ABSTRACT
Rituals tend to be both causally opaque and goal demoted, yet these two qualities are rarely dissociated in the literature. Here we manipulate both factors and demonstrate their unique influence on ritual cognition. In a 2 × 3 (action type x goal information) between subjects design 484 US adults viewed causally opaque (ritual) or causally transparent (ordinary) actions performed on identical objects. They were provided with no goal information, positive goal information (“Blessing”) or negative goal information (“Cursing”). Neither causal opacity nor goal information influenced perceptions of physical change/causation. In contrast, causal opacity increased attributions of “specialness,” whereas goal information did not. Finally, goal information interacted with action type on measures of preference, such that ordinary actions are influenced by both “blessings” and “curses,” but ritual actions are only influenced by “curses.” These findings are interpreted in light of the Ritual Stance, and the cognitive bases of the effects are described with reference to Boyer and Liénard's hazard-precaution theory of ritualized behavior. The combined value of these two theories is discussed, and extended to a causal model of developmental ritual “calibration.”
Acknowledgments
Data collection for this study began the first week of February (2015) and writing began in June. We thank both anonymous reviewers for critical insights, and we acknowledge the contribution of the anonymous mTurkers who contributed to the data. Data and syntax for the present article are available at the Open Science Framework.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. A great deal has been written on action parsing independent of these authors. See Nielbo et al., Citation2013; Nielbo & Sørensen, Citation2015; Sørensen & Nielbo, Citation2013; Zacks, Citation2004; Zacks et al., Citation2007.
2. Examples of videos are available at www.rohankapitany.com
3. This suggests that the assumption of proportional odds was violated. When an MLR was run (and the reference category was set “Drinks Always the Same”), the only significant result (p = .037) was that opacity made participants about half as likely (OR = .542) to have alternating responses across presentations (but made no difference to reporting that objects differed across both presentations). However, this analysis violated Pearson's Goodness-of-fit statistic, χ2(6)= 17.190, p = .009. Inclusion of an interaction term (as discussed in the “Exploratory 'Analysis' section) returned no significant results. Thus, for the sake of consistency, and because no analysis appears superior to any other, the results of the OLR are reported here.
4. It should be noted that the measure of effect size presented here, the Pseudo-R2 value, “should be treated with caution” (Field, Citation2013). While it can be regarded as somewhat analogous to R2 in linear regression, many have argued the measure has issues. Thus, while the effect of specialness is small, it is reliably elicited by the described methods, as it is consistent with experiments 1 and 2 of Kapitány and Nielsen (Citation2015). A p-curve analysis conducted prior to this experiment suggested the effect has evidentiary value. A p-curve analysis of these prior experiments and the present experiment affirms this result. These data can be extracted from the stated publications, or are available from the corresponding author upon request.