Criminal Procedure: Constitutional Aspects

Criminal procedure is literally at the center of the U.S. Bill of Rights, as a quick glance at the Fourth, Fifth, and Sixth Amendments makes evident. But as a subfield of constitutional law, constitutional criminal procedure stands as an anomaly, with shaky historical and conceptual foundations.

In many other areas of constitutional law, major opinions of the Marshall Court helped lay the groundwork upon which modem jurisprudence builds. In thinking about judicial review and executive power, lawyers and their fellow citizens still look to Marbury v. Madison, 5 U.S. (1 Cranch) 137 (1803); in considering the basic structure of federal jurisdiction, we ponder Martin v. Hunter's Lessee, 14 U.S. (1 Wheat.) 304 (1816); in reflecting on the scope of Congress' enumerated powers, and related issues of federalism, we refer back to McCulloch v. Maryland 17 U.S. (4 Wheat.) 316 (1819); in considering vested property rights, we return to Fletcher v. Peck, 10 U.S. (6 Cranch) 87 (1810), and Dartmouth College v. Woodward 17 U.S. (4 Wheat.) 518 (1819); and so on. But no comparable Marshall Court landmarks dot the plain of constitutional criminal procedure.

It is often thought that the explanation for this anomaly lies in another Marshall Court landmark, Barron v. Baltimore, 32 U.S. (7 Pet.) 243 (1833). Most criminal law, the argument goes, is state law: murder, rape, robbery, and the like are generally not federal crimes. Under Barron, the constitutional criminal procedure rules of the Bill of Rights did not apply against states, and so the Marshall Court predictably heard few cases raising issues of constitutional criminal procedure.

Barron is indeed part of the story, but only part. The federal government was very much in the crime-fighting business in the first century of the Bill of Rights. For constitutional scholars, perhaps the most vivid example of early federal criminal law comes from the infamous Sedition Act of 1798; but we must also not forget the territories. Perhaps the most central and sustained project of the federal government in its first century was the "Americanization" of this continent through territorial expansion, organization of territorial governments, and eventual admission to statehood of these territories. In the territories, the federal government did indeed enforce criminal laws against murder, rape, robbery, and so on. And the Bill of Rights applied to these criminal cases, even under Barron. Territorial law was, constitutionally speaking, federal law.

But for virtually the entire first century of the Bill of Rights, the U.S. Supreme Court lacked general appellate jurisdiction over federal criminal cases. This little-known fact helps explain why, for example, the Sedition Act prosecutions in the late 1790s—which raised the most important and far-reaching constitutional issues of their day—never reached the Supreme Court for ultimate judicial resolution.

By the time Congress decided to give the high court general appellate review over federal criminal cases in 1891, the sun was already setting on the territorial era. Thus, the criminal cases the Supreme Court heard under the new jurisdictional regime were a skewed lot, with disproportionately more federal customs violations, tax evasions, and bootleggings than murders, rapes, and robberies. It was during this era that the intellectual and conceptual foundations of some of today's most controversial criminal procedure rules were laid.

In the 1886 case Boyd v. United States, 116 U.S. 616, the Court melded the Fourth Amendment rule against unreasonable searches and seizures with the Fifth Amendment ban on compelled self-incrimination to suppress various papers that the government had in effect subpoenaed and sought to use in court against the target of the subpoena. The Boyd Court thus laid the intellectual groundwork for what later become known as the Fourth Amendment exclusionary rule: inculpatory evidence discovered in violation of the ban on unreasonable searches and seizures cannot be introduced in criminal cases. (Today, of course, this "rule" has many exceptions.)

The history of the exclusionary rule is interesting to trace, and nicely illustrates some of the larger problems of modern American constitutional criminal procedure—in particular, the awkward relationship between current doctrines and founding principles. Prior to Boyd, no court in America—state or federal—had ever excluded evidence on the ground that it was unconstitutionally obtained. Virtually every state had a state constitutional counterpart to the Fourth Amendment, so the sheer number of cases admitting such evidence is staggering to contemplate. Indeed, in a famous 1822 circuit court case (United States v. La Jeune Eugenie, 26F. Cas. 832), the scholarly Justice Story rejected the exclusionary rule in a dismissive opinion proclaiming that he had never heard of such an outlandish idea in Anglo-American law. (England never had anything like the modem American exclusionary rule—in the words of one famous 1861 English case (The Queen v. Leathram, 121 Eng. Rep. 589 (Q.B.)), "It matters not how you get it [evidence, that is]; if you steal it even, it would be admissible.")

Post-Boyd cases, like Weeks v. United States, 232 U.S. 383 (1914), helped crystalize the emerging American exclusionary rule in situations where various papers and other testimonial materials were seized; the key theory of these cases was that to introduce the seized papers against their owner in a criminal case would be akin to forcing the defendant to testify against himself in violation of the Fifth Amendment ban against compelled self-incrimination. Thus the exclusionary rule was born not simply to undo or remedy a past (out-of court) violation of the Fourth Amendment, but also to prevent a later (in-court) violation of Fifth Amendment principles. Only this fusion of the Fourth and Fifth Amendments can explain basic features of the exclusionary rule—for example, the fact that unconstitutionally seized evidence may be used in civil cases, as in United States v. Janis, 428 U.S. 433 (1976), or against persons other than the searchee, as in Agnello v. United States, 269 U.S. 20, 35 (1925). These limits are hard to explain in purely Fourth Amendment terms—for example, the amendment nowhere distinguishes between civil and criminal cases. But these limits are more readily explicable under a Fifth Amendment theory: under the explicit words of this amendment a person cannot be compelled to be a witness against himself in his own criminal prosecution, but he can, for example, be obliged to testify against himself in a civil case—and so the use of his seized papers in a civil context likewise presents no Fifth Amendment problem.

Post-Weeks cases then began to exclude from criminal cases even nontestimonial evidence—physical things rather than private papers—that had been unconstitutionally seized from the searchee. The theory seems to have been that a person's lawful private property was an intimate part of himself, and, therefore, forcing a person's property to "testify" against him was akin to forcing him to testify against himself. This odd property fetishism harmonized with the strong protections of property characteristic of federal case law at the turn of the twentieth century—often referred to today as the Lochner era, in recognition of the famous case Lochner v. New York, 198 U.S. 45 (1905).

By mid-century, the Court was explicitly using the exclusionary rule even in situations where the seized evidence was plainly not the searchee's property but was instead contraband or stolen goods. The Court, however, did not squarely confront the logical implications of this extension, and most major Supreme Court exclusionary rule cases from 1920 through 1960 continued to rely upon an interplay of the Fourth and Fifth Amendments, explicitly invoking both. For example, in Mapp v. Ohio, 367 U.S. 643 (1961), which held that states, too, were bound by the federal exclusionary rule, Justice Black's critical fifth vote was explicitly premised on a Boyd-brained theory of Fourth-Fifth fusion; and Justice Clark's opinion for the Court contained no less than six express or implied invocations of the Fifth Amendment.

In several more recent landmark cases, however—such as Schmerber v. California, 384 U.S. 757 (1966), Fisher v. United States, 425 U.S. 391 (1976), and United States v. Leon, 468 U.S. 897 (1984)—the Justices have explicitly and emphatically rejected the Fourth-Fifth fusion theory. Boyd itself is no longer good law, in many respects. The result, critics have argued, is a modem exclusionary rule without a firm conceptual basis. Once Fourth-Fifth fusion is rejected (and properly so, say the critics), what is the constitutional principle requiring exclusion? If the idea is that government should never profit from its own wrong, then must it return stolen goods to the thief, and a kidnap victim to the kidnapper, if it finds them in an unconstitutional search? Must it likewise refrain from using the evidence in civil cases? If evidentiary use of unconstitutionally seized material violates judicial integrity, why is such use generally admissible in civil cases, or in criminal cases against someone other than the searchee? If the rule is really justified by deterrence, aren't other possible remedial schemes capable of providing more deterrence, with more careful remedial tailoring between right and remedy, and at less overall social cost? Critics in particular try to point to the obvious conceptual and practical advantages of alternative remedial systems: damage awards and administrative remedies that would provide more compensation and justice for innocent citizens for whom the exclusionary rule rings hollow. (If the police find no incriminating evidence there is nothing to exclude; if they know they will find no evidence, but simply seek to harass, there is no deterrence whatsoever.) And many forms of unreasonable police behavior—police violence for example—have little or no causal nexus to the finding of evidence. To the critics, the exclusionary rule cannot work to remedy these Fourth Amendment violations; and so they believe there is need to devise sound alternative mechanisms. But once these mechanisms are truly in place, exclusion would not be necessary or proper—at least as a constitutional mandate. Or so the critics argue. (Many also argue that the current system of alternatives to exclusion is inadequate and should be beefed up in a variety of ways.)

The Court, however, has never squarely confronted this set of basic remedial questions, because the Justices, as explained above, did not originally conceive of exclusion in modern deterrence-based remedial terms. Indeed, as late as 1961, when the Court imposed the rule on states, there was not in place anything remotely like a proper system of federal damage remedies for state constitutional torts: A federal law allowing damages for constitutional violations had lain dormant for almost a century, and was only revitalized by the Court in the early 1960s, beginning with the landmark case of Monroe v. Pape, 365 U.S. 167 (1961). Nor had the Court yet decided that the Fourth Amendment itself provided a cause of action for damages against offending federal officials (Bivens v. Six Unknown Named Agents of the Federal Bureau of Narcotics, 403 U.S. 388 (1971)). As a result, the exclusionary rule took root in a remedially impoverished milieu, and without clear analysis of the relative advantages and disadvantages of alternative remedial schema.

The conceptual confusion surrounding the current exclusionary rule is mirrored by considerable confusion about the basic meaning and purpose of the Fifth Amendment rule prohibiting compelled self-incrimination. Soon after Boyd was decided, the Court began to develop its Fifth Amendment self-incrimination jurisprudence. In Counselman v. Hitchcock, 142 U.S. 547 (1892), the Court faced the following question: what immunity must a person receive before he may be obliged to speak outside his own criminal case, for example, in a civil case brought against him, or a grand jury proceeding, or a legislative hearing? The Counselman Court insisted that a person be given total ("transactional") immunity from prosecution in such a situation. Not only could his compelled words never be used against him, but also no physical evidence discovered as a result of his testimony could be used against him; indeed, no prosecution could ever be brought against him, even if all the evidence adduced was proved to be wholly independent and nonderivative. Building on Boyd, the Counselman Court thus held that the self-incrimination clause should be read expansively to support broad principles of evidentiary exclusion. Later cases have restricted Counselman—allowing prosecution where all the evidence introduced is shown to be independent—but none of these cases has explained exactly where the new rule of "use plus use-fruits" immunity (or the old rule of "transactional immunity," for that matter), comes from. Nor have these cases explained exactly what the Fifth Amendment stands for, and why.

Some modern critics have argued for a narrower rule of "testimonial immunity" under which a person may be compelled to speak outside his criminal case, with the secure immunity that his compelled words cannot ever be introduced against him in a criminal proceeding. But, say proponents of testimonial immunity, fruits of the immunized statement—any leads that the government tracks down as a result of the statement—should generally be admissible. The Constitution prohibits only the introduction of the words themselves; if those words are never introduced at trial, a person never will have been impermissibly compelled to be a testifying "witness" against himself in a criminal case (i.e., a trial). And the purpose of this narrow rule of immunity and of the Fifth Amendment itself, it is argued, it to protect an innocent person from erroneous conviction, rather than to protect a guilty person as such. Innocent folk, when questioned by clever prosecutors, can sometimes be made to look guilty, and so they should not be forced to speak before the jury, or in some deposition that can later be read to the jury. But if their words lead to reliable "fruit"—a murder weapon with fingerprints, for example—that fruit should itself be admissible under the letter and spirit of the Fifth Amendment, say the critics. This was indeed the dominant rule in America (and in England) prior to Boyd and Counselman. The critics' account also helps explain the intuition behind more modern cases like Schmerber (1965), which allow the government to force a criminal defendant to give a blood sample or a handwriting exemplar or voice sample, on the theory that these reliable bits of physical evidence are not the kind of unreliable "witnessing" that the Amendment seeks to protect against. Current Fifth Amendment self-incrimination doctrine, however, has not embraced the critics' revisionist account; but the Court has left the field in a state of intellectual disarray, in which it is far from clear what the main purpose of the Amendment truly is. Lacking even the most basic consensus on the big idea (or ideas) underlying the Amendment, the Court has had difficulty defending the many ways in which it has invoked or limited the clause.

Boyd and the Fourth and Fifth Amendment cases that followed it for the next two generations mostly dealt with corporate and regulatory offenses, because these amendments at the time applied only against the federal government, and the federal government had a rather limited criminal agenda at the turn of the twentieth century. These cases established controversial, pro-defendant rules, but almost none of them dealt with violent crime. It was not until the Warren Court (1953–1969) incorporated these and many other criminal procedure clauses against the states that they began to be applied regularly to violent crimes. Under an approach eventually labeled "selective incorporation"—the theory whereby the Fourteenth Amendment incorporates against the states those provisions of the Bill of Rights deemed "fundamental"—the mid-century Court began to apply the criminal procedure clauses to the states. Although most of this incorporation took place during the Warren Court, some of its roots lie in the 1930s and 1940s.

In the famous "Scottsboro Boys" case of Powell v. Alabama, 287 U.S. 45 (1932), involving poor black defendants tried in a highly racially charged and unfair proceeding, the Court held that the right to a fair trial and the right to appointed counsel in capital cases were necessary components of due process of law under the Fourteenth Amendment, and thus enforceable against the states. In the 1948 case In re Oliver, 333 U.S. 257 (1948), the Court in effect held that the right to a public trial—a right explicitly protected against the federal government by the Sixth Amendment—is also an inherent part of Fourteenth Amendment due process, and thus no state trial can ever take place in secret. And in the 1949 case Wolf v. Colorado, 338 U.S. 25, the Court made the Fourth Amendment's protection against unreasonable searches and seizures applicable to the states, but declined to saddle states with the exclusionary rule, choosing instead to allow states to fashion their own remedies for unreasonable searches and seizures.

In one of the most important criminal procedure decisions of the Warren Court, Mapp v. Ohio (1961), the Justices modified Wolf and incorporated the exclusionary rule against states. Two years later, the Court held that defense counsel is a necessary part of a fair criminal case. Thus, in Gideon v. Wainwright, 372 U.S. 335 (1963), the Court decided that the states must provide attorneys to indigent defendants in all felony cases, not just capital cases, as the Powell Court held.

The next year, in Malloy v. Hogan, 378 U.S. 1 (1964), the Court explicitly began to conceptualize its task as deciding which clauses of the Bill of Rights were so fundamental as to apply against states in every jot and tittle—the approach now known as "selective incorporation." Eventually, this process would lead to the application of almost all of the Bill of Rights against the states. This application derives strong support from the original intent of the Fourteenth Amendment, whose framers saw the provisions of the Bill of Rights as paradigmatic "privileges and immunities of citizens" that no state should be allowed to abridge. (The Justices have tended to emphasize the language of "due process" rather than "privileges and immunities," however, in part to avoid the need to confront an early emasculation of the Fourteenth Amendment's privileges or immunities clause in the Slaughterhouse Cases, 83 U.S. (16 Wall.) 36 (1873)).

Malloy held that the Fifth Amendment right against self-incrimination was made applicable to the states via the Fourteenth Amendment, and the year after that, in Pointer v. Texas, 380 U.S. 400 (1965), incorporated the defendant's Sixth Amendment right to confront witnesses against him. Soon thereafter, the Court incorporated the Sixth Amendment right to speedy trial (Klopfer v. North Carolina, 386 U.S. 213 (1967)); the Sixth Amendment right to compulsory process (Washington v. Texas, 388 U.S. 14 (1967)); the Sixth Amendment right to jury trial (Duncan v. Louisiana, 391 U.S. 145 (1968); the Fifth Amendment right against double jeopardy (Benton v. Maryland, 395 U.S. 284 (1969)); and the Eighth Amendment right to bail (Schlib v. Kuebel, 404 U.S. 357 (1971) (dictum)).

At the same time that the incorporation movement was gaining speed, the Warren Court also profoundly revised the law of police station interrogation. In Escobedo v. Illinois, 378 U.S. 478 (1964), the Court excluded a confession obtained during a prolonged interrogation of a suspect who had been advised of neither his right to an attorney nor his right to remain silent. Escobedo had asked for a lawyer several times at the police station, but his request was denied. In fact, Escobedo's mother had retained a lawyer for him, who came to the police station but was not allowed to confer with the client. Two years later, in Miranda v. Arizona, 384 U.S. 436 (1966), Chief Justice Warren went much further, holding that prosecutors could never use statements made by defendants while in custody unless the prosecutors could demonstrate the use of procedural safeguards to protect against Fifth Amendment violations. These safeguards, the Court suggested, should amount to informing the defendant of her right to remain silent, of the fact that any statement she did make could be used against her, and of her right to an attorney, either hired by her or appointed by the state. Under the Miranda regime, a defendant in custody was free to remain silent without any adverse inference of guilt. Perhaps even more exuberant than Miranda (though less politically salient) was the Court's ruling in Massiah v. United States, 377 U.S. 201 (1964), which held inadmissible statements obtained from a sting operation of an indicted defendant, who—while not in custody, and free from any compulsion whatsoever—bragged about his crimes to someone who, unbeknownst to him, was an undercover agent.

These and other Warren Court cases constituted nothing less than a revolution in criminal procedure. By the end, the Court's "selective incorporation" resulted in the application of all of federal constitutional criminal procedure against the states, except the Fifth Amendment right of grand jury indictment. In the process, the nature of the defendants invoking constitutional criminal procedure rights changed radically. Now the exclusionary rule, the right against self-incrimination, and the rest, were no longer limited to cases of corporate and regulatory offenses, but also applied to murder and robbery. To many ordinary persons and some scholars, it seemed outrageous to see murderers and thieves going free because, for instance, clearly relevant and inculpatory evidence was found in technical violation of the Fourth Amendment. Many other scholars, however, have lamented the passing of the Warren Court "revolution," which they believe provided a needed antidote to alleged governmental overreaching.

The mid-century Court's constitutional criminal procedure doctrine also seemed to some to suffer from some serious legal flaws. For instance, the Court repeatedly held that the Fourth Amendment generally requires warrants and probable cause for all searches and seizures. But the Amendment, when read carefully, does not say this. No early state constitution says this; and no leading Founder or early case or treatise says this, either. The Amendment says simply that all searches and seizures must be reasonable, without further saying that reasonableness always or even presumptively means a warrant or probable cause. Historically, a vast number of serious intrusions—arrests and border searches, to name just two—have not required warrants; and as a practical matter, there are many situations where it would not be sensible to require even probable cause. (Consider for example metal detectors at airports and in court houses.) The Warren Court itself admitted as much in the 1968 case Terry v. Ohio, 392 U.S. 1, authored by Earl Warren himself speaking for eight Justices. Terry upheld a police weapons "frisk" for self-protection even though no probable cause existed—and thus a warrant could never have been properly issued (since the Amendment does insist that all warrants be backed by probable cause). But Terry no-where explained how this result could be squared with the Court's general insistence that the Amendment did ordinarily require warrants and probable cause. Critics have argued that the warrant requirement is textually incorrect, historically mistaken, and functionally suspect, and that the Court's efforts to adhere to this requirement while also recognizing various exceptions is intellectually bankrupt. By 1991, Justice Scalia counted about twenty exceptions to the warrant "requirement" (California v. Acevedo, 500 U.S. 565, 582-83 (1991) (Scalia, J., concurring in judgment)).

As noted above, the Court's self-incrimination case law is likewise a conceptual muddle. Moreover, the Court's interpretation of the Fifth Amendment's self-incrimination clause seems to conflict with the Sixth Amendment right to compel testimony in one's defense. Under the Court's 1972 ruling in Kastigar v. United States, 406 U.S. 441, a person compelled to testify against himself outside his own criminal case (say, in a legislative hearing, or someone else's criminal prosecution) must receive "use plus use-fruits immunity"—that is, neither the testimony nor any evidence it led to could later be used against him. As a result of Kastigar, an innocent defendant will often be denied the right to compel the testimony of a witness who, the defendant claims, is in fact the guilty culprit. To give every defendant the right to compel another suspect to take the stand, as the Constitution's Sixth Amendment seems to require, would, under Kastigar, enable a guilty defendant to give all of his partners in crime a general "immunity bath"—an obviously unacceptable result. But this conflict between Fifth and Sixth Amendment rights exists only because of the broad "use plus use-fruits" immunity rule. The more narrow and historically supportable principle of testimonial immunity would restore to every defendant the right to compel defense witnesses on equal terms with the prosecutor: immunity baths are not a problem because the government loses nothing when a defendant forces another suspect to take the stand. Thus, the unhappy result of current doctrine is that an overbroad, nontextual, ahistorical reading of the Fifth Amendment requires an underprotective, nontextual, and ahistorical reading of the Sixth Amendment. Even worse, the Court's regime ends up overprotecting the interests of guilty defendants at the expense of innocent ones. And this overprotection—say critics—is a more general feature of the current landscape of constitutional criminal procedure. The Fourth Amendment exclusionary rule, Fifth Amendment immunity doctrine, Fifth Amendment double jeopardy doctrine, Sixth Amendment speedy trial rules concerning dismissal with prejudice—all these, say critics, end up in various ways protecting the guilty at the expense of (or with indifference to) the innocent.

In defense of the Warren Court, it must be said that the Justices were reacting to a regime that often seemed to run roughshod over the legitimate interests of the poor and racial minorities. Police departments were not integrated; remedial systems like section 1983 were not in place; and the third degree prevailed in a great many places. In many cases, juries were all-white because many blacks were disenfranchised, others were not allowed to be part of the jury venire, and still others were excluded by race-based peremptory challenges. One of the most impressive accomplishments of late twentieth-century constitutional criminal procedure has been a gradual integration of the jury, created by a combination of voting rights case law, voting rights statutes, jury reform statutes, and federal jury case law.

But, critics say, in many other areas of constitutional criminal procedure, the Warren Court overreacted by overconstitutionalizing, and by moving too quickly with rules that far outstripped the problems to be solved. More direct focus on civil remedies, on issues of class and poverty, on the rights of the innocent, on the nature of police brutality and police discretion, on the racial composition of police departments, and on other racial dimensions of the crime problem, may have been better than some of the indirect and overbroad strategies devised by the Court in cases like Mapp, Miranda, and Massiah. The interests of victims of crime—themselves disproportionately poor and black or brown, not to mention female—also seemed to get inadequate attention.

In sum, many of the criminal procedure rulings to come out of the Warren Court lacked firm grounding in constitutional text and structure. Key rulings ran counter to early case law both in lower federal courts and in state courts construing analogous provisions of state constitutions. Precisely because so few Marshall Court cases existed, the many breaks with Founding-era understandings were not highly visible. On key issues, the Warren Court seemed to contradict itself, laying down sweeping rules in some cases that it could not quite live by in other cases. On a political level, many of the Warren Court's constitutional criminal procedure pronouncements did not sit well with the American people. The guilty—who now included murderers, thieves, and rapists, and not merely the corporate and regulatory offenders of pre-incorporation days—too often seemed to go free because of Warren Court rules. Indeed, many of these rules seemed tailor-made to protect guilt per se.

The result was, predictably, something of a political backlash. Accusations that the Warren Court was too easy on the guilty have given way to accusations that the Burger (1969–1986) and Rehnquist (1986–) Courts are too hard on the innocent. Habeas corpus rules, expanded by the Warren Court in ways that may have overprotected the guilty, have been retracted by later Justices and by Congress in ways that may now underprotect the innocent.

As noted above, the post–Warren Courts have continued to carve out exceptions to the warrant requirement without replacing it with a clear concept of what reasonableness entails. Moreover, the Court has crafted an increasing number of exceptions to the Fourth Amendment exclusionary rule, most dramatically a habeas exception, Stone v. Powell, 428 U.S. 465 (1976); an impeachment exception, United States v. Havens, 446 U.S. 620 (1980); and a good faith exception, United States v. Leon (1984). The post–Warren Court has also chipped away at the self-incrimination clause doctrine of the Warren Court. For example, in Harris v. New York, 401 U.S. 222 (1971), a five-Justice majority held that incriminating statements made without a Miranda warning could be used to impeach the accused's testimony at trial. Miranda, the Court seemed to say, was merely a prophylactic rule, and not all Miranda-defective statements were truly "compelled" in a way that would trigger the underlying Fifth Amendment right. And in Oregon v. Elstad, 470 U.S. 298 (1985), the Court suggested that the physical evidence and other fruits generated by a "mere" Miranda violation need not be suppressed at trial.

In many other areas, however, the Burger and Rehnquist Courts have not reacted against Warren Court rulings, and, in some cases, they have furthered the Warren Court's trends. In dealing with the Sixth Amendment speedy trial clause, the Court has continued to insist that dismissal with prejudice—that is, dismissal with no possibility of refiling charges later—is the only possible remedy for violations (Strunk v. United States, 412 U.S. 434 (1973)). Perhaps most famously, since its ruling in In re Winship, 397 U.S. 358 (1970), the Court has continued to hold that the Fifth Amendment guarantee of due process of law requires that the prosecution prove guilt beyond a reasonable doubt—though defining the precise difference between offense elements governed by Winship and sentencing factors and affirmative defenses beyond the scope of Winship has not proved easy.

In one accepts the critical analysis set out above, this leaves the current state of constitutional criminal procedure in something of a mess. The Court has often seemed to proceed by indirection, at times distinguishing away precedent, at times ignoring it, at times adhering to it. When this new confusion is added to the contradictions that existed within the Warren Court, the state of constitutional criminal procedure becomes truly perplexing. The time thus seems ripe for an overhauling of the law. This seems even more likely given the current state of Court personnel: no one on the Court today served with Chief Justice Warren, and two-thirds of the Court never served with Chief Justice Burger. The question is how today's Justices should go about reorganizing the law. Precedent alone certainly cannot guide the way: precedent in this field is too often contradictory or perverse.

Some have recently suggested that constitutional criminal procedure be reevaluated with the protection of innocence as its primary purpose. Under this reading, the Fourth Amendment exclusionary rule would be discarded, as it protects only the guilty, as such. The exclusionary rule does not provide a direct remedy to innocent victims of unreasonable searches and seizures; no evidence is found in such searches, and so there is nothing to exclude. Instead, the exclusionary rule provides the most help to the most guilty—the proverbial murderer found with a bloody knife. A scheme of innocence protection would provide civil remedies, including but going beyond what the Court has done in Bivens and section 1983 cases. Such remedies would be available as a remedy for guilty defendants, as well as innocent ones, but would not aid guilty defendants because of their guilt, as does the exclusionary rule. Similarly, current Fifth Amendment self-incrimination case law could be profitably rethought. It makes sense in terms of protecting innocence (and innocents) to exclude compelled testimony itself: a cunning prosecutor may be able to make an innocent but inarticulate or unattractive defendant look guilty on the stand, even though he is not. However, the use of physical evidence does not raise these concerns. Indeed, the Fifth Amendment exclusion of physical evidence, like the Fourth Amendment exclusionary rule, protects defendants in direct proportion to their guilt. Thus, the rule of exclusion of "use fruits" as demanded by Kastigar, is dubious indeed. A scheme of only testimonial immunity—excluding possibly unreliable words without excluding reliable fruits—would continue to protect innocents while avoiding windfalls to the guilty. And as noted earlier, a regime of testimonial immunity would allow the Sixth Amendment to be construed more broadly, in ways that would aid innocent defendants who seek to prove their innocence at trial by compelling other suspects to take the stand. More generally, the due process clause, the double jeopardy clause, and the Sixth Amendment rights of speedy trial, counsel, compulsory process, and confrontation, should all be seen as having a strong unifying thread: The Constitution, rightly read, seeks to protect innocent defendants from erroneous convictions—it seeks to find the truth, not suppress it, in the interest of both the innocent, and society as a whole. Thus, by viewing innocence-protection as the primary purpose of the constitutional criminal procedure clauses, we may achieve a more unified, and more normatively attractive, interpretation linking the Fourth, Fifth, and Sixth Amendments—the centerpiece of the Constitution's Bill of Rights.