The Problem of Causation in Studies of Religiously Inspired Actors: Tracing the Cause and Effect of Interreligious Dialogue and Peacebuilding in Post-Conflict Settings

Abstract

This theoretical essay, inspired in part by the author’s own perplexities during fieldwork and post-conflict reconciliation efforts, draws on transdisciplinary studies of religion and peace science to trace the effects of religious actors and the causes of peace. The reoccurrence and determination of religiously inspired actors in post-conflict settings, where religion has historically influenced both war and peace, is rooted in the enduring effects of local religious systems. Yet the causes of these effects remain open questions. Answering them forces us to move beyond noting, for example, the interpretive association of religious actors and peaceful outcomes to inquiring whether and how religiously inspired actors cause peace. Since so much of contemporary peacebuilding depends on local religious leaders supporting peace efforts or re-traditionalizing rituals, scholars working in post-conflict settings often find themselves pulled into debates about the effects of religion on peace. An empirical task is therefore investigating the causal links between religion and peacebuilding, the effects of religious actors in their cultural context, and how to contribute to effective interreligious dialogue. To reach these ends, however, a theoretical path is needed. This essay offers such a path, firstly, by addressing faulty assumptions and residual problems in current scholarship, and, secondly, providing a pathway for making causal claims. That pathway includes aggregating commensurable data to distinguish sufficient-components in type and token cases, reconstructing the temporal order and likelihood of counterfactuals, and tracing biological universals and cultural knowledge from religious systems to peace events.

1. Introduction

It is nowadays conceded by many scholars that any satisfactory theory of religious peacemaking must give a more robust account than prior works on how religious actors effect outcomes (e.g., Appleby 2015; Isaacs 2016; Omer and Tarusarira 2023; Steen-Johnsen 2019). It is argued that unless such an account is supplied, there is no precise way of knowing that profound acts of peace and reconciliation happened because of religious actors; that alternative explanations are just as likely (Köbrich and Hoffman 2023, pp. 723–27). I do not dispute these claims. However, I wish to consider what it would take for a theory to give such a causal account in circumstances of interreligious dialogue, when violence prevention is so critical to peacemaking.

As peripheral as the question may seem, it is nevertheless timely. The manner and degree to which peace results from religious actors is a question raised in three contemporary narratives about religion. Though not yet a clarion call to action, the shared question points to a converging empirical gap.

The gap is perhaps most salient in the science of religion, which is itself an important historical reaction to the secularization thesis that came before it. Berger (1967), for instance, rebelled against phenomenological approaches that treated religion as sui generis. Going beyond Durkheim’s ([1912] 1995) thesis that “the gods are society writ large,” Berger (1967, p. 28) reduced religion to a social construct. Having run its course in human history, religion was giving way to secular forms of knowledge—a thesis about religion’s inevitable decline that Berger and others later retracted. In turn, analytical scholars in the late twentieth century pivoted to the science of religion, including how religion evolves due to its strong association with cognitive and group dynamics (for a review, see Shaver and Kavanaugh 2020).

Since then, it has become a mainstay for scientists of religion to code for the religiosity of actors as independent variables in multifactorial behavioral models. Yet a shortcoming persists. While these efforts have dispelled the secularization thesis and modeled how minds and rituals make religious phenomena socially effective (Boyer 2020), they have typically neglected actor-level causal accounts. It is for this reason that despite impressive breakthroughs in the cognitive and evolutionary sciences of religion (CESR), little is known about individual religious actors except for what is inferable elsewhere from overlapping yet disparate qualitative ethnographies (for an exception, see Xygalatas 2022). Many questions thus remain unanswered, including how religious actors vary within and between studies, how they wield rituals for social effect, and how they go about making decisions. Most critically, under what conditions do religious actors tip sociopolitical systems toward peace or conflict (Appleby 1999, 2015)?

In another popular narrative, fourth wave religious terrorism ushered in decades of research on religion as violence. The central thesis has been that religiosity is prone to conflict when it fuses individual identities with exclusive, transcendental ends (for a review, see Saiya 2026). Numerous scholars have empirically foregrounded how religious ideologies mobilize groups for violence (e.g., Whitehouse 2018). Over the last decade, however, several influential texts have called for greater emphasis on understanding the role of religious actors in public life (e.g., Odak 2022, 2024). In so doing, skeptics have harkened back to arguments that while ideologies and networks are significant for mobilizing distinct logics of enmity, religion itself is not a cause of violence (Purzycki and Gibson 2011). As the anthropologist Sosis (2020) has shown, it is rather a dynamic and complex system within which a set of actor-driven and context-dependent mechanisms contribute to group behaviors, whether peace or conflict. But this insight about religion’s complexities concerns systems, not actors. Without more empirical studies on who acts and how, we will continue to know little about religious actors as recurrent agents of social change (Koehrsen and Ives 2025).

Another narrative is religion as peacebuilding. Scholars like Gopin (2002) have challenged the thesis that religiosity is prone to conflict. Religion, he and others argue (e.g., Philpott 2012), is not a divisive force for most people but a source of morality that can motivate extraordinary ethical behavior. Advancing this claim, others have pioneered faith-based diplomacy, with the accompanying observation that religious actors can help in promoting positive peace (Johnston 2003). Empirical support for faith-based diplomacy is usually based on famous case studies including inter-church dialogues during the Troubles of Northern Ireland, faith-based activism against apartheid in South Africa, and Catholic-led interfaith resistance to authoritarianism in Central America (see Barton 2022). Others observe that religious actors have direct opportunities to influence peace through interreligious dialogue in critical regions such as Bosnia–Herzegovina, Israel–Palestine, and Myanmar (e.g., Odak 2022; Kmec and Ganiel 2019). Still, critics point out that without systematic empirical analyses, the claim that religious actors cause peace is anecdotal, if not ignores other factors, such as economic or institutional, that could just as likely explain peaceful outcomes (e.g., Glazier 2018; Svensson 2016).

Taken together, the three narratives converge on an empirical gap that motivates this essay (see also Toft 2025). Do religious actors, in fact, cause peace—and if so, who, how exactly, and to what degree? If religious actors cause peace in some contexts but not others, can we theorize in general that religious actors cause peace? Through what predictive mechanisms do they exert influence and how do we systematically account for outcomes?

To answer these questions, we must first address a core problem: most claims about religious actors causing peace, whether explicitly or implicitly, rest on two unchallenged assumptions:

• If we come to know what religious actors have done in key historical or ethnographic circumstances, then that knowledge can provide us with sufficient data for identifying the relevant factors of religious peacemaking.
• From (1) we can then make generalizable assertions for scholarly and policy-making purposes about how religious actors influence peace, which can then aid future peacemaking efforts.

As intuitive as these assumptions may seem, they are not jointly satisfied under prevailing approaches, and failing to interrogate them can lead to misleading claims and failed policies. Defending this thesis and thereby proposing a theoretical sketch for researching religious actors as causal agents is the general purpose of this article.

I start by motivating the problem. Most work on the role of religion in peacebuilding is done post hoc, usually in post-conflict settings, where researchers face the burden of reverse engineering the causes of peace and the possible effects of religious actors. To bring these difficulties to light, I reflect briefly on my own perplexities while conducting comparative post-conflict ethnography and observing reconciliation efforts in the postwar Balkans. I then address these difficulties by narrowing the definition of religious actors, reviewing current causal accounts, and showing that they rest on inconsistently defined terms, incommensurable data, and disparate methods. In doing so, I address an epistemological limit: although we can never demonstrate causation with certainty, knowledge depends on approximating it. I then propose a pathway forward by borrowing from medicine and epidemiology, which face similar uncertainty problems. Presuming the possibility of constructing layered mosaics of data, we can work between aggregated data of event-types and particular case study data of event-tokens to identify the sufficient components of peace. If religious actors turn out to play a recurrent role in peace events, then we can generalize that they in fact are causal components of peace—a potential finding with profound implications for the emerging field of peace science (Chenoweth 2025). Next, I propose how to explain religious actors as causal agents. The structure of equivalence in virtually every case in question is what Sosis (2020) describes as the religious system. Accordingly, we can approximate causation through the comparative process-tracing of transmitted knowledge and behaviors shared by religious actors and participants in religious systems linked to peace events.1

2. The Problem of Tracing Causation in Post-Conflict Settings

To illustrate the core problem of inferring the role of religious actors in the causes of peace, I reflect in this section on my own perplexities while working in the field intermittently over the last decade in the post-conflict Balkans.2 For sake of focus, I will not repeat ethnographic findings or observations of reconciliation efforts reported elsewhere (Kiper 2018a, 2018b, 2024). Instead, I recount the challenges I encountered—namely, four questions that are critical for systematically investigating the cultural, behavioral, and biological factors involved in a comparable phenomenon in diverse environments (Henrich 2016). After reflecting on these questions, I show how each connects to a residual problem in tracing religious causation post hoc, such as in post-conflict settings. By residual problem, I mean a latent assumption that is rather naturalized and easily overlooked in the discipline, despite its continued importance for advancing knowledge (Kuhn [1962] 2012; Quine 1951). I center my observations on Bosnia–Herzegovina.

2.1. A Preliminary Vignette: The Complexities of Religion in Postwar Bosnia–Herzegovina

On a wintery day in Sarajevo, my wife and I sit with a Bosniak (Muslim) veteran of the Bosnian War (1992–1995), which was the deadliest of the Yugoslav Wars (1991–2001).3 When discussing the effects of hate speech during the wars, the veteran provided an anecdote as to why he believed that religious actors caused violence. “There was a controversial priest,” he explained, “who came to Srebrenica, in front of cameras, who said, ‘This is the final blow of Serbian vengeance against the Turks.’ He said, ‘This town will never again belong to the Muslims nor the Catholics and will belong to us.” The priest then called on Orthodox Christians “to take revenge on the Turks,” which, the veteran thought, caused the Srebrenica massacre.

“But …” the veteran added, “there were positive examples, too—of priests who helped.” He then shared a story about an imam who was arrested with other Muslims by Serbian forces during the war but managed to call a nearby Serbian Orthodox priest, who rushed to the scene and successfully negotiated the Imam and the others’ release. The Serbian priest told his people, the veteran emphasized, that if they arrested Muslims, they should arrest him too. After all, he, like them, was guilty of nothing more than his religion. The veteran added that these actions sometimes carried over to peace events on the frontline, where local priests or imams would venture into so-called “no man’s land,” find fighters, and tell them to stop fighting and go home (Kiper 2018a, pp. 224–25).

This vignette is presented not as evidence for a causal claim, but as an example of the interpretive ambiguity that arises in post hoc accounts of violence and peace. For stories like those mentioned above later confused me when I observed post-conflict reconciliation efforts between Bosniak and Serb veterans (Kiper 2019). In those conversations, veterans avoided the topic of religion altogether given its post-conflict sensitivity. It was as if the incidents of religious violence eclipsed those of religious peace. But I learned that many veterans still saw the power of religiously inspired actors. Like many communities throughout the Balkans, they looked back on historical moments of interreligious dialogue and tolerance as examples of Catholics, Muslims, and Orthodox Serbs living together in peace. Still, none of my informants could explain how religiously inspired actors in interreligious dialogue caused peace—and rather embarrassingly, nor could I.

The problem was that I had read scores of inspirational accounts about religious actors involved in peace. However, the plentitude of correlational observations offered little direct or commensurable evidence linking religious actors to effects. Looking deeper into the literature, I found that I was not alone in my perplexities.

2.2. Who Counts as a Religious Actor?

A perennial question is who, or what, counts as a religious actor. Following Lederach (1997), one would assume that religious actors were collective agents—ranging from institutions to individuals—who operate within a religious tradition at one of three levels. There were the top level of leaders, the middle level of community organizers, and the lower level of followers or grassroot initiators. Given this understanding, a religious organization comprising all three—and focused explicitly on peacebuilding—would constitute an ideal set of religious actors to study. However, this broad definition complicates identifying cause and effects.

To illustrate, in Bosnia–Herzegovina, in 1997, leaders of the Muslim, Serbian Orthodox, Catholic, and Jewish communities established the Interreligious Council, with the stated goal of reconciling the country.4 Since then, the council has promoted interreligious dialogue through NGOs and state institutions. Arguably its highpoint came in 2004, when the council advanced the Law of Religious Freedom in the Bosnian parliament to protect religious sites nationwide. Nevertheless, observers have criticized the council for its top-down approach and positioning itself around legal and legislative goals, rather than substantial interreligious integration (Šarenac and Mandić 2024). But the truth is that little is known about the effects of the council. On the one hand, there is an absence of evidence linking it to peace outcomes other than passing laws, but that does not entail evidence of absence. On the other hand, the effects of interreligious dialogue in Bosnia–Herzegovina, in general, remain open questions for scholars due to their complexity.

In one of the most robust studies, for instance, Ringdal and Ringdal (2010) surveyed over 3000 Bosniaks, finding that private religious practices buffered against postwar stress, but people who engaged in communal interreligious activity, like the council’s, suffered higher rates of stress. This finding is perhaps not surprising given the nature of ethnoreligious mass violence during the Bosnian War (Sells [1996] 1998). Yet, its meaning is difficult to parse without data. With an eye toward empirical research, the root problem is having too many loose definitions, such as “religious actor,” which are treated as if they were self-evident, when in fact religion, ethnicity, identity, and politics are entangled and not easily separable in post-conflict settings like the Balkans (Helms 2013).

2.3. How Do We Avoid Faulty Comparisons?

Another question is how to compare religious actors across the subcultures of Bosnia–Herzegovina—let alone the Balkans or world regions—without making false equivalences. Adopting Appleby’s (1999) viewpoint, we might presume that similar wartime conditions in different cultures would warrant a systematic comparison. However, a common mistake in comparative anthropological research is imposing a faulty comparison of categories from one culture to another—a worry that is rather acute for scholars working in the Balkans where faulty comparisons have led to disastrous outcomes (Žižek 1999).

Since the Yugoslav Wars, for instance, policymakers have classified the Balkans as “conflict prone” due to the risk of separatist threats in Bosnia–Herzegovina and Kosovo, where wartime violence turned genocidal. But it is precisely because of the Srebrenica genocide and ethnic cleansing in Kosovo that the Balkans gets compared to similar war and postwar regions such as Rwanda, Cambodia, and the Sudan. These comparisons, however, often result in false equivalences that overlook divergent cultural histories, ignore structural violence, and rely on false analogies that have contributed to questionable top-down international policies, including failed humanitarian interventions (Petersen 2011).5 The same is often true for studies today that compare countries based on post-conflict conditions and thereby institute the same professional liberal-peacebuilding practices, despite the unique micro-level factors involved in everyday peacemaking (Eltringham 2021; Simić et al. 2012).

A similar worry occupies the attention of scholars involved in interreligious dialogue. Without controlling for numerous variables, a systematic comparison is unlikely and such studies risk making spurious observations or correlations (e.g., Vilà et al. 2020). Worse still, the outcome of inferring effects from faulty comparisons of religion, post-conflict justice, peacemaking, or reconciliation—terms that are not universal or stable categories—could propound biases, if not support unhelpful or even harmful policies (Asad 1993).

2.4. Can We Generalize That Religious Actors Cause Peace?

An additional question is whether we can generalize that religious actors are causes of peace, even when we have evidence that a particular actor, or several actors, have played roles in peaceful outcomes. I am not saying here that religious actors are not causes of peace—far from it. Religiously inspired actors are often transformational in peacebuilding (Appleby 2019, pp. 197–98; Omer 2015, p. 3). But inferring when, how, and why they are effective alongside other factors of conflict prevention remain significant questions, especially for the emerging field of peace science (Chenoweth 2025). Furthermore, some religious actors are not agents of peace but drivers of conflict, while others strive for peace but may not achieve it.

Illustrating this very point, Odak (2024, p. 18) observes that many religious leaders in Bosnia–Herzegovina have long associated themselves with antiwar statements and active peace projects, but their efforts are repeatedly curtailed by miscommunication or divisions within and between organizations. If anything, then, religious actors across the post-conflict space are motivated to become better agents of peace, but whether they are effective remains a question even for themselves (Odak 2024). Hence, asserting a causal link between well-intended religious actors and peace risks making a hasty generalization.

Such a generalization could also be categorically wrong. Despite the widespread belief that interreligious dialogue is a reliable causal force in peacebuilding, extant data suggests otherwise (Driessen 2025). Many religious actors involve themselves in peace processes simply to legitimize authorities and state-building rather than peace (Fahy 2018; Wilson 2001). Moreover, interreligious dialogue, despite its widespread support, often proves ineffective and disempowers other local peace initiatives (see also Omer 2023). Religion may thus function as a solution, a cause, a contributing factor, or a part of any peace process, but without robust causal analyses, we cannot know with much certainty (Driessen 2025, p. 13).

2.5. What Are the Mechanisms?

A longstanding question concerns the mechanisms that religious actors use to initiate peace. Putative factors typically include moral authority, cultural intelligence, and grassroots legitimacy. Beginning with moral authority, religious actors, as ethical leaders in their community, find themselves as trusted facilitators in political efforts, during which time they can (re)frame events toward peaceful ends (Driessen 2025, p. 6; Sandal 2022, p. 552). Local actors are also more likely than outside peace negotiators to possess the cultural intelligence to adapt in their native setting, which is often necessary for enabling dialogue and accountability (Kwuelum 2024, pp. 3–6). Likewise, religious actors are common to peace efforts around the world with grassroots legitimacy—although their role in successfully re-traditionalizing conflict management remains debated (for a review, see Eltringham 2021, pp. 45–67).

The problem is that these observed factors are not mechanisms per se, in which an organized system of elements and processes predictively precede a phenomenon. They are rather typologies of activity: descriptive categories of what religious actors do. Where a mechanism explains how something happens, a typology names what is observed. Unless we can move from observation to probable prediction, we face a gap between describing what may be a cause and identifying likely causal processes with relative predictability.

Returning to Bosnia–Herzegovina to illustrate, the Interreligious Council did many things after the Dayton Accords to advance interreligious dialogue, including efforts to work in communities where framing peace in religious terms was inviolable. In doing so, they served as facilitators of political efforts, interreligious dialogue, and laws. However, the Council appears to have had little influence over the many local religious actors who have supported negative peace but openly derogated religious others (Šarenac and Mandić 2024, p. 87). Such occurrences are not uncommon in post-conflict settings where religious actors demonstrate what Odak (2021) calls “theological dissonance.” This describes a mismatch between what actors say or do in favor of negative peace in one context, and what they say and do against positive peace in another (Odak 2021, p. 248). Resolving this problem dovetails with facing up to the challenges of tracing causation post hoc: both require that we complement the rich tradition of descriptive accounts in religious studies with possible causal pathways to behavioral outcomes.

2.6. A Theoretical Sketch

In what remains, I attempt to answer the above questions, respectively. Though my answers are more philosophical than empirical, they complement an emerging call for greater precision in studying religious actors. As such, my sketch is not an empirically driven study but an attempt to reveal, through extant work, what deserves further explanation in future studies. Across these examples, a consistent theme emerges: peace and violence in post-conflict settings are irreducibly multicausal, and any attempt to isolate religious actors as singular drivers risks analytical distortion.

3. Religious Actors and Interreligious Dialogue

It is convenient to begin with a key operational definition. By “religious actor” I do not mean just any organization that articulates religious meaning, community with shared religious commitments, or institution influenced by religion. For doing so would entail that any sociopolitical way of life characterized as “religious” would constitute a “religious actor.” But that approach, which is all too common, treats religion as a causal agent, and thereby collapses the distinction between religious systems and the actors who operate within them. After all, if religious actors can range from organizations to institutions, then any system of religion is an actor. But that is analytically misleading since only individuals possess intentions and make decisions, and collective entities act only through the coordinated actions of people within them. Further still, treating anything religious as a religious actor can descriptively homogenize diverse traditions and misinterpret political factors as religious.

To push the point, Roy (2017) forcefully argued that European jihadism was not the radicalization of Islam but the Islamization of radicalism. However, his thesis rested on equating individual Salafi dissidents who adopted violence for political purposes with the dangers of Salafist teachings (Dawson 2022). As a result, Roy’s definition depoliticized the motivations of the individual jihadists, collapsed the diverse traditions of Salafism into a monolithic frame, and confused religious actors for their religion.

Broad definitions, then, can be misleading if our goal is to understand and compare effects. Anthropologically, all actors are enveloped in cultures that are, in turn, contested, hybridized, and entangled with others. That is why, unlike other disciplines, anthropologists seek to understand how actors change their environment through culture rather than merely reproducing it. To that end, the elements of culture to be discussed here include religious systems and the shared knowledge that actors use for interreligious dialogue.

For greater precision then, I adopt a narrow but recognized definition: religious actors are individuals who operate on overtly religious principles or pursue certain goals in a publicly acknowledged religious capacity (Sandal 2017, p. 4). Whether acting individually or collectively, they invoke religion as the basis for their actions. The upshot of this definition, notably for post-conflict settings, is that it avoids the categorical error of mistaking religion for an actor, while allowing causal attribution to hinge on intentional action rather than social identity.

Pluralist naysayers aside, such an approach fits easily within the problem of causation and religious actors. Individual religious actors are routinely described as significant contributors to peace (Appleby 2015; Omer and Tarusarira 2023; Svensson 2016). The problem with this description is not so much a lack of general data, insofar as historical accounts are plentiful. Rather, the data we have are heterogenous and lack a defensible causal theory. There is also no consensus on how to attribute causal efficacy to religious actors in a way that is analytically precise and methodologically consistent (Steen-Johnsen 2019).

Of course, these problems are not unique to religious studies, but they are critical for interreligious dialogue. Interreligious dialogue is here understood as the intentional encounter between persons from different religious traditions for cooperative or peaceful ends (Howard 2021, pp. 6–7). Unlike religious actors, the chief intellectual problem with interreligious dialogue is not its definition but generating commensurable data. To illustrate, in a pivotal article, Lehmann (2020) underlined the importance of having databases like the Swiss Metadatabase of Religious Affiliation in Europe (SMRE) for comparative research. Without such comparisons, scholarship risks focusing on the endeavors of individual religious actors as a hermeneutic process, from which findings often get generalized for transregional peacebuilding (Lehmann 2020, p. 238). The problem is that hermeneutical interpretations are context dependent. And contextual interpretive data is, alas, valuable for contextual meaning, but is not reliably generalizable across regional differences.

4. Avoiding False Equivalences

Following Lehmann (2020), if we set for ourselves the task of moving beyond isolated interpretations and having commensurable data, then we must avoid the problem of false equivalence. In anthropology—a leading discipline for cultural analysis—scholars begin by asking whether comparing a phenomenon across different contexts identifies the same thing or simply uses the same word for different phenomena (Candea 2019). If we adopt a family resemblance to determine comparisons, we risk losing analytical precision. But if we revert to a strict functional account, we risk imposing faulty etic categories. The standard solution is to draw from multiple ethnographic and interdisciplinary perspectives to start with a middle-range concept (Merton 1968). Critically, a middle-range concept is not the same as an operational definition. Where an operational definition specifies how to observe a measurable phenomenon, a middle-range concept explains a patterned theoretical phenomenon. In scientific studies, the two go together to solve the problem of the criterion, which is that in any investigation, we are caught between method and theory and thus operational definitions and middle-range concepts—and as I address below, types and tokens (Chisholm 1973).

Here, I presume that the primary middle-range concept needed for avoiding false equivalence is not only religion (due to its logical fuzziness and ambivalence in both inciting conflict and inducing reconciliation; see Appleby 1999, p. 19) but also causation. For causation is the middle ground where science operates; it is the concept that allows us to observe and explain the world in multiple contexts (Vayda and Walters 2011, p. 5). Below, I draw from existing literature on interreligious dialogue to infer a working middle-range concept of causation, which I then develop further in subsequent sections.

4.1. Historical Counterfactuals

A strong sense of causation in many accounts of interreligious dialogue is counterfactual reasoning, in which religious actors are shown to be observable difference-makers in explanations of peace (e.g., Alfani 2019; Harvey 2016; Kmec and Ganiel 2019). Oftentimes, this approach parallels Bradford Hill’s (1965) guidelines for establishing evidence in epidemiology. But where epidemiologists attempt to explain causation through the mechanisms of pathology, the religious scholar explains causation through the intentional actions of agents. To advance this line of thinking, I shall adopt Bradford Hill’s (1965) framework. If our unit of analysis concerns religious actors R and peaceful outcomes P, then we see from the history of interreligious dialogue a consistent association, where R is a specific intervention that temporally precedes and plausibly explains P (Bradford Hill 1965, pp. 295–99).

Along these lines, Appleby (1999) has pointed to the recurrent role of religious actors in transitional contexts, such as South Africa, Cambodia, and Mozambique, where their actions appear as catalysts for stopping otherwise intractable violence. Toft et al. (2011) have likewise found an association between religious actors and peacebuilding in dozens of case studies. Similarly, Odak (2022, 2024) has shown that peace outcomes in postwar Bosnia–Herzegovina regularly occur where religious actors have intervened. The shared observation is that violence tends to go unabated in the absence of intervening religious actors while peace increases with their intervention. Hence, religious scholars parallel their compatriots in epidemiology by inferring that had religious actors not mediated, violence is likely to have persisted, suggesting that religious actors are explicitly or implicitly causal agents of peace.

4.2. Moral Responsibility

Another sense of causation concerns the intentions and, thus, moral responsibility attributed to religious actors post hoc (Glucklich 2026, p. 5). This is often done through speech acts theory (Austin [1955] 2018) where the illocutionary force of an actor’s utterances is said to demonstrate foreseeability, while the perlocutionary effect signifies the consequence of the utterance (e.g., Kiper et al. 2020; Wang 2025). Illocutionary force is the conventional action that a speaker intends to perform in saying something, while perlocutionary force is the contextual result of what was said (Austin [1955] 2018, pp. 101–2). Intention is typically inferred by the knowledge a speaker is thought to have had when saying something, as recognized by the targeted audience or reasonable observer (Buerger and Benesch 2025, pp. 87–91).

Though associated with studies of violence, inferring moral responsibility through speech acts bears implications for peace. After all, attributions of moral responsibility establish intent. And because intent is a kind of causality (see Dojčinović 2019), moral responsibility established through speech demonstrates causal agency.

As an illustration, Orjuela (2020) recently found that when Buddhist religious actors in Myanmar successfully used religious symbolism in their public speeches, they transformed otherwise symbolic utterances into declarative illocutions. When this happened, violence often increased. This is because religious symbolism can indicate to an audience a shared intention to bring about a transcendental reality, whose end can justify violent means—a rhetorical effect that civil society in Myanmar is beginning to counter (Orjuela 2020, p. 133).

Violent symbolism in religious speech is indeed dangerous (Omer 2015, pp. 10–13). This is especially true if the connotated transcendental reality involves an unwavering commitment to defend sacred land or a religiously fused identity, in which case religious symbolism can function as a readymade justification for violence (Kiper and Sosis 2020, 2021). With key felicity conditions, such as a history of intergroup conflict and perceived outgroup threats, the perlocutionary effect of a religious actor’s speech, if rich in symbolism, can shift an audience from a peace state to a crisis mentality (see also Oberschall 2011).

These observations align with several recent findings in linguistics. Experimental work shows that certain forms of propaganda, such as calls for revenge when spoken by authority figures, can increase ingroup empathy and decrease outgroup empathy (Kiper et al. 2023, 2020). Relatedly, religious speakers can tip a social system strained by socioeconomic or political crises toward violence when threats are framed as existential losses (Buerger and Benesch 2025; Di Cicco et al. 2025). Scholarship on counterspeech also reveals that some language constitutes dangerous speech—that is, utterances, symbols, or signs that statistically increase the likelihood of one group attacking members of another group (Buerger and Benesch 2025, p. 84).

Critically, scholars of dangerous speech have drawn on legal causation to infer that speakers are causal agents if three conditions are satisfied (for a review, see Dojčinović 2019). First, the illocutionary force of an utterance reveals intent (mens rea). Second, the perlocutionary effects demonstrate that an audience understood the utterance (objective foreseeability). Finally, a reasonable observer can infer that the directness of the utterance, if proximately linked to imminent violence, is a likely cause of the resulting actions (actus reus).6 Together these demonstrate proximate causation.

4.3. Multilevel Interactions

Another sense of causation is the explanation of interactions in empirical studies where religious actors serve as variables in multifactorial explanations of religion (Rüland et al. 2019; Svensson 2016). Researching the effects of religious actors in the science of religion (e.g., Sosis 2020; Whitehouse 2018) has become especially common in studies of cooperation (for a review, see Shaver and Kavanaugh 2020). However, similar models in the study of religion and peace are exceedingly rare.

Demonstrating this, Köbrich and Hoffman (2023), in an impactful review, found that of the thousands of studies on religion and peace, only 83 involved quantitative models. Of these, 80% were general survey-based attitudes of student populations, which differed so substantially that their findings were incommensurable. Furthermore, “religious actors” were typically modelled as whether participants simply had a religious identity or not. Thus, a surprising gap in CESR and the science of religion is that we lack both quantitative data on religious actors and a scientific theory explaining how they are causal agents (Köbrich and Hoffman 2023, p. 726).

4.4. Events and Causes

Still, as Liedhegener and Thieme (2015) argued over a decade ago, we need more than quantitative models and disparate interpretive accounts to understand the effects of religious actors. Drawing inspiration from medical science (Freedman 2011b, p. 160), I argue for synthesizing our most rigorous methods and anchoring our claims to stubborn facts that are testable in a variety of settings. To that end, the key outcome variable should be an event; for the most helpful causal explanations in the history of science have reduced complexity by focusing not on disconnected observations, but on events (O’Hara 2011, p. 106). An event is an occurrence that produces a change within a system (Sewell 2005, pp. 8–9). A system, in turn, is a bounded set of constituent elements that interact with one another in a coherent organization, which is distinguishable from the surrounding environment and achieves an outcome of interest (Meadows 2008, p. 11). Accordingly, the main dependent variable for researching religious actors is a peace event that produces observable changes within and from religious systems.

Let us define a peace event as a bounded change in state—or measurable set of values—in a religious or cultural system that is describable as a localized occurrence with duration, and stands in causal relation to antecedent conditions and subsequent states (Davidson 2002, pp. 181–204). Accordingly, let us take causal relation to mean an intervention that temporally precedes and plausibly explains a bounded change in a religious system and events pertaining to it, which is traceable to statistically robust behaviors and speech acts made by religious actors. With these in mind, let us now turn to making causal explanations of peace events in greater detail.

5. Toward Causal Explanations of Peace Events

Hume ([1748] 1993) famously described causality as: “if the first object had not been, the second never had existed” (as cited in Menzies and Beebee 2025). Skepticism aside, causality involves more than temporal sequencing, which is a common explanatory mistake in just-so storytelling. It also requires moving beyond outcome-based selection and instead situating individual cases within broader distributions of comparable events. For lack of better words, it demands “shoe leathering,” in which investigators bring together robust data as detectives or physicians would when retrospectively approximating the most likely cause of an event (Freedman 2011b, pp. 158–60). Doing so requires distinguishing type and token explanations, reconstructing the temporal order and probability of counterfactuals, and tracing content in systems of structural equivalence (Lewis 2011). This section addresses each of these steps.

5.1. Distinguishing Type and Token Explanations

Any causal explanation must answer two connected questions (Mackie 1974; Rothman et al. 2008). First, what is necessary and sufficient for a token event’s occurrence? Second, how does the token event relate to similar event types?

Let us address the second question first. A token event refers to a single-case occurrence. Here, it would include particular actors instantiated in a specific event. Such instantiation is usually determined by virtue of an actor’s presence and link to an event, as identified by co-participants and the actor’s intentions (Searle 2010, p. 18). An event type, which is built up from similar tokens, refers to an aggregation of occurrences, in which a summary of recurrent, profile, or population characteristics is instantiated in distinct yet similar cases (Searle 2010, pp. 32–33). Though simple in distinction, failing to distinguish tokens and types risks fallacious reasoning.

For example, if we draw hasty generalizations about populations from an analysis of a token event, we risk committing the atomistic fallacy; but if we draw inferences about particulars from aggregated data, we risk ecological invalidity (Illari and Russo 2014, p. 41). Likewise, if we assume that a causal pattern in aggregated data is uniform across underlying causes in token events, we risk making a Simpson’s paradox (Spenger and Weinberger 2021). In either case, our inferences would result in faulty conclusions—and potentially dangerous policies.

As Voas and Chaves (2016) warn, these risks are rather high in research on religious communities and actors due to the very problem at the heart of this essay. Specifically, it is all too easy to draw broad and overly extended inferences from a particular case, if not to infer the wrong characteristics about a particular case from general data. However, avoiding type-token fallacies is a longstanding practice in medicine, which offers guidance.

In medical research, explanations are constrained by observational studies of event types, meaning that aggregated data about populations provide limits for token-level explanations, but they do not determine them. Instead, aggregated data provide guidelines for evaluating tokens insofar as it offers statistical power or strength of association, control for confounding third variables, and knowledge of what is consistent (Lipton and Ødegaard 2005). Arguably the best source for doing so is Bradford Hill’s (1965) guidelines. Yet, unlike historical accounts that parallel Bradford Hill’s approach, a robust explanation takes Bradford Hill’s full set of guidelines into consideration.

Accordingly, when asking whether a token peace event was caused by local religious actors, we should first ask: what type of peace event was it? Answering this question alone will determine the aggregated data we use to guide our analysis. For instance, if the token event is a communal reconciliation ceremony, it would make little sense to base our analysis on any type of peacebuilding. Doing so would be like a doctor analyzing a particular patient’s cancer based on any form of illness. Rather, if the peace event is a reconciliation ceremony, then aggregated data on communal ceremonies is the type that informs our investigation of the token. From aggregated data, we can then delimit possibilities for the event, which point us to the most likely causes.

Keeping with the example of communal reconciliation ceremonies C and the role religious actors R, we can adapt Bradford Hill’s (1965; as cited in Illari and Russo 2014, p. 51) guidelines to ask the following questions about the overall event type:

• What is the strength of association between R and C?
• How consistent is R involved in C over time and across cultures?
• Is the involvement of R in C specific only to particular groups or sites?
• Does R’s involvement always precede C or does C ever occur first—and if so, what are the apparent reasons for R to follow C?
• Does the involvement of R increase or decrease the success of C?
• Does the claim “C requires R” seem plausible based on current knowledge?
• Is the relationship between R and C coherent with generally known facts about peace and reconciliation?
• Is there experimental evidence that R effects C?
• What analogous causal relationship describes R and C?

Answering these questions will not tell us whether R, in fact, caused C in the token event. But they do force us to answer the most fundamental question for any robust analysis: what is the most plausible explanation for the token?

The sum total of Bradford Hill’s (1965) inquiries will tell us whether a causal explanation of the token event is most plausible. If so, then we can move to the first question that began this section. What is necessary and sufficient for the token event?

Although philosophers tend to prefer Mackie’s (1974) INUS conditions to answer this query, most medical researchers prefer what is known as “sufficient-components.” According to Rothman et al. (2008), sufficient-components involve two things. First, a sufficient cause: the complete set of conditions that inevitably produce the event, which we can infer from Bradford Hill’s (1965) questions. Second, a component cause: one or more conditions from the sufficient set that is/are evident in the particular case. Since it is rare for any single case to show all sufficient-components, doctors tend to observe a minimal set of conditions that are nonetheless sufficient for the token event (Illari and Russo 2014, p. 31). Put another way, a disease occurs in individual patients because of secondary conditions that would not matter but for the primary conditions that are nonredundant in the single case.

To demonstrate the point, sufficient-components for lung cancer include an extensive list of conditions from aggregated data, such as genetic susceptibility, smoking, radiation exposure, and so forth. But any one of these sufficient-components can be a primary, secondary, or negligible condition in a particular patient. As such, a single patient’s cancer (token event) does not tell us much about the disease (event type) in general. However, with a sufficient-components list, we can bridge the two. Where aggregated data inform doctors about which conditions matter for a disease type, token cases provide additional medical data for aggregated analyses.

The same is true for peace events. Aggregated data can tell us much about what is likely for a particular peace event, while specific cases can add to our knowledge of typical occurrences. To determine what is necessary and sufficient for any peace event, we must approach peace like a potential treatment in medicine, which is to say as a dialectic between a token and type. We cannot simply describe the token but must identify what is typical about it, from which aggregated data can inform us of possible causes, and thereby explain the primary and secondary conditions of the token event. By working between aggregated and specific case study data, we can avoid type-token ambiguities and fallacies of over relying on aggregated or case-level data alone.

Multilayered Mosaics

Medicine and other sciences can give sound causal explanations because they have multilayered mosaics of data: commensurable sets of data that allow scientists to analyze an ever-expanding totality of datapoints of event types (Illari and Russo 2014, pp. 255–71). Having such mosaics enables scientists to make type-level claims (e.g., “smoking causes lung cancer”) and give causal explanations about a particular event (e.g., “the patient’s smoking caused his tumor”). Underlying multilayered mosaics is the epistemological insight that knowledge is never built from an unmediated data stream that is theoretically and methodologically disjointed—let alone a single case study. Rather, heterogenous layers of evidence come together to assemble a clearer picture of phenomena. Thus, causal explanations depend on causal pluralism (causality is not singular but a set of methods for making inferences); and the accumulation of data to answer Bradford Hill’s (1965) guidelines and infer Rothman et al.’s (2008) sufficient-components.

In the study of religion and peace, we are lacking multilayered mosaics. However, as Lane (2022, pp. 10–25) insightfully argues, building a cybernetic assemblage of data on religion is now possible in the age of big data and AI. My point here is not to sketch a multilayered mosaic. But to say that it is possible to aggregate prior and future studies into diverse layers of data—observational, experimental, ethnographic, statistical, and theoretical—and thus to assemble a library of occurrences that can move the discipline closer to making stronger causal explanations.

5.2. Reconstructing the Temporal Order and Likelihood of Counterfactuals

With multilayered mosaics, much of the legwork in building a case for causation is done. However, to give a sound explanation, we still need convincing counterfactuals (Lewis 1973). Critically, however, convincing (or sound) counterfactual explanations require more than conjecture. Such explanations demand the following (see Illari and Russo 2014; Lewis 1973, 2011).

The first is temporal ordering: the arrangement of accumulated datapoints in a chronological sequence. Effectively, this requires ordering data from multilayered mosaics, such that key peace events are temporally aligned alongside other sociohistorical data, including economic turns, environmental events, population health and morbidity, and so forth. Key ethnographic and psychological data for aligning events with the actions of individuals, social networks, and populations include identity fusion, ritual behaviors, sacred values, and more (for a review, see Lane 2022). Such data, when chronologically sequenced, reveals patterns and surprising punctuations that precede key events, where subsequent inquiries and data collection are imperative for explaining anomalous shifts in the timeline (Lewis 2011, pp. 25–28). Importantly, it would reveal whether religious actors involved in peace events were likely causally primary or secondary, if not a byproduct of events.

The second is probabilistic necessity: the use of Bayesian methods to estimate the occurrences of an event type and whether a token event would not have occurred but for certain factors. With aggregated data, this would simply require specifying priors by fitting observed cases of event types and analyzing token values to compute posterior likelihoods. The upshots of Bayesian methods are not only that they lend themselves to post hoc analyses but also calibrations of uncertainties in available data (Illari and Russo 2014, p. 83).

5.3. Tracing Content in Systems of Structural Equivalence

Of course, causal explanations depend most on mechanisms (Freedman 2011a, pp. 250–51). “Mechanisms are entities and activities organized such that they are productive of regular changes” (Machamer et al. 2000, p. 3). We know that we have identified a mechanism if by manipulating it we alter the recurrence of an outcome (Illari and Russo 2014, pp. 127–30). Alternatively, if we can trace the process that engenders the outcome, we can identify its mechanisms (Cartwright 2002, pp. 12–14).

Despite reluctance in religious studies to discuss mechanisms as opposed to human agency, a compelling example can show that agency is not peripheral to mechanistic explanations. My example is how cooperation emerges from self-interested agents. Controlling for enforceable contracts, the underlying mechanism is reciprocity by iteration and recognition, which researchers discerned through a mosaic of experimental and ethnographic data (for a review, see Matsushima 2020). Importantly, this behavior varies cross-culturally (Henrich et al. 2005). Yet, its role in cooperation is traceable to actors in socioeconomic systems, where biological universals engender culturally relative cognitive models for prosociality (see also Rossi et al. 2023). Albeit brief, this example offers a pathway forward. We can identify the mechanisms of religiously inspired peace events by tracing similar content through religious systems.

Biological Universals, Cultural Knowledge, and Religious Systems

Biological universals in humans are species typical traits such as bipedalism, binocular vision, and capacity for language (Brown 1991). To count as a universal, traits need not be invariant but rather statistically robust, which, if evident across environments, are likely produced by natural selection (Mayr 1982). Critically, the human brain, as a biological universal, engenders evolutionary universals: recurring cognitive mechanisms that underlie adaptive responses to problems faced during human evolution (Barrett 2015). These include psychological and social mechanisms that consistently actuate similar behaviors across cultures.

Given that peace evolves (Glowacki 2024; Kiper and Sosis 2024), it has a suite of mechanisms with clear evolutionary histories and recurrent presence in religious systems worldwide. These include costly signaling or hard-to-fake displays of commitment, belief in moralistic supernatural agents that punish moral transgressions, and collective rituals that arouse synchrony and collective effervescence (Alcorta and Sosis 2005). They also involve sacred values that provoke transcendent moral obligations and cosmological frameworks for giving meaning to events (Sosis 2020). Across different cultures, these mechanisms are activated in response to environmental changes and feedback to religious systems, as discussed below, and thus contribute to individual behavior and group cooperation and coordination (Sosis 2020, p. 144).

While these mechanisms have been researched individually, they have been less often investigated together and rarely with respect to peace events. As processes that link individual behavior to group outcomes, they provide the mechanistic paths for tracing the potential causes of religiously inspired peace events. Such tracing is feasible by measuring the said behaviors and group-level outcomes associated (see Alcorta and Sosis 2005). Mechanistic pathways are also measurable through observational methods that assess the relationship between individuals, groups, and events within a culture.

One such method in anthropology involves formal elicitation techniques and statistical analyses for measuring cultural consensus, otherwise known as the degree to which members of a group share the same knowledge (Dressler 2018, pp. 42–44). Because culture is shared knowledge, it is measurable to the extent that people in a culture agree about something, which is evident in a cultural model or shared cognitive schema that people use to understand the world around them. Remarkably, cultural models can identify a potential causal pathway—or “mental fingerprint” of an influential form of knowledge over time (Dojčinović 2012, p. 95). Stated formally, if a cultural model C is present among actors at time T₁ and T₂, such that it is invoked by actors to make decisions pertaining to an outcome and its absence correlates with a deficiency in the outcome, then a case for a causal mechanism is possible. Famously, Dressler (2018, pp. 184–87) has demonstrated causal pathways of cultural models pertaining to health outcomes.

The implication is significant for measuring peace events. Around the world religious actors function as cultural authorities whose statements and actions typically carry disproportionate influence relative to other actors (Odak 2022, 2024). Given this influence, their cultural knowledge about and for peace—as mechanisms that inform key decisions and behaviors—is measurable, especially where religious actors work to moralize restraint, mobilize alliances, or advocate nonviolence. Specifically, if the cultural models of persons and groups involved in peace events were used by religious actors at an earlier time, it would demonstrate a plausible causal pathway of knowledge via religious systems as a mechanism of peace.

At this point we are in a position to address structural equivalency. Recall that most scholarship on religion and peace focuses on conditions of conflict as the justification for comparison—but doing so risks false equivalence. However, as Glucklich (2026) argues, a systems approach to religion allows us to make justified comparisons of disparate religious phenomena. After all, investigating religion as a system captures the fact that religions are dynamic and complex, and provides the means of bridging layers and patterns of data, detecting surprising phenomena such as emergent properties, and comparing functions (Glucklich 2026, pp. 5–10).

Leading this approach, Sosis (2020) has demonstrated that religions maintain structural equivalency through his analysis of what he calls the “religious system.” This refers to the fact that religions are not static traditions of belief but rather complex adaptations that serve to support human cooperation and coordination (Sosis 2020, p. 144). Wherever we find religions, they typically maintain eight constituents: authority, meaning, moral obligation, myth, ritual, sacredness, supernatural agents, and taboo. We also find that religious actors participate in local versions of religious traditions that are responsive to the environment and complex. Religious systems thus have adaptive value because they integrate evolutionary universals and are dynamic. They also consist of affective, behavioral, cognitive, developmental, and neurological features that organize into a self-regulating feedback system (Sosis 2020, p. 142). Most importantly here, the religious system allows us to go beyond tracing cultural knowledge—what people believe—to how interconnected religious constituents engender behavior and knowledge with cultural effects.

Taken together, I suggest that it is possible to infer the religious causal mechanisms of peace events through the comparative process-tracing of evolutionary universals and transmitted knowledge shared by religious actors and participants in religious systems. Because energy enters a religious system through rituals (Sosis 2020, p. 145), ritual behavior should initiate upward effects from neurophysiological arousal to shared emotional states. In turn, these should reinforce cultural knowledge that motivates group behavior. But rituals also initiate downward effects back onto the group by strengthening their shared identity and cooperation. Such effects are not only observable but detectable in the bodies of participants, including neurophysiological changes, emotional valence, and health (see Xygalatas 2022). Because the system is open and responds to ecological conditions, system changes are also traceable through external pressures that alter the costs and incentives of religious participation. Finally, religious discourse will likely have effects akin to rituals by generating emotional content, encoding knowledge, and transmitting such knowledge within and outside the system (Alcorta and Sosis 2005; Sosis 2020).

6. Conclusions

As a theoretical sketch, this essay attempts to advance a concise framework for analyzing religiously inspired actors as causal agents, while remaining explicitly structured around the assumption of multicausality, where religious actors—if causal at all—function only as component conditions within complex causal configurations. These conditions occur alongside economic, institutional, or political factors. Rather than adopting latent assumptions in the field, I constructed my argument by challenging four widespread assumptions, which I characterized as residual problems to be resolved instead of being taken for granted. In doing so, I identified four problems that I addressed by drawing from the philosophy of causation and causal models in epidemiology and medicine. By clarifying key terms and centering the causal problem on events, I specified how to identify the conditions under which religious actors may plausibly function as causal agents, alongside and interacting with other causal factors. In this way, the essay contributes to ongoing debates in religious studies and moves beyond descriptive and correlational accounts toward analytically tractable causal explanations. By clarifying how causal mechanisms can be systematically studied across different levels of actors, groups, and systems, the essay also complements ongoing research on religious peacebuilding and interreligious dialogue. Finally, the framework speaks to emerging work in peace science by offering a methodologically grounded approach to identifying and testing causal claims about religious actors in different cultures and environments, yet comparable due to the structural equivalency of religious systems.