Appealing to the minds of gods: religious beliefs and appeals correspond to features of local social ecologies

ABSTRACT

How do beliefs about gods vary across populations, and what accounts for this variation? We argue that appeals to gods generally reflect prominent features of local social ecologies. We first draw from a synthesis of theoretical, experimental, and ethnographic evidence to delineate a set of predictive criteria for the kinds of contexts with which religious beliefs and behaviors will be associated. To evaluate these criteria, we examine the content of freely-listed data about gods’ concerns collected from individuals across eight diverse field sites and contextualize these beliefs in their respective cultural milieus. In our analysis, we find that local deities’ concerns point to costly threats to local coordination and cooperation. We conclude with a discussion of how alternative approaches to religious beliefs and appeals fare in light of our results and close by considering some key implications for the cognitive and evolutionary sciences of religion.

KEYWORDS:

• Cognitive anthropology
• cultural evolution
• free-list method
• gods’ minds
• human behavioral ecology
• religious systems

Acknowledgments

We thank Adam Barnett, Nicholas Chan, and Tiffany Lai for coding the data. For critical comments on earlier drafts, we thank Martin Lang, Manvir Singh, Karl Frost, Richard McElreath, John Bunce, John Shaver, William Buckner, Center for Mind and Culture’s 2021–2022 Boston Colloquium on the Scientific Study of Religion, and the Religion, Cognition, and Culture research unit at Aarhus University. T.B. and B.G.P led project development, data checks, analysis, and writing. B.G.P. managed the greater project of which this was a part, compiled data, and organized coding of the free-list data and T.B. led analysis and ethnographic background research. J.H., A.N., and B.G.P. designed the Evolution of Religion and Morality Project. C.A., Q.A., E.C., R.A.M., B.G.P., A.K.W., and D.X. collected data.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

All data and code to reproduce the present study are available at: https://github.com/tbendixen/cross-cultural-free-list-project. The main project repository including raw data, full protocols, and related materials is available at: https://github.com/bgpurzycki/Evolution-of-Religion-and-Morality.

Notes

1 Note that our account does not assume that religious beliefs and behaviors are always beneficial for a community (see e.g., Edgerton, 1992). Just as cultural and ecological pressures can push a group to beneficial behavioral patterns, so too can cultural and ecological pressures (including time lag) push groups to sub-optimal traditions (e.g., Boyd & Richerson, 1990, 1992; Colleran, 2020; Richerson & Boyd, 2005, chapter 5; for a case study, see le Guen et al., 2013). Further, appeals to gods might point to false threats (e.g., witches), but such appeals are nevertheless framed in ways as threatening to the stability of the social status quo.

2 The full protocol, summary of methods, and descriptions of the cultural samples are available at: https://github.com/bgpurzycki/Evolution-of-Religion-and-Morality.

3 We also asked participants to list up to 5 behaviors that make someone a good/virtuous/moral person and up to 5 behaviors that make someone a bad/immoral person. For an empirical report on the results of these questions, see Purzycki et al. (2018b).

4 Note that Smith’s S scores are not standardized across domains or sites. Given variation in cross-site and domain sample size as well as the breadth of category (e.g., “Morality” encompasses more behaviors than “Ritual”), treating similar or the same values across domains or sites might be misleading. These values are therefore best appreciated relative to other coded items within their respective domains and sites. There are nevertheless global properties that become clearer by using this value, as shown in Figure 1. Since these global properties (i.e., the content of god appeals) are the focus of the present study, we refrain from pursuing any population-level inferences about the distribution of these properties, such as the probability of listing certain items as a function of some predictor(s). For suggestions on how to model free-list data more formally, see Bendixen and Purzycki (2023b).

5 For the police, note also the high salience of “Don’t know” among the Hadza and Inland Tanna, two sites that are generally unfamiliar with a formal and organized police force.

6 Note that some participants considered Haine and Ishoko to be identical entities, an observation that might account for the high moral salience of Ishoko as well as the general similarity between Haine and Ishoko (see e.g., Table 4). In cases where participants said that Haine and Ishoko are the same, the free-list data from Haine were duplicated to Ishoko, a decision made by the local field research team. However, to assess how this decision impacts the main results, in the Supplements we analyze and plot the salience of the general codes separately for those Hadza participants who said that Ishoko and Haine are “different” or the “same” (Figures S3 and S4). Analyzing these two groups of participants separately does not substantially change the main results. See Sections S3, S7, and Figures S3 and S4 for further discussion.

7 Caution is also generally warranted in inferring functional behavior from cultural beliefs in lieu of rigorous theory, analysis, and evidence (see Bloch, 1983; Elster, 1983, 2015; Shariff et al., 2014; Smith & Wishnie, 2000). A related methodological issue pertains to the non-trivial task of drawing rich inferences from free-list data; since free-list responses in isolation are often subject to several interpretative etic outcomes, insight into the local emic context is crucial for valid conclusions. For instance, in the Mauritius case, many participants listed food-related items (e.g., milk, bel’leaf, flowers, water) as something Shiva likes (see Supplementary Table S10), but since certain foodstuffs play a prominent role in Hindu ritualistic tradition, these responses were coded as Ritual according to the general coding rubric.

8 These preferences “represent a synthesis of experience accumulated over generations. Violations of spirit preferences can lead to accidents, falling ill, or worse. It matters little if the supernatural threat is real or not: if people believe in it, the threat of punishment becomes a real deterrent” (le Guen et al., 2013, p. 781).

9 Purzycki and Sosis (2022, ch. 10) frame the potential costs and benefits of social drinking as a social dilemma.

Additional information

Funding

This work was made possible by the Cultural Evolution of Religion Research Consortium, funded by a SSHRC partnership grant and the John Templeton Foundation (awarded to A.N. and J.H.). B.G.P. acknowledges support from the Max Planck Society and T.B. and B.G.P. acknowledge generous support from the Aarhus University Research Foundation.