We all know the classic scene below from Monty Python and The Holy Grail: The village mob drags onto the street a woman whom they've outfitted with a witch hat and prosthetic witch nose. They bring her before Sir Bedevere the Wise, claiming that she's a witch. Bedevere walks them through the science: Witches burn. Wood burns. Wood floats. Ducks float. Therefore, if the woman weighs less than or equal to a duck, she's obviously a witch.
A couple of centuries after the age of King Arthur, much of Europe began to engage in similarly ridiculous rituals to determine guilt in cases that lacked eyewitnesses or physical evidence. These rituals, called ordeals, were usually conducted in a church by high-ranking clergy. Three ordeals in particular were popular.
In the first, the accused was to retrieve a stone ring from a pot of boiling water. If his arm emerged unscathed, he was believed to have been protected by God, and proclaimed innocent. In the second, the accused walked nine paces while carrying a piece of heated iron. Again, if he managed to avoid severe burns he was deemed innocent. The third submerged the accused in icy water. If he sunk, he was believed innocent. If he floated, he was guilty.
Absurd as these rituals sound, their intent wasn't torture. Torture will produce a confession from anyone, guilty or innocent, if it's sufficiently harsh and unrelenting. Ordeals, on the other hand, really were intended to separate the guilty from the innocent. A fun, provocative new paper (PDF) from George Mason University economist Peter Leeson, now visiting at the University of Chicago, makes a compelling argument that ordealists were also uncannily good at it.
Leeson, author of The Invisible Hook, the wonderful book on the economics of pirates, explains that during the Dark Ages, belief in the ordeal system was strong enough that even most of the guilty bought into iudicium Dei, the idea that these rituals invoked the hand of God to condemn the guilty and protect the innocent. The believing guilty, then, faced an easy choice: admit guilt, be punished, and be spared the loss of an arm or the discomfort of frigid waters; or endure the ordeal, face the discomfort and possible loss of a hand or limb, and be punished anyway. Most of the guilty, Leeson argues, admitted their guilt before proceeding to the ordeal.
Only the believing innocent, then, would actually subject themselves to the rituals, knowing that God would protect them. Here's where it gets fascinating: Leeson argues that the clergy who oversaw the ordeals knew this, and consequently rigged the process to ensure that the innocent passed the tests. A priest, for example, might ensure that the pot of water was less than scalding by the time the accused submerged his arm. With the hot iron test, he could temper the fires to keep the irons at a manageable temperature before the accused walked the nine paces. In both cases, priests could make liberal use of Holy Water, perhaps turning a ritualistic sprinkling into a dousing to protect the skin of the accused. The ordeals also often took place in large churches, where the congregation was well removed from the act, giving the priest some leeway to rig the test. In both cases, it was up to the priest to determine if the accused had been scalded to the point of guilt, again opening the system to manipulation.
Leeson finds similar room for chicanery with the cold-water tests. The instructions for ordeal rituals were vague on the amount of time a suspect had to be submerged in water before a determination could be made on whether he floated or sunk, again giving clergy wide discretion. Interestingly, Leeson points out that lean men were more likely to sink (and thus be cleared) than women. Priests, then, were much more likely to subject men to the cold water treatment than women. Again, clergy were looking for ways to help them pass. Leeson found that where we have records, those records show that an overwhelming percentage of people who chose to endure ordeals actually passed them, certainly not what you'd expect if the the ordeals had been conducted "properly."
Interestingly, while our own post-Medieval criminal justice system functions best when those in power are subject to skepticism, checks, and balances, the ordeals system did best when the community unflinchingly believed the priests in power were driven by the hand of God. Within the surrounding community, the more fervently and unquestioningly the laity embraced the superstition, the fewer innocent people the priests would need to condemn in order to stave off suspicion. A completely devoted laity would accept a 100 percent exoneration rate. A more skeptical laity wouldn't.
So the success of the ordeals relied on the guilty believing that God wouldn't intervene to save them, the innocent fully believing that God would intervene, and a surrounding community willing to accept a high clearance rate for those who allowed themselves to be tested.
As Leeson explains, when doubt entered the system, the delicate balance was thrown askew. But while they lasted—up until the Church of England withdrew its support for the notion in the 13th century—ordeals were a more efficient, likely even more accurate, way of determining guilt than expensive fact-gathering and inquisitions (which were likely subject to their own forms of manipulation).
It's doubtful there are many lessons we can glean from ordeals today. We aren't about to return to a society so reliant on divine intervention. But Leeson's paper is a fascinating look into a system that, though driven by objectively false beliefs, not only produced surprisingly accurate results, but produced results that only became more accurate the more fervently the community believed.
Radley Balko is a senior editor at Reason magazine.