his witnesses for him, as was usual to
compel their appearance against him." 4 Wu.m BLAcKSroNE,COMMENTARmS '352 (em-
phasis altered). Blackstone's formulation in turn built on the landmark Treason Act of 1696,
7 Will. 3, ch. 3, § 7 (Eng.) (emphasis added). See infra note 11. This Act gave defendants
"the like process.., to compel their witnesses... as is usually granted to compel witnesses to
appear against them." 7 Will. 3, ch. 3, § 7 (1696) (Eng.) (emphasis added). James Madison's
particular "compulsory process" phraseology in the Sixth Amendment appears to borrow
directly from Blackstone. See Peter Westen, The Compulsory Process Clause, 73 MIcH. L.
REv.71, 97-98 & n.114 (1974). On compulsory process parity more generally, see id. at 78,
same should hold true for the immunity-granting power. But in-
stead, the defendant is forced to rely on the good graces of the pros-
ecutor.11 The prosecution, however, has little incentive to grant
immunity in this situation. The prosecutor reasons as follows: 12 If
the witness confesses falsely and the defendant is acquitted, a guilty
party gets off;' 3 and if the witness confesses truthfully, the immunity
itself will make it hard to ever prosecute and convict him - and so,
here too, a guilty party gets off.' 4 The prosecutor therefore refuses
the defendant's request that the government grant the witness im-
munity. Frustrated, the defendant asks that the court require the
prosecution to grant immunity. The vast majority of courts would
reject such a request,' 5 and the reasons here are obvious. Immunity
can create a major stumbling block to prosecuting the witness on
95,116,128,140 n.331,147-48,158-59,168, and 177-82. General privileges - spousal, priest-
penitent, and so on - which put limits on the government as well as the defendant, raise
different issues and thus lie beyond the scope of our discussion.
11. Our legal forebears attempted to correct this sort of imbalance as early as the 1690s,
when Parliament enacted the Treason Act. This landmark Act granted the defendant in trea-
son cases many of the same powers the prosecution had, including the right to legal represen-
tation and compulsory process to obtain witnesses. See Langbein, supra note 2, at 1056,1067-
68. In the midst of this seventeenth-century striving for parity between prosecution and de-
fense, grants of immunity were not included because the privilege against self-incrimination
did not exist. See id. at 1074-84.
12. Although we have stipulated that the defendant is in fact innocent, the prosecutor, of
course, may not know or believe this.
13. A peijury prosecution against the lying witness is of course possible, but perjury can
be hard to prove beyond a reasonable doubt, and a perjury prosecution may be less impor-
tant than the original case from the prosecutor's perspective.
14. Tague thinks that the prosecution "has no substantive reason" to refuse to grant im-
munity to the witness in single-culprit cases. See Tague, supra note 7, at 37, 53. The reasons
seem obvious enough.
15. One of the only exceptions appears to be the Third Circuit, which has held that a
court can order a prosecutor to grant immunity and also that a court can grant the immunity
itself in certain situations. See United States v. Herman, 589 F.2d 1191, 1204 (3d Cir. 1978)
(holding that a court can order a prosecutor to grant immunity to a defense witness when the
prosecutor has withheld the grant for the purpose "of distorting the judicial fact finding pro-
cess"), cert. denied, 441 U.S. 913 (1979); Government of the Virgin Islands v. Smith, 615 F.2d
964, 969 (3d Cir. 1980) (holding that a court can grant immunity itself if a defendant is other-
wise "prevented from presenting exculpatory evidence which is crucial to his case"). Most
courts, however, have refused to grant immunity even if the testimony is crucial to a defend-
ant's case. See, eg., United States v. Heldt, 668 F.2d 1238, 1282-83 (D.C. Cir. 1981) (holding
that a trial court should not grant immunity to defense witnesses who are actual or potential
targets of prosecution).
For academic commentary on the subject, see Westen, supra note 10, at 166-70 (arguing
that courts have the constitutional ability to grant immunity or to force prosecutors to grant
immunity) and Peter Westen, Incredible Dilemmas: ConditioningOne ConstitutionalRight
On the Forfeiture of Another, 66 IowA L. Rlv. 741, 762-75 (1981) [hereinafter Westen, In-
credible Dilemmas] (providing more elaboration of this view). See also James F. Flanaghan,
Compelled Immunity for Defense Witnesses: Hidden Costs and Questions, 56 NoTRE DAME
L. REv. 447, 461-63 (1981) (arguing against judicial grants of immunity for defense witnesses
in multiculprit crimes on the practical grounds that in the eventual prosecution of the witness,
the prosecutor would have difficulty proving that she gathered evidence without the aid of
the witness's testimony).
anything related to his testimony. (The government will have to

prove that nothing in its case is in any way derived from the testi-
mony, and as we shall see in more detail below, this can be a very
hard thing to prove.) Indeed, in organized crime settings the first
defendant will be tempted to give all his partners an immunity bath.
To add insult to injury, the jury would not even hear, from the
witness or anyone else, that the witness claimed the privilege. At
trial, the defendant could introduce evidence against the witness
16
but could not force him to invoke the privilege in front of the jury.
At most the defendant could request an instruction that the jury
must not draw any inference against either side from the witness's
failure to appear. 17 On the other hand, merely forcing the witness
to assert the privilege before the jury might make collusion between
defendant and witness easier.' 8 In general, the jury would be left
with considerable uncertainty about why the witness is asserting the
privilege; jurors could learn far more from actual testimony than
from a mere assertion of the privilege.
Of course, even if immunity were conferred, the guilty witness
might take the stand and lie, denying all involvement. But the wit-
ness need not break down and deliver a full-blown confession Perry
Mason style to make all the difference for an innocent defendant.
A lying witness may well sound unconvincing or trip himself up
with inconsistent testimony; he will also be subject to impeachment
via cross-examination and defense introduction of other evidence
and witnesses. The jury would then be perfectly poised to assess
witness credibility and to resolve factual disputes. This, of course, is
what we pay jurors to do.
So in the end we are left with the following puzzle. Precisely
because Fifth Amendment immunity is so strong, innocent defend-
ants are crippled in their effort to mount a Sixth Amendment de-
fense. Might the solution be to rethink Fifth Amendment
immunity?19
16. Tague notes that every U.S. court that has considered the issue has so held. See
Tague, supra note 7, at 5. For the reasons behind this view, see id. at 13-52.
17. Id. at 2.
18. But see icL at 40-43 (discounting this concern).
19. Cf.Amar, supra note **,at 783 ("[U]njustified expansions of constitutional rights
often lead to dangerous and unjustified contractions elsewhere."). For a good illustration of
this principle in the Fifth Amendment context, see Allen v. Illinois, 478 U.S. 364, 369-75
(1986) (holding that proceedings under one state's Sexually Dangerous Persons Act were not
"criminal" despite the potential for moral stigma and incarceration in a maximum security
institution). The Allen Court's characterization of the case as "noncriminal" eliminated the
need to apply self-incrimination principles, which the Court thought undermined the reliabil-
ity of fact-finding. 478 U.S. at 375. As we shall see, however, the Self-Incrimination Clause,
rightly understood, is not at war with reliability. See infra Part II. Had the Allen Court
March 1995] Fifth Amendment FirstPrinciples
B. "Compelled"?
The Fifth Amendment does not prohibit all self-incrimination
but only compelled self-incrimination. How should we go about
construing the word compelled? Some leading Framers thought of
the Self-Incrimination Clause as a protection against torture, which
might often lead to unreliable confessions. 20 Our main concerns to-
day should still be protecting against third-degree tactics and assur-
ing reliability in evidence. Instead, at times the Justices of the
Supreme Court have become engrossed by relatively trivial forms
of compulsion; at other times they have zigged and zagged errati-
cally; and at still other times they have turned a blind eye to danger-
ous compulsion threatening our core concerns.
1. In-Court Compulsion
At times, compulsion seems to be defined as making someone
who "takes the Fifth" worse off in any way than one who sings with-
out reservation. Under this definition, the government may not
draw any adverse inference if a person insists on standing mute in
the face of an accusation of criminal wrongdoing. In its controver-
sial 1965 opinion in Griffin v. California,21 the Supreme Court es-
properly construed the clause, it might have been more inclined to admit that the case before
it was indeed criminal. Instead, a "broad" reading of the Self-Incrimination Clause ended up
eliminating the defendant's explicit constitutional right to other "criminal" procedure safe-
guards outlined in the Sixth Amendment.
20. During the debates over ratification of the Federal Constitution, several participants
expressed fears that the Constitution failed to provide common law protection against tor-
ture to extract confessions. In Virginia, Patrick Henry warned that
Congress may introduce the practice of the civil law, in preference to that of the com-
mon law. They may introduce the practice of France, Spain, and Germany - of tortur-
ing, to extort a confession of the crime.... [T]hey will tell you that there is such a
necessity of strengthening the arm of government, that they must have a criminal equity,
and extort confession by torture, in order to punish with still more relentless severity.
3 THE DEBATES IN THE SEVERAL STATE CoNVENTIONS ON THE ADOPTION OF THE FEDERAL
CoNsTiTUTION 447-48 (Jonathan Elliot ed., 1886). George Nicholas retorted that Henry's
argument about torture applied equally to the Virginia Constitution. See 3 idat 450-51. At
this point, George Mason, the drafter of Virginia's Bill of Rights, jumped into the fray and
argued that his state's Bill of Rights did prohibit torture: "[O]ne clause expressly provided
that no man can give evidence against himself, and ...[Nicholas] must know that, in those
countries where torture is used, evidence was extorted from the criminal himself." 3 id. at
452; see also 2 id. at 111 (recording remarks by Abraham Holmes in the Massachusetts de-
bates linking torture and the Inquisition with the federal government's ability to compel a
man to furnish evidence against himself). Leonard Levy has noted that references to the
privilege are scarce in the literature and debates surrounding the ratification of the Constitu-
tion and the Bill of Rights. See LEONARD W. .EVY, ORiiams OF THm FIrH AMENDMEN.
THm Rxicrr AorANsT SELF-INCRIMINATION 430 (1968). The Framers occasionally acknowl-
edged that the privilege served as a ban on torture, but "nothing can be found of a theoretical
nature expressing [an additional] rationale or underlying policy for the right in question or its
reach." Id.
21. 380 U.S. 609 (1965).
tablished this as the test of in-court compulsion. A defendant had

an absolute right not to take the stand, and no adverse inference of
guilt could be drawn from the exercise of this trial right. Any infer-
ence, said the Court, was tantamount to criminal punishment of
22
trial silence itself, a core violation of the clause's command.
But this begs the question: if the adverse inference was indeed
probabilistically sound, was it truly punishing silence, or the (more
probable) guilt signalled by the silence? On its facts, however, the
in-court rule may well make sense. As we shall see in more detail
later, many innocent defendants may decline to take the stand for 23
reasons that have little to do with their guilt in the case at hand,
and so an adverse inference might be statistically unsound and un-
reliable2 and therefore may punish silence itself, rather than under-
lying guilt on the predicate offense. Under the reliability theory we
shall outline below, Griffin can stand.
The Supreme Court, however, has failed to make clear how
Griffin should apply when a criminal court acts in the sentencing
phase rather than the guilt-determination stage of a criminal case
and seeks to draw inferences from a defendant's earlier (pre-
verdict) silence. Even if adverse inferences are unreliable to prove
guilt, perhaps they may be reliably used in the sentencing process,
after guilt has already been reliably established. The Federal Sen-
tencing Guidelines, for example, authorize a lower sentence for a
defendant who "accept[s] ...responsibility for his offense" as evi-
denced by, among other things, "truthfully admitting the conduct
comprising the offense(s) of conviction" or "assist[ing] authorities
in the investigation or prosecution of his own misconduct by taking
22. See 380 U.S. at 614-15.

23. The innocent defendant may want to avoid taking the stand because he is likely to
perform badly, being inarticulate and concerned that an experienced prosecutor, skilled in
the artificial rules governing courtrooms, will be able to trip him up. The jury, our innocent
defendant might also fear, will likely overreact to any real or perceived slip up on his part on
the stand. In addition, he may be worried that his prior convictions will come in to impeach
his credibility, wrongly inducing the jury to think that he must also be guilty in the case at
hand. The latter problem may be cured by a provision, such as MoNT.R. Evm. 609, that
states that prior convictions are not available to attack credibility. The former problem, how-
ever, cannot be cured so easily.
24. See infra section II.E; see also Stephen J. Schulhofer, Some Kind Words for the Privi-
lege Against Self-Incrimination,26 VAt.U. L. REv. 311, 330-35 (1991) (defending Griffin on
similar grounds). So too, even if silence were mildly probative, a typical jury might draw far
too large an adverse inference, and this also (by hypothesis) would be statistically unsound.
In this situation, Griffin's rule overprotects - by disallowing instructions to the jury that an
inference can be made - but a more carefully tailored instruction to "infer, but not too
much" might be unworkable.
one or both of the steps set forth in subsection (b).' '25 One of the
steps in subsection (b) involves the "timely provision of complete
information to the government concerning his own involvement in
the offense." 26 The common-sense logic and morality of these
guidelines are easy to see: those who repent and admit their mis-
deeds in open court before conviction may need less rehabilitation
and deserve less punishment in the sentencing process.
But can this everyday morality be squared with the logic of the
Self-Incrimination Clause and Griffin? Are courts in effect punish-
ing silence - and in a criminal case, no less? On the other hand, if
the Sentencing Guidelines can be upheld, and Griffin limited, on
the basis of a guilt-determination/sentencing distinction, what
would such a distinction imply about the meaning and purpose of
the Self-Incrimination Clause? (Wouldn't it, for example, suggest
that the key to the clause is reliability?) Finally, if a defendant ap-
peals, wins, and is retried, can he exclude, in the guilt-determination
stage of his second trial, any earlier statements on the grounds that
they were compelled by the "silence penalty" imposed by the Sen-
tencing Guidelines? Lower courts need to know which silence pen-
alties are valid under the Self-Incrimination Clause, and the
Supreme Court has not yet announced a theory of the clause that
can inform them.27
25. U.S. SENTENCING CoMMN., FEDERAL SENTENCING GuiDEUn4_S MANUAL § 3.E1.1

(1993 ed.).
26. Id.
27. The Supreme Court denied certiorari in a recent case raising the issue of whether the
sentencing reduction can be conditioned on defendant's admitting to uncharged conduct.
Kinder v. United States, 112 S.CL 2290 (1992). The defendant agreed to a plea bargain but
then was denied the sentencing reduction when he refused to admit to conduct to which he
had not pled guilty. As Justice white pointed out in his dissent to the denial of certiorari,
there is a circuit split on this question. See 112 S.Ct. at 2293 (White, J., dissenting).
In contrast, courts seem to agree that the sentencing reduction can be conditioned on the
defendant's admitting to conduct for which he has been convicted. See, for example, United
States v. Henry, 883 F.2d 1010 (11th Cir. 1989):
"[T]he guideline recognizes societal interest in... the increased potential for rehabilita-
tion among those who feel and show true remorse for their anti-social conduct."
... To hold the acceptance of responsibility provision unconstitutional would be to
say that defendants who express genuine remorse for their actions can never be re-
warded at sentencing. This the Constitution does not require.
883 F.2d at 1011-12 (quoting United States v. Belgard, 694 F. Supp. 1488, 1497-98 (D. Or.
1988)). But if we follow the seeming logic of Griffin's "no worse off" test, such a condition
would also violate the privilege, and no court has explained how Griffin is to be distinguished
away.
It might be claimed that intervening cases upholding plea bargaining have eroded Griffin:
a defendant who pleads not guilty may be penalized more severely than - made worse off
compared to - a defendant who pleads guilty. Cf.Brady v. United States, 397 U.S. 742,749-
54 (1970). Formally, however, pleading and witnessing are not the same thing; a plea bargain
may benefit a defendant whether or not he would have taken the stand as a witness at trial.
2. Out-of-Court Compulsion: Government Employment

In a series of cases involving public employees, the Supreme
Court tried to stretch the in-court Griffin rule to govern conduct
well outside the courtroom - indeed, outside the criminal justice
system.28 The cases held that the threat of removal from govern-
ment employment was compulsion for the purposes of the Fifth
Amendment, incorporated through the Fourteenth; they concerned
individuals such as allegedly corrupt police officers and dishonest
state contractors who successfully claimed a right to keep their
government jobs while refusing to respond to allegations of
29
wrongdoing.
But outside the courtroom, the "no worse off" test seems ex-
travagant and unworkable: the logical consequences are absurd.
Couldn't President Reagan have refused to appoint Oliver North to
be Secretary of State - or fired him from his subcabinet White
House office, for that matter - on the simple ground that North's
invocation of the Fifth Amendment raised sufficient doubt about
his worthiness to serve in a position of high honor and power?
Merely focusing preliminary investigation on someone who takes
the Fifth also makes that person worse off, and yet - so far - the
courts have allowed the government to do just that. 30 Even if
within a courtroom a jury should be barred from drawing adverse
inferences from trial silence, the real world outside the courtroom
cannot be regulated by so nice a test of compulsion; people in ev-
eryday life must draw reasonable inferences.31 A responsible pri-
vate employer may draw sensible inferences from silence and fire
32
the bank clerk who refuses to respond to accusations of pilfering.
Thus, the plea-bargaining cases may not resolve the arguably distinct witnessing issues raised
by the "silence penalty" in sentencing.
28. As we shall see in more detail, see infra section I.C, the Fifth Amendment protects
against compelling statements outside a "criminal case" if those statements are later usable
inside a criminal case - at a criminal trial.
29. See Lefkowitz v. Tbrley, 414 U.S. 70 (1973) (holding that a state could not cancel
existing contracts of and deny future contracts to a contractor who refused to testify without
immunity concerning state contracts); Gardner v. Broderick, 392 U.S. 273 (1968) (holding
that a state could not discharge a police officer who refused to testify without immunity
before a grand jury about the performance of his official duties); Uniformed Sanitation Men
Assn. v. Commissioner of Sanitation, 392 U.S. 280 (1968) (announcing a similar holding in a
companion case concerning sanitation workers).
30. See S-mrEN A. SALzauRGo & DANIEL J. CAPRA, AMERICAN CRIMINAL PROCEDURE
456 (4th ed. 1992).
31. Some of the reasons that an innocent person might stand mute in a technical legal
proceeding governed by artificial and unfamiliar rules would not apply to informal accusation
in some noncourt settings. See infra section II.E.
32. Even Dean Griswold acknowledged the appropriateness of a private employer's fir-
ing someone who refused to answer questions about his job. See ERwIN N. GruswoLD, Tam
Yet the "no worse off" test forces the government to act as an irre-
sponsible employer - ignoring common sense - or else confer
immunity and possibly lose valuable evidence in its capacity as sov-
ereign enforcer of criminal law. (Once again, it loses this evidence
unless it can conclusively prove at trial that the evidence was in no
way connected to the earlier "compelled" words.3 3) Is it possible
that, here too, this immunity rule is too broad?
So it seems that the "no worse off/no adverse inference" test
simply will not work as a global rule of impermissible compulsion
outside the courtroom. Indeed, in the 1976 case of Baxter v.
Palmigiano,34 the Court seemed to reject the "no worse off" test
even for prison disciplinary hearings. The Court allowed adverse
inferences to be drawn in a disciplinaryhearing against a prisoner
who refused to respond to questions unless he was immunized.3 5
Once we reject the "no worse off" test, however, we flounder in a
sea of murky rules. How much pressure to talk is too much? And
how can we even begin to answer this question without a theory of
why pressure is bad?
3.
Out-of-Court Compulsion: Required Records
A prime example of the confusion that results when the Court
strays from the "no worse off" test is the required records doctrine.
The Court has been unable to develop a consistent rule for the use
in criminal proceedings of records that the government requires cit-
izens to keep for administrative purposes. 36 The cases essentially
set up an open-ended test that allows the Court to swing from ad-
FIrm AMENDMENT TODAY 57-58 (1955); see also Adamson v. California, 332 U.S. 46, 60
(1947) (Frankfurter, J., concurring) ("Sensible and just-minded men, in important affairs of
life, deem it significant that a man remains silent when confronted with serious and responsi-
ble evidence against himself which it is within his power to contradict.").
33. In essence, Fifth Amendment immunity raises "causation gap" issues analogous to
those raised by the Fourth Amendment exclusionary rule. For a discussion of this causation
gap, see Amar, supra note **,at 793-95.
34. 425 U.S. 308 (1976).
35. 425 U.S. at 318 ("[Palmigiano's] silence was given no more evidentiary value than was
warranted by the facts surrounding his case."). The Court felt that other constitutional pro-
tections, such as the Sixth Amendment right to counsel, were not required in the prison
disciplinary context. See 425 U.S. at 315. In civil cases, we routinely allow adverse inferences
to be drawn when a witness "takes the Fifth." But, it might be argued, in civil cases it is
generally not the government that benefits from the inference; the beneficiary is the oppos-
ing private party, and the government merely adjudicates. By that logic, adverse inferences
from silence in civil cases could be drawn in favor of private parties, but not in favor of the
government when it is a party in a civil suit. Like the cases involving the government in its
capacity as an employer, this puts the government in an unjustifiably weaker position than
comparable private actors.
36. Again, a Fifth Amendment violation occurs at the point when compelled testimony is
introduced in a criminal case. See supra note 28; infra note 63.
mitting the records to excluding both the records and their fruits
without any principled basis.
The 1948 case that spawned the doctrine, Shapiro v. United
States,37 arose out of emergency wartime regulations and seemed to
place few limits on the required records exception to the Fifth
Amendment. Shapiro, a produce wholesaler during World War II,
was required to keep price, sale, and delivery records and to make
them available for inspection under the Emergency Price Control
Act.3 8 Shapiro complied with a subpoena to produce his records
but claimed his constitutional privilege. The government, however,
proceeded to prosecute him for illegal tie-in sales, using the records
and their fruits. A closely divided Supreme Court held that the
Self-Incrimination Clause did not apply to the subpoenaed docu-
ments; because Shapiro was required by law to keep the records,
they were unprotected public documents rather than protected pri-
vate papers. 39 In his dissent, Justice Frankfurter pointed out some
of the deep flaws in the Court's reasoning:
Subtle question-begging is nevertheless question-begging. Thus:
records required to be kept by law are public records; public records
are non-privileged; required records are non-privileged. If records
merely because required to be kept by law ipso facto become public
records [and therefore fall outside the scope of the privilege], we are
indeed living in glass houses. 4°
We must keep in mind that the rule of immunity at that time
was laid down by Counselman4l - complete immunity from prose-
cution for the crime. This no doubt encouraged the Shapiro Court's
failure to define any limits to the "public" records Congress might
require, except those limits imposed by its enumerated powers.
Government had legitimate regulatory need of the information, and
the Court was unwilling to force the government to vindicate its
legitimate regulatory needs only by abdicating - via sweeping im-
munity - its role as criminal law enforcer. As in the employment
37. 335 U.S. 1 (1948).

38. Emergency Price Control Act of 1942, ch. 26, 56 Stat. 23 (repealed 1966).
39. 335 U.S. at 32-33. Chief Justice Vinson, possibly influenced by the wartime, emer-
gency nature of the regulations, wrote for the Court:
[N]o serious misgiving that [the bounds imposed by the Fifth Amendment] have been
overstepped would appear to be evoked when there is a sufficient relation between the
activity sought to be regulated and the public concern so that the Government can con-
stitutionally regulate or forbid the basic activity concerned, and can constitutionally re-
quire the keeping of particular records, subject to inspection by the Administrator.
335 U.S. at 32. He then noted that Congress unquestionably had the constitutional power to
control commodity prices as a wartime emergency measure. 335 U.S. at 32.
40. 335 U.S. at 51 (Frankfurter, J., dissenting).
41. Counselman v. Hitchcock, 142 U.S. 547 (1892).
context, is it possible that the real problem here is that immunity

has been defined too broadly?
In later cases, the Court has struggled to establish limits to the
required records exception, but the limits have remained fuzzy. In
1968, the Court decided three required documents cases on the
same day, all authored by Justice Harlan and all reversing convic-
tions because of violations of the Fifth Amendment. The first,
Marchetti v. United States,42 controlled the rest.43 Marchetti had
been convicted of willful failure to pay the occupational tax associ-
ated with gambling and of willful failure to register as a wagerer.
Justice Harlan cobbled together bits from earlier cases into a three-
pronged pseudo-test to distinguish Shapiro, but the test failed to
undo the confusion produced by earlier cases. Under the test,
courts were to inquire, first, whether the defendant was obliged to
keep records "of the same kind as he has customarily kept;""4 sec-
ond, whether there were "public aspects" involved in the records
required;45 and third, whether a records requirement was "imposed
in 'an essentially non-criminal and regulatory area of inquiry"' or
was instead "directed to a 'selective group inherently suspect of
criminal activities.' "46 The prongs were intriguing but not well de-
fined or well justified.
Although the Court found Fifth Amendment violations in the
1968 Marchettitrilogy, in 1971 the Court lurched the other way and
upheld the constitutionality of California's hit-and-run statute in
Californiav. Byers.47 Byers had been convicted of failing to stop at
the scene of an accident and failing to leave his name and address.
The California Supreme Court required a use restriction on the in-
formation,48 but the U.S. Supreme Court held that because the stat-
ute did not violate the Fifth Amendment, no use restriction was
required.49 Chief Justice Burger, writing for a plurality, struggled to
find a rationale for holding that the privilege did not apply. He
invoked some elements of the Marchetti test, but the opinion else-
where veered away from the test. At one point, the Chief Justice
42. 390 U.S. 39 (1968).

43. The other two cases were Grosso v. United States, 390 U.S. 62 (1968) (reversing a
conviction for failure to pay an excise tax on wagers) and Haynes v. United States, 390 U.S.
85 (1968) (reversing a conviction for possession of an unregistered firearm).
44. 390 U.S. at 57.
45. 390 U.S. at 57.
46. 390 U.S. at 57.
47. 402 U.S. 424 (1971).
48. See Byers v. Justice Court, 458 P.2d 465, 477 (1969).
49. 402 U.S. at 427 n.3.
seemed to announce a new principle in self-incrimination cases, and

one completely lacking in textual support: questions about whether
the privilege should apply "must be resolved in terms of balancing
the public need on the one hand, and the individual claim to consti-
tutional protections on the other. '50
The Court further stretched the required records doctrine in the
1990 Bouknight case,51 in which Justice O'Connor relied heavily on
the exception to prevent the privilege from applying in a disturbing
case of child abuse. In Bouknight, a mother suspected of child
abuse was given custody of her injured child with extensive condi-
tions imposed by a protective order. The mother violated those
conditions, and a court ordered her to produce the child in order to
verify that the child was alive and well. When she refused, the court
held her in contempt and rejected her contention that the Fifth
Amendment protected her from having to produce him. The
Supreme Court invoked Shapiro, Marchetti, and Byers, explaining
that the requirement to produce the child involved a "broadly di-
rected, noncriminal regulatory regime"5 2 - that the state was at-
tempting to help "a child in need of assistance"53 and was not solely
concerned with criminal law enforcement.5 4
The inconsistency of these cases is striking and revealing. The
Court hems and haws and then often holds that the privilege does
not apply at all: the government often needs information for non-
penal purposes and should not be forced to let criminals go free to
get it, the Court intuits. The Court is understandably reluctant to
apply the privilege in a heinous crime such as child abuse; granting
use plus use-fruits immunity would make it difficult, and in some
cases (including the hit-and-run) almost impossible, to prosecute.
But that is what Kastigar55 currently demands. Unable to live with
that result, the Court zigs, zags, and balances, ad hoc. But the lan-
guage of the Self-Incrimination Clause does not balance: it states a
bright-line rule. Is it possible that if immunity were narrower than
50. 402 U.S. at 427 (emphasis added).

51. Baltimore City Dept. of Social Servs. v. Bouknight, 493 U.S. 549 (1990).
52. 493 U.S. at 559.
53. 493 U.S. at 559.
54. The Court, however, did not categorically state that Bouknight would not be allowed
any immunity: "We are not called upon to define the precise limitations that may exist upon
the State's ability to use the testimonial aspects of Bouknight's act of production in subse-
quent criminal proceedings." 493 U.S. at 561. As we shall see below, this caveat in effect
sidesteps the real issue in the case and under the clause generally: the scope of the immunity
that must be given in a criminal case.
55. Kastigar v. United States, 406 U.S. 441 (1972).
Kastigar indicates, judges could indeed live with the logic of the
bright-line rule?
4. Out-of-Court Compulsion: Police Interrogation

When it comes to genuine out-of-court coercion, courts at times
turn a blind eye. While Miranda56 purported to establish propriety
in police-station interrogation, our system in fact can still be quite
ugly. Despite Miranda'spromise to open up the black box of the
police station, it did not require that lawyers, magistrates, or even
tape or video recorders be present in interrogation rooms. In the
absence of these monitors, detectives and police have often en-
gaged in ingenious, but troubling, forms of interrogation.
In some instances, courts actually succeed in chilling clever tac-
tics, with a resulting loss of information for the prosecution, but in
other cases the judicial decisions have created a divergence be-
tween theory and practice. A good recent description of this diver-
gence - and of suspect interrogation in general - is found in
David Simon's 1991 book based on unlimited access to the Balti-
more homicide detective unit over the course of a year. 57 Some of
the techniques used by detectives could aptly be described as physi-
cally intimidating, teetering on the brink of violence. Simon de-
scribes detectives doing everything from slamming doors and
kicking chairs, swearing, and interrogating late at night to lying
about the evidence against the suspect and about the detective's
ability and willingness to get the suspect a lower sentence.5 8 Sev-
eral of the techniques are ingenious, most notably "polygraph by
photocopier. '59 But most people would find them uncivilized, at
56. Miranda v. Arizona, 384 U.S. 436 (1966).
57. See DAVm SmoN, HoMIc=E: A YEAR ON THE KILLING STrnErS 199-220 (1991).
Simon is a reporter for the Baltimore Sun who spent four years on the police beat before his
research leading to Homicide. Simon concludes, "[I]f the ... intent of the Miranda decision
was, in fact, an attempt to 'dispel the compelling atmosphere' of an interrogation, then it
failed miserably." Id. at 199.
58. Id. at 204-20.
59. Recently, several experienced homicide detectives in Detroit were publicly criticized
and disciplined by their superiors for using the office copy machine in purported lie detector
tests. The process worked as follows:
[T]he detectives, when confronted with a statement of dubious veracity, would some-
times adjourn to the Xerox room and load three sheets of paper into the feeder.
"Truth," said the first.
"Truth," said the second.
"Lie," said the third.
Then the suspect would be led into the room and told to put his hand against the side
of the machine. The detectives would ask the man's name, listen to the answer, then hit
the copy button.
Truth.
And where do you live?
best. We do not tolerate such behavior in civil litigation.60 Rather,

we use depositions - with court reporters, lawyers, and judicial
oversight.
Here, too, we are left with a puzzle. Perhaps the police are
tempted to subvert formal rules because the rules make no sense.
A suspect, even if ultimately innocent, is often someone close to the
action, someone who can tell the government what really happened
so that the case can be solved. 61 But formal doctrine prevents the
government from using the suspect as an involuntary testimonial
resource unless the government confers a sweeping immunity that
will often preclude prosecution. By effectively preventing formal,
civilized depositions - by creating an overly strong Fifth Amend-
ment immunity - perhaps we are driving interrogation under-
ground, in ways that make some citizens more vulnerable and the
weakest most vulnerable of all.62 So could it be that here, again, the
most sensible solution would be to rethink immunity?
C. "In Any Criminal Case"?

The government compels persons to be witnesses to their own
criminal conduct all the time - in strict conformity with the Con-
stitution, we are told. All that is required is immunity of a certain
sort, enforced by rules of exclusion "in any criminal case" against
the witness. 63 But immunity of what sort? This is the hidden key to
Truth again.
And did you or did you not kill Tater, shooting him down like a dog in the 1200 block
of North Durham Street?
Lie. Well, well ....
Id. Occasionally, a confession would result.
60. See William J. Stuntz, Lawyers, Deception, and Evidence Gathering,79 VA. L. REV.
1903, 1905 (1993) ("Deception and advantage taking are.., at the core of criminal investiga-
tion, even though legal ethics doctrine largely banishes them from the evidence-gathering
process in civil cases."). Of course, we are not proposing to abolish all governmental "decep-
tion" - sting operations, undercover agents, and the like. We largely agree with Professor
Stuntz that criminal law enforcement should not be constrained by all the rules applicable to
civil litigation. But we will argue that certain more "civilized" techniques, like depositions,
should be available to criminal law enforcement officers. See infra Part II.
61. Langbein, supra note 2, at 1055.
62. Cf. Paul G. Kauper, JudicialExaminationof the Accused - A Remedy for the Third
Degree, 30 MIcH. L. REv. 1224 (1932); Yale Kamisar, Kauper's "JudicialExamination of the
Accused" Forty Years Later - Some Comments on a Remarkable Article, 73 MiCH. L. Rv.
15 (1974).
63. Some have argued that the Fifth Amendment cannot mean what it says - that, con-
trary to its words, it must apply to compulsion outside a criminal case. Surely (the argument
goes), we cannot allow a prosecutor to recess a criminal trial; walk across the street and
compel - upon pain of contempt - a defendant to answer questions; and then walk back
across the street, reconvene the trial, and introduce as evidence both a transcript and a video-
tape of the compelled statement. What is true of a compelled deposition across the street in
the middle of the trial must be true of compulsion in pretrial legislative hearings, civil cases,
the Fifth Amendment.6a Exactly what must be excluded from "a

criminal case"? Over the years, courts have been all over the map.
As we shall see in more detail later, the earliest American courts
addressing the immunity issue required only a narrow form of im-
munity. 65 But the U.S. Supreme Court brushed these cases off in its
6
1892 decision in Counselman v. Hitchcock.6 Counselman estab-
grand jury inquests, and so on. (Mirandasays that the same should hold for informal com-
pulsion - backed by threatened police brutality rather than formal contempt - in the po-
lice station.)
All this is true - except the notion that we have somehow gone beyond the words of the
Fifth Amendment. But see Henry J.Friendly, The Fifth Amendment Tomorrow: The Case
for ConstitutionalChange, 37 U. Cni. L. Rnv. 671, 677 (1968). In all our examples, compul-
sion may not exist within the criminal case; but the introduction of the compelled statement
- the witnessing - does occur in a criminal case, and it is this introduction that violates the
Fifth Amendment. This is why out-of-court compelled testimony accompanied by formal or
informal immunity simply does not violate the amendment: unless the compelled statement'
is introduced at a criminal trial, a person has not been made a "witness" (via transcript and
videotape) against himself "in" a "criminal case." Although courts and commentators have
often been confused - and confusing - on this point, the foregoing appears to be the view
of the current Court. See Larry J.Ritchie, Compulsion That Violates the Fifth Amendment:
The Burger Court's Definition, 61 MnqN. L. Rav. 383,386,430 (1977); cf Michigan v. Tcker,
417 U.S. 433, 440-41 (1971) (declaring that to allow the in-court introduction of testimony
created by out-of-court coercion would "practically nullif[y]" the Fifth Amendment privi-
lege); Piemonte v. United States, 367 U.S. 556 (1961) (saying that once proper immunity is
offered, no Fifth Amendment violation occurs when a person is forced, upon penalty of con-
tempt, to incriminate himself outside his own criminal case); Brown v. Walker, 161 U.S. 591
(1896) (announcing a similar holding).
Put another way, deposition-like compulsion outside the courtroom is not bad; we use it
all the time in civil cases. What is bad is using compelled testimony inside a criminal case -
for reasons that have to do with the presumptive unreliability of certain types of compelled
testimony. See infra section II.E. By contrast, unregulated police-station coercion is often
bad in itself and calls for special rules. See infra section II.B.4.
64. In effect, the ability of the government to demand self-incriminating statements in
legislative hearings and the like after offering the proper immunity means that - like its
Fifth Amendment companion, the Takings Clause - the Self-Incrimination Clause in some
ways states a liability rule, not a property rule. See generally Guido Calabresi & A. Douglas
Melamed, Property Rules, Liability Rules, and Inalienability: One View of the Cathedral,85
HARv.L. Rav. 1089 (1972). Once we see this, we should see the centrality of the scope of
immunity: it establishes the all-important fixed price at which the government may buy a
person's testimony outside his own criminal case.
Another more nuanced and textual way to see the point is as follows: The clause is abso-
lute, but strictly speaking it applies only to testimony in one's own criminal case. But be-
cause one must be able to "take the Fifth" outside criminal cases, see supra note 63, the
clause must be enforced by a rule of exclusion within a criminal case. In effect, we have a
kind of "liability rule" for "taking the Fifth" in grand jury rooms, civil cases, legislative hear-
ings, and so on - enforceable by a specific performance or "property rule" of exclusion
within the criminal case, an exclusion that the government may not "buy off" with money at a
judicially fixed price, as with a pure liability rule. But to allow a person outside his own
criminal case to ignore all requests to furnish self-incriminating testimony, even after proper
immunity has been given, would plainly ignore the textual bounds of the Amendment, which
is limited to criminalcases. In effect, it would give a person an absolute right to withhold all
testimony, anywhere, anytime, and yet be free from pressure or compulsion. Such a rule is
practically unworkable as well as textually implausible. See Piemonte, 367 U.S. at 556;
Brown, 161 U.S. at 591.
65. See infra section II.C.
66. 142 U.S. 547 (1892).
lished an extraordinarily sweeping form of immunity that came to

be known as "transactional" because it absolutely prohibited prose-
cution for any criminal transaction to which a question might re-
late. 67 In effect, Counselman prevented a suspect who had been
made to sing pretrial from being a witness against himself "in any
criminal case" by preventing him from being a defendant - by
preventing the case against him from ever going forward. Put an-
other way, Counselman excluded the prosecution itself "in" a
"criminal case." But where, textually, does that rule come from?
And would its logic require that if a person is made to sing after he
has been convicted of a crime, the conviction must be vacated?
Does that make sense? 68
Formally, Counselman reigned until 1972, but for much of its
rule, the so-called dual sovereignty doctrine created an important
chink in its armor: before 1964, testimony compelled from a person
by federal officials could be used against him in a state prosecution
and vice versa. 69 The Court so held because the Self-Incrimination
Clause, and therefore exclusion, applied not to the states but only
to the federal government. 70 With its 1964 decision in Malloy v.
Hogan,71 however, the Court incorporated the clause against the
states, rejecting the notion that the Fourteenth Amendment con-
tained "only a 'watered-down, subjective version of the individual
guarantees of the Bill of Rights.' "72 The same day it decided Mal-
loy, the Court reconsidered its rule of self-incrimination dual sover-
eignty. In light of Malloy, it made little sense that two governments
acting in tandem could do what neither could do alone: extort testi-
67. 142 U.S. at 586 ("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.").
68. Although courts might require the convict's compelled statement to harm him in
some way to satisfy the requirement that he be a witness "against himself," that requirement
might be met easily. Certain avenues of postconviction collateral review - habeas corpus
and writs of coram nobis - may in some situations be limited to the factually innocent, and a
compelled postconviction confession might undercut a defendant's ability to pursue these
postconviction avenues.
69. See Akhil Reed Amar & Jonathan L. Marcus, Double Jeopardy Law After Rodney
King, 95 COLUM. L. Rlv. 1, 11-15 (1995).
70. See Barron v. Baltimore, 32 U.S. (7 Pet.) 243, 247-49 (1833); see also Feldman v.
United States, 322 U.S. 487 (1944) (holding that a defendant's compelled testimony in a state
proceeding could be used in federal court to convict the defendant of a federal crime);
United States v. Murdock, 284 U.S. 141 (1931) (holding that the federal government need
only immunize a witness from federal prosecution and that fear of state prosecution and use
of federally compelled testimony will not excuse a refusal to answer).
71. 378 U.S. 1 (1964).
72. 378 U.S. at 10-11 (quoting Ohio ex rel. Eaton v. Price, 364 U.S. 263 (1960) (Brennan,
J., dissenting)).
mony out of court and then introduce it in a criminal case against

the testifier. Thus, in Malloy's companion case, Murphy v. Water-
front Commission,73 the Court declared that "there is no continuing
legal vitality to, or historical justification for, the rule that one juris-
diction... may compel a witness to give testimony which could be
used to convict him of a crime in another jurisdiction." 74 According
to Murphy, a state witness could not be compelled to give testimony
that might be incriminating under federal law unless the compelled
testimony and its fruits could not be used in any way by federal offi-
cials in a criminal prosecution against him.75
Although the Murphy Court cited Counselman, it actually intro-
duced a narrower standard of immunity that was elaborated in a
separate opinion by Justices White and Stewart.76 Counselman
transactional immunity would create huge intersovereign friction:
government A could confer immunity and thereby in effect veto
government B's law enforcement efforts. To reduce this friction,
immunity must be rethought, argued White and Stewart. If only a
narrower immunity were required, government B could prosecute
even after A's grant of immunity so long as B could prove that all
its evidence came from truly independent sources. 77 (Note, of
course, that the difficulty of proving this meant that intersovereign
friction would be reduced but not eliminated.78 )
Eight years after Murphy, its newfangled immunity rule was for-
mally approved outside the dual sovereignty context in Kastigar v.
United States.79 Unlike Counselman,Kastigarheld that the govern-
ment at times could indeed prosecute a person who had earlier
been obliged to testify against himself under a grant of immunity.80
Kastigarheld that in order to compel incriminating testimony from
a Witness, the government merely had to grant the witness immu-
nity from the use of his testimony and of its fruits (discoveries made
as a result of that testimony) in a criminal trial against him.81 This
became known as "use plus use-fruits immunity"' 2 and is, theoreti-
73. 378 U.S. 52 (1964).

74. 378 U.S. at 77.
75. See 378 U.S. at 78-79.
76. 378 U.S. at 92-107 (White, J., concurring).
77. 378 U.S. at 106-07 (White, J., concurring).
78. Here too, we see the "causation gap." See supra note 33.
79. 406 U.S. 441 (1972).
80. See 406 U.S. at 453.
81. See 406 U.S. at 456-59.
82. Although Kastigaruses the term derivative use to describe this type of immunity, use-
fruits is more graphic and is used by leading criminal procedure scholars as well as by many
cally, the standard still in operation today. The Court aimed to find
"a rational accommodation between the imperatives of the privi-
lege and the legitimate demands of government to compel citizens
to testify" 83 and stated that the new standard "leaves the witness
and the prosecutorial authorities in substantially the same position
as if the witness had claimed the Fifth Amendment privilege. The
immunity is therefore co-extensive with the privilege and suffices to
supplant it."
In effect, this new rule prevents a suspect from being a witness
against himself "in any criminal case" by excluding his words and
all things they lead to from the "criminal case." But the Kastigar
Court failed to explain persuasively where its new rule came from.
Counselman was out, but what, precisely, was the source of Kasti-
gar's "rational accommodation"? How would introducing a physi-
cal fact (fruit), but not anything a defendant ever said, make a
defendant a "witness" against himself "in" a criminal trial? Kastigar
provided few satisfying answers - perhaps because the Court
failed to see the huge stakes involved: the entire Fifth Amendment
pivots on the precise rule of immunity required.8 Beyond its
mushy rational accommodation balancing, Kastigar's main argu-
ment seems to be that a suspect must be placed in "substantially the
same position" whether or not he sings out of court. But this is
precisely the "no worse off" test that, as we have seen, is wholly
unworkable as a global principle.
Recently, however, even Kastigarhas come under attack. Pres-
sure is growing for efficient law enforcement. When an immunized
witness is later prosecuted, Kastigar places the burden on the gov-
ernment "to prove that the evidence it proposes to use is derived
from a legitimate source wholly independent of the compelled testi-
mony.' ' 86 At times, this burden can be staggering.87 The govern-
ment may try to build a Chinese wall between prosecutors exposed
circuits. See e.g.,SALTZuR, & CAIP'A, supra note 30, at 484; United States v. Parker, 848
F.2d 61, 62 n.1 (5th Cir. 1988); United States v. Ingraham, 832 F.2d 229, 238 (1st Cir. 1987);
United States v. Perry, 788 F.2d 100, 115-16 (3d Cir. 1986); In re Grand Jury Proceedings
Larson, 785 F.2d 629, 630 n.2 (8th Cir. 1986); Grand Jury Subpoena of Ford v. United States,
756 F.2d 249, 253 (2d Cir. 1985).
83. 406 U.S. at 446.
84. 406 U.S. at 462.
85. See supra note 64; see also Baltimore City Dept. of Social Servs. v. Bouknight, 493
U.S. 549 (1990), where the Court dodged the only issue in the case: the scope of immunity,
see supra note 54.
86. 406 U.S. at 460.
87. See Letter from William Treanor to Akhil Amar (December 23, 1994) (on file with
authors). Professor Treanor was a prosecutor in the Oliver North investigation.
to the testimony and prosecutors working on the case against the

witness.88 But sometimes a Chinese wall cannot be built, and some-
times the wall leaks. Another common technique is "canning" the
results of an investigation before a potential defendant testifies,89
but this involves even greater administrative burdens. To make
matters worse, courts have adopted varying interpretations of the
independent source standard.
At one extreme is the D.C. Circuit's holding in United States v.
North.90 Oliver North, as many will recall, received a grant of im-
munity to testify in a congressional hearing involving the Iran-
Contra affair. The prosecutor had not seen the testimony, nor did
he seek to use it in the prosecution. Several prosecution witnesses,
however, had seen the testimony on their own. The court an-
nounced that the Fifth Amendment is violated "whenever the pros-
ecution puts on a witness whose testimony is shaped, directly or
indirectly, by compelled testimony, regardless of how or by whom
he was exposed to that compelled testimony." 91 The court imposed
additional administrative burdens on prosecutors in its suggestion
that the government's burden could be met by "canning the testi-
mony beforehand, just as wise prosecutors meet their burden of
showing independent investigation by canning the results of the in-
vestigation before the defendant gives immunized testimony." 92
The D.C. Circuit's superstrict approach leaves little difference be-
tween use plus use-fruits and transactional immunity.
On the other hand, several circuits have diverged from a super-
strict approach, and with good reason.93 Besides its sometimes-
88. See United States v. Schwimmer, 882 F.2d 22, 26 (2d Cir. 1989) (recommending the
use of Chinese walls); U.S. DE=T. OF JUSTICE, U.S. ATrom Ys' MANUAL § 111A0 (1987)
(suggesting that prosecution of an immunized witness should be handled by a lawyer unfamil-
iar with the substance of the witness's testimony).
89. See United States v. North, 920 F.2d 940, 942-43 (D.C. Cir. 1990) (amending United
States v. North, 910 F.2d 843 (D.C. Cir. 1990)). An investigation is "canned" by completing
all steps - such as interviewing witnesses, taking depositions, and searching for physical
evidence - before immunized testimony is given. Thus the testimony could not have af-
fected the investigation.
90. 920 F.2d 940 (D.C. Cir. 1990).
91. 920 F.2d at 942.
92. 920 F.2d at 942-43 (internal quotations and citation omitted).
93. Uses of immunized testimony are divided into two - largely artificial - categories:
nonevidentiary and evidentiary uses. Nonevidentiary use generally involves exposure of a
prosecutor to immunized testimony. The term is vague and mainly defined by lists of exam-
ples. One court has listed the following nonevidentiary uses: "assistance in focusing the
investigation, deciding to initiate prosecution, refusing to plea-bargain, interpreting evidence,
planning cross-examination, and otherwise generally planning trial strategy." United States
v. McDaniel, 482 F.2d 305, 311 (8th Cir. 1973) (finding such uses impermissible). Although
the North court did not reach the question of whether nonevidentiary use had occurred be-
cause the prosecution had not been exposed to the immunized testimony, the court suggested
crushing burdens on prosecutors, the D.C. Circuit approach at

times can embroil courts in a futile attempt to resolve never-never-
land counterfactuals: What would the world look like if Oliver
North had never testified? God knows! 94 In light of these meta-
physical imponderables, why should the law not simply presume -
irrebuttably - that somehow, some way, the truth and the fruit
might have come to light anyway? (This presumption would in ef-
fect simply expand the scope of current inevitable discovery doc-
trine and would track the approaches followed in England and
Canada. 95) But under this theory, what would be left of use-fruits
immunity? Only the testimony itself would be excluded.
In the same way that the Supreme Court in Murphy chipped
away at Counselman's transactional immunity before overruling it
in Kastigar, the Court is now chipping away at use plus use-fruits
immunity in the context of Mirandawarnings. When the police in-
terrogate a suspect without giving complete warnings, Mirandabars
that nonevidentiary use was impermissible. See 910 F.2d at 856. Several courts have come to
the opposite conclusion. See United States v. Serrano, 870 F.2d 1, 16-17 (1st Cir. 1989) (stat-
ing that Kastigar does not protect against all nonevidentiary uses of compelled testimony
because the distinction between use plus use-fruits and transactional immunity would disap-
pear otherwise); United States v. Mariani, 851 F2d 595, 600 (2d Cir. 1988) (stating that the
Fifth Amendment does not "foreclose the prosecution of an immunized witness where his
immunized testimony might have tangentially influenced the prosecutor's thought
processes"); United States v. Byrd, 765 F.2d 1524,1530-31 (11th Cir. 1985) (stating that Kastd-
gar does not protect against nonevidentiary uses of compelled testimony such as the decision
whether to indict or whether to accept a plea bargain); United States v. Pantone, 634 F.2d
716, 730-31 (3d Cir. 1980) (holding that Kastigardoes not prohibit a prosecutor's "mere ac-
cess to immunized grand jury testimony").
Evidentiary uses - uses that would somehow contribute to the evidence presented at
trial - are generally not permitted in Kastigar'sregime. The cases revolve around the defi-
nition of evidentiary. In North, the D.C. Circuit declared that use of the immunized testi-
mony by witnesses to refresh their memories - a virtually irrebuttable presumption if they
were exposed to the testimony - is evidentiary use. See 920 F.2d at 945-46. Again, several
courts have disagreed. See United States v. Helmsley, 941 F.2d 71, 82 (2d Cir. 1991) (declin-
ing to apply the evidentiary use concept to a witness who may have been exposed to immu-
nized testimony); United States v. Kurzer, 534 F.2d 511, 517 (2d Cir. 1976) (requiring the
government to prove merely that the witness's decision to testify was not influenced by im-
munized testimony).
Different approaches to use immunity are discussed in Jerome A. Murphy, Comment,
The Aftermath of the Iran-ContraTrials: The UncertainStatus of Derivative Use Immunity, 51
MD. L. REv. 1011, 1030-31, 1045-46 (1992) (discussing cases that have differed from the
North standard) and Gary S. Humble, Nonevidentiary Use of Compelled Testimony: Beyond
the Fifth Amendment 66 TEXAs L. Rnv. 351 (1987). The North case provoked a legislative
effort to overturn it. Senators Joseph Lieberman and Warren Rudman introduced a bill per-
mitting "the use of testimony based on a witness's personal knowledge, regardless of whether
the witness has been exposed to the defendant's compelled testimony, as long as such expo-
sure is in no way attributable to the prosecution." Michael Gilbert, Note, The Future of
CongressionalUse Immunity After United States v. North, 30 AM. Cmm. L. REv. 417,434-35
(1993) (citing S. 2074, 102d Cong., 1st Sess., 137 CONG. RE. S18,385 (1991)).
94. Yet again, we see a possible "causation gap." See supra notes 33 and 78.
95. See infra section II.C.
the prosecution from using any resulting confession in its case in

chief. However, the Supreme Court has repeatedly chosen to admit
the fruits of such confessions. In Michigan v. Tucker,96 the defend-
ant, arrested for rape, was informed of his right to remain silent and
right to counsel, but not of the fact that he would be given counsel
if he were indigent. The defendant told the police he was with his
friend Henderson at the time of the crime, and the police then
questioned Henderson. Henderson's information suggested that
Tucker was guilty. His testimony was admitted at trial, over the
defendant's objection, and the defendant was convicted. Emphasiz-
ing reliability, the Court held that while the failure to give the full
Miranda warning required the exclusion of defendant Tucker's
statement, it did not require exclusion of his friend Henderson's
testimony:
There is plainly no reason to believe that Henderson's testimony is
untrustworthy simply because respondentwas not advised of his right
to appointed counsel. Henderson was both available at trial and sub-
ject to cross-examination by respondent's counsel, and counsel fully
used this opportunity, suggesting in the course of his cross-examina-
tion that Henderson's character was less than exemplary and that he
had been offered incentives by the police to testify against respon-
dent. Thus the reliability of his testimony was subject to the normal
testing process of an adversary trial.97
In Tucker, the Court did not categorically declare that a mere
violation of Mirandawould always allow use of a confession's fruits,
but other decisions seem to be heading that way. In a concurring
opinion in New York v. Quarles,98 Justice O'Connor advocated a
bright-line rule that physical evidence obtained as a result of a con-
fession after a Miranda violation should be admissible.99 In
Quarles, a woman told police officers that she had just been raped,
described her assailant, and said that he had just entered a nearby
supermarket and was carrying a gun. One of the officers chased the
defendant, who matched the description given by the woman,
through the supermarket. The officer caught him and found a hol-
ster but no gun; the officer asked him where the gun was, the de-
fendant told him, and the gun was recovered there. The officer
then read the defendant his Miranda rights, and the defendant ad-
96. 417 U.S. 433 (1974).

97. 417 U.S. at 449.
98. 467 U.S. 649 (1984).
99. See 467 U.S. at 672 (O'Connor, J., concurring in part and dissenting in part) ("Limita-
tion of the Miranda prohibition to testimonial use of the statements themselves adequately
serves the purposes of the privilege against self-incrimination.").
mitted that he owned the gun.10° In the ensuing prosecution for

criminal possession of a weapon, the trial judge excluded both the
defendant's initial statement of where the gun was and the gun it-
self.101 Justice O'Connor thought the gun itself should come in:
"Admission of nontestimonial evidence of this type is based on the
very sensible view that procedural errors should not cause entire
investigations and prosecutions to be lost."102 O'Connor relied on
the Schmerber line of cases, which emphasized the difference be-
tween physical and testimonial evidence, as support for her ap-
proach.103 Writing one year later in Oregon v.- Elstad,10 4 this time
for the majority, Justice O'Connor again invoked Schmerber -
"[t]he Fifth Amendment, of course, is not concerned with nontesti-
monial evidence"' 05 - and declined to suppress the fruits of a
"mere" Miranda violation.106 Once again, the Court stressed "the
Fifth Amendment goal of assuring trustworthy evidence."10 7 This
approach has led most circuits to embrace the view that all fruits of
100. 467 U.S. at 652-53.

101. 467 U.S. at 652-53 (describing the trial court's rulings).
102. 467 U.S. at 673 (O'Connor, J., concurring in part and dissenting in part). Justice
O'Connor emphasized the importance of examining the approaches of countries such as Eng-
land, India, Scotland, and Ceylon in crafting our own rules regarding confessions and pointed
out that the Court in Miranda had explicitly looked to those countries in developing the
Miranda rule. She noted that in those countries, "nontestimonial evidence derived from all
confessions 'not blatantly coerced' was and still is admitted." 467 U.S. at 673 (O'Connor, J.,
concurring in part and dissenting in part).
Note that in expressing concern that "entire investigations" not be lost, Justice O'Connor
seemed worried about the possibility that, say, the gun might well have been found regardless
of the antecedent Mirandaviolation. See also HENRY J. FRIENDLY, BENCHMARKS 279 (1967)
(expressing concern that applying Miranda to fruits would "in effect confer immunity unless
the prosecution can meet the burden of showing that its fruits would have been discovered
anyway"). Yet again, we see the possible "causation gap." See supra notes 33 and 78.
103. Justice O'Connor stated:
Indeed, whatever case can be made for suppression evaporates when the statements
themselves are not admitted, given the rationale of the Schmerber line of cases. Cer-
tainly interrogation which provides leads to other evidence does not offend the values
underlying the Fifth Amendment privilege any more than the compulsory taking of
blood samples, fingerprints, or voice exemplars, all of which may be compelled in an
"attempt to discover evidence that might be used to prosecute [a defendant] for a crimi-
nal offense."
467 U.S. at 670-71 (O'Connor, J., concurring in part and dissenting in part) (quoting Schmer-
ber v. California, 384 U.S. 757, 761 (1966)). For more discussion of Schmerber, see infra
section I.D.
104. 470 U.S. 298 (1985).
105. 470 U.S. at 304.
106. 470 U.S. at 308-09. In Elstad, the "fruit" was itself a statement from the defendant,
which the Court found voluntary and reliable even though it was the fruit of an earlier,
Miranda-defective utterance.
107. 470 U.S. at 308.
a (merely) Miranda-defective confession are admissible - even

though the confession itself is inadmissible 08
But if Miranda is ultimately rooted in the Self-Incrimination
Clause, we need to ask what the courts have not (yet) asked: Why
not enforce the clause itself by excluding confessions and allowing
fruits? 109
D. "Witness"?
What does it mean to be a "witness" against oneself? Here, too,
the courts have been all over the map.
In the 1886 case of Boyd v. United States,1" 0 the Supreme Court
held that seizing or compelling production of a defendant's private
papers to be used in evidence against him was equivalent to com-
pelling him to be a witness against himself. Boyd's basic rule -
that a person's property could not be used against him in a criminal
proceeding - is plausible in the context of personal papers such as
diaries."1 A defendant's diary testifies: it speaks in the defendant's
108. Se4 e.g., United States v. Gonzalez-Sandoval, 894 F.2d 1043, 1048 (9th Cir. 1990)
(stating that "the reasoning of Elstad and Tucker applies as well to non-testimonial physical
evidence obtained as a result of a Miranda violation"); United States v. Sangineto-Miranda,
859 F.2d 1501, 1514-18 (6th Cir. 1988) (admitting the fruits of mere Miranda violation be-
cause "the goal of the fifth amendment's exclusionary rule is to assure trustworthy evi-
dence"); United States v. Bengivenga, 845 F.2d 593, 600-01 (5th Cir. 1988) (finding that a
"mere violation of Miranda" does not trigger the exclusion of nontestimonial fruit); United
States ex rel. Hudson v. Cannon, 529 F.2d 890, 894-95 (7th Cir. 1976) (holding, in reliance on
7cker and before Quarles, that fruits of a mere Miranda violation should not be excluded).
109. Note how various statements from Justice O'Connor's opinion explicitly speak of
the logic underlying not merely Miranda but the privilege against self-incrimination itself.
See supra notes 98-108 and accompanying text. Her invocations of Schmerber also sound in
pure self-incrimination theory, as distinct from "mere" Mirandaconcerns. See supranote 103
and text accompanying note 105.
110. 116 U.S. 616 (1886).
111. On the common law's special solicitude for private papers, see Entick v. Carrington,
19 Howell's State Trials 1029 (C.P. 1765). The idea of what was "private," however, was
construed very broadly indeed in certain English cases and went far beyond diaries. Se4 e.g.,
Regina v. Mead, 92 Eng. Rep. 119, 119 (K.B. 1703) (refusing to require the defendants, who
were charged with executing an office of trust without taking an oath, to produce books
showing the election of the corporation's officers because "they are perfectly of a private
nature"); Rex v. Worsenham, 91 Eng. Rep. 1370, 1370 (K.B. 1701) (refusing to require the
defendants in a forgery case, who were custom-house officers, to produce custom-house
books because the books were "a private concern" and therefore requiring production would
be "to compel the defendants, to produce evidence against themselves"); Chetwind v. Mar-
nell, 126 Eng. Rep. 900, 900 (C.P. 1798) (noting that the plaintiff, in an action brought on a
testator's bond that was suspected of being forged, would not be compelled to produce the
bond for inspection, if it might be the means of convicting him of a capital felony); The
Queen v. Granatelli, 7 Rep. State Trials (New Series) 979, 986 (C.C.C. 1849) (refusing to
require a witness, in a prosecution for fitting out a vessel against a friendly power, to produce
a sales agreement of the company for which he was secretary because it might incriminate
him); see also Roe v. Harvey, 98 Eng. Rep. 302, 305 (K.B. 1769) (Mansfield, L.J.) ("[I]n a
criminal or penal cause, the defendant is never forced to produce any evidence; though he
should hold it in his hands, in Court."). None of these cases involved immunity statutes, and
own words, much as would the defendant himself as a witness on

the stand. 112 Diaries often express secret thoughts that implicate
First Amendment concerns and sometimes pose reliability
problems because of the use of personal shorthand, fantasy, and so
forth.
But not all papers pose these difficulties. In particular, papers
kept in connection with legal matters such as taxes or customs du-
ties are quite different. Boyd involved a subpoena requiring the
defendant to produce not a diary but books, invoices, and papers
for enforcement of customs laws. Boyd's exclusion of such records
was inspired by a spirit akin to the spirit of Lochner11 3 that a per-
son's right to his property includes the right not to have it intro-
duced against him in a criminal case.114 (Significantly, Boyd and its
immediate progeny involved corporate crime and breaches of regu-
latory requirements, not violent crime like rape or murder.1 1 5 )
In a less extreme version of Boyd, many courts in the late nine-
teenth and early twentieth centuries held that using the defendant's
body as physical evidence was in effect compelling the defendant to
be a witness against himself.116 The question, however, was hotly
contested; some jurisdictions admitted such evidence. Courts di-
vided on whether a defendant could be compelled to undergo a
physical examination,117 to make a footprint or fingerprints, 11 8 or to
thus none focused on the precise scope of immunity necessary to overcome the self-
incrimination privilege.
112. See TELFoRD TAYLOR, Two S'rmu s IN CONSTITUTIONAL INTERPRETATION 67
(1969); Amar, supra note **, at 788.
113. See Lochner v. New York, 198 U.S. 45 (1905).
114. See Amar, supra note **, at 788.
115. See id; see also William Stuntz, The Substantive Origins of CriminalProcedure 105
YALE LJ. (forthcoming Nov. 1995).
116. One commentator states that excluding such evidence was the majority rule. See
Charles Gardner Geyh, The Testimonial Component of the Right Against Self-Incrimination,
36 CAi.-L U. L. REv. 611, 621 (1987).
117. Compare People v. Akin, 143 P. 795, 796 (Cal. Dist. Ct. App. 1914) (refusing to
compel a physical examination for venereal disease); State v. Height, 91 N.W. 935, 940 (Iowa
1902) (same); State v. Newcomb, 119 S.W. 405, 409 (Mo. 1909) (same); Bowers v. State, 75
S.W. 299, 300 (Tex. Crim. App. 1903) (same); People v. McCoy, 45 How. Pr. 216, 217 (N.Y.
Sup. CL 1873) (refusing to compel a physical examination for evidence of childbirth) with
O'Brien v. State, 25 N.E. 137 (Ind. 1890) (holding that the results of a compelled physical
examination for scars and identifying marks were admissible); State v. Miller, 60 A. 202 (N.J.
1905) (same); People v. Corder, 221 N.W. 309, 309-10 (Mich. 1928) (allowing evidence of a
voluntary physical examination); Noe v. Monmouth, 143 A. 750 (N.J. 1928) (holding that
compelling the physical examination of a driver for signs of intoxication was permissible).
118. Compare Cooper v. State, 6 So. 110 (Ala. 1889) (holding that a defendant cannot be
compelled to make footprints to have his tracks compared to those found at the scene of the
crime); Day v. State, 63 Ga. 668 (1879) (same); State v. Sirmay, 122 P. 748 (Utah 1912)
(same) with United States v. Kelly, 55 F.2d 67 (2d Cir. 1932) (permitting the State to compel
the defendant to give fingerprints); People v. Jones, 296 P. 317 (Cal. App. 1931) (same; find-
display himself to the jury or perform physical acts. 1 9 Courts

adopted various rationales for excluding or allowing the evi-
dence. 120 The reason for this uncertainty is obvious: using a de-
fendant's body is different in some ways from forcing him to take
the stand and testify as a witness, and yet in other ways it does treat
him like a witness to be observed by the jury. Is a witness someone
who testifies - or merely someone who is observed at trial? To an-
swer this question, we need an overall theory of the Self-
Incrimination Clause - and that is precisely what the legal commu-
nity seems to lack.
As the twentieth century wore on, the spirit of the Lochner era
declined, and so did Boyd and its progeny. Beginning with the
landmark case of Schmerber v. California,'21 a series of cases in the
1960s and 1970s rejected Boyd's approach. Schmerber, announced
a week after the Court's decision in Miranda,gave rise to a sweep-
ing assertion of the need to let in reliable physical evidence, via a
definition of witness that drew a sharp distinction between words
and physical evidence. The defendant was arrested at a hospital
while being treated for injuries from a car accident. A police officer
directed a doctor at the hospital to take a blood sample, and an
analysis of the alcohol percentage showed that the defendant was
intoxicated. The analysis report was introduced at trial, and the de-
ing that fingerprints are not testimonial); Magee v. State, 46 So. 529 (Miss. 1908) (allowing
the State to compel a defendant to put his foot in a track found at the scene of a crime,
because the compulsion posed no risk to truthfinding).
119. CompareBlackwell v. State, 67 Ga. 76 (1881) (holding that the State cannot compel
a defendant to show an amputated arm at trial); State v. Jacobs, 50 N.C. (5 Jones) 256 (1858)
(holding that the State cannot compel a defendant to show himself to the jury for purposes of
ascertaining his race); Ward v. State, 228 P. 498 (Okla. Crim. App. 1924) (holding that the
State cannot compel the defendant to put on a coat); Thrman v. State, 95 S.W. 533 (Tex.
Crim. App. 1898) (holding that the State cannot compel the defendant to put a cap on his
head) with Ross v. State, 182 N.E. 865 (Ind. 1932) (finding it permissible to compel a defend-
ant to grow a beard); State v. Oschoa, 242 P. 582 (Nev. 1926) (finding it permissible to com-
pel a defendant to show his body to the jurors and to put on a shirt for them); State v. Ah
Chuey, 14 Nev. 79 (1879) (holding that a compelled showing of a tatoo does not violate the
privilege, because the privilege's purpose is to help find the truth); Sprouse v. Common-
wealth, 81 Va. 374 (1886) (finding that requiring a forger to write his name does not violate
the privilege).
120. See J.A.C. Grant, Self-Incrimination in the Modem American Law, 5 TmspLE L.Q.
368,373-87 (1931); see, eg., Bruce v. State, 21 S.W. 681 (Tex. Crim. App. 1893). The court in
Bruce stated:
[Tihe ground upon which this testimony is said to be admissible is that in these cases the
physical facts speak for themselves, and no [fears] or hopes of the prisoner could pro-
duce or effect a resemblance of his track, or of the wounds or clothing, and their resem-
blance aids the jury in their search after the truth.
21 S.W. at 682.
121. 384 U.S. 757 (1966).
Michigan Law Review [VCol. 93:857
fendant was convicted of driving under the influence of alcohol. 122

Justice Brennan wrote for the Court:
We hold that the privilege protects an accused only from being com-
pelled to testify against himself, or otherwise provide the State with
evidence of a testimonial or communicative nature, and that the with-
drawal of blood and use of the analysis in question in this case did not
involve compulsion to these ends.123
In support of the Court's position, Justice Brennan cited Holt v.
United States,124 a case where a defendant had been compelled
before the trial to try on a blouse related to the crime; the fact that
the blouse fit him served as incriminating evidence. The Holt
Court, per Justice Holmes, rejected the defendant's objection as
"an extravagant extension of the Fifth Amendment;" 12 Holmes de-
clared that "the prohibition of compelling a man in a criminal court
to be a witness against himself is a prohibition of the use of physical
or moral compulsion to extort communications from him, not an
exclusion of his body as evidence when it may be material." 126 Jus-
tice Brennan summed up the distinction thus: "[T]he privilege is a
bar against compelling 'communications' or 'testimony,' but...
compulsion which makes a suspect or accused the source of 'real or
physical evidence' does not violate it."127
Although Schmerber did not explicitly overrule Boyd, its logic is
hard to square with Boyd's.128 Boyd fused the Fourth and Fifth
Amendments together, holding that the two provisions "run almost
into each other."' 2 9 Schmerber analyzed each amendment sepa-
rately. 130 In dissent, Justice Black argued that it was a "strange hi-
erarchy of values" that allowed the state to convict someone with
his own blood but not with his "lifeless papers.' 3 ' If Black is
wrong and Schmerber is right, perhaps the Fifth Amendment is not
about privacy, as is the Fourth. And if the two amendments reflect
122. 384 U.S. at 758-59.
123. 384 U.S. at 761 (footnote omitted).
124. 218 U.S. 245 (1910); see 384 U.S. at 763.
125. 218 U.S. at 252.
126. 218 U.S. at 252-53.
127. 384 U.S. at 764.
128. See Note, The Life and imes of Boyd v. United States (1886-1976), 76 MICH. L.
REv. 184, 196-98 (1977) (authored by Stan Krauss).
129. Boyd v. United States, 116 U.S. 616, 630 (1886).
130. Compare 384 U.S. at 760-65 (Part II: "The Privilege Against Self-Incrimination
Claim") with 384 U.S. at 766-72 (Part IV: "The Search and Seizure Claim"). In each part,
Boyd is mentioned only once, and briefly. See 384 U.S. at 763-64 (reading Boyd as a self-
incrimination case about "papers"); 384 U.S. at 768 (reading Boyd as a search and seizure
case about warrants and thus as "not instructive" in the case at hand).
131. 384 U.S. at 775 (Black, J., dissenting).
separate ideas, rather than the same one, then perhaps we should
exclude not things - the Fourth Amendment's concern in its "ef-
fects" language 132 - but only words, as in testimonial witnessing,
from a criminal case.
But if so, Kastigar unravels. Justice O'Connor in Quarles noted
the broad implications of Schmerber and quoted Judge Friendly's
view that "[u]se of a suspect's answers 'merely to find other evi-
dence establishing his connection with the crime... differs only by
a shade from the permitted use for that purpose of his body or his
blood.' "133 O'Connor's own words went one step further: "Cer-
tainly interrogation which provides leads to other evidence does not
offend the values underlying the Fifth Amendment privilege any
more than the compulsory taking of blood samples .... ."134
Schmerber'sprogeny continued the division between testimonial
or communicative evidence and physical evidence. A series of
135
cases allowed a defendant to be compelled to stand in a lineup,
to give handwriting exemplars, 36 to give voice-prints, 37 and to take
38
sobriety tests measuring mental acuity and physical coordination.
39
Property, too, was no longer sacrosanct. Warden v. Hayden,'
which involved the seizure of an armed robber's clothing found in a
washing machine in his house, simultaneously rejected the Fourth
Amendment mere evidence rule established under Boyd's re-
gime' 40 and reaffirmed Schmerber in holding that because the cloth-
ing was not "testimonial" or "communicative," it could be
introduced. 141 Perhaps property protections were not so enticing to
the Court when violent crime was involved.
In the 1970s, the Court began to turn away from Boyd even in
the context of nonviolent crime and to lessen protection for papers.
In a series of cases that all but overruled Boyd, the Court held that
even a defendant's subpoenaed papers - except, perhaps, personal
papers like diaries - could be introduced at a criminal trial. The
132. U.S. CONsr. amend. IV (affirming the people's right to be secure in their "persons,
houses, papers, and effects" (emphasis added)).
133. New York v. Quarles, 467 U.S. 649,671 (1984) (O'Connor, J., concurring in part and
dissenting in part) (quoting FRmeNDLY, supra note 102, at 280).
134. 467 U.S. at 670-71 (O'Connor, J., concurring in part and dissenting in part).
135. See United States v. Wade, 388 U.S. 218 (1967).
136. See Gilbert v. California, 388 U.S. 263 (1967).
137. See United States v. Dionisio, 410 U.S. 1 (1973).
138. See Pennsylvania v. Muniz, 496 U.S. 582 (1990).
139. 387 U.S. 294 (1967).
140. 387 U.S. at 306-07.
141. 387 U.S. at 302-03.
Court's theory was that although they were testimonial, and

although their production was compelled, these papers were not
compelled testimony within the meaning of the Fifth Amendment
because the defendant was not compelled to create the papers in
the first place but only to hand them over. 142 (In this last respect,
such papers differed from the "required records" whose creation
the government had mandated. 43) If this is the Court's logic, how-
ever, it applies to diaries to.o; so why did the Court pointedly leave
this question open?' 44 And in order to decide whether subpoenaed
papers are indeed compelled testimony within the meaning of the
Fifth Amendment, we need an overall theory of the clause -
which, again, is precisely what we now lack. Until we have such a
theory, it is impossible to decide whether it was the Boyd Court or
the modem Court that was playing word games where compelled
papers are at issue.
In the end, the Court has been moving away from the over-
expansive view of the word witness propounded in Boyd and has
sharpened a distinction between compelled words - testimony -
and physical evidence. This distinction maps onto one common un-
142. See Andresen v. Maryland, 427 U.S. 463, 470-77 (1976) (holding that the introduc-
tion at trial of the defendant's business records did not violate the Fifth Amendment because
the statements were "voluntarily committed to writing" and were seized pursuant to a valid
search warrant); Fisher v. United States, 425 U.S. 391, 402-14 (1976) (holding that the de-
fendant's Fifth Amendment rights were untouched because he was compelled to produce
incriminating papers and not to give self-incriminating testimony); Bellis v. United States, 417
U.S. 85 (1974) (holding that neither a partnership nor its individual partners were shielded on
self-incrimination grounds from the compelled production of partnership records); Couch v.
United States, 409 U.S. 322 (1973) (holding that a summons served on a taxpayer's account-
ant requiring him to produce the taxpayer's personal business records in his possession did
not violate the taxpayer's Fifth Amendment rights). For more discussion, see Note, supra
note 128.
In one respect, the Court has qualified the general rule allowing introduction of subpoe-
naed documents or other physical evidence. The FisherCourt briefly observed that the act of
producing documents itself might have "communicative aspects," such as indicating the tax-
payer's belief that those were the documents described in the subpoena, that might be both
"testimonial" and "incriminating" for purposes of the Fifth Amendment. See 425 U.S. at 410.
But the Court did not attempt to lay down a rule for such cases. In United States v. Doe, 465
U.S. 605, 612-14 (1984) (Doe 1), the Court relied on this caveat in Fisherto uphold exclusion
of existent documents subpoenaed from the defendant. The Court emphasized, however,
that its holding was based on deference to factual findings made by the district court. See 465
U.S. at 613-14. Doe v. United States, 487 U.S. 201 (1988) (Doe 11), specified that the privi-
lege applied in cases where producing the evidence would testify to the existence, possession,
or authenticity of the things produced. See 487 U.S. at 209.
143. See supra section I.B.3.
144. See Fisher,425 U.S. at 414 (bracketing the issue of private papers under the Fifth
Amendment); Doe 1, 465 U.S. at 610 n.7 (repeating Ftsher's Fifth Amendment caveat con-
cerning private papers); 465 U.S. at 619 (Marshall, J., concurring in part and dissenting in
part) (emphasizing n.7 as leaving open the Fifth Amendment status of private papers). But
see 465 U.S. at 618 (O'Connor, J., concurring) (arguing, contrary to n.7, that "the Fifth
Amendment provides absolutely no protection for the contents of private papers of any
kind").
derstanding of the word witness. Witnesses testify; blood does not.

The word witness comes from the Old English witnes, meaning
"knowledge" (related to the Old English witan, "to know"). 145 Ex-
cept in a poetic sense, we do not usually conceive of blood as
"knowing" anything. But if this is the key to understanding the
privilege, why not exclude the defendant's compelled pretrial testi-
mony (words) but allow in the fruits (physical evidence) that the
defendant's out-of-court compelled testimony might lead to?
E. What's the Big Idea?

What basic rationales underlie the Fifth Amendment? Many
discussions by judges and scholars have obscured the privilege be-
hind clouds of eulogy. Over a century and a half ago, Bentham
noted that the privilege had long been shielded by an "assumption
of the propriety of the rule, as a proposition too plainly true to ad-
mit of dispute." 146 In his widely influential 1968 Robert S. Marx
Lectures, Judge Friendly urged a thorough examination of the poli-
cies of the privilege: the task was "indispensable to any reconsider-
ation of the proper scope of the fifth amendment and peculiarly
necessary because of the extent to which eloquent phrases have
been accepted as a substitute for thorough thought."'147
None of the rationales typically given for the Self-Incrimination
Clause can satisfactorily explain the current scope of the privilege
and its relation to the rest of our legal and moral system.4s Some-
times, the idea behind a given rationale is simply wrongheaded. At
145. WBsTER's TmnD NEW INTERNATIONAL DICrIONARY 2627 (1971).
146. 5 JEREmY BEmNHms, RATIONALE OF JUDICIAL EVIDENCE 229 (London, Hunt &
Clarke 1827).
147. Friendly, supra note 63, at 679-81, 698.
148. Traditional rationales have long been under attack. Judge Friendly, David Dolinko,
and others have convincingly shown that various traditional rationales for the privilege can-
not support Fifth Amendment doctrine, at least as it now exists. See Friendly, supra note 63;
see also LEwis MAYERs, SHALL WE AmEND TH F= I AMENDMENT? (1959); David
Dolinko, Is There a Rationalefor the Privilege Against Self-Incrimination?, 33 U.C.L.A. L.
REv. 1063 (1986); Donald A. Dripps, Foreword. Against Police Interrogation - And the
PrivilegeAgainst Self-Incrimination,78 J. CiUM. L. & CRmINOLOoY 699 (1988). Other com-
mentators criticizing the Fifth Amendment over the years have included such luminaries as
John Henry Wiginore, Roscoe Pound, and Charles McCormick. See Friendly, supra note 63,
at 672-74. Against them are ranged the arguments of Robert S.Gerstein, Privacy and Self-
Incrimination,80 Eamcs 87, 90 (1970) (arguing that individuals should have absolute control
over revelations of guilt and remorse); Robert S.Gerstein, Punishment and Self-Incrimina-
tion, 16 AM. J. Jlins. 84,88 (1971) (similar); Thomas S.Schrock et al., InterrogationalRights:
Reflections on Miranda v. Arizona, 52 S. CAL. L. REv. 1, 49 (1978) (claiming that the purpose
of the privilege is to enhance autonomy by protecting the individual's right to choose how he
"takes responsibility"); Schulhofer, supra note 24, at 330-33 (suggesting that the privilege
protects innocent defendants from bad performances on the witness stand); and William J.
Stuntz, Self-Incriminationand Excuse, 88 COLuM. L. REv.1227,1229 (1988) (arguing that the
privilege should properly be construed as protecting "excusable perjury," not merely silence);
other times, the animating idea is valuable but proves too much or
too little (or both) and thus cannot explain why the clause goes as
far as it now does but no further.
One frequently mentioned rationale for the privilege is the
"psychological cruelty" of the so-called cruel trilemma: without the
privilege, the defendant would be forced to choose among self-
accusation, perjury, or contempt. 149 But our justice system has no
such scruples about compelling self-damaging answers from a civil
litigant both in pretrial discovery and on the witness stand. Nor
does our system object to forcing people to testify in criminal cases
against friends and family members - except spouses - even
though such compelled witnessing can be an extremely painful ex-
perience: today a mother may be forced, under penalty of con-
tempt, to testify against her son and send him to the gallows. 150
Thus, as a descriptive theory, the psychological cruelty argument
simply does not hold water. To make matters worse, it benefits
only guilty defendants: there is no trilemma if one is innocent and
says so. No other criminal procedure provision of the Bill of Rights
is designed to give special protection from conviction to guilty
defendants.' 51
Courts have also made much of the argument that the Fifth
Amendment protects a special zone of mental privacy.152 But here
cf. SALTZBURG & CAPRA, supranote 30, at 446-48 (compiling justifications for the privilege
and responses).
149. Murphy v. Waterfront Commn., 378 U.S. 52,55 (1964); see also Miranda v. Arizona,
384 U.S. 436, 460 (1966).
150. Professor Luban argues that the law should recognize broader intrafamily immunity
from compelled witnessing. See DAVID LurAN, LAWYERS AND JusncE: AN ETHICAL
STUDY 197 (1988). But the fact that Anglo-American law has never done this dramatizes the
weakness of Luban's account as a descriptive matter. Normatively, a self-incrimination privi-
lege seems much harder to justify than a family privilege: compelling an innocent mother to
send her own son to prison or death seems ruthlessly callous, but compelling him to tell the
truth and confess seems much less cruel. If he wanted to avoid this cruelty, he could have
done so by not committing the crime; his dilemma arises only because he is a criminal.
151. Cf. Amar, supra note **, at 790 n.125. Provisions like the First Amendment do of
course protect those guilty of "crimes" like heresy and seditious libel, but these provisions
sound in substance, not criminal procedure. As a matter of substantive law, heresy should
not be a crime at all; by contrast, the Fifth Amendment applies to things that should be
criminal, like rape, murder, and arson. The Eighth Amendment protects the guilty from
excessive punishment but not from convictions. The Double Jeopardy Clause does protect
the guilty via the plea of autrefois convict, but the clause and its underlying principles pro-
vide even more protection to the innocent via pleas of autrefois acquit and collateral estop-
pel. See Amar & Marcus, supra note 69, at 36-37. Moreover, autrefois convict only bars
multiple convictions of the guilty - in effect, protecting against excessive punishment, see id.
at 28-29,36 & n.184 - but of course allows the government one unfettered shot at convicting
the guilty on the basis of reliable evidence.
152. See Murphy, 378 U.S. at 55; Couch v. United States, 409 U.S. 322, 328 (1973); see
also Gerstein, Privacy and Self-Incrimination, supra note 148, at 90-91 (arguing that the self-
condemnation and remorse entailed by incriminating oneself should remain private). Justice
too, the treatment of civil litigants and witnesses belies this ration-
ale; they are often called to testify concerning intensely private,
highly embarrassing matters - in divorce cases, for example. Even
in criminal cases, immunity trumps the privilege. A witness given
immunity can be forced to testify about anything in his private
mental enclave. This treatment stands in dramatic contrast to the
true privacy privileges of wife-husband, priest-penitent, doctor-
patient, lawyer-client, and so on, for which no such trumping immu-
nity exists.
Related to the notion of protecting mental privacy is the more
convincing argument that the privilege helps to protect First
Amendment values. But the scope of the privilege on this rationale
is overbroad; its literal wording applies not just to political and reli-
gious groups but to murderers and rapists as well, whose cases lie
far from the core of the First Amendment.
A different possible rationale taps into ideas about parity and
symmetry. At the time of the Founding, a defendant was not even
allowed to take the stand and testify under oath for himself.153
Why, then, should the government have been allowed to force the
defendant to take the stand and testify under oath against him-
self?15 4 (Put another way, this disparity might seem to violate the
minimal idea of parity underlying the Compulsory Process Clause:
the government could compel the defendant to be a witness, but he
could not "compel" himself to be a witness. 55 ) But this logic is
O'Connor, however, has written that the privacy rationale does not apply to suspects in cus-
todial interrogation: "Where independent evidence leads police to a suspect, and probable
cause justifies his arrest, the suspect cannot seriously urge that the police have somehow
unfairly infringed on his right 'to a private enclave where he may lead a private life.'" New
York v. Quarles, 467 U.S. 649, 670 (1984) (O'Connor, J., concurring in part and dissenting in
part) (citing Murphy).
153. Defendants were not allowed to testify under oath at trial in America until the
midnineteenth century. See generally Joel N. Bodansky, The Abolition of the Party-Witness
Disqualification:An HistoricalSurvey, 70 Ky.L.J. 91 (1982).
154. See TsmODoRE BARLOw, THE JusnTCi OF PEACE: A TREATrsE CONTAnaN TmE
POWER AND DurY OF THAT MAGImRATE 189 (London, Lintot 1745) ("lIlt would be hard,
and unequal to rack a Man's Conscience with the Religion of an Oath, and make his Discov-
ery tend to his Condemnation, but not allow his Denial on Oath to have any Weight towards
his Exculpation or Acquittal."), quoted in Langbein, supra note 2, at 1085 n.157; cf. 3 JOHN
HENRY WIGMORE, TREA E ON THE ANGLO-AMERICAN SYSTEM OF EVIDENCE IN TRIALS
AT CoMMON LAW (3d ed. 1940). According to Wigmore,
In view of the apparent unfairness of a system which practically told the accused person,
"You cannot be trusted to speak here or elsewhere in your own behalf, but we shall use
against you whatever you may have said," it was entirely natural that the judges should
employ the only makeweight which existed for mitigating this unfairness and restoring
the balance, namely, [excluding unreliable] confessions.
Id. § 865(3), at 354.
155. See supra section I.A; see also Westen, supra note 10, at 119-20 (using the Compul-
sory Process Clause to buttress the defendant's right to testify); Peter Westen, Order of
obsolete today. The Court has held that a defendant enjoys a con-
stitutional right to testify on his own behalf. 156 In any event, this
rationale could not explain why fruits should be excluded, because
no antidefendant fruits asymmetry has ever existed. Defendants in
America have always enjoyed a general right to introduce physical
evidence or to put on third-party witnesses, even if they learned of
these things or persons from the government itself. Indeed, today
parity cuts against broad Fifth Amendment claims: if the govern-
ment is now obliged to supply a defendant with any exculpatory
evidence and information it has, why shouldn't the defendant be
obliged to supply the government with any inculpatory evidence
and information he has? 157
Another foundation of the clause might be "noninstrumental-
ization" - the notion that government impermissibly disrespects a
person when it uses him as the means of his own destruction. 15
But noninstrumentalization proves too much. The government
"uses" persons as witnesses all the time - whether they will or
no.
In general, the obligation to serve as a witness when necessary to
enforce the laws is part of the duty of citizenship; generally, the law
is entitled to every person's evidence. 59 If the government cannot
use a person against himself in a criminal prosecution, why may it
do so in a civil prosecution against him? And doesn't Schmerber160
legitimate dramatic instrumental use of a person against himself?
Though decided by the slimmest of margins in 1966, Schmerber is
an absolutely central case today - the rock on which a great many
cases and a considerable amount of crime detection policy have
Proofi An Accused's Right to Control the 71ming and Sequence of Evidence in His Defense,
66 CAL. L. REv. 935, 985 n.206 (1978) (similar).
156. See Rock v. Arkansas, 483 U.S. 44, 51-53 (1987) (holding that a defendant enjoys a
right to testify on her own behalf under the Fifth, Sixth, and Fourteenth Amendments).
157. See United States v. Agurs, 427 U.S. 97 (1976); Brady v. Maryland, 373 U.S. 83
(1963); Abraham S. Goldstein, The State and the Accused: Balance ofAdvantage in Criminal
Procedur 69 YALE L.. 1149, 1198 (1960).
158. On noninstrumentalization and government "usings," see generally Jed Rubenfeld,
The Right of Privacy, 102 HARv. L. REV. 737 (1989); Jed Rubenfeld, Usings, 102 YALE LJ.
1077 (1993); and Jed Rubenfeld, Reading the Constitution as Spoken, 104 YALE LJ. 119
(1995). On the noninstrumentalization idea in the self-incrimination context, see LUBAN,
supra note 150, at 194 ("[Mlaking me the active instrument of my own destruction signals the
entire subordination of the self to the state."). As this Luban quote indicates, individual
variations of the noninstrumentalization idea often come equipped with a set of nice distinc-
tions between active and passive use.
159. See 8 WIGMORE, supra note 154, § 2192, at 64 ("For more than three centuries it has
now been recognized as a fundamental maxim that the public (in the words sanctioned by
Lord Hardwicke) has a right to every man's evidence.") Frankly, it is hard to see how mod-
em society could operate without this general presumption.
160. Schmerber v. California, 384 U.S. 757 (1966).
been built. Can anyone now imagine even a single Justice voting
that government may not use an arrestee by forcing him to submit
to photographing, fingerprinting, and voice tests whose results may
be introduced in a criminal court? And if these instrumental uses
are okay, why is using testimonial fruits so different on instru-
mentalization grounds? 161
Occasionally, courts and scholars invoke notions such as prefer-
ring an accusatorial over an inquisitorial system' 62 or achieving a
"fair balance" between individuals and the state' 63 to explain the
privilege. These phrases, however, are more like slogans that sim-
ply restate the rule than carefully considered rationales. At times,
the fair balance idea collapses into a sporting theory of justice -
the idea that we should boost the odds for criminals just to keep the
game interesting, above and beyond the valuable and important
"handicap" that the government must prove its case beyond reason-
able doubt. 64 And the aversion to inquisitorial schemes cannot ex-
plain why the privilege applies only in criminal but not civil cases.
What's more, these notions fail to explain the key case of Schmer-
ber: Is it not arguably imbalanced or inquisitorial when a prosecu-
tor may suck blood from her adversary's very veins with needles
that invade his body, and then use his own blood to destroy him at
trial in a capital case, and then reinject those veins with lethal
poison?
Also flawed is the related notion that the government must
shoulder its entire burden of proof, or at least its prima facie case,
without any help from the defendant. 65 The moral intuition under-
161. Other problems for noninstrumentalists: Doesn't the government use a suspect as
the testimonial instrument of his own destruction when it secretly invades his house (with a
warrant), wiretaps his conversations without his consent, and then uses his own words against
him in a criminal trial? Or when it subpoenas the defendant to furnish extant documents
written in his own hand and then uses those documents at trial? Or when it compels a de-
fendant to authorize (with words) the release of his own bank statements and then uses the
authorization and the bank statements to convict him? See Andresen v. Maryland, 427 U.S.
463, 470-77 (1976); Fisher v. United States, 425 U.S. 391 (1976); Doe v. United States, 487
U.S. 201 (1988) (Doe II).
162. See Murphy v. Waterfront Commn., 378 U.S. 52, 55 (1964).
163. See 378 U.S. at 55.
164. See In re Wnship, 397 U.S. 358 (1970). This valuable safeguard does benefit some
guilty defendants but only as an incidental byproduct of achieving its direct aim of protecting
innocent defendants from erroneous acquittals.
165. See Williams v. Florida, 399 U.S. 78, 108 (1970) (Black, J., dissenting) (arguing that
the Florida notice-of-alibi rule violated the Fifth Amendment and claiming that a criminal
defendant has a "historical and constitutionally guaranteed right... to remain completely
silent, requiring the State to prove its case without any assistance of any kind from the de-
fendant himself"); see also Kevin R. Reitz, Clients, Lawyers, and the Fifth Amendment The
Need for a Projected Privilege, 41 DuKE L.J 572, 581-82 (1991) (asserting that defendant
retains "the right to doubt the justness" of the state's criminal law enforcement and thus the
894 Michigan Law Review [Vol. 93:857
lying this "look ma, no hands" idea is fuzzy -especially in light of

the government's heavy burden of proof - and once again, the
idea runs afoul of Schmerber and its progeny. In many contexts,
government can oblige a defendant to "help" in nontestimonial
ways: by showing up at trial, by allowing witnesses to point at him,
by making voice prints, by giving up his very blood, and so on.1 66
Courts have also, rightly, shown considerable concern for deter-
ring improper police practices, including physical brutality. But
does the current scope of the privilege accomplish this goal?
Although the more egregious forms of interrogation abuse such as
beatings have stopped, 167 the practices detailed by Simon indicate
that intimidation is alive and well in the police station. Far from
civilizing 68 the interrogation process, the current interpretation of
the Fifth Amendment has driven it underground. Open interroga-
tion of suspects under direct judicial supervision would serve the
goal of deterring police abuse far better, as would a better-con-
structed Fourth Amendment remedial regime for unreasonable
seizures of persons. 69
In addition, courts have stated that the privilege protects an in-
nocent defendant from a bad performance on the stand.17° But as
we have seen, the broad scope of the current privilege actually
harms the innocent defendant by denying her the Sixth Amend-
right to refuse his aid). Reitz here confuses the First and Fifth Amendments. Of course a
defendant has the right to doubt the state - and to say so vigorously and without penalty -
but the claimed right to withhold information and reliable evidence without penalty simply
does not follow.
166. Not surprisingly, Justice Black's dissent in Williams, setting out the "no help from
defendant" theory, echoes his Schmerber dissent and indicates that Justice Black understood
that to accept Schmerber - as do all the Justices today, presumably - is to reject the "no
help" theory. See Williams, 399 U.S. at 111 (Black, J., dissenting) ("[A] criminal defendant
cannot be required to give evidence, testimony, or any other assistance to the State to aid it
in convicting him of crime." (citing Schmerber,384 U.S. at 773 (Black, J., dissenting))).
We do not mean to suggest that the government has no initial burden to meet before it
can pluck someone off the street and question him or require him to produce evidence. Fish-
ing expeditions are prevented by the Fourth Amendment rule against unreasonable searches
and seizures, which typically calls for reasonable suspicion before stopping and questioning,
and probable cause before arrest. See infra section II.E.
167. See SIMON, supra note 57, at 199 ("Miranda and its accompanying decisions...
effectively ended the use of violence and the most blatant kind of physical intimidation in
interrogations.").
168. A pun is very much intended. We mean here to conjure up a "civilized" process akin
to that used today in "civil" law countries and in "civil" pretrial discovery in the United
States.
169. Some ideas for bolstering Fourth Amendment remedies are sketched out in Amar,
supra note **,at 811-19.
170. Cf.Murphy v. Waterfront Commn., 378 U.S. 52, 55 (1964); Wilson v. United States,
149 U.S. 60, 66 (1893).
ment right to compulsory process. 171 Regarding her own perform-

ance, the scope of immunity is again too broad; the problem could
be solved simply by allowing the defendant to refuse to testify at
trial for the jury to hear but eliciting information pretrial that could
generate admissible fruit.
Finally, and relatedly, courts and commentators have stressed
that coerced statements are unreliable and that the privilege there-
fore serves the goal of reliability. 172 This is indeed a worthy goal,
and courts have increasingly emphasized it over the past three de-
cades. But if this is the touchstone, again the scope of immunity
today is too broad. Why exclude the physical fruits of confessions,
when these are quite reliable and often highly probative pieces of
173
evidence?
In short, the various rationales repeatedly wheeled out to ex-
plain the privilege do not fit with the current scope of immunity. 174
Small wonder, then, that the Self-Incrimination Clause - virtually
alone among the provisions of the Bill of Rights - has been the
target of repeated analytic assault over the course of the twentieth
century from thoughtful commentators urging constitutional
amendments to narrow it or repeal it altogether. 75
In part, the current confusion about the rationale of the privi-
lege stems from historical confusion. In the past six centuries the
privilege has reflected several different fundamental ideas. Recent
scholarship has displaced the earlier heroic, teleological accounts of
171. See supra section I.A.

172. See, e.g., Michigan v. Tucker, 417 U.S. 433, 448-49 (1974); Murphy, 378 U.S. at 55;
GxswouD, supra note 32, at 10-19; Schulhofer, supranote 24; see also Withrow v. Williams,
113 S.Ct. 1745,1753 (1993) (linking the "Fifth Amendment 'trial right'" and Mirandato "the
correct ascertainment of guilt" and arguing that "Miranda serves to guard against 'the use of
unreliable statements at trial.'" (quoting Johnson v. New Jersey, 384 U.S. 719, 730 (1966))).
But see Tehan v. United States ex reL Shott, 382 U.S. 406, 415-16 (1966) (rejecting the inno-
cence-protection rationale in the course of refusing to give Griffin retroactive effect).
173. In a thoughtful essay that properly focuses on innocence and reliability, Professor
Schulhofer fails to discuss how these rationales argue against current Fifth Amendment fruits
doctrine. See Schulhofer, supra note 24, at 330-33.
174. Professor Stuntz's argument that the Fifth Amendment protects "excusable per-
jury," see Stuntz, supra note 148, is flawed for several reasons. First, it is anachronistic to
think that people at the Founding would commit perjury lightly, see infra notes 205-07 and
accompanying text, and so Stuntz's argument fails to explain the Framers' vision. Second' it
is morally inexcusable to condone lying when the lie merely compounds the liar's underlying
crime. Third, Stuntz's argument rests on a faulty Perry-Mason-like view of confessions. Lies
are still useful to prosecutors in a variety of ways. See infra notes 191-94, 208, and accompa-
nying text. In fairness to Stuntz, he claims not that his theory is historically rooted or norma-
tively appealing but only that it fits the cases.
175. See supranote 148.
the privilege's development. 176 It now appears that the privilege

evolved from a maxim of canon law imported from the Continent:
nemo tenetur prodere seipsum, "no one is obliged to produce him-
self."' 177 Originally, this meant that the duty to reveal all sins at
confession, as a condition of absolution, did not entail having to
come forward and accuse oneself in court. 178 But once a prosecu-
tion was initiated and one was accused and called on as a witness,
one had to answer truthfully. 179
The sixteenth- and seventeenth-century incarnation of the privi-
lege, closely related to the medieval version, involved protections
against religious intolerance and open-ended fishing expeditions.
In England, prerogative courts such as the Star Chamber and the
High Commission and ecclesiastical courts used the oath ex of-
ficio.180 In this procedure, a person could be plucked from the
street and forced to swear an oath that he would answer any ques-
tions the court might decide to ask him 8 ' before any charge had
been leveled against him or any probable cause had been shown to
justify singling him out. To make matters worse, the purpose of
these indiscriminate procedures was often to identify and punish
those whose only possible offense was theological disagreement
176. In the past decade, our knowledge of the origins of the privilege has been signifi-
cantly expanded and perhaps revolutionized. The earlier-received account was built on two
works: igmore's massive treatise and Levy's heroicizing work, which built on Wigmore's
version. 8 WIGMORE, supra note 154, § 2250; LnvY, supra note 20.
177. A variant of the maxim is nemo tenetur accusareseipsum, "no one is obliged to
accuse himself." Simeon E. Baldwin, PreliminaryExaminations in CriminalProceedings, 6
A.B.A. REP. 225, 229 (1883); Edward S. Corwin, The Supreme Court's Construction of the
Self-Incrimination Claus14 29 Mmcff. L. Rlv. 1, 3 (1930). TWo recent articles powerfully argue
that the nemo tenetur maxim was imported from the Continent for use in inquisitorial proce-
dure, thus casting doubt on Leonard Levy's argument that the privilege was an English in-
vention. See R.H. Helmholtz, Origins of the Privilege Against Self-Incrimination: The Role
of the EuropeanIus Commune, 65 N.Y.U. L. Rnv. 962,967-69 (1990); Michael R.T. Macnair,
The Early Development of the Privilege Against Self-Incrimination, 10 OxFoRD J. LEOAL
STUD. 66, 67-70 (1990).
178. Langbein, supra note 2, at 1072; Helmholtz, supra note 177, at 982.
179. E.M. Morgan, The Privilege Against Self-Incrimination, 34 MiNN. L. Rnv. 1, 4
(1949).
180. Coke's early-seventeenth-century discussion of the oath ex officio reveals distrust of
its use to uncover thoughts. The Privy Council on a motion from the House of Commons
asked Coke and Chief Justice Popham when the oath could properly be administered. As
part of their answer, they stated: "No Man ... shall be examined upon secret Thoughts of his
Heart, or of his secret Opinion: But something ought to be objected against him what he
hath spoken or done." An Oath before an EcclesiasticalJudge ex Officio, in 12 COKE's REP.
26 (3d ed., 1727). They were particularly concerned about questioning involving "heresy and
errors of faith." Id. They also objected to the lack of any preliminary showing of suspicion
via accusation or presentment or the like. See Corwin, supra note 177, at 7-8.
181. See Mary H. Maguire, Attack of the Common Lawyers on the Oath Ex Officio as
Administered in the EcclesiasticalCourts in England,in ESSAYS iN HISTORY AND POLrMCAL
THEORY 199 (Carl Wittke ed., 1936).
with the crown. During the sixteenth and seventeenth centuries,

the battles over the oath ex officio concerned the subject matter
jurisdiction of ecclesiastical courts and the quality of the charge
needed before the oath could be given.1 82
The privilege in a more modern form was half-heartedly in-
voked in late seventeenth-century heresy and sedition trials, but,
contrary to Wigmore's account, the privilege was not firmly estab-
fished at that time.183 Rather than emerging in the seventeenth
century to protect defendants in sedition and heresy trials, the mod-
em privilege actually grew legs in the late eighteenth century in or-
dinary criminal trials, both in England and in America. 184 It is
probably no coincidence that it was just at this time that defense
lawyers came to be widely used in criminal trials. Now that they
had lawyers, criminal defendants could afford to be silent rather
than having to conduct their own defenses; for their part, the law-
yers must have enjoyed being able to control their cases without
interference from their clients' risky testimony. 85
The privilege in anything like the form we know it today was
slow to catch on in America, even after the ratification of the Bill of
Rights, and references to the privilege during the ratification de-
bates were few. One reason is that routine pretrial procedure in-
volved questioning of an accused by a magistrate, and the accused
was expected to answer.' 86 If he did not and insisted on standing
mute in the face of accusations, this refusal to answer could be laid
before a later criminal jury for whatever inferences they might
182. Accordingly, in John Lilbume's first trial, in the Star Chamber, his objection was not
that he had a right not to answer incriminating questions but that he had a right to a proper
accusation before he did so. 8 wioMopm, supra note 154, § 2250, at 291, 298; Corwin, supra
note 177, at 8. He refused to answer to any matter not included in the information against
him. For this he was whipped, pilloried, fined, and imprisoned. With the victory of the par-
liamentary forces, the Long Parliament abolished the Star Chamber and the High Commis-
sion and forbade ecclesiastical courts to use the oath ex officio. For a discussion, see Charles
M. Gray, Prohibitionsand the Privilege Against Self-Incimination, in TUDOR RULE AND
REVOLuFON 345 (Delloyd J. Guth & John W. McKenna eds., 1982).
183. Wigmore placed the origin of the modem privilege with Lilbume's assertions in his
later trials when he was faced with questions on treason and related charges. 'Wigmore
claimed that the privilege was well established by the late seventeenth century. 8 WioMoRE,
supranote 154, § 2250, at 298-99. Now, however, the work of several historians suggests that
wigmore's evidence was flawed. See, eg., Langbein, supra note 2, at 1071-84.
184. Langbein, supra note 2, at 1065-66; Eben Moglen, Taking the Fifth: Reconsidering
the Origins of the ConstitutionalPrivilege Against Self-Incrimination, 92 Mscn. L. REv. 1086
(1994); see also J.M. BnArnE, CRmIE AND Tan CouRrs iN ENGLAND 1600-1800, at 364-66
(1986).
185. Langbein, supra note 2, at 1066-71.
186. See Moglen, supranote 184, at 1094-104; Morgan, supra note 179, at 18-19.
draw. 8 7 In the mid-nineteenth century, police departments

emerged and began to take over certain pretrial investigatory func-
tions, relying on informal coercion. The distinction between infor-
mal confession at the hands of the police and compelled self-
incrimination during formal proceedings was maintained until the
1897 case of Bran 8 8 (briefly) and the 1966 case of Miranda (per-
manently) merged the two.
Amid this tangled history and unconvincing catalogue of tradi-
tional rationales for the clause, three key questions about the scope
of the clause stand out. First, why does the privilege apply in a
"criminal case" but not in a civil case? Second, why does the privi-
lege simply dissolve once immunity is granted, whereas no true pri-
vacy privilege - spousal, priest-penitent, doctor-patient, attorney-
client, and the like - likewise dissolves in the presence of immu-
nity? Third, why does the Fifth Amendment bar only compulsory
"witness[ing]" rather than "furnishing evidence"? Any adequate
theory must explain these three central features of the clause. No
current theory does. 189
II. THE SOLUTION
Current Fifth Amendment doctrine is a quagmire. But there is

a way out - a road to firm, high ground. To clear up the confusion,
we advocate a solution remarkably like the early scope of the privi-
lege, one that borrows a page from current pretrial civil discovery.
Under our solution, the government would be able to compel all
persons to testify truthfully in a wide variety of proceedings before
the commencement of, or outside, a formal "criminal case" or trial.
These venues include grand jury rooms, legislative hearings, civil
cases, criminal cases in which someone else is on trial, and deposi-
tions organized by prosecutors. The penalty for refusing to answer
would be contempt, 90 and the penalty for lying would be per-
187. See OFFscE OF LEGAL PoUcy, U.S. DEr. OF JusTicE, "TRUTH IN CRIMINAL JUS.
rtcE" SERIES REPORT No. 1, Tmi LAW OF PRE-TRIAL INTERROGATION (1986), reprinted in 22
U. Mxcff. J.L. Rm. 437, 482 (1989).
188. Brain v. United States, 168 U.S. 532 (1897) (holding an involuntary confession made
to a police official inadmissible).
189. Professor Stuntz claims that his descriptive theory can explain the pattern of current
cases. See supra note 174. Whether or not this is so, Stuntz's theory - that one can excusa-
bly lie to cover up earlier crimes one has committed - seems historically implausible and
morally unattractive. See supranote 174. On both these grounds, our theory is superior, and
it also better fits the Court's recent trends. See infra Part II.
190. A government could choose a less compulsive scheme, of course, such as one that
"compelled" answers by allowing a later jury or fact finder to draw adverse inferences from a
suspect's pretrial silence in the face of pointed questions. Cf. supra section I.B.
jury. 191 In the case of criminal depositions, this compulsion would

take place under judicial oversight, as in the civil discovery system,
but with extra protection: the magistrate would physically preside,
rather than oversee by remote control. Defense and prosecution
lawyers and a court reporter would typically be present.- 92 In high-
profile cases, and perhaps ordinary cases as well, the magistrate
could preserve a grand-jury-like secrecy by closing the hearing to
the public and sealing the deposition.193 The prosecution would not
be able to introduce compelled out-of-court testimony "in" a "crim-
inal case" unless the defendant knowingly and intelligently con-
sented at trial, with the judge looking on. But virtually all physical
evidence and third-party testimony that the defendant's statement
191. If a defendant can be shown to have lied at his deposition, his words - or the fact
that he lied - still could not be introduced at his trial for the underlying offense. His words
could be introduced, however, in a subsequent prosecution for perjury. This caveat is neces-
sary because otherwise defendants could effectively render the entire deposition process
worthless simply by lying. The Supreme Court has adopted a similar rule in holding that
immunized grand jury testimony cannot be used for impeachment purposes at the witness's
trial involving the matter about which he testified, see New Jersey v. Portash, 440 U.S. 450
(1979); Mincey v. Arizona, 437 U.S. 385 (1978), but it can be used against the witness in a
later perjury prosecution, see United States v. Apfelbaum, 445 U.S. 115,127-32 (1980); Glick-
stein v. United States, 222 U.S. 139, 141-42 (1911). In effect, a pretrial deposition helps
freeze and lock in a suspect's story, and - via the threat of perjury charges - deters post-
hoc concoctions. Under our proposal, perjury would become a much more significant
weapon in the fight against crime - just as the scope of current self-incrimination doctrine
creates strong incentives for wiretaps and sting operations. See supra text accompanying
notes 4-5. In a case where a suspect lies, a prosecutor may be able to prove perjury even if
she cannot prove the predicate offense. If a murder suspect lies about his whereabouts at the
time of the murder, for example, a prosecutor may be able to prove he is a liar, even if she
cannot prove he is a murderer. In response to our scheme, legislatures might well choose to
boost the penalties for perjury.
192. Defense lawyers have traditionally been excluded from the grand jury room, and we
would not require their presence there. Unlike a police-station interrogation, there is little
risk of violence against a witness, and so a lawyer is not needed on that account. It is in fact
quite useful for society to have at least one nonviolent but secret interrogation place, so that
individual members of organized conspiracies can be brought in one by one, and their part-
ners in crime will never know for sure whether they ratted or stood mute. (The secrecy of the
grand jury room is in effect the wall between prisoners that creates a classic "prisoner's di-
lemma" to confess the truth.) Defense lawyers, if present in the grand jury room, can actu-
ally prevent an underling who wants to tell all from squealing, for the underling's lawyer may
really be the agent of the mob boss; such lawyers often help a group of conspirators to main-
tain a joint stonewall defense. Cf. Pamela S. Karlan, Discreteand Relational CriminalRepre-
sentation: The Changing Vision of the Right to Counse 105 HARv. L. REv. 670, 693-97
(1992). Even in the case of a lone criminal, a defense attorney may at times impede rather
than promote the truth-seeking process. See Stuntz, supra note 60, at 1944-54. Because a
lawyer-less witness in a grand jury room may be tricked or intimidated by a clever prosecutor
into making misleading or inaccurate inculpatory statements, we would not allow these state-
ments themselves to be admitted in a criminal case, unless the defendant so authorized.
193. Preserving secrecy in the magistrate hearing would protect the witness from having
to make potentially embarrassing public revelations and would prevent potential jurors in the
public from being tainted by preliminary exposure to excludable testimony.
led to would be admissible. 194 The defendant could also be subpoe-

naed to provide anything in his possession, except possibly his inti-
mate personal papers, upon pain of contempt. In a criminal trial
the subpoenaed items could be introduced, but the fact that the de-
fendant produced them could not.195
The textual argument is remarkably clean. A defendant cannot
be forced to be a "witness against himself" - to testify, with his
own words introduced against him - at trial "in [his own] criminal
case." Witness here is used in its natural sense, meaning someone
whose testimony, or utterances, are introduced at trial. Witnesses
are those who take the stand and testify, or whose out-of-court dep-
ositions or affidavits are introduced at trial in front of the jury. In-
deed, this is exactly how the word witness seems to be used in the
companion Sixth Amendment and in the Treason Clause of Article
1]1.196 Physical evidence, on the other hand, can be introduced at
trial whatever its source - even if that source is a compelled pre-
trial utterance. A witness testifies but physical evidence does not.
A thing is not a witness. Moreover, if person A takes the stand to
testify against defendant B, this is not the same as forcing B to be a
witness againsthimself, even if B's compelled pretrial statement led
the police to learn of A's existence and information. Although our
suspect has indeed been forced to testify pretrial, that testimony
occurred outside his own criminal case, beyond the earshot of the
jury. Unless these words are introduced at trial, a suspect is not a
"witness" against himself "in" a criminal "case."
This clean reading makes sense of all the words of the Self-
Incrimination Clause and shows how they fit together (and with
kindred words of the Sixth Amendment), but it also does much
more than that. It flushes out the heretofore elusive rationale of
the Self-Incrimination Clause, as best read: reliability. Compelled
testimony may be partly or wholly misleading and unreliable; even
194. An additional possibility is to allow a polygraph test to be conducted on the defend-

ant, with the proviso that the results of the test would not be admissible at trial. Polygraph
tests, while less helpful in the employment context, have been shown to be of some help in
the context of criminal investigations in enabling the police to decide which trails to follow.
OFFICE OF TECHNOLOGY AssEssMENT, U.S. CONGRESS, ScIENTIFc VALIDITY OF POLY-
GRAPH TEsTNG 8, 58 (1983).
195. This last restriction obviates the Court's concern in Fisher,Doe, and Bouknight that
the act of producing documents or objects itself has testimonial value. Cf. supra note 142;
infra note 276.
196. The Sixth Amendment provides the accused, in all criminal prosecutions, with the
right, inter alia, "to be confronted with the witnesses against him" and "to have compulsory
process for obtaining witnesses in his favor." U.S. CONsT. amend. VI. The treason provision
provides that "[n]o Person shall be convicted of Treason unless on the Testimony of two
Witnesses to the same overt Act." U.S. CoNsT. art. III, § 3.
an innocent person may say seemingly inculpatory things under

pressure and suspicion and when flustered by trained inquisitors.
But physical fruit is far more sturdy and reliable evidence, so it
should be brought before the jury. Of course, government investi-
gations and inquisitions implicate issues beyond reliability - of
bodily autonomy, of personal privacy and dignity, of freedom of
thought and conscience, and so on - but as we shall see, these
concerns are best addressed via other constitutional clauses and
principles outside the self-incrimination idea.
Our clean reading also solves many of the common-sense
conundra plaguing current doctrine. Indeed, our reading achieves
almost the same results that Judge Friendly argued for in his wise
and influential lectures on the Self-Incrimination Clause. 197 But
whereas Friendly's elaborate analysis led him to propose a six-
pronged constitutional amendment of over three hundred words to
achieve his sensible results, we reach largely the same conclusions
through a more elegant and parsimonious textual analysis of the
fifteen words of the existing clause.
A. "Person"
Under our reading of the clause, the problem of the innocent
defendant noted earlier would be solved. There would be no war
between the Fifth Amendment privilege and the Sixth Amendment
right to compulsory process to obtain witnesses. An innocent de-
fendant could use "compulsory process" (Sixth Amendment) to
"compel" (Fifth Amendment) the guilty party to be a "witness"
(Fifth and Sixth Amendments) against himself, upon pain of con-
tempt and under penalty of perjury. In short, the guilty could be
made to sing and to be subject to withering questioning by defense
counsel. The guilty witness's compelled testimony could not be in-
troduced against him in any subsequent criminal trial, unless he au-
thorized admission at trial. But officials would otherwise be free to
prosecute, even on the basis of leads (fruits) generated by the com-
pelled testimony.
The Canadians already have a similar rule in cases where the
defense calls a witness who claims he might incriminate himself by
answering. The witness is required to testify, 198 but the prosecution
197. See Friendly, supranote 63.
198. Canada Evidence Act, R.S.C., ch. E-10, § 5(1) (1970) (requiring that the witness
answer); R.S.C., ch. E-10, § 5(2) (1970) (prohibiting the prosecution from using testimonial
admissions of a witness who objects that he will incriminate himself). A coparticipant must
also testify, if he is charged separately from the defendant. Re Regan, 2 D.L.R. 135, 137-38
(N.S. 1939).
cannot use his testimonial admissions - the words themselves -

against him in a future proceeding. However, Canada does allow
the prosecution to use the fruits of his testimony against him. 199
Some might claim there is a textual problem with this interpre-
tation because the Fifth Amendment says "in any criminal case. '200
But the key "person" this clause is designed to protect is the "per-
son" on trial - the defendant in his own criminal case. A witness
required to testify at someone else's criminal trial with testimonial
immunity is in basically the same position as a suspect required to
give an immunized deposition for his own trial or to testify in a civil
case, in a legislative hearing, before a grand jury, or in any other
proceeding. 201
This testimonial immunity has several advantages over other
sorts. Transactional immunity would thwart legitimate law enforce-
ment against the guilty witness, and as a practical matter, so could
use plus use-fruits immunity. Under the latter regime, prosecutors
would need to prove that they had obtained evidence from in-
dependent sources, and, as we have seen, the costs of such a system
in terms of lost evidence and administrative burdens can be high.
Testimonial immunity, by contrast, would in no way hinder enforce-
ment efforts against the immunized witness: the government loses
nothing that it already had or might get independently. On the con-
trary, by immunizing the witness, an innocent defendant would be
helping the police, generating leads that the police would be free to
follow to convict the real culprit. The government therefore would
have no excuse for denying the innocent defendant's constitutional
right of compulsory process.
Because previous courts defined the Fifth Amendment too
broadly, they ended up betraying the explicit Sixth Amendment
right to compel witnesses. If the Fifth Amendment is restored to its
proper scope, a defendant would indeed enjoy a kind of Sixth
Amendment parity with the prosecutor. Both could subpoena;
both could immunize. This Sixth Amendment right is absolutely
essential, for it is truly at the heart of our criminal procedure. It is
199. See REPORT OF THE FEDERALIPROVINCIAL TASK FORCE ON UNIFORM RuL.ES OF
EVIDENCE § 33.3(b) (1982) (considering and rejecting proposals granting an indicted witness
use-fruits immunity); see also Tague, supra note 7, at 3 n.11.
200. See e.g., Friendly, supra note 63, at 676.
201. Current doctrine, of course, recognizes this too. A person can never be compelled,
upon pain of contempt, to witness in his own criminal trial, but he can be obliged - with
immunity - to witness in someone else's trial, civil or criminal. This immunity is enforced
by a rule of exclusion in his own case. See supra notes 63-64 and accompanying text. The
only question is how broad that immunity must be - exactly what must be excluded in his
own criminal case.
nothing less than the right to mount a defense - a right of obvious

importance to all defendants and of transcendent significance for
truly innocent defendants. 202 As the Supreme Court put the point
203
in its landmark case, Washington v. Texas:
The right to offer the testimony of witnesses, and to compel their at-
tendance, if necessary, is in plain terms the right to present a defense,
the right to present the defendant's version of the facts as well as the
prosecution's to the jury so it may decide where the truth lies.2 ° 4
Some might argue that the' right to compel a guilty witness to
take the stand would not benefit the innocent defendant very much.
The witness would simply lie. Those who framed the Fifth Amend-
ment, however, thought differently. Earlier generations believed
that perjury was a mortal sin, resulting in eternal damnation: better
to admit murder than commit perjury under oath.205 The power of
oaths several centuries ago is abundantly clear from the Constitu-
tion itself, which requires oaths in two of its most important provi-
sions, 20 6 and from landmark opinions of the Marshall Court
stressing oaths. 207
True, times have changed. Perjury has largely lost its religious
connotations and is feared mainly for its possible secular sanction.
Even today, however, perjury will sometimes be a more severe pe-
nal threat than the underlying crime, and if our solution were
adopted, legislatures might choose to increase the maximum pun-
ishment for perjury. What's more, as noted above, the witness need
not confess all on the stand for his testimony to be vital to an inno-
cent defendant. Simply forcing a witness to tell his story can be
invaluable. The innocent defendant then would have the opportu-
nity to demolish it with questions and other evidence that might
otherwise be excluded for lack of foundation. Prosecutors today,
for example, get considerable mileage out of a suspect's lies. Alibis,
denials, and explanations "can be checked and rechecked until a
'208
suspect's lies are the greatest evidentiary threat to his freedom.
202. See generally Westen, supra note 10, at 182-84.

203. 388 U.S. 14 (1967).
205. See BARLow,supranote 154, at 189 (stating that oaths "might serve instead of the
Rack, to the Consciences of some Men, although they have been guilty of Offenses"), quoted
in Langbein, supra note 2, at 1085 n.157.
206. See U.S. CONsT. art. II (elaborating the presidential oath); U.S. CONsT. art. VI, cl. 3
(requiring officeholders to take an "Oath or Affirmation" to support the Constitution).
207. See Marbury v. Madison, 5 U.S. (1 Cranch) 137, 180 (1803) (stressing the judicial
oath); McCulloch v. Maryland, 17 U.S. (4 Wheat.) 316, 416 (1819) (authorizing Congress to
add new oaths).
208. SIMON, supra note 57, at 198.
B. "Compelled"
1. In-Court Compulsion
Compared to the current morass, the compulsion line in our so-

lution would be easier to enforce. Pretrial, a suspect under our
scheme must comply with all judicially authorized depositions, in-
quests, and subpoenas. If he does not, he could be made much
worse off: held in contempt and punished. At trial, the judge
would exclude the defendant's compelled pretrial testimony, unless
the defendant affirmatively authorizes its introduction as evidence.
And for the reasons we identified above, courts could continue to
follow the nice compulsion rule laid down in Griffin concerning in-
court inferences about in-court silence. 209 The Self-Incrimination
Clause, as best read, is designed to protect a truly innocent defend-
ant who might be made to look guilty on the stand by a clever pros-
ecutor skilled in technical courtroom procedure and forensics. To
infer guilt from mere in-court silence would seem to betray the in-
nocent but unpersuasive defendant whom the clause seeks to
0
protect.21
Prosecutors could probably encourage the defendant to take the
stand at trial or to authorize introduction of pretrial statements with
bribes, such as recommendations of reduced jail time. The baseline
of compulsion would be established by the burden of proof rules at
trial, the definition of elements of the offense, the statutory sen-
tence authorized, and the Griffin rule itself. These would fix a
Coasean starting point from which defendant and prosecutor could
bargain.211 In contrast to the situation at the Founding,2 12 today the
Court has recognized the defendant's right to testify under oath; the
innocent defendant is able to calculate the risk that his testimony
will be misinterpreted, and this screening will help make the testi-
mony that the defendant voluntarily authorizes to be admitted in
court more reliable.
209. See supra section I.B. But see OFn OF LEGAL Poucy, U.S. DEPT. OF JUSTICE,
REPORT TO THE ATroRNY GENERAL ON ADVERSE INREN E FROM SILENCE, No. 8
(1989), reprinted in 22 U. MicH. J.L. RrF.1005, 1078-81 (1989).
210. See infra section II.E.
211. This in-court inducement to testify differs arguably from much out-of-court compul-
sion by the police, in that the prosecutor is merely trying to persuade the defendant to testify
- simply to take the stand - while the police often encourage a suspect to testify against
himself- to confess, to provide a certain substantive slant. The latter leads to more reliabil-
ity difficulties. It is useful here to note that a defendant's testimony is voluntary even when it
is shaped by various strategic considerations.
212. See supra text accompanying note 153.
Under this logic, the "silence penalty" imposed by the Sentenc-

ing Guidelines should pass constitutional muster. Griffin is neatly
distinguishable: there is a world of difference between using (then-
unreliable) silence to prove guilt before guilt has been indepen-
dently established and using silence once guilt has already been
otherwise proved. Put another way, by the time the Guidelines are
applied, they penalize moral culpability on the underlying offense,
rather than silence per se. If the Guidelines' common-sense moral-
ity is to be upheld on this theory, however, the logic of the Self-
Incrimination Clause must sound in reliability, as we explain in
more detail below. 213 After a successful appeal, a defendant may
have a right to insist that his earlier-induced statements be excluded
from the guilt-determination phase of the retrial - because, in law,
he reverts back to presumptive innocence and Griffin thus applies.
(In effect, his earlier-induced testimony is protected by our rule of
testimonial immunity.) But in this second trial, as in the first, the
Sentencing Guidelines' silence penalty remains in effect, constitut-
ing a lawful Coasean bribe to induce the defendant to (again) waive
his "starting point" right of silence.
2. Out-of-Court Compulsion: Government Employment

Currently, the government, when acting as an employer, is
caught between a rock and a hard place. It must choose either to
act sensibly as an employer 214 or to act efficiently in its sovereign
capacity as law enforcer. 215 Instead, the government should be put
in the same position as any other employer. It should be able to
fire stonewalling employees and otherwise to act as a normal em-
ployer, without detriment to its law enforcement function. The
only way to guarantee no detriment is to allow all fruit in, thereby
avoiding the burdens of Chinese walls, "canning," North-style ac-
quittals, and so on.
But what about the out-of-court testimony itself? When could it
come in? The issue here is cloudier - and the stakes are much
lower once fruits may come in - but one clear rule would be that
the limits on the power of the government-as-employer over its em-
ployees should be the limits on the power of normal employers.
(Private standards of coercion, such as laws against extortion,
213. See infra section II.E.
214. It could do so by immunizing its employees, making them sing, and then taking ap-
propriate employment action in light of their song.
215. It could do so by refusing to confer sweeping immunity that might compromise later
criminal prosecutions.
would apply.) Under this theory, courts could hold that no imper-
missible Fifth Amendment compulsion exists from reasonable em-
ployment decisions because no one is compelled to work for the
government in the first place. If statements made under threat of
dismissal are reliable enough to come in when dismissal is
threatened by a private employer, why not for a public employer
too? 216 (Note that in asking the "How much pressure is too much
217
pressure?" question, we now have a touchstone: reliability. )
Any action beyond these limits should be construed as impermissi-
ble coercion, with the result that the coerced statement itself - but
not the fruits - would be excludable. This, in effect, would treat
coercion beyond responsible employment sanctions exactly the
same as coercion in depositions, hearings, and the like. Deposition
coercion is backed by contempt and imprisonment - powers no
private party enjoys - and so the coercion in depositions is clearly
the coercion of a "sovereign." The same should be true when a
government employer acts more coercively than a reasonable pri-
vate employer would; in effect, leveraging its sovereign power'218
in
the employment context is an "unconstitutional condition.
3. Out-of-Court Compulsion: Required Records

In required records cases the Court has adopted a balancing ap-
proach to the privilege, unwarranted by constitutional text, in order
to avoid the substantial burdens imposed by, and the broad logic
underlying, Kastigar'suse plus use-fruits immunity. That form of
immunity would have practically foreclosed any prosecution of
Bouknight for child abuse 219 and prevented the police from discov-
ering the identity of, and therefore prosecuting, the hit-and-run
216. Cf Friendly, supra note 63, at 707-08 (drawing an analogy between public and pri-
vate employers).
217. See infra section H.E.
218. The argument here is not that government employment action is not state action,
nor that mimicking the market can never violate constitutional provisions, such as the First
Amendment. Rather, it is that market mimicry should not be understood as the kind of
compulsion that offends the letter or spirit of the Self-Incrimination Clause. This conclusion
is not the product of unreflective labels about "natural baselines" distinguishing between
government as "sovereign" law enforcer and government as "mere" employer, but instead it
follows Professor Sunstein's invitation to reflect self-consciously on the "baseline" that best
vindicates a particular provision's purposes. See Cass R. Sunstein, Lochner's Legacy, 87
COLUM. L. Rv. 873 (1987); Cass R. Sunstein, Why the UnconstitutionalConditionsDoctrine
is an Anachronism (With ParticularReference to Religion, Speech, and Abortion), 70 B.U. L.
Rv. 593 (1990). For an earlier exposition that makes very similar points, see Westen, In-
credible Dilemmas, supra note 15.
219. See Baltimore City Dept. of Social Servs. v. Bouknight, 493 U.S. 549 (1990).
driver in Byers.220 The Court responded by creating exceptions to

the privilege in the Shapiro-Byers-Bouknightline of cases. But the
need for such exceptions shrinks dramatically if one gets at the root
of the problem: the scope of immunity.
Under our solution, required records, just like any other com-
pelled, incriminating statement, would generally receive full Fifth
Amendment protection. They themselves could not be introduced
"in" a "criminal case" - that is, at trial - but any fruits they led to
could be introduced. 221 Thus, the body of the child in a case like
Bouknight would have to be produced and could be introduced in
the trial of his abuser, but not the fact that the defendant's informa-
tion led to him. The hit-and-run driver could still be prosecuted.
The quandary the Court currently faces - possibly making the gov-
ernment worse off as criminal law enforcer whenever it needs
records for legitimate noncriminal purposes - would thus be ele-
gantly solved.
220. See California v. Byers, 402 U.S. 424 (1971); see also Shapiro v. United States, 335
U.S. 1 (1948); supra section I.B.3.
221. This interpretation of the Fifth Amendment arrives at Judge Friendly's solution
without the need for a constitutional amendment. Friendly advocated "[r]equiring registra-
tion or reporting reasonably necessary for a proper governmental purpose, provided that no
registration or report so compelled shall be admissible as evidence of any crime revealed
therein." Friendly, supra note 63, at 722. But, he argued, "the government should not be
forced to show it would have been able to prosecute quite apart from the information fur-
nished by the registration or report or be prohibited from using leads obtained therefrom."
Id. at 720.
222. Even with this quandary solved, however, the required records doctrine still presents
thorny issues. Sometimes, the government will need to introduce the records themselves to
get a conviction. At least two theories are plausible here. First, if a required record is of a
type a person would have kept anyway, the government could make a kind of inevitable
discovery argument as follows: the defendant would have kept the record voluntarily, so the
government did not really "cause" or "compel" its creation, and it should be subject to sub-
poena under Fisher v. United States, 425 U.S. 391 (1976), with immunity provided merely for
the testimonial act of compelled production itself. See infra text accompanying notes 275-76.
This theory builds on Justice Harlan's first prong in Marchetti v. United States, 390 U.S. 39
(1968), discussed supra in section I.B.3, but tries to give it more rigor. (Even if a new,
noncustomary record is required as a condition of doing a certain kind of business, perhaps
the government could also argue here that no real Fifth Amendment compulsion exists, be-
cause a person is free not to engage in that business. Cf. supra section II.B.2.) Second, it
might be argued that certain kinds of records, required of broad classes of persons not sus-
pected of criminal wrongdoing, and not involving face-to-face encounters with interrogators,
need not be seen as akin to criminal "witnessing" even though these records are testimonial
and ultimately introduced in a criminal case. This theory repackages Justice Harlan's third
Marchetti prong as a textual argument about "witness" and reflects the Big Idea that at the
Founding compelled criminal witnessing would often be unreliable because of the imbalance
of power, interruptions, traps, and the like in face-to-face exchanges between citizens and
prosecutors. See infra section II.E. We do not necessarily embrace either theory here; their
elaboration must await another day.
Note finally that corporate required records have long been given only minimal Fifth
Amendment protection. See Hale v. Henkel, 201 U.S. 43, 75 (1906) ("While an individual
may lawfully refuse to answer incriminating questions unless protected by an immunity stat-
4. Out-of-Court Compulsion: Police Interrogation

Our deposition approach would limit abusive police tactics. The
basic insight uniting pre-Warren Court voluntariness cases like
Brown v. Mississipp223 and Warren-era landmarks like Miranda22 4
and Escobedo225 would be preserved and strengthened: we need to
rein in unsupervised police officers who might be tempted to abuse
suspects. The best way to do this is to shift interrogation from po-
lice stations to magistrates' hearing rooms. On the civil discovery
side, we generally do not try to get damning admissions from de-
fendants by using third-degree tactics. We use subpoenas and dep-
ositions with lawyers, loosely supervised by judges. 226 Miranda
failed to require the use of lawyers, magistrates, and recorders -
and from a civil libertarian perspective, this has been its undoing.
Miranda also failed to create strong incentives for suspects to talk
and to tell the truth - and from a crime control perspective, this
has been its undoing. The deposition model would combine both
perspectives: the suspect would be protected from abuse and intim-
idation but must answer truthfully.
The question of how to treat confessions that are obtained by
police outside the formal deposition-like process remains.22 7 Fortu-
nately, our solution would make the problem much less pressing
because the stakes would shrink: a civilized process would exist
outside the police station to compel suspects to talk truthfully, and
so the police would be less tempted to force the issue. In light of
this civilized alternative, courts might well choose to police the po-
ute, it does not follow that a corporation, vested with special privileges and franchises, may
refuse to show its hand when charged with an abuse of such privileges.").
223. 297 U.S. 278 (1936).
224. Miranda v. Arizona, 384 U.S. 436 (1966).
225. Escobedo v. Illinois, 378 U.S. 478 (1964).
226. In civil discovery, magistrates are not typically present during depositions but over-
see the general deposition process. In our proposal, magistrates would generally be physi-
cally present for criminal depositions of suspects. See supra text accompanying notes 192-95.
227. Rogue police behavior will call for special remedies, including punitive damages
against the police department and strict administrative disciplinary mechanisms to punish
abusive officers. An appropriate remedial regime is sketched in Amar, supra note **,at 811-
16. Neither the Fourth Amendment, nor the Fifth, properly construed, requires exclusion of
reliable fruits of unreasonable seizures of persons. Moreover, since, in theory, a suspect can
be lawfully obliged to truthfully tell all to a magistrate under our approach, the fruits should
have come to light anyway and thus would fall under the "inevitable discovery" doctrine, see
infra section II.C. Legislatures, of course, would nonetheless be free to require exclusion if
they were determined to "teach the cops a lesson." But since abusive cops must be punished
and deterred even when they expect to find no evidence, and (unsurprisingly) find no evi-
dence, exclusion is not constitutionally sufficient. A proper punitive damage and administra-
tive disciplinary scheme, by contrast, is both constitutionally necessary and constitutionally
sufficient - it protects the innocent but avoids windfalls for the guilty.
lice even more strictly than today, enforcing a prophylactic rule that
no police-station confession by a defendant is ever allowed in, un-
less volunteered by a suspect in the presence of an on-duty defense
lawyer or ombudsman in the police station, or unless the defendant
consents to its introduction at trial. Rooted in a legitimate concern
about unsupervised police compulsion, this strict regime would cre-
ate powerful incentives to conduct interrogation before magistrates
rather than in police stations.
On the other hand, more relaxed schemes are also compatible
with our approach. For example, each suspect in custody could be
told that he must be brought before a magistrate and a lawyer
within a short time (say, five hours) and that he has an absolute
right to remain silent until then; but he should also understand that
if he stands mute until then, a later jury can be told of his pre-
magistrate, pre-lawyer silence and might view more skeptically any
story he later tries to offer at trial. The range of possible police
station schemes compatible with our model is hardly unique or em-
barrassing. Rather, it reflects the fact that the Self-Incrimination
Clause historically addressed formal testimonial compulsion in judi-
cial settings, and so applying the clause to the informal compulsion
of the modern police station requires creative adaptation of Found-
ing principles.
And however we regulate the police station itself, nice problems
will arise concerning suspects' statements and silences before sta-
tionhouse custody commences - at the scene of the crime, on the
street corner, in the squad car, and elsewhere. The wide variety of
these encounters may well mock any effort at comprehensive
rulemaking. But, in keeping with our treatment of employment
compulsion, we propose one global rule of compulsion: reasonable
adverse inferences from suspicious silence outside courtrooms need
not always be treated as Fifth Amendment "compulsion." 228
C. "In Any Criminal Case"
Textually, the Fifth Amendment speaks to witnessing within the

criminal case, not beyond. Therefore, the key question is what
"witnessing" is excludable "in" a "criminal case" - that is, at
228. Again, silence in the face of some kinds of informal accusation may be far more
suspicious than silence in certain formal, legalistic, forensic judicial settings. See supra notes
23, 31. But see supra note 211. See also Jenkins v. Anderson, 447 U.S. 231, 243 (1980) (Ste-
vens, J.,concurring) (stating that jurors may draw a "reasonable inference from [suspicious
pre-arrest] silence in a situation in which the ordinary citizen would normally speak out").
910 Michigan Law Review [Vol. 93:857
trial.22 9 This question is the same as what the scope of immunity
229. Pretrial proceedings are not best read as included within a Self-Incrimination Clause
"case." The clause is concerned with a "witness" in a "case" who in effect testifies before the
jury. Miranda does not hold otherwise. To be sure, it holds that the Self-Incrimination
Clause applies to pretrial police compulsion, but only because the clause applies to all sorts
of compulsion outside a criminal case. See supra note 63. The rule of exclusion, however,
exists within a criminal case, see id., and Miranda does not hold that the police station is
somehow a courtroom. Escobedo does contain language linking police stations and court-
room rights, but later Supreme Court cases broke with Escobedo here, and Escobedo's lan-
guage, read literally, is hard to take seriously. See Henry J. Friendly, The Bill of Rights as a
Code of CriminalProcedure 53 CAL L. REv. 929, 946-47 (1965). A police station is not a
courtroom, and nothing is tried there. There is no judge, no jury, no public, no confrontation
right, and no compulsory process right; and if a right of counsel exists in the police station, it
is best understood as a due process right, not a Sixth Amendment right. See id at 944-46; see
also United States v. Gouveia, 467 U.S. 180, 187-88 (1984) (holding that the Sixth Amend-
ment right to counsel does not attach until the initiation of formal adversary criminal pro-
ceedings); Moran v. Burbine, 475 U.S. 412,428-30 (1986) (following Gouveia and finding no
Sixth Amendment right to counsel in a police station).
Nor does the grand jury proceeding fall within a Self-Incrimination Clause "case." lypi-
cally, a grand jury sits before an indictment has issued and often before a clear suspect has
emerged. Historically, persons were obliged to tell all in a grand jury hearing, with testimo-
nial immunity enforced only in a later criminal trial. See infra text accompanying notes 234-
35. In Article III, the word case refers not to a factual transaction but to formal legal pro-
ceedings with parties to a lawsuit: a "case" begins when a prosecutor or plaintiff files an
indictment or complaint. Indeed, the word case is linked to the word cause, as in a formal
legal "cause of action." See Akhil Reed Amar, Law Story, 102 HARv. L. Rav. 688,718 n.154
(1988). Functionally, it does not make much sense to bar postindictment depositions while
allowing preindictment hearings, because indictments could often be postponed or dismissed
in order to conduct the desired depositions. The key exclusion is thus exclusion of testimony
from the trial, where the jury sits to find the facts.
This notion of what a Self-Incrimination Clause "case" is fits with the way we apply rules
of evidence, such as hearsay. Rules of evidence, of course, apply at trial - but they do not
apply in pretrial proceedings such as depositions and grand jury hearings. See FLEMNGN
JAMES, JR., ET AL, CiVIL PROCEDURE § 5.3, at 238 (4th ed. 1992) (stating that at a deposition
the matter inquired into "need not itself be admissible evidence"); see also Costello v. United
States, 350 U.S. 359 (1956) (refusing to quash an indictment based exclusively on hearsay
testimony). Like the privilege against compelled self-incrimination, rules of evidence such as
hearsay are meant to improve reliability. Reliability of individual bits of information is criti-
cal at trial, where final decisions are made, but not so critical where the goal is simply to
gather as much relevant information as possible before sifting, as in pretrial proceedings.
The different burdens of proof at the pretrial and trial stages - probable cause for an indict-
ment as opposed to proof beyond a reasonable doubt for a criminal conviction - lead to
differences in the need for rules emphasizing reliability. See infra section II.E; cf. United
States v. Williams, 112 S. Ct. 1735, 1743 (1992) (suggesting that evidence previously obtained
"in violation of the privilege against self-incrimination" may be used in the grand jury, but
not in a criminal trial, a suggestion that would be hard to sustain if the grand jury were itself a
"case" within the meaning of the Self-Incrimination Clause).
Note finally that the question is not whether the word case must mean "at trial but not
before" but whether it most sensibly should mean this to achieve maximum textual coher-
ence, structural harmony, common sense, and so on. Our reading of the word case enables
the words of the Self-Incrimination Clause to fit together and make good policy sense; it
coheres with the idea of "witness" "in" a "case"; it fits with the cognate words and principle
of the Sixth Amendment, which is about "witnesses" at trial (there is no right to confront
grand jury witnesses or those who give investigators pretrial statements that are never intro-
duced at trial); it meshes tolerably with the wording of Article III; and it draws support from
American history. See infra text accompanying notes 233-49. Overall, this is more than
other readings of case can do, especially when unsupported by a clear and coherent theory of
the overall meaning and purpose of the Self-Incrimination Clause.
should be. The Kastigar23o rule for "true" Fifth Amendment viola-
tions - use plus use-fruits immunity - should be trimmed back
and brought into line with Justice O'Connor's suggested approach
in Quarles231 for "mere" Mirandaviolations. Compelled testimony
should be excluded from a criminal case - unless the defendant
authorizes its introduction at trial - but not fruits. A rule exclud-
ing compelled testimony would be much easier to enforce than the
Kastigarrule excluding fruits - a rule that has led to many difficult
issues of proof, but-for causation, and never-neverland counterfac-
tuals, as in the North case. A statement by the defendant is self-
identifying, but other physical evidence and witnesses are not: they
do not come with "courtesy of defendant" labels stating how the
government got them or whether the chain of causation ran through
the defendant's pretrial statements. The recurrent friction between
different government authorities with the power to grant immunity
(states, federal prosecutors, Congress, and so on) that led the Court
in Murphy232 to move toward Kastigarwould be eliminated. If gov-
ernment A grants immunity, government B would be no worse off.
Currently, A can to some extent veto B's law enforcement by im-
posing North-like litigation burdens on B requiring it to prove it got
all its evidence independently. So the very same internal logic that
drove the Court in Murphy from Counselman to proto-Kastigar
should now drive it all the way home to a Quarles-like rule.
Testimonial immunity also has some important history on its
side; it is consistent with many courts' view of the privilege before
the Counselman decision in 1892. Prior to Counselman, case after
case in state after state and in lower federal courts held that the
privilege was satisfied by excluding compelled pretrial testimony
but not the fruits of that testimony.233 Before 1892, the leading
American case was People v. Kelly,234 decided by New York's high-
230. Kastigar v. United States, 406 U.S. 441 (1972), discussed in supra section I.C.
231. New York v. Quarles, 467 U.S. 649, 660-74 (1984) (O'Connor, J., concurring in part
and dissenting in part), discussed in supra section I.C.
232. Murphy v. Waterfront Commn., 378 U.S. 52 (1964), discussed in supra section I.C.
233. See New York ex reL Hackley v. Kelly, 24 N.Y. 74 (1861); infra note 238 (citing
cases). Early manuals for justices of the peace also stressed testimonial immunity for pretrial
examinations, which were routine. A Georgia manual announced: "No man shall be com-
pelled to give evidence against himself. Hence it is that if a criminal be sworn to his examina-
tion taken before a justice, it shall not be read against him." AUGUsTIN S. CLAYTON, THE
OFa-cE AND Dur, OF A JusTcn OF TmE PEAcE 132 (Milledgeville, Ga., S. Grantland 1819)
(emphasis added).
234. New York ex rel. Hackley v. Kelly, 24 N.Y. 74 (1861). Early federal cases shed little
light on the scope of the privilege. The most often cited is the opinion of Chief Justice Mar-
shall in United States v. Burr, 25 F. Cas. 38 (C.C.D. Va. 1807) (No. 14,692e). In Aaron Burr's
trial for high treason, the prosecution sought to examine Burr's secretary, a Mr. Willie, about
est court in 1861. Kelly involved a witness called by a grand jury

and asked to answer incriminating questions. The court held that as
long as the compelled testimony was excluded from any subsequent
criminal trial against the witness, the privilege was satisfied.23 5 In
addressing the argument-that the privilege required broader immu-
nity, the court announced:
[N]either the law nor the Constitution is so sedulous to screen the
guilty as the argument supposes. 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... [T]he statute makes it impossible that his testimony given on
that occasion should be used by the prosecution on the trial. It can-
not, therefore, be said that in such a criminalcase he has been made a
witness against himself....236
This rule was widely followed among the states; in his opinion in
Kelly, Judge Denio observed that "[i]f the case is so situated that a
repetition of [the witness's admission] on a prosecution against him
is impossible, as where it is forbidden by statute, I have seen no
237
authority which holds or intimates that the witness is privileged."
whether he understood a letter in code. Willie objected, claiming his Fifth Amendment privi-
lege and suggesting that he feared prosecution for misprision of treason. After hearing argu-
ment on the point for two days, the Chief Justice required Willie to answer, on the somewhat
flimsy basis that the question only referred to current knowledge of the code, and current
knowledge would not justify an inference that Willie knew the code at the time the letter was
sent. 25 F. Cas. at 40. Marshall went on to state a general rule that "the court ought never to
compel a witness to give an answer which discloses a fact that would form a necessary and
essential part of a crime which is punishable by the laws." 25 F. Cas. at 40. United States v.
Burr was quoted at length in Counselman v. Hitchcock, 142 U.S. 547, 565-66 (1892), but the
case does not support the Counselman, or even the Kastigar,position. There were no immu-
nity statutes at the time, and so Marshall's opinion does not concern the scope of immunity.
Anything Willie said in Burr's trial could be repeated against him at Willie's subsequent trial.
If Willie's testimony were considered self-incriminating, compelling him to testify in Burr's
case would have been tantamount to compelling him to testify against himself in his own
case. Thus, in Marshall's era, the key question was not the scope of immunity but the defini-
tion of self-incriminating. Marshall's broad language does not apply to a world where a wit-
ness can be given testimonial immunity. (Interestingly, Marshall's holding is far narrower
than his language. Willie was compelled to answer after all. Marshall's definition of self.
incriminatingcould be quite narrow in practice.)
The later merging of the rule against coerced confessions and the privilege against self-
incrimination makes Marshall's language all the more obsolete. See infra note 249. When a
coerced confession occurred, the words were excluded from the defendant's trial but the
fruits came in.
235. 24 N.Y. at 84. The New York Constitution of 1846 tracked the federal Fifth Amend-
ment virtually word for word. See N.Y. CONST. art. I, § 6 (1846) (amended 1938, 1949, 1959
& 1973) ("No person ... shall... be compelled in any criminal case, to be a witness against
himself .... ).
236. 24 N.Y. at 83-84 (emphasis added).
237. 24 N.Y. at 83-84.
Numerous immunity statutes providing for testimonial but not use-

fruits immunity were upheld.238
Furthermore, testimonial immunity was Congress's idea of the
scope of the privilege at the time of Kelly.2 39 An 1857 act provided
broad transactional immunity for anyone who "shall be required to
testify before ... Congress."2 4 o Senator Trumbull criticized the
scope of immunity as too sweeping, allowing "the greatest criminal
[to] escape." 241 Subsequent events confirmed Trumbull's fears,2 42
and the Act was pointedly amended in 1862 to read simply: "The
testimony of a witness ... shall not be used as evidence in any
criminal proceeding against such witness." 243 Congressional debate
over this bill "shows conclusively" that Congress believed that the
Constitution required only testimonial immunity. 244 Senator Benja-
min Wade described the scope of immunity under the new statute
with crystal clarity:
238. For federal cases upholding such statutes, see, for example, United States v. McCar-
thy, 18 F. 87, 89 (C.C.S.D.N.Y. 1883) (relying on Kelly for proper interpretation of the privi-
lege and scope of immunity) and In re Counselman, 44 F. 268 (C.C.N.D. Il1.1890), revd sub
nom. Counselman v. Hitchcock, 142 U.S. 547 (1892). The Counselman court held:
If, through threats or fears of violence, a man confesses that he has committed murder,
and states who was present at the time, and where the weapon and the dead body may
be found, and he is afterwards put on his trial for the offense, he cannot be confronted
with his confession; but the person who saw the crime committed is a competent witness,
although the prosecutor might never have known there was such a witness but for the
confession, and it may be shown by others that the weapon and dead body were found
where the defendant said they could be found.
44 F. at 271. For state cases, see, for example, Exparte Rowe, 7 Cal. 184 (1857) (upholding a
testimonial immunity statute); Higdon v. Heard, 14 Ga. 255,259 (1853) (holding that immu-
nity satisfies the federal constitutional privilege if the answers given by witnesses "cannot be
read in evidence against them, in any criminal case whatever"); Bedgood v. State, 17 N.E.
621, 623-24 (Ind.1888) (holding that a testimonial immunity statute fully guarantees the priv-
ilege); Wflkins v. Malone, 14 Ind. 153, 156-57 (1860) (holding that facts revealed by immu-
nized testimony may be proved against the witness in his later criminal trial, "although the
confessions are wholly inadmissible," and that even if the fruit comes in, so long as the testi-
mony is excluded, "the party cannot be said, in any just sense, 'to be compelled to testify
against himself' in the criminal prosecution"); Ex parte Buskett, 17 S.W. 753, 754-55 (Mo.
1891) (relying on Kelly and holding that a witness is not protected from the introduction of
the fruits of immunized testimony); Lathrop v. Clapp, 40 N.Y. 328, 332 (1864) (approving a
testimonial immunity statute after Kelly); LaFontaine v. Southern Underwriters Assn., 83
N.C. 132, 141-43 (1880) (relying on and quoting Lathrop).
239. Not all federal immunity statutes were as clear as the Act of 1862, discussed infra
text accompanying notes 243-46. For a list of the relevant statutes, see J.A.C. Grant, Immu-
nity From Compulsory Self-Incrimination in a FederalSystem of Governmen4 9 TEmP.L.Q.
57, 64 n.47 (1934).
240. Act of Jan. 21, 1857, ch. 19, § 2, 11 Stat. 155, 156.
241. CONG. GLOBE, 34th Cong., 3d Sess. 437 (1857); cf.United States v. North, 920 F.2d
940 (D.C. Cir. 1990), discussed in supra text accmpanying notes 90-92.
242. See CONG. GLOBE, 37th Cong., 2d Sess. 364, 428-29 (1862).
243. Act of Jan. 24, 1862, ch. 11, 12 Stat. 333.
244. Grant, supra note 239, at 64.
You may inquire; [a witness] may testify and may be compelled to

testify [out of court], but whatever he says shall not be used as evi-
dence against him in any court. That is all that a rascal ought to have
at the hands of justice ....
*.. [If his [out-of-court] testimony is given, though it cannot be
used directly against him, it may lead to other testimony that may
throw light on the subject, whereby in the concatenation of events he
may be convicted of crime. Well, sir, I hope it will be so.2 45
This was Congress's understanding of the privilege at the time it
proposed the Fourteenth Amendment, making the Bill of Rights
applicable against the states and reglossing the federal provi-
sions.246 (This clear understanding and legislative history loom es-
pecially large in light of the absence of much clarity or legislative
history underlying the initial adoption of the Self-Incrimination
Clause in the 1780s.)
The early and middle nineteenth century, the period in which
the above cases were decided and narrow immunity statutes
adopted, is the key period in the development of the modem privi-
lege. Once defense lawyers began to be widely used in the late
eighteenth and early nineteenth centuries, defendants at trial could
keep quiet and let their lawyers speak for them - as is the case
now.247 As the privilege began to take hold in the pretrial phase of
magistrate examination, the criminal justice system needed a way to
find out what happened - to tap the suspect as a valuable testimo-
nial resource. 248 Immunity statutes proved to be the answer, and
they sprang up in abundance during this period. 4 9 This is also the
245. CONo. GLoBE, 37th Cong., 2d Sess. 429 (1862).
246. See Akhil Reed Amar, The Bill of Rights and the FourteenthAmendment, 101 YALE
LJ.1193 (1992). On the "feedback effect" of the Fourteenth Amendment on a proper un-
derstanding of the original Bill of Rights and its limits on federal power, see id. at 1281-82.
247. See Langbein, supra note 2.
248. See id. at 1055.
249. Note that under our regime, in which truthful answers are required under threat of
punishment but (testimonial) immunity is automatically granted, immunity statutes are tech-
nically unnecessary. The privilege becomes self-executing; compelled self-incriminating
statements are automatically excluded from the person's own criminal trial. The privilege
thus operates exactly like the common law rule excluding coerced confessions. So why were
immunity statutes thought to be necessary in the midnineteenth century? Because, formally,
the rule against coerced confessions and the privilege against compelled self-incrimination
were distinct legal doctrines. See supra text accompanying note 188. Coerced confessions
made outside the courtroom were automatically excluded from one's own criminal case, but
testimony within courtrooms was not. Before immunity statutes, therefore, courts had no
way to threaten courtroom witnesses with punishment if they did not answer and still pre-
serve the privilege. (The trial of Aaron Burr illustrated the difficulties of trying to get self-
incriminating testimony from a witness in the pre-immunity-statute era. Indeed, Chief Jus-
tice Marshall strained the definition of non-self-incriminating testimony to get Willie's testi-
mony in. See supra note 234.) But Miranda merged the two doctrines. See supra text
accompanying note 188. After the merger, any compelled self-incriminating statement,
whether made inside or outside a courtroom, would automatically trigger immunity, without
period that saw the emergence of professional police charged with

systematic investigation of crime. Thus, it was in the early and mid-
dle nineteenth centuries that courts and legislatures hammered out
the true scope of the modem privilege.
Unfortunately, the U.S. Supreme Court in Counselman chose to
follow those few cases that called for transactional immunity rather
than the large number that supported testimonial immunity. Wig-
more noted only three cases before Counselman that required im-
munity broader than testimonial;250 all three of these involved state
constitutional provisions that, unlike the federal Fifth, protected
against being compelled to give "evidence" against oneself. The
most influential of these was the Massachusetts case Emery v. Corn-
monwealth.251 The second followed Emery and involved a very
broad protective immunity statute, 252 and the third was later ques-
253
tioned by the court that initially decided it. Indefensibly, the
U.S. Supreme Court in Counselman followed Emery, which was
based explicitly and self-consciously on the peculiar wording of the
Massachusetts Constitution, and rejected Kelly, which was based on
a New York provision that tracked the Federal Constitution virtu-
ally word for word. 254 When Kastigareffectively overruled Coun-
selman, the Supreme Court adopted a new standard - use plus
the need to invoke immunity statutes. (A finding that the statement had been compelled -
formally or informally - would of course still be necessary.) Courts have been unwilling to
recognize this logic under the regimes of transactional and use-fruits immunity because those
immunities are so broad, so they have clung to the requirement of immunity statutes. But,
with testimonial immunity, practice can be reconciled with theory.
250. 8 WIGOORE, supra note 154, § 2283, at 527 & n.6.
251. 107 Mass. 172,182 (1871). Emery explained that because the Massachusetts Consti-
tution forbade that one "be compelled to accuse, or furnish evidence against himself," 107
Mass. at 180 (emphasis added), the privilege was broader in Massachusetts than in New York
and protected a witness "so long as he remains liable to prosecution criminally for any mat-
ters or causes in respect of which he shall be examined or to which his testimony shall relate."
107 Mass. at 185. The case thus required transactional immunity.
252. The highest court of New Hampshire followed Emery in State v. Nowell, 58 N.H.
314, 314 (1878), which required transactional immunity. The court noted that the New
Hampshire constitution provided that no one "shall.. . be compelled to accuse or furnish
evidence against himself." 58 N.H. at 314 (quoting N.H. CONST. art. I, § 15 (emphasis ad-
ded)). In addition, the New Hampshire immunity statute was very broad; it provided that
"no testimony ...given by [the witness] shall, in any prosecution, be used as evidence, either
directly or indirectly, against him, nor shall he be thereafter prosecuted for any offence so
disclosed by him." 58 N.H. at 315.
253. The Virginia Supreme Court required transactional immunity in Cullen v. Common-
wealth, 65 Va. 624, 633 (1873). The Virginia Bill of Rights stated that no one could "be
compelled to give evidence against himself." VA. CoNsT. art. I, § 8. In any event, Cullen was
questioned by a member of the court in Temple v. Commonwealth, 75 Va. 892, 902 (1881).
254. See Counselman v. Hitchcock, 142 U.S. 547, 586 (1892). The Court noted the differ-
ences in the wordings of the federal and Massachusetts constitutional provisions but stated
that "there is really, in spirit and principle, no distinction arising out of such difference of
language." 142 U.S. at 586.
use-fruits immunity - but offered no good reason for failing to

return to Kelly's testimonial immunity.255 In short, Fifth Amend-
ment doctrine today is the unconvincing and half-hearted residue of
an 1870s opinion from Massachusetts that explicitly relied on state
constitutional phrasing that the Federal Fifth Amendment im-
pliedly rejected.
Counselman also relied on the landmark Boyd case,- 6 whose
Fourth and Fifth Amendment mishmash has now been emphatically
rejected. Indeed, Boyd's role as the godfather of Counselman is
visible in two of Boyd's key - and now repudiated - passages.
The first fused together the Fourth and Fifth Amendments:
We have already noticed the intimate relation between the two
amendments. They throw great light on each other. For the "unrea-
sonable searches and seizures" condemned in the fourth amendment
are almost always made for the purpose of compelling a man to give
evidence against himself, which in criminal cases is condemned in the
fifth amendment2 57
The second also explicitly read the Fifth Amendment as if it prohib-
ited compulsion to "furnish evidence" and implied that all criminal
subpoenas of a defendant violated the Fifth.2 8s But these two inter-
pretations were decisively renounced by the Supreme Court in
United States v. Leon2 59 and Fisher v. United States,260 respectively.
In each of these cases, the Court used the same phrase: Boyd had
'261
not stood "the test of time."
Counselman also rejected the English rule, which was testimo-
nial immunity. In the coerced confession context, which Miranda
merged into the Self-Incrimination Clause, this rule was firmly es-
tablished in the 1783 case The King v. Warickshall.262 In Warick-
shall, the defendant, charged with being an accessory after the fact
255. Ironically, New York's highest court repudiated Kelly as a result of the Counselman
decision. See People v. O'Brien, 176 N.Y. 253 (1903) (requiring transactional immunity and
explicitly following Counselman).
256. Boyd v. United States, 116 U.S. 616 (1886).
258. 116 U.S. at 637 (stating that the subpoena in this case "is surely compelling [the
defendant] to furnish evidence against himself" (emphasis added)).
259. 468 U.S. 897, 905-06 (1984) (citing Boyd's Fourth-Fifth Amendment fusion
progeny).
260. 425 U.S. 391,408 (1976) ("[T]he Fifth Amendment does not independently proscribe
the compelled production of every sort of incriminating evidence but applies only when the
accused is compelled to make a testimonial communication that is incriminating.").
261. Leon, 468 U.S. at 906 ("The Fifth Amendment theory has not withstood critical
analysis or the test of time .... "); Fisher,425 U.S. at 407 ("Several of Boyd's express or
implicit declarations have not stood the test of time.").
262. 168 Eng. Rep. 234 (K.B. 1783).
for knowingly receiving stolen property, made a confession because

of "promises of favour,"263 which in those days was rightly consid-
ered to be a coerced confession. 26" In her confession, she revealed
where the stolen goods were hidden, and so they were found in her
mattress in her lodgings. The question to be decided was whether
the stolen goods and their location could be admitted into evidence;
the court held that they were admissible. The court stated with
assurance:
[T]his subject has more than once undergone the solemn considera-
tion of the Twelve Judges; and a majority were clearly of the opinion,
That although confessions improperly obtained cannot be received in
evidence, yet that any acts done afterwards might be given in evi-
dence, notwithstanding they were done in consequence of such
confession. 265
Coerced confessions, said the court, were unreliable, but the fruits
here posed no reliability problem. 266 To exclude these fruits simply
in order to prevent a suspect from being "made the... instrument
of her own conviction" would be "novel in theory," "dangerous in
practice," and "repugnant to the general principles of criminal
law."'267 This was apparently the leading English case on point
when the U.S. Bill of Rights was adopted in 1791.
263. 168 Eng. Rep. at 234.

264. The crown witness system was a major force leading to the rule against coerced
confessions. John H. Langbein, Shaping the Eighteenth-Century CriminalTria. A Vew from
the Ryder Sources, 50 U. CHL L. REv. 1, 105 (1983). Under this system, witnesses were
granted immunity from prosecution if they testified against their partners in crime. If several
gang members tried to become crown witnesses, the one who could reveal the most was
granted immunity. This led to contests between gang members to tell all, and more than all.
Incentives to lie were great, because the crimes involved usually carried the death penalty. If
a suspect tried and failed to become a crown witness, his confession could and would be used
against him. Thus the exclusionary rule for tainted confessions was born of a concern for
reliability.
265. 168 Eng. Rep. at 235 (footnote omitted). The court emphasized: "The rules of evi-
dence which respect the admission of facts, and those which prevail with respect to the rejec-
tion of parol declarations or confessions, are distinct and independent of each other." 168
Eng. Rep. at 235. In America, it was hornbook law as late as 1960 that courts would not
exclude the fruits of coerced confessions. See 2 FNcas WHARTON, WHARTON'S CRDMINAL
EVIDENCE §§ 357-58 (Ronald A. Anderson ed., 12th ed. 1955); 3 WiGMoRE, supra note 154,
§§ 856-59; Yale Kamisar, Wolf and Lustig Ten Years Later: Illegal Evidence in State and Fed-
eral Courts, 43 MiNN. L. Rnv. 1083, 1115 n.109 (1959). In fact, we are aware of no U.S.
Supreme Court case - before or after 1960 - that actually excludes physical fruits of a
coerced confession that occurred outside formal proceedings. Miranda does contain an am-
biguous sentence about fruits, see 384 U.S. 436, 479 (1966) (speaking of "evidence obtained
as a result of interrogation"), but that sentence has since been repudiated. See supra section
I.C. But cf. Wong Sun v. United States, 371 U.S. 471, 484-88 (1963) (excluding, on Fourth
Amendment grounds, physical evidence as fruit of illegal arrest).
266. 168 Eng. Rep. at 235.
267. 168 Eng. Rep. at 234 (emphasis added). Note the court's explicit and emphatic re-
jection of a broad noninstrumentalization theory.
Seventy years later, in The Queen v. Leatham, 26s the English

judges made clear that a similar result applied to self-incrimination
under formal immunity statutes: only testimonial immunity, not
transactional or use plus use-fruits immunity, was required. In
Leatham, decided under an immunity provision of the Corrupt
Practices at Elections Act, the defendant's immunized testimony
before election commissioners led to discovery of a preexisting let-
ter that was introduced against him at his subsequent trial for brib-
ery. The court held that only the defendant's statements should be
excluded and that the letter was admissible. Crompton J. noted:
In the analogous case of confessions by persons accused of crimes,
they cannot be used against such persons if obtained from them under
the compulsion of a threat, or the inducement of a promise; but mat-
ters to which such a confession gives a clue may nevertheless be
unexceptionally put in evidence. For instance, if stolen goods or a
murdered body are or is found in a 269
place indicated by the confession,
this fact may be given in evidence.
Several judges mentioned the great difficulties involved in adminis-
tering a rule of use plus use-fruits immunity. As Crompton put the
point, a legislative grant of use-fruits immunity
would have introduced great inconvenience, giving rise in every case
to the necessity for an inquiry, in all subsequent proceedings, as to
whether or not the clue which led to them was obtained from some-
thing let fall by the defendant when before the Commissioners, and
thereby opening as wide a field for investigation as can possibly be
conceived. 270
The problems noted by Crompton would be avoided by restor-
ing the narrow scope of the privilege and allowing fruits to come in.
This result could also be reached by simply expanding the Supreme
Court's inevitable discovery doctrine, expounded in Murray v.
United States.271 The fruits could have come to light anyway, so the
268. 121 Eng. Rep. 589 (Q.B. 1861).

269. 121 Eng. Rep. at 593 (Crompton, J.). Contemporaneous American cases explicitly
following the same logic in the self-incrimination/testimonial immunity context include Wil-
kins v. Malone, 14 Ind. 153, 156-57 (1860) and In re Counselman, 44 F. 268, 271 (1890), revd
sub nor. Counselman v. Hitchcock, 142 U.S. 547 (1892).
270. 121 Eng. Rep. at 592-93 (Crompton, J.); see also 121 Eng. Rep. at 594 (Blackburn,
J.) ("[A]n enactment that nothing, the first clue to which was given by a witness under exami-
nation by the Commissioners, should be provable against him by evidence aliunde, would
have been very unwise; would have encouraged rather than checked the corrupt practices
which the Act seeks to put a stop to; and would have introduced excessive practical incon-
venience .... ).
For further statements of the English rule of testimonial immunity, see Commonwealth v.
Knapp, 26 Mass. (9 Pick.) 495, 510-11 (1830) and 1 J.F. ARcrmota, PRACrICE, PLEADINO,
AND EVIDENCE IN CRIMINAL CAsEs 424 (Thomas W. Waterman ed., 7th ed. 1860).
271. 487 U.S. 533, 536-44 (1988). The Canadians have a well-developed inevitable dis-
covery doctrine and draw sharp distinctions between physical evidence and testimony for
argument would go, and the presumption to that effect would be

irrebuttable.
D. "Witness"
Schmerber2 7r was right to emphasize the distinction between
testimony and physical evidence, but later decisions have failed to
follow its logic to the end. Only the defendant's compelled testi-
mony should be protected by the Amendment. The "witnessing"
that the defendant has a right to exclude from the criminal trial
includes both communicating on the stand at trial and introducing
at trial any earlier compelled depositions. This definition of witness
closely tracks what seems to be the best definition of witness under
the Confrontation and Compulsory Process Clauses of the Sixth
Amendment. 273 Unlike some state constitutions, such as the Mas-
sachusetts Constitution of 1780, the Fifth Amendment does not
prohibit the government from compelling a defendant to "furnish
evidence against himself." 274 Compelled fruit is admissible, but
compelled testimony is not.
purposes of exclusion. See R. v. Collins, [1987] 1 S.C.R. 265, 284 (Can.) (Lamer, J.) ("Real
evidence that was obtained in a manner that violated the Charter [constitutional protections
against self-incrimination, unreasonable search and seizure, and so on] will rarely operate
unfairly for that reason alone. The real evidence existed irrespective of the violation of the
Charter and its use does not render the trial unfair."); Mellenthin v. Regina, [1992] 3 S.C.R.
615 (Can.) (drawing a distinction between real evidence that the defendant has been forced
to create by constitutional infringement and real evidence that the defendant has been forced
to locate or identify by constitutional infringement, and also distinguishing between indepen-
dently existing evidence that could have been found without compelled testimony and inde-
pendently existing evidence that would have been found without compelled testimony);
Black v. Regina [1989] 2 S.C.R. 138 (Can.) (holding that physical evidence obtained as a
direct result of a statement taken in violation of a defendant's right against self-incrimination
is admissible where the evidence would have been uncovered by the police in any event).
The author of a leading Canadian treatise has noted the "overwhelming tendency of our
courts to characterize any tangible evidence such as weapons or drugs as real evidence not
going to the fairness of the trial and hence, under the Collins regime, generally admissible."
DoN STUART, CHARTER JUSTICE IN CANADIAN CRIMINAL LAW 401 (1991). He does cite a
few exceptions, however, mainly in cases where the constitutional violation was "flagrant."
Id. at 406-07, 414.
273. A sensible definition is found in Justice Thomas's concurrence, joined by Justice
Scalia, in White v. Illinois, 502 U.S. 346, 358 (1992). This definition includes "any witness
who actually testifies at trial" and anyone who gives "extrajudicial statements... contained
in formalized testimonial materials, such as affidavits, depositions, prior testimony, or confes-
sions." 502 U.S. at 365 (Thomas, J., concurring). The extension to extrajudicial statements
prepared for trial is analogous to the Self-Incrimination Clause's application to out-of-court
compulsion. See supra note 63. Obviously, the government should not be able to evade a
defendant's right to confront government witnesses by recessing a trial, walking across the
street and deposing a witness without the defendant or his lawyer present, and then returning
to court and introducing into evidence a transcript and videotape of the deposition. Cf.id.
274. Compare U.S. CONST. amend. V with MASS. CONST. pt. I, art. XII (1780) ("furnish
evidence against himself") and PA. CoNST.art. I, § 9 (1776) ("give evidence against him-
self"). Several other state constitutions use the phrase give evidence. See 8 WiOMoRE, supra
Fisher,275 too, fits our approach. Obliging a suspect to hand

over incriminating words or things already in existence can be dis-
tinguished from obliging him to be a witness - to testify in re-
sponse to clever questions put by a prosecutor. As we shall see,
testimony extorted by clever prosecutors can raise distinct reliabil-
ity concerns - prosecutors tricking suspects into misleading and
hasty concessions that look like confessions - that are not gener-
ally present when a suspect is merely obliged to hand over an al-
ready extant object (with testimonial immunity for the act of
276
handing over itself).
Two borderline questions noted earlier remain concerning de-
fendants' bodies and diaries. First, what, if any, protections should
a defendant have if the prosecution would like to use his body,
rather than his communications? 277 The answer lies in the Fourth
Amendment, not in the Fifth. The use of a defendant's body as
physical evidence is not testimonial, and therefore it is not covered
by a privilege that protects a person from "witnessing" against him-
self. But the Fourth Amendment guards against "unreasonable
searches and seizures" of "persons," and its protections should be
applied in these situations.278 Properly construed, the requirement
of "reasonableness" invites judges to engage in balancing the state's
and the individual's interests; 279 factors to be considered should in-
clude the gravity of the offense charged, the invasiveness or humili-
ation of the examination or act, and the importance of the evidence
to the prosecution's case.2 80 (Unlike the Fifth Amendment, the
note 154, § 2252, at 321-24 n.3. Wigmore thought that these differences were "immaterial"
and that "[tihese various phrasings have a common conception, in respect to the form of the
protected disclosure." 8 id. § 2263, at 362. However, Wigmore thought that New York's
Kelly case should be the uniform rule. 8 id. § 2283, at 525-28.
275. Fisher v. United States, 425 U.S. 391 (1976), discussed in supra section I.D.
276. Under testimonial immunity, the subpoenaed object could be used at trial but not
the testimonial fact that defendant supplied the object. This clean rule contrasts dramatically
with the Kastigarregime, under which the proper scope of "act of production" immunity is a
huge unsolved - and perhaps metaphysically insoluble - puzzle. To get a sense of the
Kastigarcomplexities here, see Kenneth J. Melili, Act-of-ProductionImmunity, 52 Omno ST.
L.J. 223 (1991); Reitz, supranote 165, at 618-27.
277. See supra section I.D.
278. In the lineup case of United States v. Wade, 388 U.S. 218 (1967), Justice Fortas
articulated a rationale that sounds like a Fourth Amendment, not a Fifth Amendment, inter-
pretation: the lineup was "an incident of the State's power to arrest, and a reasonable and
justifiable aspect of the State's custody resulting from arrest." 388 U.S. at 260 (Fortas, J.,
concurring).
279. See Amar, supra note **, at 801-11.
280. If the defendant refuses to comply, there are several options available. Depending
on the circumstances, the defendant could either be held in contempt or evidence of the
refusal could be introduced at trial. See South Dakota v. Neville, 459 U.S. 553 (1983). There,
the Court held that "a refusal to take a blood-alcohol test, after a police officer has lawfully
Fourth Amendment approach would protect a person's body even

in civil cases - indeed, especially in civil cases.) Under this ap-
proach, courts are free to distinguish sensibly between, say, a pain-
ful and intrusive stomach pumping and a quick and virtually
painless taking of a blood sample.281
Second, what protection should diaries enjoy? Unlike bodies,
diaries are clearly communicative and testimony-like. At a mini-
mum, the search for and seizure of diaries should be governed by a
Fourth Amendment reasonableness test. This test should be in-
formed by the probability that a search for a diary will be intrusive,
the broad freedom of thought principles of the First Amendment,
and the special treatment the Fourth Amendment accords to "pa-
pers." 282 What's more, reading a person's diary (even if lawfully
obtained) in open court, civil or criminal, can be seen as an addi-
tional invasion of privacy - an incremental "search" of a man's
soul, an additional "seizure" of a woman's most intimate secrets -
that once again calls for a careful judicial inquiry into the reasona-
bleness of this public reading s3 Above and beyond these Fourth
Amendment concerns is a key Fifth Amendment concept - relia-
bility. Writers of diaries often fantasize or write in a personal short-
hand easily misinterpreted. Though not compelled testimony in
exactly the same way that forcing the witness to take the stand is
compelled testimony, diaries may raise sufficiently distinct reliabil-
ity issues to justify treating them differently from all other volunta-
rily created documents that the government wants to search for or
subpoena. 2s4 Therefore, we can see why the Court has intuited that
diaries might differ on Fifth Amendment grounds from, say, volun-
tarily created business records. 28 5
requested it, is not an act coerced by the officer, and thus is not protected by the privilege
against self-incrimination." 459 U.S. at 564. Moreover, the officer was not required to give
warnings that evidence of refusal might be introduced at trial. 459 U.S. at 566.
281. Compare Rochin v. California, 342 U.S. 165 (1952) (invalidating stomach pumping)
with Schmerber v. California, 384 U.S. 757 (1966) (upholding the taking of blood). See also
Winston v. Lee, 470 U.S. 753 (1985) (disallowing unnecessary surgery).
282. See Amar, supra note **,at 806-07; see also TAYLOR, supranote 112, at 66-68; Eric
Schnapper, Unrebsonable Searches and Seizures of Papers, 71 VA. L. RaV. 869 (1985).
283. Cf.Fisher v. United States, 425 U.S. 391, 401 n.7 (1976) (noting "[s]pecial problems
of privacy which might be presented by subpoena of a personal diary").
284. See Henry E. Smith, Two Reliability Rationales for the Privilege Against Self-In-
crimination (Mar. 1995) (unpublished manuscript, on file with authors) (proposing "commu-
nicative" reliability rationale based on concern about risk of misunderstanding, and applying
rationale to case of diaries).
285. See supra text accompanying notes 142-44.
E. The Big Idea(s)

The Self-Incrimination Clause has long been a mandate in
search of a meaning, and the scope of immunity permitted under it
has varied widely over the centuries. The rule of testimonial immu-
nity has powerful advantages, especially when compared with trans-
actional and use plus use-fruits immunity. It safeguards defendants
and emphasizes truth-finding.
How, one might ask, is such a positive-sum solution - reducing
the risk of both false convictions and false acquittals - even possi-
ble? As statisticians have long understood, one can simultaneously
reduce both false negatives and false positives only by bringing
more information into a system. That is precisely what our propo-
sal does - by bringing in fruit and facilitating civilized pretrial
questioning. Our current system throws out too much information,
and in the end, this hurts both truth-seeking prosecutors and inno-
cent defendants.
But why does it make sense for the Self-Incrimination Clause to
distinguish between compelled testimony and compelled fruit?
First and foremost, fruits and physical evidence are more relia-
ble than coerced testimony itself.28 6 This analysis is similar to the
logic underlying the substantive criminal law doctrine of corpus
delicti. Truth is a preeminent criminal procedure value in the Bill
of Rights: most procedures were designed to protect innocent de-
fendants from erroneous conviction. 28 7 Especially when pressured,
people may confess - or seem to confess - to crimes they did not
commit. As Blackstone put the point in his bestselling Commenta-
ries, out-of-court confessions "are the weakest and most suspicious
of all testimony; ever liable to be obtained by artifice, false hopes,
promises of favor or menaces; seldom remembered accurately, or
reported with due precision; and incapable in their nature of being
disproved by other negative evidence. 288 A leading modem com-
286. Cf.Friendly, supra note 63, at 712 n.176 (observing that physical leads are often
more important to law enforcement than getting statements for use in court). Because of
huge leaps in technology, physical evidence can yield far more reliable information today
than at the Founding. This enhanced reliability only strengthens the wisdom of respecting
the testimony-fruits distinction established as early as 1783. See supra text accompanying
notes 262-67. "Fruit" in our model of course also encompasses reliable third party testimony.
287. See supra note 151.
288. 4 BLActsrozNE, supranote 10, at *357. Blackstone, always influential, is particularly
important in this context because of the paucity of legislative history concerning the Self-
Incrimination Clause at the ratification of the Bill of Rights. He retained an enormous legal
,influence at the time the Fourteenth Amendment was ratified. Note how Blackstone's em-
phasis on reliability parallels the analysis in the 1783 case of The King v. Warickshall, 168
Eng. Rep. 234 (K.B. 1783), discussed suprasection II.D. See also BARLow, supranote 154, at
189 (claiming that using "the Rack or Torture to compel Criminals to accuse themselves" is
mentator, Judge Friendly, echoed Blackstone's concern and noted

that the main motivation behind extending the privilege to informal
proceedings must have been "the truly dreadful risk of the false
confession.... [T]here is thus good reason to impose a higher stan-
dard on the police before allowing them to use a confession of mur-
der than a weapon bearing the confessor's fingerprints to which his
confession has led .... ,,2s9
Though the prospect of an unreliable confession (or perceived
confession) occurring on the witness stand itself may seem to some
unlikely in a criminal trial today, we must remember that from 1789
until well into this century, many innocent defendants in noncapital
cases could not afford lawyers and were not furnished lawyers by
the government. If forced to take the stand, they might be bullied
or bamboozled by a professional prosecutor into assenting to un-
true or misleading propositions that would - wrongly - seal their
fate in the minds of the jury. While the concern about a slip-up on
the stand applies to other witnesses as well, slip-ups by the defend-
290
ant are particularly damaging in jurors' minds.
In a post-Gideon291 world the defenseless defendant scenario is
less likely, but even after Gideon, many defendants enjoy only nom-
inal assistance of counsel. Moreover, even a good lawyer cannot
always save an innocent but unpersuasive-sounding client from be-
ing demolished on the stand. As the Supreme Court observed a
century ago in one of its earliest self-incrimination opinions:
It is not every one who can safely venture on the witness stand though
entirely innocent of the charge againsthim. Excessive timidity, ner-
vousness when facing others and attempting to explain transactions of
a suspicious character, and offenses charged against him, will often
confuse and embarrass him to such a degree as to increase rather than
remove prejudices against him. It is not every one, however honest,
who would, therefore, willingly be placed on the witness stand.292
not only "cruel" but "at the same Tme uncertain, as being rather Trials of the Strength and
Hardiness of the Sufferer, than any Proof of the Truth"), quoted in Langbein, supra note 2, at
1085 n.157.
289. FRmNDLY, supra note 102, at 281-82.
290. Compare Blackstone's suggestion that statements from the defendant's own lips
have a peculiarly devastating impact. See supra text accompanying note 288.
291. Gideon v. Wainwright, 372 U.S. 335 (1963) (establishing the right of indigent defend-
ants in noncapital cases to have the assistance of appointed counsel).
292. Wilson v. United States, 149 U.S. 60, 66 (1893) (emphasis added). Note how these
concerns are not present in the typical case where the government secretly wiretaps a suspect
in his home and later introduces his words - his "testimony" - in open court against his
will. Although noninstrumentalists might have a hard time distinguishing between using a
defendant on the stand and using him in a wiretap, see supranote 161, a reliability approach
shows wh these two cases are different.
More recently, the Supreme Court's repeated emphasis on trust-

worthiness has led it to sharpen the difference between testimony
and physical evidence; the logical culmination of this trend is testi-
monial immunity. We need not say every coerced statement is un-
reliable, or every physical fact reliable; the Fifth Amendment lays
down a bright-line rule, and as with any rule the rationale need not
explain every instance of the rule's application. It is enough that as
a category, a criminal defendant's compelled testimony is particu-
larly troubling on reliability grounds. 293 On the other hand, volun-
tary testimony is likely to be more reliable because innocent
defendants have the opportunity to screen themselves on the basis
of whether they might be misunderstood.
Focusing on reliability answers our three basic questions that
any theory of the clause must confront. First, why does the privi-
lege apply only in a "criminal case" and not in civil ones? The an-
swer lies in the higher degree of certainty required for a criminal
conviction. We insist on proof beyond a reasonable doubt in crimi-
nal but not civil cases precisely because we are so much more con-
cerned about erroneous criminalconvictions. For the same reason,
we are particularly concerned with unreliable evidence being intro-
duced against a criminal defendant. (This also helps explain why
corpus delicti is a criminal doctrine. 294) Second, why can immunity
overcome the privilege and "compel" a witness to speak? Because
the fruits, which can be introduced in a criminal trial, will be more
reliable than the speaker's immunized words, which must be ex-
cluded. Finally, why does the Fifth Amendment concern "witness-
ing" rather than furnishing evidence? Again, pieces of physical
evidence - a gun, the defendant's own body, tax records - are
more reliable than forced words.
Although our approach begins with reliability, it does not end
there. Government interrogation and investigation implicates other
293. Nor are reliability concerns always cured by a physical corroboration test, because
many confessions may concern internal mental states, where misunderstandings are quite
likely. In short, physical evidence can at best partially rather than fully corroborate a state-
ment. To the extent the physical evidence partially corroborates, it can be introduced itself.
To introduce the confession in addition risks introduction of unreliable and uncorroborated
aspects of the confession - say, about the defendant's mens rea. Perhaps, however, the trier
of fact could be told merely that "something defendant said" led the police to the victim's
body, the stolen goods, or what have you. On the other hand, this paraphrase looks rather
like defendant witnessing - it is an account of defendant's own words - and, as with all
paraphrases, introduces reliability concerns of its own.
294. Under our reading, the Self-Incrimination Clause strongly harmonizes with the doc-
trine of corpus delicti. Both sharply distinguish, on reliability grounds, between a suspect's
self-accusatory words on the one hand, and reliable physical fruit and third-party witnesses'
reliable testimony, on the other.
constitutional values, and our approach respects these values. For

example, a distinction between testimony and physical evidence in-
directly promotes First Amendment values. A rule excluding from
criminal cases only defendants' compelled utterances (and perhaps
intimate personal papers such as diaries) would lead to acquittals
for most eighteenth-century thought crimes - including blas-
phemy, heresy, and sedition295 - but it would not be nearly so big
a stumbling block to prosecution of murderers and rapists, whose 296
crimes generate nontestimonial fruits like knives and bodies.
Thus we do justice to both the Fifth Amendment's literal applica-
tion beyond the thought crime and to its obviously weaker intuitive
appeal in cases of violent crime.
Related to the First Amendment freedoms of thought and con-
science is a certain narrow but important protection of "mental pri-
vacy" and "noninstrumentalization." A criminal defendant's
compelled utterances may often reveal interior mental states - of
guilt, conscience, joy, and so on. These compelled words and
thoughts may reflect a defendant's inner self - his soul - and are
far more likely to emerge in criminal interrogation than in the typi-
cal civil case. Under our reading, government at trial may not di-
rectly use a person's compelled inner life as the means of that
person's own destruction. The government may directly use only
tangible things - fruits - typically focusing on what happened
rather than the personal meaning the defendant wishes to attach to
these events. In this limited but important sense, our reading pro-
tects a person from being compelled in a criminal case to be a wit-
ness against his own inner "self." The government is entitled to,
and may use, every man's evidence, but not his soul. A straightfor-
ward focus on reliability reaches the same result. Reports of inte-
295. Thought crimes have not been confined to the eighteenth century. Persecution of
thought crimes, especially political crimes, has continued to generate praise for a right to
remain silent. The anti-Communist hearings and trials of the 1950s, for instance, provoked
fervent defenses of the privilege, most notably from Erwin Griswold. See GruswowD, supra
note 32. For a more general discussion of how the Fourth and Fifth Amendments have been
pressed into the service of First Amendment values, see Stuntz, supra note 115.
296. To fully protect mere "thought criminals," we need to supplement the Fifth Amend-
ment with an absolute privilege of associational privacy for purely political and religious
confederates, lest a person be compelled to divulge the identity of his thoughtmates who
could then be forced to testify against him. Cf.Shelton v. Tucker, 364 U.S. 479 (1960) (hold-
ing that states cannot compel teachers to disclose their associational ties because to do so
deprives teachers of their right to associational freedom); NAACP v. Alabama, 357 U.S. 449
(1958) (holding that a political association may assert its members' constitutional right not to
be compelled to disclose their affiliation with the association). In the absence of this absolute
privilege, testimonial immunity could not protect a political or religious group very well. But
neither could Counselmanor Kastigar: one person with immunity could be obliged to furnish
a list of members, and each member could then be obliged to testify against each other.
rior mental states are easily misunderstood, notoriously imprecise

(depending on a person's mood when reporting), and hard to verify.
"How did you feel when you killed her?" is a very different kind of
question from "Where is her body buried?"
The "psychological cruelty" of a compelled self-incrimination
now appears in a very different light. Conventional accounts fo-
cused on guilty defendants faced with a "cruel" choice of contempt,
self-accusation, or perjury. Earlier, we rejected this misplaced ef-
fort to shield evildoers from the consequences of their crimes. But
once we see that compelled testimony from an innocent defendant
may be misleading, and may destroy a man who has done no wrong,
a different cruelty emerges: the cruelty of forcing someone who
knows he is innocent but unpersuasive to take the stand, knowing
that his truthful testimony may well hang him because a clever pros-
ecutor can twist his words, and make him look guilty before the
jury. A desire to protect the innocent defendant from erroneous
conviction - and from the added injury, insult, and cruelty of be-
ing forced to cut his own throat with an honest but unpersuasive
performance - is wholly consistent with the deep structure of our
Bill of Rights.
An added advantage of our approach is that no "emergency ex-
ception" is needed to prevent imminent harm. The need to get
fruits in emergency situations - a loaded gun in Quarles297 or In-
nis,298 or a child in Bouknight299 or Brewer3oo - will not create an
exception that impermissibly allows compelled testimony itself to
be introduced at trial. Justice O'Connor wisely warned against cre-
ating a "finespun new doctrine on public safety exigencies incident
to custodial interrogation, complete with the hair-splitting distinc-
tions that currently plague our Fourth Amendment
jurisprudence. 301
Finally, the use of depositions and pretrial judicial examination
would curb the temptation to police abuse. Questioning would be
accomplished in a relatively civilized setting, as in civil litigation,
rather than in the rough-and-tumble atmosphere of interrogation at
the police station, with its attendant intimidation, sleeplessness, and
other physical and psychological pressures. The process would be
judicially supervised; refusals to answer and outright lies would be
297. New York v. Quarles, 467 U.S. 649 (1984).

298. Rhode Island v. Innis, 446 U.S. 291 (1980).
299. Baltimore City Dept. of Social Servs. v. Bouknight, 493 U.S. 549 (1990).
300. Brewer v. Williams, 430 U.S. 387 (1977).
301. Quarles,467 U.S. at 663-64 (O'Connor, J., concurring in part and dissenting in part).
punished with judicial contempt or adverse inferences, 302 not with

fists banging on desks and shouted four-letter words. These deposi-
tions and hearings would act as the equivalent of the nineteenth-
century questioning of the accused by a magistrate. Fourth Amend-
ment standards would constrain both the government's right to de-
mand answers in general - the government must justify its
decision to single a person out for detention (seizure) and interro-
gation (search) - and the government's right to ask any particular
question. Irrelevant questions, questions for which no foundation
had been laid, intrusive or embarrassing questions, repetitive ques-
tions - all these should be subject to a general Fourth Amendment
3
test of reasonableness. 30
The Founding-era history of the self-incrimination slogan in
America was bound up with concerns about torture, and our solu-
tion attacks the problem of torture on several fronts. While torture
can be lawless and unregulated, leaving a suspect's body and mind
subject to the whims of examiners, our approach civilizes and con-
trols the process of evidence-gathering. While torture produces un-
reliable confessions, our approach emphasizes reliability. While
torture is cruel, we provide a check by focusing on unreasonable or
intrusive Fourth Amendment seizures of the person. (And the fact
that even criminal witnesses and civil witnesses and parties - and
everyone else, too - must be protected against torture proves that
the root antitorture idea is largely a Fourth Amendment idea and
not a Fifth Amendment idea.) Finally, torture has at times been
particularly likely to occur in cases involving religious and political
opponents of the government. In response, we stress the First
Amendment values protected when a defendant's words cannot be
introduced at trial.
CONCLUSION
While our approach may at first glance seem like a startling

break from current interpretations, in fact it is merely an extension
of any one of four current doctrines or trends. First, it is a logical
extension of Schmerber's3°4 distinction between physical evidence
302. See supra text accompanying notes 190-91.
303. Eben Moglen quotes a Boston pamphlet from 1769 attacking ex officio oaths as in-
viting customs commissioners in Vice Admiralty courts to harass their political opponents
with vexatious, impertinent, and groundless questions such as "Pray Sir, when did you kiss
your maid Mary? - Where? and in what manner? ... Did you lay with her in a barn? or in
your house?" Moglen, supra note 184, at 1116. On the obviously gendered nature of this
example and of privacy discourse generally, see Amar, supra note **, at 808-09.
and testimony. Second, the Court could replace Kastigar305 by fol-

lowing the same reasoning that it followed in Murphy306 as it
moved toward replacing Counselman307 with Kastigar in the first
place: intersovereign friction would be eliminated. Third, Justice
O'Connor has pointed the way in Quarles,308 where she suggested
always allowing fruits in for mere Miranda violations. And finally,
an extension of the inevitable discovery doctrine would simply pre-
sume that fruits have an independent source and thus can be intro-
duced at trial.
Fifth Amendment jurisprudence, like that of the Fourth Amend-
ment, needs to regain coherence. In Fourth Amendment case law,
Boyd's3°9 (con)fusion of the Fourth and Fifth Amendments led to a
constitutionally flawed exclusionary rule, motivated by the mis-
taken notion that the Fourth, like the Fifth, requires that certain
items be excluded from a criminal case. 310 In Fifth Amendment
case law, Boyd's (con)fusion has led to the mistaken notion that the
Fifth, like the Fourth, is about things - "effects" or fruits - and
not just words. Instead, the privilege requires immunity that is ab-
solute but narrow. Fact gathering should be accomplished under
circumstances that are both more civilized and more apt to produce
the truth. Instead of the current wide divergence between civil and
criminal discovery practice, the two should be brought closer to-
gether. Finders of fact in criminal cases should not be deprived of
reliable, highly probative evidence. A legal system that ignores the
truth is simply not doing its job, and neither is a court that cannot
make the Constitution cohere.
305. Kastigar v. United States, 406 U.S. 441 (1972).

306. Murphy v. Waterfront Commn., 378 U.S. 52 (1964).
307. Counselman v. Hitchcock, 142 U.S. 547 (1892).
308. New York v. Quarles,.467 U.S. 649, 660-74 (1984) (O'Connor, J., concurring in part
and dissenting in part).
309. Boyd v. United States, 116 U.S. 616 (1886).
310. See Amar, supranote **, at 787-91. For a long list of cases documenting that the so-
called Fourth Amendment exclusionary rule was really a joint Fourth-Fifth affair, under a
theory that the Court has now rightly discredited, see Amar & Marcus, supra note 69, at 16
n.97. In addition to the eleven U.S. Supreme Court cases cited there, see also Adams v. New
York, 192 U.S. 585,594,597-98 (1904); Perlman v. United States, 247 U.S. 7, 13 (1918); Bur-
deau v. McDowell, 256 U.S. 465, 474-76 (1921); Marron v. United States, 275 U.S. 192, 194
(1922); United States v. Lefkowitz, 285 U.S. 452, 466-67 (1932); and Ker v. California, 374
U.S. 23, 30 (1963). See generally Corwin, supra note 177, at 1-2, 13-16, 203-04 (carefully
reading exclusionary case law as based on a Fourth-Fifth fusion).

Fifth Amendment First Principles - The Self-Incrimination Clause

Uploaded by

Copyright:

Available Formats

Fifth Amendment First Principles - The Self-Incrimination Clause

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fifth Amendment First Principles - The Self-Incrimination Clause

Uploaded by

Copyright:

Available Formats

Yale Law School

Yale Law School Legal Scholarship Repository

Fifth Amendment First Principles: The Self-

Renee Lettow Lerner

Follow this and additional works at: http://digitalcommons.law.yale.edu/fss_papers

2. See John H. Langbein, The HistoricalOrigins of the Privilege Against Self-Incrimina-

"in" a "criminal case" - but the fruits of these compelled pretrial

the Self-Incrimination Clause, as (mis)construed over the years. By

American implementation, from plain meaning, and from common

Perhaps the most striking flaw in current application of the priv-

6. U.S. CoNsr. amend. VI (emphasis added).

Let's consider a simple situation - a single defendant truthfully

anything related to his testimony. (The government will have to

tablished this as the test of in-court compulsion. A defendant had

22. See 380 U.S. at 614-15.

25. U.S. SENTENCING CoMMN., FEDERAL SENTENCING GuiDEUn4_S MANUAL § 3.E1.1

2. Out-of-Court Compulsion: Government Employment

37. 335 U.S. 1 (1948).

context, is it possible that the real problem here is that immunity

42. 390 U.S. 39 (1968).

seemed to announce a new principle in self-incrimination cases, and

50. 402 U.S. at 427 (emphasis added).

4. Out-of-Court Compulsion: Police Interrogation

best. We do not tolerate such behavior in civil litigation.60 Rather,

C. "In Any Criminal Case"?

the Fifth Amendment.6a Exactly what must be excluded from "a

lished an extraordinarily sweeping form of immunity that came to

mony out of court and then introduce it in a criminal case against

73. 378 U.S. 52 (1964).

to the testimony and prosecutors working on the case against the

crushing burdens on prosecutors, the D.C. Circuit approach at

the prosecution from using any resulting confession in its case in

96. 417 U.S. 433 (1974).

mitted that he owned the gun.10° In the ensuing prosecution for

100. 467 U.S. at 652-53.

a (merely) Miranda-defective confession are admissible - even

own words, much as would the defendant himself as a witness on

display himself to the jury or perform physical acts. 1 9 Courts

fendant was convicted of driving under the influence of alcohol. 122

Court's theory was that although they were testimonial, and

derstanding of the word witness. Witnesses testify; blood does not.

E. What's the Big Idea?

lying this "look ma, no hands" idea is fuzzy -especially in light of

ment right to compulsory process. 171 Regarding her own perform-

171. See supra section I.A.

the privilege's development. 176 It now appears that the privilege

with the crown. During the sixteenth and seventeenth centuries,

draw. 8 7 In the mid-nineteenth century, police departments

II. THE SOLUTION

Current Fifth Amendment doctrine is a quagmire. But there is

jury. 191 In the case of criminal depositions, this compulsion would

led to would be admissible. 194 The defendant could also be subpoe-

194. An additional possibility is to allow a polygraph test to be conducted on the defend-

an innocent person may say seemingly inculpatory things under

cannot use his testimonial admissions - the words themselves -

nothing less than the right to mount a defense - a right of obvious

202. See generally Westen, supra note 10, at 182-84.

Compared to the current morass, the compulsion line in our so-