EX PARTE JOSEPH DAVID HORVATH v. << (2025)

DISSENTING OPINION

Today, the Court declines Applicant Joseph David Horvath's suggestion that we reconsider the denial of his application for writ of habeas corpus. See Ex parte Horvath, No. WR-88,478-01, — S.W.3d —, 2025 WL 1699335 (Tex. Crim. App. June 18, 2025) (per curiam) (denying relief); Tex. R. App. P. 79.2(d). I disagreed with the decision, Horvath, 2025 WL 1699335, at *5 (Walker, J., dissenting), and I still believe that, had the jury been given the new scientific evidence, he would not have been convicted. I would have this Court reconsider the application. Because the Court does not, I respectfully dissent.

Applicant asks that the Court reconsider its order denying relief, which stated, in its entirety, that:

Applicant was convicted of capital murder and sentenced to life imprisonment. The Tenth Court of Appeals affirmed his conviction. Horvath v. State, 10-07-163-CR (Tex. App.—Waco Feb. 25, 2009). Applicant filed this application for a writ of habeas corpus in the county of conviction, and the district clerk forwarded it to this Court. See Tex. Code Crim. Proc. art. 11.07.

Applicant contends that new scientific research exists since his trial, and if that research would have been known or presented during his trial, it is more likely than not that he would not have been convicted. Based on the record, the State and trial court agree that Applicant should be granted a new trial.

We disagree. In order to obtain relief under Texas Code of Criminal Procedure art. 11.073, an applicant must show that new scientific evidence is available which contradicts scientific evidence presented at trial and had such new evidence been presented at trial, on the preponderance of the evidence, the applicant would not have been convicted.

Based on this Court's independent review of the entire record, we deny relief.

Id. at *1 (per curiam order). The order did not explain why the Court disagreed, only that it did so “[b]ased on this Court's independent review of the record[.]” Id.

Traditionally, this Court “will defer to and accept a trial judge's findings of fact and conclusions of law when they are supported by the record.” Ex parte Reed, 271 S.W.3d 698, 727 (Tex. Crim. App. 2008). So when this Court disagreed and denied relief, it must have determined that the trial judge's findings of fact and conclusions of law were not supported by the record.

Yet the Court did not explain how or why the parties’ agreement and the habeas court's recommendation that Applicant be granted relief was not supported by the record. Instead, the Court's order summarily declared that it disagreed. Certainly, it is this Court's prerogative to agree or disagree. See Ex parte Chavez, 371 S.W.3d 200, 207 (Tex. Crim. App. 2012) (“the convicting court is the ‘original factfinder,’ and this Court is the ultimate factfinder.” (quoting Reed, 271 S.W.3d at 727)). But this left Applicant, the State, and the habeas court to wonder what was wrong with the record that they all believed supported relief. It also leaves future applicants, their lawyers, prosecutors’ offices, and judges scratching their heads as to how to avoid summary disagreement from this Court, where they agree that relief should be granted and believe the record supports it.

What lessons can be learned from the order denying relief? Can anything be learned from it? What makes Applicant's case not worthy of article 11.073 relief, unlike Roark? Some may say that this case is unlike Roark because Applicant admitted to shaking the baby and causing the baby's death. See Horvath, 2025 WL 1699335, at *5 (Yeary, J., concurring) (noting factual distinctions between Roark and Applicant's case).¹

Others may suggest that this was not a “shaken baby” case.² The medical examiner testified the injury could have been caused by slamming the baby onto the bed. But the fact that there could be another way does not mean this case was therefore not a “shaken baby case.” Immediately after the medical examiner agreed with the prosecutor's question that slamming the baby down on a bed could have caused the injury, the examiner also agreed that shaking the baby from side to side could have caused the injury. Indeed, the examiner stated that “if one were dealing with a baby on a bed, began shaking it violently, threw it down on the bed,” that could be the kind of impact he was testifying about.³

The jury also heard that, when interviewed by police, Applicant admitted to the police detective he shook the baby and caused his death. Furthermore, during Applicant's trial testimony, he repeated this statement and said he shook the baby and caused his death, though he did not intend to cause harm.

During closing arguments, the prosecution argued Applicant knowingly caused the baby's death because: “[H]e just told you not two hours ago. He told you, he testified, he admitted that he knew that shaking a baby was dangerous and could cause death. He knew that and he shook him anyway. He knew all those things.”⁴ The prosecution returned to this in their rebuttal argument, telling the jury: “[Y]ou know, for a fact, he told you on the stand he knew it would kill the baby to shake him and he did it. That's the legal test. It's tough. It's hard, but that's what he did.”⁵

But the closing argument misattributes when Applicant knew this. “He knew it; he did it,” strongly suggests that Applicant knew it first, and then did it after. While it is true that Applicant testified that he knew shaking could cause death, and he testified that he shook the child, he did not testify that he knew shaking could cause death at the time he did the shaking. Rather, he knew it at the time he testified at trial, because the medical examiner had earlier testified that it could. See Horvath, 2025 WL 1699335, at *6 (Walker, J., dissenting).

Returning to the topic of the prosecution's argument, it contended that not only did Applicant know shaking could cause death, but that he intended it; the prosecution contended the medical testimony, stating the “force that was necessary to cause those brain injuries[,]” was “not from gently shaking. It's from a large significant force.”⁶

Before those arguments, the jury heard the court's charge which repeatedly asked the jury to find whether Applicant committed the charged capital murder offense, or any of the lesser included offenses, by “shaking him,” striking him or causing him to strike an unknown object, or by an unknown manner and means.⁷ The charge itself reiterated what the jury heard at the very beginning of the trial when the indictment was read, that Applicant was accused of causing the baby's death by “shaking him,” striking him or causing him to strike an unknown object, or by an unknown manner and means.⁸

If this was not a “shaken baby case,” the jury did not get the message. Instead, “shaken baby” was at the forefront of their thinking; they sent a note to the trial court during deliberations: “Question, did you know shaking a baby would kill it? And another question, did you know shaking the, slash, your baby would kill it?”⁹

“Shaken baby” was a significant part of the case against Applicant. I would have the Court reconsider Applicant's -01 habeas application. Even if the Court persists in denying him relief, reconsideration would at the very least give the Court a second opportunity to explain why it disagrees with all of the parties and with the habeas court, rather than summarily denying relief because, “We disagree.” Horvath, 2025 WL 1699335, at *1 (per curiam order).

For my part, I still believe Applicant is entitled to relief in the form of a new trial. If the new science relating to “shaken baby syndrome” was presented to the jury, it would have changed the complexion of the trial and it would have been much harder for the jury to reach “guilty beyond a reasonable doubt.” Instead, it is more likely than not that the jury would have harbored a reasonable doubt, and they would not have convicted Applicant of capital murder. While his case would certainly be stronger had he shown that it is more likely than not that he is innocent, that is not the question that article 11.073 asks. The statute asks whether, “had the scientific evidence been presented at trial, on the preponderance of the evidence the person would not have been convicted.” Tex. Code Crim. Proc. Ann. art. 11.073(b)(2). In other words, is it more likely than not that he would have been found not guilty. Applicant has met the standard.

I would reconsider Applicant's -01 writ application. Because the Court does not, I respectfully dissent.

FOOTNOTES

1.   As I explained in my dissenting opinion to the Court's denial of habeas corpus, in his trial testimony, Applicant agreed to a question posed to him, that the act of shaking could have caused the baby's death. The fact that he said he knew this at trial does not necessarily mean he knew it at the time of the baby's death. Instead, he may well have not known that it was possible until the same time that the jury did, when the medical examiner testified. Horvath, 2025 WL 1699335, at *6 (Walker, J., dissenting).

2.   In regards to another “shaken baby” Roark claim, see, e.g., Ex parte Roberson, — S.W.3d —, No. WR-63,081-03, 2025 WL 2858392, at *17 (Yeary, J., concurring and dissenting) (arguing that Roberson's case was not a shaken baby syndrome case, but rather, a blunt force trauma case), at *21 (Finley, J., concurring and dissenting) (“it is unquestionable that the State's underlying theory was a shaken baby death.”), at *37 (Parker, J., concurring and dissenting) (“[Roberson's] case is not a ‘shaken baby’ case.”)

3.   RR 4: 129.

4.   RR 5: 147.

5.   RR 5: 170.

6.   RR 5: 147.

7.   RR 5: 133 (capital murder), 134 (acquit of capital, proceed to murder), 135 (murder), 136 (acquit of murder, proceed to manslaughter), 137 (manslaughter), 139 (criminally negligent homicide).

8.   RR 4: 13.

9.   RR 5: 172.