The wrongful convictions in this archive share a bleak feature: almost none of them came apart through the ordinary machinery of appeal. Something outside the system intervened, usually far too late.
Nine other entries show the machinery working exactly as designed, and they are worth reading alongside those, because the result is not what you might expect.
In each of these cases a conviction was overturned. Not on a technicality in the dismissive sense, but on real defects of the kind that are supposed to matter. Rudy Bladel's statement to police had been improperly obtained. Joseph Christopher's competence to stand trial had never been properly assessed. Elmer Henley's trial had been held in the wrong venue. Leslie Irvin's conviction was overturned by the United States Supreme Court in a landmark ruling on pre-trial publicity, which is still taught. The Florida Supreme Court reversed Frank Athen Walls and ordered a new trial. An appeal court ordered a retrial for Christine Paolilla.
Every one of them was convicted again.
Bladel was retried and convicted, and died in prison in 2006. Christopher was retried in November 1985 and convicted again, receiving sixty years. Henley was retried at Corpus Christi in 1979 and convicted again on all six counts. Irvin was convicted again at his retrial. Walls was convicted again in 1992 and sentenced to death again. Paolilla was convicted again in 2015. Juan Corona, retried in Hayward in 1982 after a first conviction that had drawn heavy criticism, was convicted again of all twenty-five murders.
Seven appeals succeeded. Seven second juries reached the same verdict as the first.
Two cases in this group run the other way, and they show what the same machinery does when it is pointed in the opposite direction.
Angus Sinclair could not be retried for the World's End killings of 1977 because Scotland had a double jeopardy rule, and he had already been acquitted when a judge stopped the first trial. Scotland changed the law. He was retried in 2014 and convicted, thirty-seven years after the deaths. That is the only case in this archive where a legislature dismantled a protection specifically so that one man could be tried a second time.
Gilles de Rais was convicted in 1440 and executed. In 1992 a mock retrial in France, staged with historians and lawyers, concluded that he would not be convicted on the same evidence today. His conviction followed a confession obtained under threat of excommunication, in proceedings brought by men who stood to gain his lands. Nothing legal follows from a mock trial, but it makes the point that the system's confidence in itself is a function of its era.
Put the two groups side by side and the picture is uncomfortable in both directions.
An appeal that succeeds on procedure fixes the procedure. It does not disturb the verdict, and in nine cases here the second jury returned the same one, sometimes decades later, sometimes with less evidence than the first had. Meanwhile the men and women who were actually innocent. Kathleen Folbigg, Billy Glaze, Frank Lee Smith, Rolando Cruz, mostly did not get their convictions overturned on appeal at all. Cruz was convicted three times. Smith was refused DNA testing and died on death row.
The appeals system is good at catching the wrong kind of error. It exists to police how a trial was run, and it does that well enough that a defect gets a man a fresh trial for improper questioning or a prejudiced jury pool. What it is not built to ask is whether the person in the dock did it.