C. WE HAVE AUTHORITY TO ORDER A LIMITED REMAND
Tennessee Code Annotated section 27-3-128 authorizes this court to remand a case for further proceedings whenever "complete justice cannot be had by reason of some defect in the record, want of proper parties, or oversight without culpable negligence." The statute empowers appellate courts to direct limited proceedings to ensure that their judgments rest on a record that is both accurate and complete.
Our supreme court has long applied this provision when significant factual or legal developments arose after the trial or post-conviction hearing. Our opinion cites Pruett v. State, 501 S.W.2d 807 (Tenn. 1973), but there are other precedents as well. In Caesar v. Harris, 69 S.W. 731 (Tenn. 1902), the court reversed a refusal to remand after counsel presented an affidavit establishing new facts following judgment. The court explained that when "injustice will be done by dismissing, but [a party's] suit fails from a neglect not culpable, it is the duty of the court to remand, that justice may be done." Id. at 731 (quoting Wood v. Neely, 66 Tenn. 586, 590 (1874)). It further clarified that "culpable negligence"
means something more than ordinary neglect: "As a matter of course, there must always be some neglect in every oversight; therefore the qualification of the statute requiring it to be culpable." Id.
The supreme court reaffirmed this principle in Sartain v. Dixie Coal & Iron Co., 266 S.W. 313, 320 (Tenn. 1924). In that case, the court held that although an appellate court ordinarily acts on the record as it stands, it may remand "in a proper case, and in furtherance
5
of justice," when the proof is "not altogether satisfactory" and additional development is necessary "to the end that it may be heard again."
Together, I believe that these cases establish three guiding principles: (1) remand serves an equitable function, ensuring appellate judgments rest on a complete and reliable record; (2) "culpable negligence" means willful or inexcusable neglect, not mere delay or oversight; and (3) remand is proper when a material factual or legal development arises that could affect the fairness or reliability of the judgment. To be sure, an appellate court is not required to remand under section 27-3-128 whenever new information appears. The statute confers a discretionary, equitable power—not a mechanical duty. Its exercise depends on the gravity of the issue raised, the potential prejudice to the parties, and whether any omission resulted from true culpable negligence.
In my view, the record does not support any finding that the Petitioners acted with culpable negligence in failing to obtain and present Dr. Levy's affidavit earlier. For example, the record does not reveal when the parties were able to locate Dr. Levy in Pennsylvania or whether the Petitioners were able to contact him before the hearing. We also do not know when the District Attorney's Conviction Review Unit first communicated with him or what other circumstances may have led to the timing of his affidavit. Instead, what we do know is that the State provided the affidavit to the Petitioners and that the affidavit was filed with the court on the same day Dr. Levy executed it. In the absence of evidence establishing heightened negligence by the Petitioners, I would not presume it. It might be suggested that the Petitioners should have anticipated the issue and presented evidence addressing the absence of "culpable negligence." However, the record does not indicate that the parties were on notice that the timing of the affidavit itself would be a bar to remand. The State did not raise this issue in its response to the Petitioners' motion, and our prior order denying relief did not refer to this possibility. If clarification of the circumstances surrounding Dr. Levy's affidavit were needed, additional briefing or submissions could be requested from the parties.
In any event, there is no prejudice to the State in remanding the matter for limited proceedings, particularly where the interests of justice favor a full and accurate factual record. The District Attorney's office secured Dr. Levy's affidavit and provided it to the Petitioners. In these circumstances, where the prosecuting authority itself has secured this evidence for consideration, concerns about sandbagging or gamesmanship by the Petitioners are either wholly absent or entitled to little weight.
6
Section 27-3-128 grants us the authority to remand this case; the only remaining question is whether the facts here compel its use. They do, as I explain below.
D. WHY THIS CASE WARRANTS REMAND
At the argument of this case, the State suggested that Dr. Levy's affidavit is merely another round in an endless "battle of the experts." I view it differently. This case presents an extraordinarily rare circumstance: the State's own chief forensic pathologist, whose testimony established that a homicide occurred, has now reviewed additional evidence, including evidence not available to him previously, and concluded that his original determination was mistaken. Several reasons exist why we should grant the motion for a limited remand.
1. Dr. Levy's Recantation Is Not Just Another Expert Opinion
In my view, the characterization of Dr. Levy's affidavit as being simply a part of a continuing "battle of the experts" misapprehends the nature of the evidence. There is a distinction between an expert's critique of another's testimony and an expert's withdrawal of his own.
The seven medical witnesses who testified at the post-conviction hearing served the typical role of retained experts: to challenge and attempt to impeach the scientific basis of the State's trial evidence. Their testimony gave the court additional perspectives to weigh, but the underlying evidence supporting the convictions remained intact. Dr. Levy's position is altogether different. As Tennessee's Chief Medical Examiner, he was not a partisan advocate, but a public official charged by statute with determining cause and manner of death in cases of suspected criminal violence. He performed the autopsy, examined the victim's body, reviewed the medical records, and testified that the victim's death was a homicide caused by Shaken Baby Syndrome arising from the May 1999 event. His testimony did not merely supplement the State's case; it constituted the forensic foundation upon which it stood.
Now, Dr. Levy has withdrawn the very conclusions that sustained that foundation. In his affidavit, he attests that the cause of death is "undetermined," the manner "natural,"
and the death unrelated to the May 1999 incident. This is not impeachment of another witness's opinion; it is the retraction of the State's indispensable proof. As the Sixth Circuit has explained, such a recantation does "not merely add to the defense, but also deduct[s]
7
from the prosecution," because it removes the testimony on which the verdict was built. Souter v. Jones, 395 F.3d 577, 593 (6th Cir. 2005).
If Dr. Levy's new conclusions are credited, the record no longer contains proof of a homicide at all. After all, "It is hard to imagine any reasonable jury's returning a conviction when no one can even say confidently that a murder has been committed." Ex parte Robbins, 478 S.W.3d 678, 692 (Tex. Crim. App. 2014) (so concluding in the context of recanted testimony from a medical expert who now opined that the cause of death was
"undetermined"); see also Keen v. State, 398 S.W.3d 594, 612 n.18 (Tenn. 2012) ("'Actually innocent of the offense' means that one was never factually guilty of the crime.").
Consequently, Dr. Levy's affidavit is not the mere continuation of a forensic debate. If credited, it is the dismantling of the forensic premise upon which the State's homicide conviction rested. The post-conviction court should have the opportunity to hear this revised testimony, assess its credibility, and determine whether the structure of the State's case can still stand once its cornerstone has been removed.
2. Dr. Levy's Recantation is Itself New Scientific Evidence
I also respectfully disagree that Dr. Levy's affidavit reflects only a long-standing divergence in the medical community's understanding of Shaken Baby Syndrome and therefore does not constitute "new scientific evidence." In my view, the medical examiner's recanting of his testimony establishing the cause and manner of death, separate from the reasons for the recantation, is itself the new scientific evidence warranting reconsideration.
The question is not whether theories of Shaken Baby Syndrome have been debated in medical journals or explored in courtrooms for decades. Nor is it whether a retained defense expert could have disagreed with Dr. Levy's trial testimony at the time. The proper question is whether the specific scientific evidence that formed the basis of the State's homicide conviction has now been withdrawn by the very witness who supplied it. It has. Two points illustrate why.
First, it is Dr. Levy's revised opinion itself—not merely the reasons underlying it— that constitutes the new scientific evidence. As the Sixth Circuit has explained in a similar context, the expert's opinion at trial—not its underlying rationale—is the evidence the
8
factfinder considers. Thus, when a forensic pathologist later recants and changes that opinion, "the evidence itself has changed and can most certainly be characterized as new."
Souter, 395 F.3d at 592; see also Robbins, 478 S.W.3d at 692 (concluding that, when a medical examiner revised her cause-of-death determination from "homicide" to
"undetermined," the new opinion—grounded in updated scientific knowledge—is itself new scientific evidence). That same reasoning governs here: Dr. Levy's new conclusions, drawn from his professional reassessment and new information, represent different scientific evidence than what was available and offered at trial. Second, the fact that Dr. Levy's revised conclusions draw, in part, upon information previously discussed by the defense does not diminish their status as new scientific evidence. This point has again been emphasized by the Sixth Circuit, which has recognized that the recantation of a medical examiner's testimony, even if cumulative of prior defense evidence, "does not minimize its effectiveness in weakening the prosecution's case."
Souter, 395 F.3d at 593 n.8. What matters is not whether some of the same materials were available to others in the past, but that the State's own expert—the very witness whose testimony supplied the forensic foundation of guilt—has now, under oath, withdrawn that foundation.
In other words, the significance of the recantation lies in its source. When a neutral state medical examiner, rather than a retained partisan expert, repudiates his own prior scientific conclusions, the consequence of that change does not depend on whether others might have voiced similar doubts before. The recantation carries its own independent probative force precisely because it comes from the official whose conclusions the State relied upon to prove that a homicide occurred. It is that change in scientific judgment by the State's own forensic authority that gives the evidence its character as new—and its potential to alter the outcome.
To be sure, Dr. Levy's reasons for changing his view—his review of materials not previously available to him, his engagement with updated medical literature, and his professional reassessment—are relevant to credibility, which the post-conviction court must evaluate. Cf. Cleveland v. Bradshaw, 693 F.3d 626, 640 (6th Cir. 2012). But those reasons do not alter the central fact: Dr. Levy's sworn change of professional judgment as to the manner and cause of death constitutes new scientific evidence within the meaning of our law.
The issue, then, is not whether other doctors could have reached this conclusion earlier, but whether the expert whose testimony established the cause and manner of death
9
has now, through renewed scientific judgment, repudiated that testimony. He has. Dr. Levy's recantation is itself new scientific evidence of actual innocence.
3. The Impact on the Post-Conviction Court's Assessment of the Case Before Us Now
Given Dr. Levy's central role and the nature of his recantation, the post-conviction court should now be permitted to evaluate his affidavit directly. That court presided over the trial, the post-conviction hearing, and the present proceeding. It is intimately familiar with the medical evidence, the testimony of numerous experts, and the factual circumstances of this case. It is therefore uniquely positioned to assess Dr. Levy's credibility, to weigh his recantation against his trial testimony, and to determine whether the new opinion alters its prior findings.
The record suggests that such testimony would have been critical. In denying relief, the post-conviction court discounted the testimony of seven defense experts—at least in part because it continued to credit Dr. Levy's original autopsy findings. The court expressly noted that it "diminishe[d] the value of the newly presented evidence where fresh opinions were offered but not probed" through cross-examination.
That reasoning rests on an implicit premise: that Dr. Levy's original testimony remains valid and unrebutted. But if Dr. Levy himself now disavows those conclusions under oath, then the factual foundation of the court's ruling has shifted. His recantation calls into question the very findings that the court treated as authoritative. A limited remand would not require the post-conviction court to accept Dr. Levy's new conclusions. But it would allow that court—the judicial officer most familiar with the case—to assess whether this evidence, if found credible, changes its previous analysis. Proceeding without that assessment risks affirming a judgment whose foundation is now fractured by the very expert who constructed it.
4. Finality and Judicial Efficiency Favor a Limited Remand
I recognize that these considerations favoring remand must be weighed against important countervailing concerns. Chief among them is finality, which serves essential functions in our justice system. Convictions must reach a point of repose. Victims and society need closure. Law enforcement and prosecutors must be able to rely on jury verdicts. Defendants must be afforded reasonable opportunities to challenge their
10
convictions, but those opportunities cannot be unlimited. The Post-Conviction Procedure Act reflects this balance by strictly limiting the grounds for post-conviction relief and imposing procedural requirements designed to bring cases to a conclusion. The question is not whether finality matters—of course it does. The question is whether the circumstances of this case warrant a narrow exception to the general rule against piecemeal litigation. For several reasons, I believe they do. First, remand in this case would not open the floodgates to endless litigation. The circumstance presented here—the State's chief medical examiner recanting his determination that a death was a homicide—is exceptionally rare. Recognizing remand authority in this unique situation would not establish a precedent permitting defendants to continually seek new expert opinions and present them piecemeal over decades. The very rarity of this circumstance limits its potential for abuse.
Second, remanding the case now will promote finality and expeditious resolution. The post-conviction court could hold a focused hearing on Dr. Levy's affidavit, make appropriate findings, and either grant relief or reaffirm its previous denial. Appellate review would follow in the ordinary course. This approach resolves the issue in its proper sequence and forum, avoiding the procedural complexity and delay that would otherwise result from multiple separate proceedings.
Third, principles of judicial restraint favor remand. Our role as an appellate court is to review the decisions of the trial court, not to make factual findings in the first instance. The post-conviction court's decision was based on the record available to it at the time. On that record, the court's reasoning was measured and thorough. But the factual foundation has now shifted in a material way, through the submission of evidence that post-dates the court's ruling. Judicial restraint counsels deference to the trial court—particularly when new evidence may alter the factual foundation of a case—rather than affirming an order that may rest on premises now called into question.
Finally, I see no prejudice against the State in ordering a limited remand. If, upon remand, the post-conviction court finds Dr. Levy's recantation credible and material, that finding would serve the State's interest in ensuring that its prosecutions rest on accurate factual foundations. If the court finds the recantation lacking in credibility or materiality, the conviction will stand on a more complete record. Either outcome promotes the interests of justice.
11
Remand is not a retreat from finality. It is a reasoned response to new and specific concerns. As such, before closing the door, we should be clear about what it means to open it—even slightly.
E. DEFINING THE SCOPE OF A TARGETED REMAND
The concern is expressed that permitting remand would allow the Petitioners
"another proverbial bite at the apple" and would transform post-conviction proceedings into "an open- and possibly never-ending approach." In my view, these concerns can be fully addressed through appropriate limits on the scope and standard applicable on remand. A remand under Tennessee Code Annotated section 27-3-128 need not authorize a wholesale reopening of the post-conviction proceedings. This court can direct that the remand be limited to consideration of Dr. Levy's affidavit and related evidence bearing on the subjects raised therein. The post-conviction court would not reconsider all issues previously decided, but only whether Dr. Levy's recantation affects the court's prior findings.
As to the applicable standard, Tennessee law requires that a petitioner establish actual innocence based on new scientific evidence by clear and convincing proof. See Tenn. Code Ann. § 40-30-110(f). This demanding standard would apply on remand. The question would not be merely whether Dr. Levy has changed his opinion, but whether his recantation—if credited—establishes by clear and convincing evidence that the Petitioners did not commit the offenses of which they were convicted.
In assessing Dr. Levy's credibility and the weight of his revised opinion, the post- conviction court would necessarily consider factors including the following:
• the factual basis for Dr. Levy's changed opinion, including what medical records he reviewed that were not previously available to him and what role those records played in his revised conclusions;
• the scientific basis for his conclusion that medical understanding of Shaken Baby Syndrome has changed since the trial in ways that affect his analysis;
• Dr. Levy's explanation for his original findings and testimony, and his account of why those were erroneous;
12
• the consistency or inconsistency between Dr. Levy's revised opinion and the physical evidence, medical records, and other evidence in the case;
• the nature and extent of any external influences on Dr. Levy's revised opinion, including the passage of time, his review of other experts' reports, or other factors that might bear on reliability; and
• the weight to be accorded to Dr. Levy's recantation in light of the contrary evidence presented at trial, including non-medical evidence of guilt. These are quintessentially factual determinations that trial courts make regularly. The post-conviction court is well-equipped to conduct this analysis. If, after hearing from Dr. Levy and considering all relevant evidence, the court concludes that his recantation is credible and establishes actual innocence by clear and convincing proof, relief may be warranted. If not, the convictions would stand.
With these considerations in mind, I would grant a limited remand—not to reopen the case, but to ensure it has not been closed too soon.
CONCLUSION
In summary, the post-conviction court denied relief after hearing from seven defense experts who challenged the Shaken Baby Syndrome diagnosis. The court was unpersuaded, in part, because it credited Dr. Levy's original testimony as establishing the facts of the case. The court doubted that Dr. Levy would recant. Dr. Levy has now recanted.
Whether his recantation is credible, scientifically sound, and sufficient to establish actual innocence are questions the post-conviction court should have the opportunity to evaluate. That court knows this case intimately, and it has the tools to evaluate Dr. Levy's testimony. It can probe the basis for his changed opinion and determine whether it withstands scrutiny.
We should let it do so.
Remanding this case does not guarantee relief. It offers only a hearing—narrow in scope, governed by law, and subject to review. Yet when the State's own medical examiner
13
withdraws the testimony that made its case, and when justice can be served without prejudice to the State, prudence counsels pause, not closure. I respectfully dissent.
S/ Tom Greenholtz
TOM GREENHOLTZ, JUDGE
14