http://orcid.org/0000-0002-8499-7535
![Sample our Bioscience journals, sign in here to start your access, Latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5925/TOC-Subject-Sample-2021-Bioscience.jpg"})
ABSTRACT
Understanding the expansion of human sociality and cooperation beyond kith and kin remains an important evolutionary puzzle. There is likely a complex web of processes including institutions, norms, and practices that contributes to this phenomenon. Considerable evidence suggests that one such process involves certain components of religious systems that may have fostered the expansion of human cooperation in a variety of ways, including both certain forms of rituals and commitment to particular types of gods. Using an experimental economic game, our team specifically tested whether or not individually held mental models of moralistic, punishing, and knowledgeable gods curb biases in favor of the self and the local community, and increase impartiality toward geographically distant anonymous co-religionists. Our sample includes 591 participants from eight diverse societies – iTaukei (indigenous) Fijians who practice both Christianity and ancestor worship, the animist Hadza of Tanzania, Hindu Indo-Fijians, Hindu Mauritians, shamanist-Buddhist Tyvans of southern Siberia, traditional Inland and Christian Coastal Vanuatuans from Tanna, and Christian Brazilians from Pesqueiro. In this article, we present cross-cultural evidence that addresses this question and discuss the implications and limitations of our project. This volume also offers detailed, site-specific reports to provide further contextualization at the local level.
Acknowledgments
We thank Nicholas Chan, Tiffany Lai, Jonathan Lanman, Jessica McCutcheon, Caitlyn Placek, Edward Slingerland, Harvey Whitehouse, and Colin Xu. We also give special thanks to the editors at Religion, Brain & Behavior for their encouragement and hard work on making this special issue a pleasure to produce and to the anonymous reviewers for their invaluable feedback. QDA thanks a Rutherford Discovery Fellowship for support, JH thanks the Canadian Institute for Advanced Research for support, and AN thanks the James McKeen Cattell Foundation for sabbatical support. DX thanks the Velux core group “Technologies of the Mind” at the Interacting Minds Centre at Aarhus University, and Jakub Cigán, Eva Kundtová Klocová, Silvie Kotherová, Radek Kundt, Martin Lang, and Peter Maňo. The authors give extra-hearty thanks to Adam Barnett.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
Quentin D. Atkinson http://orcid.org/0000-0002-8499-7535
Notes
1. Note that religious and secular sources of trust and cooperation may be to some extent interchangeable. As indicated by the literature on secularization (Norenzayan & Gervais, Citation2015; Norris & Inglehart, Citation2012; Solt et al., Citation2011), when the effectiveness of secular institutions increases in some societies, and when wealth and decision-making power are distributed more equally across sectors of society, religiosity dwindles. Ultimately, the selection pressures that favor prosocial religions are minimized when effective secular social institutions take over the role of religious communities and services. Indeed, previous research (Hruschka et al., Citation2014; McNamara et al., Citation2016) shows that the more people feel the police are effective, the less they will cheat others to favor themselves and their local communities. Additionally, playing by the rules of experimental games increases as a function of both GDP per capita and government effectiveness (Hruschka et al., Citation2014). Resource insecurity at local and national levels as well as confidence in secular forms of social regulation such as the police should therefore positively contribute to the expansion of cooperation as well. But in the absence of effective secular monitoring, certain religious concepts may also alter the course of social interaction and may significantly contribute to this expansion. We return to this point below.
2. We note that a god’s status as being “High” or as a creator deity has no psychological potency in the expansion of the cooperative sphere in our theoretical model. Note, too, that the studies assessing the emergence of moralistic gods using the Standard Cross-Cultural Sample necessarily includes this constraint in their analyses. In other words, these studies assess the emergence of creator deities that may or may not be “specifically supportive of human morality” rather than concerned with morality regardless of their capacity as creators. See next paragraph.
3. Note that nonmoralistic gods or spirits may nevertheless prime latent moral cognition, perhaps by virtue of their agency and/or concern for human behavior (Piazza et al., Citation2011; Purzycki, Citation2013). Explicitly associating moral domains with gods’ concerns may expedite behavioral corollaries, especially if “good” behavior is a widely shared expectation of conduct endorsed by powerful agents.
4. This section heavily draws from the supplementary materials of Purzycki et al. (Citation2016c).
5. Note that we removed the Hadza data for these questions and their difficulty with scales suggest the data are suspect. We nevertheless include the data in our public data set for further examination.
6. As the Hadza have difficulties with continuous scales, these questions and response options were altered to accommodate them (e.g., “yes,” “no,” and “I don’t know”). These alterations include our basic religiosity questions, questions about the gods, material security questions, and perceptions of their religious ingroup, distant co-religionists, and religious outgroups. As such, where noted, all models considering such variables exclude the Hadza. Additionally, local god questions were not asked to Lovu participants because of ceiling effects and a lack of variation found in the Religious Landscape interview.
7. Because of practical problems with our original frequency scale, we altered the protocol during our field season. As such, two sites (Pesqueiro, Tyva Republic) answered this question on a five-point Likert scale: (4) Every day or multiple times per day; (3) A few times per week; (2) A few times per month; (1) A few times per year; and (0) Very rarely/never. Among the Tannese, Indo-Fijians, Mauritians, and Yasawans, we used a four-point scale: (0) Never; (1) Sometimes; (2) Frequently; and (3) All the time. Note, too, that at the Coastal Tanna site there was a clerical error in translation on the last two points of this scale (2 and 3) which rendered them virtually indistinguishable (see Atkinson, this volume). We therefore normalized individuals’ responses by dividing them by the maximum possible value on the scales used. For all regressions including this variable, we remove the Coastal Tannese from the analyses. Note, however, that in Purzycki et al. (Citation2016c), there were no major differences across models with or without the Coastal Tannese. As the questions regarding gods’ frequency of punishment of immoral behavior were the source of the confusion and not as central to our concerns as the amount of importance people claim their gods place on the punishment of immoral behavior, we focus here on the latter.
8. Note that this question was not asked of most of the Hadza due to their difficulty with more open-ended questions.
9. Because of coin visibility, we removed a single individual from the Local Co-Religionist Game from both the Coastal and Inland Tanna sites. As such, the sample size in this game is 589.
10. All focal models presented in the main text were run in R (R Core Team, Citation2012). For statistical coding scripts for R and STATA, see supplements for present article as well as those in Purzycki et al. (Citation2016c) and http://hecc.arts.sites.olt.ubc.ca/files/2016/01/Purzycki_et-al._Nature_Script.R.zip. Purzycki et al. (Citation2016c) report similar models, but without interactions and the OUTGROUP variable.
11. As material insecurity and material confidence were negatively correlated (r = −.41, p < .001; see Supplementary Table S1), and the Hadza answered different material confidence questions using a dichotomous scale, we focus here on material insecurity. See online supplements for correlation table for all target variables.
12. Elsewhere (Purzycki et al., Citation2016c), we report a variety of other analyses including sites as clustered robust standard errors clustering by site, and combined game models using individuals and clusters and mixed effects. Here, we report the most conservative models using sites as fixed effects and also emphasize intergroup variation in allocations.
13. Note that while the breadth of the confidence intervals for moralistic gods’ knowledge shifts across models, its coefficients do not seriously fluctuate. This suggests that its effect is stable, but it is not well estimated.
14. These data are drawn from the 2013 World Bank data at: http://data.worldbank.org/indicator/NY.GDP.PCAP.CD?order=wbapi_data_value_2013+wbapi_data_value+wbapi_data_value-last&sort=asc (accessed 10 April 2015).
15. Note, however, that if we regress allocations to DISTANT by standardized measures of government effectiveness, there are no obvious relationships in either the Local Co-Religionist Game (F (1, 6) = .54, p = .49, standard R2 = .08), or the Self Game (F (1, 6) = .38, p = .56, standard R2 = .06).