The power of mind

This post was written in 2006 by Karim N'Diaye. It was originally published on the Alphapsy blog.

Social psychologists from Harvard and Princeton campuses report on belief in magical causation from lay people (as far as students from those places can be considered as laid lay persons representative of the general population…). Reinforcing an already strong case for magical thinking, their studies yet deserve some attention for they try to draw a link between causal inferences in mental and physical domains.



Can one alter the course of external physical events by the sole power of one's mind? That is what Emily Pronin and her colleagues investigated. Well not exactly if one can, but if one thinks he or she can. In a series of lab and field experiments, they asked participants to think of some event which effectively occured, either because the event was under the experimenters' control or by interviewing a posteriori those of the participants that had made the correct predictions.



In the first experiment, particpiants were recruited to participate on a inquiry on non-conventional medicine and "psychosomatic symptoms". Each participant was paired with a confederate that behaved in the most contumelious manner (showing up late, throwing the information sheet at the dust bin and missing it, etc.). In the control condition, the confederate behaved neutrally. Of course, the subject was (open quotes) "randomly" (close quotes) designated as the one putting the curse on the other. After the so-called voodoo imprecation and the famous pin-stabbed doll, the confederate reported headache. Participants paired with the offensive confederate reported more evil thoughts and felt more causally responsible for the pain in the other than those paired with the neutral confederate. The interesting point is that despite physical actions were the same in both treatment groups (sticking pins in a voodoo doll) those of the subjetcs experiencing more ill thoughts also believed more in the causal relation between their acts (and thoughts) and the physical state of the victim. A similar effect was observed when participants were merely asked to generate evil thoughts towards a neutral victim.

In a second study, Pronin and her colleagues made "spectator" subjects to watch a blindfolded confederate making basket-ball shoots. Before each throw, those "spectator" participants were instructed to mentally visualize either bad performance (inconsistent condition) or good performance (consistent condition) — in a between-subject design. Another participant, the "witness", also read the instructions but was simply asked to watch the interaction between the spectator and the shooter. The shooter's success rate was similarly high in both conditions. When asked about their causal influence on the performance, mentalizing spectators reported increased influence in the consistent compared to the inconsistent condition, replicating the results from the previous study using positive rather than negative influence. Interestingly, witnesses were also more likely to see an influence from the spectator's thoughts in the consistent condition.

Two field studies nicely complement this lab research. The first used people attending a real basket-ball game and prompted, beforehand, to have either positive or neutral thoughts about the game. The last study inquired people watching the famous Super Bowl game about their thoughts during the game. In both conditons, thoughts consistent with the final results were perceived to have influenced the game. In the last case, supporters of the loosing team even felt responsible for having ruminated about the defeat.

This series of experiments thus provide conclusive evidence that even educated people not especially proned to superstition may express magical thinking such as mental causation. However their claim that this behaviour is rooted in a more general causality inference process needs further investigation. It is noteworthy that in all experiments, mental influence was directed at other people's mind (in order to alter either their behaviours or feelings). Quite a lot of speculation has been made by cognitive anthropologists such as Pascal Boyer on how religious thinking could be rooted in the very nature of our cognitive systems. One of this claim states that modular organization of the human mind leads to domain-specific inferential processes. I therefeore wonder if people would as readily express magical thinking when faced with cross-domain causation, such as thoughts altering physical properties of objects. Conversely, does adding acts (such as stabbing a doll with pins) to thoughts increase feeling of magical influence?

Another interesting issue concerns the inferential process itself, recent work in cognitive psychology has indeed investigated how the temporal relationships between two physical events lead to causal inference. Are these temporal constraints the same in perceived mental causation? Finally, it would be worth collecting personality traits in parallel with magical thinking tendency, as obsessive compulsion has been shown to be correlated with magical thinking in children and adolescents. Does this link hold in adults?

Old comments on this post:

Olivier Morin wrote: "These experiments are really nice ! It makes me think of the mechanisms for collaborative action investigated by Nathalie Sebanz (and many others). It can be shown that when you're engaged in a collaborative task, you tend to adopt the point of view of the person you're collaborating with, and mentally rehears her moves. Since mental imagery of action shares many properties of real action, it comes as no surprise that the mind should mistake other people's action for his own. That could explain Pronin's result in the base ball experiment, especially the fact that mental influence is felt as being stronger in the congruent condition. SO my bet would be that it is not domain-general causal cognition that is playing here, but rather doimain-specific mechanisms for joint action."

Alberto Masala wrote: There’s a well documented case that human mind is massively prone to illusions of authorship (see for example wegner “the illusion of conscious will”). For example, it happened to me several times to play an entire game at street fighter thinking it was me as Ryu against my friend as Ken, when actually the opposite was true…I can testify that the illusion is very strong and can last several minutes without any doubt. If there’s so a widespread problem with authorship in general, it’s difficult to interpret the experiments as being specifically about “magical causation”.

Olivier Morin wrote: I agree with you completely Alberto, both on the Street FIghter Illusion (I had exactly the same, which might explain my poor performance at this game…), and on the importance of distinguishing illusions of authorshipo from magical thinking. Anthropologists made a useful distinction between witchcraft and wizardry, that is relevant here. Witchcraft is power springing directly from the mind, whereas wizardry is power mediated by some object. Pronin's first experiment may be a case of wizardry, the second case of witchcraft.

Hugo Mercier wrote: I'm concerned about the significance of the results (at least those of the two lab studies). The authors find the difference they expect, but the baseline is very low. What I mean is that when people are asked to rate their causal influence on a scale from 1 to 9 (first exp) or from 1 to 7 (second exp), with 1 labelled 'not at all' [no causal relation], they are very close, on average, to 1 (2 in the control condition of the first exp, and 1.6 for the second). The authors should have given the variance, or the proportion of people who choose 1. It is important to know if the effect is due to each participant showing a very weak belief in their 'magical' causal power, or to most of the participants showing no belief whatsoever, and some of them showing a strong belief. Without these data, it's hard to figure out what are the actual psychological mechanisms explaining the results.

Karim N'Diaye answered: Alberto> As you might have noticed, Wegner is one of the authors. The witness in exp #2 is there to show that the effect doesn't stem from an egocentric bias or so as it survives in a third-person point-of-view situation. However, as underscored by Hugo, the authors go pretty fast on numerical data and the witnesses' responses make no exception.

No comments yet