How I Actually Run a Paranormal Investigation

Right, so here’s a case I come back to whenever someone asks how I actually investigate, rather than how it looks on the telly. It’s an old one. Cock Lane, in the City of London, 1762. Most people who’ve heard of it only know the headline, a scratching ghost that accused a man of murder. The method behind sorting out what actually happened there is the same one I use on every case I write up here, so I thought it was worth setting it out properly, start to finish.

Start With the Earliest Account You Can Find

The account I lean on is in Charles Mackay’s Extraordinary Popular Delusions and the Madness of Crowds, published in 1841, which gathered up the pamphlets and newspaper reports printed at the time and set the whole business out in order. A young girl called Elizabeth, living in Cock Lane with her father Richard Parsons, began hearing scratching and knocking at night. The knocks were said to answer questions, one for yes, two for no, and through that code the ghost identified itself as Fanny Lynes, a woman who had died two years earlier, and claimed she had been poisoned by the man she’d lived with, William Kent.

That’s the earliest version, and it’s a detailed one. Named people, a specific accusation, a method of communication anyone present could test for themselves. That’s not evidence of anything paranormal, but it is interesting, because it means the story didn’t start vague and grow detail later. The detail was there from the start, which changes what I’m looking for next.

Check It Against the Record

Let’s go back to the record. An accusation of poisoning against a named, living man is the sort of claim you can actually test, and at the time plenty of people tried. A committee of gentlemen was put together to look into it properly, and Samuel Johnson was one of them. Johnson sat up through the night with Elizabeth to hear the knocking for himself, and the committee had Fanny Lynes’s body exhumed so it could be examined for signs of poison.

The exhumation found nothing to support the poisoning claim. That’s a record doing its job: it doesn’t prove Kent innocent on its own, but it removes the one piece of physical evidence the whole accusation depended on. When a claim rests on something checkable, checking it is the whole point. If the record had shown signs of arsenic, I’d be writing a very different piece.

Rule Out the Ordinary Explanations First

With the medical claim gone, the next question is simpler: where were the knocks actually coming from? The committee arranged for Elizabeth to be watched closely, away from the usual house, and the sounds stopped whenever she couldn’t move freely. Eventually she was caught with a small wooden board hidden under her clothes, which she used to produce the scratching and knocking herself.

Once you’ve got the mechanism, you look for the motive, and here there was a plain one. Richard Parsons owed William Kent money from an earlier loan and the two men had fallen out badly over it. A ghost that accused Kent of murder was a very effective way to ruin a man you had a grudge against without ever having to prove anything in a court. Parsons was later convicted of conspiracy over the whole affair and pilloried for it, along with his wife and another woman involved in the hoax.

What the Record Leaves You With

So that’s the method. Earliest account first, because a detailed origin tells you something a vague one doesn’t. Then the record, because a claim that can be checked should be checked before anything else happens. Then the ordinary explanations, the mechanism and the motive, because most hoaxes fall apart once you ask who had a reason to tell the story that way. It’s the same four steps whether the case is from 1762 or from last year, and it’s the same four steps behind every verdict I give on this site.

Cock Lane happens to have a clean ending: a confession, a conviction, a board found hidden under a girl’s clothes. Most cases I look at don’t tie up that neatly. The record runs out, a witness is sincere but can’t be checked, or the building’s history is too thin to confirm either way. I tell you when that happens too. What could it be, or is it a ghost? You decide.