concurring in part and dissenting in part:
I concur in the majority’s holdings in Parts II-V of the majority opinion. That is, I agree that the district court erred in concluding that Lopez’s ineffective assistance of counsel claims for failure to investigate and present mitigating evidence were unexhausted, and I agree that Lopez is not entitled to remand or relief on his *1045claims that the Arizona Supreme Court denied him meaningful appellate review, that his lawyer was deficient in failing to challenge Dr. Hobeich’s testimony, or that his lawyer was deficient in failing to object to evidence of prior injuries.
However, because I conclude that Lopez has successfully demonstrated an unreasonable application of Eddings v. Oklahoma, 455 U.S. 104, 102 S.Ct. 869, 71 L.Ed.2d 1 (1982), based on both the trial judge’s and the Arizona Supreme Court’s failure to consider his mens rea as a mitigating factor, I respectfully dissent from Part I of the majority’s opinion.
I
I share the concern raised by Chief Justice Feldman in his special concurrence, namely the lack of any trial court findings on mens rea. The lack of such a finding creates a distinct constitutional problem under Eddings because there is no evidence that the Arizona courts actually considered mens rea as a mitigating factor.
This case is an outlier in our death penalty jurisprudence because the record is undisputed — and all parties seem to concede — that Lopez never intended to kill his son. Throughout trial, the prosecution emphasized that child abuse felony murder does not include an intent element. During the sentencing hearing, after Lopez’s counsel stressed Lopez’s lack of specific intent as a mitigating factor, the prosecution’s only argument in rebuttal was the following analysis: that Lopez knew from experience that children are vulnerable to injury, that Lopez intended to injure Anthony, and that “we can infer that [Lopez] intended the damage [the injuries] caused because of his very familiarity with children.” (emphasis added). Then, while arguing before the Arizona Supreme Court on direct appeal, the prosecution “conceded that there was little or no evidence of intent to kill and that the most probable explanation for the child’s death was that Defendant became enraged, lost control, and beat the child quite seriously.” State v. Lopez, 174 Ariz. 131, 847 P.2d 1078, 1092 (1992) (Feldman, C.J., specially concurring). There was also evidence introduced at trial that Lopez attempted to administer CPR when Anthony stopped breathing and that Lopez was visibly upset by Anthony’s death. This behavior is inconsistent with an intentional homicide.
Although the formation of a specific intent to kill is not a prerequisite to the imposition of a death penalty, Tison v. Arizona, 481 U.S. 137, 107 S.Ct. 1676, 95 L.Ed.2d 127 (1987), the lack of a specific intent to kill is relevant to whether the death sentence is appropriate for an individual defendant. Notably,' most child abuse felony murderers do not receive a sentence even approaching death, even though all such murders are undoubtedly death penalty eligible.
In short, this case is an unusual one. Lopez received the harshest sentence that the state is empowered to impose even though he never intended for his victim to die. This fact does not exempt him from the death penalty by any means. However, it does call into question the state courts’ assessment of mitigation.
II
Under Eddings, a sentencer contemplating the death penalty has a constitutional obligation to consider, “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.” 455 U.S. at 110, 102 S.Ct. 869 (quoting Lockett v. Ohio, 438 U.S. 586, 604, 98 S.Ct. 2954, 57 L.Ed.2d 973 (1978)). Applying Eddings to this case, both the sentencing judge and the Arizona Supreme Court had a constitutional duty, before choosing or *1046affirming a death sentence, to consider all mitigating factors that Lopez’s trial counsel proffered.
During Lopez’s sentencing hearing, his trial counsel spent considerable time arguing that Lopez’s lack of specific intent militated against the death sentence. There is therefore no doubt that Eddings required the trial judge and the Arizona Supreme Court to consider Lopez’s mens rea when deciding whether to impose and whether to affirm the death penalty. The only question before us, therefore, is whether or not the state courts did, in fact, consider Lopez’s argument.
In reviewing the state court record for a potential Eddings violation, this court must, as the majority correctly notes, presume that the trial judge and the state Supreme Court actually considered all of the mitigating arguments that were presented. This presumption arises, first and foremost, from a blanket presumption that state judges know and follow the law. It arises secondarily in this case from the trial judge’s multiple assurances on record that she was considering all of Lopez’s proffered evidence and arguments. Parker v. Dugger, 498 U.S. 308, 315, 111 S.Ct. 731, 112 L.Ed.2d 812 (1991) (“We must assume that the trial judge considered all this [mitigation] evidence before passing sentence. For one thing, he said he did.”). Under Parker, we must presume that the trial judge was neither disingenuous nor mistaken when she stated that she considered all mitigating arguments.
The Parker presumption, however, is neither conclusive nor irrebuttable. In Parker itself, the Supreme Court did not rely on the trial judge’s assurance that he had considered all mitigating evidence. Rather, the Court examined the entire record, reconstructing the trial judge’s decision-making process to determine whether or not that process included consideration of non-statutory mitigating factors. The Court then rejected the Eddings claim, not because the trial judge had made a bare statement that he had considered mitigating evidence, but because the trial judge’s final decision was inexplicable without reference to that evidence. Specifically, the Supreme Court pointed out that the trial judge had overridden the jury’s recommended life sentence for only one of the two murders involved in the case. Although Parker had committed a dual homicide, he received the death sentence for one murder and a life sentence for the other. Because the aggravating circumstances applied to both homicides, the Supreme Court reasoned that the only explanation for the differential sentencing was that the trial judge had considered non-statutory mitigating circumstances and had determined that those circumstances justified leniency with respect to only one of the murders. Parker, 498 U.S. at 316— 17, 111 S.Ct. 731. In Parker itself, therefore, the initial presumption that the trial judge considered all mitigators was only a starting point. The presumption was not — and is not — conclusive.
Furthermore, the Parker presumption is easily rebuttable. As Justice O’Connor wrote in her Eddings concurrence, the qualitatively different nature of a death sentence requires reviewing courts “to remove any legitimate basis for finding ambiguity concerning the factors actually considered by the trial court.” 455 U.S. at 119, 102 S.Ct. 869 (O’Connor, J., concurring). As a result, if there is any legitimate reason to believe that the trial judge excluded a mitigating factor from her consideration, then we should remand for re-sentencing. The consequences of error in a death case are too great to allow for speculation or fudging.
In the end, then, the Parker presumption is merely a rule for allocating burdens, which will determine the outcome *1047only if the defendant or habeas petitioner offers no legitimate reason to doubt its validity in a particular case. If, however, a habeas petitioner can point to any legitimate evidence in the record that indicates the trial judge’s exclusion of any factor or evidence, then this court should grant the writ and remand for resentencing.
Ill
In this case, Lopez offers two theories to rebut the Parker presumption. First, he points to one instance in the record in which the trial judge stated that she would consider only those mitigating factors that were presented during the aggravation/mitigation hearing, impliedly to the exclusion of mitigating evidence adduced at trial. Lopez argues that that one statement should suffice, per Justice O’Con-nor’s admonition against speculation, to undermine this court’s confidence in the trial judge’s Eddings compliance.
As the majority correctly concludes, Lopez takes that statement out of context. The trial judge’s only point was that she would, consistently with Lopez’s motion to strike, ignore certain statements that were included in the presentence report. As she made clear at several other points in the proceeding, the trial judge did not mean to imply that she would neglect mitigating evidence adduced at trial. The statement that Lopez emphasizes, therefore, does not introduce speculation or confusion when it is read in the context of the entire trial record. Lopez’s first attempt to rebut the Parker presumption fails.
But Lopez offers a second theory, which is more compelling. Lopez directs our attention to Chief Justice Feldman’s separate concurring opinion in the Arizona Supreme Court’s decision of his direct appeal, which noted that the trial judge failed to make a statutorily required finding as to the potential mitigating effect of Lopez’s mens rea. State v. Lopez, 174 Ariz. 131, 145-16, 847 P.2d 1078 (1992) (Feldman, C.J., specially concurring). In his separate opinion, Chief Justice Feldman noted that A.R.S. § 13-703(D) requires trial judges to articulate specific findings with respect to any statutory mitigating factor that is present in the case. He also noted that the statute lists as a mitigating factor “ ‘any of the circumstances of the offense.’ ” Id. at 145, 847 P.2d 1078 (quoting A.R.S. § 13-703(G)). Chief Justice Feld-man then observed that a highly relevant circumstance of Lopez’s offense was the possibility that Lopez “may never have formed an intent to kill.” Id.
Contrary to the statutory articulation requirement, Lopez’s trial judge never made a specific finding as to whether Lopez did, in fact, intend to kill Anthony or as to whether Lopez’s lack of specific intent was or was not sufficiently mitigating to call for leniency. Because the trial judge did not articulate a specific finding related to the statutory mitigating factor of Lopez’s intent, Chief Justice Feldman wrote that he preferred to remand the sentence for entry of that finding.
In our determination of whether the trial judge actually considered all of Lopez’s proffered mitigating theories, the Arizona Chief Justice’s opinion is highly relevant. The trial judge’s failure to comply with statutory articulation requirements indicates that she might have failed to consider Lopez’s lack of specific intent as a mitigating factor in the § 13-703 balance. That is, if the trial judge had determined either that Lopez actually intended to kill Anthony or that Lopez’s lack of specific intent to kill was insufficiently mitigating to call for leniency, then she should have articulated a § 13-703(D) finding to that effect. Because the trial judge did not make any such finding, we must conclude that one of two things happened: either she made a state statutory error by failing *1048to enter a required finding or she made a federal constitutional error by failing to consider intent as a mitigating factor.1
I would conclude that the latter possibility is the more likely, for two reasons. First, as noted, Eddings requires us to remand if there is any legitimate basis for finding that the sentencer neglected a mitigating factor. Even taken alone, the trial judge’s statutory failure raises some ambiguity as to whether she included Lopez’s mens rea in her aggravation/ mitigation balance. Because we should err on the side of caution, avoiding speculation, the statutory failure alone is enough to justify a remand.
Second — and more compellingly — there is a strong commonsensical likelihood, which gains support from the record, that neither the trial judge nor the state Supreme Court believed that intent should count as a mitigating factor. At the time of Lopez’s sentencing, the United States Supreme Court had recently decided Enmund, v. Florida, 458 U.S. 782, 102 S.Ct. 3368, 73 L.Ed.2d 1140 (1982), and had even more recently decided Tison v. Arizona, 481 U.S. 137, 107 S.Ct. 1676, 95 L.Ed.2d 127 (1987). Both of those cases considered the role that a defendant’s intent should play in an Eighth Amendment determination of whether the death penalty is an excessive punishment for a particular defendant. Specifically, Enmund held that, under the Eighth Amendment, the defendant must have a general intent to kill before he may be given a death sentence. Tison then clarified Enmund, holding that a reckless disregard for someone’s life, combined with an intent to engage in conduct that endangers that person’s life and ultimately results in his death, is sufficient to allow for the death penalty.
As the federal district court found in this case, there can be little doubt that Lopez’s intentional child abuse included a reckless disregard for Anthony’s life, sufficient to satisfy Eighth Amendment requirements under Tison. But the conclusion that Lopez had a sufficiently culpable mens rea to satisfy Tison is not the same as the conclusion that the holistic circumstances of the crime and of the defendant, taking a limited mens rea into account, justify a death penalty. That is, the defendant’s intent should play a dual role at the sentencing phase. Intent is relevant in the first instance to determine whether the death penalty is even on the table, and it is relevant in the second instance to determine whether, on balance, the aggravating and mitigating characteristics of the defendant and the crime militate in favor of (or against) leniency. The individualized analysis that Eddings requires must include holistic analysis of whether the characteristics of the defendant and the circumstances of the crime — including the defendant’s mens rea — justify a death penalty.
In this case, the record strongly implies that the trial judge considered Lopez’s intent argument only in its first role, as a Tison challenge. The Arizona Supreme Court then repeated that error by neglecting the argument altogether, presumably because the court assumed that intent was no longer relevant in light of Lopez’s failure to bring a Tison challenge on direct appeal.
*1049A
The trial judge’s only reference to Lopez’s intent was her statement that Lopez “willfully and intentionally physically abused the child entrusted to his care and caused his death.” This statement makes perfect sense as a Tison finding; it clearly concludes that Lopez intended to engage in conduct that endangered Anthony’s life, and it clearly concludes that the dangerous conduct caused Anthony’s death. Under Tison, that is all that the Eighth Amendment demands.
As Chief Justice Feldman concluded, however, the trial judge’s statement falls far short of a § 13-703(D) finding. First, the grammar of the statement makes it difficult if not impossible to determine whether the intent finding attaches only to abuse or also to homicide. That is, the trial judge might have meant only that Lopez willfully and intentionally abused Anthony and that the intentional abuse actually (though perhaps unintentionally) caused Anthony’s death. On the other hand, she may have meant that Lopez willfully and intentionally physically abused Anthony and willfully and intentionally caused Anthony’s death.2 Because of its phrasing, the finding is, as Chief Justice Feldman believed, too ambiguous to satisfy § 13-703(D). The judge’s statement does not draw a clear conclusion as to whether Lopez intended for his son to die.
Second, the statement does not, as § 13-703(D) requires, include any articulation of the finding’s effect on the overall § 13-703 balance. That is, even assuming that the trial judge’s statement sufficiently clarified that Lopez did, in fact, intend to kill his son, the trial judge should also have stated — explicitly on record — the resulting conclusion that Lopez’s proffer failed to provide a mitigating circumstance. In other words, if the trial judge had considered Lopez’s argument as a mitigating factor— and not just as a Tison challenge — she would have said more.
In sum, the trial transcript raises a real possibility that the judge took Lopez’s argument only as a Tison challenge, failing to consider the same argument in its second role as a statutory mitigator. Indeed, the transcript provides no support — except by speculation — that the trial judge specifically considered mens rea as a mitigating factor.
*1050B
On direct appeal, the Arizona Supreme Court’s majority opinion implicitly confirmed this view of the trial record and ultimately repeated the error. The Supreme Court’s opinion strongly implied that the majority found no specific intent to kill, but the opinion proceeded to exclude that finding from its consideration of mitigating factors.
1
Although the Arizona Supreme Court made no specific intent finding,3 the majority opinion did strongly indicate its belief that Lopez did not intend to kill Anthony. The court’s implication to that effect came in its analysis of a statutory aggravating factor that was based on the finding that Lopez was in a “heinous and depraved state of mind when he beat Anthony.” Lopez, 174 Ariz. at 144, 847 P.2d 1078. As the Supreme Court noted, there are five circumstances that may support a finding of heinousness and depravity: “(1) the murderer’s relishing of the murder; (2) the infliction of gratuitous violence on the victim; (3) the needless mutilation of the victim; (4) the senselessness of the murder; and, (5) the helplessness of the victim.” Id. (citing State v. Gretzler, 135 Ariz. 42, 52, 659 P.2d 1, 11 (1983)). The first three of those factors, if present, would certainly demonstrate a specific intent to kill; only someone intent on killing would relish the act, gratuitously beat the victim, and needlessly mutilate the body. But the last two circumstances — the senselessness of the murder and the helplessness of the victim — do not indicate anything about the defendant’s specific intent. A vehicular manslaughter of an infant passenger, for example, clearly would involve the senseless death of a helpless victim without involving a malicious intent on the part of the intoxicated driver.
For our purposes, it is therefore highly relevant that the Arizona Supreme Court relied solely on the last two factors when upholding the finding of heinousness and depravity in Lopez’s case. The court’s conclusion that none of the first three factors were present in this case strongly implies that the court found little to no evidence of specific intent to kill. Id. In fact, the evidence that the Arizona Supreme Court cited to support its finding of senselessness was the very evidence that Chief Justice Feldman cited when arguing that Lopez probably did not intend to kill his son. In the words of the Supreme Court’s majority opinion: “The murder of Anthony was senseless. Lopez admitted that after he had given Anthony a bath and was putting lotion on him, the infant ‘peed, so I smacked him.’ ” Lopez, 174 Ariz. at 144, 847 P.2d 1078. This finding is exactly the same as the concurring opinion’s finding that “the most probable explanation for the child’s death was [not that Lopez intended to kill Anthony, but rather] that Defendant became enraged, lost control, and beat the child quite seriously.” Id. at 145, 847 P.2d 1078 (Feldman, C.J., specially concurring). In short, the Supreme Court majority apparently could not support the aggravating factor of heinousness and depravity by reference to evidence of intent; on the contrary, the majority opinion implicitly agreed with the concurrence’s view that the murder, while senseless, was likely unintentional.
The upshot of this discussion is that the Arizona Supreme Court’s logic confirms the view that this case probably involved— and the trial judge probably found — intentional abuse resulting in death, not intentional murder. Although Lopez’s mens *1051rea was sufficiently culpable to allow a death sentence under Tison and to support a finding of heinousness and depravity, neither of the state courts found that Lopez actually intended for his son to die.
2
After implicitly accepting that Lopez did not intend to kill his son, the Arizona Supreme Court’s majority opinion then completely excluded that factor from its consideration of mitigating evidence. The majority opinion noted only that “[t]he vast majority of the mitigation evidence presented by Lopez centered on his being a good parent and on the fact that he cared for children and never acted inappropriately with them.” Lopez, 174 Ariz. at 145, 847 P.2d 1078. The opinion then easily disposed of that mitigating argument — affirming the trial court’s rejection of Lopez’s character evidence — by noting that Lopez had another conviction for child molestation, which was before the trial judge at the time. But the Arizona Supreme Court’s review and rejection of Lopez’s character evidence comprised its entire consideration of mitigating factors. In the three paragraphs in which the majority opinion considered mitigation, there is not a single mention of Lopez’s mens rea.
The absence of Lopez’s intent argument from the Supreme Court’s analysis is particularly telling in light of the special concurrence. The entire raison d’etre of Chief Justice Feldman’s opinion was to point out that § 13-703(G) required the trial judge to consider Lopez’s mens rea as a mitigating factor. The majority opinion completely (though silently) rejected that argument. As noted, however, the majority implicitly agreed with — and certainly made no effort to rebut — the concurrence’s conclusion that Lopez probably did not intend to murder his son, thereby indicating that the basis for disagreement between the majority and the concurrence could not have been the factual conclusion. Rather, as indicated by the majority’s complete failure to include intent in its discussion of mitigating evidence, the basis for disagreement must have been a question as to whether mens rea counts as a statutory mitigator at all. The majority of the Arizona Supreme Court apparently believed that it did not. As previously noted, that belief was not unfathomable at the time, given that the United States Supreme Court had recently given mens rea a unique role in death penalty decision making. The Arizona Supreme Court might have believed that Emnund and Tison removed mens rea from the aggravation/mitigation balance, giving that factor an entirely separate role.
Nevertheless, by excluding Lopez’s intent argument from its consideration of mitigating factors, the majority opinion committed an even clearer Eddings violation than the trial court. Although the trial transcript does not include any listing of mitigating factors — thereby making it somewhat difficult to determine which factors actually entered the trial judge’s thought process — the Supreme Court’s majority opinion specifically mentions the mitigating factors that it considered and clearly neglects Lopez’s (and Chief Justice Feldman’s) intent argument.
C
In conclusion, Chief Justice Feldman’s concurring opinion sheds important light on the trial judge’s and the majority opinion’s thought processes. Neither the trial court nor the appellate court found that Lopez did intend to kill Anthony; the Supreme Court’s majority opinion implied that Lopez probably did not intend to kill Anthony; and neither court made any finding as to the role that Lopez’s potential or actual lack of intent played in its analysis of mitigating factors. Because a finding as to mitigation is statutorily required *1052whenever the courts consider mens rea as a mitigating factor, the failure to make the required findings is sufficient in itself to rebut the Parker presumption that the courts actually considered Lopez’s proffer. Furthermore, in this case, the record indicates that both the trial judge and the Arizona Supreme Court took Lopez’s intent argument only as a Tison challenge, neglecting to reevaluate the same argument in its second role as a mitigating factor. There is legitimate — indeed, strong — record evidence to support the conclusion that both courts, contrary to Eddings, failed to consider Lopez’s mens rea in the § 13-703 balance.
Particularly given Justice O’Connor’s admonition against speculation, I would conclude that Lopez has demonstrated an Eddings violation.
TV
The final question is whether, under AEDPA, the state courts’ failure to consider intent as a mitigating factor constitutes an “objectively unreasonable,”4 rather than simply an “incorrect or erroneous,” application of Eddings. See Lockyer v. Andrade, 538 U.S. 63, 75, 123 S.Ct. 1166, 155 L.Ed.2d 144 (2003) (citing Williams v. Taylor, 529 U.S. 362, 410, 412, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000)). In this case, there can be little doubt that it does.
Eddings held, without qualification, that due process requires a sentencer to consider all mitigating factors that a defendant proffers. 455 U.S. at 114, 102 S.Ct. 869 (“The sentencer, and the [appellate court] on review, may determine the weight to be given relevant mitigating evidence. But they may not give it no weight by excluding such evidence from their consideration.”). The Eddings opinion also relied on Lockett v. Ohio, 438 U.S. 586, 98 S.Ct. 2954, 57 L.Ed.2d 973 (1978), which explicitly held that the sentencer must consider “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.” 438 U.S. at 604, 98 S.Ct. 2954. There is no potential for reasonable minds to differ on the question of whether the defendant’s intent falls within the scope of Eddings and Lockett. It does.
The dual role that mens rea must play in death penalty cases does not undermine this conclusion; in fact, it strengthens it. The fact that the Eighth Amendment contains a minimum culpability requirement demonstrates the critical importance of the defendant’s mens rea in a death penalty decision. Although intent’s dual role makes it understand,able that a sentencing judge might fail to switch hats after making an Enmund finding, the dual role does not make the same failure reasonable. Even after establishing eligibility for the death penalty, Eddings clearly requires the sentencing judge to continue evaluating the whole crime, including mens rea, to determine whether the death penalty is actually appropriate for the individual defendant.
In this case, the applicable state law and the arguments presented at sentencing make the state courts’ failures all the more unreasonable. First, to the extent that the Lockett requirement might have been ambiguous, the same requirement was codi-*1053fled in Arizona’s criminal statutes. As Chief Justice Feldman noted, A.R.S. § 13-703(G) specifically requires Arizona trial judges to consider “any of the circumstances of the offense.” The trial judge’s neglect of Lopez’s mens rea, therefore, was not only a misapplication of Supreme Court precedent but also a defiance of state statute, enhancing the unreasonableness of the Eddings violation.
Second, Lopez clearly proffered — indeed, emphasized — his mens rea as a mitigating factor. Although Lopez focused primarily on his good character during the aggravation/ mitigation hearing, his counsel focused almost exclusively on intent throughout the sentencing hearing. In fact, in his last statement to the trial judge before the sentence was imposed, Lopez’s attorney said:
I think the one single most important thing here, in this case, is that there’s been no showing, at all, that George ever intended that baby to die[.] ... I would say that everything points to the contrary. That if that were-his intent, there would be different facts before you and, under these circumstances, I think he’s going to be punished enough by getting the sentence that the Court has to impose.
Given that Lopez’s intent argument was the last thing — and “the one single most important thing” — that his attorney offered in mitigation, the trial judge’s and the Arizona Supreme Court’s failure to consider it was objectively unreasonable.
I therefore would conclude that Lopez is entitled to relief under AEDPA on his certified Eddings claim, and I would grant the writ of habeas corpus with respect to that claim, remanding his case to the Arizona courts for resentencing.
. The majority here concludes: “Of course, the trial court did not specifically discuss any of the mitigating evidence that Lopez presented, so a failure to discuss this factor does not necessarily indicate it was not considered.” Maj. Op. at 1038. As should be clear from this discussion and from Chief Justice Feld-man’s opinion, the failure to make an intent finding, unlike the failure to make findings related to non-statutory mitigators, is a legal error, not mere churlishness. That failure, therefore, has far more significance than the majority opinion assigns to it.
. Interestingly, the sentencing transcript and the minute entry, both of which include the intent finding, punctuate the finding slightly differently; an important comma is missing from the minute entry. In the sentencing hearing transcript, the trial judge is quoted with the following commas: "[Tjhe court finds beyond a reasonable doubt that the defendant, the father of the child, Anthony, willfully and intentionally, physically abused the child entrusted to his care and caused his death.” In the minute entry, however, the trial judge wrote: “THE COURT FINDS, as to Count 2, FIRST DEGREE MURDER, beyond a reasonable doubt that the defendant, father of the child Anthony, willfully and intentionally physically abused the child entrusted to his care and caused his death.” The comma following "willfully and intentionally” in the sentencing transcript gives that phrase an independent and lasting significance, such that it might be read to attach to both halves of the following phrase. The absence of the comma in the written order is then significant since "willfully and intentionally” becomes a seamless part of the immediately following phrase, "physically abused the child entrusted to his care." That phrase is then separated from the murder finding by the conjoining "and,” such that the "willfully and intentionally” language might not carry over the "and” to attach to the homicide. Reading the minute entry, which is the finding that the judge herself might have been involved in punctuating, it seems highly possible that the intent finding attached only to abuse, not to murder.
. As noted, Lopez did not raise a Tison challenge on direct appeal. The Supreme Court, therefore, did not even make a specific Tison finding.
. Because neither the trial judge nor the Arizona Supreme Court made any specific mention of the Eddings standard, I will not analyze the question under AEDPA’s "contrary to” clause. See 28 U.S.C. § 2254(d)(1). Although it appears that the state courts applied an entirely incorrect legal standard by assuming that intent is not a mitigating factor under Eddings, I cannot conclude with certainty that they did so. I will therefore confine my analysis to the stricter standard, which allows reversal only if the state courts’ decisions "involved an unreasonable application” of the correct federal rule. Id.