Comments on Cognitive factors underlying paranormal beliefs and experiences

Posted on February 21, 2007


For someone like me, who is setting out to do research on superstition, the French and Wilson article published in Tall Tales about the Mind and Brain is just about the best place to find out what kind of psychological research is being done. Of course, the article goes through the various areas at the speed of a two-week European tour – memory research on the left, probabilistic reasoning on the right. But, if you just want to find out what the basics are, this would seem to be a great place to start. Of course, I should add that not being a psychologist I am not in the position to spot problems that would be obvious to a psychologist. Having said that, there is a big issue that is close to my heart and previous research.

My worries with the article are at the level of the relationship between superstitious beliefs and rationality that is assumed by French and Wilson – just the kind of issue for a philosopher to get worked up about and for empirical psychologists not to be bothered by. Indeed, when I discussed the matter with Chris French he felt it was merely a case of different viewpoints without significant practical differences. I must beg to differ. French and Wilson talk about superstitious beliefs as due to cognitive biases, where these biases are seen as examples of human failure to reach some standard. This is, of course, the standard view of superstition so the authors are in no way unusual in assuming it. None-the-less, I feel that it is the wrong view of to hold and it is out of this worry that my limited problems with their article arise.

The view which I would argue for is to speak in terms of cognitive illusions rather than biases. The use of the term ‘biases’ suggests an unbiased norm that is veered away from. This is problematic as it is far from clear what this norm would be. The obvious but profoundly problematic answer would be that rational norms are to be read off the laws of logic. The view is problematic as despite many centuries of effort no-one has been able to come up with a satisfactory ‘dictionary’ and, indeed, once we invented systems capable of very effective use of logic, i.e. computers, we found them, to put it delicately, somewhat lacking in reason. Speaking in terms of cognitive illusions allows us to understand superstitions in a very different way. Here, the analogy is to perceptual illusions which, although capable of leading us into error, are generally due to the functioning of very useful assumptions made by our perceptual mechanisms. An example may be the human tendency to view pictures as representations of three dimensional scenes – our visual instruments assume that they are dealing with three dimensional input and given even weak clues interpret what is seen in that way. The significant and practical difference is that talking of illusions immediately suggest a way of explaining why people are superstitious. It would be hard to see why people should be biased to hold superstitious beliefs and even harder to see why they should persist with their biases once they are pointed out to them. However, if we think of superstitious beliefs are due to cognitive illusions that, themselves, are due to the functioning of normally useful assumptions it what used to be problematic becomes clear. People are superstitious as a byproduct of their relying of generally useful strategies/assumptions. Their superstitions tend to be maintained as they naturally fall back into using those generally effective strategies which then lead them back to the same conclusions – even though they are incorrect in that particular case.

The argument between what I am calling here the biases and the illusions view has been at the core of the recent dispute regarding how Kahnemann and Tversky’s work on heuristics should be interpreted – with Kahnemann and other representing the “these are failures to be fully rational” view and Gerd Gigerenzer and others representing the “that’s what it means to be rational” view. Clearly, I fall on the Gigerenzer side of that debate.

In the context of the paper, the most important thing is the issue of what is examined and what is not examined. The studies French and Wilson site standarly look into the differences between the characteristics of ‘superstitious’ and ‘sceptical’ people – identifying correlations between particular traits and superstitiousness. While this research is of course both interesting and useful it also has particular limitations we should be aware of. The first is that the correlations tend to be fairly weak. It isn’t that certain kinds of people believe in superstitions while others do not but, rather, that certain kinds of people are often only slightly more likely to believe a higher number of superstition. People in general are more or less superstitious, with the sceptical group being merely comparatively less likely to hold superstitious beliefs and comparatively more likely to reject them when they are identified as such. This aspect remains hidden in the paper due to the fact that French and Wilson do not generally discuss how strong the correlations they identify are. Given an illusions view, weak correlations is what one would expect against a background of fairly ubiquitous superstition. More importantly, on the illusions view it becomes natural to think of superstition not in an undifferentiated way but as specific kinds of superstitious beliefs arising due to the application of specific heuristics that lead to specific cognitive illusions. This, it seems, opens the way to a completely different way of examining superstition empirically.

The second important practical issue is that understanding superstition in terms of illusions rather than biases also opens the way to a research programme which could look at the evolutionary basis for superstion – this being spelled out in terms of the evolutionarily comprehendible selection of reasoning strategies which are powerful and useful despite their limited unfortunate by-products that take the form of succeptibility to superstion.