Court of Criminal Appeals of TexasDec 19, 2007
240 S.W.3d 882 (Tex. Crim. App. 2007)

Nos. PD-1597-05, PD-1598-05, PD-1599-05.

November 7, 2007. Rehearing Denied December 19, 2007.

Appeal from the 278th Judicial District Court, Walker County, Erwin Ernst, J.

Kelly Gatewood, Huntsville, for Appellant.

Melinda Mayo, Amarillo, Matthew Paul, State's Atty., Austin, for State.

KELLER, P.J., delivered the opinion of the Court in which KEASLER, HERVEY, HOLCOMB and COCHRAN, JJ., joined.

We must determine whether a defendant may collaterally attack a prior judgment of conviction used to enhance a new offense, in the trial of that new offense, on the ground that the prior judgment was too lenient. We hold that he cannot.


Appellant was serving time in prison for burglary (three years) and aggravated sexual assault (forty-five years). He was bench-warranted to Smith County to answer for a theft offense. While in the Smith County jail, he escaped. He was ultimately convicted on both the theft and escape charges. He was sentenced to two years in state jail for the theft and ten years in prison for the escape. The escape sentence was ordered to be run concurrently with the theft sentence. The written judgment for the escape conviction is silent as to whether the escape sentence was to run concurrently to or consecutively with appellant's prior burglary and aggravated sexual assault sentences. This written judgment is also silent concerning whether or not there was a plea agreement.

Appellant later committed more crimes and was ultimately charged with new felony offenses of escape, burglary of a habitation, and theft. The State alleged two prior judgments of conviction for enhancement purposes. One of those prior judgments was for the Smith County escape case. Appellant filed a motion to quash that enhancement allegation. Relying upon Article 42.08 of the Code of Criminal Procedure, he claimed that the judgment was void because the escape sentence was run concurrently with his other sentences when the statute required that it be cumulated with the earlier burglary and sexual assault sentences he was serving at the time he committed the escape.

Article 42.08(b) provides: "If a defendant is sentenced for an offense committed while the defendant was an inmate in the institutional division of the Texas Department of Criminal Justice and the defendant has not completed the sentence he was serving at the time of the offense, the judge shall order the sentence for the subsequent offense to commence immediately on completion of the sentence for the original offense."

The State presented three responses to appellant's claim: (1) Article 42.08(b) did not require the escape judgment to be stacked because appellant was not actually "in" the prison system at the time of the escape, since he had been bench-warranted to the Smith County jail. (2) Even if the judgment had to be stacked, the absence of stacking did not make the judgment void and the deficiency could be corrected in a nunc pro tunc order. (3) Equitable notions of justice should prevent the defendant from gaining the benefit of concurrent sentencing and then turning around to complain about it later.

The trial court denied appellant's motion, the enhancement allegations were subsequently found to be true, and appellant was sentenced as an habitual offender on all three offenses, receiving sentences of thirty-three years, twenty-five years, and twenty-five years, respectively.

B. Appeal

In his sole point of error on appeal, appellant re-urged his contention that the Smith County escape judgment was void. In response, the State re-urged the first argument it had made at trial. The court of appeals found that Article 42.08(b) required that the sentence for the Smith County escape be stacked onto the sentence for the prior aggravated sexual assault, and it found that the failure to stack rendered the prior judgment void and thus unusable for enhancement purposes. As a result, the court of appeals reversed the judgment of the trial court and remanded the case for a new punishment hearing.

The court of appeals found that the State conceded in its brief that, under Fullbright v. State, 818 S.W.2d 808, 810 (Tex.Crim.App. 1991), "if the trial court erred in failing to stack the Smith County escape sentence, then it cannot rely upon that conviction to enhance Rhodes's convictions that are on appeal here." Rhodes v. State, 175 S.W.3d 348, 352 (Tex.App.-Houston [1st Dist.] 2004). In its brief, the State said, "The State does not disagree with the Appellant on the nature of the law as to the use of convictions for purposes of enhancements as stated in Fullbright v. State." We are uncertain that this concession is as expansive as the court of appeals has characterized it, but in any event, we are not bound by a party's concession on a matter of law. Long v. State, 931 S.W.2d 285, 289 (Tex.Crim.App. 1996).

Because appellant was serving time for both the burglary and the aggravated sexual assault offenses at the time of the escape, Article 42.08 required the escape sentence to be stacked on the much longer sentence for the aggravated sexual assault conviction because it would be the last to terminate. See Basden v. State, 897 S.W.2d 319 (Tex.Crim.App. 1995).

Id. at 355.

The State filed a motion for rehearing, repeating its original argument and making the following additional arguments: (1) a prior conviction can be valid for enhancement purposes even if the sentence is void, (2) the trial court had no evidence before it that appellant met the criteria outlined in Article 42.08(b), (3) appellant bargained for the punishment obtained and cannot now complain of an action he earlier requested, and (4) a cumulation order is not part of the sentence. Appellant filed a response to the State's motion. In responding to argument (3) above, appellant acknowledged that he "did enter into a plea bargain for his sentence in the Smith County escape case" but claimed that he did not "invite" the error as a result. The motion for rehearing was denied.

En banc consideration was also requested but denied. In an opinion dissenting from the denial of en banc consideration, Justice Keyes argued that a cumulation order is not part of a sentence, so its absence could not render a sentence void.

Id. at 356-59 (Keyes, J., dissenting from denial of en banc consideration).

C. Discretionary Review

In its petition for discretionary review, the State raises several issues, which can be summarized as follows: (1) whether appellant was estopped from complaining that his sentence was illegal, (2) whether a defendant's constitutional rights would be violated by a nunc pro tunc order that would make the sentences run consecutively, (3) what is the legal effect of a failure to order a sentence to run consecutively as required by Article 42.08(b), and (4) what amount of evidence is required to invoke the mandatory provisions of Article 42.08(b).

The State's grounds for review are as follows:

First: This Court needs to decide whether a defendant can be estopped from complaining about an error capable of rendering a judgment or sentence void.

Second: This Court needs to decide whether a defendant can be estopped from complaining, after he enjoyed its benefits, that his sentence was illegal or void because it was below the statutory minimum.

Third: This Court needs to decide whether a defendant's constitutional rights are violated when a trial court corrects a clerical error by changing a concurrent sentence to a statutorily mandatory consecutive sentence years after the defendant has begun to serve the sentence.

Fourth: This Court needs to determine the legal effect of the failure to order a sentence to run consecutively as ordered by 42.08(b) of the Texas Code of Criminal Procedure. Fifth: This Court needs to make a practical review of the question of what amount of evidence is required to invoke the mandatory provisions of Article 42.08(b).

In arguing the estoppel question, the State claims that appellant entered into a plea agreement for his sentence on the Smith County escape charge. In his response brief, appellant contends that there is no evidence in the record to prove that his conviction was the result of an agreement. Our review of the trial record in the present case reveals that there was indeed no indication that the parties entered into a plea agreement. The trial record contains no court reporter's record or plea papers that might have shed light on whether a plea agreement existed, and the trial record does not otherwise contain any indication of a plea agreement. However, nothing in the trial record specifically refutes the existence of a plea agreement either, and as we observed above, appellant conceded the existence of a plea agreement in a pleading before the court of appeals. But even if we assume there was a plea agreement, we have no information concerning its terms, and more specifically, whether the terms included concurrent sentencing, unless we construe the State's contention that there was a plea agreement as an admission to that effect.

Appellant's main argument in his response brief is that the State forfeited review of its current claims by failing to raise them on original submission to the court of appeals. Arguably, the State's fifth ground for review is an extension of the arguments it did present on original submission to the court of appeals. Nevertheless, because the State prevailed at trial, we may in our discretion address subsidiary arguments that support the trial court's ruling that were not presented to the court of appeals. Volosen v. State, 227 S.W.3d 77, 80 (Tex.Crim.App. 2007).
Judge Price contends that Volosen's holding is inconsistent with prior holdings in Rochelle v. State, 791 S.W.2d 121 (Tex.Crim.App. 1990), and Sotelo v. State, 913 S.W.2d 507 (Tex.Crim.App. 1995), and he contends that the holdings in those latter cases would bar our disposition of the current case. But Rochelle and Sotelo are distinguishable because, in both cases, the petitioner's ground for review did not challenge any holding of the court of appeals, while Volosen and the present case involve grounds that do challenge a holding of the court of appeals, but for reasons not addressed by the lower appellate court.
In Rochelle, the State advanced a procedural default ground for review that did not challenge the court of appeals's holding that the indictment was fatally defective. 791 S.W.2d at 122, 125. In Sotelo, the defendant's claim that a remand for a new punishment hearing violated double jeopardy principles was not a challenge to the court of appeals's holding that the trial court had erroneously quashed the State's enhancement allegations. 913 S.W.2d at 508-09. By contrast, in Volosen, the State's grounds for review directly challenged the court of appeals's holding that the evidence was legally insufficient to support the defendant's conviction because the defendant had established a defense as a matter of law — though the argument supporting the State's challenge had not been considered by the court of appeals. 227 S.W.3d at 79-80. Similarly, in the present case, the arguments we consider today respond directly to the court of appeals's holding that the prior conviction was not available for enhancement purposes, although those arguments were not discussed in the court of appeals's opinion.
This is the same sort of position the parties are in when the case is before a court of appeals. The party who loses at the trial level must address both the trial court's ruling and the rationale for that ruling, see TEX.R.APP. P. 33.1(a)(1)(A) (party complaining about trial proceedings must have stated grounds for ruling), (2)(A) (party complaining about trial proceedings must have obtained a ruling), but the party who wins at the trial level can prevail on appeal on the basis of a rationale not addressed by the trial court, see State v. Herndon, 215 S.W.3d 901, 905 n. 4 (Tex.Crim.App. 2007)("the general rule is that a trial court's ruling will be upheld if it is correct on any applicable legal theory, even if the court articulated an invalid basis"). The point of Volosen is that the parties' positions in this regard are fixed at trial and do not change during the subsequent course of appellate litigation. Thus, the party who loses at the trial level who then complains about a decision of the court of appeals must address both the holding and the reasoning of the court of appeals, but the party who wins at the trial level who complains about a decision of the court of appeals need only address the holding of the court of appeals.

An agreement to concurrent sentencing would be to the State's disadvantage if estoppel does not apply, because if there is no agreement, then the sentence is reformable. See discussion below.

Although the ten-year sentence for the Smith County escape was the maximum for a third-degree felony, the record suggests that the offense could have been enhanced to a second-degree felony with either the prior burglary or the prior sexual assault conviction. See TEX. PEN. CODE § 12.42(a)(3) (West 1998).


We need not decide whether to treat this case as involving a plea agreement with respect to the concurrent sentencing issue. As we shall explain, if there was no plea agreement on the concurrent sentencing issue, then the judgment is not void, and thus not subject to collateral attack, because it was at some point in time subject to reformation. If there was a plea agreement on the concurrent sentencing issue, then appellant is estopped from challenging the judgment. We detail our reasoning on these points below.

Because of this holding, we need not address whether a defendant could be estopped to attack a judgment even when there was no plea agreement.

A. No Agreement Means Judgment Is Not Void

Appellant's challenge to the enhancement allegation in this case constitutes a collateral attack on the prior judgment of conviction. Such a collateral attack is permitted only if the prior judgment is void, and not merely voidable.

New v. State, 583 S.W.2d 781, 782 (Tex.Crim.App. 1979).


Initially, we point out that there is a difference between an entire judgment being "void" and a portion of a judgment being `Void." For example, a judgment may contain two or more sentencing elements (e.g. imprisonment and fine), one of which may be valid while the other is void. When only one of the sentencing elements is void, the judgment is rendered void only if the judgment cannot be reformed to cure the infirmity (i.e. the infirmity cannot be cured without resort to resentencing). Some cases addressing other types of judgment defects have like-wise indicated that the judgment is not void if the defect could have been reformed on the direct appeal of the judgment in question or in a nunc pro tunc order.

See Ex parte Johnson, 697 S.W.2d 605 (Tex.Crim.App. 1985) (imprisonment was within the range of punishment, but fine exceeded the range of punishment).

Id. at 607-08; Barker v. State, 169 Tex. Crim. 277, 278-79, 334 S.W.2d 182, 183-84 (1960).

Rhodes v. State, 560 S.W.2d 665, 667 (Tex.Crim.App. 1978) (quoting Barker, 169 Tex. Crim. at 279, 334 S.W.2d at 184)("A judgment or sentence containing an irregularity which may be reformed on appeal or by nunc pro tunc entry is not void, and may not be collaterally attacked."); Smothermon v. State, 383 S.W.2d 929, 931 (Tex.Crim.App. 1964)("could have been reformed in a proper proceeding"); Ex parte Brown, 145 Tex.Crim. 39, 42, 165 S.W.2d 718, 720 (1942)("upon the original appeal of this case, it could have been reformed").

This Court has addressed at least three different curable judgment defects relating to punishment that arose from a contested trial, in which there was no plea agreement. In Williams v. State, we confronted on direct appeal a claim that a cumulation order was void. We agreed, and we deleted the cumulation order but otherwise affirmed the conviction.

675 S.W.2d 754, 760 (Tex.Crim.App. 1984).


In Ex parte Johnson, an applicant claimed on habeas corpus that his conviction for aggravated robbery was void because it contained an unauthorized fine. The jury had rendered a verdict assessing punishment at fifty years of confinement and a $10,000 fine, but according to the habitual-sentencing statute under which the applicant was punished, no fine was authorized. We held that the trial court was empowered to reform the judgment by deleting the unauthorized fine.


Id. at 607-08.

In Barker v. State, the defendant did what appellant did here: he challenged a prior conviction used for enhancement purposes at the trial of a new offense. He claimed that the prior judgment was void because, in addition to a valid three-day jail sentence and fifty-dollar fine, the judgment contained an invalid provision commuting the sentence to a six-month term of probation. We rejected the defendant's contention. First, we relied upon Ex parte King for the proposition that "a judgment or sentence containing an irregularity which may be reformed on appeal or by nunc pro tunc entry is not void, and may not be collaterally attacked." Second, we cited Lenore v. State for the proposition that "a prior judgment providing for a fine of $25 when the minimum punishment provided by law was a $100 fine was erroneous, but not void, and was available for enhancement of punishment in a prosecution for a subsequent offense."


Id. at 279, 334 S.W.2d at 184.

Id. (citing Ex parte King, 156 Tex.Crim. 231, 240 S.W.2d 777 (1951)).

Id. (citing Lenore v. State, 137 Tex.Crim. 417, 129 S.W.2d 657 (1939)).

Clearly, where concurrent sentencing is not a part of a plea agreement, reforming the judgment to make the sentence consecutive can be accomplished without resort to resentencing. We have already held that a cumulation order can be deleted without disturbing the remainder of the judgment. Imposing a cumulation order where one is required is simply the flip side of the coin. This type of situation does not involve a range of valid possible punishment options that would require a new sentencing hearing to be held. If concurrent sentencing between two sentences (here the escape and the prior aggravated sexual assault) is invalid, then only one valid option remains (consecutive sentencing), making reformation possible, at the very least, on direct appeal.

We need not decide whether the defect here is itself of the void or voidable variety, whether it is even voidable at all, or whether a cumulation order (or lack thereof) is part of the "sentence." However one characterizes the defect of failing to cumulate appellant's sentence in the prior escape judgment, the reformable nature of the defect means that it cannot render the prior judgment void.

See Ex parte Townsend, 137 S.W.3d 79, 80-82 (Tex.Crim.App. 2004) (challenge to stacking order as invalid due to defendant's status not cognizable on habeas corpus when it could have been raised on direct appeal).

See Langs v. State, 183 S.W.3d 680, 686 (Tex.Crim.App. 2006) (double jeopardy, multiple punishment violation can be forfeited when the violation is not clear on the face of the record and forfeiture would serve the interests underlying the rules of procedural default).

See State v. Ross, 953 S.W.2d 748, 750 (Tex.Crim.App. 1997) (for purpose of State's appeal, "sentence" includes whether the term of imprisonment is concurrent); but see Speth v. State, 6 S.W.3d 530, 535 (Tex.Crim.App. 1999)(Womack, J., concurring) (criticizing use of statutory definition of "sentence" to determine what this Court means when it says a sentence is void).

We need not and do not decide whether the State now has any remedy that would enable it to cumulate the prior escape sentence, or whether the State's own use of the prior judgment to enhance the current offense would estop the State from utilizing such a remedy.

B. Agreement Means Appellant Is Estopped from Challenging the Judgment

This Court has not, in a majority opinion, addressed the application of the estoppel doctrine in circumstances such as these. But the issue was discussed in a plurality opinion on rehearing in Heath v. State and in a concurring opinion in Ex parte Williams, both of which involved a defendant challenging a prior plea-bargained judgment of conviction on the ground that the trial judge imposed probation without the authority to do so. The Heath plurality decided that the doctrine of estoppel would not bar a plea-bargaining defendant's challenge to an illegal sentence. By contrast, the Williams concurrence concluded that the plea-bargaining defendant in that case was estopped from challenging the trial court's judgment because he had accepted the benefits of probation. Although a majority of the Court in Williams chose to deny relief upon a different ground, this Court has subsequently suggested that the estoppel issue raised by the concurring opinion is an open question.

817 S.W.2d 335, 337-40 (Tex.Crim.App. 1991) (plurality opinion on rehearing).

65 S.W.3d 656, 658-60 (Tex.Crim.App. 2001)(Keller, P.J., concurring).

See Heath and Williams, supra. The opinion on original submission in Heath has been overruled on the basis that probation is not a part of the sentence. Williams, 65 S.W.3d at 657-58 (Court's op.).

65 S.W.3d at 658-60 (Keller, P.J., concuring).

Id. at 658 (Court's op.)("probation" is not part of the sentence and the defendant failed to show harm).

Mizell v. State, 119 S.W.3d 804, 806 n. 8 (Tex.Crim.App. 2003) (citing concurrence in Williams and citing Collins v. State, 509 N.E.2d 827 (Ind. 1987)).

An examination of the plurality opinion in Heath reveals that its holding on the estoppel issue was based upon a faulty premise. Immediately before reaching its conclusion, the plurality discussed Shannon v. State. The Heath plurality pointed out that, in Shannon, the defendant was sentenced to two years of confinement in accordance with a plea agreement, but he contended on appeal that the offense for which he was sentenced was actually a Class A misdemeanor with a maximum punishment of one year. The court of appeals in Shannon agreed, but remanded the case for resentencing. In a petition for discretionary review, the State correctly argued that because there was a plea bargain, the proper remedy was to undo the entire plea. We agreed. The court of appeals in Shannon had erroneously applied to the negotiated plea setting the rule that error affecting only punishment results in only a new sentencing hearing.

Heath, 817 S.W.2d at 339-40 (discussing Shannon v. State, 708 S.W.2d 850 (Tex.Crim.App. 1986)).

Id. at 339 (citing Shannon).

Id. at 339-40 (citing Shannon, 708 S.W.2d at 851).

Id. at 340 (citing Shannon, 708 S.W.2d at 851).



The Heath plurality then made an illogical leap: it assumed that the mistake made by the court of appeals in Shannon (which never applied the doctrine of estoppel) was somehow an assumption underlying the doctrine of estoppel. Without any explanation other than the unsupported assumption that Shannon somehow had something to do with the issue at hand, the Heath plurality reached the odd conclusion that applying estoppel to bar a plea-bargaining defendant's complaint about an illegally lenient sentence would result in "treating those errors as punishment errors only" and somehow involving resentencing. But this proposition makes no sense. The Heath plurality took the premise that it is unfair to bind just one party to a plea bargain, and concluded from this that when a sentence is void, both parties must be unbound from the plea bargain. The alternative the Heath plurality overlooked was that both parties could be bound by the plea bargain — which is precisely what estoppel would accomplish. We thus conclude that the Heath plurality lacks persuasive force.

See Heath, 817 S.W.2d at 340.

Acknowledging a "dearth of Texas case-law" on the subject, the Williams concurrence cited Pry stash v. State, which applied to the somewhat different "invited error" context, and out-of-state cases that were more on point. We now turn to a full examination of the authorities relating to this subject.

Williams, 65 S.W.3d at 659 (Keller, P.J., concurring).

3 S.W.3d 522 (Tex.Crim.App. 1999).

See Williams, 65 S.W.3d at 658-660.

We have previously dealt in Prystash v. State with a type of estoppel involving unilateral requests that result in "invited error," but estoppel is a flexible doctrine that manifests itself in various forms that are not limited to unilateral requests. The variant of estoppel at issue here is "estoppel by judgment." In this type of estoppel, "[o]ne who accepts the benefits of a judgment, decree, or judicial order is estopped to deny the validity or propriety thereof, or of any part thereof, on any grounds; nor can he reject its burdensome consequences." The only exception to this principle is for challenges to the subject-matter jurisdiction of the court rendering the judgment.

31 C.J.S. Estoppel Waiver § 130.

Id., p. 565.

Id., p. 566; see also Williams, 65 S.W.3d at 659-60 (Keller, P.J. concurring) (discussing cases).

Moreover, a close cousin of "estoppel by judgment" is "estoppel by contract," where a party who accepts benefits under a contract is estopped from questioning the contract's existence, validity, or effect. So estoppel can apply not only to unilateral requests (as in Prystash), but also in the context of agreements, where the "request" is mutual. As is true with most contracts, it is typical that both parties to a plea bargain will benefit from the judgment.

31 C.J.S. Estoppel Waiver § 124, p. 554.

Indeed, courts in several other jurisdictions have held that a defendant cannot enter a plea agreement that imposes an illegal sentence, benefit from that sentence, and then attack the judgment later when it is suddenly in his interests to do so. In rejecting a defendant's claim for overturning a too-lenient, illegal sentence, the Mississippi Court of Appeals anticipated exactly what applicant wishes to accomplish in the present case:

People v. Hester, 22 Cal.4th 290, 295, 92 Cal.Rptr.2d 641, 992 P.2d 569, 572 (2000)("defendants who have received the benefit of their bargain should not be allowed to trifle with the courts by attempting to better the bargain through the appellate process"); Collins, 509 N.E.2d at 833 ("a defendant may not enter a plea agreement calling for an illegal sentence, benefit from that sentence, and then later complain that it was an illegal sentence"); Punta v. State, 806 So.2d 569, 570 (Fla.App. 3rd DCA 2002) (quoting Bashlor v. State, 586 So.2d 488 (Fla.App. 1st DCA 1991))("sentences which are imposed in violation of statutory requirements, which are to the benefit of the defendant and to which he agreed, may not be challenged after the defendant has accepted the benefits flowing from the plea, but has failed to carry out the conditions imposed on him"); Graves v. State, 822 So.2d 1089, 1091 (Miss.App. 2002) (defendant "cannot stand mute when he is handed an illegal sentence which is more favorable than what the legal sentence would have been, reap the benefits of that illegal sentence, and later claim to have been prejudiced as a result thereof"); Pruitt v. State, 846 So.2d 271, 274 (Miss.App. 2002)("A convicted felon may not quietly enjoy the benefits of an illegally lenient sentence, and later attack the sentence when suddenly it is in his interest to do so").

On the other hand, a defendant should not be allowed to reap the benefits of an illegal sentence, which is lighter than what the legal sentence would have been, and then turn around and attack the legality of the illegal, lighter sentence when it serves his interest to do so. Allowing such actions would [wreak] havoc upon the criminal justice system in this state. For example, all subsequent convictions and sentences of that defendant which are reliant upon the conviction concomitant with the illegal sentence would have to be set aside. This would result in a number of enhanced and habitual offender sentences being set aside for the very offender who had already enjoyed greater leniency than the law allows.

For the same reasons, that court refused to grant relief in a later case when the defendant challenged a prior unauthorized probation after it had been revoked and used to enhance a later offense.

Pruitt, 846 So.2d at 274.

We find these cases persuasive. A defendant who has enjoyed the benefits of an agreed judgment prescribing a too-lenient punishment should not be permitted to collaterally attack that judgment on a later date on the basis of the illegal leniency. Here, appellant received a judgment that was illegally lenient by having his sentence run concurrently instead of consecutively. Had he complained about the illegal leniency at the time of trial, or even on direct appeal, the State could likely have obtained a legal judgment that would now be available for enhancement purposes. But instead, appellant quietly enjoyed the benefits of the illegally lenient judgment, challenging it now only because, due to his own subsequent criminal conduct, the judgment can be used to enhance his punishment for a new offense. If he agreed to the concurrent sentencing provision, then through his own conduct he helped procure and benefit from the illegality and he should not now be allowed to complain.

We express no opinion today on whether estoppel could apply to bar a challenge to a judgment imposing a punishment that is illegally harsh, nor do we express an opinion on whether estoppel would bar a direct attack on a judgment imposing an illegally lenient punishment. See Williams, 65 S.W.3d at 660 (Keller, P.J., concurring) (observing that it could be persuasively argued that a defendant who appeals his conviction has not yet accepted the benefits of the illegal judgment).

The judgment of the court of appeals is reversed, and the judgment of the trial court is affirmed.

MEYERS, WOMACK and JOHNSON, JJ., joined Parts I and IIA and dissented to Part IIB.

PRICE, J., filed a dissenting opinion.

PRICE, J., filed a dissenting opinion.

I respectfully dissent to the Court disposing of this case on the basis of an issue that was never decided by the court of appeals. For a long time we have said that, in our capacity as a discretionary review court, we review only "decisions" of the courts of appeals. We have not always been able to agree on what constitutes a "decision" of the court of appeals. But we have been fairly consistent in the requirement that the court of appeals must reach a "decision" with respect to an issue before we will review it on the merits. There are some indications in the case law that we may even lack jurisdiction to do otherwise. In cases in which the direct appeals court should have reached the merits of an issue, but did not, and the failure to reach the merits is brought to our attention in a petition for discretionary review, we remand so that the appellate court can reach a "decision" in the first instance. But we do not address the merits ourselves unless the issue is brought back to us after remand.

E.g., Arline v. State, 721 S.W.2d 348, 353 n. 9 (Tex.Crim.App. 1986); Lee v. State, 791 S.W.2d 141, 142 (Tex.Crim.App. 1990); Holland v. State, 802 S.W.2d 696, 700-01 (Tex.Crim.App. 1991); George E. Dix Robert O. Dawson, 43A TEXAS PRACTICE: CRIMINAL PRACTICE AND PROCEDURE § 44.21, at 858-59 (2d ed.2001).

E.g., Barfield v. State, 63 S.W.3d 446, 449 n. 6 (Tex.Crim.App. 2001).

If the court of appeals failed to render a decision on an issue properly brought before it, we will sometimes grant discretionary review to determine whether the court of appeals erred in failing to decide the issue. But when we find error on the part of the court of appeals in these instances, we remand to the court of appeals rather than reach the merits ourselves. E.g., Lee v. State, supra; Dix Dawson, supra, § 44.23, at 862-63.

E.g., Owens v. State, 827 S.W.2d 911, 917-18 n. 7 (Tex.Crim.App. 1992); Dix Dawson, supra, § 44.21, at 859. See also State v. Consaul, 982 S.W.2d 899, 902 (Tex.Crim.App. 1998) (Price, J., concurring) ("This court's jurisdiction is limited to review of decisions by the courts of appeals.").

In the past several years, Presiding Judge Keller has begun to advocate an exception to this rule, allowing the State to raise issues of error preservation for the first time in petitions for discretionary review. In Alonzo v. State, she first argued that, because issues of error preservation are "systemic," and should be reached by a first-tier appellate court whether raised by the parties or not, the failure of the State to explicitly raise procedural default in its reply brief on direct appeal should not prevent it from raising the issue in a petition for discretionary review. In that event, she has argued, this Court "should either remand the case to the court of appeals to consider the issue or [we] should consider the preservation issue on discretionary review." In Haley v. State we implicitly adopted Presiding Judge Keller's position, and even reached the merits of the procedural default question ourselves, rather than to remand it, though the court of appeals had never rendered a "decision" with respect to that question.

158 S.W.3d 515 (Tex.Crim.App. 2005) (Keller, P.J., dissenting).

Archie v. State, 221 S.W.3d 695, 698 (Tex.Crim.App. 2007); Jones v. State, 942 S.W.2d 1, 2 n. 1 (Tex.Crim.App. 1997).

Alonzo v. State, supra, at 518 (Keller, P.J., dissenting).

Id. at 519.

Haley v. State, 173 S.W.3d 510, 515 (Tex.Crim.App. 2005).

In Haley, the same case in which we implicitly adopted the Presiding Judge's position in Alonzo, she filed another separate opinion, concurring in the result. This time she argued that her Alonzo exception should not apply just to procedural default questions, but should also apply to any issue that the prevailing party at trial neglected to raise in its reply brief on direct appeal. She maintained that:

the party that prevailed at trial should never be required to advance an argument before the Court of Appeals as a predicate for raising that argument on discretionary review. This conclusion is consistent with the view that I articulated in Alonzo v. State with regard to preservation of error. I would further conclude that the reasoning in my Alonzo opinion should apply also to all types of arguments that could be made by a party that prevailed at trial.

Id. at 519 (Keller, P.J., concurring).

Quite recently, and without fanfare, the Court adopted this view as well. In Volosen v. State, the Court held that the failure of the prevailing party in the trial court to make an argument in its reply brief in the first-tier appellate court will not prevent this Court from entertaining that argument on the merits when raised for the first time in a petition for discretionary review. I dissented in Volosen, and I dissent again today.

227 S.W.3d 77, 80 (Tex.Crim.App. 2007).

The evolution of this exception to the requirement that there be a "decision" of the court of appeals for this Court to review illustrates how inconsistencies creep into our case law. In her dissenting opinion in Alonzo, Presiding Judge Keller candidly acknowledged that her position was at odds with this Court's opinion in Rochelle v. State, the "viability" of which she argued we should "re-examine." In her separate opinion in Haley, she did not revisit this inconsistency. She simply argued that her position in Alonzo, which the Court adopted sub silentio in Haley, ought to be extended. And in Volosen, that is just what the Court did, but again without ever addressing the apparent conflict with, much less overruling, our binding precedent in Rochelle.

791 S.W.2d 121 (Tex.Crim.App. 1990).

In Rochelle we held that when the State (the prevailing party in the trial court on the facts of that case) raises an issue for the first time in a motion for rehearing, the court of appeals is not obligated to address the issue, and if it does not, there is no "decision" with respect to that issue for the State to raise in a petition for discretionary review. That is precisely what happened in the instant case; Rochelle is directly on point. In its reply brief on direct appeal, the only argument that the State made in response to the appellant's point of error on appeal was that Article 42.08(b) of the Code of Criminal Procedure did not apply to him, since he was not in the custody of the penitentiary at the time of his escape and was therefore not an "inmate" in contemplation of that provision. The State made no other argument, and was content to accept that Fullbright v. State, upon which the appellant principally relied, was controlling. Then, for the first time in its motion for en banc rehearing, the State made a number of new arguments, including the argument that the appellant should be estopped from claiming that an illegally lenient punishment, of which he has enjoyed the benefits, is void. The court of appeals did not see fit to grant rehearing. Under Rochelle, then, which has never been overruled, there is nothing for this Court to review.

Id. at 124-25; See also Sotelo v. State, 913 S.W.2d 507, 509 (Tex.Crim.App. 1995).

818 S.W.2d 808 (Tex.Crim.App. 1991).

Rhodes v. State, 175 S.W.3d 348, 355 (Tex.App.-Houston [1st] 2004).

Two justices dissented to the denial of the en banc rehearing. Id. at 356-59. Those justices did not address the State's estoppel argument. Instead they addressed another issue altogether, viz., whether Fullbright should control the question whether an order that illegally fails to cumulate sentences renders the sentence "void." The dissenters believed that it should not, and that, absent a void sentence, the prior conviction was available to enhance the appellant's instant punishment. Even if it could somehow be said that, because two justices filed a written dissent, the court of appeals "decided" this question when it denied rehearing, the issue thus "decided" is still not the same as the one the majority decides today.

I can agree that Rochelle ought to be overruled to the extent that it applies to issues of procedural default. Questions of error preservation are systemic, and first-tier appellate courts are required to address them regardless of whether the parties join issue in their appellate briefs. If a court of appeals fails to reach an obvious issue of procedural default, the party that prevailed in the trial court ought to be able to complain about it for the first time in a petition for discretionary review, and we ought to remand the cause for the court of appeals to address the issue in the first instance. And indeed, I joined the majority opinion in Haley, in which we implicitly adopted Presiding Judge Keller's exception as it applied to error-preservation issues. But I do not think the rationale justifying the exception extends beyond that context, which is the reason I dissented in Volosen. Consistent with Rochelle, we ought to hold that any non-systemic issue that the party who prevailed at trial neglects to raise in a reply brief and which is therefore not necessary to the court of appeals's decision, and not in fact decided by the court of appeals, cannot be raised for the first time in a petition for discretionary review. Because the Court continues down this path without ever re-examining the viability of Rochelle (at least as it applies to issues other than procedural default) and reaches out to decide a non-procedural default issue that the court of appeals in this case never addressed, much less decided, I continue to dissent.

In retrospect, however, I think we should have remanded the cause in Haley to the court of appeals to address the issue of error preservation in the first instance, rather than deciding it ourselves for the first time on discretionary review. See note 3, ante.

If in fact Rochelle has been de facto over-ruled in its entirety, and Volosen now controls, such that the party that prevails in the trial court but loses on appeal can now bring any issue it wants in a petition for discretionary review, regardless of whether the court of appeals was called upon to decide that issue on appeal, I can only presume that in practice the Court will apply that holding even-handedly to both the State and criminal defendants.