State v. ToneyAnnotate this Case
979 S.W.2d 642 (1998)
The STATE of Texas, v. Mark Anthony TONEY, Appellant.
Court of Criminal Appeals of Texas, En Banc.
October 7, 1998.
*643 Kurt B. Wentz, Houston, for appellant.
Alan Curry, Asst. Dist. Atty., Houston, Matthew Paul, State's Atty., Austin, for State.
Before the court en banc.OPINION ON STATE'S PETITION FOR DISCRETIONARY REVIEW
PRICE, Judge, delivered the opinion of the Court, in which BAIRD, OVERSTREET, MEYERS and MANSFIELD, Judges, joined.
Appellant was convicted of delivery of a controlled substance. Upon pleading "true" to two enhancement paragraphs, he was sentenced by the jury to 40 years imprisonment. On appeal, his fourth point of error alleged that the trial court committed reversible error when it failed to include a complete definition of "reasonable doubt" in the jury instruction in accordance with Geesa v. State, 820 S.W.2d 154 (Tex.Crim.App.1991). The Court of Appeals agreed and reversed the conviction, holding that under Reyes v. State, 938 S.W.2d 718 (Tex.Crim.App.1996), the trial judge's omission of the required definitional jury instruction mandated automatic reversal without a harm analysis. Toney v. State, 942 S.W.2d 750, 751-752 (Tex.App. Houston [14th Dist.] 1997, pet. granted). We granted the State's petition for discretionary review to determine whether the Fourteenth Court of Appeals erred in this holding.
In Geesa v. State, 820 S.W.2d 154 (Tex. Crim.App.1991), this court abrogated the "reasonable hypothesis of innocence" analytical construct. Id. at 160-161. However, in doing so, we determined that such an abrogation required the adoption of an instruction on reasonable doubt. Id. at 161-162. We reasoned that the requirement of a definitional instruction was implicated by the Supreme Court decisions of Holland v. United States, 348 U.S. 121, 75 S. Ct. 127, 99 L. Ed. 150 (1954) and Jackson v. Virginia, 443 U.S. 307, 99 S. Ct. 2781, 61 L. Ed. 2d 560 (1979). Geesa, 820 S.W.2d at 161. Therefore, we set out the following instruction:All persons are presumed to be innocent and no person may be convicted of an offense unless each element of the offense is proved beyond a reasonable doubt. The fact that a person has been arrested, confined, or indicted for, or otherwise charged with, the offense gives rise to no inference of guilt at his trial. The law does not require a defendant to prove his innocence or produce any evidence at all. The presumption of innocence alone is sufficient to acquit the defendant, unless the jurors are satisfied beyond a reasonable doubt of the defendant's guilt after careful and impartial consideration of all the evidence in the case. The prosecution has the burden of proving the defendant guilty and it must do so by proving each and every element of the offense charged beyond a reasonable doubt and if it fails to do so, you must acquit the defendant. It is not required that the prosecution prove guilt beyond all possible doubt; it is required that the prosecution's proof excludes all "reasonable doubt" concerning the defendant's guilt. A "reasonable doubt" is a doubt based on reason and common sense after a careful and impartial consideration of all the evidence in the case. It is the kind of doubt that would make a reasonable person hesitate to act in the most important of his own affairs. Proof beyond a reasonable doubt, therefore, must be proof of such a convincing *644 character that you would be willing to rely and act upon it without hesitation in the most important of your own affairs. In the event you have a reasonable doubt as to the defendant's guilt after considering all the evidence before you, and these instructions, you will acquit him and say by your verdict "Not guilty".
Id. at 162. We ultimately held that this instruction should be submitted to juries in all criminal cases, regardless of whether a request or an objection is made by either party, or whether the evidence is circumstantial or direct. Id.
Consequently, when the trial judge in Reyes v. State, 938 S.W.2d 718 (Tex.Crim. App.1996) totally omitted the reasonable doubt instruction from the charge at the guilt/innocence phase of trial, we held that the reasonable doubt instruction set out in Geesa created an absolute systemic requirement that could be neither waived nor forfeited. Id. at 720-721. We further held that such a requirement was not subject to a harm analysis under Rule 81(b)(2). Id. at 721.
Nonetheless, only a year later in Cain v. State, 947 S.W.2d 262 (Tex.Crim.App.1997), the subject of a meaningful harm analysis arose within the context of a trial court's failure to admonish a defendant concerning the deportation consequences of a plea in accordance with TEX.CODE CRIM. PROC. ANN. art. 26.13(a)(4) (Vernon 1989). Id. at 263. There we broadly held that:Except for certain federal constitutional errors labeled by the United States Supreme Court as "structural," no error, whether it relates to jurisdiction, voluntariness of a plea, or any other mandatory requirement, is categorically immune to a harmless error analysis. Of course, where the error involved defies analysis by harmless error standards or the data is insufficient to conduct a meaningful harmless error analysis, then the error will not be proven harmless beyond a reasonable doubt under Rule 81(b)(2). Hence, it may be true that some kinds of errors (particularly jurisdictional ones) will never be harmless under the Rule 81(b)(2) test and that some other kinds of errors will rarely be harmless. But, appellate courts should not automatically foreclose the application of the harmless error test to certain categories of error. Where an error is shown to be harmless, it is not a ground for reversal, regardless of the category or label attached to that particular error.
Id. at 264 (footnote omitted). Although Cain did not deal with a reasonable doubt instruction and did not cite to Reyes, its broad language concerning harmless error analysis appeared to at least somewhat limit Reyes. The limitation appeared to be at most indirect, as Cain focused on the harm analysis under TEX.R.APP. P. 81, whereas error in a jury charge is usually analyzed under TEX. CODE CRIM. PROC. ANN. art. 36.19 (Vernon 1981).
These limitations were borne out recently in our plurality opinion of Mann v. State, 964 S.W.2d 639 (Tex.Crim.App.1998). In Mann, the jury was properly instructed on reasonable doubt and the burden of proof in eight different places throughout the charge, but in one paragraph of the charge, the jurors were erroneously instructed, through the unintended omission of the terms "do not," that they could find the defendant guilty even if they had a reasonable doubt as to one element of the offense. Id. at 641 (plurality opinion). There, we distinguished Reyes, since that case involved a total omission of the instruction on reasonable doubt, and was thus incapable of harm analysis. Id. at 641-642 (plurality opinion). Yet in Mann, since the jury was properly instructed on reasonable doubt and on the burden of proof throughout the charge, except in one paragraph, we held that the error was to be analyzed under the standards set forth in Abdnor v. State, 871 S.W.2d 726 (Tex.Crim. App.1994), Almanza v. State, 686 S.W.2d 157 (Tex.Crim.App.1984), and TEX.CODE CRIM. PROC. 36.19. Mann, 964 S.W.2d at 642 (plurality opinion).
Harmonizing Geesa, Reyes, Cain and Mann then, if there is a total omission of the instruction on reasonable doubt, such error defies meaningful analysis by harmless error standards. However, if the jury is given a partial or substantially correct charge on reasonable doubt, then any error *645 therein is subject to a harm analysis under Abdnor, Almanza, and TEX.CODE CRIM. PROC. art. 36.19. In the present case, appellant received the following first three paragraphs of the Geesa instruction:All persons are presumed to be innocent and no person may be convicted of an offense unless each element of the offense is proven beyond a reasonable doubt. The fact that a person has been arrested, confined, or indicted for, or otherwise charged with, the offense gives rise to no inference of guilt at his trial. The law does not require a defendant to prove his innocence or produce any evidence at all. The presumption of innocence alone is sufficient to acquit the defendant, unless the jurors are satisfied beyond a reasonable doubt of the defendant's guilt after careful and impartial consideration of all the evidence in the case. The prosecution has the burden of proving the defendant guilty and it must do so by proving each and every element of the offense charged beyond a reasonable doubt and if it fails to do so, you must acquit the defendant. It is not required that the prosecution prove guilt beyond all possible doubt; it is required that the prosecution's proof excludes all reasonable doubt concerning the defendant's guilt.
Here, as in Mann, appellant failed to object to the partial instruction given. Therefore, we hold that appellant must show that he suffered "egregious harm" as a result of this error. See Mann, 964 S.W.2d at 641-642 (plurality opinion). Because the Court of Appeals did not have the benefit of this decision on its first hearing, we vacate the judgment of the court of appeals and remand this cause to that court for further proceedings consistent with this opinion. If the Court of Appeals determines that appellant has not suffered "egregious harm", it should then address the remaining points of error originally brought before it by appellant.
MANSFIELD, J., files a concurring opinion.
KELLER, J., files an opinion concurring in the judgment, in which McCORMICK, P.J., and HOLLAND, J., join.
WOMACK, J., files a concurring opinion in which McCORMICK, P.J., and KELLER and HOLLAND, JJ., join.
MANSFIELD, Judge, concurring.
I join the opinion of the Court. However, I am concerned that an elimination of the requirement that the jury be given the reasonable doubt instruction mandated by Geesa might breathe new life into the so-called reasonable hypothesis analytical construct eliminated by Geesa, and rendered largely meaningless by implication by Clewis v. State, 922 S.W.2d 126 (Tex.Crim.App.1996).
KELLER, Judge, concurring.
I agree with the majority opinion that the "egregious harm" standard applies to the error before us. However, instead of relying upon a plurality opinion to harmonize Reyes v. State, 938 S.W.2d 718 (Tex.Crim.App.1996) with other caselaw, I would take this opportunity to overrule Reyes.
Precedent should not be overruled lightly: often it is better to be consistent than right. Awadelkariem v. State, 974 S.W.2d 721, 725 (Tex.Crim.App.1998). But there are times when precedent must be overruled because the precedent creates inconsistency, contrary to the interests embodied by the rule of stare decisis, and/or because the interests of justice outweigh the benefits of consistency. Id. at 724-25; see also Malik v. State, 953 S.W.2d 234, 236-240 (Tex.Crim.App.1997). And, the interests underlying stare decisis are less weighty for judicially-made rules than for interpretations of statutory enactments. Awadelkariem, at 726. Some factors militating in favor of overruling precedent are: (1) when the original decision was flawed from the outset, due to flawed reasoning, lack of authority, or misplaced reliance upon cited authorities, (2) when the precedent flowing from the decision conflicts with other, newer precedent, (3) when the justifications proffered for the rule established by the decision have been undercut with the passage of time, (4) when the rule produces inconsistency and confusion in the law, (5) when the rule consistently creates unjust results or places unnecessary constraints or *646 burdens upon the system, and (6) when the rule creates differences between criminal and civil practice when no principled reason exists for the disparity. Awadelkariem, 724-25 and 726; see also Malik, 953 S.W.2d at 236-240. I find that the first, second, fourth, and fifth factors are present in the case at bar and are weighty enough to compel overruling Reyes.1. Reyes was a flawed decision.
In Reyes, this Court held the instruction mandated by Geesa v. State, 820 S.W.2d 154 (Tex.Crim.App.1991) to be an "absolute systemic requirement that cannot be waived or forfeited." 938 S.W.2d at 721. The Court concluded that "[s]uch requirements are not subject to a harm analysis." Id. In support of that conclusion the Court cited Sodipo v. State, 815 S.W.2d 551 (Tex.Crim.App.1990), Powell v. State, 897 S.W.2d 307 (Tex.Crim. App.1994), and Stine v. State, 908 S.W.2d 429 (Tex.Crim.App.1995)(plurality opinion), and the Court pointed out that it did not engage in a harm analysis in Geesa. Reyes, 938 S.W.2d at 721. Reyes was a bare 5-4 majority opinion with only half-hearted support from the fifth voteJudge Meyers. Although he characterized the holding as "troubling," he felt compelled to join the majority opinion. 938 S.W.2d at 722 (Meyers, J. concurring). Judge Mansfield dissented from the Court's conclusion that the failure to include a Geesa instruction is per se reversible error. Id. at 722 (Mansfield, J. dissenting). This writer also dissented, complaining that the majority had erroneously equated non-waivable error with automatic reversible error and that the two were not synonymous. Id. at 724.
I remain firmly convinced that Reyes erroneously equated two different concepts: preservation of error and harm. That a particular kind of error may be non-waivable does not mean that the error is necessarily harmful. The Court gave no reason in Reyes for equating the two, except for its citation to Sodipo, Powell, and Stine and its remark that no harm analysis was performed in Geesa. Hence, an examination of those cases is in order.
In Sodipo, the State moved to amend the indictment on the date of trial before trial began. 815 S.W.2d at 555 (opinion on rehearing). Defense counsel objected to the amendment and requested ten days to prepare. Id. On original submission, we held that the ten day preparation requirement found in Article 28.10(a) was not subject to a harm analysis "because the record will not reveal any concrete data from which an appellate court can meaningfully gauge or quantify the effect of the error." Id. at 554. We cautioned, however, that "[w]e do not decide today whether all `mandatory' statutes are immune to a harmless error analysis." Id. On rehearing, we held that our analysis of the ten day preparation requirement was "improvident" and proceeded instead to analyze whether there was a violation of Article 28.10(b). Id. at 555. We found that amendments on the date of trial violated subsection (b), and we held that the error was not subject to a harm analysis. Id. at 556. We found that a harm analysis should not be conducted because conducting such an analysis would prevent giving effect to the full meaning and intent of an unambiguous statute. Id.
But Sodipo cannot support Reyes' holding that absolute, systemic requirements are immune from harm analyses because Sodipo did not involve an absolute, systemic requirement. The statute at issue in Sodipo established forfeitable rights: a defendant was entitled to ten days preparation only upon request and an indictment could not be amended on the date of trial in the face of a defendant's objection. 815 S.W.2d at 553 and 556. Neither the opinion on original submission *647 nor the opinion on rehearing contains any language that exempts an error from a harm analysis on the ground that the error is an "absolute, systemic requirement."
Reyes did not utilize either of the rationales actually contained in Sodipo. Any attempt to do so, however, would have been misguided. The "concrete data" approach taken in the opinion on original submission does not automatically render Geesa error immune from a harm analysis. Even where a Geesa instruction is wholly omitted, the record in a case may contain "concrete data" from which a meaningful harm analysis could be conducted. Geesa offered defendants a quid pro quo: instead of requiring the reasonable hypothesis charge for circumstantial evidence cases, this Court formulated a definition of reasonable doubt to be uniformly administered. 820 S.W.2d at 161 and 165. Hence, defendants whose convictions are supported by direct evidence would not be harmed by a failure to obtain a Geesa instruction. Reyes, 938 S.W.2d at 724 (Keller, J. dissenting). Moreover, other developments during trial may provide concrete data from which to determine whether the absence of a Geesa instruction was harmful. For instance, in Reyes, the prosecutor, during argument to the jury, defined reasonable doubt in a manner closely tracking the definition mandated by Geesa. Reyes, 938 S.W.2d at 724 (Mansfield, J. dissenting). The record contained concrete evidence that the defendant was not harmed because the jury was in fact given correct information concerning the definition of reasonable doubt.
The "clear language of a mandatory statute" reasoning found in Sodipo's opinion on rehearing also does not justify holding a harm analysis to be inapplicable to Geesa errors. The Geesa instruction is not a statute at all. There is no Legislative intent that we must attempt to effectuate. Nor is there any constitutional imperative requiring the instruction. See Geesa, 820 S.W.2d at 163 (The rules are not of constitutional dimension per se; rather, the rules serve to implement the constitutional requirement that a criminal conviction cannot stand "except upon proof beyond a reasonable doubt"); Reyes, 938 S.W.2d at 723 n. 1 (Mansfield, J. dissenting)(citing Victor v. Nebraska, 511 U.S. 1, 5, 114 S. Ct. 1239, 127 L. Ed. 2d 583 (1994)). The Geesa definition is simply a court-made rule, and there is no indication in the language of Sodipo that the opinion was ever meant to apply to such rules.
Further, even if one engaged in the strained analogy that a mandatory court-made rule should be treated in the same fashion as a mandatory statute, the "clear, mandatory statute" rationale in Sodipowas shaky ground for immunizing a court-made instruction from a harm analysis. Before Reyes was decided, we had explained that the applicability of a harm analysis was not contingent upon whether the statute was mandatory, but whether the record will contain concrete data to enable conducting a harm analysis. Warmowski, 853 S.W.2d at 577. In other words, we indicated that Sodipo's opinion on original submission, not the opinion on rehearing, contained the correct test for determining when a harm analysis should be conducted. We had further explained that determining whether a harm analysis should apply to a violation of a mandatory statute should be a case-by-case inquiry. Adanandus v. State, 866 S.W.2d 210, 219 (Tex.Crim.App.1993). By categorically holding that Geesa error is not subject to a harm analysis, Reyes conflicted with that case-by-case approach.
In Powell, we held that a party could not waive the "deliberateness" special issue, required under the death penalty scheme applicable to the defendant. 897 S.W.2d at 315-319. However, Powell did not hold that all systemic requirements were immune from a harm analysis. At most, Powell held that the particular systemic requirement at issue was immune because "the verdict [was] incomplete" and an affirmative answer to the issue was required to authorize the death penalty. Id. at 317-318 (plurality opinion); id at 318 (Clinton, J. concurring). Further, another interpretation of Powell's holding is equally viable: because the verdict was incomplete, the error was necessarily harmful. Rather than being immune from a harm analysis, the error in Powell simply failed the harm analysis. See id. at 318 ("The sentence of death is likewise tainted" (emphasis added)). Of *648 course, given the nature of the specific kind of error in Powell, that error will probably always be harmful. That is different, however, from saying that the error is immune from harm analysis.
In Stine,a portion of the trial proceedings had occurred outside the county seat. 908 S.W.2d at 430. This Court held that no harm analysis need be done because the Texas constitutional provision at issue was "clear and unambiguous," "mandatory," and "non-waivable." Id. at 431. In support of this proposition, the Court cited Sodipo. Stine, 908 S.W.2d at 431. But, as discussed above, that a provision is clear and mandatory did not necessarily render violations immune from a harm analysis under Sodipo. And, as also discussed above, Sodipo did not hold "non-waivable" errors immune from a review for harm. Stine's reliance upon Sodipo was misplaced.
Finally, Reyes's reliance upon Geesa's reversal and remand is misplaced. Geesa was apparently a circumstantial evidence case, and the Court of Appeals in that case found the evidence to be insufficient under the "reasonable hypothesis" construct. Geesa, 820 S.W.2d at 162-163. Hence, the error was harmful because the case implicated the quid pro quo that the Geesa opinion set up. The defendant would have prevailed under the old sufficiency standard, so the Court was required to give him the benefit of the new jury instruction. As the above discussion shows, none of the authority cited by Reyes supports its holding.2. Reyes conflicts with more recent precedent (Cain).
In Cain v. State, 947 S.W.2d 262 (Tex. Crim.App.1997), this Court held:Except for certain federal constitutional errors labeled by the Supreme Court as "structural," no error, whether it relates to jurisdiction, voluntariness of a plea, or any other mandatory requirement, is categorically immune to a harmless error analysis.
Id. at 264 (footnote omitted). Cain's plain language shows that the Court intended to make a broad, sweeping change in harm analysis jurisprudence by including within the reach of a harm analysis all errors which we had the power to so include. The phrase "Except for certain federal constitutional errors labeled by the Supreme Court as structural" is followed by a footnote, which notes that the Supreme Court has mandated that certain types of federal constitutional errors are immune from a harm analysis. See 947 S.W.2d at 264 n. 5. The phrase and footnote is a recognition that we are bound by the pronouncements of the Supreme Court concerning federal constitutional errors. But the opinion unambiguously holds that, except where the Supreme Court has held otherwise, all errors are subject to a harm analysis. See also Atkins v. State, 951 S.W.2d 787, 797 (Tex.Crim.App.1997)(citing Cain). Given Cain's language, Reyes was effectively overruled by Cain. Geesa error is not federal constitutional error, much less error labeled by the Supreme Court as "structural." So, under the clear language of Cain, Geesa error is subject to a harm analysis, period.
The majority, however, contends that Cain only indirectly limits Reyes because Cain focused upon former Tex.R.App. P. 81(b)(2) instead of the harm analysis utilized for jury charge error under Article 36.19 and Almanza v. State, 686 S.W.2d 157 (Tex.Crim.App. 1984). A harm analysis for jury charge error does differ from the harm analyses for other errors in two respects: (1) the former flows from statute while the latter are court-made rules, and (2) the defendant has the burden to show harm while the harm analyses for other errors require a showing of harmlessness. Compare Almanza with Tex.R.App. P. 44.2(a)(formerly Rule 81(b)(2)) and 44.2(b). These differences, however, are not material to the issue at hand. Jury charge error that rises to the level of a state or federal constitutional violation is analyzed under Rule 44.2(a)(former Rule 81(b)(2)). Abdnor v. State, 871 S.W.2d 726, 732 (Tex.Crim.App. 1994). The majority cannot deny that Cain, given its sweeping language, at least applies to all errors subject to Rule 44.2(a). It would be highly anomalous to hold Cain applicable to jury charge error that rises to the level of a constitutional violation but hold the case inapplicable to errors that do not rise to that level.
*649 Moreover, Cain's holding and reasoning undermine the foundations upon which Reyes was built. As discussed above, Reyes relied upon the notion that absolute, systemic errors were automatically immune from a harm analysis. But, in announcing our holding in Cain, subjecting virtually all errors to a harm analysis, we explicitly included at least one example of an absolute, systemic requirement: jurisdiction. And the statute in question in Cain, Article 26.13, is also an absolute, systemic requirement. See Matchett v. State, 941 S.W.2d 922, 928 n. 6 (Tex.Crim.App.1996)(plurality opinion); see also Flowers v. State, 935 S.W.2d 131, 134 (Tex.Crim.App.1996).
Moreover, Cain demolishes the authority upon which Reyes relied. In announcing our holding in Cain we stated: "To the extent that Marin, Morales, Whitten, and any other decision conflicts with the present opinion, they are overruled."Cain, 947 S.W.2d at 264. Marin expressly relied upon Sodipo in holding that certain errors are not subject to harmless error review. Marin, 851 S.W.2d at 281. In overruling Marin, then, we necessarily overruled Sodipo, and by implication, Stine as well. Moreover, the reasoning in Cain is clearly at odds with the reasoning in Sodipo. While Sodipo held that an error is immune from a harmless error analysis if concrete data from which to conduct such an analysis is absent, Cain held that a harmless error test must nevertheless be conducted, and the absence of data is simply taken into account in determining whether the harmless error test is passed or failed. Cain, 947 S.W.2d at 264 ("where ... the data is insufficient to conduct a meaningful harmless error analysis, then the error will not be proven harmless ... under Rule 81(b)(2)" (ellipses inserted)).3. Retaining Reyes would promote inconsistency in the law.
Having subjected all errors to a harm analysis in Cain, except where the Supreme Court has dictated otherwise, this Court will create confusion by upholding Reyes. If the total failure to give a Geesa instruction is not subject to a harm analysis, then Cain's language becomes inaccurate, and an exception is made to the sweeping nature of Cain's holding. Courts of appeals will be left wondering whether there might be other exceptions to the literal language in Cain. Those courts will be left wondering whether a given error falls under the rubric of Cain's sweeping language or constitutes an exception, like Reyes. And finding one exception here may encourage this Court to find other exceptions later. Language in the majority opinion reinforces this confusion by relying upon the idea that an error is immune from a harm review if it "defies meaningful analysis" even though that very reasoning was rejected in Cain.4. Immunizing errors from a harm analysis will produce unjust results.
Permitting reversal without regard to harm where a Geesa instruction is omitted is likely to lead to unjust results. Reyes itself is a perfect illustration. Neither party requested a Geesa instruction or objected to its absence, but the proper definition of reasonable doubt was presented during argument to the jury. The case contained direct evidence, which would have made the evidence sufficient to support the conviction even before Geesa. The quid pro quo that motivated our imposing the instruction in Geesa was not present, the jury was at least told the proper definition, and there was no reason to believe the jury failed to utilize the proper standard of proof that the term "reasonable doubt" embodies. The State's conviction should not have been overturned in that case.
But, the unjust results are not limited to the Geesa context, but will likely occur in any context which immunizes error from a review for harm. The unjustness of immunizing errors from a harm analysis was also illustrated by the scenario found in Cain: even though a trial judge is absolutely required to inform a person of the immigration consequences of a plea, a person who is in fact a *650 United States citizen suffers no adverse consequences if he is not so informed. See Cain, 947 S.W.2d at 264. The idea that conducting a harm analysis is a waste of judicial resources was "exposed ... as plainly wrong." Matchett, 941 S.W.2d at 928 (ellipsis inserted).
Stine presents another illustration of the unjustness of precluding a harm analysis. In that case, the complaining witness was unable to appear in court because he had been hospitalized as a result of the charged assault. 908 S.W.2d at 430, 908 S.W.2d at 434 (McCormick, P.J. dissenting). With the consent of both parties, the trial court reconvened court at the hospital to take the witness' testimony, apparently in hopes that, in case the victim died, the trial would be over before the defendant could be charged with murder. Id. All other proceedings in the trial took place at the courthouse. Id. at 430-431. This Court held that error occurred because the hospital where the testimony was taken was located outside the county seat. Id. at 431. The State argued that the error was harmless, but this Court declined to conduct a harm analysis. Id. Because a mere portion of a trial was conducted in an unauthorized location, we held the entire trial void. We could have simply held that portion to be void and then considered whether the unauthorized testimony contributed to the conviction or punishment. But finding a systemic error, the Stine court blindly reversed without even considering whether that error affected the outcome of the proceedings, most of which were conducted in the proper forum.
For the reasons articulated above, Reyes should not be upheld in any form. We should recognize that Cain effectively overruled Reyes, and we should erase all doubt about Reyes' status by expressly overruling it now. Because the majority holds otherwise, I can only concur in the result.
McCORMICK, P.J., and HOLLAND, J., join.
WOMACK, Judge, concurring in which McCORMICK, Presiding Judge and KELLER and HOLLAND, Judges, join.
I agree with the Court and with Judge Keller that the omission of this definition, if error, should be analyzed for harm under the Almanza standards.
But I would undertake to "reexamine Geesa," as Judge Meyers suggested two years ago. See Reyes v. State, 938 S.W.2d 718, 721 (Tex.Cr.App.1996) (Meyers, J., concurring). Such a reexamination convinces me that Geesa should not be followed. The instruction it required was unnecessary and confusing, the justification it gave for requiring an instruction was fallacious, and it was not a legitimate exercise of the judicial function.I.
The district court's instruction in this case should be thought of as a fundamental improvement rather than a fundamental error. It was better than the one required by Geesa.
The paragraphs of the Geesa instruction that were omitted in this case combined three definitions of "reasonable doubt." The first definition said that "reasonable doubt" is "a doubt based on reason and common sense." Geesa v. State, 820 S.W.2d 154, 162 (1991). The second said that "reasonable doubt" is "the kind of doubt that would make a reasonable person hesitate to act in the most important of his own affairs." Ibid. The third definition said that "proof beyond a reasonable doubt" is "proof of such a convincing character that you would be willing to rely and act upon it without hesitation in the most important of your own affairs." Ibid. The theory of the Geesa opinion seems to be that, if one definition of reasonable doubt would be good, three definitions would be three times as good. It should be obvious that the theory is wrong.
To make things worse, two of the definitions are in opposition to each other. "The `willing to act without hesitation' definition of reasonable doubt suggests action, while the `hesitate to act' definition implies inaction.... The use of three different formulations is likely to confuse jurors, while the use of two definitions which are in opposition to each other is almost certain to mislead." Craig Hemmens et al., "Doubt over `Reasonable Doubt' in Texas," 59 Texas Bar Journal 130, *651 135 (1996). No other state or federal appellate court has combined three definitions, or combined the two opposing definitions. Ibid.
It is by no means clear that even one definition of "reasonable doubt" would be an improvement over our former practice. The predecessor of this Court decided in Black v. State, 1 Tex.App. 368 (1876), that the term did not require definition. For at least 115 years, Texas courts were not required to give any definition of reasonable doubt. "We th[ought] such definition was not necessary, same being words of commonly accepted meaning and well understood by everyone." Gallegos v. State, 152 Tex. Crim. 508, 510, 215 S.W.2d 344, 346 (1948).
The process by which a definition of "reasonable doubt" was elevated from the status of "not necessary" to that of an "absolute systemic requirement" was remarkable.II.
Geesa was convicted on circumstantial evidence. On appeal he argued that the evidence was legally insufficient. The court of appeals agreed because it found that "a rational trier of fact could not have excluded every reasonable hypothesis except Geesa's guilt." Geesa v. State, No. 2-88-140-CR (Tex.App.Fort Worth 1990) (unpublished), reproduced in Geesa v. State, 820 S.W.2d 154, 177 (Tex.Cr.App.1991) (Appendix 1). This Court rejected the use of "the reasonable-hypothesis-of-innocence analytical construct" as a method of appellate review of the legal sufficiency of evidence, and it held that a definition of reasonable doubt should be given juries in the future. Geesa v. State, 820 S.W.2d 154 (Tex.Cr.App.1991). Geesa's case was remanded to the district court for new trial because his jury was not given the definition. Id. at 163. There was no further discussion of the sufficiency of the evidence.
As an exercise of the judicial function, the opinion fails.
First, how could the case have been remanded to the trial court with Geesa's claim of legally insufficient evidence unresolved? No matter what the error in failing to instruct the jury on the definition of "reasonable doubt," a second trial would be barred by the Double Jeopardy Clause of the Fifth Amendment if the evidence were found legally insufficient on appealas the court of appeals had held it was. See Watson v. State, 605 S.W.2d 877, 880 (Tex.Cr.App.1979). Accord, e.g., Hooker v. State, 621 S.W.2d 597, 598 (Tex.Cr.App.1981). And if the evidence was legally insufficient under a proper analysis, there would be no occasion to take up the question of whether the jury should have had a different charge.
Second, no question of a jury charge was before the Court. The appellant made no claim regarding the jury charge. There is no rational relationship between defining "reasonable doubt" in the jury charge and eliminating the "the reasonable-hypothesis-of-innocence analytical construct" to determine the sufficiency of evidence. For this reason, Judge Mansfield's concern that eliminating the requirement that "reasonable doubt" be defined might bring back the former standard of appellate review (see ante) is unfoundedas far as law and logic go.
Which brings me to the final point.III.
Law and logic were not the basis of the Geesa Court's decision to require an instruction on "reasonable doubt." The Geesa Court was not functioning as a court when it promulgated the instruction; it was engaging in a form of log-rolling that was essentially legislative. A majority of the members of the Court struck a bargain with the State to abandon the "the reasonable-hypothesis-of-innocence analytical construct"in consideration of being allowed to require a definition of "reasonable doubt" in the jury charge.
In the background of this bargain was Hankins v. State, 646 S.W.2d 191 (Tex.Cr. App.1983), in which this Court held that it was improper to instruct a jury that circumstantial evidence must exclude, to a moral certainty, every reasonable hypothesis except the defendant's guilt. Three members of the Hankins Court pointed out that in other jurisdictions whose courts had made the same holding, "reasonable doubt" was defined *652 for the jury. See id. at 200 (opinion of Miller, J.); id. at 203 (opinion of Onion, P.J.); id. at 217 (opinion of Clinton, J.). These dissenters suggested that the abolition of the circumstantial evidence instruction was not valid without a definition of reasonable doubt in the charge.
Geesa came before this Court with no issue involving the jury charge. After oral arguments, the State filed a brief which proposed, in effect, that Hankins could be rewritten if the Court would agree to abandon "the reasonable-hypothesis-of-innocence analytical construct" on appellate review:We also acknowledge that dissenting judges in Hankins felt that the absence of the instruction [on circumstantial evidence] left the jurors rudderless as they sought to apply the basic reasonable-doubt standard in a circumstantial case. In light of the concerns expressed in that case, it has occurred to the State that the absence of an explanatory instruction on reasonable doubt in Texas jury charges conceivably colors the thinking of some judges regarding the State's argument here. Should that be the case, we have no objection to the Court's announcing, if the Court believes it can and should, that trial judges are henceforth, on request, to give juries the standard federal instruction describing a reasonable doubt as the kind of doubt that would make a persons hesitate to act in the most important affairs of his everyday life.
State's Post-submission Brief at 5, n. 3. The State acknowledged, "Frankly, we don't see the logical relationship between this issue and the appellate test." State's Reply to the Brief of Amicus Curiae, The Texas Criminal Defense Lawyers Association, at 15.
The State offered, not an argument of law or logic, but a quid for a quo: if the Court agreed to change the method of appellate review, an unrelated change could be made in the jury charge to answer a question that the Hankins dissents raised. Although there was no issue in the Geesa case about the jury charge, the Court accepted the bargain. It used Geesa to effectively rehear Hankins, revisiting the unrelated issue of what the jury should be told about the evidence.
Such an opinion has little claim to the respect ordinarily given to obiter dictum, much less to that accorded the reasoned resolution of the legal issues presented in an appeal.IV.
I would overrule the parts of Geesa that require this misbegotten definition of "reasonable doubt." I would hold that the omission of the definition was not error in this case. I would remand the case to the Court of Appeals for the resolution of the appellant's other points of error.NOTES
 Appellant raised three other points of error in the Court of appeals but these were dismissed as moot due to the disposition of point of error four, which was sustained.
 The precise ground for review we granted is as follows: The Fourteenth Court of Appeals erred in holding that it was automatic reversible error for the trial court to fail to instruct the jury in accordance with the judicially mandated definitional instruction on the term "reasonable doubt," without regard to whether an objection was made by the appellant, and without regard to whether the direct evidence established the appellant's guilt beyond a reasonable doubt.
 All references to articles are to the Texas Code of Criminal Procedure unless otherwise provided.
 Arguably, the opinion on original submission in Sodipo is not binding authority because it was found to have been "improvidently" issued. The language concerning the ability to "meaningfully gauge" the effect of an error does not appear in the opinion on rehearing. Nevertheless, subsequent cases have relied upon the opinion on original submission (although those cases failed to recognize its apparent lack of precedential value). See Marin v. State, 851 S.W.2d 275, 281 (Tex.Crim.App.1993); Warmowski v. State, 853 S.W.2d 575, 577 (Tex.Crim.App.1993).
 The named cases in the quotation refer to Marin v. State, 851 S.W.2d 275 (Tex.Crim.App. 1993), Morales v. State, 872 S.W.2d 753 (Tex. Crim.App.1994), and Whitten v. State, 587 S.W.2d 156 (Tex.Crim.App.1979).
 Reyes v. State, 938 S.W.2d 718, 720 (Tex.Cr. App.1996).