Richard Gerry Drinkard, a Texas death row inmate, seeks a certificate of probable cause (“CPC”) to appeal the district court’s denial of his petition for a writ of habeas corpus. Construing his application for CPC as an application for a certificate of appeala-bility (“COA”) under 28 U.S.C. § 2253, as amended by section 102 of the Antiterrorism and Effective Death Penalty Act of 1996 (the “AEDPA”), Pub.L. No. 104-132, 110 Stat. 1214 (1996), we grant the COA because Drin-kard has made a substantial showing of the denial of a constitutional right.
Turning to the merits of his appeal, the central issue we decide today is whether a special instruction addressing temporary insanity caused by intoxication, which was given during the sentencing phase of Drinkard’s trial under section 8.04(b) of the Texas Penal Code, unconstitutionally prevented the jury from considering mitigating evidence of intoxication that did not rise to the level of temporary insanity. Based on our review of Drinkard’s appeal, we conclude that the special instruction did not have such an effect. Alternatively, and in view of the cogent dissent of Judge Garza, we are compelled to address the question whether 28 U.S.C. § 2254(d)(1), as amended by section 104(3) of the AEDPA, applies to our review of Drin-kard’s appeal. Holding that the AEDPA does apply, we conclude that it bars relief because the state court’s decision on Drin-kard’s claim was neither “contrary to, [n]or ... an unreasonable application of, clearly established Federal law, as determined by the Supreme Court.” AEDPA, § 104(3) (to be codified at 28 U.S.C. § 2254(d)(1)). We therefore affirm the district court’s denial of Drinkard’s habeas petition.
I
A Texas jury convicted Drinkard of capital murder in the deaths of Louann Anthony, Ladeen Hendrix, and Jerry Mullins.1 Evidence of Drinkard’s intoxication at the time of the murders was presented at both the guilt and the sentencing phases of his trial. At the close of the sentencing phase, the trial court submitted two special issues to the jury.2 The trial court gave the following general instruction concerning the two statutory special issues:
[I]n determining each of these Special Issues, you may take into consideration all of *755the evidence submitted to you in the full trial of the case, that is, all of the evidence submitted to you in the first part of this case wherein you were called upon to determine the guilt or innocence of the defendant, and all of the evidence, if any, admitted before you in the second part of the trial wherein you are called upon to determine the answers to Special Issues hereby submitted to you.
Over Drinkard’s objection, the trial court also gave the following special instruction after the general instruction:
Evidence of temporary insanity caused by intoxication may be introduced by the defendant in mitigation of the penalty attached to the offense for which he is being tried. Intoxication means disturbance of mental or physical capacity resulting from the introduction of any substance into the body. Temporary insanity caused by intoxication means that the defendant’s mental capacity was so disturbed from the introduction of the substance into the body that the defendant did not know that his conduct was wrong. Therefore, if you find that the defendant at the time of the commission of the offense for which he is on trial was temporarily insane as a result of intoxication, then you may take such condition into consideration in mitigation of penalty attached for the offense for which the defendant is being tried.3
The jury answered both special issues affirmatively, and the trial court sentenced Drin-kard to death. On direct appeal, the Texas Court of Criminal Appeals affirmed. Drinkard v. State, 776 S.W.2d 181 (Tex.Crim.App.1989). Drinkard did not petition the United States Supreme Court for writ of certiorari.
After being denied habeas relief by the Texas Court of Criminal Appeals, Drinkard filed a federal habeas petition, along with a motion to stay his execution. The district court granted the motion to stay and ordered the state to respond to Drinkard’s petition. After Drinkard filed an amended federal petition for habeas relief, the state filed a motion for summary judgment, and Drinkard filed a motion for partial summary judgment. The district court granted the state’s motion for summary judgment, denied Drinkard’s motion for partial summary judgment, and vacated the stay. Drinkard filed a notice of appeal and a motion for a CPC to appeal the district court’s denial of his petition. The district court denied the motion. Drinkard applied for a CPC with this court in September 1994, which was carried with this appeal. We granted an emergency motion for stay of execution in December 1995.
II
In determining whether a CPC should issue in this case, we must consider the question in the light of some relevant statutory amendments under the AEDPA. Before the President signed the AEDPA into law on April 24,1996, a petitioner could not appeal a district court’s ruling on a habeas petition that concerned detention arising from state court proceedings unless a district or circuit judge issued a CPC. 28 U.S.C. § 2253, amended by AEDPA, § 102; Fed.R.App.P. 22(b), amended by AEDPA, § 103. In Barefoot v. Estelle, 463 U.S. 880, 103 S.Ct. 3383, 77 L.Ed.2d 1090 (1983), the Court stated the standard governing the issuance of a CPC: the applicant must make “a substantial showing of the denial of a federal right.” Id. at 893, 103 S.Ct. at 3394-95 (emphasis added). A “substantial showing” requires the applicant to “demonstrate that the issues are debatable among jurists of reason; that a court could resolve the issues (in a different manner); or that the questions are adequate to deserve encouragement to proceed further.” Id. at 893 n. 4, 103 S.Ct. at 3394 n. 4 (internal citations and quotation marks omitted).
Section 102 of the AEDPA amended 28 U.S.C. § 2253 to require that a petitioner obtain a “certificate of appealability” from a circuit judge.4 AEDPA, § 102 (to be codified *756at 28 U.S.C. § 2253(c)(1)). Section 2253 now requires that a circuit judge issue a COA “only if the applicant has made a substantial showing of the denial of a constitutional right.” AEDPA, § 102 (to be codified at 28 U.S.C. § 2253(c)(2)) (emphasis added).
The Tenth Circuit recently held that “Congress drafted the plain language of the newly enacted § 2253(c)(2) to codify the Barefoot standard for issuance of a certificate of probable cause.” Lennox v. Evans, 87 F.3d 431, 434 (10th Cir.1996). Disagreeing with the Ninth Circuit’s holding in Williams v. Calderon, 83 F.3d 281 (9th Cir.1996),5 the court explained:
Although the Court [in Barefoot ] used the word “federal,” an applicant seeking a certificate of probable cause to appeal a district court’s denial of a § 2254 petition for writ of habeas corpus must demonstrate a substantial showing of constitutional error underlying the state conviction. We have always read the Barefoot standard to require a habeas petitioner to make a substantial showing of the denial of a federal constitutional right. Indeed, in the context of federal habeas review of a conviction entered in state court, it is the only intelligible reading.
87 F.3d at 434. We agree with the Tenth Circuit. Accord Reyes v. Keane, 90 F.3d 676, 679-80 (2d Cir.1996). “Because the standard governing the issuance of a certificate of appealability requires the same showing as that for obtaining a certificate of probable cause, application of § 102 of the [AEDPA] to Petitioner’s request for a certificate of probable cause would not constitute retroactive application of a statute under Landgraf [v. USI Film Products, 511 U.S. 244, 114 S.Ct. 1483, 128 L.Ed.2d 229 (1994)].” 87 F.3d at 434. We will therefore treat Drinkard’s application for CPC as an application for COA.
A
Drinkard first argues that the jury instruction concerning “temporary insanity caused by intoxication” given during the penalty phase of his trial violated the Eighth Amendment. Drinkard contends that the jury charge precluded the jury from considering evidence of lesser degrees of intoxication in mitigation of his sentence.6
(1)
The Eighth Amendment requires an individualized determination of sentencing in death penalty cases, based on the character of the defendant, the record of the defendant, and the circumstances of the offense. Woodson v. North Carolina, 428 U.S. 280, 304, 96 S.Ct. 2978, 2991, 49 L.Ed.2d 944 (1976) (plurality opinion). In Lockett v. Ohio, 438 U.S. 586, 98 S.Ct. 2954, 57 L.Ed.2d 973 (1978), the Supreme Court reversed a death sentence on Eighth Amendment grounds because the Ohio death penalty statute limited the consideration of mitigating evidence. According to Lockett, a statute cannot constitutionally preclude a sentencer “from considering, as a mitigating factor, any aspect of a defendant’s character or record and any of the circumstances of the offense that the defendant proffers as a basis for a sentence less than death.” Id. at 604, 98 S.Ct. at 2964 (plurality opinion). In Eddings v. Oklahoma, 455 U.S. 104, 102 S.Ct. 869, 71 L.Ed.2d 1 (1982), a majority of the Court embraced *757Lockett’s plurality rule in striking down a death sentence on Eighth Amendment grounds because the trial judge limited his consideration of mitigating evidence. According to Eddings, a sentencer cannot “refuse to consider, as a matter of law, any relevant mitigating evidence.”7 Id. at 114, 102 S.Ct. at 877. The Eddings Court additionally noted that the trial judge’s actions were “as if the trial judge had instructed the jury to disregard the mitigating evidence.” Id., at 114,102 S.Ct. at 877.
Drinkard does not, and could not, argue that the Texas special issues standing alone prevented the jury from considering his intoxication at the time of the offense. Lackey v. Scott, 28 F.3d 486, 489 (5th Cir.1994) (“[T]he Texas sentencing scheme does not preclude the jury from giving mitigating effect to evidence of a defendant’s voluntary intoxication at the time of the offense”), cert. denied, — U.S. -, 115 S.Ct. 743, 130 L.Ed.2d 644 (1995); Cordova v. Collins, 953 F.2d 167, 170 (5th Cir.) (same), cert. denied, 502 U.S. 1067, 112 S.Ct. 959, 117 L.Ed.2d 125 (1992); Kelly v. Lynaugh, 862 F.2d 1126, 1133 (5th Cir.1988) (same), cert. denied, 492 U.S. 925, 109 S.Ct. 3263, 106 L.Ed.2d 608 (1989). Instead, he challenges the effect of the special instruction on the special issues. Drinkard argues that the challenged instruction on “temporary insanity caused by intoxication” prevented the jury from considering and giving effect to evidence of his intoxication if the jury concluded that it did not rise to the level of temporary insanity, evidence that the jury otherwise could have considered through the two special issues standing alone.8
The proper standard for reviewing a challenged jury instruction in the capital sentencing context is “whether there is a reasonable likelihood that the jury has applied the challenged instruction in a way that prevents the consideration of constitutionally relevant evidence.” Boyde v. California, 494 U.S. 370, 380, 110 S.Ct. 1190, 1198, 108 L.Ed.2d 316 (1990). This “reasonable likelihood” standard does not require the petitioner to prove that the jury “more likely than not” interpreted the challenged instruction in an impermissible way; however, the petitioner must demonstrate more than “only a possibility” of an impermissible interpretation. Id. at 380, 110 S.Ct. at 1198. We must analyze the challenged jury instruction within the context of the overall jury charge. Cupp v. Naughten, 414 U.S. 141, 146-47, 94 S.Ct. 396, 400, 38 L.Ed.2d 368 (1973). “In evaluating the instructions, we do not engage in a technical parsing of this language of the instructions, but instead approach the instructions in the same way that the jury would — with a ‘commonsense understanding of the instructions in the light of all that has taken place at the trial.’ ” Johnson v. Texas, 509 U.S. 350, 368, 113 S.Ct. 2658, 2669, 125 L.Ed.2d 290 (1993) (quoting 494 U.S. at 381, 110 S.Ct. at 1198).9
*758(a)
Focusing initially on the language of the challenged instruction itself, we conclude that there is not a “reasonable likelihood” that the jury applied it so as to place consideration of non-insane intoxication beyond its effective reach.10 The instruction reads:
Evidence of temporary insanity caused by intoxication may be introduced by the defendant in mitigation of the penalty attached to the offense for which he is being tried. Intoxication means disturbance of mental or physical capacity resulting from the introduction of any substance into the body. Temporary insanity caused by intoxication means that the defendant’s mental capacity was so disturbed from the introduction of the substance into the body that the defendant did not know that his conduct was wrong. Therefore, if you find that the defendant at the time of the commission of the offense for which he is on trial was temporarily insane as a result of intoxication, then you may take such condition into consideration in mitigation of penalty attached for the offense for which the defendant is being tried.
In attempting to understand the significance of this instruction to Drinkard’s claim of a deprivation of a constitutional right, we must consider whether there is a reasonable likelihood that this instruction, within its four corners, actually precluded the jury from considering Drinkard’s non-insane intoxication as a mitigating factor. We must first set out, therefore, what the instruction actually states.
The first sentence clearly indicates that the instruction is about temporary insanity caused by intoxication not about intoxication in general. It reads “[e]vidence of temporary insanity caused by intoxication,” not “evidence of intoxication.” (Emphasis added.) The second sentence defines “intoxication” as the “disturbance of mental or physical capacity resulting from the introduction of any substance into the body.” According to the third sentence, “temporary insanity caused by intoxication means that the defendant’s mental capacity was so disturbed from the introduction of a substance into his body that the defendant did not know that his conduct was wrong. ” (Emphasis added.) The instruction concludes, “Therefore, if you find that the defendant at the time of the commission of the offense for which he is on trial was temporarily insane as a result of intoxication, then you may take such condition into consideration in mitigation of the penalty_” (Emphasis added.) This concluding sentence directs the sentencer to take into account a defendant’s “temporary insanity caused by intoxication” if it meets the definition contained in the preceding sentence.
The instruction effectively tells the jury how to go about evaluating a defendant’s claim that, at the time he committed the crime, his intoxication rendered him temporarily insane; that is, that because of his temporary insanity caused by intoxication; he could not have deliberately caused the death of the deceased—a specific response to the first special issue under the Texas capital sentencing scheme, which asks whether the conduct was deliberate and whether it was committed “with the reasonable expectation that the death of the deceased or another would result.”11 The instruction thus represents a permissible attempt to structure how *759the sentencing jury will consider a particular mitigating circumstance, namely, temporary insanity caused by intoxication. See, e.g., Boyde, 494 U.S. at 377, 110 S.Ct. at 1196 (1990) (“States are free to structure and shape consideration of mitigating evidence ‘in an effort to achieve a more equitable administration of the death penalty’ ” (quoting Franklin v. Lynaugh, 487 U.S. 164, 181, 108 S.Ct. 2320, 2331, 101 L.Ed.2d 155 (1988))). “In other words, the challenged special instruction invited the jury affirmatively to consider, as a mitigating factor, any evidence that the crime had been committed while Drinkard was temporarily insane as a result of intoxication.” Dist.Ct.Op., at 36.
Drinkard argues that the use and placement of the term “such condition” in the fourth sentence of the challenged instruction plausibly informs a jury that it can only consider intoxication (“such condition”) if it rises to the level of temporary insanity. We cannot agree, however, that there is a reasonable likelihood that the jury interpreted the term “such condition” as referring to the single word, “intoxication.” The focus of the instruction from the first is on “temporary insanity caused by intoxication” as a mitigating factor, not “intoxication” as a mitigating factor. Within the concluding sentence itself, “such condition” naturally refers, as a matter of grammatical construction, to the entire antecedent phrase, “temporary insanity caused by intoxication.” Thus, we cannot say there is a reasonable likelihood that the jury interpreted “such condition” as referring to a truncated part of the preceding phrase, i.e., “intoxication,” as opposed to the entire antecedent phrase, “temporary ihsanity caused by intoxication.”
Although we cannot say that there is not some remote possibility that the jury, as a whole, could have interpreted the instruction standing alone so as to preclude consideration of non-insane intoxication, or that a single, isolated, hypothetical “reasonable juror” could not have interpreted the instruction in such a manner,12 these are not the touchstones of our inquiry. Applying the Boyde standard, we simply cannot say that there is a reasonable likelihood that the jury as a whole, with “[djifferences ... in interpretation ... thrashed out in the deliberative process,” Boyde, 494 U.S. at 381, 110 S.Ct. at 1198, construed the instruction standing alone as precluding consideration of intoxication that did not rise to the level of temporary insanity.13
(b)
Turning to the instructions as a whole, Cupp v. Naughten, 414 U.S. at 146-47, 94 S.Ct. at 400, we cannot say that there is a reasonable likelihood that the jury interpreted the instructions as precluding the consideration of Drinkard’s intoxication if it did not rise to the level of temporary insanity. Prior to the challenged instruction, the trial court clearly and unambiguously charged the jury to consider all of the evidence in answering the special issues:
[I]n determining each of these Special Issues, you may take into consideration all of the evidence submitted to you in the full trial of the case, that is, all of the evidence *760submitted to you in the first part of this ease wherein you were called upon to determine the guilt or innocence of the defendant, and all of the evidence, if any, admitted before you in the second part of the trial wherein you are called upon to determine the answers to Special Issues hereby submitted to you.
(Emphasis added.) This general instruction necessarily and undeniably directed the jury to consider Drinkard’s evidence of intoxication in answering the special issues. The fact that the charge included this affirmative instruction to consider all the evidence strongly supports our conclusion that there is not a reasonable likelihood that the jury understood the instructions, as a whole, as precluding consideration of non-insane intoxication.
The inclusion of this general instruction in the charge also undercuts the possibility of concluding that there is a reasonable likelihood that the jury interpreted the existence of an explicit instruction mentioning intoxication in the context of temporary insanity as implicitly foreclosing the consideration of lesser forms of intoxication. In other words, we cannot say that it is reasonably likely that the jury, instructed to consider “evidence of temporary insanity caused by intoxication,” would interpret this instruction as meaning that it “could consider evidence of intoxication only if it produces temporary insanity.” This variation on the canon of statutory interpretation expressio unius — mentioning one thing implies the exclusion of another — is particularly inapt where the implication of exclusion flies in the face of an affirmative direction not to exclude consideration of any evidence. Cf. Blystone v. Pennsylvania, 494 U.S. 299, 308, 110 S.Ct. 1078, 1084, 108 L.Ed.2d 255 (1990) (holding that specific mitigating factor providing for consideration of “extreme” disturbance, “substantial” impairment, or “extreme” duress did not foreclose jury’s consideration of lesser degrees of disturbance, impairment, or duress because trial court “made clear to the jury that [list of statutory mitigating factors] were merely items it could consider” and trial court instructed jury that it could consider “any other mitigating matter”). Although the Court in Boyde explicitly distinguished “those instances where we have found broad descriptions of the evidence to be considered insufficient to cure statutes or instructions which clearly directed the sentencer to disregard evidence,” 494 U.S. at 384, 110 S.Ct. at 1200 (citing Hitchcock v. Dugger, 481 U.S. 393, 398-99, 107 S.Ct. 1821, 1824-25, 95 L.Ed.2d 347 (1987), and Lockett, 438 U.S. at 608, 98 S.Ct. at 2966),14 the challenged instruction in this case did not clearly direct the sentencer to disregard intoxication for all reasons except to the extent that it supported temporary insanity caused by intoxication.
(c)
Furthermore, the interplay between the challenged instruction and the special issues also leads us to conclude that there is not a reasonable likelihood that the jury applied the challenged instruction so as to preclude consideration of non-insane intoxication. The challenged instruction, by its own terms and as a matter of common sense, is relevant only to the first of the two special issues that the jury was required to answer under the Texas capital sentencing scheme and thus would have no effect on the jury’s consideration of the second special issue.15
*761The first special issue requires the jury to look back in time and determine whether the defendant acted deliberately in committing the murder. Tex.Code Crim.Proc.Ann. art. 37.071(b) (West 1981) (“[Wjhether the conduct of the defendant that caused the death of the deceased was committed deliberately and with the reasonable expectation that the death of the deceased or another would result.”). The second special issue requires the jury to look forward to the defendant’s future dangerousness. Art. 37.071(b) (“[Wjhether there is a probability that the defendant would commit criminal acts of violence that would constitute a continuing threat to society.”). The challenged instruction itself asks the jury to consider whether the defendant was temporarily insane (or, more specifically, “did not know his conduct was wrong”) as a result of intoxication “at the time of the commission of the offense,” The focus of the challenged instruction, like that of the first special issue, is backward looking to the time of the offense. We thus think that it is not reasonably likely that the jury would have applied the instruction to the second special issue. In other words, even if there is a reasonable likelihood that the jury somehow interpreted the challenged instruction as removing from its consideration evidence of Drinkard’s non-insane intoxication in answering the first special issue, we cannot say that there is a reasonable likelihood that the jury applied the challenged instruction to the second special issue so as to preclude consideration of evidence of non-insane intoxication in answering that issue.
As in Boyde, “[e]ven were the language ... less clear than we think, the context of the proceedings would have led reasonable jurors to believe that evidence of [Drinkard’s non-insane intoxication] could be considered in mitigation” in answering the second special issue. 494 U.S. at 383, 110 S.Ct. at 1199. In Boyde, the Court pointed to “[a]ll the defense evidence presented at the penalty phase” to support its conclusion that there was not a reasonable likelihood that the jury misinterpreted the instruction challenged in that case. Id. at 383-84, 110 S.Ct. at 1199-1200. On the other hand, in Graham, the Court emphasized that “both of Graham’s two defense lawyers vigorously urged the jury to answer ‘no’ to the special issues based on the evidence” in denying habeas relief. Graham v. Collins, 506 U.S. 461, 475, 113 S.Ct. 892, 902, 122 L.Ed.2d 260 (1993). Here, as in Graham, we examine the context within which the instruction was given — specifically, the arguments of Drinkard’s attorneys — to understand the effect of the instruction. Drinkard’s two attorneys described in great detail at the sentencing hearing how the jury could take into account Drinkard’s intoxication.16
At the sentencing hearing, the state waived its right to open. Thus, Drinkard’s attorneys, Mr. Heath and Mr. Taylor, argued first. Mr. Heath first introduced the subject of intoxication evidence:
I also want you to think about the long talks we had about intoxication. Each and every act of wrongdoing that Mr. Drinkard has committed since his release from the penitentiary at least by 1979 has been related to excessive intoxication. The incidents with his wives that were gone into by the State, you think about it. Every time the ultimate issue was Mr. Drinkard was extremely intoxicated when these occurred.
The evidence in this particular case was, at least by everyone that saw him, was that Mr. Drinkard was drinking heavily on the evening in question, and we are told in this jury charge that you can consider *762intoxication in mitigation of punishment, and I’m sure the first thought that comes to your mind is how are you going to do that in this ease?
Trial tr., vol. 36, at 5. Mr. Heath then related the intoxication evidence to special issue number one, arguing that intoxication had rendered Drinkard temporarily insane:
... Special issue number one talks about a deliberate act. I submit to you— and I’m still not convinced Mr. Drinkard by his own hand took all three of those lives. I’m still not convinced there weren’t other individuals involved.
[Objection by the state sustained.]
But where intoxication to the point of temporary insanity comes in is when we talk about an act deliberately done. That’s what logically comes to mind. Mr. Drin-kard was drunk to the point of temporary insanity. The State would want you to believe that Michael Watson was stumbling drunk that night but not Richard Drin-kard. It’s amazing. They.spent hours together drinking Schnapps, Miller Lite, and then Mr. Drinkard topped it off with a Mandrax.
Trial tr., vol. 36, at 5-6. Then Mr. Heath related the intoxication evidence to special issue number two, arguing that Drinkard would not be dangerous if not intoxicated:
One thing that you can utilize sitting as a juror is your common sense. Common sense dictates that on the night in question Mr. Drinkard was drunk out of his mind, and then let’s talk about this intoxication relationship to all of the offenses that Mr. Drinkard has committed. How does that tie in to issue number two? Real simply. Mr. Drinkard is not a dangerous individual when he is not intoxicated. I submit to you if Richard Drinkard spends a life sentence in the Texas Department of Corrections he is not going to get intoxicated, and if he’s not intoxicated he’s not dangerous. Think about it. Every offense that these individuals got up on the stand and talked about, every offense, a DWI, the BB gun incident, all the recent incidents were alcohol and drug-related.
Trial tr., vol. 36, at 6-7. After discussing other aspects of the case, Mr. Heath returned to the intoxication evidence to conclude his argument, again arguing temporary insanity with regard to special issue number one, and intoxication generally with regard to special issue number two:
I think the record is clear that Ricky Drin-kard was temporarily insane on the night in question, and I anticipate Mr. Gotschall standing up at sometime and arguing how can anyone commit acts like these, and I submit to you they can’t in their right mind. Mr. Drinkard was not in his right mind that night, and I beg each and every one of you to consider the facts and the evidence in this case, and if you do you will come to the proper conclusion, and that is that the acts were not done deliberately by reason of temporary insanity and further that if Mr. Drinkard is locked up, not allowed to take drugs and not allowed to drink to excess, he will not be a continuing threat to society. Thank you.
Trial tr., vol. 36, at 11. After Mr. Heath concluded his argument, Mr. Taylor further argued on behalf of Drinkard. He organized his discussion of the intoxication evidence in a manner similar to that argued by Mr. Heath. First, he introduced the subject of intoxication evidence:
Intoxication, alcohol, drugs is almost at an epidemic stage in our society. It is— constitutes a social disease, the toll of which is enormous, not only in this ease but in society in general. Just look at some of the people that have been on this witness stand. Look at Jerry Michael Watson. He contributes little or nothing to society. He works very little and sits around and gets drunk every day. Doug Bailey drinks every day.
You know, obviously Ricky Drinkard suffers from the social disease of alcohol and drugs. From the evidence, from Rick’s statements you know that on the night in question that there were at least two 12-packs of beer. Ricky in his statement, which is in evidence, which you can read again, stopped off on the way to his brother’s house, bought a 12-paek of beer. They consumed that. They consumed two pints of Schnapps. He went and bought another 12-pack of beer. They consumed *763that. There was marijuana. Then there was a Maudrax above all that.
Trial tr., vol. 36, at 13-14. Mr. Taylor then related the intoxication evidence to special issue number one, arguing that intoxication rendered Drinkard temporarily insane:
... I suggest to you, as Mr. Heath has talked to you about, that there’s no way that anybody can consume those quantities of alcohol, ingest drugs into their system and be conscious of what they’re doing, and there’s no way anyone under those circumstances can deliberately do anything.
The State would have you believe that Ricky Drinkard deliberately, intentionally with forethought, went to Louann Anthony’s townhouse to take the lives of at least two individuals, if not three, because they tried to elicit testimony that tried to show you that Ricky knew not only Louann Anthony would be there but her sister or cousin with her boyfriend; and they would have you believe he deliberately went there with the idea of killing three people. I suggest to you based on the evidence and based on alcohol and drugs there’s no way that Richard Drinkard could have in a moment of sanity been involved in such.
When you read and look at special issue number one, I suggest to you that the only possible answer that this jury could put down is “no.”
Trial tr., vol. 36, at 14-15. Mr. Taylor then related the intoxication evidence to special issue number two, arguing that Drinkard would not be dangerous if not intoxicated:
... Some of the acts that the State brought to you in punishment, the burglaries were all done by a young man 16 and 17 years of age; and after that the problem -with Ricky Drinkard has been alcohol and drugs. Mr. Heath — during voir dire you were told that when you read these questions, if you get that far, that some of the terms are not going to be defined for you. In fact, in the two special issues none of the terms are going to be defined for you. It’s whatever or however you wish to define it.
One of those terms was “society.” It can be whatever you want it to be. If Ricky Drinkard by your verdict received life imprisonment, his society is going to be prison life.
[Objection by the state overruled.]
And there are not drugs and there is not alcohol available in prison life, and I suggest to you that the social disease of alcohol and drugs are not going to be available to Ricky Drinkard in the society of prison life and that there’s no evidence in the record whatsoever that would have you answer special issue number two “yes.” Take away the drugs. Take away the social disease of alcohol. There’s no evidence of violence, and I suggest to you that a proper verdict with respect to special issue number two would be “no” based on the evidence and based on the law in this case.
Trial tr., vol. 36, at 15-16.
Drinkard’s two attorneys each explained in great detail exactly how the jury could account for intoxication in mitigation of Drin-kard’s sentence in answering both special issues. We think that their explanations would have led the jury to believe that it could consider Drinkard’s intoxication in answering the second special issue even if it did not rise to the level of temporary insanity as defined by the challenged instruction.17 At a *764minimum, then, we can say that there is not a reasonable likelihood that the jury applied the instructions so as to preclude consideration of lesser forms of intoxication in answering the second or “look-forward” issue.
(d)
In sum, our larger task is to assure that all relevant evidence that Drinkard submitted in mitigation of the death penalty was within the effective reach of the jury, so that it had some opportunity to consider that evidence and to give to it whatever mitigating effect it deemed appropriate. Reading the challenged instruction standing alone, in connection with the general instruction to consider all the evidence and the special issues themselves, and, finally, in the light of the proceedings at trial, specifically, the arguments of Drinkard’s attorneys, we simply cannot agree- with Drinkard that there is a reasonable likelihood the jury interpreted the instructions in such a way as to exclude consideration of his non-insane intoxication.
(2)
While this appeal was pending, the President signed the AEDPA into law. Title I of the AEDPA contains a series of amendments to existing federal habeas corpus law. The insistence of Judge Garza’s dissent compels an alternative holding in this case, which requires our determining whether the standards of review for state court decisions on the merits contained in 28 U.S.C.
§ 2254(d)(1), as amended by title I of the AEDPA, § 104(3), applies to our review of Drinkard’s appeal. Paragraph (d), as now amended, reads as follows:
(d) An application for a writ of habeas corpus on behalf of a person in custody pursuant to the judgment of a State court shall not be granted with respect to any claim that was adjudicated on the merits in State court proceedings unless the adjudication of the claim—
(1) 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
(2) 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.
AEDPA, § 104(3) (to be codified at 28 U.S.C. § 2254(d)) (emphasis added).
The state argues that the new standards of review contained in subsection (d)(1) apply to all habeas cases pending before us when the AEDPA was signed into law because they are jurisdictional and procedural in nature. On the other hand, Drinkard relies on the Tenth Circuit’s decision in Edens v. Hanni-gan, 87 F.3d 1109, 1112 n. 1 (10th Cir.1996), and a number of district court cases to argue that the standards of review do not apply to his appeal.18 For the following reasons, we *765agree with the state and hold that the new standards of review contained in § 2254(d)(1) apply to our review of Drinkard’s appeal.19
(a)
Landgraf v. USI Film Products, 511 U.S. 244, 114 S.Ct. 1483, 128 L.Ed.2d 229 (1994), provides the framework for answering the retroactivity question presented in this case. There, the Supreme Court addressed the circumstances under which statutory amendments apply to lawsuits based on events occurring before those amendments.20 The Court declared that when Congress has not “expressly prescribed the statute’s proper reach,” we must determine whether the new statute has a “retroactive effect,” 511 U.S. at -, 114 S.Ct. at 1505, that is, “whether the new provision attaches new legal consequences to events completed before its enactment.” 511 U.S. at -, 114 S.Ct. at 1499. In other words, the question is “whether [the statute] would impair rights a party possessed when he acted, increase a party’s *766liability for past conduct, or impose new duties with respect to transactions already completed.” 511 U.S. at -, 114 S.Ct. at 1505. If we conclude that the statute does not have a retroactive effect, we should apply the new statute in rendering a decision in the case before us. 511 U.S. at -, -, 114 S.Ct. at 1501, 1505.
(b)
Because Congress has not “expressly prescribed” the reach of the new habeas standard of review contained in § 2254(d)(1), as amended by § 104(3) of the AEDPA, Reyes v. Keane, 90 F.3d 676, 678-79 (2d Cir.1996), we must turn to determine whether the new standards of review contained in § 2254(d)(1), as amended by the AEDPA, have a retroactive effect in this ease. The Court in Landgmf explained, “The conclusion that a particular rule operates ‘retroactively1 comes at the end of a process of judgment concerning the nature and extent of the change in the law and the degree of connection between the operation of the new rule and a relevant past event.” 511 U.S. at -, 114 S.Ct. at 1499. The change in law at issue here has no plausible connection to Drinkard’s conduct on the night of the murder. Drinkard cannot argue that the new standards of review attach new legal consequences to that conduct by increasing his liability for that conduct or by imposing new duties on him based on that conduct. In other words, Drinkard obviously cannot argue that he relied on the existence of federal de novo review of claims adjudicated on the merits in state court proceedings the night he killed his three victims. This provision instead speaks to the power of the federal courts to grant habeas relief to state' prisoners.
As standards of review governing our own review of Drinkard’s appeal, subsection (d)(1) is easily classified as procedural in nature. Cf. United States v. Mejia, 844 F.2d 209, 211 (5th Cir.1988) (citation omitted) (“A change in the standard of review is properly characterized as procedural rather than substantive because it neither increases the punishment nor changes the elements of the offense or the facts that the government must prove at trial.”). Pointing to “the diminished reliance interest in matters of procedure” and the fact that “rules of procedure regulate secondary rather than primary conduct,” 511 U.S. at -, 114 S.Ct. at 1502, the Court in Land-graf recognized that “[cjhanges in procedural rules may often be applied in suits arising before their enactment without raising concerns about retroactivity.” 511 U.S. at -, 114 S.Ct. at 1502.
Here, the change in procedural rules governing federal habeas review raises no concerns of retroactivity. Because the new rules involve federal standards of review of state court decisions, Drinkard must be able to show that he relied to some extent on the former federal standards of habeas review in making strategic, tactical, or other decisions during the state court litigation. Although during his state post-conviction proceedings, Drinkard may well have expected that the federal courts would review claims adjudicated on the merits in those proceedings de novo, “[a] statute does not operate ‘retrospectively’ merely because it is applied in a ease arising from conduct antedating the statute’s enactment, or upsets expectations based in prior law.” 511 U.S. at -, 114 S.Ct. at 1499 (internal citation and footnote omitted). In short, Drinkard cannot argue credibly that he would have proceeded any differently during his state post-conviction proceedings had he known at the time of those proceedings that the federal courts would not review claims adjudicated on the merits in the state court proceedings de novo. Because the new standards of review do not have a retroactive effect, we hold that they apply to our review of Drinkard’s appeal from the district court’s denial of his petition for writ of habeas corpus. We thus turn to the task of applying these new standards to Drinkard’s appeal.
(3)
Drinkard turns the task of statutory interpretation on its head by arguing summarily that § 2254(d)(1), as amended, is essentially only a codification of Teague v. Lane, 489 U.S. 288, 109 S.Ct. 1060, 103 L.Ed.2d 334 (1989), and thus constitutes no change in *767federal habeas law.21 Instead, “[a]s with any statutory question, we begin with the language of the statute.” Matter of Greenway, 71 F.3d 1177, 1179 (5th Cir.) (citation omitted), cert. denied sub nom., Boyce v. Greenway, - U.S. -, 116 S.Ct. 2499, 135 L.Ed.2d 191 (1996).
(a)
Subsection (d) limits the ability of the federal courts to grant habeas relief to state prisoners:
(d) An application for a writ of habeas corpus on behalf of a person in custody pursuant to the judgment of a State court shall not be granted with respect to any claim that was adjudicated on the merits in State court proceedings unless the adjudication of the claim—
(1) 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
(2) 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.
AEDPA, § 104(3) (to be codified at 28 U.S.C. § 2254(d)) (emphasis added). It applies when a state prisoner is seeking relief on the basis of a “claim that was adjudicated on the merits in State court proceedings.” A federal court may grant habeas relief on the basis of such a claim only if the “decision” resulting from that adjudication (1) “was contrary to ... clearly established Federal law, as determined by the Supreme Court of the United States,” or (2) “involved an unreasonable application of[ ] clearly established Federal law, as determined by the Supreme Court of the United States.” Because a decision that is “contrary to” law is in some sense a decision “involv[ing] an unreasonable application of’ law, the language of subsection (d)(1) on its face suggests at least one reading that would render the first clause a nullity. We, however, must read these two clauses in such a way as to give effect to both. United States v. Nordic Village, Inc., 503 U.S. 30, 36, 112 S.Ct. 1011, 1015, 117 L.Ed.2d 181 (1992).
Our analysis of these two clauses begins with the fundamental proposition that judicial decisions rest on answers to one or more of three types of questions: questions of law, questions of fact, and mixed questions of law and fact (i.e., questions that require the application of law to facts). In order properly to understand section (d)(1), it should be read in conjunction with subsection (d)(2). See United Sav. Ass’n of Texas v. Timbers of Inwood Forest Associates, 484 U.S. 365, 371, 108 S.Ct. 626, 630, 98 L.Ed.2d 740 (1988) (indicating that we must read statute holistically, interpreting each of its portions in light of other portions). Although not at issue in this case, subsection (d)(2) of § 2254 applies to a state court’s factual determinations. It permits federal court relief if the state court adjudication of the claim “resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence.” AEDPA, § 104(3) (to be codified at 28 U.S.C. § 2254(d)(2)). Subsection (d)(2) thus supplies the applicable standard of review for the second type of question — a question of fact. It is clear to us, therefore, when the statute is read holistically, that subsection (d)(1) provides standards of review for questions of law and mixed questions of law and fact.
The second clause of subsection (d)(1), by its own language, refers to mixed questions of law and fact because it speaks of an “unreasonable application of[ ] clearly estab*768lished law.” Thus, when reviewing a mixed question of law and fact, a federal court may grant habeas relief only if it determines that the state court decision rested on “an unreasonable application of[] clearly established Federal law, as determined by the Supreme Court,” to the facts of the case. We read the first clause, on the other hand, as referring to questions of law. When reviewing a purely legal question, a federal court may grant habeas relief only if, it determines that a state court’s decision rested on a legal determination that was “contrary to ... clearly established Federal law, as determined by the Supreme Court.” Thus, the standard of review will vary depending on whether the question before the federal court is one of fact, one of law, or mixed.
With this understanding of the language of subsection (d)(2), we now proceed to apply it to Drinkard’s appeal.
(b)
In applying § 2254(d)(1), as amended by the AEDPA, we must first determine whether Drinkard’s claim regarding the challenged instruction during the sentencing phase of his trial was adjudicated on the merits in state court proceedings. Our review of the state post-conviction record indicates that there is no question that this claim was in fact adjudicated on the merits. Drin-kard’s petition for habeas relief in the state trial court challenged, inter alia, this instruction. The claim appeared in the state trial court’s order designating issues as one of five that “this Court will resolve.” The trial court entered findings of fact and conclusions of law, recommending to the Texas Court of Criminal Appeals that it should deny relief. In conclusion of law number 29, the court held:
The trial court’s instruction on the law of temporary insanity as a result of intoxication was sufficient to allow the jury to consider such in mitigation of punishment; evidence, if any, of voluntary intoxication could be given full mitigating effect within the scope of the special issues without additional jury instructions. Moreover, the trial court’s charge on the law of temporary insanity as a result of intoxication did not preclude the jury from considering other types of mitigating evidence, did not mandate additional instructions, and did not impermissibly limit the jury’s consideration of the applicant’s alleged voluntary intoxication by requiring that it rise to the level of temporary insanity.
(Internal citations omitted). The Court of Criminal Appeals denied relief based on “the findings and conclusions of the trial court.”22
We now must apply the new standards of review to determine whether we are permitted to grant relief to Drinkard under the AEDPA. The first question we ask is whether the state court’s resolution of any legal questions underlying its decision on this claim was contrary to clearly established federal law. It is clear from conclusion of law number 29 that the state court made no error involving purely legal questions. The court correctly determined the law applicable to Drinkard’s claim — that a sentencing jury cannot be precluded from considering any relevant mitigating evidence. Lockett; Ed-dings. We thus cannot say that the decision of state court was “contrary to” clearly established law as determined by the Supreme Court.
The next question before us is a mixed question of law and fact. In specific terms, we must decide whether the state court’s determination — that the special instruction on temporary insanity caused by intoxication did not place beyond the reach of the jury’s consideration the mitigating evidence of intoxication — involved an unreasonable application of this law to the facts of this case.
This “unreasonable application” standard of review of a state court decision must mean more than that a federal court may grant habeas relief based on its simple disagreement with the state court decision; this would amount to nothing more than a de novo review. See H.R.Conf.Rep. No. 518, 104th Cong., 2d Sess. 111 (1996), reprinted in 1996 U.S.C.C.A.N. 944, 944 (indicating in no uncertain terms that § 2254(d)(1) “requires *769deference to the determinations of state courts that are neither ‘contrary to,’ nor an ‘unreasonable application of,’ clearly established federal law” (emphasis added)). The use of the word “unreasonable” in formulating this restrictive standard of review implicitly denotes that federal courts must respect all reasonable decisions of state courts. Thus, given the statutory language, and in the light of legislative history that unequivocally establishes that Congress meant to enact deferential standards, we hold that an application of law to facts is unreasonable only when it can be said that reasonable jurists considering the question would be of one view that the state court ruling was incorrect. In other words, we can grant habeas relief only if a state court decision is so clearly incorrect that it would not be debatable among reasonable jurists.
In this case, the majority has applied the law of Lockett and Eddings, using the Boyde reasonable likelihood standard, to the specific facts of this case, analyzing the special instruction standing alone and in conjunction with the general instruction, the special issues, and the arguments of counsel. The majority has unequivocally concluded that the instruction at issue did not place mitigating evidence of intoxication beyond the reach of the jury. Judge Garza, on the other hand, has concluded that the challenged instruction removed the mitigating evidence of intoxication from the jury’s consideration. It follows that when the jurists considering the state court ruling disagree in this manner, the application of the law by the state court is not unreasonable. The AED-PA therefore bars us from granting relief to Drinkard on this claim.
B
Drinkard also contends that the trial court should have instructed the jury to consider convicting him on a lesser-ineluded offense. A defendant is entitled to a lesser-ineluded offense instruction only if the evidence warrants such an instruction. Beck v. Alabama, 447 U.S. 625, 685-38, 100 S.Ct. 2382, 2388-90, 65 L.Ed.2d 392 (1980); Cordova v. Lynaugh, 838 F.2d 764, 767 (5th Cir.), cert. denied, 486 U.S. 1061, 108 S.Ct. 2832, 100 L.Ed.2d 932 (1988). To support such a claim, a petitioner must make “a showing that the facts of the case and the laws of the State warrant such an instruction.” Andrews v. Collins, 21 F.3d 612, 629 (5th Cir.1994), cert. denied, — U.S. -, 115 S.Ct. 908, 130 L.Ed.2d 790 (1995). Drinkard makes no showing on appeal that such evidence was produced at trial. Accordingly, we find this claim to be without merit.23
C
Although Drinkard also challenged the constitutionality of the trial court’s jury instruction regarding voluntary intoxication given during the guilt-innocence phase of his trial,24 he conceded in supplemental briefing to this court that “the U.S. Supreme Court’s recent decision in Montana v. Egelhoff, [— U.S. -, 116 S.Ct. 2013, 135 L.Ed.2d 361 (1996) ], forecloses [his] challenge to Tex.Pen. Code sec. 8.04(a), under the Due Process Clause of the Fourteenth Amendment.”25
*770III
To sum up, we hold today that the standard for granting a certificate of appealability under the AEDPA is the same as the Barefoot standard for granting a CPC. Because Drinkard has made a substantial showing of the denial of a constitutional right with respect to the application of the special instruction on temporary insanity caused by intoxication during the sentencing phase, we GRANT Drinkard’s COA. We also hold that the special instruction on temporary insanity caused by intoxication given under § 8.04 of the Texas Penal Code did not violate Drin-kard’s Eighth Amendment rights by placing mitigating evidence of non-insane intoxication beyond the effective reach of the jury. We therefore AFFIRM the district court’s denial of habeas relief. In the alternative, we hold that the new federal standards of review contained in 28 U.S.C. § 2254(d)(1), as amended by § 104(3) of the AEDPA, do not have a retroactive effect and thus are applicable to habeas eases pending at the time the President signed the AEDPA into law. Applying those new standards of review to Drinkard’s appeal, we conclude that § 2254(d)(1) bars relief because the state court’s decision on Drinkard’s claim was neither “contrary to, [n]or ... an unreasonable application of, clearly established Federal law, as determined by the Supreme Court.” We therefore VACATE our earlier stay of execution.
COA GRANTED, Judgment AFFIRMED, and Stay VACATED.
. The three victims were murdered in Ms. Anthony’s home. All three victims received circular head wounds, consistent with wounds inflicted by the head of a carpenter’s hammer. In addition, Anthony was stabbed three times in the chest; Hendrix was stabbed several times in the back and abdomen; and Mullins was stabbed fifteen times in the back.
. At the time of Drinkard’s trial, the Texas Code of Criminal Procedure required the submission of the following issues to the jury:
(1)whether the conduct of the defendant that caused the death of the deceased was committed deliberately and with the reasonable expectation that the death of the deceased or another would result;
(2) whether there is a probability that the defendant would commit criminal acts of violence that would constitute a continuing threat to society; and
(3) if raised by the evidence, whether the conduct of the defendant in killing the deceased was unreasonable in response to the provocation, if any, by the deceased.
TexCode Crim.Proc Ann. art. 37.071(b) (West 1981). Since the issue of provocation was not “raised by the evidence," the third special issue was not submitted to Drinkard’s jury.
. This instruction was given pursuant to section 8.04(b) of the Texas Penal Code, which is a provision applicable to capital and non-capital cases in both the guilt and sentencing phases. See Tex.Penal Code Ann. § 8.04(b) (West 1994).
. There appears to be a discrepancy between the amended version of 28 U.S.C: § 2253 and the amended version of Rule 22(b) of the Federal Rules of Appellate Procedure. Section 103 of the AEDPA amended Rule 22(b) to reflect the change in terminology contained in § 2253. AEDPA, *756§ 103 (to be codified at Fed.R.AppP. 22(b)). Even after the amendment, however, Rule 22(b) still permits either a district or circuit judge to issue a COA, AEDPA, § 103 (to be codified at Fed.R.App.P. 22(b)), as opposed to only a circuit judge under § 2253(c)(1). The posture of this case obviates the need to address this discrepancy.
. The Ninth Circuit concluded summarily that the standard for obtaining a certificate of appeal-ability is “more demanding” than the Barefoot standard. Id. at 286. The court then “as-sumefd], without deciding, that section 2253(c)(2) does not apply retroactively.” Id.
. Although the state argues that we have already considered and rejected the challenge that Drin-kard raises in this case, our review of the cases cited by the state convinces us otherwise. See Lackey v. Scott, 28 F.3d 486 (5th Cir.1994), cert. denied, - U.S. -, 115 S.Ct. 743, 130 L.Ed.2d 644 (1995); Barnard v. Collins, 958 F.2d 634, 639 (5th Cir.1992), cert. denied, 506 U.S. 1057, 113 S.Ct. 990, 122 L.Ed.2d 142 (1993); James v. Collins, 987 F.2d 1116 (5th Cir.), cert. denied, 509 U.S. 947, 114 S.Ct. 30, 125 L.Ed.2d 780 (1993); and Cordova v. Collins, 953 F.2d 167 (5th Cir.), cert. denied, 502 U.S. 1067, 112 S.Ct. 959, 117 L.Ed.2d 125 (1992).
. See also Penry v. Lynaugh, 492 U.S. 302, 319, 109 S.Ct. 2934, 2947, 106 L.Ed.2d 256 (1989) C'[I]t is not enough simply to allow the defendant to present mitigating evidence to the sentence. The sentencer must also be able to consider and give effect to that evidence in imposing sentence.”).
. The prohibition against the announcement of new constitutional rules of criminal procedure on collateral review contained in the line of cases beginning with Teague v. Lane, 489 U.S. 288, 109 S.Ct. 1060, 103 L.Ed.2d 334 (1989), does not bar relief in this case. Granting the relief Drinkard requests would not be a "new rule” under the Teague line of cases becáuse it would represent the application of "a well-established constitutional principle to govern a case which is closely analogous to those which have been previously considered in the prior case law.” Penry, 492 U.S. at 319, 109 S.Ct. at 2947 (internal quotation marks and citation omitted); see also Wright v. West, 505 U.S. 277, 304, 112 S.Ct. 2482, 2497, 120 L.Ed.2d 225 (1992) (O’Connor, J., concurring in judgment) ("If a proffered factual distinction between the case under consideration and pre-existing precedent does not change the force with which the precedent's underlying principle applies, the distinction is not meaningful_”).
.Although parts of the following analysis may appear contrary to Johnson’s admonition, the dispute between the majority and the dissent on the possibility of an impermissible interpretation requires our close examination of the challenged instruction. In any case, our final conclusion does not rest upon parsing the language of the instruction, but instead upon a review of the instruction in the context of the instructions and special issues as a whole, and in the light of the’ proceedings before the jury.
. Evidence that Drinkard was intoxicated at the time of the murders is clearly "constitutionally relevant." Evidence implicates the Eighth Amendment under Lockett and Eddings if it concerns “any of the circumstances of the offense that the defendant proffers as a basis for a sentence less than death.” Lockett, 438 U.S. at 604, 98 S.Ct. at 2964. As argued by Drinkard’s counsel, a jury could find that a defendant who was intoxicated at the time of the commission of a dangerous offense would not be dangerous in prison, where alcohol is not available. See Parker v. Dugger, 498 U.S. 308, 314, 111 S.Ct. 731, 736, 112 L.Ed.2d 812 (1991) (stating that evidence that defendant "was under the influence of large amounts of alcohol and various drugs ... during the murders” was mitigating evidence); Skipper v. South Carolina, 476 U.S. 1, 5, 106 S.Ct. 1669, 1671, 90 L.Ed.2d 1 (1986), the Supreme Court held that evidence concerning the defendant’s good behavior in jail while awaiting trial was mitigating evidence, on the theory that a jury could opt to impose life in prison instead of a death sentence if convinced that the defendant would not be dangerous in prison.
. See supra note 2.
. Prior to Boyde's “reasonable likelihood" standard, we judged jury instructions in this context by the “reasonable juror” standard. See California v. Brown, 479 U.S. 538, 541, 107 S.Ct. 837, 839, 93 L.Ed.2d 934 (1987).
. We are unpersuaded by dicta in Tucker v. State, 771 S.W.2d 523 (Tex.Crim.App.1988), which suggests that the plain language of the special instruction does not allow a jury to consider mitigating evidence of intoxication unless it rises to the level of temporary insanity. Id. at 534 & n. 4. The Texas Court of Criminal Appeals itself has not followed this dicta in subsequent cases. See, e.g., Ex parte Rogers, 819 S.W.2d 533, 536-37 (Tex.Crim.App.1991) (Clinton, J., dissenting) (citing Tucker to argue that court should have granted petitioner relief because special instruction on temporary insanity did not enable jury to give effect to non-insane intoxication).
We are similarly unpersuaded by the dissent in Nethery v. Collins, 993 F.2d 1154 (5th Cir.1993), which argued that the special instruction precluded consideration of non-insane intoxication based on the "reasonable juror” standard. Id. at 1163-65 (King, J., dissenting). Our holding rests on the application of the more stringent "reasonable likelihood” standard. Finally, although some language in our recent decision in Rogers v. Scott, 70 F.3d 340 (5th Cir.1995), possibly could be read to support a contrary conclusion, id. at 343-44, the court clearly did not reach the ultimate question before us today. Id. at 344.
. In Hitchcock, the petitioner challenged the jury charge given to an advisory jury at the penalty phase of his capital murder trial. The charge, pursuant to Florida statute, listed seven specific mitigating factors for the jury to consider. 481 U.S. at 396 n. 3, 107 S.Ct. at 1823 n. 3. The judge instructed the jury that " '[t]he mitigating circumstances which you may consider shall be the following [list of statutory mitigating circumstances].’ ” Id. at 1824, 107 S.Ct. at 1824 (quoting Tr. of Advisory Sentence) (second alteration added). The petitioner argued that none of the seven factors allowed the jury to consider evidence of his background, character, and potential for rehabilitation in mitigation of penalty. Id. at 396-98, 107 S.Ct. at 1823-24. The Court found that "it could not be clearer that the advisory jury was instructed not to consider, and the sentencing judge refused to consider, evidence of nonstatutory mitigating circumstances,” and held that the petitioner’s death sentence violated the Eighth Amendment. Id. at 398-99, 107 S.Ct. at 1824-25.
. The Texas sentencing scheme does not violate the Constitution if a jury can give effect to a particular type of mitigating evidence only when answering one of the special issues. See Graham *761v. Collins, 506 U.S. 461, 475-76, 113 S.Ct. 892, 902, 122 L.Ed.2d 260 (1993) ("Even if Graham’s evidence, like Penry's, had significance beyond the scope of the first special issue, it is apparent that Graham’s evidence — unlike Peniy's — had mitigating relevance to the second special issue concerning his likely future dangerousness.... This distinction leads us to conclude that neither Penry nor any of its predecessors "dictates ” the relief Graham seeks within the meaning required by Teague.”).
. The arguments of counsel are relevant to a jury's interpretation of challenged jury instructions, but the court's instructions cany substan-tiaEy more weight. 494 U.S. at 384-85, 110 S.Ct. at 1200. To the extent relevant, the arguments of counsel must also be analyzed in context. Id. at 385, 110 S.Ct. at 1200.
. The prosecutor mentioned the relevance of the intoxication evidence to the second issue only in passing. His focus was instead on the intoxication evidence as it related to the first special issue:
The Defense talks to you about this issue of temporary insanity due to intoxication, and I suppose that comes in mostly — they connected up somehow with both special issues, but to consider that at all — and I suggest after you look at the evidence you won’t consider that at all. To consider that at all you have to decide, one, that at the time of the deaths Mr. Drin-kard was intoxicated. This is 3:00 o'clock in the morning. The — -Mike Watson testified when he dropped his brother off it was around midnight or so or when he last saw his brother it was around midnight. There was obviously drinking and marijuana smoking and that sort of thing. We don’t have any doubt that Mr. Drinkard was intoxicated. That’s not the question.
You have to decide from the evidence, one, whether Mr. Drinkard was intoxicated and, *764two, that by reason of this voluntary intoxication he didn’t basically know right from wrong, he didn’t know what he was doing when he killed these three people was wrong. Okay? You might find, well, maybe he wouldn't have hit him so many times if he wasn’t drunk. That doesn’t make any difference. You have to find that his intoxication rendered him to such a state that the defendant — in the charge, the defendant did not know that his conduct was wrong, and we know that’s not true, because look at the evidence as to what he did after he killed these three people.
Trial tr., vol. 36, at 22-23 (emphasis added). We do not think that this single statement negates the voluminous arguments of Drinkard’s attorneys concerning intoxication, as it relates to answering the second special issue, in determining whether there is a reasonable likelihood that the jury interpreted the charge in such a way as to preclude consideration of non-insane intoxication. See Boyde, 494 U.S. at 385, 110 S.Ct. at 1200 (" ‘[A] court should not lightly infer that a prosecutor intends an ambiguous remark to have its most damaging meaning or that a juty, sitting through a lengthy exhortation, will draw that meaning from the plethora of less damaging interpretations.' ” (quoting Donnelly v. DeChristoforo, 416 U.S. 637, 647, 94 S.Ct. 1868, 1873, 40 L.Ed.2d 431 (1974))).
. We must say that we are unpersuaded by the cases cited by Drinkard. The Tenth Circuit in a footnote concluded summarily that the new ha-beas provisions do not apply because only § 107 contains an effective date provision. Id. In Grady v. Artuz, 931 F.Supp. 1048, 1054 n. 1 (S.D.N.Y.1996), the court summarily concluded that the new provisions do not apply for the same reason. Id. (citing United States v. Trevino, No. 96 C 828, 1996 WL 252570, at *2 n. 1 (N.D.Ill. May 10, 1996)). In Trevino, the district court concluded that the traditional presumption *765against retroactivity applies because the new provisions would have "a truly retroactive effect.” 1996 WL 252570, at *2 n. 1 (citing Maitland v. University of Minnesota, 43 F.3d 357, 363 (8th Cir.1994)). The court makes no attempt to explain why the provision at issue in Maitland, an amendment to Title VII barring a person from challenging an action taken pursuant to consent decrees if that person had actual notice of a proposed consent decree and a reasonable opportunity to participate, 43 F.3d at 361, is analogous to the new habeas provisions for retroactivity purposes. Finally, Drinkard cites Warner v. United States, 926 F.Supp. 1387, 1390 n. 4 (E.D.Ark.1996). The court in Warner summarily concluded that the new provisions do not apply retroactively because only § 107 contains an effective date provision, and thus "[ajccordingly, the Court need not consider what effect, if any, the amendments ... might have in this case.” Id. (citations omitted).
Since Drinkard submitted his supplemental briefing on the AEDPA, the Second Circuit has also held that the habeas provisions do not apply to cases pending on appeal at the time of the enactment of the AEDPA. Boria v. Keane, 90 F.3d 36 (2d Cir.1996). We are also unconvinced by the Second Circuit's reasoning. The Second Circuit appears to have interpreted the following language in the Supreme Court's decision in Landgraf v. USI Film Products, 511 U.S. 244, -, 114 S.Ct. 1483, 1499, 128 L.Ed.2d 229 (1994), as requiring an outcome-determinative test to ascertain retroactivity: "[T]he court must ask whether the new [statute] attaches new legal consequences to events completed before its enactment.” The Second Circuit declared with no analysis, "Assuming ... that the new statute would require a different outcome [in this case], application of the new statute to these circumstances would be retroactive." 90 F.3d at 37. Once it determined that the statute was retroactive, the court looked for a "clear signal from Congress” that the habeas provisions were to apply retroactively. Id. at 38. Finding none, the court held that the new habeas provisions did not apply to the case before it. Id. The Second Circuit, in one sentence, reduced the Supreme Court’s extended attempt in Landgraf “to reconcile two seemingly contradictory statements found in our decisions concerning the effect of intervening changes in the law,” 511 U.S. at -, 114 S.Ct. at 1496, to a simple test: if an intervening change in the law alters the outcome of a case before a court, it does not apply retroactively unless Congress has given some “clear signal" to the contrary. As much as the Second Circuit's proffered test would happily simplify the task facing courts in this area, it is not a correct synthesis of the applicable law. See infra.
. The Seventh Circuit utilized an analysis similar to the one that follows and reached this same conclusion in Lindh v. Murphy, 96 F.3d 856 (7th Cir.1996) (en banc). In Lindh, the Seventh Circuit first found that the AEDPA amendment to the federal habeas corpus provision lacked an effective date provision and so should be given effect with respect to pending appeals in the absence of any retroactive impact. Id. at 862. The court then held that under Landgraf's established framework for determining retroactivity of a statute, the amendments were not retroactive and thus should be applied to cases pending on appeal at the time of the passage of the AEDPA. Id. at 867. Specifically, the court concluded that the amendments did not "impair rights a party possessed when he acted, increase a party’s liability for past conduct, or impose new duties with respect to transactions already completed," thus, the court found that the habeas provision, as amended, controlled consideration of the pending habeas petition. Id. (citing Landgraf v. USI Film Prods., 511 U.S. 244, -, 114 S.Ct. 1483, 1505, 128 L.Ed.2d 229 (1994)).
. In Landgraf, the Supreme Court was attempting to harmonize "two seemingly contradictory statements in [its] decisions concerning the effect of intervening changes in the law.” Id. at -, 114 S.Ct. at 1496. In Bradley v. School Board of City of Richmond, 416 U.S. 696, 711, 94 S.Ct. 2006, 2016, 40 L.Ed.2d 476 (1974), the Court declared, "[A] court is to apply the law in effect at the time it renders its decision." In Bowen v. Georgetown University Hosp., 488 U.S. 204, 208, 109 S.Ct. 468, 471, 102 L.Ed.2d 493 (1988), the Court declared, "[Congressional enactments and administrative rules will not be construed to have retroactive effect unless their language requires this result.” (Citations omitted).
. We see more than a little irony in the suggestion that, after all the years of failed attempts by Congress to adopt a deferential standard of review in this area, Wright v. West, 505 U.S. at 295 n. 9, 112 S.Ct. at 2491 n. 9 (opinion of Thomas, J.); 505 U.S. at 305-06, 112 S.Ct. at 2498 (O’Connor, J., concurring in judgment), the passage of subsection (d)(1) represents no more than the codification of existing Supreme Court precedent, the meaning of which even the Supreme Court at times has difficulty explaining in a coherent manner. See, e.g., Wright (exhibiting three different interpretations of the Teague line of cases in five opinions, none of which received a majority). We will not complicate the task of statutory interpretation before us by turning first to the murkiness that is Teague retroactivity doctrine to determine whether the language of the statute somehow parallels Supreme Court precedent in this area.
. In his brief to this court, Drinkard himself admits that "the Texas courts clearly addressed the merits of the claim " concerning the challenged instruction.
. Drinkard also challenges the factfinding procedure used by the state habeas court. However, as counsel conceded at oral argument, none of the claims addressed in this appeal turn on factual findings. Accordingly, any errors in the state's factual findings did not prejudice Drinkard.
Drinkard raised other claims before the district court, but failed to brief them on appeal. He instead requested us to "consider the discussion of all claims and arguments contained in prior pleadings.” Whether we consider issues not briefed on appeal is a matter of discretion. Compare Black v. Collins, 962 F.2d 394, 399 (5th Cir.) (addressing arguments made in district court even though not obligated to do so), cert. denied, 504 U.S. 992, 112 S.Ct. 2983, 119 L.Ed.2d 601 (1992) with. Hobbs v. Blackburn, 752 F.2d 1079, 1083 (5th Cir.) (refusing to review "matters [that] have not been cited as error on appeal and have not been briefed”), cert. denied, 474 U.S. 838, 106 S.Ct. 117, 88 L.Ed.2d 95 (1985). We find Drinkard's non-briefed claims to be without merit, and we decline to address them further.
. As required by § 8.04(a) of the Texas Penal Code, the trial court instructed the jury that "[v]oluntary intoxication does not constitute a defense to the commission of a crime.”
. Drinkard argues summarily in his supplemental brief that the Supreme Court’s decision in Egelhoff does not foreclose his “distinct Eighth Amendment challenge to the operation of section 8.04(a) during the guilt-innocence phase of [his] *770trial.” A careful review of his briefing to this court reveals no "distinct Eighth Amendment challenge.” He relies principally on Sandstrom v. Montana, 442 U.S. 510, 99 S.Ct. 2450, 61 L.Ed.2d 39 (1979), and In re Winship, 397 U.S. 358, 90 S.Ct. 1068, 25 L.Ed.2d 368 (1970), both of which rested on the Due Process Clause of the Fourteenth Amendment, not the Eighth Amendment.
. The relevant sentence in the instruction reads, "Therefore, if you find that the defendant ... was temporarily insane as a result of intoxication, then you may take such condition into consideration....” The grammatical definition of an antecedent is "a word, phrase, or clause, usually a substantive that is replaced by a pronoun usually at a later point.” Random House College Dictionary 56 (1980) (emphasis added). The word “insane” in the above-quoted sentence does not serve as a substantive (i.e., noun), but rather as an adjective. Consequently, the term “temporarily insane as a result of intoxication” is not a substantive phrase, but an adjectival phrase. If *772we analyze the sentence in the strictest grammatical sense, the only substantive "condition” in the instruction to which “such condition” could properly refer is the norm "intoxication.” The majority avoids this obstacle, for purposes of its grammatical analysis, by substituting the phrase "temporary insanity caused by intoxication," the term that is used in the first sentence of the instruction, for the phrase "temporarily insane as a result of intoxication,” the term that is used in the sentence relevant to the referent “such condition."