Untestable explanations, magic and skill

Posted on January 28, 2008


In the Trondheim talks I put forward the idea that superstitions are the result of bias in the heuristics we use, the distinguishing feature being that, while the practices have a manifest function which is capable of being tested, the explanations which are offered for why this function is supposedly achieved are not, in themselves, effectively testable. The two reasons for the effective untestability of superstitious explanations I gave were the actual content of the explanations and the attitudes attending the superstitions. This way of thinking about superstitions leads to the conclusion that the category is not particularly deep, especially if one treats the explanations given as not particularly important. While this fits nicely the way in which belief in superstitions correlates with, for example, pseudoscientific beliefs, it does lead to a problem. The difficulty, pointed out to me by Trevor Case, is that, as Malinowski observed, people in traditional societies do tend to separate off superstitious practices from whatever rudimentary scientific knowledge they have, clearly distinguishing skill and magic and only using the second where the first is insufficient. In effect, the category of magic appears to be more important to people than my approach to it suggests that it actually is. The difference requires an explanation.

The initial intuition would be to see if the answer has something to do with why people chose to form untestable ontological claims. However, medical myths lacking in explanations and pseudoscientific claims which generally rely on at least prima facie testable claims, seem to fit along with superstition on one side of the distinction Malinowski talks about. While this allows me to retain the view that the category of superstition is not very important, it does leave open the question of what is so special about the broader category that contains both superstition and pseudoscience.