I received a personal message from a paranormal investigator who thought it was a shame I didn’t believe in the validity of spirit communication. He pointed me to a video he made that he said was the clearest responses he’s ever received in an EVP recording. I’m always looking to either be impressed or spot an obvious hoax so I checked it out. Upon noticing that he had included subtitles in the video, I quickly put my hand up to cover the lower portion of the screen while watching so I wouldn’t be primed to “hear” what (he interpreted) the spirit voice had said. I did hear the first sound he interpreted, a very rapid “What?” in response to his opening inquiry because I saw the subtitle signalling it. It was so soon after his question, it felt out of place and I think I would not have noticed it had he not pointed it out. For the rest of the video, without the priming, I could not hear any anomalous voice, just background noise of insects or wind outside the structure. When I told him I was not impressed, he seemed stunned. To him, these voices were crystal clear.
I would not have thought much about this exchange again except it served as a great real life example of the concepts put forward in Daniel Kahneman’s Thinking, Fast and Slow. As I was reading it today, the well-tested observations of human perception and thinking habits he explained applied directly to paranormalists and their methods of reinforcing their paranormal worldview. Here are some examples.
In my book, Scientifical Americans, I call what paranormal investigators do “sham inquiry”. That is, they work from the conclusion (that paranormal activity exists) backward to place their collected data onto this preconceived framework and it appears valid. Everything they identify as an anomaly – environmental variables, sounds, smells, shadows – are strung together as a coherent story that supports their belief that spirits (or whatever they think the agent is) are the cause.
Kahneman talks about our two natural systems of thinking. System One is swift and intuitive; System Two is deliberate and does the calculating. System One incorporates available information (no matter how incomplete) and the active ideas of the individual (such as spirits cause bumping noises) into a coherent story. The quality and quantity of the information doesn’t matter. We naturally make these snap conclusions to save us the effort of thinking about it which would be a lot of extra cognitive work. System Two is harder to engage and can cause discomfort. Wrangling with doubt and uncertainty is much more taxing than assuming you know something and going with that decision. This is what I think paranormalists do: they reinforce their existing worldview with the personally satisfying story they create. Then, they vehement reject the comments from those of us who share our System Two output.
Paranormal investigators have been using basically the same methods for the past decade. They listen to the witnesses and use themselves as tools to examine the situation. They also use instruments that record variables or generate their own anomalies. Taken individually, these data would not be anything special. But strung together, collectively, the brain’s system creates a neat story of a haunted location. It’s a method that works very well for the individual. Response from others around him is typically positive. So they keep doing that same thing. No outside review or critique is needed because, in their view, everything is making perfect sense. As Kahneman describes it, what you see is all there is (WYSIATI). More information would spoil their awesome story.
The EVP session in my example above showed how the pro-paranormal investigator easily smoothed over the ambiguity in the noise to make something meaningful to him. He automatically clarified it to suit his needs and rejected my view that it was just misinterpreted background noise. To those of us with different active ideas (such as how easily we can fool ourselves and how equipment readings require adept interpretation), his is not a coherent story. The pieces don’t add up at all.
To be fair, one could say that I’m not open to hearing spirit voices, therefore, I don’t. EVP proponents should blind test their “unambiguous” recordings with a large sample of believers and nonbelievers. They would find that most wouldn’t hear anything on these recordings at all. But, again, that would wreck the story and they don’t want that.
Note also that EVP evidence violates the scientific norm of universalism. Everyone should be able to get a similar result with the same setup if the cause was consistent. Why don’t all the various ghost groups get the same voices saying similar things every time? Because it’s not really spirit voices coming through, it’s people finding meaning in random noise. We do this all the time. It’s normal.
Kahneman mentions a concept that can be truly insidious for human society: our susceptibility to influence. We are socially and mentally coerced to hear what we are told we should hear in an EVP (which is why I blocked the subtitles), and we are compelled to agree with the majority. In the case of paranormal investigators, they all share their experiences at a location and that input tends to coalesce into one coherent story as they place their pieces into building the preferred narrative. Often, the leader of the group will say they have heard, seen or felt something that prompts the others to report the same experience. We defer to a perceived authority and accept what s/he tells us we should be feeling. It’s not bad, really, it just IS. But when we are aiming for a reliable investigation, we must recognize and account for these biases in the conclusions. That’s why SCIENCE was invented – to try to eliminate these annoying, misleading biases!
If investigators really want to get the best answer, they have to engage System Two thinking. For a start, observations should be made independently and documented well before discussions with others. The data should be examined by unbiased reviewers (don’t tell them what they are supposed to observe before they are exposed to it). Make a checklist that deliberately requires assessment of the completeness of data and alternative explanations. Figure out what additional information would be needed to support a conclusion. Is the existing information a complete picture or just pieces “consistent” with what you wish it shows. Tailor the confidence in your conclusions accordingly. In most cases, the data provided by investigators is far too sparse, subjective, and unconfirmed to make any conclusion at all. Therefore, you can’t justifiably make that leap.