OPINION

delivered the opinion of the Court

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.”1 In *337Bauder 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.”2 We granted review to reexamine Bander’s holding.3 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 Miranda4 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?”5 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? 6

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 — 7

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.8 All three objections were sustained, and instructions to disregard were requested and given in connection with the second two sets of questions.9 A mistrial was not requested on the first question, was requested and denied with regard to *338the second set of questions, and was granted after the third set of questions.10

Appellant later filed a pretrial habeas application, claiming that any subsequent prosecution was barred under double jeopardy principles, but the trial court denied relief.11 After discussing this Court’s latest opinion on the Bauder standard,12 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.” 13

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.”14 Precedent can be overruled, however, if the reasons for doing so are weighty enough.15 Some factors supporting the overruling of precedent are: (1) that the original rule or decision was flawed from the outset,16 (2) that the rule’s application produces inconsistent results,17 (3) that the rule conflicts with other precedent, especially when the other precedent is newer and more soundly reasoned,18 (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,19 and (5) that the reasons that support the rule have been undercut with the passage of time.20

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.21 If this interpretation were adopted, *339the 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.22 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.23

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,24 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,”25 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.”26

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.27

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.28 His opinion quotes the state provision with the following in italics, suggesting that he believed that the italicized clause modifies the entire provi*340sion 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.” 29

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.”30 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.31 A defendant was considered to be placed twice in jeopardy upon a second trial only if there existed a prior conviction or acquittal.32 Essentially, the doctrine embodied “a res judicata policy” for criminal cases.33 That policy required an actual acquittal or conviction for its implementation.34 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.35

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.36 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.”37

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.”38 The rule was originally an absolute command that “once banded to*341gether a jury should not be discharged until it had completed its solemn task of announcing a verdict.”39 Even during deliberations, the jury was required to be kept together “unfed and without drink” even “till death if they do not agree.”40 The harshness of this rule was soon mitigated with two exceptions: necessity41 and consent of the defendant.42 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.43

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.44 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.”45 “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.”46

In Bretz, the Supreme Court majority acknowledged this history but responded, “But this constitutional understanding was not destined to endure.”47

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.” 48 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.”49

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.50 But North Carolina reaffirm*342ed its earlier holding as a correct articulation of double jeopardy principles,51 and it was joined in its position by Pennsylvania, Tennessee, South Carolina, and Ohio.52 Subsequently, several other jurisdictions switched sides to join the trend towards incorporating Lord Coke’s rule into double jeopardy jurisprudence.53 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.54

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.55 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.56 “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.”57 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 *343entitled to be also discharged.”58 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.”59

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.60 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.”61 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.62

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.63 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).64 Although most of the *344cases in which relief was granted involved the discharge of the jury during deliberations or later,65 several state courts expansively pronounced that jeopardy attached when the jury was impaneled, sworn, and charged with the case.66 By “charged with the case,” these authorities meant charged in the sense of being charged as jurors at the beginning of trial.67

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.68 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.69

*345In 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.” 70 But the Supreme Court made no clear pronouncement to that effect in the nineteenth century,71 and in fact, it implied in an 1874 decision — Ex parte Lange — that Perez was indeed a double jeopardy case.72

In the meantime, states on both sides of the issue cited Perez in support of, or at least in discussion of, their respective positions.73 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.74 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.75

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,76 or expressly claimed that it did not,77 or claimed that a new trial might be barred if prejudice occurred.78 Others avoided general pronouncements, saying simply that there was no bar if necessity existed79 or the defendant consented.80

*346D. 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.” 81 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.”82 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.83 With the defeat of the Confederacy, and the resultant new constitution of 1866, the semicolon was changed back to a comma.84 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.” 85 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,”86

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 *347exempts 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.87

The provisions of Section 19 were later carried forward as Section 20 of the 1879 Code of Criminal Procedure.88

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.89 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.90 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.”91 The court observed that the defendant’s position was maintained by “Lord Coke ... and the courts of Pennsylvania, Tennessee and some other states,”92 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.” 93 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.94 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.95 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.” 96

The next year, the Texas Supreme Court revisited the issue in Taylor v. State.97 The defendant in that case was initially indicted for the murder of N. Evans, but the proof at trial showed that he *348killed Morgan Evans.98 After the State’s opening argument, the prosecutor entered a nolle prosequi (a nonsuit of the prosecution), and the jury was discharged.99 The State subsequently filed a new indictment charging the defendant with the murder of Morgan Evans.100 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.”101 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.102 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.”103

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).104 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.105

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.”106 The jury was discharged over the protest of the defendant, and a new indictment, alleging the correct name, was filed.107 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.108 “[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 *349accused 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.109

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.110

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.111 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.112

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.113

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.114 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.115 The trial judge sent the jurors out to deliberate for another hour, after which they returned to again say they could never agree.116 The trial judge sent the jurors out a third time, and they returned an hour later with the same results.117 The trial judge then discharged the jury over the defendant’s objection.118

In attempting to ascertain the meaning of the double jeopardy provision found in the Texas Constitution, the Court of Ap*350peals 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).119 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.”120 But the Court held that the Legislature had no authority to construe a constitutional provision “which has become fixed and settled by judicial determination.”121 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.122

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.123 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.”124 The Court found that this view was “uniformly” supported by “the decided weight and respectability of authority.” 125 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.126 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.127

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.”128 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.129 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.130 As a result, the Court reversed the trial court’s judgment and dismissed the prosecution.131

E. Evaluation

We first address the argument that the Texas double jeopardy provision ap*351plies 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.132

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 trial133 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 *352Supreme 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.134

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.135 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 *353within 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 sup*354port 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.136 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.137 Nor did those opinions cite early Texas cases construing the 1876 provision.138 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.139 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.140 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.”141 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.”142

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 *355the proceedings in order to abort the trial.”143 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.”144 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.145 The defendant immediately moved for a mistrial, and the trial court granted the motion.146 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.”147 The Pennsylvania Supreme Court held that retrial of the first-degree-murder charge was barred by the state double jeopardy provision.148

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.”149 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.”150

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.”151 Other courts stated the rule differently, encompassing “deliberate and intentional misconduct” 152 or conduct “designed to avoid an acquittal.”153

*356In 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.154 The Court recognized an exception for “governmental actions intended to provoke mistrial requests and thereby subject defendants to the substantial burden of multiple prosecutions.”155 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.” 156 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.”157

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.158 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.”159 Other courts, though, took the view that intent to provoke a mistrial was an essential ingredient to relief on this type of claim.160 However, while recognizing a potential conflict with Dinitz, the Fifth Circuit nevertheless appeared to cling to its earlier “gross negligence” articulation of the standard,161 while several other courts, including our own, followed the Fifth Circuit formulation without any apparent awareness of the conflict.162

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.163 The Supreme Court then *357adopted the current federal rule for reasons that we will discuss later.164 This Court later recognized the Supreme Court’s clarification of the rule.165

Since Kennedy was decided, only seven state high courts, including this Court, have adopted a broader standard.166 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,167 (2) permitting a lesser culpable mental state with respect to the occurrence of the mistrial, followed by Oregon, Arizona, Texas and New Mexico,168 and (3) the “intent to avoid an acquittal” approach, followed by California.169 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.170

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.

*3582. 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.171 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.172 Rather, the question is whether the defendant retained primary control over the course to be followed.173

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,174 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.”175 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.”176 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.”177 The question, for double jeopardy purposes, is not whether the defendant’s trial was “fair” but whether requesting a mistrial was ultimately his decision.178 The Bander Court suggested that a defendant’s decision in a “recklessness” situation would not be a “free” decision,179 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.180 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 prose*359cutor 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.” 181 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 issue182 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 appeal183 while the New Mexico case involved a post-verdict motion for new trial.184 In a later case, the Arizona Supreme Court extended its own rule to appellate reversals.185 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.186

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 *360is overturned later, on appeal.187 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.”188 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 achieved his purpose. Thus far, Texas appears to be the only jurisdiction adopting this illogical position.

C. Practical Considerations

1. The Kennedy Standard

In Kennedy, the Supreme Court criticized as too vague certain proposed general standards for determining what type of prosecutorial conduct should result in a jeopardy bar after a mistrial on the defendant’s motion. It found that standards such as “bad faith,” “harassment,” and “overreaching” offered “virtually no standards for ... application.”189 “By contrast,” the Court said, “a standard that examines the intent of the prosecutor, though certainly not entirely free from practical difficulties, is a manageable standard to apply. Inferring the existence or nonexistence of intent from objective facts and circumstances is a familiar process in our criminal justice system.”190

The Bander Court suggested, however, that the “intent” standard was inadequate because it was too difficult to distinguish between intent and recklessness with respect to a prosecutor’s culpable mental state regarding the occurrence of a mistrial.191 In support of this claim, the Court contended, “we do not perceive a distinction of constitutional significance between conduct of prosecuting attorney in which he intends to cause a mistrial and conduct of a prosecuting attorney which he is aware is reasonably certain to result in a mistrial.”192 But, as an examination of our own Penal Code illustrates, “reasonable certainty” is the hallmark of a “knowing” mental state and is quite different from the “substantial risk” standard of recklessness.193 The Court claimed that the distinction between culpable mental states was “far too insensitive a criterion for decision in these cases,” but the Court’s conclusion seems to reflect its own confusion regarding which mental state it was really talking about.

In his concurring opinion, Judge Baird offered a more straightforward statement of what the Court may have really been concerned with: that intent to cause a mistrial would be “virtually impossible to *361prove.”194 In this vein, Judge Baird commented that his independent research “failed to reveal even a single case where the Kennedy burden was met.”195 Our research of high state court opinions has uncovered seven, five of which were in published opinions before Bander was decided.196 One of the seven cases is War-field, which, as we have already discussed, presciently anticipated the standard before Kennedy was decided, but the other six cite Kennedy and follow it.197 In two of the seven cases, the State conceded the prosecutor’s intent,198 while the other five involved a trial court finding that the prosecutor intended to cause a mistrial.199 In all seven cases, a finding of the requisite intent was supported by the record.200

*362Obviously, obtaining relief under the Oregon v. Kennedy standard is rare. That is understandable, however, when one considers that prosecutors do not ordinarily attempt to “throw” their cases, even when problems are encountered. Moreover, the double jeopardy sanction renders such conduct self-defeating.

And while we are not aware of any decisions granting relief under Kennedy where the proceedings contain neither a state concession nor a favorable trial court finding, the absence of any such cases does not necessarily point to any inadequacy in the standard. For questions that are highly fact-intensive, the prevailing party has a significant advantage in the appellate forum, and if the prevailing party is not the one charged with the burden of proof, the advantage may be especially great.201 That, however, is the nature of the system, where the trial proceedings are the “main event, rather than a tryout on the road.”202 Trial courts are in the best position to determine whether a prosecutor’s conduct evinces an intent to cause a mistrial, and Texas provides defendants with the opportunity to litigate the question in the trial forum (before the second trial),203 and even before a jury (on retrial).204

Finally, one should expect that such an extreme remedy — what is essentially an acquittal, “the greatest form of relief in the criminal system” 205 — invoked under the most inhospitable of circumstances — a request for relief as a result of an action the requesting party procured, which would ordinarily give rise to estoppel 206— would be difficult to obtain and would seldom be granted.

2. The Bauder Standard

a. Inception

The Bauder Court suggested that its standard would have “practical advantages” because it was “less subjective.”207 But, as this Court would later recognize in Peterson II, the proffered “recklessness” standard turned on the prosecutor’s mental state just as much as Kennedy’s, intent standard, and therefore was “no less subjective.” 208 The former was more lenient, and thus, would at least seem easier to meet, but “easier” in this context is not necessarily “better.” Our later opinion suggested that the “more objective” rationale might carry some force “if trial and appellate courts focus primarily upon the objective facts and circumstances of the prosecutor’s conduct and the events which led to that conduct.” 209 But, as discussed above, the Supreme Court had previously recognized in Kennedy under its own standard that intent could be assessed in light of the “objective facts and circum*363stances.”210 The Bauder standard simply offers no advantage in this respect.

As a well-recognized culpable mental state, “recklessness” would at least seem to be reasonably specific, avoiding the Supreme Court’s criticism of a generalized “overreaching” standard, but the Bauder formulation did not completely adhere to the familiar definition found in the Penal Code because it did not require that the “risk” be a “substantial” one.211 Even if that element were implied, confusing language in the Bauder opinion suggested that the standard’ application would not be that simple. Statements by the Court that the defendant’s right to a “fair trial” was an important consideration, that the prosecutor’s specific intent was “irrelevant,” that the double jeopardy question might turn on whether the prosecutor was “reasonably certain” that a mistrial would result, and that the rule formulated was somehow “less subjective” than one that turned on specific intent all suggested that the standard would be more complex than simply evaluating whether the prosecutor was reckless with regard to whether a mistrial would occur.212

The Bauder opinion also contained language that actually suggested that granting a defendant’ requested mistrial would usually result in a double jeopardy bar. It did so by emphasizing that “trial conditions must be extreme before a mistrial is warranted in Texas,” that “[accordingly, the line between legitimate adversarial gamesmanship and manifestly improper prosecutorial methods should be difficult for most prosecuting attorneys to cross unless they do it on purpose,” and that crossing that line, “either deliberately or recklessly,” results in a bar to further prosecution.213 Such a statement harkens back to the generalized “overreaching” standard that the Supreme Court warned against and turns double jeopardy jurisprudence on its head by making jeopardy preclusion from a defense-requested mistrial a common occurrence rather than the rare exception. That was probably not the Court’s intent, but the language was there, a trap waiting to be sprung by an unwary trial or intermediate appellate court.

Setting aside confusing language in the Bauder opinion, the recklessness standard poses some practical problems in this context. “Every act on the part of a rational prosecutor during a trial is designed to ‘prejudice’ the defendant” in front of the jury so that it will convict him.214 The prejudice should be fair rather than foul, but in the heat of battle, prosecutors may overstep their bounds. While a prosecutor clearly knows he should not try to cause a mistrial, he may be a lot less certain what conduct an appellate court would decide carried a substantial risk of a mistrial that the prosecutor is then found to have consciously disregarded. In his dissent, Presiding Judge McCormick doubted that prosecutors would “know with any certainty what conduct is prohibited by this rule.”215

*364Another concern is that a broad standard is likely to unduly deter the granting of mistrials when otherwise warranted by the circumstances because a trial judge would rather leave the decision to reverse to an appellate court than create a situation that would forever bar prosecution of a defendant who may well be guilty.216 In Kennedy, the Supreme Court isolated two ways in which a defendant could be harmed as a result: (1) the defendant loses “some of the advantages secured to him” by the double jeopardy protection because, even if he ultimately obtains an appellate reversal, he will be subject to additional time, expense, and anxiety that might have been avoided by obtaining a mistrial, and (2) the defendant fails to obtain any relief at all (which he would otherwise have obtained through a mistrial) because of the appellate courts’ practice of deferring to trial court rulings.217 As discussed earlier, most jurisdictions that have adopted a broader rule have taken the logical step of extending the double jeopardy sanction to appellate reversals, which to some extent ameliorates this concern. If the defendant obtains relief on appeal that he should have obtained by mistrial, he will receive the same double jeopardy protection. At least theoretically, the trial court may have less incentive to deny a meritorious mistrial motion since the double jeopardy consequences of mistrial and appellate reversal are the same. But practically, the trial court would understand that its decision to grant or deny the mistrial might affect the appellate court’s determination regarding the flagrancy of the misconduct, and so these jurisdictions do not completely escape the issue. Of course, in Texas, the concern applies with full force because, due to Davis and Mitchell, the trial court does know that reversal on appeal carries none of the double jeopardy risks of granting a motion for mistrial. This problem is similar to that posed by the disparate treatment under Texas law of punishment errors at the motion for new trial stage and on appeal, and a trial court’s refusal to grant a new trial on the basis of punishment error (and our upholding of that refusal)218 comports with the type of conduct foreseen by the Supreme Court in Kennedy. But the problem is even worse here because, the more egregious the pros-ecutorial misconduct, the greater the incentive the trial court has to ensure the completion of the trial to avoid a double jeopardy bar.

b. Subsequent Cases

Problems with applying the Bauder standard began with this Court’s remand in Bauder itself. In explaining the recklessness standard, one passage in this Court’s opinion stated: “Under this rule, the prosecutor is not accountable for mistrials when the trial judge need not have *365granted the defendant’s motion.”219 On remand, the San Antonio Court of Appeals concluded that this sentence created the first prong of a two-prong test: (1) whether the motion needed to be granted, and (2) whether the prosecutor’s conduct was intentional or reckless.220 The court of appeals further concluded that this first prong concerned “whether the trial court would have abused its discretion if it had denied [the] motion for mistrial” which in turn required determining “whether any reasonable view of the record supports not granting the mistrial.”221 After discussing the trial incident that led to the mistrial, a police witness’s description of the defendant engaging in a sex act in response to a prosecutor’s question about what the officer observed at a certain location, the appellate court held that the trial court did not have to grant the motion for mistrial because any harm flowing from the testimony could have been cured by an instruction to disregard.222 The court of appeals did express some concern that this Court’s holding required it to ignore the trial court’s findings.223

The defendant petitioned for review, and this Court decided that the Court of Appeals applied the wrong standard: “The question is not the correctness of the ruling granting the mistrial. The question under the Double Jeopardy Clause is whether the defendant truly consented to the mistrial.” 224 Quoting from the Supreme Court cases in Jom, Dinitz, and Kennedy, and from this Court’s earlier decision in Bauder, this Court held in Bau-der III that the issue was whether the defendant’s motion for mistrial was “made in response to ordinary reversible error” or was “required ... because the prosecutor deliberately or recklessly crossed ‘the line between legitimate adversarial gamesmanship and manifestly improper methods’ ... that rendered the trial before the jury unfair to such a degree that no judicial admonishment could have cured it.”225 Consequently, this Court reversed the court of appeals’s decision in Bauder II and remanded for reconsideration.226 Judge Baird concurred, expressing “fear that the majority opinion potentially blurs the critical distinction between mistrials granted in response to ‘ordinary’ reversible error and those granted because of ‘prosecutorial overreaching.’ ” 227 Judge Keller dissented, arguing that the standard articulated by this Court in Bauder III did not appear to differ materially from the standard applied by the court of appeals in Bauder I/.228

On the second remand, the court of appeals acknowledged Bauder III and proceeded to address the double jeopardy question for a third time.229 This time, the court discussed evidence that the prosecutor was surprised by the witness’s response to his question.230 Consequently, *366the court concluded that the prosecutor had not been aware of a risk that an event for which he was responsible would cause a mistrial and had not “deliberately or recklessly crossed the line between legitimate adversarial gamesmanship and manifestly improper methods.”231 In a footnote, the intermediate appellate court reiterated its belief that the mistrial was unnecessary because an instruction to disregard would have cured the error, and it expressed perplexity at what it saw to be conflicting statements on the matter by this Court in Bauder III.232 Affirming the trial court’s denial of habeas relief, the court of appeals finally succeeded in ending Bauder’s appellate orbit.233

Before Bauder III was decided, the Dallas Court of Appeals handed down a decision in State v. Lee.234 In that case, tried in October 1995, a mistrial was granted after the prosecutor said in opening statement that the defendant had told a police detective that he did not want to talk to him and that the detective should contact the defendant’s lawyer.235 When the State sought to retry the case, the defendant filed a habeas action, and in December 1995, the trial judge denied relief. But, the defendant subsequently filed a motion to reconsider in light of Bauder, and the trial judge thereafter granted relief, dismissing the prosecution.236 Relying upon Bauder II, the court of appeals held that the double jeopardy analysis entailed a two-step process: (1) whether the mistrial was “properly granted,” and (2) whether the mistrial was “made necessary by the deliberate or reckless conduct of the prosecutor.” 237 The Lee court expressed concern about the first prong because “many times, because of concern for an accused’s right to a fair trial, a trial judge will err on the side of caution and grant a mistrial. Many considerations factor into this decision and many of them cannot be adequately reflected in the record.”238

The appellate court nevertheless proceeded to the first prong, holding that it would be met if it were shown that error was committed and the error could not be cured by an instruction to disregard.239 Finding the prosecutor’s statement to be a comment on the right to counsel in violation of caselaw and Article 38.38 of the Code of Criminal Procedure and finding that the comment “could well have shaped the jury’s entire view” of the defendant and the case, the appellate court concluded that incurable error had occurred.240 With regard to the second prong, the court of appeals concluded that the trial court was within its discretion to believe that the experienced prosecutor, although not intending to cause a mistrial, was reckless with regard to whether her comment would require a mistrial at the defendant’s request.241

The State petitioned for review, and this *367Court reversed (Lee J/).242 First, this Court determined that the prosecutor’s statement was properly characterized as a comment on the right to silence rather than the right to counsel.243 Second, the Court determined that the appropriateness of the type of comment at issue — on pre-arrest, pre-Miranda silence — was a question of first impression in Texas and was unsettled in the federal system.244 Finally, the Court concluded that, in view of the “state of the law, the prosecutor’s actions could not have been intentional or reckless.” 245

In explaining that conclusion, the Court offered what can only be described as a troubling array of definitions of “intentional” and “reckless” conduct. According to the opinion, a prosecutor engages in “intentional” conduct when:

(1) believing that he cannot obtain a conviction under the circumstances with which he is confronted, and given the admissible evidence then at his disposal, deliberately offers objectionable evidence which he believes will materially improve his chances of obtaining a conviction, and the law considers the prejudicial effect of such objectionable evidence to be incurable even by a firm judicial admonishment to the jury

or ...

(2) the objectionable conduct of the prosecutor was intended to induce a motion for mistrial.246

The first definition of “intentional” conduct is not a definition of the “intent” standard found in Oregon v. Kennedy, nor does the definition clearly fall within the “recklessness” standard found in Bander. Rather, the so-called definition appears to be from language in Bander that took on a life of its own. The Court also offered three definitions of the “reckless” standard:

(1) 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,

(2) he is aware his conduct is reasonably certain to result in a mistrial, or ...

(3) he is aware that his conduct creates a risk that a mistrial is reasonably certain to occur, but consciously disregards that risk.247

Although all three of these formulations were derived from language in Bander, only the first seems to comport with the original standard, but, as with the original formulation, it does not require the risk to be “substantial” and, therefore, is not the definition of “reckless” as traditionally understood in Texas. The second definition appears to be a “knowledge” standard, rather than one of “recklessness,” while the third appears to be some sort of hybrid between the two culpable mental states.

After this Court’s decision in Lee II, the Dallas Court of Appeals confronted another Bander claim in Ex parte Peterson.248 In that case, the prosecutor elicited testi*368mony that the trial court believed was barred as a result of a discovery violation.249 Although the prosecutor contended at trial that she believed the matter was contained in an offense report the defendant had timely received, the trial court granted the defendant’s motion for mistrial “to give [the prosecutor] another opportunity to give discovery to the defendant.”250 However, the defendant later filed a pretrial habeas application, alleging double jeopardy, which the trial court granted.251

The Dallas Court of Appeals affirmed, concluding that the prosecutor had “deliberately or recklessly cross[ed] the line between legitimate adversarial conduct and manifestly improper methods” and that the trial judge could have concluded that the error was incurable.252 In its discussion, the court of appeals relied upon the array of definitions recited in Lee il.253 The court explained that “the prosecutor knew the tapes were ‘severely damaging’ to [the defendant] and that admission of the tapes ‘would have substantially increased the State’s chances of securing a conviction.’ ”254

We vacated that decision in Peterson I/.255 Acknowledging that the Bauder standard “has not always proven easy to apply,” 256 we proceeded to “clarify the standards” under which the Texas double jeopardy protection could be invoked.257 We set forth a three-pronged test:

(1)Did manifestly improper prosecuto-rial misconduct provoke the mistrial?

(2) Was the mistrial required because prejudice produced from that misconduct could not be cured by an instruction to disregard? and

(3) Did the prosecutor engage in that conduct with the intent to goad the defendant into requesting a mistrial (Kennedy standard) or with conscious disregard for a substantial risk that the trial court would be required to declare a mistrial (Bauder standard)?258

We also gave a nonexclusive list of factors to consider in evaluating cases under the third prong of the test: (1) whether the trial was “going badly for the State,” (2) whether the misconduct was repeated despite admonitions from the trial court, (3) whether the prosecutor provided a reasonable, “good faith” explanation for the conduct, (4) whether the conduct was “clearly erroneous,” (5) whether there was a legally or factually plausible basis for the conduct, and (6) whether the prosecutor’ actions leading up to the mistrial were consistent with inadvertence, lack of judgment, or negligence, or instead, were consistent with intentional or reckless misconduct.259

The Court’s opinion in Peterson II obviously went to great lengths to clarify the messy jurisprudence flowing from Bauder. It brought the standard closer to true recklessness by explicitly requiring that the “risk” be “substantial,” implicitly jettisoned various other confusing formulations of the “intent” and “recklessness” standards articulated in Lee II, clarified that *369the error in question must have been incurable, and sought to establish a clear framework by formulating a three-pronged approach and suggesting a number of relevant factors to consider with regard to the third prong, which it considered to be “the most problematic.” 260 But the Court’s candid admission that the Bauder test was “no less subjective” than the intent test articulated in Oregon v. Kennedy led Judge Hervey to say that the Bauder holding has “been transformed into a made-up constitutional rule in search of a rationale to justify its existence.”261 And despite the attempted clarifications, Judge Keasler still maintained that the Bauder rule is “ill-conceived, historically unsound, and imprecise.” 262 Judge Hervey also concluded that the attempted clarification was ultimately futile: “Instead of pursuing a quixotic attempt to clarify this jurisprudence, the Court should reexamine it.” 263 Despite the new prongs and factors, Judge Hervey said, the Bauder standard “will continue to be difficult to apply and will do little to promote (if not actually frustrate) the interests protected by double jeopardy principles.”264

Courts from other jurisdictions have mounted similar criticisms of the Bauder standard. The Supreme Court of Connecticut complained that the tests in jurisdictions recognizing a standard broader than found in Oregon v. Kennedy, “with the possible exception of California, lack the requisite clarity to achieve an optimal balance between the defendant’s double jeopardy rights and society’s interest in enforcing criminal laws.” 265 And the Supreme Court of California, while contending that the Oregon v. Kennedy standard did not completely protect the interests underlying California’s double jeopardy provision, found the recklessness/willfulness/indifference standards, such as the one articulated in Texas, to be “less than satisfactory, because none articulates the precise double jeopardy basis for a conclusion that the principles underlying a defendant’s double jeopardy interest have been violated.” 266 Consequently, the California Supreme Court concluded that “as applied to different factual settings, each test improperly may mandate double jeopardy relief (that is, barring any trial) for instances of prosecutorial misconduct that more appropriately should be remedied by reversal and retrial.”267

One indication that these criticisms are on target is that we have more cases pending before us. On remand from Peterson II, the Dallas Court of Appeals analyzed the case under the Pederson II framework and concluded that habeas relief should be denied, and that case has not again come *370before us.268 However, three other Ban-der cases, including the present one, were pending before us during the past year. In Ex parte Wheeler and the present case, the trial courts denied relief, the courts of appeals reversed and granted relief, we vacated the courts of appeals’ decisions in light of Peterson II, and the courts of appeals maintained their position that relief should be granted.269 In State v. Ma-sonheimer, the trial court granted relief, but the court of appeals reversed, holding that relief should be denied.270 We granted petitions for discretionary review in all three cases. Recently, we issued an opinion in Wheeler, concluding that the court of appeals had erred in granting relief because the evidence supported the habeas trial court’s finding that the prosecutor did not possess the culpable mental state required for a double jeopardy violation.271

If we were discussing a standard that “acquire[d] content only through application,” such as “reasonable suspicion” to conduct a stop,272 or even a standard that is expected to be utilized in many cases, we would not be surprised that the high court of a state would be called upon numerous times to clarify and expound the standard’s meaning and reach. But the type of standard at issue is one that should very rarely be implicated, so the frequency with which we are called upon to construe the standard points to a problem with the standard itself.

One problem seems to be that the Court has never really been able to describe adequately what it believes double jeopardy should protect that is not already protected under Oregon v. Kennedy. The recklessness standard in Bander appears to have been only an approximation, and in subsequent cases, the Court has added conditions (e.g. conduct improper from an objective standpoint, harm flowing from conduct not amenable to cure) that it has believed made the approximation closer, but it never quite gets to the unarticulated (and unarticulable) ideal that the Court seems to have been striving for. To be sure, we have suggested that these other conditions also attach to the Kennedy standard, but the United States Supreme Court has not said so.273 It may be that these conditions do indeed attach, but it has never been necessary to say that they do because, when the prosecutor actually intends to provoke a mistrial, the other conditions will almost certainly be present. But when the standard varies from the Kennedy “intent to provoke a mistrial” test, other conditions must be expressly added to “guard against an unwarranted imposition of the double jeopardy bar.”274

The problem is that the refinement never seems to end. If we continue down the Bander path, we must either accept at *371some point that some defendants who are not entitled to a double jeopardy acquittal will nevertheless obtain one under the Bauder standard, or we must continually refine the standard to reach for that elusive unarticulated ideal — overturning every grant of relief under Bauder along the way except on the rare occasion when relief would also be supported by Oregon v. Kennedy. The simple explanation for the never-ending path toward this “separate” state constitutional ideal is that it does not exist, because the real ideal is the Oregon v. Kennedy standard.

D. Evaluation

The Bauder opinion was flawed in a number of respects. It was not based upon an historical understanding of double jeopardy available to the framers of the Texas Constitution, and the standard it formulated does not accurately reflect the purposes of the double jeopardy protection. The opinion’s justifications for the new standard were faulty, and the standard, as articulated, was confusing. Practical difficulties in applying the standard have followed. Trial courts and courts of appeals have had difficulty correctly interpreting and applying the Bauder standard to various fact situations, and this Court has struggled to clarify it. Further, the Bauder standard conflicts logically with this Court’s treatment of appellate reversals in Davis and Mitchell, and partly as a result, undoubtedly carries the risk of unduly discouraging trial courts from declaring mistrials when warranted. And, the conflict between Bauder and Davis/Mitchell is relevant not only for its practical consequences but also as a conflict in legal precedents that must be resolved. By contrast, the Kennedy standard is workable, appropriately narrow, and comports with the purpose of the double jeopardy provision’s application to the mistrial setting. Consequently, we overrule Bauder and its progeny (Bauder III, Lee II, Peterson II). As a matter of state constitutional law, we adopt the standard articulated by the United States Supreme Court in Oregon v. Kennedy for determining when to grant double jeopardy relief after a defense-requested mistrial, and we reaffirm the holdings in Davis and Mitchell.

The judgment of the Court of Appeals is reversed, and the case is remanded for analysis of appellant’s claim under the standard articulated in Oregon v. Kennedy.

COCHRAN, J., filed a concurring opinion.

concurring.

I join the majority opinion. I write separately to set out additional reasons for overruling Bauder1 And, as the author of the opinion in Ex parte Wheeler;21 also fall upon my sword.

*372Swanda Lewis was charged with murdering her husband. During the trial, the prosecutor asked her, on three separate occasions, whether she had told the 911 operator, the crime scene officer, or the detective to whom she had given a post-arrest statement, anything about her trial-time testimony that her husband had raped her immediately before she killed him. Each time the defense objected, stating that these questions improperly commented on her post-arrest silence.3 After the third such question, the trial court granted a defense-requested mistrial. Ms. Lewis then filed an application for a writ of habeas corpus, arguing that retrial was barred by the Texas constitutional double jeopardy provision under this Court’ decision in Ex parte Bauder. The trial court denied relief, but the court of appeals held that the trial court abused its discretion in failing to find that the prosecutor’ questions were manifestly improper and were asked “with conscious disregard for a substantial risk that the trial court would be required to declare a mistrial.”4 The court of appeals held that the State was prohibited from retrying Ms. Lewis. We granted review to decide, inter alia, whether to reconsider our decision in Bau-der. I join the Court in overruling Bau-der.

First, however, I agree with the majority that the Texas constitutional double-jeopardy provision does bar retrial when a mistrial is neither a manifest necessity nor consented to or requested by the defendant. Although the State and several members of this Court have argued that the double-jeopardy provision of the Texas Constitution provides no protection in this context,5 the Presiding Judge’s historical inquiry is precisely the type of analysis necessary to determine the independent content of the Texas Constitution. And I agree with her conclusion: Although the Texas Constitution does not explicitly address the double-jeopardy consequences of a mistrial, Texas courts have, for well over 100 years, assumed (and held) that retrial is barred after jeopardy has attached if the jury is discharged without a manifest necessity unless the defendant consents.6

The problem with Bauder is that it undertook no historical analysis of the Texas Constitution. It did not look to the framers’ intent, uniquely Texan social, political, legal, and jurisprudential developments throughout the late nineteenth and twentieth centuries, or any other factors in con-*373eluding that the Texas double-jeopardy provision provides a thicker shield than the corresponding federal provision.7

In Bauder, a bare majority of this Court ignored the United States Supreme Court’s lengthy historical analysis of the federal double-jeopardy provision in its Oregon v. Kennedy8 decision. The Bauder majority did not find any inaccuracies in the Supreme Court’s recitation of the common-law historical development of the Anglo-American double-jeopardy protection. It did not point to any Texas deviations from that hoary lineage.9 Quite the contrary, it expressly stated:

The Texas Double Jeopardy Clause, like its federal counterpart, is meant to restrain the government from subjecting persons accused of crimes to the mental, emotional, and financial hardship of repeated trials for the same offense.10

If the purpose is the same, then why would the meaning and application of our provision differ? What special and unique factors exist in Texas history or even in its current social and legal milieu that would call for a different constitutional rule under the same conditions and fact patterns as those set out by the Supreme Court? In Bauder, this Court pointed to nothing uniquely Texan. Instead, this Court’s answer to the Supreme Court’s decision in Kennedy was: We think the highest court in the land is wrong.11 It was not that the *374Bauder court concluded that the Texas and federal double-jeopardy provisions embodied different purposes or enjoyed a separate historical lineage, it was simply that this Court disagreed with the specific bright-line that the Supreme Court drew. So it drew its own double-jeopardy fine in the sand and called it the Texas constitutional fine.

Indeed, this Court does have the authority to draw such constitutional lines. It is not the power to draw new constitutional lines that is at issue, it is the “oughtness” of such an endeavor. At the Alamo, Colonel Travis drew a line in the sand, and his men had a choice — step over it or stand pat. When this Court draws new lines in the constitutional sand, the citizens of Texas have no choice — they must step over it. Their only recourse is to have the legislature propose a constitutional amendment for the citizens’ approval at a statewide election to erase that new line upon which they were never consulted.

One member of the U.S. Supreme Court, Justice Jackson, famously said, “We are not final because we are infallible, but we are infallible only because we are final.” 12 Perhaps because that Court recognizes that its word on constitutional issues is the final one in America, it is appropriately deferential to the other branches of government, recognizing that “[w]e sit as judges, not as legislators[.]”13 The Supreme Court may, at times, intentionally draw its constitutional lines parsimoniously to leave the citizens of the various states free to make their own choices on whether to draw a different line.14 Justice Bran*375deis, noting that the Supreme Court sets the floor on constitutional rights and obligations, urged local experimentation with social and economic reforms: “It is one of the happy incidents of the federal system that a single courageous State, may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country.”15 Justice Brandeis’s emphasis was upon the right of the citizens to choose, not the power of state judges to choose for them.

In Kennedy, the Supreme Court recognized that its prior precedent lacked “crystal clarity” and could be read to grant broader protection in the mistrial context than did the “bright-line” rule it adopted in Kennedy itself.16 It adopted the rule that double jeopardy would bar a second trial after a mistrial “[o]nly where the governmental conduct in question is intended to ‘goad’ the defendant into moving for a mistrial” precisely because its prior pronouncements (and the rule sought by Justice Stevens in his concurring opinion) were essentially “standardless” and thus would offer little practical guidance to the lower courts.17

Indeed, that has been precisely the problem with Bauder. It followed the rationale of Justice Stevens’ concurring opinion which the Kennedy majority criticized as “amorphous,” 18 and has led to precisely the type of judicial clarifications and re-*376clarifications that the Supreme Court predicted would occur without a clear, simple “black letter law.”19 The Kennedy rule may not be Platonically perfect, but it does have the virtues of clarity and simplicity. It provides notice and certainty to the bench and bar on how to apply that rule. It ensures that similarly-situated defendants will be treated the same. It constrains judicial discretion. It avoids complicated tests and idiosyncratic judicial balancing factors.

Because Texas courts have dealt with Bauder for ten years and this Court has clarified and reclarified it several times,20 stare decisis might counsel against jettisoning it now. It is said, in defense of stare decisis, that sometimes “it is better to be consistent than right.”21 But Ban-der is neither consistent nor right. Stare decisis most frequently does and should apply to “black letter law,” the type of holding that has, among its main virtues, historical familiarity, ease of application, long-standing reliance, predictability, and precision.

As the Supreme Court stated in Payne v. Tennessee, “Stare decisis is not an inexorable command,” rather “it is a principle of policy.”22 That policy is least applicable in constitutional adjudication, because

“ ‘correction through legislative action is practically impossible’ ” in those cases.23 In Payne, the Supreme Court overruled its relatively recent precedent which judicially created a constitutional prohibition against the admissibility of victim impact evidence in a capital sentencing hearing.24 It noted that its original holdings in Booth v. Maryland25 and South Carolina v. Gathers 26

were decided by the narrowest of margins, over spirited dissents challenging the basic underpinnings of those decisions. They have been questioned by Members of the Supreme Court in later decisions and have defied consistent application by the lower courts.27

The same is true of Bander, and we appropriately overrule Bauder for precisely the same reasons that the Supreme Court overruled Booth and Gathers.

With these comments, I join the Court’s opinion.

dissenting in which MEYERS and HOLCOMB, JJ., joined.

I agree that Article I, Section 14 of the Texas Constitution has been, and should continue to be, read to apply in the mistrial setting. I therefore agree with Parts I through III of the Court’s opinion. I dis*377agree, however, that any of the justifications for ignoring stare decisis apply to support overruling Bander v. State.1 I therefore dissent to Part IV, and to the majority’s disposition of this case. I would reject the State’s contentions that Bauder ought to be overruled, and proceed to its third ground for review, which the majority instead dismisses, to determine whether the court of appeals properly applied Bau-der to the facts of the instant case. Because the majority does not, I ultimately dissent.

I. FEDERALISM

It has by now been established beyond serious debate that in construing provisions of our own constitution “we must ultimately follow our own lights.”2 We may construe our own constitutional provisions either more protectively than their federal constitutional counterparts, as we did in Bauder, or less protectively, as we did in Hulit v. State.3 In Bauder we construed Article I, Section 14 of the Texas Constitution4 to be more solicitous of a defendant’ “valued right to have his trial completed by a particular tribunal”5 than the Supreme Court interpreted the Fifth Amendment to afford in Oregon v. Kennedy6 Now, little more than ten years after Bauder, the majority declares our construction of Article I, Section 14 to represent an elusive “ideal,” and declares that the “real ideal” is the Supreme Court’ standard in Kennedy,7 I do not believe the majority1 belated arguments in favor of the federal “ideal” are sufficiently compelling to justify disregarding stare decisis.8

As the majority develops in Part III of its opinion, this Court’s predecessor, the Texas Court of Appeals, long ago held that the concept of jeopardy, as embodied in what is now Article I, Section 14 of the Texas Constitution, is broad enough to protect a defendant’s right to proceed to a final verdict, once the jury has been impaneled and sworn.9 It did so long before the United States Supreme Court first expressly recognized, in 1949, that a criminal defendant has a “valued right” under the Fifth Amendment “to have his trial completed by a particular tribunal.”10 But *378once the Fifth Amendment’ double jeopardy protection was found to be applicable to the states by incorporation through the Fourteenth Amendment, in 1969,11 this Court, like many state courts across the country, allowed the Supreme Court to take the lead in developing double jeopardy doctrine. Thus, in the 1970s and 1980s, until Bander came along, we were content to follow the dictates of federal precedent, including Oregon v. Kennedy, as a matter of federal jeopardy law, without addressing the independent question of whether we ought to construe our analogous state constitutional provision any differently.12 Federal jeopardy law had become the template for decision. Bander required us to decide whether, in applying our own jeopardy provision in a case of first impression, we would conform to the federal template or “follow our own lights.”

The majority acknowledges that an historical analysis of our own jeopardy provision would not shed any light on the question, either now or at the time Bander was decided.13 The question whether double jeopardy protection is implicated by prose-cutorial misconduct that provokes a defendant to request a mistrial arose during the interim when state courts were deferring to the federal template. Cases like United States v. Jom,14 United States v. Dinitz,15 and Arizona v. Washington,16 which formed the doctrinal backdrop for Oregon v. Kennedy, were all decided during this interim. It was a new gloss on double jeopardy law. The question in Kennedy was how far to go with that new gloss under the Fifth Amendment, consistent with the principles of double jeopardy. And squarely presented to this Court for the first time in Bander was whether to adopt that gloss at all under Article I, Section 14, and, if so, how far to take it consistently with our own understanding of the proper scope of double jeopardy protections. That is all the majority has done today. It simply disagrees with the majority that decided Bander.

II. PROSECUTORIAL MISCONDUCT

The criminal defendant’s right to have his trial completed by the tribunal originally selected to decide his fate, once jeopardy has attached, has never been regarded as absolute. Indeed, when first recognized by the Supreme Court in Wade v. Hunter, that right was immediately balanced against “the public’s interest in fair trials designed to end in just judgments.”17 This is the reason that “manifest necessity” will justify a mistrial, even in the event of ordinary judicial or prosecutorial error. *379As Justice Stevens explained for the Supreme Court in Arizona v. Washington:

Because of the variety of circumstances that may make it necessary to discharge a jury before a trial is concluded, and because those circumstances do not invariably create unfairness to the accused, his valued right to have the trial concluded by a particular tribunal is sometimes subordinate to the public interest in affording the prosecutor one full and fair opportunity to present his evidence to an impartial jury. Yet in view of the importance of that right, and the fact that it is frustrated by any mistrial, the prosecutor must shoulder the burden of justifying the mistrial if he is to avoid the double jeopardy bar. His burden is a heavy one. The prosecutor must demonstrate “manifest necessity” for any mistrial declared over the objection of the defendant.18

Courts need not inquire about manifest necessity for a mistrial, however, when the defendant himself asks for or consents to it. The defendant’s consent signals his election to forego his right to proceed to verdict with the first tribunal even though prejudicial error may have been injected into the proceedings, and there is no jeopardy bar to reprosecution.19

The defendant may reasonably conclude that a continuation of the tainted proceeding would result in a conviction followed by a lengthy appeal and, if a reversal is secured, by a second prosecution. In such circumstances, a defendant’s mistrial request has objectives not unlike the interests served by the Double Jeopardy Clause — the avoidance of the anxiety, expense, and delay occasioned by multiple prosecutions.20

The choice must be the defendant’s, to-wit: whether he believes his interest in avoiding anxiety, expense, and delay is better served by proceeding to verdict with the first tribunal, and possible acquittal, or by cutting the first prosecution short in order to proceed more expeditiously with a second. And in this context, “[t]he important consideration, for purposes of the Double Jeopardy Clause, is that the defendant retain primary control over” that decision.21

At some point, of course, ordinary prose-cutorial error may rise to the level of prosecutorial misconduct. And, at some point, prosecutorial misconduct may become so egregious that it cannot fairly be said that the defendant has retained primary control over the decision whether to proceed to verdict or abort the proceedings. The question for decision in Oregon v. Kennedy was how to identify prosecuto-rial misconduct that is so manipulative that it deprives the defendant of a “meaningful” choice of which option best protects his interest in avoiding as much as possible the anxiety, expense, and delay inherent in criminal prosecution.22 In making this determination, a court must be mindful of the competing values that are at stake — on the one hand, the defendant’s valued right to proceed to verdict with the first tribunal, and on the other, the State’s equally valued right to “one full and fair opportunity to present [its] evidence to an impartial *380jury.”23 At what point does prosecutorial misconduct cause the scale to tip in favor of the defendant’s right, notwithstanding that it was he who requested the mistrial?

In Oregon v. Kennedy, the Supreme Court answered this query for Fifth Amendment purposes by holding that a criminal defendant loses primary control over the critical choice whether to proceed to verdict or abort only “where the governmental conduct in question is intended to ‘goad’ the defendant into moving for a mistrial[,]” and that only then “may a defendant raise the bar of double jeopardy to a second trial after having succeeded in aborting the first on his own motion.”24 But why should this be so? It is not self-evident that the specific intent to provoke a mistrial request should be the only degree of prosecutorial culpability sufficient to reinvigorate the defendant’s valued right to proceed to verdict with the first tribunal despite the fact that he asked for the mistrial. The State is entitled to one full and fair opportunity to present its evidence to an impartial jury. It can reasonably be argued that other, lesser degrees of prosecutorial culpability should suffice to justify the conclusion that the State has abused, and therefore forfeited, that opportunity. When the prosecutor intentionally commits misconduct he knows will seriously compromise the fairness of the trial, he has arguably squandered his one full and fair opportunity to present his case, so that the State’s interest can no longer be said to outweigh the defendant’s — even if he did not harbor a specific intent to provoke a mistrial. If his intention was to inject manifest unfairness into the proceeding, and he was consciously indifferent with respect to whether this intentional misconduct illegitimately increased his chances of gaining a conviction or provoked the defendant into asking for a mistrial, the argument is practically as compelling that he has forfeited his one full and fair opportunity to present his case as when it was his specific intent to provoke a mistrial. Either way, a reasonable argument can be made that the prosecutor has manipulated the defendant’s choice to such an extent that it is no longer primarily the defendant’s, and the State can no longer show that its interest outweighs the defendant’s in the constitutional balance.

This is, in essence, what we held in Bauder,25 It was not an unreasonable or an outlandish holding. It was certainly not a unique holding.26 It may even have been the more logical holding, given the constitutional principles involved.27 In any event, it was not such a manifestly erroneous holding that we can justify overruling it just because there is a majority of the Court presently willing to do so. I next *381address what I understand to be the majority’s asserted reasons for doing so.

III. DISREGARDING STARE DECISIS

As I break down the majority’s opinion, it has identified essentially six flaws in the Bauder analysis that, taken together, are serious enough to justify its demise. I do not share any of those concerns. I will respond to them in turn.

A. Only a Specific Intent to Provoke Mistrial Triggers Jeopardy Concerns

Like the Supreme Court in Oregon v. Kennedy, the majority believes that a specific prosecutorial intent to provoke a mistrial “is critical to determining whether [the prosecutor], rather than the defendant, has exercised primary control over whether a mistrial is sought.”28 For reasons I have already explained, a specific intent to provoke mistrial is not the only degree of prosecutorial culpability that can reasonably be found to trigger jeopardy protection. The majority does not explain why a slightly lesser degree of prosecutorial culpability could not also justify a court in concluding that the prosecutor has forfeited his one full and fair opportunity to present his case, thus tipping the balance in the defendant’s favor on the double jeopardy scale. In Bauder we essentially held that a prosecutor who consciously disregards a known risk that his deliberate misconduct will provoke a mistrial, even if a mistrial was not necessarily his conscious objective, nevertheless harbors a degree of culpability sufficient to forfeit his one full and fair opportunity.29 The majority does not convince me that this cannot reasonably be regarded as a sufficient level of culpability to justify jeopardy protection.

The majority argues that our earliest case law deemed the premature termination of a trial to implicate jeopardy concerns only because it was the functional equivalent of an acquittal,30 and a prosecutor whose conscious objective is not necessarily to provoke a mistrial cannot be said to have sought the functional equivalent of an acquittal.31 But this is not self-evident to me either. A prosecutor who has consciously disregarded a substantial risk that his deliberate misconduct would provoke a mistrial has also consciously disregarded the substantial risk his conduct would cause the functional equivalent of an acquittal, even if that was not necessarily his conscious objective. To reiterate: It is not unreasonable for a court to conclude, as we did in Bauder, that such a prosecutor has abused, and therefore forfeited, his one full and fair opportunity at a verdict favorable to the State.

B. Bauder Is Due Process in Double Jeopardy Clothing

The majority claims that, by focusing on the fact that a slightly lesser degree of prosecutorial culpability than specific intent may also compromise the defendant’s right to a fair trial before the first tribunal selected, the Court in Bauder “conflates the double jeopardy protection with more generalized notions of due process and due *382course of law.”32 Elsewhere it has been suggested that to apply a lesser standard of culpability than specific intent to justify jeopardy relief is merely a means for courts to punish prosecutors, rather than to scrupulously serve double jeopardy principles.33 I disagree on both counts.

Both the defendant and the State are entitled to one full and fair opportunity for trial. Ordinarily, double jeopardy entitles the defendant to proceed to verdict with the first tribunal selected. Manifest necessity or the defendant’s own consent may suffice to defeat his constitutional interest, but not otherwise. This means that sometimes the defendant must experience the anxiety, expense, and delay of a second trial even when his first trial was rendered unfair for reasons unattributable to him. But he should not necessarily have to suffer that consequence when the retrial was attributable to deliberate misconduct on the part of the prosecutor. When that misconduct so compromised the fairness of trial as to render mistrial inevitable, and the prosecutor was at least consciously indifferent to that result, the State may reasonably be said to have abused its one full and fair opportunity to present its evidence to an impartial tribunal, and it can no longer carry its burden to demonstrate that its interest in the jeopardy balance outweighs the defendant’s, even when it was the defendant who requested a mistrial. Thus, jeopardy principles are vindicated. It is true that the fairness that due process and due course of law guarantee may also be vindicated and that the prosecutor may feel he is being made to pay a heavy price for his misconduct. But these consequences are incidental to, and do not by any means displace, the jeopardy analysis.

C. Later Case Law Is Inconsistent with Bauder

The majority complains that if we were correct in Bauder to find a lesser prosecu-torial culpability to be sufficient to trigger jeopardy protection in the context in which the trial court grants a defendant’s motion for mistrial, we should also have held, as other jurisdictions have done, that even when the trial court erroneously denies the mistrial, we should bar retrial after the defendant successfully challenges his conviction on appeal.34 In Ex parte Davis35 and Ex parte Mitchell,36 we expressly declined to do so. We held that the defendant had not been deprived of his valued right to proceed to verdict with the first tribunal, since his trial was prosecuted to a conclusion, albeit a conviction. This may well represent a logical inconsistency in our case law.37 But if so, it applies with equal illogic whether the prosecutor consciously intended to cause a mistrial or he consciously disregarded a substantial risk his misconduct would provoke a mistrial.38 *383Thus, to the extent these cases are logically unfaithful to Bauder, they are equally unfaithful to the Oregon v. Kennedy standard. It seems to me that if any of our precedent deserves closer scrutiny, it would be Davis and Mitchell, not Bauder.

D. The Bauder Standard Is Too Amorphous

I do not disagree with the majority that mistrials that result in a jeopardy bar ought to be relatively rare occurrences.39 The Bauder standard, particularly as it was later elaborated in Ex parte Peterson, 40 should produce jeopardy bars only in a very slightly greater number of cases than the Oregon v. Kennedy standard. The majority complains that some of the language in Bauder seemed to blur the concept of recklessness, creating a danger that it might be understood as something akin to the prosecutorial “overreaching” standard that the Supreme Court flatly rejected in Oregon v. Kennedy.41 Again, whatever imprecision may have inhered in the specific language of Bauder itself, and whatever confusion it may have engendered, was later remedied in our opinion in Peterson. I do not think any reasonable prosecutor would now mistake the standard explicitly set out in Peterson as a rule imposing a jeopardy bar on the basis of mere prosecutorial “overreaching.” It makes little sense to discard a legal standard barely three years after, in the ordinary evolution of our decisional law, we have finally perfected it.

I cannot agree with the majority that, even as perfected, the standard articulated in Peterson will not adequately equip prosecutors to be able to tell that deliberate misconduct sufficient to trigger double jeopardy protection from that deliberate misconduct which is not.42 A prosecutor knows he should never engage in deliberate misconduct — that is to say, it should never be his design and conscious objective to commit what he recognizes to be misconduct. Under Bauder/Peterson, if he does so with the conscious objective to provoke a mistrial, or with a conscious disregard for the substantial risk it would cause a mistrial, double jeopardy protections will apply. Either state of mind can be established (or refuted) by an express assertion from the prosecutor, or by evidence of the circumstances surrounding his misconduct such as those that we enumerated in Peterson.43 The degree of risk that the prosecutor’s deliberate misconduct will result in a mistrial is largely a function of the amount of unfair prejudice it injects into the proceedings. The more unfair prejudice his deliberate misconduct injects, the more compelling will be the inference that he was aware of the risk *384(because it was, objectively speaking, more obvious), and consciously disregarded it.44 This standard is not so amorphous that it cannot fairly be imposed upon responsible prosecutors.

E. Trial Court Will Stop Granting Mistrials

The majority fears that the overly “broad” standard in Bauder will cause trial courts to unduly hesitate to grant meritorious mistrials on account of the jeopardy consequences.45 I do not believe the Bau-der standard, especially as solidified in Peterson, is so much broader than that of Oregon v. Kennedy that it will substantially increase whatever public or media pressure trial courts feel to eschew meritorious mistrials. In any event, I do not share the majority’s (and the Kennedy Court’s) cynical distrust of trial judges to follow their oaths to uphold the law.46 A conscientious trial court should always exercise its discretion to grant any motion for mistrial that it fairly judges to be meritorious, regardless of the potential for later double jeopardy consequences. The presumption that we should institutionalize in our case law is that a trial court will do so, not that it will not.47

The majority observes that this tendency it perceives in trial courts to want to avoid mistrials could have been “ameliorate[d].”48 We could have held, in Davis and Mitchell, that double jeopardy protections also apply when a conviction is reversed on appeal due to the failure of the trial court to grant a meritorious mistrial following misconduct that the prosecutor committed with conscious disregard of the substantial risk that it would cause a mistrial. Such a holding would have created at least some disincentive for trial courts to avoid granting meritorious mistrials.49 If I shared the majority’s premise that we cannot trust trial courts to follow the law, this observation would cause me to question the correctness of our decisions in Davis and Mitchell, not the correctness of our decisions in Bauder and Peterson.

F. Bauder Has Proven Unworkable

The majority documents at some length the “messy jurisprudence flowing from Bauder.”50 It is true that, since Bauder, *385the Court has not always been entirely consistent in its articulation of the standard.51 But a certain amount of fine-tuning is inevitable in the evolution of deci-sional law.52 I do not understand the majority to hold that our most recent clarification was a failure — in fact, quite the opposite, since it is at least acknowledged that what Bander meant by “recklessness” was firmly nailed down in Ex parte Peterson.53 I cannot agree that a case barely three years old CPeterson) ought to be overruled on the ground that it has proven unworkable. We had no problems applying it in our recent decision in Ex parte Wheeler,54 In that case we overruled the court of appeals and reinstated the trial court’ ruling, which had denied double jeopardy relief. The only real point of contention between this Court and the court of appeals was whether the court of appeals had erred in failing to give proper deference to the trial court’s application of the Bauder/Pe-terson standard.55 There was no confusion evident at any level of the proceedings as to the substance of that standard. At this point, only time can tell whether further “refinement” will be necessary.56 I cannot fathom the majority’s haste and determination to dispatch a standard so soon after we have managed, at least with apparent success, to work the kinks out of it.

IV. APPLICATION OF BAUDER/PETERSON

We originally remanded this cause to the court of appeals for its reconsideration in light of Peterson, which had come down after the court of appeals’ original opinion.57 Thus, the court of appeals was the first to conduct an analysis under the refined Peterson standard. The court of appeals meticulously applied the Peterson factors to conclude that our state constitutional jeopardy protection applied to bar retrial. In its petition and its brief on the merits, the State disagrees with the court of appeals’ conclusion, but does not identify any substantial flaw in its application of the law as expounded in Peterson. I do not find any, and would therefore affirm its judgment.

CONCLUSION

The majority’s bottom-line seems to be that the Bauder/Peterson standard is too broad, and not adequately tethered to legitimate double jeopardy principles.58 But, as I have endeavored to develop in Part II of this opinion, the doctrinal basis for that standard is more than evident enough, and it embodies an eminently reasonable construction of Article I, Section 14, our state constitutional jeopardy provision. I therefore dissent to Part TV of the Court’s opinion and to its disposition of the case. I would proceed to the State’s third ground for review and hold that the court *386of appeals did not err in its application of the Bauder/Peterson standard. Because the Court’s disposition instead moots that inquiry, I respectfully dissent.