Petitioner-Appellant David Allen Gardner appeals the district court's denial of his petition for a writ of habeas corpus pursuant to 28 U.S.C. 2254. He was convicted of capital murder in the course of a kidnaping and was sentenced to death. After exhausting his remedies at the state level, Gardner applied for a writ of habeas corpus which the district court denied. We granted a Certificate of Appealability (COA) limited to determining whether the State's psychiatrists' pre-examination warnings to Gardner were sufficient to ensure that his consent to be examined was "informed," thereby negating any potential violation of his Fifth Amendment right against compulsory self-incrimination that might otherwise have resulted from the punishment phase admission ---- over timely objection ---- of the assertedly prejudicial testimony of the psychiatrist who conducted that exam. As we conclude that this constitutional right was violated by the sentencing-phase admission of the testimony of the psychiatrist who examined Gardner on behalf of the State of Texas and that Gardner suffered prejudice from that violation, we reverse the decision of the district court, grant Gardner's petition for a writ of habeas corpus, and remand for entry of an appropriate judgment vacating his sentence and allowing the State a reasonable time within which to conduct a new, constitutionally valid sentencing proceeding or, alternatively, to resentence Gardner to life imprisonment in conformity with Texas law.

The facts of Gardner's crime of conviction are set forth in the opinion of the Texas Court of Criminal Appeals (CCA) disposing of Gardner's direct appeal.1 Gardner stopped and picked up a pair of fourteen-year old runaway hitchhikers, turned down a gravel road, and pulled off beside a bridge. After he told the teenagers to get out of the car, the three walked down an embankment where Gardner stabbed the male numerous times and left him for dead, then took the female to a nearby lake where he stabbed her numerous times, hit her in the head with a rock, and abandoned her as well. The male lived but the female died.

The state procedural history of Gardner's case is highly significant to our consideration today, so we review it in detail. During the sentencing phase of Gardner's murder trial, the State introduced only two evidentiary matters: (1) evidence that, some years prior to committing the instant crime, Gardner had fled the state of Kentucky after being released on bond while awaiting trial on two charges of theft and (2) testimony of Dr. Clay Griffith, who had, pursuant to a court order, conducted a pre-trial psychiatric evaluation of Gardner. After telling the jury that he had testified in "[p]robably three thousand" criminal trials,2 Dr. Griffith stated with "one hundred percent certainty" that, in his professional opinion, Gardner would "commit violent acts in the future," he was "super dangerous, and [he would] kill [again] given any chance at all." Dr. Griffith's testimony also included his professional opinion that Gardner would "continue to be violent even if placed in incarceration; and this would not prevent his violence and his brutality." And, added Dr. Griffith, Gardner "showed absolutely no remorse through the interview" and his tearful in-court confession of the murder was not credible because Gardner could "turn tears on and off" at will.

After his objection to the admission of Dr. Griffith's testimony was overruled and the psychiatrist was allowed to testify, defense counsel cross-examined Dr. Griffith extensively and also presented three favorable character witnesses on Gardner's behalf. Two of Gardner's former co-workers testified that he was a good employee and that they had never seen him exhibit any improper or violent conduct. The chief jailer of the Parker County Sheriff's Office, where Gardner was held while awaiting trial, testified that Gardner was a model prisoner who had never caused any problems.

The record shows that, even though Gardner was already represented by defense counsel, his attorney was not present either when Gardner consented to the psychiatric examination or at any time during the course of the examination itself. In fact, it is clear from the record that defense counsel had no knowledge that his client was to be examined and that the State made little or no effort to inform counsel in advance.3

During his testimony at the punishment phase of Gardner's trial, Dr. Griffith stated that he informed

the Defendant . . . what he was coming for, for a psychiatric examination; that this was ordered by Judge Hopkins. We informed him that [1] a report would have to be sent to the Court stating our findings so far as whether he was competent to stand trial, whether he, in our opinion, was sane or insane at the time of the alleged offense; [2] that in the State of Texas, there is no confidentiality so that anything that he might say could be used against him, or could be used for him at some later date in the courtroom (emphasis added).

Counsel for Gardner timely objected to the admission of Dr. Griffith's testimony at the punishment phase. In Gardner's direct appeal, counsel contested the admission of Dr. Griffith's testimony on the grounds that he had unlawfully induced Gardner's consent by telling him that the examination "could be used against him or could be used for him at some later date in the courtroom." Unpersuaded, the CCA affirmed Gardner's conviction and death sentence.

Gardner petitioned for a writ of habeas corpus in state court. He reiterated his objection to the admission of Dr. Griffith's testimony, this time emphasizing that the warnings given prior to the examination were constitutionally deficient under Estelle v. Smith4 because he was not adequately informed that the results of the exam could be used against him (1) during the punishment phase of the trial (2) to secure the death penalty.5 The state trial court, after entering its findings of fact and conclusions of law, recommended that habeas relief be denied. The CCA denied relief, stating that Gardner had already raised his Estelle v. Smith claim on direct appeal.

Gardner filed a second state habeas petition, stressing that he had not raised his Estelle v. Smith claim on direct appeal and that the earlier decision of the CCA was therefore erroneous. As a result, the CCA ordered a state trial court to conduct an evidentiary hearing to clarify the content of the warnings given to Gardner by Dr. Griffith prior to the psychiatric examination. At the hearing (held in 1995, fifteen years after Dr. Griffith's psychiatric examination of Gardner), Dr. James P. Grigson,6 who had aided Dr. Griffith in conducting the examination of Gardner, testified to what he (Grigson) had told Gardner before the examination:

Prior to the beginning of the examination I introduced myself, my name, explained that I was a medical doctor, a psychiatrist. Introduced Dr. Griffith. Also explained that he was a medical doctor, also a psychiatrist. I did read him the court order signed by Judge Hopkins. And then I explained to him that it was not confidential because we would be sending back a report discussing it. I explained to him that the motion had been filed by the district attorney . . . to have the examination. And the purpose was to examine him in three areas, competency, sanity, and dangerousness. And I explained to him at that time that competency did mean whether or not he had sufficient present mental ability to consult with his attorney with a reasonable degree of rational understanding, and whether he had a factual as well as a rational understanding regarding the proceedings against him. He told me he understood that. And then I explained sanity or insanity was defined as whether or not he was suffering from a severe mental disease or defect that prevented him from knowing the difference between right and wrong. And he understood that. I told him dangerousness meant whether or not he represented a continuing threat to society (emphasis added).

In denying Gardner's habeas petition for a second time, the CCA ruled that his Estelle v. Smith claim was procedurally barred and, in the alternative, that it was without merit.7 Gardner filed a motion for rehearing, pointing out that the CCA had again incorrectly recounted the procedural history to come up with the conclusion of procedural bar. In apparent recognition of its mistake, the CCA granted Gardner's motion and issued a new opinion which corrected the factual errors of the previous opinion; the court did not, however, formally withdraw its earlier opinion, instead leaving it "on the books."

The CCA's new opinion reaffirmed its prior ruling that Gardner's Estelle v. Smith claim was procedurally barred, but the court failed to address the merits of his constitutional claim at all. Inasmuch as (1) all parties to the case agree that the procedural bar rule used by the CCA was novel and thus inapplicable on federal habeas review, and (2) the perfunctory discussion of the merits of Gardner's Estelle v. Smith complaint was never withdrawn, the CCA's scant reasoning and ruling on the merits is what is before us today.

Having exhausted the remedies available at the state level, Gardner filed a petition for a writ of habeas corpus in federal district court, which was denied. He appealed that decision to this court, and we granted Gardner's application for a COA on his Fifth Amendment Estelle v. Smith claim.

As Gardner filed his federal petition for habeas review in 1998, well after the effective date of the 1996 Antiterrorism and Effective Death Penalty Act ("AEDPA"), we review his petition under the standards specified in that act.8 The AEDPA forbids us to issue a writ of habeas corpus with respect to "any claim that was adjudicated on the merits in State court proceedings" unless the state court's adjudication of that claim resulted in "a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States . . . ; or resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding."9 A decision is contrary to clearly established federal law "if the state court arrives at a conclusion opposite to that reached by [the Supreme Court] on a question of law or if the state court decides a case differently than [the] Court has on a set of materially indistinguishable facts."10 A decision is an unreasonable application of federal law "if the state court identifies the correct governing legal principle . . . but unreasonably applies that principle to the facts of the prisoner's case."11 Factual findings of the state court are presumed to be correct, so we defer to them "unless they were 'based on an unreasonable determination of the facts in light of the evidence presented in the state court proceeding.'"12

Gardner claims that his Fifth Amendment right against self-incrimination, as interpreted by the Supreme Court in Estelle v. Smith, was violated by the introduction of Dr. Griffith's testimony at the punishment phase of his trial. In Smith, the Supreme Court "held that a capital defendant's right against compelled self-incrimination precludes the state from subjecting him to a psychiatric examination concerning future dangerousness without first informing the defendant that he has a right to remain silent and that anything he says can be used against him at the sentencing proceeding."13 The warnings required under Miranda v. Arizona,14 ---- "including that [the defendant] has 'a right to remain silent' and that 'anything said can and will be used against the individual in court'"15 ---- are not sufficient to satisfy the more stringent requirements set forth in Estelle.16 To apprise a capital defendant fully of his Fifth Amendment rights before subjecting him to a court-ordered psychiatric examination, the defendant must be told that it will "be used to gather evidence necessary to decide whether, if convicted, he should be sentenced to death."17 We conclude that the warnings provided by Drs. Griffith and Grigson, whether viewed separately or in combination, were insufficient fully to apprise Gardner of his constitutional rights; moreover, we conclude that the CCA decision that held those warnings to be sufficient was an "unreasonable application of existing federal law."

The CCA made a factual determination that Dr. Griffith had made the warning in 1980 that he testified to during Gardner's 1981 trial, and that Dr. Grigson had made his warning in 1980 that he testified at the 1995 evidentiary hearing to having made. Although, given the vast number of trials at which Dr. Grigson testified, we find remarkable his ability to remember his specific warning to Gardner some fifteen years after the fact, we are constrained by the AEDPA to conclude that the CCA's factual findings that the doctors "said what they said they said" are not unreasonable and thus must be accorded the specified deference.

On the strength of these factual findings, the CCA made two rulings on the merits of Gardner's Smith claim. In the first, the CCA held that Dr. Griffith's warnings to Gardner that statements he would make in the course of the examination "could be used against him . . . at some later date in the courtroom" "sufficiently informed [Gardner] that his statements could be used against him at the punishment stage of his capital murder trial since that went on 'in the courtroom'" and that "a warning that a statement 'may be used against' a defendant conveys that the statement could be used at the punishment stage of a capital murder trial."18 The CCA's second ruling added that Dr. Grigson's warnings, in combination with those given by Dr. Griffith, were clearly sufficient under Estelle v. Smith.19 At the 1995 evidentiary hearing, Dr. Grigson testified that in 1980 he had informed Gardner that he was being examined for "dangerousness" which, Grigson explained to Gardner, meant "whether or not he represented a continuing threat to society." The CCA concluded that, in conjunction with Dr. Griffith's statement, "this more than complies with Estelle v. Smith."20

As noted, the AEDPA mandates that habeas petitions be granted only if the State court adjudication of the claim either "resulted in a decision that [1] was contrary to, or [2] involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States."21 Being disjunctive, each of these two prongs is to be accorded independent meaning, so habeas relief can be granted if the prisoner prevails on either prong.22 Although we cannot say that the decision of the CCA was "contrary to" the law as established in Estelle v. Smith because the CCA clearly did apply the correct legal rule to the pertinent facts, the CCA's application of that rule obviously produced an incorrect result. Thus our relevant inquiry is whether the CCA's application "of clearly established Federal law, as determined by the Supreme Court of the United States"23 produced a result that is not merely wrong but is so wrong that it is "unreasonable." We conclude that it did.

The Supreme Court, in the recent case of Williams v. Taylor,24 clarified the standard of review of habeas petitions under the AEDPA. In her majority opinion, Justice O'Connor does not purport to define the term "reasonable" but does offer useful guidance. Her opinion first makes clear that the standard is an objective one, specifically rejecting25 our previously employed, subjective "all reasonable jurists" standard.26 She then describes a relatively broad range along the "reasonableness" continuum at any point on which a state court decision might be held to be an "unreasonable application of Federal law": To be unreasonable, the state decision must be more than merely incorrect but can be something less than the stringent "all reasonable jurists" standard (under which the mere fact that reasonable jurists may disagree about the result requires the state court decision to be upheld).27

Although we have addressed Williams's "unreasonable application" rule on several occasions, we have done little to clarify the Supreme Court's standard in that case.28 Other circuits have addressed the issue too and in several instances have offered helpful clarifications of the Williams standard. For instance, the Tenth Circuit, after noting the importance of the objective nature of the standard, held that "the fact that one court or even a few courts have applied the precedent in the same manner to close facts does not make the state court decision 'reasonable.'"29

The Ninth Circuit in Van Tran v. Lindsey,30 held that the Williams "unreasonable application" standard "generally allows for reversals only where the court of appeals is left with a 'definite and firm conviction' that an error has been committed."31 As the Van Tran court went on to explain,we must reverse a state court's decision as involving an 'unreasonable application' of clearly established federal law when our independent review of the legal question does not merely allow us ultimately to conclude that the petitioner has the better of two reasonable legal arguments, but rather leaves us with a 'firm conviction' that one answer, the one rejected by the [state] court, was correct and the other, the application of the federal law that the court adopted, was erroneous.32

We also note another insightful observation made in Van Tran: The Ninth Circuit emphasized that in Williams the Supreme Court

rejected the interpretation, adopted in various forms by the Fourth, Fifth, Seventh, and Eleventh Circuits, that defines reasonableness on the basis of whether 'reasonable jurists' could disagree about the result reached by the state court. Instead, the Court adopted an 'objectively unreasonable' standard, employing language used in decisions by the Third and Eight Circuits.33

This is doubly significant when viewed in the context of the Third and Eighth Circuits' adoption of the same test as Williams because both circuits found the "reasonable jurists" standard to be too deferential to state courts, clearly implying that the Supreme Court preferred a more stringent habeas review of state court decisions.

Although Williams teaches that state court decisions should not be reversed merely because they are incorrect ---- i.e., just because we would have reached a different conclusion ---- Justice O'Connor's opinion makes equally clear that neither should such decisions be upheld when we conclude that the state court has not just misapplied the law to the facts but has done so in an objectively unreasonable manner. Stated another way, even though the AEDPA requires the federal courts to show more deference to state court decisions than they would in a de novo review, this cannot be interpreted to mean that an "objectively unreasonable" application of federal law should be allowed to stand. Even though we cannot reverse a decision merely because we would reach a different outcome, we must reverse when we conclude that the state court decision applies the correct legal rule to a given set of facts in a manner that is so patently incorrect as to be "unreasonable."

That is clearly the case here. The CCA first held that Dr. Griffith's warning to Gardner that his statements during the examination "could be used against him . . . at some later date in the courtroom" "sufficiently informed [Gardner] that his statements could be used against him at the punishment stage of his capital murder trial since that went on 'in the courtroom.'"34 This warning ---- given well in advance of trial, to a layman with no legal training, out of the presence of his counsel ---- simply cannot be stretched to the point of having "apprise[d] [Gardner] of his rights" and allowed him "knowingly [to] decide to waive them."35 Not, at least, when the rights in question are those recognized in Estelle v. Smith as clearly requiring warnings to the defendant that the adverse use in question means use (1) at the punishment stage (2) to accomplish the state's goal of obtaining the death penalty.

Many events take place in a courtroom during the protracted course of a criminal proceeding ---- the arraignment, various evidentiary hearings, the guilt/innocence phase of the trial ---- before and in addition to the sentencing phase of the trial. Such a vague reference ---- "in the courtroom" ---- cannot possibly suffice to fulfill Estelle v. Smith's strict requirement that the defendant be informed that his words and the results of the psychiatric examination can and will be used against him at the sentencing phase to secure the death penalty. In like manner, the general phrase "may be used against him" in no way narrows or identifies the point in the criminal proceeding at which Gardner's statements and the results of the examination could and would be used against him or for what specific purpose.

We repeat for emphasis that Dr. Griffith's vague and ambiguous words could reasonably be interpreted by an uninitiated layman to mean that (1) the results of the examination could be used (a) by the court (no mention of the prosecution) (b) to determine mental competency to stand trial; and (2) the defendant's own statements (no mention of the test results or the examining psychiatrist's opinions) could be used against him in court.36 The CCA then supplemented its holding with an alternative conclusion: When taken together, the warnings given by Drs. Griffith and Grigson gave Gardner adequate notice of his rights under Estelle v. Smith. Although Dr. Grigson's warnings ---- specifically his reference to determining Gardner's "dangerousness," as explained to mean whether Gardner "represented a continuing threat to society" ---- might be viewed by some as slightly more informative than those given by Dr. Griffith, it is patently unreasonable to say that they meet the standards of Estelle v. Smith. As Dr. Grigson testified, his warnings were gleaned from the language of opinions authored by a Texas state court judge and this court. An experienced defense counsel or even, perhaps, a career criminal well-versed in "jailhouse legalese," might recognize this language as a reference, however oblique, to the arcane terms of art in the Texas special issue of "future dangerousness" which in turn signal reference to the penalty phase of a capital trial. To most laymen, however, particularly unsophisticated and undereducated members of society with no legal training or experience,37 and unaccompanied by counsel, this language cannot reasonably be read to satisfy even minimally the strictures of Estelle v. Smith. Indeed, if layman such as Gardner could be expected to grasp the hidden significance of such legal "buzz words" and thus be deemed to have been adequately informed of their constitutional rights in such settings, then prophylactic warnings such as that mandated by Estelle v. Smith would be wholly unnecessary. We can only speculate that the CCA's extensive treatment of procedural bar, coupled with the relatively short shrift that it gave the Estelle v. Smith issue, might account for that court's otherwise inexplicable conclusion that the "warnings" given by Drs. Griffith and Grigson, long before the commencement of even the guilt/innocence phase of Gardner's trial, were sufficient under Estelle v. Smith even though neither doctor mentioned or even hinted at the possibility of the introduction of the results of the examination at the punishment phase of the trial to secure a sentence of death. Those elements are so clearly required by the Supreme Court in Estelle v. Smith and its progeny that they are indispensable elements to a conclusion of informed consent, itself an indispensable requisite for the waiver of such a basic constitutional right.

The State nevertheless contends that, even if the warnings given by Drs. Griffith and Grigson were inadequate to meet the Estelle v. Smith standard, Gardner was not prejudiced by the admission of Dr. Griffith's testimony at sentencing. We are well aware that we cannot grant habeas relief to a petitioner unless he can show that he suffered "actual prejudice" from the trial error at issue.38 Actual prejudice results when "the error had substantial and injurious effect or influence in determining the jury's verdict."39 Texas argues that Gardner was not prejudiced by the introduction of Dr. Griffith's testimony because (1) Dr. Griffith was thoroughly cross-examined at the sentencing phase of the trial, and (2) the evidence demonstrated that Gardner's crime was such a heinous and brutal one that the jury undoubtedly would have sentenced him to death, even if they had not been exposed to the testimony of Dr. Griffith. We disagree entirely.

Dr. Griffith's testimony was the centerpiece of the evidence presented by the State during the punishment phase of Gardner's trial. After being introduced to the jury as a medical expert with extensive experience in evaluating the future dangerousness of criminal defendants, Dr. Griffith testified, with "one hundred percent certainty," that Gardner would "commit violent acts in the future"; that he was "super dangerous, and [would] kill [again] given any chance at all"; and that he would be a danger to others even if incarcerated. "Would": not "might," not likely "would," but absolutely "would." Dr. Griffith further testified that Gardner exhibited no remorse for his crimes and that any behavior to the contrary should not to be believed.

Those words, spoken by a highly credentialed and experienced expert bearing the imprimatur of the State, constitute as great if not greater prejudice to Gardner than that suffered by the criminal defendants in Satterwhite v. Texas40 and Vanderbilt v. Collins,41 both cases in which Estelle v. Smith violations were found to have actually prejudiced the habeas petitioner. In each of those cases, the State bolstered the testimony of the psychiatric examiner by presenting many witnesses who testified to the bad character of the defendant. Here, Dr. Griffith was the one and only character witness presented by the State at the punishment phase. In addition, the substance of Dr. Griffith's testimony and the vehemence with which he presented it were at least as damaging, if not more so, than that of the psychiatric examiners in Satterwhite and Vanderbilt. In the words of our opinion in Vanderbilt, "it would strain credulity to conclude that Dr. [Griffith]'s testimony, which was quite lengthy and bore the imprimatur of an expert's opinion, did not have substantial, injurious effect on the outcome of [Gardner's] penalty phase."42 We are satisfied that this prejudice was in no way diminished, much less eliminated, by the valiant efforts of Gardner's trial counsel to mitigate through cross-examination the devastating effects of the expert testimony of the experienced and clearly biased psychiatrist for the State.

Likewise, the State's stereotypical fall-back argument ---- that the heinous and egregious nature of the crime would have ensured assessment of the death penalty even absent the psychiatric testimony about future dangerousness ---- cannot carry the day here. First, that argument cannot prevail without eviscerating the Supreme Court-approved Texas "special issues" scheme. To permit a jury to impose the death sentence solely because the facts are heinous and egregious would be to return to the days of inflicting capital punishment based on emotion and revenge, supplanting altogether the questions of deliberateness and future dangerousness which make the Texas scheme constitutional. Second, in this particular case, the details of the crime, as horrific as they are on an absolute scale, are not significantly more egregious than those in, for example, Vanderbilt.43 Except for there being a second teenage victim here (who survived), the crimes are amazingly parallel; yet the equally heinous facts in Vanderbilt were insufficient to negate prejudice. Finally, our decades of experience with scores of 2254 habeas cases from the death row of Texas teach an obvious lesson that is frequently overlooked: Almost without exception, the cases we see in which conviction of a capital crime has produced a death sentence arise from extremely egregious, heinous, and shocking facts. But, if that were all that is required to offset prejudicial legal error and convert it to harmless error, habeas relief based on evidentiary error in the punishment phase would virtually never be available, so testing for it would amount to a hollow judicial act. We are satisfied that here, Dr. Griffith's testimony cannot conceivably be said to have had no substantial, injurious effect on the outcome of the penalty phase of this case: There was Estelle v. Smith error and it was legally prejudicial.

As Estelle v. Smith teaches, the Fifth Amendment requires that the defendant in a capital trial who is subjected to a court-ordered psychiatric examination be informed that he is free to refuse to participate in that examination because its results can be used against him at the sentencing phase of the trial to secure the death penalty. Even though no magic words are required to be incanted talismanically, we nevertheless conclude that the "warnings" given here were so vague and ambiguous that it would not merely be erroneous but indisputably would be unreasonable to conclude that they could possibly have informed Gardner adequately, for purposes of satisfying Estelle v. Smith, that the psychiatric examination to be conducted by Dr. Griffith could and would be so used. We are thus satisfied that the CCA's conclusion ---- that the warnings given by Drs. Griffith and Grigson were sufficient under Estelle v. Smith ---- constitutes an "unreasonable application of federal law" to the facts that out of deference we are constrained to accept. We therefore reverse the decision of the district court and grant Gardner's petition for the writ of habeas corpus. Inherent in this holding is our conclusion that Gardner was actually prejudiced by this violation of his Fifth Amendment rights. The judgment of the district court is reversed and the case remanded for that court to enter an appropriate judgment directing the State of Texas either to (1) conduct a new sentencing proceeding within a reasonable time specified by the district court on remand, or (2) vacate Gardner's death sentence and impose the automatic life sentence specified by Texas law for a defendant who is convicted of capital murder but not sentenced to death.

REVERSED; Petition GRANTED; Case REMANDED with instructions.

I concur in the conclusion reached by the majority. Respectfully, however, I am unable to subscribe to its reasoning. I believe that ultimately the correct result in this case can be reached swiftly, without a prolix effort to further define "unreasonable."

Succinctly stated, this is the way I see this case: Estelle v. Smith, 451 U.S. 454 (1981), requires that, before undergoing a psychiatric examination concerning future dangerousness, a defendant must be "informed. . .that he has a right to remain silent and that anything he says can be used against him at the sentencing proceeding." Powell v. Texas, 492 U.S. 680, 681 (1989). The Texas Court of Criminal Appeals applied this legal principle in Gardner's case. The record shows that Gardner was advised of his right to remain silent and told that his statements during the psychiatric exam could be used for or against him in the courtroom to determine dangerousness.1 These statements constitute the undisputed facts to which the court of criminal appeals applied the Estelle legal principle. The court of criminal appeals determined that the warnings given Gardner were sufficient to comply with the requirements of Estelle.

Under Williams v. Taylor, 529 U.S. 362, 120 S. Ct. 1495, 1523 (2000), we are to grant the habeas writ only if the court of criminal appeals' determination was "an unreasonable application" of the Estelle principle. While Estelle requires that a defendant be warned that his statements could be used against him in a sentencing proceeding, Gardner was only told that his statements could be used "in the courtroom" to determine his dangerousness. As a matter of law, I believe the warnings given Gardner failed to convey the express message specifically required by Estelle as applied in death cases ---- that any statement a defendant makes could be used against him for the purposes of sentencing. The warning given Gardner only conveys that the statements could be used during the course of the trial, at whatever point in the trial dangerousness may become relevant. This broad warning does not convey the admonition that specifically addresses the sentencing phase so as to inform a reasonably minded defendant that what he says can be used against him to put him to death. Such specificity is required, as a matter of law, under Estelle. The court of criminal appeals, therefore, unreasonably applied the law when it determined that Gardner's warnings complied with Estelle.

Thus, I fail to see the relevance of the majority's repeated reference to Gardner's alleged status as an "uninitiated layman" and "unsophisticated and undereducated member[] of society." Is the majority saying that at some point a defendant's education renders a warning under Estelle unnecessary, or that the law applies differently to defendants based on their socio-economic and intellectual status? Or is the majority's emphasis on the fact that Gardner was not "experienced defense counsel" or "well-versed in jailhouse legalese" suggesting that the failure to give proper warnings under Estelle is subject to a harmless error exception if the defendant is an experienced attorney? In my view, the socio-economic and intellectual status of the defendant is irrelevant in a case like this, where the express statements fail, as a matter of law, to convey the warnings required under Estelle.

In sum, deciding the case in the way I suggest obviates the need to try further to define "unreasonable application" ---- a task undertaken by the majority with little success when it suggests that "we must reverse when we conclude that the state court decision applies the correct legal rule to a given set of facts in a manner that is so patently incorrect as to be 'unreasonable.'" The majority's analysis here is a tautology ---- it simply substitutes one protean phrase (patently incorrect application) for another (unreasonable application). In the end, the majority's lengthy journey to define "unreasonable" is a circular one, and we are left at the point at which we started. In my opinion, here we are better off not wandering down this road, especially when the excursion is unnecessary. Although I cannot subscribe to the majority's "unreasonable application" analysis, I respectfully concur in the conclusion reached by the majority.