According to a Times story from this week, another convicted felon has been shown almost certainly innocent by DNA evidence. “Convicted of robbing and raping an 83-year-old woman, Joseph A. Buffey at 19 … was sentenced to 70 years in the maximum security prison here in southern West Virginia.”
District attorneys continue to resist reopening cases for consideration of DNA evidence. This is understandable: the more doubt there is about the verdicts courts have reached, the less confident victims and their families will be that justice has been served and the less effective the system will be at deterring crime. At an extreme, convicted felons could appear to be the unlucky victims of a blind and random process.
They often are the victims of a blind and random process.
The felon in question was convicted on the strength of his confession. Perhaps the fact that he was “a marijuana smoker and high school dropout” also had something to do with it. To a certain kind of juror, the distance between smoking a joint and forcibly sodomizing an 83-year-old woman is not that great.
To me, all of these considerations point to the extraordinary flaws in the way criminal cases are tried. The issue of the character of the defendant hardly needs much discussion; it should be plain that, while some real relationship between certain traits such as schizophrenia and certain violent behaviors probably can be found, your average juror is not the person to find them.
The confession is just as much a problem, but perhaps this deserves more discussion. It seems clear to me that confessions should not be admissible evidence. They should not be admissible for the very reason that they seem to be strong evidence. But are they really?
I think the question was answered with some finality by the Duke of Brunswick in the 17th century. Europe was at that time still crazed with a fear of witches, and busy burning them. Many of these witches had confessed. From Charles Mackay’s Memoirs of Extraordinary Popular Delusions and the Madness of Crowds we have these:
- The rack was displayed, and the two poor creatures extended upon it. In reply to various leading questions from their tormentors, they owned, in their agony, that they were in the constant habit of meeting the devil, that they had sold their souls to him, and that at their command he had raised the tempest. (1487)
- In 1571, the celebrated sorcerer, Trois Echelles, was burned in the Place de Grêve, in Paris. He confessed, in the presence of Charles IX, and of the Marshals de Montmorency, De Retz, and the Sieur du Mazille, physician to the King, that he could perform the most wonderful things by the aid of a devil to whom he had sold himself. He described at great length the saturnalia of the fiends—the sacrifices which they offered up—the debaucheries they committed with the young and handsome witches, and the various modes of preparing the infernal unguent for blighting cattle.
- Gilles Garnier was put to the rack, after fifty witnesses had deposed against him: he confessed everything that was laid to his charge.
This was the state of criminal justice, as it were, that the Duke of Brunswick encountered. Mackay writes of him:
It is related of the Duke of Brunswick that he invited two learned Jesuits to his house, who were known to entertain strong opinions upon the subject of witchcraft, with a view of showing them the cruelty and absurdity of such practises. A woman lay in the dungeon of the city accused of witchcraft, and the Duke, having given previous instructions to the officiating torturers, went with the two Jesuits to hear her confession. By a series of artful leading questions, the poor creature, in the extremity of her anguish, was induced to confess that she had often attended the sabbath of the fiends upon the Brocken—that she had seen two Jesuits there, who had made themselves notorious, even among witches, for their abominations—that she had seen them assume the form of goats, wolves, and other animals; and that many noted witches had borne them five, six, and seven children at a birth, who had heads like toads and legs like spiders. Being asked if the Jesuits were far from her, she replied that they were in the room beside her.
The Duke of Brunswick led his astounded friends away, and explained the stratagem. This was convincing proof to both of them that thousands of persons had suffered unjustly; they knew their own innocence, and shuddered to think what their fate might have been, if an enemy, instead of a friend, had put such a confession into the mouth of a criminal. One of these Jesuits was Frederick Spee, the author of the Cautio Criminalis, published in 1631. This work, exposing the horrors of the witch trials, had a most salutary effect in Germany: Schonbrunn, Archbishop and Elector of Menz, abolished the torture entirely within his dominions, and his example was imitated by the Duke of Brunswick and other potentates.
It will be objected that present-day prosecutors do not use the rack. But they do use pressure, and that should be enough to contaminate the whole enterprise. We know empirically that confessions are unreliable, and we know that pressure (in addition to other psychological factors) can lead to false confessions. In what way are confessions evidence then?
But there is a more general problem, the profound unreliability of the judgment of those deciding the case–both the judges at the bench and the jurors. A recent study showed that judges are more and more likely to deny prisoners’ requests as they become more fatigued. A prisoner requesting parole will have the most luck if he is heard by the judge right after the judge’s break. This is a consequence of the fact that the human mind is simply not well-suited to tasks like deciding a person’s guilt or a person’s probability of committing crime. If a judge, a professional at making these decisions, is hardly competent to make them, what of a juror?
Trial by jury
The virtues of trial by jury are so fundamental that the framers of the constitution felt no need to state them. Alexander Hamilton wrote in The Federalist Paper #83:
It would be altogether superfluous to examine to what extent it deserves to be esteemed useful or essential in a representative republic, or how much more merit it may be entitled to as a defence against the oppressions of an hereditary monarch, than as a barrier to the tyranny of the popular magistrates in a popular government. Discussions of this kind would be more curious than beneficial, as all are satisfied of the utility of the institution, and of its friendly aspect to liberty.
This is true; the system of trial by jury is a great benefit to a Democratic society.
But it would be a mistake to conclude, as many do, that it is similarly beneficial to the individual defendant. Trial by jury protects the system from hereditary monarchs, and protects individual freedom fighters from persecution by monarchs, but in the more usual run of criminal cases, persecution by monarchs is not the main concern of your average defendant. Your average defendant has more to fear from the prejudices of jurors.
If you watch this YouTube video and pay attention to the juxtaposition of the juror’s reasoning about her verdict in the Rodney King case and the main piece of evidence in the trial, the video of four police officers beating the fallen King for 15 minutes, you will have to wonder what really motivated the juror’s decision. In this case, of course, the defendants were the police officers, and if prejudice was involved, the white police officers mostly benefited from it. Another case in which the jury was popularly suspected of being swayed by race was the O.J. Simpson trial.
Regardless of the merits of these individual verdicts, there is significant room for doubt, and the common disagreement of DNA evidence with juries’ verdicts makes more room. According to The Innocence Project, 301 convictions have been proven false by DNA evidence.
Toward a solution
In Oliver Wendell Holmes Jr.’s discussion of murder cases (in The Common Law), he shows that in order to determine whether an act that resulted in death was really murder or was manslaughter, we need to have some idea of the probability, given what the defendant knew, that the act would result in a death. It is the beginning of an acknowledgement of the epistemic limitations in which all parties are working. But even if it is only a beginning, it is strong: an implication of this consideration is that not even the defendant necessarily knows whether he is guilty of murder!
But Holmes points to a superior way of navigating this epistemic darkness: probability. In Holmes’s day, there could not have been sufficient data to make exact calculations of probabilities useful to a courtroom, but that is no longer the case. What is more, we may have the algorithmic tools to put them to excellent use.
In Thinking Fast and Slow, Daniel Kahneman extols the superiority of algorithms over human judgment for deciding everything from the prognosis of a cancer to the relative values of vintages of wine (see the chapter “Intuition versus Formulas”). Simple numerical algorithms yield more accurate results consistently. Why should the same thing not be true of the guilt or innocence of a defendant? Of course, there can be no single algorithm that could handle the multiple dimensions of an actual criminal case. But it may be possible to build a software package that could reliably put confidence intervals on a defendant’s guilt.
A juror’s job would then be one of quality assurance: A jury would hear several cases, and be given the probability of guilt in each case. The jury would then perform its constitutional duty of ensuring that the government was not corrupt by confirming that the estimates showed no political bias.