When a veteran seeks disability benefits, federal law provides that ties go to the applicant. But if the Veterans Administration decides it’s not a tie—that is, the preponderance of the evidence comes out against the veteran—then it has no occasion to apply this tiebreaking rule. That leads to a question only an appellate lawyer would ask: What standard of review applies to the VA’s determination that the evidence isn’t even: The de novo standard generally used for legal questions or the clear error one used for findings of fact? In Bufkin v. Collins (No. 23-713), a seven-Justice majority held that this is a best seen as a mixed question of law and fact where the fact piece dominates, meriting clear error review. That prompted a dissent from the two Justices perhaps most likely to favor the little guy against the big-bad government—Justices Jackson and Gorsuch—who thought the whole point of this tie-breaking rule was to thwart the VA’s historical reluctance to award veterans the disability benefits they should receive.

Joshua Bufkin and Norman Thornton are two veterans who applied for disability benefits for PTSD caused by their time in the military. Their claims began in local VA regional offices (the first port of call for veterans seeking disability benefits), where Bufkin’s claim was denied entirely, while Thornton received lower benefits than he sought. Both then appealed to the Board of Veterans’ Appeals, an Article I court that reviews the benefits decisions of VA regional offices. The Board affirmed both regional offices’ decisions. In doing so, it acknowledged that whenever “there is an approximate balance of positive and negative evidence” on any issue material to a veteran’s claim, the VA must “give the benefit of the doubt to the claimant.” 38 U.S.C. § 5107(b). But the Board concluded that the evidence was not approximately balanced, so Bufkin and Thornton weren’t entitled to that deferential standard.

Bufkin and Thornton then appealed their respective cases to the U.S. Court of Appeals for Veterans Claims (the “Veterans Court”), another Article I tribunal, which reviews decisions from the Board. There, they argued that the evidence supporting their claims was about equal to the evidence against them, and that they were therefore entitled to get the benefit of the doubt. Federal law provides that in reviewing Board decisions, the Veterans Court must “take due account” of this benefit-of-the-doubt rule. But the Veterans Court concluded that the account that was “due” wasn’t much: Seeing no clear error in the Board’s decision that evidence weighed more strongly against the veterans, it affirmed the Board.

Bufkin and Thornton then appealed the Veterans Court’s decisions to the U.S. Court of Appeals for the Federal Circuit, a genuine Article III court that (among a great many other things) reviews decisions from the Veterans Court. It agreed with the Veterans Court that clear error applies to the Board’s decision that the evidence wasn’t roughly 50-50, so it too affirmed the denial of benefits. These three rounds of appeals weren’t enough for Bufkin and Thornton, though, as they successfully convinced the Supreme Court to grant cert to address the appropriate standard of review.

Unfortunately for our persistent appellants, the Court affirmed all the courts below it in a 7-2 opinion authored by Justice Thomas. It began with the language of the statute which, as discussed above, requires the Veterans Court to take “due account” of the “benefit-of-the-doubt” rule in reviewing the Board’s decisions. But the phrase “due account” doesn’t have a lot of content on its own, so Thomas concluded the general standards of review called for by the veterans statutes are all the “account” that is “due.” Those statutes prescribe the ordinary standards of review appellate lawyers know well, calling for the Veterans Court to review conclusions of law de novo and findings of fact for clear error. So in which bucket fell the Board’s conclusion that the evidence wasn’t about equal, meaning there’s no “doubt” for the veteran to benefit from? For Thomas and majority, weighing up the evidence involves both legal and factual work, making it a mixed question of fact and law. And because this particular mixed question “is about as factual sounding as any question gets,” Thomas thought it was appropriately reviewed only for clear error.

Justice Thomas then brushed aside two objections to this reasoning. First, Bufkin and Thornton argued this interpretation of the legislative command that the Veterans Court take “due account” of the benefit-of-the-doubt rule made the “due account” provision surplusage. Thomas acknowledged that this objection was “a serious one,” but the problem was that it’s just as true if you apply the de novo standard Bufkin and Thornton asked for: Either way, you’re simply following the statute’s default standards of review. Thomas thus concluded that this wasn’t a context where the rule against surplusage could do any work. Second, the veterans observed that some mixed questions of law and fact—like probable cause determinations—are reviewed de novo. But for Thomas, probable cause determinations dwelt in the “constitutional realm,” giving rise to heightened scrutiny. The “benefit-of-the-doubt” standard, by contrast, was a create of statute. And further, probable cause asks the legal-sounding question of what the hypothetical reasonable man might think of a particular set of facts. The question here—whether the evidence is about equal—was just too fact-like for an appellate court to conduct de novo review.

In dissent, Justice Jackson, joined by Justice Gorsuch, disagreed on both points. In her view, the statutory mandate that the Veterans Court “take due account” of the benefit-of-the-doubt rule should be understood as superseding the general standard of review found in the statute, thereby mandating de novo review. And even if one were to apply the baseline standards of review, Jackson thought that the Board’s determination about whether the benefit-of-the-doubt rule applied looked more like a probable cause determination, meriting de novo review. Although couched in the language of textualism, Jackson’s dissent relied heavily on legislative history, pointing to past drafts of the statute and testimony from veterans groups to Congress, all of which suggested that the whole point of the “due account” provision was to override the Veterans Court’s perceived record of being too deferential to the Board. Finally, Jackson bolstered her ultimate conclusion with the so-called veterans canon, which provides that statutory provisions for the benefit of veterans should be construed in the beneficiary’s favor. It is notable that Justice Gorsuch signed on to a dissent that made such heavy use of legislative history. Perhaps he simply thought veterans should get the benefit of the doubt.