Full opinion text
OPINION KELLER, P.J., delivered the opinion of the Court in which WOMACK, KEASLER, HERVEY, and COCHRAN, JJ., joined. In Oregon v. Kennedy, the United States Supreme Court held that the Fifth Amendment’s Double Jeopardy Clause barred retrial after a defendant successfully moved for mistrial only when it was shown that the prosecutor engaged in conduct that was “intended to provoke the defendant into moving for a mistrial.” In Bauder v. State, we interpreted the Double Jeopardy provision of the Texas Constitution more expansively, to cover “reckless” conduct, holding that retrial would also be barred “when the prosecutor was aware but consciously disregarded the risk that an objectionable event for which he was responsible would require a mistrial at the defendant’s request.” We granted review to reexamine Bander’s holding. We conclude that Bauder should be overruled and that the proper rule under the Texas Constitution is the rule articulated by the United States Supreme Court in Oregon v. Kennedy. I. BACKGROUND Appellant called 911 after killing her husband. When officers arrived, they placed her in a patrol car, and eventually she was taken to the police station. At the scene and at the station, appellant gave statements after receiving Miranda warnings. At trial, the prosecutor asked three sets of questions that the court of appeals deemed relevant to its analysis. First, the prosecutor asked the crime scene officer, “When you met with [Appellant] Swanda Wiley, is that the name that she was giving you then?” Second, while the prosecutor cross-examined appellant, the following occurred: Q. Did you ever tell the 911 operator [Kenneth Wiley] had been raping [you], he had been attacking [you]? A. No. Q. In fact, you never told any law enforcement about the rape? The next day, during further cross-examination of appellant, the following occurred: Q. After speaking with [Detective] John McCaskill on August 10th of the year 2000, did you have occasion to learn the next day, on August 11th of the year 2000, John McCaskill wanted to speak with you again? A. Yes. Q. And you denied him opportunity to speak — After the first question, and after the last question in each of the two succeeding sequences, defense counsel objected that the prosecutor had commented on the defendant’s post-arrest silence in violation of Article 38.08 of the Texas Code of Criminal Procedure, Article I, § 10 of the Texas Constitution, and the Fifth and Fourteenth Amendments to the United States Constitution. All three objections were sustained, and instructions to disregard were requested and given in connection with the second two sets of questions. A mistrial was not requested on the first question, was requested and denied with regard to the second set of questions, and was granted after the third set of questions. Appellant later filed a pretrial habeas application, claiming that any subsequent prosecution was barred under double jeopardy principles, but the trial court denied relief. After discussing this Court’s latest opinion on the Bauder standard, the court of appeals reversed, holding that “the prosecutor, at the least, engaged in this conduct with conscious disregard for a substantial risk that the trial court would be required to declare a mistrial.” II. STARE DECISIS In conducting a re-examination of precedent, we keep in mind the strong preference for adhering to past decisions: “Often it is better to be consistent than right.” Precedent can be overruled, however, if the reasons for doing so are weighty enough. Some factors supporting the overruling of precedent are: (1) that the original rule or decision was flawed from the outset, (2) that the rule’s application produces inconsistent results, (3) that the rule conflicts with other precedent, especially when the other precedent is newer and more soundly reasoned, (4) that the rule regularly produces results that are unjust, that are unanticipated by the principle underlying the rule, or that place unnecessary burdens on the system, and (5) that the reasons that support the rule have been undercut with the passage of time. The State points to, and disputes, two currently accepted legal propositions upon which Bander’s holding rests. The first is that the Texas double jeopardy protection embraces the mistrial setting. The second is that the Texas double jeopardy protection imposes a different standard than its Fifth Amendment counterpart for determining when a defense-requested mistrial can properly be attributed to the State for the purpose of barring further prosecution. Overruling either of these legal propositions would result in eliminating the rule announced in Bauder. We will examine each proposition in light of the factors articulated above. III. MISTRIALS A. The Issues The State contends that, properly construed, the Texas double jeopardy provision does not apply to the mistrial setting. This contention has also been advanced in dissenting opinions in Bauder and its progeny. If this interpretation were adopted, the result would be to hold that, in some respects, the Texas constitutional provision actually provides less protection than its Fifth Amendment counterpart, and as a result of the way in which it provides less protection, would obviate any inquiry in this ease into whether it might be more expansive in other respects. In the language of Hulit, if the mistrial setting is not part of the state double jeopardy “building,” then one has no occasion to determine whether that building contains a recklessness “floor” not found in the federal double jeopardy building. The State makes two basic arguments in support of its position. First, the State claims that the “mistrial species” of double jeopardy jurisprudence was not part of the common law that formed the basis for the Texas constitutional provision. Relying upon Justice Powell’s dissenting opinion in Crist v. Bretz, the State contends that, instead, the “mistrial species” traces its lineage through English common law to an independent rule of jury practice, formulated by Lord Coke, that prohibited needless discharges of the jury. The State further asserts that “a bar to re-prosecution because of a premature termination of the first trial because of mistrial did not even emerge as a constitutional principle of double jeopardy jurisprudence until 1949 when the Supreme Court delivered its opinion in Wade v. Hunter,” which imported Lord Coke’s rule. Consequently, the State concludes, “it becomes almost ludicrous for one to truly believe that the framers of the Texas constitution in 1876 had contemplated a double jeopardy protection that did not even come into existence until more than seventy years later.” In line with the State’s position, the dissent in Peterson II had concluded that “during the approximately 150 years before our decision in Bander I, our state constitutional double jeopardy provision had never been interpreted as having any application to the mistrial setting.” Second, the State contends that legislation passed in 1856 — defining double jeopardy solely by conviction or acquittal— reflected the intent of the framers of the Texas Constitution. The State points out that this legislation was passed a mere eleven years after the Texas Constitution of 1845 (containing a predecessor of the current double jeopardy provision) and twenty years before the Constitution of 1876 (containing the current double jeopardy provision). The dissents in Peterson II and Lee II made the same argument. Finally, we include in this discussion a third argument, made by former Presiding Judge McCormick in his dissent in Ban-der: that the state double jeopardy provision’s language suggests that it applies only to acquittals. His opinion quotes the state provision with the following in italics, suggesting that he believed that the italicized clause modifies the entire provision rather than the immediately preceding clause: “no person, for the same offense, shall be twice put in jeopardy of life or liberty, nor shall a person be again put upon trial for the same offense, after a verdict of not guilty in a court of competent jurisdiction.” B. Before the 19th Century The Fifth Amendment’s Double Jeopardy Clause provides: “nor shall any person be subject for the same offence to be twice put in jeopardy of life or limb.” This provision was framed in light of a long history of the concept of double jeopardy in English common law. The development of double jeopardy law in England and the history of its incorporation into the United States Constitution is discussed in various Supreme Court and state court opinions. There seems to be no dispute between the majority and dissent in Bretz regarding the historical developments, so we rely heavily upon Justice Powell’s dissent, along with information from other sources. In English common law, “jeopardy” referred to the principle underlying the doctrines of autrefois acquit and autrefois convict. A defendant was considered to be placed twice in jeopardy upon a second trial only if there existed a prior conviction or acquittal. Essentially, the doctrine embodied “a res judicata policy” for criminal cases. That policy required an actual acquittal or conviction for its implementation. The debates in 1789 on the Bill of Rights confirmed that the framers of the United States double jeopardy provision understood that it would operate in such a manner. There is also some historical indication that the phrase “life or limb” in the Fifth Amendment was intended to perform a limiting function. One commentator has argued strenuously that the phrase was intended literally to encompass only capital cases. But that interpretation seems to discount the words “or limb,” which, even under a literal interpretation, suggests application to lesser punishment. Several old state court cases indicated that the phrase “or limb” was used to refer to a category of serious offenses, punished long ago in England by dismemberment, that came to be known as “felonies.” Separate from pleas in bar that formed the basis for the doctrine of double jeopardy, a rule was laid down by Lord Coke prohibiting the discharge of juries: once the jurors were “retorned and sworn, their verdict must be heard, and they cannot be discharged.” The rule was originally an absolute command that “once banded together a jury should not be discharged until it had completed its solemn task of announcing a verdict.” Even during deliberations, the jury was required to be kept together “unfed and without drink” even “till death if they do not agree.” The harshness of this rule was soon mitigated with two exceptions: necessity and consent of the defendant. In the seventeenth century, Lord Coke’s rule became a useful defense against Crown oppression— precluding the “tyrannical practice” of permitting discharge of the jury and reindictment when acquittal seemed likely. However, this rule against discharge was one of jury practice and was not a basis for the pleas in bar from which the doctrine of double jeopardy arose. The early cases and treatises announced no clear standard regarding the effect of failing to follow the rule, and it “seems never to have been pleaded successfully in bar of a second prosecution in the period of the Year Books, when the rule is said to have arisen.” “In any event, it seems never to have furnished the basis for a plea of autrefois acquit. Rather, it was viewed as a matter committed to the discretion of the trial judge, from which no writ of error would he nor any plea in bar of a future prosecution would be allowed.” In Bretz, the Supreme Court majority acknowledged this history but responded, “But this constitutional understanding was not destined to endure.” C. 19th Century Developments in Other Jurisdictions In 1795, less than four years after the Bill of Rights was ratified, the Supreme Court of North Carolina invoked Lord Coke’s rule against unnecessary discharges to bar retrial in a capital case after the premature discharge of a jury, and the court specifically stated that it would “not again put [the defendant’s] life in jeopardy.” The court explained that, during the reign of the Stuart family in England, the Crown often dismissed a prosecution during the middle of the trial “for the purpose of having better evidence against [the defendant] at a future day,” and the court condemned that practice as “so abhorrent to every principle of safety and security, that it ought not to receive the least countenance in the courts of this country.” After that opening salvo, state courts in the nineteenth century split on whether mistrials implicated their state constitutional protections against double jeopardy. Led by New York, a number of state courts hewed to the traditional understanding that double jeopardy was implicated only after a previous conviction or acquittal. But North Carolina reaffirmed its earlier holding as a correct articulation of double jeopardy principles, and it was joined in its position by Pennsylvania, Tennessee, South Carolina, and Ohio. Subsequently, several other jurisdictions switched sides to join the trend towards incorporating Lord Coke’s rule into double jeopardy jurisprudence. By 1876, the Supreme Court of Nevada would confidently say that “the rule now seems to be pretty well settled in American courts” that double jeopardy protections extended to the premature termination of a trial. In deciding to import Lord Coke’s rule into double jeopardy jurisprudence, several courts looked to the plain meaning of the word “jeopardy” in finding that the protection necessarily extended to proceedings occurring before verdict: someone was in “jeopardy,” or peril, of his life or liberty when he was put to trial, not after the verdict was delivered. One of the opinions issued in the Waterhouse case by the Supreme Court of Tennessee acknowledged the tension between the “common acceptation” of the word “jeopardy” and the history of its usage in English common law. “But for a long course of judicial opinion to the contrary,” the Tennessee justice wrote, “I should be at a loss to attach a different meaning to the expression.” Ultimately, he appeared to harmonize ancient English law with ordinary language by concluding that the term “acquittal” could “include the case of illegal discharge of the jury — in which case the defendant is virtually acquitted, and is entitled to be also discharged.” Even one of the jurisdictions adhering strictly to the common law of England acknowledged the conflict between popular usage and established tradition: “If this were a question of first impression, grave doubts might be entertained as to its proper solution. The constitutional prohibition, interpreted in its popular sense, would seem to bear the construction put upon similar provisions, in Pennsylvania, North Carolina, [and] Tennessee.” Several jurists in these state cases also reasoned that the prohibition against placing a person “twice in jeopardy of life or limb” must necessarily mean more than the age-old common law principle that pri- or judgments were given preclusive effect — a principle that applied even in civil cases — because that common law principle was in no need of special protection. In addition, one of the justices in the Pennsylvania case opined that according double jeopardy effect to a wrongful discharge of the jury was also necessary to effectuate the state constitutional provision requiring that the right to trial by jury “remain inviolate.” Finally, some courts expressly found that according such double jeopardy effect was necessary to protect the very interests of the double jeopardy clause itself, to safeguard against the possibility that the trial court would discharge a jury and order a retrial simply because a prosecution witness was absent or the proof offered was insufficient to support the conviction. Most of the state decisions importing Lord Coke’s rule against discharges into double jeopardy jurisprudence contained the express pronouncement that an illegal discharge of the jury operated as an acquittal. Virtually all of the importing cases were in agreement, however, that a discharge was legal (and thus did not operate as an acquittal) if there was necessity or the defendant consented (assuming he had an attorney). Although most of the cases in which relief was granted involved the discharge of the jury during deliberations or later, several state courts expansively pronounced that jeopardy attached when the jury was impaneled, sworn, and charged with the case. By “charged with the case,” these authorities meant charged in the sense of being charged as jurors at the beginning of trial. Although the United States Supreme Court did not start the trend of incorporating Lord Coke’s rule against unnecessary discharges into double jeopardy jurisprudence, the Court issued a decision that had the effect of fueling it. In United States v. Perez, the jury was discharged because it could not agree upon a verdict, and the defendant claimed that the discharge operated as a bar to further prosecution. The Supreme Court held that it did not, but in language that suggested that the absence of a bar might turn upon the existence of “manifest necessity” for the jury’s discharge: We are of opinion, that the facts constitute no legal bar to a future trial. The prisoner has not been convicted or acquitted, and may again be put upon his defence. 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 fav-our 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. We are aware that there is some diversity of opinion and practice on this subject, in the American Courts; but, after weighing the question with due deliberation, we are of opinion, that such a discharge constitutes no bar to further proceedings, and gives no right of exemption to the prisoner from being again put upon trial. In the late twentieth century, the Supreme Court would acknowledge that “a close reading” of the above passage “could support the view that the Court was not purporting to decide a constitutional question, but simply settling a problem arising from the administration of federal criminal justice.” But the Supreme Court made no clear pronouncement to that effect in the nineteenth century, and in fact, it implied in an 1874 decision — Ex parte Lange — that Perez was indeed a double jeopardy case. In the meantime, states on both sides of the issue cited Perez in support of, or at least in discussion of, their respective positions. Citing or discussing Perez and state cases, at least three respected nineteenth-century American commentators concluded that jeopardy attached at the time the jury was impaneled and sworn and that termination of the trial without a verdict — absent necessity and against the defendant’s wishes — resulted in a bar to future prosecution. Two of those commentators expressly recognized that there were decisions to the contrary but concluded, from the language of various double jeopardy provisions, the consequences of a contrary rule to a defendant’s double jeopardy interests, or the prevailing common-law practice, that their view was the better one. In light of the various authorities, some states took intermediate positions. Some claimed that the improper dismissal of a jury resulted in barring a future trial but did not claim that the bar flowed from the protection against double jeopardy, or expressly claimed that it did not, or claimed that a new trial might be barred if prejudice occurred. Others avoided general pronouncements, saying simply that there was no bar if necessity existed or the defendant consented. D. 19th Century Developments in Texas Texas adopted its first double jeopardy provision in 1836, when it was an independent republic. The ninth section of the Declaration of Rights of the Constitution of the Republic of Texas provided: “No person, for the same offence, shall be twice put in jeopardy of life or limbs. And the right of trial by jury shall remain inviolate.” Upon joining the United States, Texas drafted a new constitution, adopted in 1845. This new state constitution deleted the “s” from “limbs,” added a new clause, and made some minor changes in punctuation, causing the section to then read: “No person, for the same offense, shall be twice put in jeopardy of life or limb, nor shall a person be again put upon trial for the same offence after a verdict of not guilty; and the right to trial by jury shall remain inviolate.” When Texas became part of the Confederate States during the War Between the States, it drafted a new constitution, adopted in 1861. The text of the section containing the double jeopardy provision remained the same, except that the comma after “limb” became a semicolon. With the defeat of the Confederacy, and the resultant new constitution of 1866, the semicolon was changed back to a comma. In 1869, under a new constitution adopted as a result of reconstruction, the words “or limb” were deleted, a semicolon was placed after “life,” and a new comma was inserted after the second appearance of the word “offense,” causing the section to read as follows: “No person, for the same offense, shall be twice put in jeopardy of life; nor shall a person be again put upon trial for the same offense, after a verdict of not guilty; and the right of trial by jury shall remain inviolate.” Finally, the present state constitution, adopted in 1876, inserted additional language in two different places, changed a semicolon to a comma, and moved the “jury trial” clause to a separate provision, resulting in what is now the present double jeopardy provision: “No person, for the same offense, shall be twice put in jeopardy of life or liberty, nor shall a person again be put upon trial for the same offense, after a verdict of not guilty in a court of competent jurisdiction,” In the midst of these constitutional changes, the Legislature enacted laws articulating its own view of double jeopardy. The 1856 Code of Criminal Procedure provided: Art. 18. No person for the same of-fence can be twice put in jeopardy of life or limb. This is intended to mean that no person can be subjected to a second prosecution for the same offense, after having been once prosecuted in a Court of competent jurisdiction and duly convicted. Art. 19. The foregoing article will exempt no person from a second trial, who has been convicted on an illegal instrument or information, and the judgment thereupon arrested, nor where a new trial has been granted to the defendant, nor where a jury has been discharged without rendering a verdict, nor for any case other than that of a legal conviction. Art. 20. By the provisions of the Constitution, an acquittal of the defendant exempts him from a second trial, or a second prosecution for the same offense, however irregular the proceedings may have been; but if the defendant shall have been acquitted upon trial, in a Court having no jurisdiction of the offense, he may, nevertheless, be prosecuted again in a Court having jurisdiction. The provisions of Section 19 were later carried forward as Section 20 of the 1879 Code of Criminal Procedure. In the 1871 decision of Moseley v. State, the Texas Supreme Court addressed whether the state double jeopardy protection applied when a trial was terminated prematurely. The defendant was tried for a capital offense, and his case was submitted to the jury, but the jury was discharged (without his consent) because the jurors could not agree upon a verdict. In response to the defendant’s claim that the discharge created a double jeopardy bar to retrial, the Texas Supreme Court acknowledged that the issue involved an area of disagreement among various jurisdictions: “All authorities agree that the word jeopardy, in its common and legal signification, means danger or hazard, but many disagree as to the time when that danger begins to a person charged with an offense, and when it ends.” The court observed that the defendant’s position was maintained by “Lord Coke ... and the courts of Pennsylvania, Tennessee and some other states,” but the court expressed its own belief that the question of double jeopardy protection had “finally been settled by the highest authority in England and America to have reference to the trial and the verdict” so that no person could “claim an exemption from a second trial under this maxim, unless he has once been tried ... and acquitted or convicted.” Nevertheless, the court also pointed out that even the authorities finding double jeopardy principles to be applicable in this context found an exception in the case of necessity. And the court further concluded that, if necessity be considered an exception, then, logically, the issue of whether to discharge a jury is simply a matter addressed to the trial court’s sound discretion. In support of that conclusion, the court cited the United States Supreme Court’s decision in Perez and its own conclusion that a contrary decision would be based upon the unacceptable notion that “it is preferable to confine and perhaps starve a jury for an indefinite period, and thereby force from them a reluctant verdict against their judgment rather than permit the court to exercise that dangerous discretion to discharge a jury, when it became morally certain that they could not arrive at an intelligent and honest verdict.” The next year, the Texas Supreme Court revisited the issue in Taylor v. State. The defendant in that case was initially indicted for the murder of N. Evans, but the proof at trial showed that he killed Morgan Evans. After the State’s opening argument, the prosecutor entered a nolle prosequi (a nonsuit of the prosecution), and the jury was discharged. The State subsequently filed a new indictment charging the defendant with the murder of Morgan Evans. Citing Moseley, the court reiterated its holding that the double jeopardy protection could not be invoked in a second prosecution unless there had been a prior “trial and verdict.” The court observed that “Mr. Bishop, in his valuable work on criminal law, seems to entertain a somewhat different view,” but because of the numerous exceptions cited, the court concluded that the discussion really lent support to its own view that discharging the jury in the first trial was simply a matter to be left within the trial judge’s discretion, with no attendant double jeopardy consequences. However, the court also observed that double jeopardy might not in any event apply because the murders of N. Evans and Morgan Evans constituted “separate and distinct offenses.” Although the Texas Supreme Court had aligned this state with those jurisdictions hewing strictly to the common law of England, that would soon change when criminal appeals were handled by our predecessor, the Court of Appeals (assigned that responsibility by the Constitution of 1876). In the 1877 case of Parehman v. State, the defendant claimed that double jeopardy barred retrial after the case was dismissed pursuant to a nolle prosequi. After the jury had been impaneled and sworn, and during the testimony of the State’s witness, it was discovered that the indictment erroneously referred to the victim as “H. Franks” when his name was in fact “H. Frank.” The jury was discharged over the protest of the defendant, and a new indictment, alleging the correct name, was filed. The Court of Appeals observed that there had “been quite a conflict of opinions in this country” regarding whether the double jeopardy protection was implicated by the premature termination of a trial. “[AJfter a careful examination of the authorities,” the court concluded that the double jeopardy protection did apply, but with exceptions for when the defendant consented or in a variety of circumstances that could be reasonably characterized as necessity: We believe ... that if the court had no jurisdiction of the cause, or if the indictment was so defective that no valid judgment could be rendered upon it, or if by any regular necessity the jury are discharged without a verdict — which might happen from the sickness or death of the judge of the court, or the inability of the jury to agree upon a verdict after sufficient deliberation and effort — or if the term of court as fixed by law comes to an end before the trial is finished, or the jury are discharged with the consent of the defendant, expressed or implied, or if, after verdict against the accused, it has been set aside on his motion for a new trial or in arrest of judgment, the accused may, in all such cases, again be put upon trial for the same facts charged against him, and the proceedings had will constitute no protection. But, when the legal bar has once attached, the government cannot avoid it by varying the form of the indictment. If the first indictment was such that the accused might have been convicted under it on proof of the facts by which the second is sought to be sustained, then the jeopardy which attached to the first must constitute a protection against a trial on the second. With regard to the case before it, however, the court concluded that a second trial was not barred by double jeopardy because, due to a material variance in the name of the victim, the second indictment charged a different offense from the first. The next year, in Vestal v. State, the Court of Appeals was confronted with whether parol evidence was admissible to show the actual status of a first trial in connection with a defendant’s special pleas of autrefois acquit and former jeopardy. In its discussion, the Court quoted with approval the above passage from Parch-man, and it also quoted a statement from the defendant’s brief in that case regarding the attachment of jeopardy at the beginning of trial and the acquittal consequences of a premature termination of the proceedings: [Wjhen a party is once placed upon his trial for a public offense, involving life or liberty, on a valid indictment, before a competent court, with a competent jury impaneled, sworn, and charged with the case, he has then reached and is placed in jeopardy ... and, after the jeopardy has once so attached, a discharge of the jury without the consent of the defendant, before they have reached a verdict, is equivalent to a verdict of acquittal. Arguably the pronouncements in Parch-man and Vestal were dicta because they were not necessary to the resolution of those cases and because neither case referred to the conflicting holdings in Moseley and Taylor. Any cloudiness in the law on that account would disappear in 1884, however, with the advent of Powell v. State. In Powell, the defendant claimed that the Texas double jeopardy provision prevented his retrial after the jury was discharged in his first trial for failure to agree on a verdict. At the first trial, the jurors returned to the courtroom after deliberating for one-and-a-half hours to say that they could not — and would never be able to — agree on a verdict. The trial judge sent the jurors out to deliberate for another hour, after which they returned to again say they could never agree. The trial judge sent the jurors out a third time, and they returned an hour later with the same results. The trial judge then discharged the jury over the defendant’s objection. In attempting to ascertain the meaning of the double jeopardy provision found in the Texas Constitution, the Court of Appeals first examined Articles 18 and 19 of the Texas Code of Criminal Procedure of 1856 (and the successor Article 20 in the Code of Criminal Procedure of 1879). The Court observed that “if it could so be done, a fixed and definite meaning has been given by the Legislature to the words ‘former jeopardy,’ and that meaning, as declared, is that ‘former jeopardy’ is nothing short of a prior legal conviction.” But the Court held that the Legislature had no authority to construe a constitutional provision “which has become fixed and settled by judicial determination.” Acknowledging “a diversity of [judicial] views” on the meaning of jeopardy, the Court nevertheless maintained that resolving uncertainty on the matter was not for the Legislature but for the courts to decide. The Court then cited with approval the views of Cooley and Bishop, and the second edition of Bennett and Heard’s note to United States v. Perez, all three of which advocated application of double jeopardy protection to the premature termination of a trial. The court quoted from Cooley’s treatise for the proposition that jeopardy attached at the time the jury was “impaneled and sworn,” that the defendant at that time became entitled to a verdict that would bar a new prosecution, and that the defendant could not “be deprived of this bar by a nolle prosequi entered by the prosecuting officer against his will, or by a discharge of the jury and continuance of the cause.” The Court found that this view was “uniformly” supported by “the decided weight and respectability of authority.” The Court observed that Bennett and Heard’s note contained the “most thorough discussion and elaboration of authorities upon this subject” which “fully sustain[ed]” the Court’s conclusion. The Court also quoted with approval from its prior opinion in Vestal (quoted above herein), and, recognizing the prior Texas cases expressing a contrary view, the Court overruled Moseley and Taylor. The Court then recognized Parchman (also quoted above herein) as setting forth the circumstances under which a jury could be discharged without creating a bar to future trial (i.e. consent and necessity), which included “where [the jurors] have been kept together for such time as to render it altogether improbable that they could agree.” The Court found that a mere three-and-a-half hours was not enough time to determine that the jury could never agree upon a verdict. Consequently, the Court concluded that the trial judge “abused his discretion” in discharging the jury and, therefore, that the defendant’s plea of double jeopardy was correct. As a result, the Court reversed the trial court’s judgment and dismissed the prosecution. E. Evaluation We first address the argument that the Texas double jeopardy provision applies only to acquittals. As explained above, the provision, remaining unchanged since 1876, states: “No person, for the same offense, shall be twice put in jeopardy of life or liberty, nor shall a person again be put upon trial for the same offense, after a verdict of not guilty in a court of competent jurisdiction.” At best, the conclusion that this provision applies only to acquittals is based upon one possible (but not necessarily the most likely) construction of an ambiguously punctuated sentence. That construction is problematic because it would render wholly redundant the entire clause of the statute referring to “jeopardy of life or liberty.” Moreover, earlier versions of the constitutional provision make clear that the language containing the words “put upon trial ... after a verdict of not guilty” was intended to comprise a separate clause from the clause referring to “jeopardy of life or liberty”: it was added as a separate clause to the Constitution of 1845 and all prior versions containing the language were punctuated (through the omission of a comma or the timely placement of a semicolon) in a manner that made this obvious. Further, none of the old judicial decisions have interpreted the provision to apply only to acquittals, and in fact, a 1900 decision held that the Texas double jeopardy provision actually contained two “jeopardy” protections: one prohibiting a second jeopardy of life or liberty and the other prohibiting a second trial after a verdict of not guilty. Finally, there are at least two reasons why the framers might have wanted two separate protections. By its language, the “verdict of not guilty” clause extends to all types of criminal prosecutions, regardless of the type of punishment, but the language “jeopardy of life or liberty” suggests application only to death or imprisonment — excluding offenses for which the only punishment is a fine or the forfeiture of property. Thus, an express acquittal— a verdict of not guilty — would have preclu-sive application in a broader spectrum of cases than do convictions or implied acquittals. In addition, if the framers were aware of various conflicts in double jeopardy jurisprudence among the states, having the two clauses would ensure that verdicts of not guilty would broadly be given the desired preclusive effect regardless of how other aspects of double jeopardy law were ultimately decided. We next address the State’s claim that the “mistrial species” of double jeopardy jurisprudence was not a part of the common law that formed the basis for the Texas constitutional provision. The State’s claim that this species of double jeopardy jurisprudence did not even emerge until 1949 is correct, if at all, only as a matter of federal constitutional law. As the above discussion shows, application of double jeopardy protection to the premature termination of trial occurred at the state level as early 1795 and became an accelerating trend in state jurisdictions during the nineteenth century. While Justice Powell’s criticism in Bretz that the Supreme Court ignored the pre-1791 history of double jeopardy jurisprudence in its own decisions possesses strong logical force, it must be remembered that the earliest Texas double jeopardy provision did not appear until 1836, and the current Texas provision was adopted in 1876. The Texas Constitution’s “Johnny-eome-lately” status raises the possibility that the framers did consider the emerging jurisprudence in other states in framing the Texas double jeopardy provision. Moreover, the 1876 version reflects substantial alterations from the text originally contained in the 1836 document, and for that matter, from the text of the counterpart provision contained in the Fifth Amendment to the United States Constitution. That significant alterations of language are present suggests that the framers of the Texas Constitution did not simply pattern the state double jeopardy provision after its federal constitutional counterpart but gave independent thought to its crafting. That the framers gave independent thought to the crafting of the provision suggests they would also have been cognizant of the developing double jeopardy jurisprudence and likely crafted the provision with that jurisprudence in mind. In 1836, the framers of the provision in the Republic of Texas Constitution would have had available the Pennsylvania double jeopardy case that drew a connection between “jeopardy of life or limb” and having the right to a jury trial “remain inviolate.” While it might have been a coincidence, the placement of these two protections in the same section of the Republic's constitution appears to be unusual, especially given the fact that the Republic’s constitution did contain a separate provision patterned after the jury trial guarantee found in the United States Constitution. Similarly, the addition of the words “or liberty” to the 1876 version of the state double jeopardy provision could have been a reaction to Ex parte Lange, decided two years earlier. In Lange, the United States Supreme Court decided that the federal double jeopardy protection extended to all crimes, regardless of the severity of the contemplated punishment. With “or liberty,” the framers of the Texas Constitution may have signaled their agreement with the notion that the double jeopardy protection extended even to misdemeanors but at the same time expressed minor disagreement with the Supreme Court’s opinion by carving a very limited exception for crimes that do not carry a risk of incarceration and for which the defendant was not acquitted by a verdict of not guilty. If the Texas framers considered Lange in drafting the 1876 version of the double jeopardy provision, that at least raises the possibility that they also considered the then-mushrooming state jurisprudence regarding the application of double jeopardy to mistrials and the implication in Lange itself that the double jeopardy protection might have some applicability in that context. Of course, the 1856 code provisions and the decisions in Moseley and Taylor are evidence to the contrary: both the Legislature and the Texas Supreme Court expressed the view that jeopardy was not implicated by the premature termination of trial, and these views were expressed within a relatively short time before the Constitution of 1876 was adopted. However, neither of these developments survived Powell (decided within a relatively short time after the adoption of the Constitution of 1876), which declared the code provisions unconstitutional and overruled the Texas Supreme Court cases. And neither of these earlier developments were unassailable on their own merits. The Legislature may have revealed its own confusion regarding double jeopardy jurisprudence when it indicated that the term “jeopardy” applied only when there was a prior conviction — a position that clearly did not comport with English common law, which also applied the concept of “jeopardy” to prior acquittals. Moseley (and Taylor, as it was based upon Moseley) grounded its decision in part upon what we now know is a false dilemma: characterizing the issue as a choice between keeping jurors together indefinitely until, induced by starvation, they issue a reluctant verdict, or, giving the trial court absolute, unreviewable discretion to declare a mistrial with no attendant double jeopardy consequences. The decision in Powell pointed to an approach between those extremes: give the trial court discretion but allow review for abuse of discretion — a practice with which appellate courts are now intimately familiar. The upshot of this discussion is that the applicability of the Texas double jeopardy provision to mistrials depends upon the vitality of Powell, and we cannot say with any confidence that Powell was wrongly decided, much less that the decision was flawed from the outset. At most, we can say that the issue was disputable, and that Moseley and Powell each advanced reasonable positions. Even if we decided that Moseley’s position was more likely correct as an historical matter, that would not be sufficient to overturn a precedent that has existed unmolested for over 120 years. Nor do practical considerations counsel otherwise. The framework of barring retrial when a mistrial has occurred without the defendant’s consent and absent manifest necessity has proven to be consistent and workable. While the framework does have its cost — allowing the occasional guilty person to go free — it serves to protect defendants from multiple harassing prosecutions, an important interest underlying the double jeopardy clause, and the exceptions of consent and necessity serve to reasonably limit any adverse impact. In accordance with stare decisis, we decline to overturn Powell’s holding that the Texas double jeopardy provision, with exceptions, protects a defendant against the premature termination of trial. IV. DEFENSE-REQUESTED MISTRIALS A. The Issues The State’s contentions can be accurately sorted into four categories. First, the State attacks the Bauder standard as inconsistent with the legal theory and purpose of the “mistrial species” of double jeopardy protection. The State claims that the Bauder standard goes awry by operating as a penal sanction against the prosecution rather than as a shield against a prosecutor’s attempt to abort a trial to prevent an impending acquittal. The State further claims that the penal nature of the sanction conflicts with subsequent caselaw declining to accord double jeopardy effect to appellate reversals. Second, the State attacks the opinion in Bauder as poorly reasoned. The State claims the opinion failed to examine Texas history, law, or jurisprudence but arrived at its holding based solely on the Court’s subjective notion of “fairness.” Third, the State contends that there is no historical support for the standard announced in Bau-der. The State points out that no Texas cases endorsed the standard as a matter of state constitutional law before the United States Supreme Court applied the federal double jeopardy clause to the states. And finally, the State contends that the Bauder standard is too amorphous for practical application. The State claims that current Texas caselaw reflects the confusion the standard has generated. B. Legal Underpinnings 1. Historical Developments The majority and concurring opinions in Bauder did not attempt to show that the framers of the Texas Constitution intended the standard set forth by the Bauder decision. Those opinions did not cite legal materials (cases, treatises, statutes, etc.) preceding the framing of the 1876 constitution that might have influenced the wording of the state double jeopardy provision or in some way reflected the intent of the framers. Nor did those opinions cite early Texas cases construing the 1876 provision. In other words, Bauder did not conduct, with regard to the “recklessness” standard, the kind of review we have conducted in part III regarding the mistrial question. Our research suggests that Bauder could not have been based upon such a review because the supporting evidence simply does not exist. As discussed above, the nineteenth century cases applying double jeopardy protection to the mistrial setting uniformly held that a defendant could be tried anew if he had consented to the mistrial. The first time an exception to that principle appears to have been mentioned in caselaw was in 1964 by the Supreme Court in United States v. Tateo. In that case, the defendant claimed that he was deprived of an opportunity to obtain a verdict of acquittal due to comments by the trial judge that coerced him into pleading guilty. The Supreme Court suggested that, had the defendant requested a mistrial on the basis of the judge’s comments, “there would be no doubt that if he had been successful, the Government would not have been barred from retrying him,” and further said it would be “strange” for the defendant “to benefit because of his delay in challenging the judge’s conduct.” In a footnote, the Court added, “If there were any intimation in a case that prosecutorial or judicial impropriety justifying a mistrial resulted from a fear that the jury was likely to acquit the accused, different considerations would, of course, obtain.” Two years later, in an early foreshadowing of the Kennedy standard, the Pennsylvania Supreme Court suggested that double jeopardy would bar retrial after a defense-requested mistrial if “the prosecution intentionally sought to infect the proceedings in order to abort the trial.” But, in the case before it, the prosecutor’s improper remarks “were not calculated to precipitate the mistrial,” and thus, the appellate court was not confronted with a case “in which the prosecution invited the mistrial in order to secure another, possibly more favorable opportunity to convict the accused.” The Pennsylvania Court would confront that case a year later in Commonwealth v. Warfield, where, after the defendant’s motion to suppress his confession was granted, the prosecutor, in opening statement, told the jury that the defendant had made a confession to the police. The defendant immediately moved for a mistrial, and the trial court granted the motion. The parties agreed that the prosecutor’s remark “was made for the specific purpose of causing a mistrial so that a ruling might be obtained from the Supreme Court of Pennsylvania upon the correctness of the trial judge’s suppression of the confession.” The Pennsylvania Supreme Court held that retrial of the first-degree-murder charge was barred by the state double jeopardy provision. In 1971, the United States Supreme Court suggested in dictum in a plurality opinion that “where circumstances develop not attributable to prosecutorial or judicial overreaching, a motion by the defendant for mistrial is ordinarily assumed to remove any barrier to reprosecution, even if the defendant’s motion is necessitated by prosecutorial or judicial error.” In a footnote, the Supreme Court stated as the “converse” proposition: “where a defendant’s mistrial motion is necessitated by judicial or prosecutorial impropriety designed to avoid an acquittal, reprosecution might well be barred.” Without a clear, definitive holding to guide them, lower courts struggled to define the precise contours of the rule suggested in dicta by the Supreme Court. The Fifth Circuit held that prosecutorial error would result in a jeopardy bar to retrial after a defense-requested mistrial if the error amounted to “gross negligence or intentional misconduct.” Other courts stated the rule differently, encompassing “deliberate and intentional misconduct” or conduct “designed to avoid an acquittal.” In 1976, the Supreme Court handed down United States v. Dinitz, which held that a defendant’s motion for mistrial ordinarily removed any double jeopardy bar to retrial. The Court recognized an exception for “governmental actions intended to provoke mistrial requests and thereby subject defendants to the substantial burden of multiple prosecutions.” Retrials would be barred “where ‘bad-faith conduct by judge or prosecutor’ ... threatens the ‘[harassment] of an accused by successive prosecutions or declaration of a mistrial so as to afford the prosecution a more favorable opportunity to convict’ the defendant.” With regard to the case before it, the Court held that the conduct in question “was not done in bad faith in order to goad the respondent into requesting a mistrial or to prejudice his prospects for an acquittal.” The negative disjunctive in the preceding sentence, at least in isolation, raised an ambiguity concerning whether the standard was met if the prosecutor intended (1) either to goad the defendant into moving for a mistrial or to prejudice the prospect of acquittal, or (2) both to goad the defendant into moving for a mistrial and to prejudice the prospect of acquittal. Taking the Supreme Court’s “application” statement in context with its abstract discussion of the law suggests the latter, but at least one state court opinion interpreted it as the former. The Supreme Court of Hawaii had an even broader reading of the rule as protecting against “misconduct designed to avoid an acquittal, or ... deliberate misconduct which has for its intended purpose the denial of the defendant’s constitutional right to a fair trial.” Other courts, though, took the view that intent to provoke a mistrial was an essential ingredient to relief on this type of claim. However, while recognizing a potential conflict with Dinitz, the Fifth Circuit nevertheless appeared to cling to its earlier “gross negligence” articulation of the standard, while several other courts, including our own, followed the Fifth Circuit formulation without any apparent awareness of the conflict. Against this backdrop, the United States Supreme Court decided Oregon v. Kennedy. The Court acknowledged that its previous cases had phrased the rule “with less than crystal clarity” and that this state of affairs had caused confusion in the lower courts. The Supreme Court then adopted the current federal rule for reasons that we will discuss later. This Court later recognized the Supreme Court’s clarification of the rule. Since Kennedy was decided, only seven state high courts, including this Court, have adopted a broader standard. Various nuances appear in each of these decisions, but they can be sorted roughly into three categories: (1) the “fair trial” approach, followed by Pennsylvania and Hawaii, (2) permitting a lesser culpable mental state with respect to the occurrence of the mistrial, followed by Oregon, Arizona, Texas and New Mexico, and (3) the “intent to avoid an acquittal” approach, followed by California. Ironically, the Arizona and Pennsylvania holdings represented a departure from the previous adoption of a standard essentially identical to the one espoused in the Oregon v. Kennedy case before Oregon v. Kennedy had been decided. As the present discussion shows, the rule according double jeopardy consequences to a defense-requested mistrial (under certain limited circumstances) was a relatively recent innovation. There was no authority in any jurisdiction for such a rule at the time the Texas Constitution of 1876 was adopted, despite the fact that there existed at the time numerous authorities applying double jeopardy protections to the mistrial context in general (but only when the defendant did not consent to a mistrial). When this Court originally adopted the rule, it did so in response to federal jurisprudence rather than as an effort to independently construe our state’s own double jeopardy protection. 2. Purpose of the Mistrial Double Jeopardy Protection Absent any reason to believe the framers of the Texas Constitution intended for the double jeopardy protection to apply to a de/ense-requested mistrial, why recognize a rule in that context at all, however narrowly the rule is crafted? The answer must be that the posited circumstances show the defendant’s consent to a mistrial to be a sham. In Kennedy, the Supreme Court held that the defendant’s valued right to complete his trial before the first jury would be a “hollow shell” if retrial were permitted after the prosecution, through its conduct, intentionally precipitated a mistrial. According to the Court, it is not enough, however, that the defendant faced a “Hobson’s choice” between giving up his first jury and continuing with a tainted trial. Rather, the question is whether the defendant retained primary control over the course to be followed. The Bander Court claimed that situations encompassed by its “recklessness” standard were “constitutionally indistinguishable” from those encompassed by the specific-intent standard of Oregon v. Kennedy, but the Court’s justifications for that contention fall short. With no authority whatsoever, the Bander Court contended that “the right to a trial before the jury first selected is the right to a fair trial before that jury.” That statement, however, conflates the double jeopardy protection with more generalized notions of due process and due course of law. As the California Supreme Court has recognized, “[Djouble jeopardy is neither another form of due process protection ensuring the propriety of the criminal trial nor a means to protect against outrageous government conduct.” The “remedy of a new trial” is “sufficient to vindicate both the citizen’s interest in a fair trial and the societal interest in bringing those properly found guilty to punishment.” The question, for double jeopardy purposes, is not whether the defendant’s trial was “fair” but whether requesting a mistrial was ultimately his decision. The Bander Court suggested that a defendant’s decision in a “recklessness” situation would not be a “free” decision, but the question is not whether the decision was “free” in the sense of being unconstrained but whether the decision was his own, albeit in the face of a dilemma. To say that the decision was not the defendant’s own is to say that the decision was in reality made by someone else, e.g. the prosecutor. But when a prosecutor is merely reckless, one cannot say the prosecutor has made the decision to seek a mistrial. Only when the prosecutor intends to provoke the defendant’s mistrial motion can it be said that the prosecutor, rather than the defendant, has exercised primary control over the decision to seek the trial’ termination. The Bauder Court also claimed that, for double jeopardy purposes, the distinction between intent and recklessness was “fuzzy and imponderable,” and the Court did “not believe that the purpose of the constitutional right here in issue really has anything to do with the prosecutor’s intent.” But actually, the distinction between the two culpable mental states is clear, and the requirement of intent is important. As discussed above, whether the prosecutor intends to bring about a mistrial is critical to determining whether he, rather than the defendant, has exercised primary control over whether a mistrial is sought. In addition, the different culpable mental states reflect different purposes of the prosecutorial misconduct, and those different purposes are important in the double jeopardy context. As discussed in part III, one of the common threads of the nineteenth century double-jeopardy mistrial cases was that the declaration of mistrial, without necessity or the defendant’ consent, constituted an implied, or virtual, acquittal. When the prosecutor’s purpose is to produce a mistrial, he has, in essence, sought that acquittal. But when the prosecutor’s purpose is to produce a conviction, even at the substantial risk of mistrial, the prosecutor has not sought an acquittal. One method of bringing the distinction into focus is to ask what happens if the prosecutor succeeds in his purpose. Under the Kennedy standard, a prosecutor who succeeds in causing a mistrial also succeeds (presumably to his dismay) in barring further prosecution. To be consistent, one would expect, under the Bauder standard, that a prosecutor who succeeds in obtaining a conviction through his reckless conduct would also be faced with a double jeopardy bar to retrial when that conviction is overturned, because of that conduct, pursuant to a post-verdict motion for new trial or on appeal. Indeed, the five other jurisdictions recognizing a broader rule than that articulated in Oregon v. Kennedy that have addressed the issue have expressed the view that the rule applies even when a mistrial was not granted, a verdict was obtained, and the case was overturned at a subsequent stage of the proceedings. The lead cases in Pennsylvania and Hawaii actually involved convictions that were reversed on appeal while the New Mexico case involved a post-verdict motion for new trial. In a later case, the Arizona Supreme Court extended its own rule to appellate reversals. And while we have not found an Oregon decision applying the rule to a case that proceeded to verdict, the original statement of the rule — referring to the “resulting mistrial or reversal” — strongly suggests that it does. Under this Court’s subsequent cases— Ex parte Davis and Ex parte Mitchell— retrial is not barred under the Texas Constitution if the case proceeds to verdict and is overturned later, on appeal. This is so because the defendant “was not denied his right under the double jeopardy clause to have the charges against him tried to a verdict before the first tribunal.” So, if the reckless prosecutor succeeds (in obtaining a conviction through dubious means), then in Texas no double jeopardy bar arises. But, if the reckless prosecutor fails, because a mistrial is declared, then the defendant obtains greater relief, a bar to future prosecution, than he would have obtained if the prosecutor had