What is this thing called superstition? – the outline
Jahoda’s 1956 Psychology of Superstition remains an often referred to classic in the area. His ‘definition’ of superstition also remains relevant: “the kind of belief and action a reasonable man in present-day Western society would regard as being ‘superstitious’”. Of course, this is not so much definition as the honest avowal that we lack one and are, for the moment, forced to proceed largely on the basis of our intuitions as to what is a superstition.
A satisfactory definition will most likely have to await a consensus on what causes superstitions. None-the-less, in the meanwhile it is possible to take the substantive characterisations that have been put forward by superstition researchers and to relate them to each other in order to identify promising and not-so-promising lines of inquiry. The two main disciplines to look to are anthropology and psychology, with the examples from anthropology coming from the early 20th century and the psychological examples starting with Skinner’s 1948 work.
The element that comes through most unequivocally is that superstitions – beliefs as well as practices – form under conditions of uncertainty: a claim borne out by experiments (Vyse 1991) and historical studies (Padgett, Jorgenson 1982). Why they should do so, however, is a more controversial question. The most common anthropological explanation is that given by Malinowski (1925) who holds that superstitious rituals are a mechanism for reducing anxiety. Another theory is that the purpose is to communicate the willingness to cooperate in situations requiring intensive cooperation (Palmer 1989). Psychological explanations tend to focus on pattern-seeking as a general human trait which, in turn, may be given an emotional basis (Vyse 1997), an essentially game-theoretic one (Killeen 1977, 1981), and, finally, an evolutionary one.
One danger is that something like the naïve inductivist view of the pattern-seeking methods is accepted (Vyse 1997), even though studies of conditioning have shown selective associations are the norm (for example Cook, Mineka 1990) and the view is philosophically unacceptable (Hume 1748, Goodman 1955). Another is that the difference between superstition and potentially rational beliefs is made at the level of whether the beliefs are true (Maller, Lundeen 1933) – the problem being that this would obscure the difference between superstitious beliefs and other false beliefs as well as, in effect, denying the fallibility of the best of human reasoning.
A seemingly attractive way to make the distinction is in terms of superstition involving supernatural claims. The first problem with this view is that societies which do not distinguish supernatural from natural beliefs none-the-less identify superstition (Martin 2004) or, more neutrally, magic (Durkheim 1921). The second problem is that this fails to differentiate between superstitions and religious beliefs: unless one assumes that (the proper) religious beliefs are the true supernatural beliefs (Aquinas 1265). This distinction, however, may be made in terms of the institutionalisation or function of the beliefs or practices (Durkheim 1921, Wilson 2002). The third problem is that there is empirical evidence (Saher, Lindeman 2005) that people susceptible to superstitious beliefs which involve supernatural claims are also likely to accept pseudoscientific claims, as are made by homeopaths. This is compatible with the possibility that people only opt for supernatural explanations of patterns they believe to have observed when natural explanations are not forthcoming.
The most promising approach to differentiating between superstitions and other kinds of error has been to identify superstition is terms of the cognitive processes leading to it. This has usually taken the form of reifying the difference between superstitious and paradigmatically rational scientific beliefs in terms of a difference between modes of reasoning. Durkheim (1912) and Lévy-Bruhl (1910) saw the difference in terms of logical versus pre-logical thinking and Piaget (1929) in childhood versus adult modes of thought, with more recent writers also turning to dual-aspect reasoning (Epstein, Pacini, Denes Raj, Heier 1996) and essentialist accounts (Gelman 2003) of childhood intuitive reasoning (Hood, Bloom 2007, Lindeman, Aarnio 2006). These approaches share two weaknesses. The first is that, as the writers themselves often recognise, the difference between ‘scientific’ and ‘superstitious’ reasoning is not between mutually competing forms of reason but, at best, between mutually-supportive modes of reasoning that are usually applied together. The second is that, if Simon (1956) is correct, reasoning needs to fit the specific problems it is applied to – an issue ignored by these attempts to understand superstition. This leaves open the, thus far uninvestigated, alternative that superstitious beliefs and practices are the result of a mismatch between the modes of reasoning and the situations they are applied to.
Certainly the overly obsessive pattern matching behavior (al la Skinner) must play some part, for other intelligent animals display the exact same tendancy to notice false correlations more quickly than they can discard them: pigeons and dogs can spontaneously develop “superstitous” fixations, connecting particular movements and actions with the idea that they produce things they want, even when there is no connection.
The research on selective association is in the tradition that starts with Skinner’s pigeon studies, including his 1948 superstition study. Vyse’s work on how people look for patterns in random data is also an extension of that work but applied to people. His chapter on this research in the 1997 book is pretty good, not surprisingly.
Ritualized behavior – the set of which probably includes what you describe as ’superstitious practices’ performed with no real connection to their purported goal and that have been developed under conditions of uncertainty – is approached in an alternative way to Malinowski et al. and that of Craig Palmer and other (costly) signaling theorists. Specifically ritualized behavior is made up of rigid, stereotypic, goal-demoted, repetitive actions that are a behavioral output of cognitive systems triggered by perceived POTENTIAL danger. Most notably Boyer and Liénard in their 2006 BBS article entitled “Why ritualized behavior: Precaution systems and action parsing in developmental, pathological and cultural rituals.” Other researchers in varying disciplines are also interested in ritualized behavior as indicators of precautionary system activity include Henry Szechtman, Erik Woody, David Eilam etc.
Thanks for the suggestion, Joel, I’ll definitely follow up on all of the names you list. It so happens that I already have the Pascal and Lienard article on my desk, on top of my ‘to read now!’ pile. It was recommended to me just a week or so ago by Christophe Heintz. It is a shame that I will not be able to read and think it through before my talk at the KLI as it is in a few hours and I am yet to finish the presentation slides. As I keep saying, I should just read each issue of BBS through from cover to cover. Unfortunately, I am basically ignorant of the work on ritualised behaviour at this point.
Should one perceive the distinction between science and religion in terms of rationality/rationality or supernatural/natural?