We develop a cultural evolutionary theory of the origins of prosocial religions and apply it to resolve two puzzles in human psychology and cultural history: (1) the rise of large-scale cooperation among strangers and, simultaneously, (2) the spread of prosocial religions in the last 10–12 millennia. We argue that these two developments were importantly linked and mutually energizing. We explain how a package of culturally evolved religious beliefs and practices characterized by increasingly potent, moralizing, supernatural agents, credible displays of faith, and other psychologically active elements conducive to social solidarity promoted high fertility rates and large-scale cooperation with co-religionists, often contributing to success in intergroup competition and conflict. In turn, prosocial religious beliefs and practices spread and aggregated as these successful groups expanded, or were copied by less successful groups. This synthesis is grounded in the idea that although religious beliefs and practices originally arose as nonadaptive by-products of innate cognitive functions, particular cultural variants were then selected for their prosocial effects in a long-term, cultural evolutionary process. This framework (1) reconciles key aspects of the adaptationist and by-product approaches to the origins of religion, (2) explains a variety of empirical observations that have not received adequate attention, and (3) generates novel predictions. Converging lines of evidence drawn from diverse disciplines provide empirical support while at the same time encouraging new research directions and opening up new questions for exploration and debate.

Today's major religions are moralizing religions that encourage material sacrifice for spiritual rewards. A key issue is whether moralizing religions gradually evolved over several millennia to enable cooperation among genetic strangers in the spiraling competition between increasingly large groups occupying Eurasia's middle latitudes, or whether they emerged only with the onset of the Axial Age, about 2,500 years ago, as societal wealth increased to allow privileging long-term goals over immediate needs.

Norenzayan et al. propose that religious beliefs with incidental prosocial effects propagated via a long-term process of cultural evolution. Applying their model, I explore a possible candidate target of cultural selection: the teleological view – often culturally elaborated as a belief in karma or fate – that life events occur to punish or reward individuals’ moral behavior.

Norenzayan et al. are praised for choosing to deal with significant questions in the understanding of religion. They are then criticized for refusing to define religion and for relying on problematic theoretical concepts. The authors discuss Abrahamic religions as the best-known prosocial religions, but the evidence shows that the case does not fit their conceptual framework. Finally, an extension of the authors’ ideas about the meaning of priming effects is proposed.

In their analysis, Norenzayan et al. completely ignore memetics, which, unlike other theories, treats memes as replicators and looks to memetic as well as genetic advantage. Now that memes are evolving ever faster, genetic advantage is less relevant. So when religious and secular values are at odds, we need a memetic analysis to understand what is going on.

The proposed narrative relies on an anachronistic projection of current religions onto prehistorical and historical cultures that were not concerned with prosocial morality or with public statement of belief. Prosocial morality appeared in wealthier post-Axial environments. Public demonstrations of belief are possible and advantageous when religious diversity starts interacting with coalitional recruitment dynamics in large-scale societies, a typical feature of modern, so-called WEIRD societies.

We suggest an alternative explanation for the emergence of Big Gods that places less emphasis on the role of cognitive tendencies and selection of prosocial cultural variants. Instead, we argue that the fundamental motivation to reduce uncertainty and increase long-term predictability provides a better account for the rise of Big moralizing Gods in a complex and heavily regulated social environment.

Norenzayan et al. argue that prosocial religion develops through cultural evolution. Surprisingly, they give little attention to developmental accounts of prosocial religious beliefs. A consideration of the developmental literature supports some, but not all, of the authors’ conclusions.

Norenzayan et al.’s account for the spread of monotheistic “Big God” religions sees these religions originating as by-products of innate cognitive biases. These biases produce polytheistic rather than monotheistic systems, however, and so do not explain the origin of monotheism. Accounts where monotheism arises from polytheism (for political reasons, for example) appear better able to explain the spread of monotheism.

We dispute the target article that belief in Big Gods facilitated development of large societies and suggest that the direction of causality might be inverted. We also suggest that plain theory of mind (ToM), although necessary, is not sufficient to conceive Big Gods. Grasp of other aspects of the mind is required. However, this theory is useful for the teaching of religion.

The gods of monotheistic religions, which began amongst pastoralists and defeated exiles, are closer to Big Gods than those associated with ancient city-based polities. The development of Big Gods is contingent upon a need to reduce uncertainty and negative feelings in combination with a relatively high level of prosociality, rather than a need to induce or assess prosociality.

Whereas Norenzayan et al. describe extravagant displays as a reliable means of belief transmission, this commentary reviews three emerging hypotheses about a direct connection between the awe elicited by extravagant displays and prosocial behaviour. If some of these hypotheses are correct, extravagant displays enhanced prosociality even among nonbelievers. Methodological suggestions are made for future experimental research on the awe–prosociality pathway.

Big Gods are described as having a “prosocial” effect. However, this conflates parochialism (group cohesion) with cooperation extended to strangers or out-group members. An examination of the cited experimental studies indicates that religion is actually associated with increased within-group parochialism, rather than extended or universal prosociality, and that the same general mechanisms underlie both religious and secular effects.

The target article develops an account of religious prosociality that is driven by increases in self-control. We suggest this account is incomplete. Although religion might increase prosociality to the in-group, it decreases it to the much larger out-group. Rituals, for example, lead to out-group derogation. We also challenge the link between religion and improved self-control, offering evidence that religion hinders self-control.

Norenzayan et al. suggest that Big Gods can be replaced by Big Governments. We examine forms of social and self-monitoring and ritual practice that emerged in Classical China, heterarchical societies like those that emerged in pre-Columbian Mesoamerica, and the contemporary Zapatista movement of Chiapas, and we recommend widening the hypothesis space to include these alternative forms of social organization.

The Big Gods model focuses on belief in an authoritarian God as a psychological mechanism that inhibits antisocial behavior and facilitates the formation of tight, cohesive groups. Recent empirical evidence suggests, however, that belief in a benevolent God is more likely to inspire helping and inclusivity. Both kinds of beliefs are necessary to explain the development of large-scale societies.

The theory of group-selected Big God religions is a master narrative of cultural evolution. The evidence is a positive manifold of correlated assumptions and variables. Although provocative, the theory is overly elastic. Its critical ingredient – belief in Big Gods – is neither necessary nor sufficient to account for in-group prosociality and discipline. Four specific issues illustrate this elasticity.

Norenzayan et al.’s theoretical synthesis is highly plausible and commendable. However, the empirical evidence for the arguments on mentalizing, cognitive biases, and religious belief is currently not as strong as the writers suggest. Although certainly abundant and compelling, this evidence is indirect, contradictory, and weak and must be acknowledged as such. More direct studies are needed to support the theory.

Cognitive by-product theorists maintain that standard cognitive development facilitates the acquisition of religion. Citing secularization, Norenzayan et al. qualify that theory, proposing that gods and good governments are psychologically and culturally interchangeable. That contention, though, occasions questions about the psychological dynamics involved, about what qualifies as religiosity, and about asymmetries between gods and good governments in the face of catastrophes.

The term prosocial has often been taken to mean nice or neighbourly, but many acts that further in-group interests are hostile and aggressive towards out-groups. According to Norenzayan et al., religion's ability to foster social cohesion within religious groups has been a key factor in the human transition to complex societies. But what are the prospects for nonparochial “religious prosociality”?