Does Religion Promote Trust? The Role of Signaling, Reputation, and Punishment

Scholars of religion, including such luminaries as Durkheim, Rappaport, Turner, and Weber, have widely assumed that religion promotes intragroup trust among adherents. Recent applications of signaling theory to religious behavior among economists, cognitive scientists, and evolutionary anthropologists further endorse this assumption. However, trust has not been rigorously or consistently defined across authors, making generalizations difficult to evaluate. Here I follow Bacharach and Gambetta’s (2001) behavioral definition of trust and show that the conditions for intragroup trust are often not met in religious communities, especially isolationist and closed communities to which high levels of trust are typically ascribed. Rather, in such communities, cooperation is maintained through institutional structures that effectively punish cheaters and enhance the value of an honest reputation. These groups gainfully facilitate collective action by offering a circumscribed social arena in which reputations can be built, evaluated, rewarded, and efficiently punished. While face-to-face reciprocal relations obviate the need for trusting behavior within closed religious communities, when social groups are fluid, religious practices and symbolic markers are successful in promoting trust among in-group members and anonymous coreligionists who reside in different communities. In addition, these religious badges of identity may be used by non-group members as signals of trustworthiness. *I thank Candace Alcorta, Joseph Bulbulia, Sharon Feldstein, Bradley Ruffle, and two anonymous reviewers for very helpful comments on earlier drafts of this manuscript. I additionally thank the Russell Sage Foundation (Initiative on Trust) and the University of Connecticut for generous funding of this project. Sosis: Does Religion Promote Trust? 3 Trust none; for oaths are straws, men’s faiths are wafer-cakes ... William Shakespeare The Life of King Henry the Fifth (Pistol at II, iii) [A] group within which there is extensive trustworthiness and extensive trust is able to accomplish much more than a comparable group without that trustworthiness and trust. Coleman (1988: S101) While it is broadly recognized that both religion and trust are directly relevant to current world affairs, surprisingly little attention has been given to how they are related to each other. This is particularly notable because recent theoretical developments in the study of religion posit that trust plays an important role in generating the cooperatively derived benefits that religious groups are presumed to offer (see Sosis and Alcorta 2003). Economists (Berman 2000; Carr and Landa 1983; Iannaccone 1992, 1994), cognitive scientists (Atran 2002; Bulbulia 2004a, 2004b), and evolutionary anthropologists (Cronk 1994; Irons 1996a, 1996b, 1996c; 2001, 2004; Sosis 2003, 2004; Sosis and Alcorta 2003) have converged in applying signaling theory to model and explain religious behaviors that are puzzling from the perspectives of the egoistic-based models that serve as these disciplines’ foundations (e.g., rational choice, evolutionary game theory). These researchers have argued that religious behaviors signal group commitments, thus enhancing intragroup trust and facilitating collective action. Various authors working outside the signaling framework have also asserted that religion enhances trust among community members (e.g., Boudon 1987; Collins 2004; Shield 2002; Steadman and Palmer 1995; Weber 1958: 302–322). However, inconsistency in definitions of trust have limited the ability to make generalizations about this relationship. Even in our everyday use of the term, there is ambiguity. 1 Colloquially, we may employ the word trust to characterize the attitude of agents toward other agents. Thus trust sometimes denotes an attitude of confidence about another agent’s reliability. However, we also use trust to describe behavior. Accordingly, to trust is to act on this attitude of confidence about another agent’s reliability. The statement “Richard trusts Joseph” may refer to a specific attitude that Richard maintains about Joseph, or it may refer to an actual behavior that manifests itself as a consequence of the attitude—for instance, that Richard trusts Joseph with his wallet. In the discussion that follows, all references to trust refer to its behavioral usage, not its attitudinal one. This focus on the behavioral aspects of trust is motivated by the fact that observers of religion have generally commented on the remarkable levels of trusting behavior 1 In my trusty American College Dictionary, trust is defined as a noun, a verb, and an adjective, and 23 definitions are offered. No wonder this term has caused so much confusion! 4 Interdisciplinary Journal of Research on Religion Vol. 1 (2005), Article 7 exhibited by members of certain religious communities, with less emphasis on the more difficult to observe and measure attitudes and values that encourage these behavioral patterns. To address the question posed in the title of this article, I first offer a definition of trust that can be easily incorporated into signaling models of religious behavior and delineates between various types of trust. I clarify the benefits that enhanced trust has for individuals and groups and the conditions under which we would expect religious communities to exhibit high levels of intragroup trust. I then examine the prevalence of trust within several well-studied religious communities. To summarize the main argument, while many researchers have maintained that religion increases trust, I argue that this may in fact be illusory. Most of what observers claim is trust in religious communities, especially those communities that demand exclusive commitment from their members, is actually a consequence of institutional frameworks that effectively punish cheaters and support gains for those who maintain an honest reputation. These communities, which I refer to as “closed” (see below), can efficiently punish untrustworthiness and offer reputational histories because social relations are bounded. This circumscribed social arena is created by religious rituals and symbolic markers (e.g., clothing, hairstyles) that clearly signal and demarcate inand out-group membership. However, when group memberships are fluid and individuals maintain multiple group identities, punishment and reputational mechanisms are less influential. Consequently, these “open” religious communities often rely on religiously endorsed practices and symbolic markers to foster trust. Religious behaviors and badges also promote trust between anonymous in-group members residing in different communities and generate epiphenomenal benefits from increased perceived trustworthiness by nongroup members.