This chapter examines the largely failed efforts of the Supreme Court to find a coherent set of double jeopardy doctrines. The first three sections in the chapter coincide with the legislative-prerogative principle and supporting presumptions developed in chapter 2. Then I briefly describe the best of the solutions offered by commentators. None is any better than the Court’s efforts, save that of Michael Moore,1 and Moore’s theory can be usefully deferred to chapter 5 for its relevance in developing a detailed set of presumptions about the same-offense issue.
The jeopardy aspect of the problem is a good place to start, partly because it is easier to describe than the same-offense problem. Moreover, trying to determine when jeopardy ends is where the Court began. No Supreme Court case reached the merits of a same-offense claim until 1887.2 As early as 1824, however, the Court was deciding jeopardy-related cases. That 1824 case is today universally understood to state a test for determining when a mistrial ends jeopardy. But this may be a modernist misreading of the 1824 case; that Court, after all, relied on Blackstone’s pleas in bar, and these pleas offered no protection in the absence of a verdict.
As long as the courts understood double jeopardy to incorporate Black-stone’s pleas in bar, they were on firm interpretational ground, whatever one might think of the underlying policy issues. A mistrial would never bar a second trial, irrespective of the linguistic implications of “jeopardy,” if defendants had to show a former conviction or acquittal as a threshold requirement of a Blackstone-like jeopardy plea. When United States v. Perez3 was decided in 1824, the Court was only thirty-three years from the Bill of Rights ratification and only fifty-five years from the publication of Book IV of Blackstone’s Commentaries. The presence of a strong Blackstone influence is not surprising, particularly on so astute a scholar as Justice Story, the author of Perez.
The jury was unable to agree in Perez’ capital case and was discharged by the trial court “from giving any verdict upon the indictment, without the consent of the prisoner, or of the attorney for the United States.”4 Story noted “some diversity of opinion and practice on this subject, in the American courts,”5 but the Supreme Court seemed to find the issue quite easy. Perez could not prevail because he had “not been convicted or acquitted, and may again be put upon his defence.”6 This certainly sounds as if the Double Jeopardy Clause was limited to Blackstone’s pleas of former conviction and former acquittal. So limited, of course, the mistrial doctrine would be easy to apply, just as it would have been under Madison’s prohibition of more than one trial. The sets of outcomes would be different under the two approaches, because Blackstone’s emphasis was on verdict rather than trial, but both rules are hard edged.
The twentieth-century reading of Perez is precisely the opposite from what Story seems to have intended. Rather than incorporate Blackstone’s hard-edged requirement of a verdict to end jeopardy, modern courts have used Perez to justify a mistrial rule that has extremely soft edges—edges so soft they are almost invisible. The modern mistrial rule purports to protect the “valued right” of each defendant to each jury but, unlike Madison’s proposed language, operates without any substantive principles to guide the inquiry. It seems likely that the modern mistrial rule is a product of the general movement in the twentieth century toward seeking justice in individual cases. This movement has favored soft-edged rules that give judges discretion to decide for defendants when injustice seems pronounced. The legal realist approach discussed in chapter 1 is an outgrowth of this movement.7 The question remains, however, how courts managed to use Perez to achieve the rule demanded by modern legal culture.
Looking for rules that promise perfect justice in individual cases, courts read Story’s dicta as if they were part of the holding. The dicta seem much more likely, however, to be Story’s effort to give guidance to trial courts about how to exercise what he viewed as their unreviewable discretion. Following the remark that because Perez lacked a conviction or acquittal, he “may again be put upon his defence,” Story wrote:
We think, that in all cases of this nature, the law has invested courts of justice with the authority to discharge a jury from giving any verdict, whenever, in their opinion, taking all the circumstances into consideration, there is a manifest necessity for the act, or the ends of public justice would otherwise be defeated. They are to exercise a sound discretion on the subject; and it is impossible to define all the circumstances, which would render it proper to interfere. To be sure, the power ought to be used with the greatest caution, under urgent circumstances, and for very plain and obvious causes; and, in capital cases especially, courts should be extremely careful how they interfere with any of the chances of life, in favor of the prisoner. But, after all, they have the right to order the discharge; and the security which the public have for the faithful, sound and conscientious exercise of this discretion, rests, in this, as in other cases, upon the responsibility of the judges, under their oaths of office.8
Story could hardly have been clearer that the decision to discharge a jury when it could not agree on guilt or innocence is completely discretionary with the trial judge. The manifest necessity discussion seems intended to impress upon trial judges the seriousness of the decision to discharge a jury. Or perhaps Story was suggesting how seriously judges already take this responsibility as a way of justifying the Court’s refusal to impose a standard of review. But, in either event, he said clearly that discharge was the right of trial judges, the only limitation being their oaths of office.
It is really quite perverse that Perez is now known and cited as the case establishing the principle that trial courts can discharge a jury only upon a finding of manifest necessity, which finding can be reviewed in the appellate system. That may be a better system than the one Story envisioned, but it is decidedly not Story’s system. When he used the “manifest necessity” language, he modified it with “in their [trial judges’] opinion.” Moreover, no reason existed to mention that judges were limited by their oaths of office if Story intended for appellate courts to review the mistrial decisions of trial courts.
More evidence in favor of a Blackstonian interpretation of Perez is Justice Washington’s presence on the Perez Court (the decision was unanimous). The year before, while sitting as a circuit judge, Washington had rejected a mistrial bar to a second trial on the ground that “the jeopardy spoken of in [the Fifth Amendment] can be interpreted to mean nothing short of the acquittal or conviction of the prisoner, and the judgment of the court thereupon.”9 Washington’s statement seems clear enough, and his joining of the Perez opinion suggests at least how he viewed Story’s language about “manifest necessity.”
The most telling evidence that Perez manifests a Blackstonian, verdict-or-nothing proposition about the ending of jeopardy is Story’s treatment of Perez nine years later in his Commentaries on the Constitution.10 Story wrote there that the Double Jeopardy Clause “does not mean, that [a person] shall not be tried for the offense a second time, if the jury shall have been discharged without giving any verdict.” In that case, Story concluded that the defendant’s “life or limb cannot judicially be said to have been put in jeopardy.” As authority for that proposition, Story cited Blackstone, Justice Washington’s opinion quoted earlier, and Perez.11 However others have understood Perez, Story understood it to state Blackstone’s rule requiring a verdict to end jeopardy.
Many years after the Court began to use manifest necessity as a test for appellate review, it admitted doubt that Story intended to state a double jeopardy rule. In Crist v. Bretz,12 the Court sought to justify “manifest necessity” as a rule necessary for the “administration of federal criminal justice.”13 While conceding that Story might not have meant to locate manifest necessity in the Double Jeopardy Clause, the Court insisted that Story meant to state a rule that, whatever its origin, could be conveniently packaged into the modern Double Jeopardy Clause.14 But no rule is apparent in Story’s Perez opinion, except the rule that there is no rule. Trial courts decide whether to grant mistrials; appellate courts have no role to play.
The modernist use of “manifest necessity” as a standard for reviewing mistrials granted by trial courts has not gone well. For many years, the Court ignored the indefiniteness of the standard, perhaps assuming that decades of case law would provide some meaning to the meaningless. In 1978 the Court gave up and acknowledged that “manifest necessity” does not provide sufficient guidance. In Arizona v. Washington,15 the issue was whether the trial judge violated the Double Jeopardy Clause when he granted the prosecutor’s motion for mistrial to cure improper remarks made in defense counsel’s opening argument. In discussing the Perez “test,” the Court wrote,
Mr. Justice Story’s classic formulation of the test has been quoted over and over again to provide guidance in the decision of a wide variety of cases. Nevertheless, those words do not describe a standard that can be applied mechanically or without attention to the particular problem confronting the trial judge. Indeed, it is manifest that the key word “necessity” cannot be interpreted literally; instead, contrary to the teachings of Webster, we assume that there are degrees of necessity and we require a “high degree” before concluding that a mistrial is appropriate.16
The Court tells us that, in the manifest necessity standard, it is “manifest” that “necessity” cannot be interpreted literally. Is this a joke? And what does the Court mean by a “high degree” of “necessity”?
The Court acknowledged that, in the case before it, “some trial judges might have proceeded with the trial after giving the jury appropriate cautionary instructions.”17 Well, if it was not only possible but plausible to continue the trial, there can be no doubt that the “high degree” of “necessity” was missing in Washington. But that would be wrong.
The Court conceded that “[i]n a strict, literal sense, the mistrial was not ‘necessary.’” But, the Court, continued, the “overriding interest in the evenhanded administration of justice requires that we accord the highest degree of respect to the trial judge’s evaluation of the likelihood that the impartiality of one or more jurors may have been affected by the improper comment.”18 This time the Court articulated a standard phrased in terms of “highest degree.”
In sum, a mistrial is a jeopardy bar unless justified by a “high degree” of necessity. But because the Court has no idea how to apply that standard, it developed a deferential attitude toward what judges do: appellate courts must accord the “highest degree of respect” to a trial judge’s estimate of the effect of trial error on the jury. As we should expect, the “highest” degree standard trumps mere “high degree,” and Washington lost his manifest necessity claim. Whenever a judge grants a mistrial to cure prejudicial error heard by the jury, the highest degree of respect for the trial judge’s evaluation will always (or almost always) provide the high degree of necessity to justify a mistrial.
The Washington gloss on “manifest necessity” can be read as a pragmatic reinstatement of Blackstone (and my reading of Story’s opinion in Perez) for this category of mistrials: defendants are going to lose mistrial claims in which prejudicial error is heard by the jury and the judge decides the best course of action is to grant a mistrial. But if this is the set of results we want, why not achieve it more directly (and honestly) by requiring a verdict to end jeopardy? The Washington gloss on Perez could manifest a policy that mistrials are necessary unless they interfere with a likely acquittal—which would not have been the case in Washington—but the Court has never put the policy in this way. Thus, courts are left on their own to puzzle out if they should ever refuse to give the “highest degree or respect” to the trial court’s mistrial judgment.
Moreover, Washington provides no guidance for mistrials that do not involve the jury hearing prejudicial error. What kind of respect is due the trial judge in those cases? The Court has provided only the vaguest of clues, which I will seek to explicate in Chapter 8. Chapter 8 also contains an argument that the best mistrial formulation begins with the hard-edged Blackstone-Story rule and then amends the rule in a way that makes clear when double jeopardy policy requires a different outcome: when the mistrial manifests a likelihood of acquittal had the trial gone to verdict. The resulting rule is both more sensitive to double jeopardy policy and far more determinate than the present manifest necessity “rule.” “Manifest necessity” should be reburied in the original grave dug by Joseph Story.
Beyond “manifest necessity,” the Court has done little damage in the jeopardy area. The mechanical concept of jeopardy “attaching” is unnecessary if my jeopardy presumption is followed, but “attachment” has rarely produced an unjustified result. The Court also demonstrated confusion for a short time about when a dismissal should count as a jeopardy bar—in 1975 the Court unanimously held that any dismissal barred a second trial that would require new factfinding, but overruled that case by a 5-4 vote three years later.19 If only verdicts or endings like verdicts are an end to jeopardy, the Court was right the second time to reject this rule, which is based on preventing generalized harassment and not on protecting the defendant’s right to proceed when an acquittal is likely.
But the Court has otherwise been sensitive to the fine distinctions between an outcome that has resolved the defendant’s blameworthiness and one that has not. Burks v. United States 20 is a good example. The defendant moved for a new trial after conviction, on the ground that the state presented insufficient evidence of his sanity and, therefore, that the trial judge should have directed an acquittal. The court of appeals agreed; it vacated the conviction and granted the motion for a new trial at which the state and defendant would have a new chance to litigate the sanity issue.
The Supreme Court considered conflicting precedents in Burks. The earliest Supreme Court case on criminal appeals held that a successful appeal (in that case on the ground of an insufficient indictment) removed the conviction as a bar to a second trial.21 A second kind of precedent supporting the court of appeals was the rule that “a defendant who requests a new trial as one avenue of relief may be required to stand trial again, even when his conviction was reversed due to failure of proof at the first trial.”22 The theory behind this rule is simple: by reversing and remanding, an appellate court is simply giving a defendant what he requested.
On the other side was the precedent that an acquittal always bars a new trial. The reason Burks was not an easy case, of course, is that the trial judge did not grant the motion for a directed verdict of not guilty. The appellate court held that the judge should have granted the motion, but does that have the same double jeopardy effect as the trial judge granting the motion?
The two events might be different on the following argument. When the trial judge grants a motion for acquittal, he is acting as a surrogate for the jury; he has heard the evidence, witnessed the demeanor of the witnesses, and his ruling that no conviction could result is informed in a way that no appellate decision could be. When an appellate court rules that the judge should have granted a motion for acquittal, perhaps the case should be resubmitted for a new fact-finding.23
Burks easily distinguished the precedent about flawed indictments. An error in an indictment is different in kind from an error in failing to grant a directed verdict; the first error goes to the procedure for determining guilt and innocence, the second to the very question of guilt and innocence in the case presented by the state. The second precedent, giving the defendant the new trial that he requested, was more difficult to distinguish. This doctrine had direct support in Supreme Court cases and thus could have resulted in the Court denying certiorari on the grounds that the case was an easy one. The Court had to reason to a conclusion that the easy answer was the wrong one. The easy answer seems, at one level, fair. How is a defendant injured if a court gives him exactly what he requested?
But this precedent is logically defective. Writing for a unanimous Court, Chief Justice Burger noted that an appellate reversal for insufficient evidence “means that the government’s case was so lacking that it should not have even been submitted to the jury. Since we necessarily afford absolute finality to a jury’s verdict of acquittal—no matter how erroneous its decision—it is difficult to conceive how society has any greater interest in retrying a defendant when, on review, it is decided as a matter of law that the jury could not properly have returned a verdict of guilty.”24
This quote discloses the premise that underlies the Court’s decision. The Court assumed that an acquittal by a jury is always the end of jeopardy. As this has been the rule since A.D. 1200, it required no more than a passing mention—“[s]ince we necessarily afford absolute finality to a jury’s verdict of acquittal.” Once that premise is securely in place, however, everything else about Burks follows nicely.
Burks thus supports a presumption that jeopardy ends only upon a formal verdict or an outcome that demonstrates the defendant should have been (or would have been) acquitted. If the Court simply applied the same rule to mistrials, it would have a coherent jeopardy principle. But trapped as it is in the modernist misreading of Perez, the Court does not seem even to appreciate that mistrials, dismissals, and reversals on lack of evidence can be analyzed under the same substantive frame-work—whether the defendant was denied an acquittal to which he was entitled. This doctrinal unity manifests the same policy against erroneous convictions.
In sum, the Court’s jeopardy jurisprudence suffers largely from a failure to understand mistrials as a type of dismissal that should be analyzed by looking to the substantive effect of denying the defendant a chance to proceed to verdict. A more difficult task for the Court has been finding clarity in the “same offense” part of the Double Jeopardy Clause. To find both clarity and a test sensitive to double jeopardy policy is a daunting challenge.
Defining “jeopardy” is not easy, but linguistics and a long-accepted common law understanding are available to assist the inquiry. There are no similar aids to interpreting “same offense.” Bishop agreed, noting that the “same-offense” decisions are even more “discordant” than the twice-in-jeopardy cases.25 Jay Sigler predicts that, without recodification of the substantive criminal law, “the problem is virtually insoluble.”26 The problem is serious but not insoluble. Linguistics is not helpful, as we have seen, because the most logical referent suggested by “same offense” is “same statutory offense.” Alternatively, same offense could include only greater- and necessarily-included offenses, but this, too, suffers from problems of overinclusiveness and underinclusiveness. Moreover, the Blockburger definitional test has absolutely nothing to say about limits on dividing the relevant conduct to prove multiple counts. Is robbing six victims a single offense? One would think not. Is stealing four items from a dresser a single larceny? One would think so. But we have no rule for this aspect of same offense.27
History offers some clues to the best understanding of same offense, and we can begin there.
The thirteenth-century commentators—Bracton, Britton, Fleta, and the author of The Mirror of Justices —did not perceive a “same offense” problem. The self-evident nature of “same offense” in the 1200s reflected the rudimentary nature of criminal law at that time, which recognized only a few broad categories of conduct as separate crimes. This, of course, is consistent with an act-equals-offense presumption.
Hale has the most comprehensive discussion of same offense prior to 1800. He recognized that different offenses can occur close together. But he saw the distinctiveness of acts as the key.
If A commit a burglary … and likewise at the same time steal goods out of the house, if he be indicted of larciny for the goods and acquitted, yet he may be indicted for the burglary notwithstanding the acquittal. And e converso, if indicted for the burglary and acquitted, yet he may be indicted of the larciny, for they are several offenses, tho committed at the same time.28
Hale offered a test of sorts for determining when offenses may be separate, though closely connected in terms of acts: “burglary may be where there is no larciny, and larciny may be where there is no burglary.”29 A version of Hale’s test was eventually adopted by the Supreme Court in the 1915 case of Morgan v. Devine:“the test is not whether the criminal intent is one and the same and inspiring the whole transaction, but whether separate acts have been committed with the requisite criminal intent and are such as are made punishable by the act of Congress.”30
While the distinctiveness in blameworthy acts is critical to an act-equals-offense presumption, “act” here is not the metaphysical “act” that some courts (and some commentators) assume. Indeed, even if we could avoid the skeptical critique Austin lodged against metaphysical accounts of acts (as Michael Moore tries to do in Act and Crime), we would still be left with odd double-jeopardy results. An example is the Court’s very first same-offense case, In re Snow.31 The issue was whether Snow could be prosecuted for three counts of cohabitation with the same women over an unbroken period of thirty-five months, all of which occurred before the first indictment.32 It is, of course, true that different cohabitations are different offenses, just as different robberies are different offenses. To cohabit with person X in 1990 must be a different offense than to cohabit with person Z in 1991. Otherwise, a person who committed one particular statutory offense could commit other violations of the same statutory provision “for free,” since only one double-jeopardy offense would be committed. Similarly, cohabiting with X in 1988 must be a different offense than cohabiting with X in 1990 if the actor did not cohabit with X in 1989.
Recognizing this obvious truth, the prosecutor in Snow sought to create different “acts” of cohabitation by dividing the thirty-five-month period of cohabitation into three distinct periods. This division makes the conduct at least technically distinct. As cohabitation required living with a woman, it includes smaller periods of time. Perhaps Snow “cohabited” each time he returned home after an absence of any period of time. Or perhaps he committed an infinite number of metaphysical cohabitations because there were an infinite number of points during the thirty-five months when he could have terminated the cohabitation.
The Court rejected at least the second of these interpretations, noting:
The division of the two years and eleven months is wholly arbitrary. On the same principle, there might have been an indictment covering each of the thirty-five months, with imprisonment for seventeen years and a half and fines amounting to $10,500, or even an indictment covering every week, with imprisonment for 74 years and fines amounting to $44,400; and so on, ad infinitum, for smaller periods of time.33
The Court seemed to appreciate that the issue was the scope of the proscribed act created by the statute: “There can be but one offence between such earliest day [of cohabitation with the same women] and the end of the continuous time embraced by all of the indictments.”34 This puts a particular reading on “cohabitation,” but it is a plausible reading. Implicit in the Snow opinion is that this is the reading that Congress meant.
Two years later, the Court turned to the same-offense definitional problem, with yet another case applying the federal antipolygamy provisions to the Utah territory. (These cases suggest that the Court was uncomfortable with the enforcement of the antipolygamy laws.) In In re Nielsen,35 different statutory crimes were involved: adultery and cohabitation. The question is whether looking at the way the acts are proscribed in each statute is helpful in resolving the same-offense question.
Each offense required proof of an element the other did not—adultery required proof of marriage to a third person, and cohabitation required proof of living together as man and wife. This distinctiveness of statutory elements makes the offenses separate under the necessarily-included-offense test, but Nielsen held that a conviction of cohabitation barred a trial for adultery. Many explanations for Nielsen have been offered, but the soundest is that both statutes proscribed the same blameworthy act.
Determining which elements proscribe acts and which proscribe what the Model Penal Code calls “attendant circumstances” is no easy matter, as I show in chapter 5. For the present, it is enough to note, as the Court did, that both adultery and cohabitation have sexual intercourse as their central conduct element. Whether “living together” is an additional act and, if so, whether it is a blameworthy act are nice questions, but the Court read the statutes to be directed at sexual intercourse outside marriage: “Living together as man and wife is what we decided was meant by unlawful cohabitation under the statute. Of course, that includes sexual intercourse. And this was the integral part of the adultery charged in the second indictment, and was covered by and included in the first indictment and conviction.”36
If both offenses proscribe sexual intercourse outside marriage, these offenses fail Hale’s test, for there could not be proof of adultery (sex outside marriage) without proof of cohabitation (sex outside marriage). Stated differently, the act of sex outside marriage with particular partners is the same blameworthy act as sex outside marriage with those same partners.
If Nielsen is read in this fashion, it is consistent with the inclusion test of both Hale and Blackstone. By 1911 the Court still saw the key question as whether each offense proscribed the same blameworthy act. Like many state courts at the time, the Supreme Court in Gavieres v. United States sought to uncover the “gist of the offense” proscribed by two different municipal ordinances.37 The Philippines ordinances prohibited (1) behaving in an indecent manner in a public place, and (2) insulting or threatening public officials. As the Court noted, the proscribed act of behaving in an indecent manner is not the same as the proscribed act of insulting or threatening public officials. It did not, therefore, matter that in the particular case one act happened to violate both ordinances. As I discuss in chapter 5, seeking the “gist” or “gravamen” of the offenses is consistent with my blameworthy-act account of same offense. Thus, Gavieres is at one level consistent with both Nielsen and my account. But in applying the act-based “gist” principle, the Court phrased the test as if any distinct element made the “gist” different,38 which was inconsistent with the state court approach and with Nielsen.39
The difference between these applications of the “gist” test may appear unimportant (may have appeared so to the Gavieres Court), but it is crucial in many cases, particularly those that involve modern statutes with many elements.
The best example of the difference between what the Gavieres Court seemed to permit, on the one hand, and what Hale and Blackstone seemed to envision, on the other, is Blockburger v. United States.40 The Court held that sale of narcotics not from the original stamped package could be punished in addition to sale of the same narcotics not pursuant to a written order because neither offense was included within the other. Each required a distinct element, indicated by italics. If, however, both offenses in Blockburger proscribe the blameworthy act of selling narcotics41 and if “same offense” means “same blameworthy act,” these offenses would be the same offense.
Not so under Blockburger, which treats the presence of the distinct elements of “not from the original stamped package” and “not in pursuance of a written order” as evidence that Congress intended to create cumulatively punishable crimes.42 Justice Scalia claims that the Blockburger test is historically and textually indicated as a definition of same offense. As this book makes clear, Scalia is right about much in double jeopardy. But he is wrong about Blockburger.
Scalia asserts a textual argument on behalf of Blockburger that is clever but ultimately unsatisfying. He argues that “if each [offense] contains an element the other does not, i.e., if it is possible to violate each one without violating the other, then they cannot constitute the ‘same offence.’”43 To evaluate Scalia’s argument, we first examine its buried premise that all offense elements are created equal when considering what constitutes a double jeopardy offense. But there is nothing in the language of the Double Jeopardy Clause that compels that conclusion, just like there is nothing in the Clause that is helpful in separating civil “offenses” from criminal offenses. To defend Scalia’s account, one must accept that a single armed robbery in a dwelling is constitutionally transformed into two robberies in a state that proscribes “armed robbery” separately from “robbery in a dwelling.” Or that it is two robberies when a robber wearing a white shirt robs a victim on Sunday in a state proscribing “robbery while wearing a white shirt” and “robbery on Sunday.”
Not all “offenses” are double jeopardy offenses (torts are not, for example). There is, then, no reason to accept as inevitable Scalia’s premise that all offense elements in a concededly criminal offense are capable of making offenses different under the Double Jeopardy Clause.
Obviously recognizing that his textual argument is not conclusive, Scalia turns to English history. The authority on which he relies for longstanding acceptance of Blockburger is the 1796 case of King v. Vander-comb.44 But Scalia’s use of Vandercomb does not withstand even a superficial critique. He quotes a statement from the opinion that is the “precise equivalent of our statement in Blockburger.”45
What he fails to quote is the English court’s formulation of the principle that it deduces from the common law, which is demonstrably not the Blockburger test. Near the end of its opinion, the English court wrote, “These cases establish the principle, that unless the first indictment were such as the prisoner might have been convicted upon by proof of the facts contained in the second indictment, an acquittal on the first indictment can be no bar to the second.”46
This is not Blockburger for several reasons. It applies only to acquittals, and it relies on the contents of the indictment rather than the statutory (or common law) elements. Blockburger operates on the statutory elements and applies regardless of whether the first trial ended in a conviction or an acquittal.47 Moreover, the Vandercomb formulation appears to make the order of prosecution crucial, while Blockburger applies regardless of the order of prosecution.
On Vandercomb’s literal holding, if a greater offense is prosecuted first, no bar would arise to a necessarily-included offense because the defendant could not have been convicted under the first indictment on proof of the lesser offense.48 For example, Vandercomb’s holding would permit the state to prosecute murder and then manslaughter; the facts of manslaughter would not convict of murder. The state could not prosecute, however, in the order of manslaughter and murder. The facts of murder as contained in the second indictment would convict of manslaughter. So the state could prosecute murder-manslaughter but not manslaughter-murder.
This is an idiotic same-offense test; Friedland is more charitable, calling Vandercomb “totally unreasonable.”49 The fatal flaw is that it ignores the principle of correlative identity. When A is the same offense as B, B must always be the same offense as A. If one were going to dispense with this correlative principle, it would be in the direction opposite from the one Vandercomb chose: manslaughter and murder could conceivably be different offenses if prosecuted in that order, because manslaughter leaves unprosecuted an element in the greater offense of murder. Not so in the direction of murder-manslaughter. Once the greater offense is proven, all the elements of the lesser are always proven. Vandercomb gets it backwards, making offenses the same when at least one element has never been prosecuted and different when all elements are reprosecuted.
Yes, the Vandercomb court at one point phrased a test consistent with Blockburger. But what inference can be drawn from that when the court phrased a different test as the principle on which it based its holding? That the English courts were floundering in 1796 for a satisfactory account of same offense is further indicated by Vandercomb’s reinterpretation of Turner’s Case. The court in Turner’s Case found an acquittal of the burglary of T’s house and stealing the goods of T to be a bar to the charge of burglary of T’s house and stealing the goods of H.50 The Vandercomb judges classified this statement of law as a “manifest mistake” and pronounced themselves not bound by it since it was not the holding in Turner’s Case.51
To find in Vandercomb a basis to assert a long-standing acceptance of Blockburger is, therefore, to ignore the English confusion over same offense that is manifest in Vandercomb itself. No Blockburger understanding of same offense existed in Blackstone’s day or at the time the Double Jeopardy Clause was written. Blockburger is neither logically entailed by the constitutional text nor historically indicated.
Scalia’s attempt to ground Blockburger in the Court’s early jurisprudence is only marginally more successful. To be sure, Gavieres applied a version of Blockburger when it held that disorderly conduct is not the same offense as insulting a public official.52 But one gets the same result by comparing proscribed blameworthy acts, which were different in Gavieres.
To assert, as Scalia does, that the 1902 case of Burton v. United States53 is support for Blockburger is simply a mistake. The Court merely held in Burton that two acts of receiving money would be two offenses under a statute that prohibited certain receipts of money. An obvious difficulty with Scalia’s argument that Burton presaged Blockburger is that both charges were made under the same statutory offense, which means the only same-offense issue is how the prosecutor divided up the conduct (the In re Snow issue) and not whether two statutes define the same offense. To be sure, Burton applied Vandercomb to the language in the indictment, concluding that “the jeopardy is not the same when the two indictments are so diverse as to preclude the same evidence from sustaining both.”54 This analysis is gratuitous and, more important, focuses on the particular way the offenses are charged, a test that Scalia has rejected.55 How Scalia finds support in Burton for his understanding of Blockburger is mysterious.
Blockburger has nothing to say about the Burton kind of case. Comparing different offense descriptions has nothing to do with the conduct dimension of same offense. For example, the issue in Snow was the proper number of convictions of cohabitation for a single thirty-five-month period of living with the same women. That issue cannot be decided by comparing the elements in two statutes, for there is only one statute.56
If Blockburger is used in this context, then either (1) multiple charges under a single statute are always the same offense, which is obviously wrong, because robbery of X cannot be the same offense as robbery of Z, or (2) the Blockburger formulation should be expanded to examine the underlying indictment. While (2) is more appealing than (1)—and usually gives the right results in cases with more than one victim—it remains seriously flawed. Looking at the contents of the indictments, Martin Friedland has noted, makes the result of a double jeopardy plea “often depend on the manner in which the indictments happen to be framed.”57 So, for example, drunk driving would be the same offense as vehicular homicide if the latter indictment alleged killing by drunk driving and not if the indictment simply alleged killing by reckless driving. But the Court has essentially abandoned that approach, insisting that Blockburger be applied only to the statutory formulations.58
If the Court is right that Blockburger should be applied only to the statutory formulation (and in chapter 5 I agree with the Court on this point), then Blockburger provides answers to only half the same-offense puzzle. Of course, whether Blockburger has the pedigree of precedent, logical entailment, or historical inevitablity, whether it solves all the puzzle or only part of it, the Court can read it into the Double Jeopardy Clause. Miranda v. Arizona59 is a good example of a solution that does not depend on a long-accepted textual meaning. The only requirement for these Court-constructed solutions is that they work—that is, that they are relatively hard edged and offer a reasonable fit between the policies and the results produced.
Though Blockburger is mostly hard-edged,60 it fails to deliver a set of results consistent with double jeopardy policies.61 Blockburger’s failures are documented throughout the book, but the conceptual flaw can be summarized in the following hypothetical criminal offenses: (1) selling narcotics on Sunday; (2) selling narcotics from an automobile; (3) selling narcotics that have been unlawfully imported; (4) selling narcotics within a thousand feet of a church; (5) selling narcotics that have not been inspected by the FDA; (6) selling narcotics without having a state pharmaceutical license. On the Blockburger analysis, all six of these offenses are different double jeopardy offenses.
The actor who sells unlawfully imported narcotics from an automobile on Sunday, within a thousand feet of a church, without FDA inspection, and without the proper license can be prosecuted and punished six separate times. And, of course, there is no reason to stop at six. We could make it a separate crime to sell narcotics on street corners, in certain areas of the city, at night, to buyers who are not already addicted, to buyers who are unemployed, to buyers who are younger than twenty-five, to more than one buyer in a single transaction, to buyers who have crossed state lines in the past twenty-four hours, to buyers who have not passed a certified drug education course at the high school level, to buyers who are not physicians—.
Stop. It cannot be that every element is a relevant part of the double jeopardy offense description. Moreover, whatever the value of Block-burger’s definitional test as a starting point, it is unresponsive to the same-conduct question, and the Court’s approach to the latter issue is both ad hoc and unsatisfying.
Two related claims underlie this chapter and the rest of the book. First, the Supreme Court has missed the point of why a mistrial should be a double jeopardy bar to a second trial, abandoning Blackstone’s pleas in bar in favor of a test with edges so soft it fails to deliver answers. Second, the Court has not understood its own same-offense precedents, which carried over Blackstone’s and Hale’s premise that one proscribed blameworthy act could be only one offense.
Now that we have examined the Court’s doctrinal approach to jeopardy and same offense, we can examine the Court’s approach to the overarching principle of legislative prerogative. With a single blind spot, caused by the slavish devotion to the notion that double jeopardy prevents harassment, the Court has acknowledged the legislative-prerogative principle in its double jeopardy cases.
The earliest Supreme Court double jeopardy case, in 1820, involved dual sovereignty.62 Dual sovereignty, as I discuss in more detail in chapter 6, permits a state and the federal government to prosecute and punish the same offense. In the process of analyzing this issue, Justice Washington’s opinion announcing the Court’s judgment seems to say that the Double Jeopardy Clause same-offense analysis is wholly dependent on legislative intent. Despite his view that penalties could not be imposed on the same offense by both the state and federal government, he noted that a single legislature could “impose upon the same person, for the same offense, different and cumulative punishments” because that would be “the will of the same body to do so” and there would be “no opposition of wills.”63
The only way to make sense of Washington’s theory is to understand it as imputing legislative intent. If a single legislature imposes different penalties on the same offense, Washington was willing to infer that the legislature intended all the penalties to apply. But Washington was apparently not willing to infer that different legislatures were aware of what the other was doing. Whatever the value of these presumptions, which I consider in chapter 6, the point here is that at least one member of the 1820 Court viewed the creation of distinct criminal blameworthiness to be wholly a legislative prerogative.
If the legislature can impose as many different and cumulative punishments as it wishes, it obviously can define the end of jeopardy any way it pleases.
A legislative-prerogative principle is relatively uncontroversial when applied to defining the end of jeopardy. This is so for two reasons. First, legislatures usually create same-offense problems inadvertently, by criminalizing the same conduct in several overlapping criminal statutes without specifying how those statutes should relate to each other. But legislatures do not go around enacting hundreds of statutes regulating verdicts. Once a process for producing verdicts is set in place, there is little need to revisit it. Thus, we are much more likely to get a considered legislative view of what constitutes the end of jeopardy.
Legislative prerogative here is not controversial for another reason I have already noted—the legislature is hemmed in here by other Bill of Rights guarantees—right to trial by jury, to be found guilty beyond a reasonable doubt, and so forth. Moreover, the legislature has much less reason to decrease the protections attending verdicts than it does to enact harsh new penalties that can be cumulated. Though some defendants (including some highly publicized defendants) seem to escape “justice” with an acquittal against the weight of the evidence, there is little concern with the accuracy of acquittals generally. We have, in this regard, come a long way since Henry II was concerned about fraudulent acquittals by the water ordeal.64 Thus, the legislative view about the end of jeopardy is almost certainly going to manifest the common-law view of verdict. No controversy can be found in that view.
So it is only in the peripheral areas that the issue of legislative intent presents a live end-of-jeopardy issue. An example is the unusual procedure in Swisher v. Brady.65 The Maryland juvenile system required a master to enter proposed findings of fact that were reviewed by the juvenile judge, who entered judgment after either accepting or rejecting the master’s findings. In Swisher, the master made findings of fact that were favorable to the juvenile, and the judge rejected them. The juvenile appealed the adverse judgement, claiming that the master’s favorable findings of fact were equivalent to an acquittal, thus barring government conduct seeking to disturb the blameworthiness determination. The Supreme Court did not agree. At first blush, this seems wrong: favorable findings of fact seem functionally the same as an acquittal.66
But “functionalism” is a mostly failed legal realist approach. Trying to make different things look the same by examining the overarching policies permits good lawyers to make competing arguments that are almost equally good, and it permits courts the luxury of choosing the argument they think best in the particular case. As one commentator conceded when rejecting a double jeopardy historical analysis in favor of a functional one, “[A] functional approach may not make double jeopardy questions easier to resolve, particularly when a court must balance the defendant’s interest against the state’s prosecutorial interest.”67 Quite so. If hard-edged rules are set in place and followed, greater certainty can be achieved.
Except in the rare case when the prosecutor terminates a trial that likely would have resulted in an acquittal, the account of jeopardy offered in this book does not ask whether anything is functionally the same as anything else. It asks the more basic question: is this a verdict? That question has a procedural, not a functional, answer. The procedural answer is that a verdict is the outcome to which the process ascribes “verdict status.” In Maryland, the process did not define the master’s proposed findings of fact as a verdict, and the Double Jeopardy Clause contains no independent basis to ascribe “verdict status” to this event.68
Because of the procedure created by the Maryland legislature, the judge could accept the master’s findings or reject them, but the case is not over, the verdict is not rendered, until the judge acts. If juveniles had the right to a jury trial,69 the Maryland procedure would be unconstitutional under the Sixth Amendment. And the Court noted that the only issue raised in the lower courts was double jeopardy,70 concluding that “if there are any objections to such a system, they do not arise from the Double Jeopardy Clause.”71 There can be no double jeopardy harm because there is but one verdict under the procedural law of the jurisdiction.
When the Court got a chance to examine legislative prerogative in the context of defining the end of jeopardy, it decided the issue correctly. In the same-offense area, the results are more muddled. The Court has granted the primacy of legislative intent in part of the same offense doctrine, but not all of it. But when the Court has directly faced legislative prerogative in a same-offense context, it has always held in favor of the primacy of legislative intent.
Following Justice Washington’s 1820 opinion in the dual sovereignty case, the Court did not again address the role of legislative intent in defining same offense for more than one hundred years. When the Court addressed the issue again, Justice Brandeis spoke for a unanimous Court, in Albrecht v. United States.72 In answering the question of whether possessing untaxed liquor was the same offense as selling it, Brandeis wrote: “There is nothing in the Constitution which prevents Congress from punishing separately each step leading to the consummation of a transaction which it has power to prohibit and punishing also the completed transaction.”73 Brandeis and Washington can easily be counted in the camp of those who subscribe to the primacy of legislative intent in the same-offense context.
The issue did not arise again until the 1960s, when the Court began to draw a distinction between multiple punishment (more than one conviction imposed in a single trial) and successive prosecutions. The suggestion, often drawing on Justice Black’s famous dicta about the embarrassment, expense, and anxiety caused by successive prosecutions, was that a greater protection existed against successive prosecutions.
In the 1977 case of Brown v. Ohio, the Court went to considerable length to explain both the role of legislative intent in evaluating multiple punishments and how that relates to the successive prosecution issue. The Court began with multiple punishment: “Where consecutive sentences are imposed at a single criminal trial, the role of the [Double Jeopardy Clause] is limited to assuring that the court does not exceed its legislative authorization by imposing multiple punishments for the same offense.”74 The Court then reiterated the rationale for its legislative intent analysis, while hinting at the connection between legislative intent and successive prosecution doctrine:
[T]he Fifth Amendment double jeopardy guarantee serves principally as a restraint on courts and prosecutors. The legislature remains free under the Double Jeopardy Clause to define crimes and fix punishments; but once the legislature has acted courts may not impose more than one punishment for the same offense and prosecutors ordinarily may not attempt to secure that punishment in more than one trial.75
This analysis implies the proper relationship between the multiple punishment and successive prosecution doctrines.76 We begin with the notion of defining crimes and fixing punishments (creating criminal blameworthiness). This tells us when offenses are the same. The successive prosecution issue is simply whether reason exists to permit the prosecutor to use two trials to secure the singular blameworthiness judgment that is permitted. (What these reasons might be concerns us in chapters 8 and 9. One obvious example is a situation in which the defendant successfully resists joinder of two offenses that are the same offense.)77
To be sure, Brown did not state the implication of same-offense unity between multiple punishment and successive prosecution as I just did. Quite the contrary. Brown suggested that the greater harms of successive prosecution might lead to a different, and more expansive, test of same offense.78 But Brown made clear that the analysis at least begins with legislative intent.
In 1980 the Court applied Brown to a federal multiple punishment issue. Acknowledging that the Double Jeopardy Clause protects “against multiple punishments for the same offense,” the Court reiterated the Brown teaching: “the question whether punishments imposed by a court after a defendant’s conviction upon criminal charges are unconstitutionally multiple cannot be resolved without determining what punishments the Legislative Branch has authorized.”79 The Court used a similar formulation the next year in Albernaz v. United States, adding one sentence: “Where Congress intended, as it did here, to impose multiple punishment, imposition of such sentences does not violate the Constitution.”80
It is difficult to be much clearer than the Albernaz opinion, but if there is any doubt as to the Court’s meaning, it is dispelled by the opinion of Justice Stewart, concurring in the judgment (joined by Justices Marshall and Stevens). These Justices read Blockburger to state a constitutional rule. While agreeing with the Albernaz outcome, because Blockburger pronounced the offenses to be different, Stewart criticized the majority’s assertion about legislative intent: “These statements are supported by neither precedent nor reasoning and are unnecessary to reach the Court’s conclusion.”81 On Stewart’s view, “No matter how clearly it spoke, Congress could not constitutionally provide for cumulative punishments unless each statutory offense required proof of a fact that the other did not, under the criterion of Blockburger v. United States.”82
The state courts in Missouri failed to get the Albernaz message, even when the United States Supreme Court vacated and remanded more than one hundred Missouri cases for noncompliance with this dicta.83 The state court intransigence led to the ultimate test of the Albernaz reasoning. Stewart was right that, prior to 1983, the legislative intent analysis merely reinforced the Blockburger result. What was needed was a case in which legislative intent was clear and contrary to Blockburger.
Missouri v. Hunter,84 involving a statute that defined a new state crime of armed criminal action, was that case. Like felony murder, armed criminal action required proof of an underlying crime. Clear legislative intent to cumulate penalties for the underlying crime and armed criminal action was set out in the latter statute: “The punishment imposed pursuant to this [statute] shall be in addition to any punishment provided by law [for any other crime committed with a firearm].”85 On the Supreme Court’s felony murder application of Blockburger, however, the armed criminal action statute defined the same offense as the underlying crime that it required.86 The issue of legislative prerogative was joined.
The Supreme Court turned its Albernaz reasoning into a holding with ease. Only Justices Marshall and Stevens dissented.87 The Court noted the state court’s emotional formulation of the issue:
Until such time as the Supreme Court of the United States declares clearly and unequivocally that the Double Jeopardy Clause … does not apply to the legislative branch of government, we cannot do other than what we preceive to be our duty to refuse to enforce multiple punishments for the same offense arising out of a single transaction.88
The Court responded: “This view manifests a misreading of our cases.… [W]e need hardly go so far as suggested to decide that a legislature constitutionally can prescribe cumulative punishment for violation of its first degree robbery statute and its armed criminal action statute.”89
Whether Hunter is right turns out to be a question about limiting the form of punishment. If the rule were that the legislature cannot turn a single robbery into more than one offense, regardless of how clearly it spoke, that would not stop the legislature from achieving the same result through its punishing power. In Gore v. United States,90 for example, the Court upheld consecutive sentences based on a single sale of narcotics that violated three statutes that proscribed different ways of selling narcotics. The Court noted that Congress could have created a three-tier sentencing scheme that authorized a sentence of fifteen years in prison for selling drugs (1) not in the original package, (2) not pursuant to a written order, and (3) with knowledge of its unlawful importation. The authorized punishment could then be reduced by five years for each aggravating circumstance not present, creating a sentencing scheme indistinguishable from the consecutive sentences in Gore.
Should the Double Jeopardy Clause forbid in one form what Congress can clearly accomplish in another? If the legislative intent to cumulate penalties on the same act is “crystal clear,”91 the difference between doing it in one statute or in multiple statutes is the existence of more than one conviction. Obviously, the legislature can authorize the same penalty either way. Though multiple convictions create more onerous collateral consequences (e.g., parole eligibility, habitual offender status), the clarity of the legislative intent makes this difference constitutionally irrelevant. If the legislature wants two convictions, then it wants whatever collateral consequences attend two convictions, and it can have those consequences as long as the total punishment is not cruel and unusual under the Eighth Amendment.92
So the question comes down to the following: is there anything in the Double Jeopardy Clause definition of “offense” that precludes the legislature from authorizing two convictions for a Blockburger greater and lesser offense? The answer is negative on the Court’s Hunter analysis and on the argument presented in this book. Though Hunter implies that successive prosecutions for these offenses would be forbidden by Blockburger, that remains an open question.
A robust legislative-prerogative principle implies that legislative intent solves the same-offense problem in both contexts. The Court’s earliest and seminal multiple-punishment case, Ex parte Lange,93 is consistent with this view. The statute in Lange prescribed a fine of not more than two hundred dollars or imprisonment for not more than one year. Notwithstanding the disjunctive phrasing of the statute, the judge sentenced Lange to the maximum fine and the maximum imprisonment. After Lange paid the fine and served five days of the jail term, the judge vacated the original sentence and resentenced Lange to a single sanction of a year’s imprisonment.
The Court read the disjunctive phrasing of the statute literally and held that Lange’s payment of the fine was a fully executed alternative penalty that deprived the judge of the power to later sentence Lange to a jail term.94 Justice Scalia is right to observe that the Lange Court mixed due process and double jeopardy notions.95 Given the Court’s construction of the statute, the Court could simply have held (and in places did seem to hold) that due process prohibited the imposition of both penalties. About a hundred years later, the Court noted the obvious: “The Due Process Clause of the Fourteenth Amendment … would presumably prohibit state courts from depriving persons of liberty or property as punishment for criminal conduct except to the extent authorized by state law.”96
But the Lange Court drew an explicit (and I think correct) parallel between a judge imposing two punishments based on one conviction and a judge imposing two convictions for the same offense. If the latter is barred by double jeopardy, the Court reasoned, so is the former. The Court observed that it was not the threat of a second trial but, instead, “the punishment that would legally follow the second conviction which is the real danger guarded against by the Constitution.”97
This reading of Lange suggests that the protection against more than one trial for the same offense is derived from the protection against more than one judgment for the same offense. In the past thirty years or so, the opposite relationship has been asserted by many commentators (myself included), as well as by the Court.98 Indeed, as we saw in chapter 2, the Court patched together dicta from various cases and held in Grady v. Corbin that a greater protection exists against successive prosecutions than against multiple punishment. The nature of this greater protection was, as Brown intimated, a more expansive definition of same offense that essentially prohibited reprosecuting the same conduct.
In chapter 2 I endorsed Michael Moore’s view that double jeopardy protection does not vary by single trial/successive prosecution context.99 As we do not think it vexatious to prosecute successively for embezzlement and assault, for example, we should not think it vexatious to prosecute for two offenses arising out of the same transaction as long as we have a good account of when offenses are truly different. Justice Scalia dissented in Grady and wrote the opinion in Dixon, overruling Grady. Nonetheless, Scalia seems to believe that a greater double jeopardy protection exists in successive prosecutions.
But what would greater protection look like? Scalia suggests that it would be an inflexible Blockburger test that could not be rebutted by legislative intent.100 There are two arguments that can be made here. First, on Moore’s view, it makes no sense to treat successive prosecutions of truly different offenses differently from how we treat multiple punishments for those offenses. Second, the Court has for the past twenty-five years characterized Blockburger as a way of discerning legislative intent.101 How is it that a tool for discerning legislative intent somehow becomes a limit on the legislature?
If Scalia denies that Blockburger is a tool of statutory construction, then he is left without a principle to use in the single trial/multiple punishment analysis. If he admits that Blockburger is a tool to infer legislative intent, he must argue that it also functions as an inflexible rule in the context of successive prosecutions. But this is a truly odd position: a tool to infer legislative intent sometimes (somehow) becomes a rule of constitutional law that can deny actual legislative intent. As Eli Richardson puts it, viewing Blockburger as having a dual nature is a “strange characterization.”102
Moreover, there is no historical reason to create greater protection against successive prosecutions. The common law sought to prohibit multiple punishment, not successive prosecutions. The crux of the dispute between Henry II and Becket was the right of the royal courts to sentence clerics; there was no apparent concern with the harm of successive prosecutions. The common law as it developed from the twelfth century focused on the harm of the punishment. The reason former attainder was the broadest common-law plea was that the offender received the ultimate punishment and was thus mort in ley (dead in law). The new proceeding, in Blackstone’s words, “cannot be to any purpose; for the prisoner is dead in law by the first attainder, his blood is already corrupted, and he hath forfeited all that he had; so that it is absurd and superfluous to endeavor to attaint him a second time.”103 No second trial was permissible because further punishment would be superfluous, not because there was a need for a protection against successive prosecutions.104
This can be seen in the existence of the plea of former conviction, a plea that was available only when no judgment (and hence no punishment) had been imposed. As noted in chapter 2, the plea of former conviction offered a much less expansive protection. But if the crux of double jeopardy protection were directed against successive trials, rather than multiple punishments, no reason existed to treat convictions differently depending on whether punishment was imposed.
Then why did Blackstone and the Framers use “twice in jeopardy”? The answer lies in the procedural law of this time. A felony could not be joined with a misdemeanor until 1915; no conviction of a misdemeanor could be entered if a felony was charged; a defendant who was charged with a misdemeanor could not be convicted if the facts showed a felony; judges typically required the prosecutor to elect one felony offense if the indictment showed more than one.105 The combined effect of these procedural rules, all of which were in effect in Blackstone’s day, was to prevent multiple convictions in a single trial and thus to restrict the pleas in jeopardy to the second trial context.
So it is not surprising that no reference to multiple convictions in a single trial can be found in Coke, Hale, Blackstone, or Hawkins. The question about applying the double jeopardy pleas in a single trial would simply never have arisen. But the underlying rationale of not trying a defendant twice for the same offense was to avoid more than one punishment for the same offense. Madison’s original double jeopardy language was to that effect, forbidding “more than one punishment or trial for the same offense.”106 The Court was thus right in Lange to draw a parallel between multiple trials for the same offense and multiple convictions in the same proceeding:
For of what avail is the constitutional protection against more than one trial if there can be any number of sentences pronounced on the same verdict? Why is it that, having once been tried and found guilty, he can never be tried again for that offence? Manifestly it is not the danger or jeopardy of being a second time found guilty. It is the punishment that would legally follow the second conviction which is the real danger guarded against by the Constitution.107
Policy strongly supports the Lange unity between multiple punishment and successive prosecution. If two felony judgments of guilt could not be obtained in two trials, it would make no sense to allow them in one trial. Once the Lange unity is accepted, it requires unity in analyzing questions of multiple punishment and successive prosecution. On a legislative-prerogative account, both issues can be solved with legislative intent. It is a fundamental conceptual error to use legislative intent in the multiple punishment doctrine and then cast the successive prosecution doctrine adrift with the inferior Blockburger test.
The reason the Court created the ill-fated Grady test in 1992 is that Blockburger does not provide enough protection against successive prosecutions. But that means, on a unitary view of same offense, that Block-burger does not provide enough protection against multiple punishments. The solution is not to give more protection against successive prosecutions, as Scalia is apparently willing to do, but to give more (and more coherent) protection than Blockburger provides against multiple punishment.
If all same-offense questions are answered by reference to legislative intent, the Double Jeopardy Clause still has a “plain meaning” that should please Justice Scalia. A legislative-prerogative account reads “same offense” to mean “what the legislature meant to be punished once.” This is just another way of saying that the legislature creates substantive criminal offenses and thus necessarily controls by means of substance the procedural question of when offenses are the same. Offenses are the same when the legislature meant only one conviction to result; offenses are not the same when the legislature meant to authorize multiple convictions. It is just that simple.
Scalia’s view that Blockburger is an inflexible limit on successive prosecutions causes doctrinal difficulties. Recall Garrett v. United States108 from chapter 1. Decided before Scalia joined the Court, Garrett found that the Blockburger test pronounced the “compound” offense of engaging in a continuing criminal enterprise (CCE) and the “predicate” felony of a drug offense as the same double jeopardy offense.109 The Court then held that Blockburger created a presumption about same offense, which was rebutted by expressions of contrary congressional intent under the facts of Garrett. But, on Scalia’s view, Congress could not, regardless of the clarity of its intent, permit separate prosecutions of offenses that were the same Blockburger offense.
This outcome is particularly suspect with respect to CCE and similar offenses (such as RICO). As the Court noted in Garrett, it is extremely unlikely that Congress meant to create an offense that is aggravated because it includes prior crimes and then forbid prosecution of the aggravated offense if one of the prior crimes has been prosecuted.110 Yet this odd result is precisely the one dictated by Scalia’s view that Blockburger is an inflexible definition of same offense when separate trials are involved.
If legislative intent is an integral part of the analysis in successive prosecution cases, it must be the source of same offense. How could a court consider legislative intent and then decide the case in a contrary fashion? What would be the basis to ignore legislative intent? It cannot be Block-burger, which is a tool for divining legislative intent. It cannot be some metaphysical notion of harassment, for that is too indeterminate to provide concrete answers to real cases. The legislature thus determines not only what constitutes blameworthiness but also when blameworthiness is to be considered singular.
So far this chapter has identified the problems with Supreme Court doctrine on the three crucial jeopardy issues: the principle of legislative prerogative and the two presumptions that serve that principle (one proscribed blameworthy act is presumed to be only one double jeopardy offense, and jeopardy does not end without a verdict or an outcome that suggests that an acquittal would have resulted). Now we can examine briefly what other commentators have offered as solutions to the problems identified in this chapter.
Given the inconsistencies and complexities demonstrated so far, one might think that the woods would be full of proposals to bring coherence to double jeopardy. Oddly enough, that is not true. Only two books take double jeopardy doctrine as their subject. Both books accept the 1962 Model Penal Code view that the overriding double jeopardy policy is to prevent harassing prosecutions. Both books describe the then current law of double jeopardy (both were published in 1969) and offer occasional critiques and prescriptions. But neither book offers a framework for analysis that is significantly different from the 1962 Model Penal Code.
Martin Friedland’s book is an excellent comparative law study on double jeopardy, but its normative thrust is to require mandatory joinder as a way of solving the harassing prosecution problem in England and Canada. Jay Sigler’s book contains a wealth of material on state and federal approaches to many double jeopardy issues, along with frequent trenchant observations. Ultimately, however, Sigler views the Double Jeopardy Clause “as a series of problems in social policy.”111 He criticizes courts for not “fac[ing] the critical policy issues which lie deep beneath the surface” of double jeopardy law.112 He criticizes reliance on “the ‘same offense’ dogma,”113 by which he seems to mean the courts’ failure to substitute policy statements for the language of the Clause. In place of “same offense dogma,” Sigler calls for “a more complete consideration of the social policies which underlie double jeopardy.”114 While Sigler is right that the Court has not solved the same-offense problem, his reliance on underlying policies is unhelpful.
An example of what Sigler considers useful policy analysis is Justice Black’s statement of double jeopardy policies. Sigler identifies it as an exception to the refusal of courts to face the deeply buried policy issues.115 But answers to complex, specific double jeopardy questions cannot be found anywhere in Black’s exhortations about “embarrassment, expense and ordeal” or living in a “continuing state of anxiety and insecurity.”116 Like Friedland, Sigler recommends a mandatory joinder solution.117
But my project is not about how to avoid double jeopardy problems through statutory solutions such as mandatory joinder. My project is to find the best constitutional interpretation of “jeopardy” and “same offense.” Neither Friedland nor Sigler offers an account of how best to interpret these concepts.
The law review commentary is disappointing. For example, Larry Simon’s student note, Twice in Jeopardy, is an example of the power and failure of skepticism. Prior to the publication of Twice in Jeopardy in 1969, decades of court decisions and academic commentary had suggested that there was (or could be) some relatively clear answer to the question of when two offenses were the same, and Simon blew apart this smug view of double jeopardy in a tour de force. He was right to conclude that double jeopardy doctrine was filled with “fictions and rationalizations” that are “the characteristic sign of doctrinal senility.”118 He was right to skewer courts for the idea that double jeopardy rules were “expressions of self-evident moral principles.”119 Concluding that the answers are not self-evident does not, of course, help in finding a useful set of principles.
It is here, at the stage of providing principles, that Simon’s skeptical effort flounders. In the area of successive prosecutions, he, too, recommends mandatory joinder.120 To solve the single-trial problem, Simon borrows the proposal of Otto Kirchheimer and argues that convictions under different statutes should be permitted only when courts believe that the statutes were designed to prevent distinct social evils.121 Though this idea is conceptually helpful, its edges are far too soft to be useful as a doctrinal rule.
Peter Westen and Richard Drubel would avoid double jeopardy difficulties by rejecting all rules about twice in jeopardy and same offense. Courts should recognize instead that the Double Jeopardy Clause generates a set of values that can be arranged in a hierarchical manner. Paying close attention to this hierarchy of values in the context of each case will lead judges to reach the right result, whether the issue is same offense or second jeopardy.122
The Westen-Drubel hierarchy of double jeopardy values puts a premium on jury nullification, thus elevating an acquittal over a conviction as a bar to a second trial. The least important double jeopardy value, they assert, is finality. The middle value is a rebuttable presumption against double punishment, which they define as punishment that exceeds the total amount authorized by the legislature.
This legal realist solution thus does not protect against double jeopardy for the same offense. Rather, it protects absolutely jury nullification. In a few cases, it preserves finality, though Westen and Drubel offer no guidance beyond joining the chorus condemning generalized harassment of defendants by multiple trials. The Westen-Drubel solution protects against double punishment to the extent one can figure out what punishment was authorized by the legislature and then only as a deduction from the total authorized by the greater offense.
While Westen and Drubel offer a useful description of the Court’s doctrine, it is necessarily limited by that doctrine and thus partakes of the Court’s flawed conceptual framework.123 To the extent the Court has failed in its double jeopardy project, Westen and Drubel have failed as well.
If complexity and the lack of hard-edged rules is the problem with much of the current approach to double jeopardy, Akhil Amar and Jonathan Marcus have the antidote—a genuine identity test based on the “plain meaning” of same offense.124 Their basic rule is so clear that no one could have serious problems applying it. The only offense that is the same as first-degree premeditated murder is first-degree premeditated murder. But this approach offers almost no protection against double jeopardy. The medicine is worse than the disease.
The Amar-Marcus approach is more protective than it first appears because it embraces the Court’s implied acquittal doctrine—a prosecution for murder, for example, that results in a manslaughter conviction implicitly acquits that defendant of murder.125 Amar and Marcus would also supplement double jeopardy protection by finding mandatory joinder (prohibition of vexatious prosecutions) in the Due Process Clause. Finally, they relocate collateral estoppel in the Due Process Clause as well.
Once all the relocating to the Due Process Clause is done, Amar and Marcus have a plain-meaning test of same offense. Inevitably, though, the same offense issues left in their conception of the Double Jeopardy Clause are trivial. First-degree premeditated murder is the same as first-degree premeditated murder; yes, of course. A conviction of manslaughter on a trial of first-degree premeditated murder acquits of murder; but otherwise no other offense is the same as first-degree premeditated murder, not felony murder or second-degree murder or aggravated murder, and certainly not manslaughter.
Blackstone was of a different mind on this issue, as we have seen, noting that a conviction under a manslaughter indictment was a jeopardy bar to a murder trial.126 One might contend that we can improve on Blackstone’s understanding and that, in any event, prosecutors are rarely going to bring manslaughter and murder trials separately. I doubt that we can improve on Blackstone’s understanding of very much about double jeopardy, and the pragmatic response misses the point of finding the best conception of same offense. Moreover, anyone wishing to change the rule about lesser included offenses must contend not only with Blackstone but also, as we saw in chapter 1, with a line of cases and authorities from Blackstone to the present.127 As Bishop said in 1923, the name of the offense “in the two indictments may differ, and within our constitutional guaranty the offences be the same.”128 To justify rejecting more than two centuries of settled law would require a showing that the replacement doctrine is much better. Amar and Marcus offer, at best, a different justification for the same doctrine and at worst a jumbled due process test that would produce even less clarity than the Court has provided.
Perversely, a plain-meaning approach to same offense requires Amar and Marcus to force much of what has been thought for centuries to be double jeopardy into a soft-edged due process analysis. This is the worst of both worlds. Amar and Marcus achieve a plain meaning of same offense that is far too narrow to replicate the common-law protection. Then they leave the rest of Blackstone’s double jeopardy protection mired in due process, to be solved with a vexatious-prosecution analysis that has almost no contours or bounds.
For example, Amar and Marcus express uncertainty about whether their due process test would bar successive prosecutions for the greater-lesser offenses of auto theft and joyriding that the Court found to be the same offense in Brown. In their own words,
The state’s reasons for two trials did not seem particularly malicious, but sounded in local autonomy: the police in one county, where defendant was caught, apparently did not know of the plans of the prosecutor from another county, where defendant had stolen the car nine days earlier. Under an intent-based due process analysis, the state’s actions might pass muster as nonvexatious; but under a stricter effects-based due process test, the fact that a highly organized state could have brought a single case might be enough for Brown to win.129
The state’s reasons did not “seem particularly malicious” and therefore the state’s actions “might pass muster,” but, then again, on a different version of their test, Brown “might” have won. This due process swamp must be avoided if double jeopardy doctrine is to have coherence.130 Moreover, unless they are willing to argue for a plain-meaning interpretation of “jeopardy of life or limb,” Amar and Marcus achieve only a partial double jeopardy plain meaning.131 This is hardly enough to justify rejecting the line of cases and authorities from Blackstone to Brown, as well as the certainty of a single conviction for a single homicide.
To this point, I have argued for the plausibility of a legislative-prerogative account of double jeopardy. In this account, “same offense” means “same legislatively created blameworthiness.” Now a new difficulty arises. I do not believe, for example, that the legislature intends assault to be a different offense than assault with intent to murder, yet these offenses do not describe precisely the same blameworthiness. Similarly, premeditated murder describes a greater blameworthiness than manslaughter; yet there is only one killing and, Blackstone thought, only one offense.
I need an account of legislatively created criminal blameworthiness that tracks legislative intent to create distinct offenses as opposed to overlapping offenses. The difference between the two is suggested in the terminology: distinct offenses manifest blameworthiness that the legislature meant to be separately punishable; overlapping offenses are merely alternative ways of prosecuting the same blameworthiness. Murder and manslaughter are examples of alternative, overlapping offenses. Murder and rape are examples of distinct offenses.
Justice Rehnquist recognized in Garrett the double jeopardy difference between offenses that are distinct and those that are overlapping alternatives. He wrote, “In view of this legislative history, it is indisputable that Congress intended to create a separate CCE offense. One could still argue, however, that having created the separate offense, Congress intended it, where applicable, to be a substitute for the predicate offense.”132
A second analytic problem has to do with what counts as double jeopardy blameworthiness. Two judgments cannot manifest the same offense unless both judgments are for an “offense” in the first place. Tort law and regulatory law create a type of blameworthiness, as does the motor vehicle code, but it is not clear that this type of fault is same-offense blameworthiness.
Because understanding what constitutes double jeopardy blameworthiness is helpful in developing my account of distinct blameworthiness, the first task is to untangle tort and crime, as well as regulatory offense and crime. I take on that task in chapter 4. Then, in chapter 5, I address the problem of separating distinct blameworthiness from merely overlapping blameworthiness.