The cultural evolution account proposed by Norenzayan et al. does a nice job of integrating multiple lines of scientific research. We mostly agree with the authors’ theoretical framework. However, there are two points that have been overlooked in their model and that warrant further discussion. First, in considering the evolution of religious behaviors, specifically costly ritual displays, the authors focus on intragroup prosociality, but they have little to say about how religious ritual increases out-group hostility. Second, the link between religion, self-control, and prosocial behavior, as outlined in the article, does not account for recent neurophysiological evidence showing that religious mind-sets predict brain states associated with less (not more) self-control. We discuss these two points in relation to research from our lab and others.

Intergroup competition has helped shape the cultural evolution of religious belief and practice (e.g., Bulbulia Reference Bulbulia2004). According to the model, cultural pressures of intergroup conflict fuel prosociality among a group's adherents, galvanizing in-group ties and fostering solidarity. But how far does this prosocial behavior extend? This prosociality, the authors posit, is within the in-group, and a central feature of the authors’ model is that by prosocial, they mean parochial altruism (Choi & Bowles Reference Choi and Bowles2007), or affiliation and prosocial behaviors toward in-group members, coupled with hostility toward out-group members. It logically follows, then, that the current account is as much about out-group hostility as it is in-group cooperation. The cultural evolution of antisocial religions is the other, less appealing side of the coin and one that we feel has been overlooked in the cognitive science of religion literature (e.g., Atran & Ginges Reference Atran and Ginges2012; Ginges et al. Reference Ginges, Atran, Medin and Shikaki2007; Reference Ginges, Hansen and Norenzayan2009; Neuberg et al. Reference Neuberg, Warner, Berlin, Hill, Johnson, Filip-Crawford, Millsap, Thomas, Winkelman, Broome and Taylor2014) and, perhaps as a result, in the target article. The authors recognize that there is a dark side of prosocial religions and state that we ought to understand “the conditions under which prosocial religions become accessories to intergroup intolerance, conflict, and violence” (sect. 5.3, para. 3). Beyond this, however, there is little mention of the relationship between the prosocial and antisocial elements of religion. And although they hint at it in their model, the authors fall short in making explicit these divergent effects of in-group versus out-group, giving considerably more weight to the prosocial (i.e., in-group) element than the antisocial (i.e., out-group) one. A complete picture of religion is therefore missing.

Recent evidence from our lab shows that ritual – even mock ritual that is devoid of cultural meaning – leads to heightened out-group discrimination (Hobson Reference Hobson2013; Hobson et al. Reference Hobson, Norton, Gino and Inzlicht2015). In a series of studies, we show that ad hoc collective ritual is capable of promoting discriminatory attitudes and behaviors toward out-group members and that this bias is amplified as the ritual behavior (the sequences involved) becomes more effortful and onerous. What is more, these socially motivated out-group biases appear to be biologically rooted in the reward systems in the human brain, where group rituals appear to tune people to the punishment of out-groups. Across these studies, we find that extravagant ritual display, one of the hallmarks of prosocial religions, might in fact act as a signal of not only in-group allegiance but also of out-group hostility and separation. Religious rituals embolden the in-group, but by doing so maintain the sense of “us” versus “them” (e.g., Allport & Ross Reference Allport and Ross1967; Hunsberger & Jackson Reference Hunsberger and Jackson2005). More research is needed here to understand the full picture.

As a second point, the authors briefly outline the link between religious prosociality and self-control, taking the position that religion leads to prosociality through improvements in self-regulation. We question whether there is a direct link here between religiosity and greater self-control. For example, although widely discussed (e.g., McCullough & Willoughby Reference McCullough and Willoughby2009; Rounding et al. Reference Rounding, Lee, Jacobson and Ji2012), the evidence in favor of religion supporting self-control is thin. The association, for example, between religiousness and conscientiousness (a personality trait associated with self-control) may be driven by a person's need for orderliness rather than his or her industriousness or trait self-control (Eisenberg Reference Eisenberg1992) – a pattern of effects that has been found in conservative personalities (Hirsh et al. Reference Hirsh, DeYoung, Xu and Peterson2010). Moreover, numerous neuroscience studies in our lab have shown that religious primes predict brain states associated with less control (Inzlicht et al. Reference Inzlicht, McGregor, Hirsh and Nash2009; Inzlicht & Tullett Reference Inzlicht and Tullett2010; see Inzlicht et al. Reference Inzlicht, Tullett and Good2011 for a review). In a recent study (Good et al. Reference Good, Inzlicht and Larson2015), for example, we found that reminders of God's forgiving nature diminished the amplitude of the error-related negativity, an evoked brain potential thought to reflect performance monitoring, critical for control. We further found that such reminders decreased, not increased, actual behavioral control. Importantly, we found no evidence that reminders of God's punishing nature increased performance monitoring or behavioral control (even on a religiously important task), which directly contradicts the authors’ model of Big punishing Gods keeping people honest. Perhaps, then, religious prosociality (targeted at the in-group) does not come about because of simple increases in self-control, but through some other route. Recent fMRI work complements these findings, showing that certain features of religious interactions and group ritualized behaviors limit people's executive resources by narrowing the focus of attention toward emotional, low-level action units (Schjoedt et al. Reference Schjoedt, Sørensen, Nielbo, Xygalatas, Mitkidis and Bulbulia2013). Religious experiences turn down (not up) the brain's self-control system, making people less self-oriented and more likely to go along with the beliefs and practices of the group.

As a final, more general point, although the cultural evolution model provides a plausible ultimate explanation of the function of religion, it does little to address proximal explanations.

A comprehensive psychological theory ought to consider how ultimate, evolutionary accounts map onto the underlying proximate mechanisms. How does the authors’ ultimate account explain religious prosociality in terms of basic cognitive and affective processes? We think that much of the work in this field would benefit from using neuro- and psychophysiological tools to arrive at questions related to process. Indeed, if we are to agree with the authors’ view, then a methodological approach focused on proximal mechanism is needed.