I occasionally get appointed to what are called “cold record” appeals, so called because the lawyer handling the case on appeal (that’s me) is not the same person, or from the same office, who handled the case at trial, plea, or sentencing. In those cases I have a letter I send that explains just what a direct appeal is and what it isn’t. Part of that spiel is that in the United States courts of appeals generally don’t review the facts of cases. They’re looking for legal error (for which the facts may be relevant), but generally the question of “did this person actually do it?” has already been answered and the court of appeals isn’t going to revisit that.

It’s called “finality” and it is, in many ways, the hobgoblin of our criminal justice system. It’s the idea that once someone is convicted of a crime it becomes really difficult to even get a court to look at the question of guilt/innocence, much less in a way that might actually result in a conviction being reversed. It’s something that a lot of lay people don’t understand and find frustrating – which I do, too!

I was thinking about this while listening to season one of the podcast Bone Valley.

The subject is Leo Schofield, who was convicted of the 1987 murder of his wife, Michelle. The podcast covers not only the initial investigation and Schofield’s trial (for which the evidence presented is pretty slight), but also his attempts to get a new trial after evidence turns up pointing to another man, Jeremy Scott, as the killer. Scott eventually confesses, but even that’s not good enough for Florida courts, who denied Schofield’s requests for relief (he was eventually released on parole in 2023).

Bone Valley does a pretty good job as a “whodunnit,” but there’s a gaping hole in the middle of it when it comes to the law. Nobody ever comes out and says just what Schofield has to show to get a new trial. What is the standard of review? Is it enough to show newly discovered evidence or does he have to show it’s of a kind that it completely undermines his conviction? Does it matter that Scott tells a half dozen different stories and isn’t particularly reliable as a witness, but appears to get some details right that only the killer would know? The show presents its case as if courts now are looking at the question of guilt anew, but they aren’t.

Let me give you an example. The state, of course, has a theory as to how Schofield killed his wife and can marshal some facts to back it up (not enough, in my opinion, but a jury disagreed). A couple of times Bone Valley dings courts for taking the state’s gloss on the facts as gospel truth, as if they’re just gullible doofuses.

Only that’s how review of facts in criminal cases generally work. On appeal I can technically challenge the sufficiency of evidence supporting a conviction, but in doing so the court of appeals has to take the facts of the case in the light most favorable to the prosecution. In particular, that means if a witness whose credibility was sketchy testified to X and X supports conviction, the court of appeals has to accept X as true.

For example, there’s a witness in Schofield’s trial who testified that she saw him loading something heavy, perhaps a body, in the back of his car the night his wife disappeared. Damning if true! Only her sister’s testimony suggested she had her dates wrong and that was really a couple of weeks before the murder (when, Schofield says, he was loading an amp into his car to go play a gig). You can choose to believe that witness or not, but courts of appeal can’t – her testimony supports the verdict and so, taking things in the light most favorable to the prosecution, her testimony is reliable.

Finality is, like all the various procedural hurdles inmates have to jump through to get back to court, about prioritizing an answer already gotten over making sure it’s the right answer. Antonin Scalia famously wrote in an opinion that there’s nothing in the Constitution that “forbids the exe­cu­tion of a con­vict­ed defen­dant who has had a full and fair tri­al but is lat­er able to con­vince a habeas court that he is ​’actu­al­ly’ innocent.” He wasn’t wrong – our system is designed to produce a winner and a loser, rather than struggle to get to the “right” answer (the same’s true in civil cases, too).

This is not how it has to be. Lots of other countries, usually working in variants of the civil/Napoleonic system, review the factual bases of convictions fresh on appeal. I honestly don’t know if those systems work out any better, on average, than ours when there are genuine questions of guilt, but at least the system isn’t structured to entrench mistakes in the name of not upsetting the status quo. What’s for certain is that any system that allows Alford and nolo contender guilty pleas (where a defendant pleads guilty even though they maintain they didn’t do it) or excludes reliable evidence because it was found or seized in the wrong way has other priorities than getting the ultimate question of guilt “right.”

All this is important if you really care about legal outcomes, not just the factual issues around the case. It’s one thing to ask, as a factual question, “did Leo Schofield kill his wife?” You can weigh the evidence however you want and make your best case. It’s an entirely different thing to ask, as a legal matter, “should Leo Schofield get a new trial based on newly discovered evidence that somebody else killed his wife?” The judges who will make the decision have law that constrains them in terms of how they view the facts of a case.

It’s surely not the best system the world has ever known. I’m not even certain it is (to borrow a phrase) “the worst . . . except for all the others.” But it is the one we’ve got, so know what it is you’re working with. Still and all, sometimes Dickens was right.