Review : The Witches (II)

Welcome back to summary-review of Stacy Schiff's Big Book Of Witches, an incredibly thorough (though often torturous) examination of the Salem witch trials. In part one I set out what made Salem different to other witch cases in my previous reading material : principally the routine claims of spectacular magical powers, the grand designs of a new world order, the extreme accusation and conviction rate. Now we need to look at why things ended up like this. Let's start with the court system, ordinarily able to keep the everyday fears of the populace in check... but in this case they actually made things much, much worse.

Evaluation of Evidence : Or, How To Catch A Witch

It's important to remember that the proverbial sort of witch hunts were rare events, and in Salem things seem to have gone exceptionally badly wrong. Even here though, the judicial system wasn't purely one-sided – it was more like a mixture of the good, the bad, and the exceedingly ugly. Often the boundaries were altogether blurry.

In Salem's favour, it should be acknowledged that not everybody was convinced of the evidence in all cases. A few even seem to have been almost diehard skeptics; doubt about the reality of witchcraft itself was in its infancy but it was in existence. If such fundamental doubts weren't much raised in Salem, then at least a few were concerned about all the specific cases and the difficulty of proof. Even some of the most virulently anti-witch were at least considering whether capital punishments were required or beneficial. Which is chilling in its own way... these people didn't go to knee-jerk "burn the witch !" chanting, but sat down and thought about it quite carefully (albeit in a deeply flawed way), and decided that hanging nineteen people was something they really wanted to do.

And there was a requirement, even among the most fervent, that evidence should be presented and persuasive. It was not enough to simply accuse, something of substance had to be shown. There were definite levels of evidence, with the hard, tangible sort greatly preferred. Confessions were also a gold standard : again, a mere accusation wasn't (quite) enough. Particularly contentious was whether hallucinations ("spectral" visions that only a few people could see) should count. Though not everyone dismissed such claims, everyone preferred the situation where there were multiple witnesses to supernatural phenomena; there was an awareness that it was possible some people were just seeing things.

But this is pretty much everything that can be said in favour of the Salem trials; beyond that they were legally about as rigorous as a puppy on cocaine. The impressions I get from Schiff is that the great and the good in Salem sincerely believed, by and large, that their standards were sufficient and they were driving at truth, but by any reasonable definition the legal "protections" afforded by the judicial system were scarcely more than a fig leaf. 

For one thing, the forensic skills of these imbecilic villagers was appalling, almost literally to the degree of not being able to tell arse from elbow : when examining internal organs, they had difficulty in distinguishing the heart from the stomach. Good luck with looking for supernatural marks if you can't even understand the most basic regular biology.

Confessions were treated with extreme confirmation bias, with no attempt to question their validity whatever. Incredibly obvious questions were overlooked : not only such basic issues as how powerful witches were apparently easily subdued, but also specific ones like how one woman claimed to be a witch for the last 40 years when she was only 38. The procedure was ludicrously one-sided and absolutely goal-driven, and it could also be corrupted to ensure that the judges' friends were never in danger of being in the dock.

While there were definitely standards of evidence, practically anything could count to some degree. Muttering, kneeling, expressing one's own virtues too loudly, being too strong... all could be blamed on witchy powers. If the court wanted evidence, evidence would be found : not manufactured exactly, but anything you said or did could be twisted to suit the agenda. The accused weren't given the evidence until the trial began, and even if they were found innocent of the crime they were initially accused of, they could still be found guilty of another crime at the same trial.

It look a while to understand Schiff's description of the Salem villagers as "logical", but by the end it becomes clear. Having accepted a "fact", they would be proceed to deduce its implications in great detail... but they would rarely question the validity of the information given them, and always steer the analysis to fit the agenda rather than looking for the most likely implications independently. They were intensely analytical but only rarely ever critical. 

For example, judges and juries might equally well decide that a story was too fantastic not to be true as it was to declare the opposite. Inconsistencies in accounts of multiple witnesses were first seen as problematic, then later "realised" to be actually quite helpful, since real people have faulty memories* – actually, they decided, these differences constituted evidence that something had indeed happened rather than the accusers collaborating.

* It's interesting to see the glimmers of a better approach to assessing evidence slowly coming through in Salem. But not only did this happen much too slowly to have any meaningful impact on the trials, it did so in a highly perverted way, with methodologies changing to ensure conviction (such as deliberately picking the most promising case to bring to trail first) rather than out of any interest in the truth.

This twisting, agenda-driven approach was one reason conviction rates soared. Another factor was at work which drove the accusation rates through the roof : the idea that criticising the procedure was itself evidence of witchcraft. Once that took hold, there was a runaway growth in the accusation rates. True, this "evidence" was hardly top tier, but it was enough.

The other factor helping both conviction and accusation rates reach absurd levels was very much Salem-specific. Whereas elsewhere a confession might save your immortal soul but would do nothing to save your physical body from the noose, here it almost always led to the capital sentence being rescinded in favour of a much lesser punishment, or even nothing at all. That alone is a clear, major factor why so many confessed to things that were absolutely impossible. Couple this with the horrendous conditions in which prisoners were kept, sometimes for months*, and the runaway witch hunt becomes all too easy to understand.

* Direct, deliberate torture was not quite absent, but was extremely rare. Keeping the accused in absolute squalor, in temperatures which were at first unbearably cold and then oppressively hot, with little food and less sanitation, might as well amount to much the same thing, however. The lack of deliberate torture to extract information in some ways makes the thing all the worse : ordinary people were willing to believe their neighbours were guilty of the most outlandish and horrifying crimes without recourse to forcing them to confess.

In many ways the Salem witch trials resemble the output of an especially crappy LLM : extremely confident, founded on falsehoods but carried to a logical, analytical, uncritical extreme; they were also of course heavily reliant on a great deal of literal hallucinations. One notion I rather like from programming is what I call thinking at scale. If you find a bug, first you look for typos, then you check things like whether a variable is being set or used correctly, then you move up to structural issues like whether loops are nested properly, and finally you might question whether what you've done is really a good idea after all. These people seemed to have incredibly narrow-scale thinking limited to a single level, a single premise that couldn't itself be much questioned. They could analyse to the nth degree, but never question the fundamental premise of what they were doing.

Another analogy might be the Borg of Star Trek. These are not a species, but something which happens to any suitable species that encounters Borg hive-mind technology : assimilated into a collective delusion. These kinds of witch hunts were not the result of some crazed individuals like Hopkins, but a process which took on a life of its own, under the right conditions able to escape the control of any one person. Salem was truly in the grip of mass hysteria.

---

All this covers most of the major points as to why, when the witch hunts began, they spiralled out of control. But the above doesn't really address the more fundamental question of why people were like this, why they thought and acted in the way they did, why they were so afraid of witches and why they began the trials in the first place. That's what we need to look in the concluding part of the trilogy, along with what happened when it all came to a sudden, spontaneous halt.