dissenting.
Appellant argues in his ninth point of error that the trial court erred in failing to let him inform the jury, as per his explicit request, that if assessed a life sentence, he would be required by law to serve a minimum of fifteen years confinement in the penitentiary before he would even be eligible to be paroled by the Board of Pardons and Paroles. The plurality rejects this contention on the following reasoning. First, the plurality cites the “policy” of this state, grounded in principles of separation of powers, to prohibit jury deliberation on parole in capital cases, and “not to officially inform jurors of the actual consequences of a life sentence.” Op. at 849. Second, notwithstanding the opinion of the United States Supreme Court in Simmons v. South Carolina, — U.S. -, 114 S.Ct. 2187, 129 L.Ed.2d 133 (1994), the plurality concludes that the Due Process Clause of the Fourteenth Amendment does not require informing the jury as appellant requested in this cause, and thus does not trump our own state-eonstitutionally-based policy against it.
After careful consideration, I find I must disagree with the plurality in both its conclusions. First, I believe we have been mistaken over the last twenty years to conclude that principles of separation of powers somehow prohibit (we have never expressly explained how) informing jurors in a capital case that a capital defendant sentenced to life imprisonment must serve a certain minimal length of that sentence before he becomes eligible for parole. They do not. Second, that a capital defendant sentenced to serve a life sentence must actually serve a certain minimum length of that sentence before becoming eligible for parole seems to me to be “indisputably relevant” to the determination whether he will “commit criminal acts of violence that would constitute a continuing threat to society” under former Article 37.071(b)(2), V.A.C.C.P. See Simmons, supra, at-, 114 S.Ct. at 2194, 129 L.Ed.2d at 142. Extrapolating from the principles announced in the plurality opinion in Simmons, the conclusion seems inescapable to me that failure to avail the jury of this “indisputably relevant” information violates due process — at least where, as here, there is evidence that by the time the defendant has served the statutory minimum penitentiary time, his potential “threat to society” at large *858will be greatly reduced. In my view, then, state separation of powers does not prevent, and federal due process positively requires, at least upon request, that the jury be informed of, and allowed to consider for its relevance to future dangerousness, the minimum length of time a capital defendant sentenced to life must serve before becoming parole-eligible.
I.
Appellant committed the instant offense on May 15,1990. An accused convicted of committing a capital murder on that date is required by statute to serve a minimum of fifteen years in the penitentiary, exclusive of good time credits, before becoming eligible for parole. See former Article 42.18, § 8(b), V.A.C.C.P., prior to its amendment by Acts 1991, 72nd Leg., ch. 652, §§ 10 & 15(b), pp. 2396-97, eff. Sept. 1, 1991. This means that appellant would not even become eligible for release on parole until well into his upper thirties.
At the punishment phase of trial appellant called Dr. Fred Fason, a board certified psychiatrist, to testify. Fason had examined appellant prior to trial for purposes of determining his sanity at the time of the offense and his competency to stand trial. In the course of this examination Fason concluded that appellant was “suffering from antisocial, what is now referred to as antisocial reaetion[,]” formerly known as “sociopathie disturbance.” Fason elaborated that antisocial reaction:
“is characterized by ... the rebelliousness, self-centeredness, concern for oneself and difficulties with the law of antisocial behavior. In essence, it’s the person’s mental problems manifested in their behavior rather than in their thinking or their feelings.
Q How does a person in effect become that way? Again for want of a better phrase, is this something that a person is born with or is it something that the person develops during an individual development as a human being?
A There are somewhat different theories. I will try to explain it to the best that I can. * * * [T]he first thing about the sociopathie personality disorder or the antisocial individual has to do with what occurs at relatively an early age of development. Most of us when we get to — those of us who have children, remember when our children were a year and a half or two years old. Before the age of two, children that are that age are prone to be very narcissistic. And what I mean by that I mean they view the world and their whole life as when they cry and vomit and mama appears and fees [sic] them. If they are cold, they cry and mama appears and covers them up, etcetera. Now, their views are prone to see the mother as kind of a light, the genie from Aladdin’s lamp. And around the age of two and a half or somewhere in there the child becomes aware that mother looks after him because she loves him, not because she has to. And it’s a whole new ball game. Mother with children experience this. They say that this is when the terrible twos become the terrific threes and the child understands now that Mother is looking after him not because he makes her do that but because she loves him and it’s a whole new ball game.
Now, the narcissistic individual is left in a sense like the year and a half old baby that he feels entitled to whatever he wants. And people get started at this stage of development they become self-centered, they don’t really are [sic] about other people. And this is a sense almost of the sociopath or psychopath. There is an additional factor that gets superimposed and without — and I hope I’m not being disruptive for the Court or the ladies and gentlemen of the jury. I will have to quote what goes on inside the mind of the psychopath to relieve the stress that others of us experience. And it is an attitude that is expressed by what people say to themselves in the phrase — and this is based upon the examination of hundreds of individuals who have fallen into the diagnostic category. Again if you will pardon the expression, it’s fuck it, I don’t care. And this is what the *859individual says to himself about the things that are normally stressful to him. And in order to make it work he has to act like it and many individuals that, as with Mr. Smith when I asked him if he remembers when he started saying that to himself and he clearly remembered it. And it was a way of discounting. It’s a way of avoiding shame. It’s a way of avoiding the ordinary drive control that we use to control our impulses.
* * ⅜ * * *
The sociopathic personality at some point in the development not only is started at the narcissistic level but adopts as a life-style the phrase to himself that enables him to discount the rules. And if you say well, aren’t you ashamed of that, you get that response. Are you concerned that it hurts that person, and again it’s that same phrase. And this frees the sociopathic personality from having to struggle with conflicts that most of us struggle with in life. It comes out in their behavior.
Q Let me clarify that for myself. You used the term drive controls—
A Yes.
Q —that are in place for the normal individuals who interact with other people on a daily basis in our society; is that right?
A Yes.
Q Those drive controls would be shame?
A Yeah. The three drive controls that most of us use to control our impulses are, number one, shame. That is what it says about us. I might get an impulse to do something and it might be a natural impulse, as Freud said, but I might think I’m not that kind of person. I don’t want to be the kind of person that would do something like that. And if I did that I would be ashamed of doing that and I wouldn’t want to feel that way and that would keep me from acting on that impulse. Now, that’s one.
The second one would be guilt. I don’t like to feel guilty. I don’t know about you but I don’t think most normal people like to feel guilty about something. And I don’t like to do things that make me feel guilty. * * *
The third drive control that we normally use is concern about the consequences. * * * So that you have those three types of drive control that people use to fit in this society.
Q All right. Go ahead. I’m sorry.
A Mr. Smith doesn’t have those. He quit using those a long time ago.
Q And as I understand, that has been replaced, the concern for those three areas or drive controls have been basically either replaced or covered up with — again my apologies to the jury — the attitude of the phrase fuck it, I don’t care?
A That is right. When his conscience started to speak to him years and years ago the little voice inside says you shouldn’t do that. But he started saying that to his conscience so many years ago that he even quit listening to it and quit hearing it.”
Asked what appellant’s “prognosis” would be, Fason replied:
“... [Y]ou see a lot of people in their twenties with this diagnosis. Anyone in the criminal justice system who has been here a while sees an awful lot of people who are 18 to 29 that would fall into that category. When people get past the age of 30 you don’t see that many in the criminal justice system. And past age 40 you rarely ever see someone with this diagnosis past the age of 40.
⅜ * ⅜ ⅜ ⅜ ⅝
... [Sociopaths] get into their late twenties or early thirties or sometimes mid-thirties and they start realizing they do care and telling themselves they don’t care doesn’t work anymore. And when it doesn’t work anymore then they have to start changing their whole lifestyle.
... What happens to them when they get to their forties? Occasionally you will *860see one but it is relatively rare or at least uncommon in comparison to the number of individuals that fall into this diagnostic category in their twenties.
When appellant next attempted to ask Dr. Fason how long appellant would be required to serve in the penitentiary before he would become eligible for parole, the State objected, and a lengthy bench conference ensued. Ultimately the trial court ruled that the answer to this question was inadmissible because it would invite jurors to disregard the conventional jury instruction charging them not to consider any action the Board of Pardons and Paroles might take on some future occasion. Dr. Fason was ultimately allowed to testify in response to questioning by appellant, however, that “the length of time that a person is locked up might be a factor in determining whether or not that person is a danger to this community!)]”1
Later, at the jury charge conference, appellant requested “that the parole instruction be withdrawn and that the defendant be permitted through counsel to instruct the jury and tell the jury as to the length of time the defendant in all probability will be forced to spend ... [t]hat it would be 15 years.” This request was denied. The trial court instructed the jury:
“During your deliberations, you are not to consider or discuss any possible action of the Board of Pardons and Paroles Division of the Texas Department of Criminal Justice or of the Governor, or how long the defendant would be required to serve to satisfy a sentence of life imprisonment.”
On one occasion during his final summation, appellant’s counsel attempted an oblique rhetorical allusion to the fact that appellant would not be a danger to the general public by the time he might be paroled, viz: “I assure you that by the time the bus from Texas Department of Corrections drops him off at the bus stop most of us won’t be around any more.” The State’s objection to this argument was sustained, and the jury instructed to disregard it.
II.
The plurality observes that the jury in a capital ease may not “consider parole in any manner when considering whether a capital defendant should be sentenced to life or death.” Op. at 849. It is true that we have consistently held that judicial consideration of parole in sentencing would be improper in capital as well as non-capital cases. We have said so as early as Freeman v. State, 556 S.W.2d 287, at 304 (Tex.Cr.App.1977), and as late as Garcia v. State, 887 S.W.2d 846, 860 (Tex.Cr.App.1994). As a general proposition this is undoubtedly correct. “The evil to be avoided is the consideration by the jury of parole in assessing punishment.” Rose v. State, 752 S.W.2d 529, at 535 (Tex.Cr.App. 1987) (Opinion on original submission), citing Clark v. State, 643 S.W.2d 723, at 725 (Tex. Cr.App.1982).
In a capital ease in Texas, however, the jury does not “assess punishment.” See Boyd v. State, 811 S.W.2d 105, at 120 (Tex. Cr.App.1991). Indeed, as of the time of appellant’s trial, capital juries did not even decide “whether a capital defendant should be sentenced to life or death.” They simply answered special issues, questions of fact relating to the deliberateness with which the defendant acted to cause the death of the deceased, the probability he would be a continuing threat to society, and whether his killing of the deceased was a reasonable response to provocation, if any. See former Article 37.071(b), V.A.C.C.P. That the jury in a capital punishment proceeding might *861consider the minimum length of time a capital defendant serving a life sentence must actually serve before becoming eligible for parole, relevant to assessing the probability of his future dangerousness, does not mean it is considering parole “in assessing punishment.”
A
The Court has long held in non-capital cases that jury consideration of when the Board of Pardons and Paroles might release a convict on parole would constitute a judicial encroachment upon a manifestly executive function, in violation of Article II, § 1 of the Texas Constitution. This is surely correct. In Sawders v. State, 580 S.W.2d 349, at 351 (Tex.Cr.App.1978), we explained:
“[i]t would be improper for punishment to be based on an expectation that clemency powers would be exercised, and it would be unconstitutional to attempt to delay exercise of the clemency powers or to avoid the possible granting of parole by increasing punishment in anticipation thereof.”
See also Rose v. State, supra, at 532. It is for this reason that jury consideration of parole in a non-capital case is considered always to constitute jury misconduct under former Article 40.03(8), V.A.C.C.P., and now, Tex.RApp.Pro., Rule 30(b)(8). Sanders v. State, supra, at 352.
In 1985 the Legislature added Section 4 to Article 37.07, V.A.C.C.P. See Acts 1985, 69th Leg., ch. 576, § 1, p. 2195, eff. Sept. 1, 1985. This provision for the first time required trial courts to instruct juries in non-capital felony trials generally about the law of parole, but specifically that “[y]ou are not to consider the manner in which the parole law may be applied to this particular defendant.” Nevertheless, the statutory instruction did authorize the jury in some unspecified sense to “consider the existence of the parole law and good conduct time.” We therefore declared § 4(a) of Article 37.07 unconstitutional, on the basis, inter alia, that it violated separation of powers under Article II, § 1, supra. Rose v. State, supra, at 535.
The voters of Texas approved an amendment to Article IV, § 11 of the Texas Constitution in 1989 that for the first time authorized the Legislature to enact “laws that require or permit courts to inform juries about the effect of ... eligibility for parole ... on the period of incarceration served by a defendant convicted of a criminal offense.” See Acts 1989, 71st Leg., S.J.R. 4, § 1, p. 6414, approved Nov. 7, 1989. At the same time the Legislature reenacted Section 4 of Article 37.07, supra, thus reinstating the jury charge provisions we had struck down in Rose. See Acts 1989, 71st Leg., ch. 103, § 1, p. 442, effective upon approval of the constitutional amendment, Nov. 7, 1989. This Court later held that, Rose notwithstanding, a jury instruction given pursuant to the reenacted provision did not violate separation of powers. Oakley v. State, 830 S.W.2d 107, at 110 (Tex.Cr.App.1992).
By its terms, reenacted Article 37.07, § 4 expressly exempts capital cases from its operation. Thus, the Legislature has not expressly provided any law that permits or requires courts to inform juries of the effect of eligibility for parole on the period of incarceration that a capital defendant sentenced to life imprisonment must serve.2 Article IV, § 11, supra. It may be argued accordingly *862that, in the absence of such an express provision, the separation of powers principle of Sanders and Rose still prevails, preventing the trial court in a capital case from instructing the jury at punishment in any manner other than as the jury in this cause was in fact instructed, viz: that it may not consider any possible action of the Board of Pardons and Paroles or how long the defendant may actually have to serve to satisfy a life sentence.
This argument presupposes, however, that any jury instruction regarding parole in a capital case will violate separation of powers in the same way we perceived it would in Sanders. It presumes that a jury instruction about the parole law in a capital case will invite the jury to encroach upon the executive prerogative to grant parole in much the same way we feared that a jury so instructed in a non-capital ease would. It seems to me, however, that that presumption merits greater scrutiny, and to that I now turn.
B.
The body of caselaw culminating in Sanders has since been incorporated uncritically into our capital jurisprudence. Without stopping to ask whether, and if so to what extent, jury consideration of parole in the punishment phase of a capital case really would encroach upon the executive function, the Court has simply assumed that it would, and ruled accordingly. In O’Bryan v. State, 591 S.W.2d 464 (Tex.Cr.App.1979), for example, the defendant requested a jury instruction that explained the parole process, and function of the Board of Pardons and Paroles, and then admonished the jury “not to consider in their deliberations parole or whether [the defendant] might be paroled at some date in the future.” Id., at 478. In holding the trial did not abuse its discretion in fading to give the requested instruction, the Court first observed that “[t]he matter of parole or a defendant’s release thereon is not a proper consideration for a jury’s deliberations on punishment.” Id. Without explication, the Court then concluded: “This is also true in a capital case where the jury’s task at the punishment stage is to answer the special issues.” But is this necessarily so?
The earliest case in which a capital defendant expressly requested an instruction educating the jury as to the minimum time he would have to serve before becoming eligible for parole is Franklin v. State, 693 S.W.2d 420 (Tex.Cr.App.1985). There the defendant requested that the jury be instructed that:
“you are not to consider how much of his sentence the defendant would be required to serve before being released on parole, if his punishment is assessed at life imprisonment. You are instructed, however, that under the law, a person who is serving a sentence of life imprisonment for the offense of capital murder may not be considered for release on parole until his actual calendar time served, without consideration of good conduct time, is twenty (20) years.”
The trial court refused this instruction, and Franklin complained of it on appeal. This Court simply quoted the above language from O’Bryan v. State, supra, as if it were self-evidently dispositive of Franklin’s contention, and overruled this ground of error. But O’Bryan did not purport to address the question whether a specific jury instruction on minimum parole eligibility might be appropriate insofar as it relates to the jury’s resolution of the question whether the accused will probably constitute a continuing threat to society. With all due respect, the matter is not as simple as we thought in Franklin.
Likewise in Andrade v. State, 700 S.W.2d 585 (Tex.Cr.App.1985), the defendant argued the trial court erred in failing to instruct the jury that, if sentenced to life imprisonment, he would not become eligible for parole until serving at least twenty calendar years. This *863Court rejected this contention on authority of its earlier decision in Williams v. State, 668 S.W.2d 692 (Tex.Cr.App.1983). But in Williams the Court had a different question before it. There the defendant had objected to an instruction that was submitted to the jury cautioning it not to consider “how long this defendant would be required to serve to satisfy a sentence of life imprisonment.” We held this instruction was not erroneous. That a capital jury may properly be admonished not to speculate as to the length of time a defendant will actually serve before he is paroled, however, does not necessarily mean that same jury may not also be told that he will have to serve a certain actual calendar time before he even becomes eligible for parole. The two instructions are not mutually exclusive, and propriety of the former does not logically establish impropriety of the latter. In my view this is a critical distinction, but one which has yet to be examined by the Court. Williams does not support the holding in Andrade any more than O’Bryan supported the holding in Franklin.
In Knox v. State, 744 S.W.2d 53 (Tex.Cr. App.1987), the defendant requested an instruction like the one we held Andrade was not entitled to. His request was denied. On appeal he invited this Court to “revisit” its holding in Andrade in view of the promulgation of Article 37.07, § 4, supra. We rightly observed that this provision on its face does not apply to capital eases. But Knox argued further that an instruction on the minimum period a capital convict sentenced to life imprisonment must serve before becoming eligible for parole “is necessary to guide the jury in answering the [future dangerousness] special issue.” The Court’s response to this argument is a puzzle. We said:
“[I]f it is clear that Article 37.07, § 4(a), does not apply to capital felonies, then it is also clear that jurors in capital cases should focus solely on the special issues submitted to them during the punishment phase. * * * Since jurors do not actually ‘sentence’ a defendant in Texas capital cases, their attention should be directed only to answering the special issues without regard to the sentence that will ultimately be imposed.”
Id., at 64 (emphasis in the original). It is of course true that the jury should focus on resolving the special issues at the punishment phase of a capital case, since that is its only function at that juncture. It is equally true that the jury should answer the special issues without regard for the outcome. In short, the jury’s role is simply to find facts. To point this out, however, was hardly sufficient to answer Knox’s contention. He argued that his requested instruction should have been given because minimum parole eligibility is in fact relevant to the jury’s fact-finding function insofar as it relates to the issue of future dangerousness. We did not agree or disagree with this contention. We apparently ignored it.
We were once again confronted with a claim, in Jones v. State, 843 S.W.2d 487, at 495 (Tex.Cr.App.1992), that a capital defendant should have been allowed to inform the jury that “he would be confined a minimum of 20 years” in the event he were sentenced to life imprisonment. Citing O’Bryan and its progeny for the familiar but not-dispositive proposition that parole in general is not a proper consideration for capital juries, we overruled Jones’ claim. Once again we failed squarely to address the contention that knowledge of the minimum parole eligibility date would be relevant to the jury’s resolution of the second special issue. In other cases we have rejected generic claims that capital juries should have been instructed on “the parole laws.” E.g., Stoker v. State, 788 S.W.2d 1, at 16 (Tex.Cr.App.1989); Elliott v. State, 858 S.W.2d 478, at 489-490 (Tex.Cr. App.1993); Garcia v. State, supra. We have yet, however, to offer a coherent explanation of why the minimum period a capital convict who is assessed a life sentence must serve to become eligible for parole should not be relevant to the issue of future dangerousness. Nor have we explained precisely how an instruction on minimum parole eligibility would violate Article II, § 1, supra. In my view the information is relevant, and inform*864ing the jury of it would not violate separation of powers.
c.
Relevance
Evidence is relevant, under Tex.R.Cr. Evid., Rule 401, if it has “any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence.” Moreover, “[a]ll relevant evidence is admissible,” except as otherwise provided by, inter alia, the state constitution. Tex.R.Cr.Evid., Rule 402. If the fact that appellant must serve at least fifteen years of a life sentence has any tendency to make more or less probable that he will commit future acts of violence so as to constitute a continuing threat to society, then that fact cannot be excluded from the jury’s consideration, at least not on the basis that it lacks relevance.
Is minimum parole eligibility relevant in this sense? It seems to me that, at least in combination with other evidence, it often is. We have consistently held that “society” is to be understood in its common acceptation in ordinary English, and that as such, its meaning is not limited either to the “society” existing outside the walls of the Texas Department of Criminal Justice, Institutional Division, or to that “society” that exists within those walls. We have construed “society” to embrace both. See, e.g., Sterling v. State, 830 S.W.2d 114, at 120, n. 5 (Tex.Cr.App. 1992); Caldwell v. State, 818 S.W.2d 790, at 798-99 (Tex.Cr.App.1991); Boyd v. State, supra, at 118, n. 12; Rougeau v. State, 738 S.W.2d 651, at 660 (Tex.Cr.App.1987). In predicting “the fact of consequence,” viz: whether a particular capital defendant will constitute a continuing threat to “society,” it would undoubtedly aid the jury to know which aspect of society he will be exposed to, and when.
We often see evidence, for example, that a capital defendant presents a much greater threat to the “society” at large than to the prison population. Sometimes, as in the instant case, there is evidence that in time the accused will not present much of a threat to either the prison population or the general public. See also, e.g., Matson v. State, 819 S.W.2d 839 (Tex.Cr.App.1991). In combination with this kind of evidence, knowing that the accused, if assessed a life sentence, will spend at least a certain minimum period of time incarcerated helps the jury to decide the probability that he will be a continuing threat to society, because it helps the jurors identify where he will be. Thus they can tell what portion of society may or may not be at risk. If the defendant must spend at least fifteen years in prison to satisfy a capital sentence of life imprisonment, the jury must decide the probability he will pose a threat to the prison population for at least that long, and then, the probability he will threaten the general public after that. Moreover, if there is evidence, as here, that in fifteen years the accused will not constitute a significant danger either inside or outside prison walls, the fact that he will be incarcerated until then tends to minimize the possibility of his ever constituting a threat, at least to the public at large. In short, appellant’s location, to the extent it can be foretold with relative certainty, is an evidentiary fact from which the elemental fact, whether he will be a continuing threat to “society,” may be, in part, derived. This evidentiary fact is “indisputably relevant” even under the provisions of our rules of criminal evidence.
Separation of Powers
Unless imparting this relevant evidentiary fact to the jury would violate separation of powers under Sanders, then, it ought to be accessible to the factfinder. In my view, the danger we noted in Sanders simply does not pertain to the question whether a capital jury should be informed how long a defendant who is sentenced to life imprisonment will have to serve before he becomes eligible for parole. Jury consideration of the minimum parole eligibility date as it relates to the issue of future dangerousness is manifestly a judicial function. It does not encroach upon the executive prerogative at all.
*865At the punishment phase of a non-capital criminal case, the jury’s function is not, as in a capital case, to make discrete findings of fact. Instead, the jury decides what term of punishment is appropriate within the statutorily prescribed range. “Deciding what punishment to assess is a normative process, not intrinsically factbound.” Murphy v. State, 777 S.W.2d 44, at 63 (Tex.Cr.App.1988) (Plurality opinion on State’s motion for rehearing). There is always the danger in a non-capital punishment proceeding that the jury, cognizant of the parole law, will abuse its wide discretion to decide what punishment is appropriate by imposing a sentence that takes the possibility of parole into account, and tries to adjust for it.3 This encroaches unconstitutionally upon the executive function in that it attempts to anticipate and circumscribe exercise of the executive authority to grant parole. Sanders v. State, supra, at 351.
In a capital punishment proceeding the jury’s role is more limited. The capital jury does not “assess punishment.” Rather, it determines the existence of discrete facts, such as the probability the accused will commit criminal acts of violence that would constitute a continuing threat to society. Punishment is assessed according to the jury’s resolution of these fact issues. See former Article 37.071(e), V.A.C.C.P. Capital jurors have no occasion to anticipate and circumscribe the executive parole authority. Because they do not “assess punishment” at all, they cannot assess death instead of a life sentence in order to prevent the eventual release of the accused on parole. Capital jurors just find facts. They weigh relevant information and answer special issues accordingly. Thus, when capital jurors consider the minimum parole eligibility date of a capital defendant serving a life sentence, they only use that information to more accurately gauge the probability of his future dangerousness — manifestly a judicial function. They do not try to predict when he will actually be released on parole.4 Indeed, they do not pretermit, second-guess or otherwise interfere with any power or prerogative of the executive branch in any way. It therefore cannot be said that they violate the principle of separation of powers articulated in Sanders.
D.
An instruction on minimum parole eligibility in a capital case does not offend Article II, § 1, supra. Our handful of cases tacitly assuming that it does, such as Franklin, An-drade, and Knox, all supra, ought to be disapproved. Moreover, that Article 37.07, § 4, supra, does not expressly authorize a parole instruction in a capital case does not mean a capital jury cannot be informed as to minimum parole eligibility. That provision, in tandem with the amendment to Article IV, § 11, supra, was meant to authorize parole instructions notwithstanding any violation of separation of powers. Because informing a *866capital jury on minimum parole eligibility does not violate separation of powers in the first instance, express legislative authorization for an instruction, pursuant to Article IV, § 11, supra, is unnecessary.
Thus, there is no constitutional or statutory impediment in Texas to informing capital jurors of the minimum period of time an accused sentenced to life imprisonment for a capital offense must serve to become eligible for parole.5 There is no real “policy” against it, as the plurality perceives. Minimum parole eligibility is certainly relevant, under Rules 401 and 402 of the Rules of Criminal Evidence, and any evidence or instruction informing the jury of it ought to be permitted for that simple reason, if no other. Moreover, federal due process may require it, under Simmons v. South Carolina, supra. If so, there is no conflict with Texas law.
III.
Appellant contends there is in fact such a federal due process requirement. It is less than evident, reading Simmons on its face, that due process requires informing a capital jury of the statutory minimum period of service on a life sentence before parole becomes available under the law. On its face Simmons seems only to require informing the jury that parole is unavailable whenever that is true as a matter of state law and the issue of future dangerousness has been injected into the case. We have no provision for life-scms-parole in Texas. Nevertheless, logically extending the due process principle announced in Simmons leads me inexorably to conclude that the trial court erred in failing to allow appellant to inform the jury as he requested to do in this case.
Speaking for a plurality in Simmons, Justice Blackmun announced at the outset:
“that where the defendant’s future dangerousness is at issue, and state law prohibits the defendant’s release on parole, due process requires that the sentencing jury be informed that the defendant is parole ineligible.”
— U.S. at -, 114 S.Ct. at 2190, 129 L.Ed.2d at 138. Because our legislature has made future dangerousness one of the statutory criteria for imposition of the death penalty, it is “at issue” in every capital punishment proceeding in which the State seeks a sentence of death. Thus, it appears that the first prong of the Simmons due process test is always met in Texas. However, under Texas law capital murder defendants sentenced to life imprisonment are not ineligible for parole. Justice O’Connor, in her concurring opinion, strongly suggests that the due process principle extends only when parole is not an available alternative.6 Several jurisdictions have already interpreted Simmons to be limited in this fashion.7 In the instant case, without any analysis, the plurality does *867too. Op. at-. On that basis the plurality concludes that the holding in Simmons has no application in Texas. In my view, however, the principle of Simmons transcends its factual context, and does apply in Texas in some instances.
A
Future dangerousness is not automatically at issue in South Carolina as it is in Texas. In Simmons, however, the prosecutor had expressly invited the jury in his closing argument to factor the defendant’s future dangerousness into its decision whether to assess life or death. Id., — U.S. at-, 114 S.Ct. at 2190-2191, 129 L.Ed.2d at 139.8 The defendant had proffered evidence that he was only a threat to elderly women, a class of people he would not encounter in the penitentiary. Id., — U.S. at-, 114 S.Ct. at 2191, 129 L.Ed.2d at 139. Out of the jury’s presence he presented public-opinion evidence that jury-eligible adults in South Carolina largely assumed a life-sentenced murderer would be paroled at some point. Id., — U.S. at -, 114 S.Ct. at 2191, 129 L.Ed.2d at 140. Nevertheless, Simmons’ counsel was prevented from voir diring the jury panel on the issue of parole eligibility. Arguing that the public misperception about parole eligibility needed to be dispelled in his cause, Simmons requested jury instructions that would inform the jury of the law. The trial court refused, instructing the jury instead that it should disregard parole and parole eligibility, and that it should construe “life imprisonment” according to its ordinary meaning. Id., — U.S. at-, 114 S.Ct. at 2192, 129 L.Ed.2d at 140.
In holding that the trial court violated due process in failing to inform the jury of the unavailability of parole, the plurality in Simmons observed:
“In assessing future dangerousness, the actual duration of the defendant’s prison sentence is indisputably relevant. Holding all other factors constant, it is entirely reasonable for a sentencing jury to view a defendant who is eligible for parole as a greater threat to society than a defendant who is not. Indeed, there may be no greater assurance of a defendant’s future nondangerousness to the public than the fact that he never will be released on parole. The trial court’s refusal to apprise the jury of information so crucial to its sentencing determination, particularly when the prosecution alluded to the defendant’s future dangerousness in its argument to the jury, cannot be reconciled with our well established precedents interpreting the Due Process Clause.”
Id., — U.S. at-, 114 S.Ct. at 2194, 129 L.Ed.2d at 142-43. Accordingly, the plurality held that denying the jury “a straight answer” regarding availability of parole impaired Simmons’ ability to “deny or explain” the State’s showing of his future dangerousness, and thereby violated his right to due process of law. Id., — U.S. at-, 114 S.Ct. at 2195 & 2196, 129 L.Ed.2d at 144 & 145-46, quoting Gardner v. Florida, 430 U.S. 349, at 362, 97 S.Ct. 1197, at 1207, 51 L.Ed.2d 393, at 404r405 (1977).
Simmons thus stands for the proposition that whenever future dangerousness is in issue in a capital punishment proceeding “the actual duration of the defendant’s prison sentence is indisputably relevant.” Id.,-U.S. at-, 114 S.Ct. at 2194, 129 L.Ed.2d at 142. When parole is not an option at all, that fact is relevant in the sense that it tells the jurors where the capital defendant will spend the rest of his natural life, so that they may confine their deliberations on future dangerousness to the risk he poses to prison society. It tends to make less probable that he will be a continuing threat at least to the general public. Moreover, to the extent the capital *868defendant can show, as could Simmons himself, that he does not pose a threat even to the inmate population, unavailability of parole tends to make less probable that he will be a threat to any facet of society at all.
This logic is not limited to the no-parole situation. On the contrary, it extends to any evidence of “the actual duration of the defendant’s prison sentence.” How much of his life the capital defendant will spend in prison is relevant to the future dangerousness determination in exactly the same way as the fact that he will spend the rest of his life there. It informs the jurors where the capital defendant will be so that they can tell what “society” may or may not be at risk. See Part IIC, ante. If you add other evidence that by the time the capital defendant becomes eligible for parole he will not likely pose a threat to any facet of society, then the fact that he must serve a specific minimum period of time in prison before he can be paroled takes on practically the same significance as no parole eligibility at all. It tends to show he will never be a threat to the public at large. Therefore, if failing to inform the jury of the “indisputably relevant” fact that a capital defendant can never be paroled violates due process, as the plurality opinion in Simmons establishes, then so, at least under some circumstances, does failing to inform the jury of the minimum parole eligibility date.9 We should so hold.
B.
The plurality mentions California v. Ramos, 463 U.S. 992, 103 S.Ct. 3446, 77 L.Ed.2d 1171 (1983), almost in passing. Op. at 849. The suggestion seems to be that when a state makes parole available to a capital defendant sentenced to life imprisonment, the Supreme Court will not interfere with the state’s decision whether to inform juries of the parole law. Admittedly, in Simmons the plurality itself did observe:
“that Ramos stands for the broad proposition that we generally will defer to a State’s determination as to what a jury should and should not be told about sentencing. In a State in which parole is available, how the jury’s knowledge of parole availability will affect the decision whether or not to impose the death penalty is speculative, and we shall not lightly second-guess a decision whether or not to inform a jury of information regarding parole. States reasonably may conclude that truthful information regarding the availability of commutation, pardon, and the like, should be kept from the jury in order to provide ‘greater protection in [the States’] criminal justice system than the Federal Constitution requires.’ [California v. Ramos], 463 U.S., at 1014, 103 S.Ct., at 3460.”
As I read them, however, neither Ramos itself, nor the plurality’s discussion of Ramos in Simmons, justifies failing to inform a capital jury of minimum parole eligibility.
In the first place, we have never “reasonably” concluded in Texas that information about parole eligibility should be kept from capital juries, at least not in order to “provide greater protection than the Federal Constitution requires.” We have basically assumed that jury consideration of parole in general would violate separation of powers, even in a capital case, and is therefore prohibited. That assumption does not “protect” a capital defendant at all, much less does it provide greater protection than the Federal Constitution. Quite the contrary, it insulates the jury from information that is “indisputably relevant” to rebut the State’s evidence *869on the issue of future dangerousness. In any event, the assumption itself is incorrect. It is precisely because information about minimum parole eligibility is relevant to the future dangerousness determination that giving it to the jury does not violate separation of powers. See Part IIC, ante.10 We have yet to offer any principled reason for keeping information about parole from a capital sentencing jury in Texas.
Second, and more importantly, today’s plurality cannot fairly plead “speculativeness” to argue that informing a capital jury about minimum parole eligibility does not fall within the Simmons due process requirement. There is nothing remotely speculative about the fact that, if sentenced to life imprisonment, appellant would have to serve a minimum of fifteen years before becoming eligible for parole. See former Article 42.18, § 8(b), supra. Nor is there anything speculative about “how the jury’s knowledge of parole availability will affect the decision whether or not to impose the death penalty.” Jurors in Texas do not decide whether to impose a life or death sentence. They answer special issues. That we can even say that minimum parole eligibility is “indisputably relevant” to the jury’s determination of the future dangerousness issue means we know precisely “how” the jurors will consider that fact. They consider it to the extent, and only to the extent, that it militates either for or against a finding of future dangerousness. When it militates against a finding of future dangerousness, Simmons dictates that the jury hear it, Ramos notwithstanding.11
C.
Finally, the plurality maintains that other procedural safeguards during the course of trial prevented the parties from placing parole eligibility in issue, and therefore appellant did not need to present the fact of minimum parole eligibility to “deny or explain” anything. Op at 850-853. Texas law precludes either party from broaching the subject of parole during voir dire, except to ascertain whether veniremen can follow instructions requiring jurors to ignore parole. It also prevents prosecutors from arguing expressly that the jury should answer special issues affirmatively in order to obviate the possibility a capital defendant may ever be paroled. Under Texas law the jury is positively instructed, as in this case, to disregard parole in its deliberations. Protections against jury misconduct guarantee that if his jury nevertheless does consider parole, the capital defendant may obtain a new trial. These legal safeguards, the plurality assures us, are sufficient to keep eligibility for parole from entering into the jury’s deliberations, and nothing is injected into the sentencer’s field of vision that the capital defendant is prevented from addressing. The plurality relies in large measure upon Gardner v. Florida, supra.
The plurality fails to acknowledge that the principle of due process articulated in Gardner was later expanded in Skipper v. South Carolina, 476 U.S. 1, 106 S.Ct. 1669, 90 L.Ed.2d 1 (1986), and in Simmons itself. It is true that the holding in Gardner itself extends no further than its facts. In Gardner the sentencing judge rejected an advisory jury’s recommendation of life imprisonment in favor of a sentence of death based in part upon a confidential presentence report that the defendant was given no chance to review. A plurality of the Supreme Court held that due process was violated because the defendant was afforded no chance to *870“deny or explain” the contents of the report, and hence no chance to participate meaningfully in the adversarial process by which Florida decides whether to impose a life or death sentence. Id., 480 U.S. at 360 & 362, 97 S.Ct. at 1206 & 1207, 61 L.Ed.2d at 403 & 404. But the principle in Gardner has since been extended beyond its factual context. In Skipper v. South Carolina, supra, the majority observed:
“Where the prosecution specifically relies on a prediction of future dangerousness in asking for the death penalty, it is not only [the Eighth Amendment] that requires that the defendant be afforded an opportunity to introduce evidence on this point; it is also the elemental due process requirement that a defendant not be sentenced to death ‘on the basis of information which he had no opportunity to deny or explain.’ Gardner v. Florida, [supra].”
IdL, 476 U.S. at 6, n. 1,106 S.Ct. at 1671, n. 1, 90 L.Ed.2d at 7, n. 1 (emphasis added).
On the strength of this passage from Skipper, the plurality in Simmons concluded that any time future dangerousness is put in issue, the defendant must be allowed to rebut with information that he can never be paroled, so that whatever risk he poses is limited to the prison population. — U.S. at -, 114 S.Ct. at 2194, 2194-2196 & n. 5,129 L.Ed.2d at 142-143, 143-144 & n. 5.12 Justice O’Connor’s concurring opinion agreed with the plurality that simply placing future dangerousness in issue is enough to trigger the due process requirement. Id., — U.S. at-, 114 S.Ct. at 2200-2201, 129 L.Ed.2d at 150-151. One of Justice Scalia’s principal complaints in dissent was that seven members of the Court had unduly imposed “a rule at least as sweeping as this: that the Due Process Clause overrides state law limiting the admissibility of information concerning parole whenever the prosecution argues future dangerousness.” Id., — U.S. at-, 114 S.Ct. at 2203, 129 L.Ed.2d at 164 (emphasis in the original). In Texas, of course, the prosecution always argues future dangerousness. If it is correct, therefore, that the Simmons principle extends also to informing the jury in some cases about minimum parole eligibility, see Part IIIA, ante, the plurality cannot escape that principle by reverting to the earlier decision in Gardner. For under Simmons, minimum parole eligibility may serve, together with other evidence, to “deny or explain” the State’s case for future dangerousness in every capital punishment proceeding in Texas.
Moreover, it does not satisfy Simmons to implement state procedures designed to insulate the jury entirely from any consideration of parole. Due process requires that a capital defendant be afforded “a meaningful opportunity to present a complete defense.” Id., — U.S. at-, 114 S.Ct. at 2194, 129 L.Ed.2d at 143, quoting Crane v. Kentucky, 476 U.S. 683, at 690, 106 S.Ct. 2142, at 2146, 90 L.Ed.2d 636, at 645 (1986). Preventing either side from broaching the subject of minimum parole eligibility when that fact is “indisputably relevant” to rebut the State’s evidence of future dangerousness denies a capital defendant that opportunity. Simmons, supra, — U.S. at-, 114 S.Ct. at 2198, 129 L.Ed.2d at 147.13 This denial might be constitutionally acceptable if it *871somehow served to enhance the overall fairness of the proceeding or to heighten the reliability of the jury’s resolution of the special issues. See Crane v. Kentucky, supra, 476 U.S. at 690, 106 S.Ct. at 2146, 90 L.Ed.2d at 644. But as I have labored to show in the balance of this opinion, it does not. The plurality errs to conclude that when procedural steps are taken to prevent a capital jury from considering parole for any purpose, the defendant’s due process rights as defined by Simmons have been preserved.
IV.
Turning to the instant cause, I am compelled to conclude that appellant was denied due process under Simmons. Dr. Fason testified that appellant was a sociopath, and that the likelihood of recidivism for sociopaths by the time they reach their thirties is low, and by the age of forty, even lower. Compare Matson v. State, supra. Had it been informed that by law appellant was not even eligible for parole on a life sentence until serving fifteen years, by which time he would be well into his upper thirties, the jury would have been justified in concluding that his threat to the public at large was insignificant.14 Without that information, the jury was probably left wondering exactly what the relevance of Dr. Fason’s testimony was. See n. 9, ante. Indeed, without that information, but given Dr. Fason’s testimony, the jury might have been tempted to disregard the trial court’s instruction not to consider parole for any purpose, and engaged in the speculative enterprise of trying to guess when appellant would become eligible for parole. Surely it is preferable to inform capital jurors accurately on the law of parole eligibility.
For the reasons I have given in Part III, ante, I conclude that due process requires that a capital defendant be allowed, on request, to inform the jury on the law of minimum parole eligibility, at least whenever there is other evidence, as in this ease, to indicate he may not constitute a continuing threat to society by the time he is parole-eligible. Nothing in the law of Texas should be construed to stand in the way. Part II, ante. Accordingly, the trial court erred in failing to allow appellant to inform the jury as he requested.
The Supreme Court did not indicate whether Simmons error is subject to a constitutional harmless error analysis. See Chapman v. California, 386 U.S. 18, 87 S.Ct. 824, 17 L.Ed.2d 705 (1967). The Simmons plurality simply reversed the cause and remanded it to South Carolina for unspecified further proceedings. Simmons, supra, - U.S. at-, 114 S.Ct. at 2198, 129 L.Ed.2d at 147. Assuming Simmons error is subject to a harm analysis, I cannot say to a level of confidence beyond a reasonable doubt that it was harmless in this case. Had the jury in appellant’s trial been informed that he could not become eligible for parole for fifteen years, they might well have considered that information, together with Dr. Fason’s testimony, sufficient to justify a negative answer to the special issue whether he would commit criminal acts of violence that would constitute a continuing threat to society.
Accordingly, I would vacate the trial court’s judgment in this cause and remand the cause for a new punishment proceeding.15 *872Because the plurality does not, I respectfully dissent.16
OVERSTREET, J., joins this opinion.. Appellant also attempted to elicit from Dr. Fa-son that convicted capital murderers serving life sentences in maximum security units have not generally proven to be violent. Counsel’s questions were inartful, however. The most he established was that Fason was "familiar with studies on that.” Fason did not say whether he believed those studies were accurate. Nor did he have personal knowledge that capital murderers sentenced to life would necessarily serve those sentences in maximum security units. In short, counsel was less than successful in establishing appellant would not pose a threat of future violence to the prison "society” if given a life sentence.
. In his concurring opinion Judge Mansfield opines that by providing that the particular instructions in Article 37.07, § 4, are not to be given in capital cases, the Legislature has “also determined that the jury in a capital case is not to be charged as to the law relating to parole and/or good time in any case in which the defendant has been convicted of a capital felony.” Op. at 857. The Legislature has made no such broad determination. Article 37.07 does not purport to cover punishment proceedings in capital cases. Such proceedings are governed by Article 37.071, V.A.C.C.P. That provision neither expressly authorizes nor expressly prohibits informing a jury as to minimum parole eligibility. But subsection (a) of 37.071 does authorize the trial court to admit any evidence it “deems relevant to sentence." Minimum parole eligibility is "indisputably relevant" to the issue of future dangerousness. See Part IIC, post. Moreover, were we to construe Article 37.07, § 4 to prohibit informing a capital jury of minimum parole eligibility in any case in which the defendant is able to present evidence he will present a significantly *862reduced threat to society by then, it would operate to deprive him of his right to present a complete defense, in violation of due process. See Part III, post. We should eschew potentially unconstitutional interpretations of a statute.
. Of course, under current Article 37.07, § 4(a), supra, this would only be an abuse of discretion if the jury tried to anticipate "the manner in which the parole law may be applied to [the] particular defendant" on trial. It is not now an abuse of discretion for the jury generally to “consider the existence of the parole law."
. In any case, it is not clear to me that jury consideration of when a capital defendant might actually be paroled would itself violate separation of powers. The actual date a life-sentenced inmate is paroled is undoubtedly relevant to the future dangerousness issue in the same way and to at least the same extent that minimum parole eligibility is. A capital punishment jury considering actual parole date would thus be simply performing a judicial factfinding role, and not attempting to circumscribe an executive function. The difference is, of course, that actual parole date is contingent upon future facts and actions of the Board of Pardons and Paroles, and cannot be determined with certainty as can the minimum parole eligibility date. That the actual parole date is speculative does not mean it is not relevant, however, or that jury consideration of same would violate separation of powers. The speculativeness of the enterprise might be reason to hold that evidence of actual parole date, while relevant and not violative of separation of powers, is nevertheless excludable because it is more prejudicial than probative. See Tex.R.Cr.Evid., Rule 403. We need not reach these issues in the present case, however, since appellant only requested that the jury be informed of the minimum parole eligibility date.
. In Rose v. State, supra, we also held that a jury instruction authorizing consideration of the existence of parole in a non-capital case violates due course of law under Article I, §§ 13 and 19 of the Texas Constitution. We found intolerable "the risk that punishment will be based on extraneous considerations.” Id., at 537. No such risk exists, however, when jurors in a capital punishment proceeding are informed of the minimum number of years a life-sentenced capital convict must serve to become eligible for parole. Minimum parole eligibility is not an “extraneous" consideration in determining the issue of future dangerousness. Indeed, as I hope I have demonstrated in Part IIC, ante, it is "indisputably relevant” to that determination. For this reason, informing the jury as appellant requested in this cause does not violate due course of law any more than it violates separation of powers.
. Justice O'Connor opined that:
"[D]espite our general deference to state decisions regarding what the jury should be told about sentencing, I agree that due process requires that the defendant be allowed to [bring his parole ineligibility to the jury’s attention] in cases in which the only available alternative sentence to death is life imprisonment without possibility of parole and the prosecution argues that the defendant will pose a threat to society in the future."
-U.S. at-, 114 S.Ct. at 2201, 129 L.Ed.2d at 151.
. E.g., Allridge v. Scott, 41 F.3d 213 (CA5 1994); State v. Price, 337 N.C. 756, 448 S.E.2d 827 (1994); Ramdass v. Commonwealth, 248 Va. 518, 450 S.E.2d 360 (1994).
. The plurality contends that the prosecutor in Simmons "urged the jury to vote for the death penalty because Simmons would be released from prison." Op. at 851 & n. 19. Literally this is not true, at least as far as the plurality opinion in Simmons discloses. Id.,-U.S. at-, 114 S.Ct. at 2190, 129 L.Ed.2d at 139. Although the prosecutor's having raised future dangerousness no doubt led the jury to ponder whether the defendant would be released from prison, and if so when, the plurality’s disposition in Simmons is not contingent on the fact that the prosecutor literally invited them to (if he did).
. In Matson v. State, 819 S.W.2d 839 (Tex.Cr. App.1991), we held the trial court violated the Eighth Amendment in failing to admit the defendant’s proffered evidence that recidivism rates are "extremely low" for inmates who serve long sentences their first time through the prison system. Absent some evidence that the capital defendant on trial will in fact be required to serve a relatively lengthy period of time incarcerated before he may be released, this evidence would serve no practical mitigating purpose vis -a -vis the special issues. If our holding in Matson is to make any sense at all, it is necessary also to inform the jury how long the accused will actually be imprisoned — or at least the minimum period we know he will be required by law to be imprisoned.
. Nor does informing the jury of minimum parole eligibility in a capital case violate due course of law. See n. 5, ante.
. It may well be the case that under Ramos we could prevent capital jurors from trying to predict when a life-sentenced defendant will actually be paroled. The actual parole date is, of course, purely speculative. Therefore, an instruction to capital jurors not to try to predict that date would not seem to violate due process, Simmons notwithstanding. This is not to say that the actual parole date is not relevant to the issue of future dangerousness — undoubtedly it is. See n. 4, ante. And while Ramos suggests that we could prohibit jury consideration of when a capital defendant sentenced to life imprisonment might actually be paroled consistent with due process, we have yet to articulate a good reason why we should. See Part II, ante.
. The plurality in Simmons also alluded to two other Supreme Court cases in support of the broader proposition that due process guarantees defendants "a meaningful opportunity to present a complete defense.” Id.,-U.S. at-, 114 S.Ct. at 2194, 129 L.Ed.2d at 143, quoting Crane v. Kentucky, 476 U.S. 683, at 690, 106 S.Ct. 2142, at 2146, 90 L.Ed.2d 636, at 645 (1986). See also, Ake v. Oklahoma, 470 U.S. 68, 105 S.Ct. 1087, 84 L.Ed.2d 53 (1985).
. In deciding whether the trial court’s instruction in Simmons had been adequate to protect the defendant’s due process rights, the plurality observed:
”[E]ven if the trial court’s instruction successfully prevented the jury from considering parole, petitioner’s due process rights were still not honored. Because petitioner’s future dangerousness was at issue, he was entitled to inforni the jury of his parole ineligibility. An instruction directing the jury not to consider the defendant’s likely conduct in prison would not have satisfied due process in Skipper, supra, and, for the same reasons, the instructions issued by the trial court in this case does not satisfy due process."
Id., — U.S. at --•, 114 S.Ct. at 2198, 129 L.Ed.2d at 147.
. It does not matter that appellant was unable to show he would not be a threat to prison society. See n. 1, ante. That the State can argue appellant may pose a danger to the prison population does not vitiate the relevance of other evidence that in time he will not pose a danger to anyone. As long as evidence has a tendency to show appellant will not pose a threat to some facet of society, due process requires that appellant be allowed to present it. Simmons, supra, -U.S. at-, 114 S.Ct. at 2194-2195, n. 5, 129 L.Ed.2d at 143-144, n. 5.
. The offense was committed in May of 1990. Because the trial court committed "error affecting punishment only," we remand the cause for a new punishment hearing only. See Articles 44.251 and 44.29(c), V.A.C.C.P., as amended by Acts 1993, 73rd Leg., ch. 781, § 5, p. 3059, eff. Aug. 30, 1993, which made fully retroactive the amendments in Acts 1991, 72nd Leg., ch. 838, § 2, p. 2900, eff. Sept. 1, 1991, notwithstanding § 5(a) of the 1991 amendment, which had originally made § 2 applicable only to offenses committed after September 1, 1991.
. Because I would remand for a new punishment hearing, I need not decide whether appellant’s other arguments claiming error affecting punishment only are meritorious. I must comment briefly, however, on the plurality's disposition of appellant’s sixteenth point of error. Op. at 842-843. The plurality acknowledges that the trial court should not have sustained the particular objection that the State leveled at trial to appellant's proffered evidence. Nevertheless, the plurality holds there was no error because the trial court could have ruled (though manifestly it did not) that the probative value of appellant’s proffered evidence was substantially outweighed by the danger of unfair prejudice, under Tex. R.Cr.Evid., Rule 403. There are so many problems with this disposition, one hardly knows where to begin. I will limit myself to two complaints. First, I would have thought it clear after Montgomery v. State, 810 S.W.2d 372 (Tex.Cr. App.1991) (Opinion on rehearing on Court's own motion), that it is the trial court’s job to decide in the first instance whether relevant evidence is nevertheless inadmissible under Rule 403, subject only to appellate review for abuse of discretion. Here this Court makes that determination for the first time on appeal. This brings me to a second complaint. The plurality justifies its disposition under the rubric, "if the trial court’s decision was correct on any theory of law applicable to the case....” Op. at 843. Since the State made no Rule 403 objection, I do not see how the theory that appellant's evidence was more-prejudicial-than-probative is “applicable to the case." What the plurality is really saying, it seems to me, is that ordinary notions of procedural default, as codified in Tex.R.App.Pro., Rule 52(a), do not apply to the State. The State need not object at trial or make the same objection on appeal that it did at trial in order to prevail on appeal. The plurality is willing enough to declare appellant's own arguments on appeal to have been procedurally barred. See, just as a convenient example, the plurality's disposition of appellant’s very next point of error, point seventeen. Op. at 844. This is a particularly unsightly juxtaposition, and one that does not speak well of the evenhandedness of our supposedly adversarial criminal justice system. As for the plurality’s disposition of appellant's eighteenth point of error, Op. at 853-855, complaining of the trial court’s tardy correction of its instructions to the jury at the punishment phase, I am dubitante.