Clever Hans The Crime-Fighting Horse Is Getting Nervous

While I'm a big proponent of the idea---which I intend to hit on more in my next book that I'm writing now---that skepticism and liberalism should be strong allies, rarely do I get an opportunity to point out how woo can feed directly human rights violations. But via LGM, here's a story about how the woo-based belief that animals have supernatural abilities led to a series of unjust convictions, most of which may still be standing.

A wonder dog helped convict all three men: a German shepherd named Harass II, who wowed juries with his amazing ability to place suspects at the scenes of crimes.

Harass could supposedly do things no other dog could: tracking scents months later and even across water, according to his handler, John Preston.

If it sounds hard to believe, there's a good reason.

At least three men who've been convicted because of Harass and his owner's testimony have been set free after it was discovered that Preston's claims about his dog's superpowers were full of shit. To be entirely fair, I doubt that Preston, like many people who feel their animals can do things those animals can't do, may have been unconsciously signaling the dog to give him affirmations of what Preston secretly believed was true. Dogs may not be able to smell over water or smell someone's presence 6 months after the fact, but they can read their owners really well and produce behavior that will result in rewards off very subtle cues. It's the Clever Hans problem. The judge who finally forced the issue and had the dog's tracking abilities independently tested said that Preston's services came into play when the state didn't have a case, but felt like they had to have that conviction. This doesn't mean Preston was deliberately faking the results, but it's possible he inherited the prosecution's preconceived notions and sent unconscious signals to his dog. I'd in fact argue this is likelier than pure malice.

So far, the three cases---a murder, a rape, and a rape/murder---had no real evidence outside of the dog's confirmations, and in the rape case, the man convicted didn't even fit the victim's description. The problem now is that two cases were only overturned because the prisoners had legal counsel who demanded DNA testing to set them free. The state of Florida has shown zero interest in going back and looking at all the cases where the dog was used as evidence, even though they have a strong record of using this dog in lieu of real evidence. Who knows how many innocent people this dog has been used to finger?

I personally was shocked, though I shouldn't have been, that a dog can be used as an expert witness to a crime like this. Already dogs are way overused to overcome the limits put on human law enforcement's ability to invade your privacy, but obviously there should be some situations where hounds sensing something can be used to create enough evidence to conduct a search. Well-trained bomb-sniffing dogs come to mind, as does using dogs to track people who are trying to escape justice. But I fail to see any excuse for using a dog as a witness in court, especially since you need a trainer to interpret the signals who could be sending the dog unintentional positive reinforcement. This is basic animal psychology, not something that's hard to find out. In cases where a dog smells something, that should only, at the most, be evidence for a warrant, and if you don't turn anything up, too bad for you that your dog was wrong. Or, more likely, you were.

What this case also tells me is that people who're interested in skepticism need to think more about where the criminal justice system could be using woo to convict innocent people, and how we can help stop that from happening by demanding that evidence be vetted through scientific research and understanding as much as possible.