Do minds switch bodies? Dualist interpretations across ages and societies

ABSTRACT

Researchers explain cultural phenomena ranging from cognitive biases to widespread religious beliefs by assuming intuitive dualism: humans imagine minds and bodies as distinct and separable. We examine dualist intuition development across two societies that differ in normative focus on thinking about minds. We use a new method that measures people’s tendency to interpret ambiguous stimuli using mind-body dualist thinking. We recruited 180 Canadian children (2–10 yrs) along with 42 Indigenous iTaukei Fijian children (5–13 yrs) and 38 Indigenous iTaukei Fijian adults (27–79 yrs) from a remote island community. Participants tracked a named character within ambiguous animations that could be interpreted as a mind-body switch. Animations vary “agency cues” that participants might rely on for dualistic interpretations. Results indicate early emerging dualistic inclinations across populations and reliance on “agency cues” of body proximity and appearance of eyes. “Agency cues” increase dualist interpretations from 10% to 70%, though eyes mattered more for Westernized participants. Overall, statistical models positing that dualist interpretations “emerge early and everywhere” fit our data better than models positing that dualism “develops gradually with exposure to Western cultural traditions.” Fijian participants, who normatively avoid focus on minds, offered even more dualistic interpretations when they had less Western cultural exposure (via formal education).

KEYWORDS:

• Dualism
• mind-perception
• body-switching
• developmental
• cross-cultural
• psychology

Acknowledgements

Joseph Henrich would like to thank the Canadian Institute for Advanced Research (CIFAR) for their support. Rita McNamara would like to thank the Yasawan villages for their participation in this research. She would also like to thank the Psychology Department of the University of British Columbia in Vancouver, Canada as well as the School of Psychology at Victoria University of Wellington, New Zealand, for their support during preparation of this manuscript. We further wish to thank previous anonymous reviewers for their extensive commentary in helping us improve previous versions of this article.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Maciej Chudek http://orcid.org/0000-0002-6381-1703

Notes

1 All research and consent procedures were reviewed and approved by the University of British Columbia's Behavioural Research Ethics Board. In Vancouver, parents gave their signed consent and children their verbal assent to participate. In Yasawa, participants only gave their verbal consent because of variable literacy. Yasawan adults’ consent was obtained in three phases: first from the village elders and chief, then from each household, and finally from each participant prior to participation. Yasawan parents gave consent for their children to participate and Yasawan children assented. Consent and assent were documented on our data sheets. No data were collected without all three levels of consent, and we maintained contact with the villagers to allow them to revoke their data if they so choose after the study ended.

2 For simplicity, we calculate the predicted probabilities of indicating a body switch (pointing to the triangle as Penny) from the quadratic formula from the full data set, though the predictions from the exponential model are almost identical. The biggest difference between the two models is for younger adults in Yasawa. The quadratic model predicts a 59% probability that a 25-year-old Yasawan will indicate a body switch in the eyes condition, while the exponential predicts a 69% probability for the same individual. We do not have sufficient data to detect a difference between these models with age alone. However, as we discuss with the results on exposure to Western culture through years of formal Western education, adding education to these models adds significant prediction above and beyond age. Education also tends to be highest among young adults, which may indicate that the quadratic model may be a better theoretical model for this data.

3 An increase from a 4.4% pointing rate to a 50% rate is about a 22-fold increase in the odds of pointing. Dividing the natural logarithm of this odds ratio over the number of years over which the change occurs yields the predicted logistic regression coefficient.

4 One way to examine the relative fit of CAD and ID is to consider three possible variants of CAD. Experience living in a Western, Cartesian-influenced society (Canada) could cause the odds of switch-perceiving to increase by a small amount (OR = 1.2). The likelihood of this CAD model, given our data, is three times smaller than ID's prediction of no cultural difference. Strong variants of CAD fare even worse. A medium (OR = 2) cultural effect is 180 times less likely, a large one (OR = 3) is about 15,000 times less likely.

5 Keep in mind that we are not advocating for or trying to disprove either of these positions. Rather, they are rhetorical devices to more easily discuss what the regression coefficients in our models say about the age- and education-dependence of dualist inferences.

6 It is tempting to conclude that dualist inferences are a baseline of human reasoning, and that the very recent shift to scientific thinking, buttressed by formal schooling, is actually pushing reasoning away from our dualist roots. However, since we only studied two populations, our data do not permit this conclusion. It is equally plausible that there is something peculiar about Fijian culture that facilitates more dualist thinking – e.g., ancestor gods, spirits, and witchcraft concerns are a regular part of village life.