It’s Who You Know: Social Networks, Interpersonal Connections, and Participation in Collective Violence

Although popularly perceived as a positive force, social capital may also produce socially undesirable outcomes. Drawing on Rwanda’s 1994 genocide, this article shows that participation in its violence was partly determined by the features of individuals’ social networks. Perpetrators possessed larger networks in general and more connections to other perpetrators in particular. The quality as well as quantity of connections also mattered. Strong ties generally, and kinship and neighborly ties specifically, were strong predictors of participation. In contrast, possession of countervailing ties to nonparticipants was not significant. In explaining these findings, I suggest participants’ networks fulfilled functions of information diffusion, social influence, and behavioral regulation. The findings point to the importance of social structure and suggest that relational data should complement individual attribute data in predicting participation in collective violence. 1 London School of Economics 2 Ever since Putnam lamented the significance of bowling alone for the quality of democracy in the United States, the concept of social capital has entered popular consciousness as a beneficial force to be promoted in communities. Yet this popular perspective overlooks that the social networks and connections in which this capital resides also have a dark side. Today, a small but growing body of research evidences the vicious potential as well as the virtuous value of these networks. They can promote coordination, cooperation, and cohesion to solve dilemmas of socially undesirable collective action. Participation in criminal gangs, terrorist cells, deadly riots, and ethnic violence has been linked to the interpersonal ties that bring and bind participants together. These social bonds then can be harnessed for ill as well as for good. This article extends the exploration of social capital’s dark side to a particularly heinous and uniquely human form of collective behavior: genocidal collective violence. It examines the extraordinary events observed in Rwanda in 1994 in which an estimated one in five ethnic Hutu men mobilized to commit intimate acts of violence, overwhelmingly in groups, and primarily against their ethnic Tutsi neighbors. It asks the simple question of what determined the participation of one-fifth and the non-participation of four-fifths of this segment of the population? It finds that the answer lies in part in the particular interpersonal networks in which individuals were embedded that facilitated or constrained their participation. Participants had social networks and social connections that differed to those of nonparticipants. It was the particular features of these participants’ networks and connections that facilitated their selection into the violence, while it was the characteristics of non-participants’ networks and connections that constrained their recruitment. Specifically, the article makes three central claims. First, the size of an individual’s network is an important determinant. Compared with non-participants, participants in genocidal collective violence possess more social connections within their communities in general and more connections to other participants in particular. Second, in addition to overall network characteristics, the characteristics of individual connections also matter. Not all connections are equal and some connections matter more 3 than others. In particular, kinship and neighborly ties – when compared with economic, social (friendship), political, and religious ties are the strongest predictors of participation. Related to this, the strength of a connection also matters. The stronger the connection an individual has to a participant, the more likely he is to be drawn into the violence. Third, countervailing ties to nonparticipants in an individual’s network does not lower the likelihood of participation. Individuals may have connections to both non-participants and participants, but it is the connections to other participants that prevail. I suggest several mechanisms may be at work in explaining the significance of social connections. Networks fulfill the functions of information diffusion, social influence, and behavioral regulation. Overall, the evidence suggests that social interaction is a strong predictor of differential participation. Who one knows matters for whether one participates. The article begins with the theoretical framework for the relational approach and sets out several hypotheses based on this. It then describes the research design, data, and methods before presenting the results and alternative interpretations of the data. It concludes with a discussion of the theoretical implications of these findings. Theoretical Framework The article’s theoretical point of departure is two simple but fundamental observations from sociobiology and sociology respectively. First, humans are a social species and naturally seek connections to each other (Morris 1967). Second, through social organization humans are able to accomplish more collectively than as individuals (Durkheim 1960). Social Capital The contemporary notion of social capital flows naturally from these twin axioms. The concept, however, has proved susceptible to diverse formulations. Its classical conceptualizations, originating with Coleman and Bourdieu, confounded the sources and resources of social capital. Whereas its sources lay in the connections, networks, and structures linking individuals and families, its resources 4 encompassed trust, norms, sanctions, information access, obligations, and expectations among other things (Coleman 1988). Thus Bourdieu defined social capital not only as “what ordinary language calls connections” (Bourdieu 1993) but also “the aggregate of the actual or potential resources which are linked to possession of a durable network” (Bourdieu 2008). Similarly, Coleman defined social capital not only as “some aspect of social structures” but also as “a particular kind of resource available to an actor” or the substance that “inheres in the structure of relations between actors and among actors” (Coleman 1988). The more contemporary notion, introduced by Putnam, did not resolve the conceptual confusion between the sources and resources of social capital, but did imbue it with a distinctly positive connotation and extended its application from individuals and families to whole communities and nations. For Putnam, who originally focused on the civic-mindedness created by participation in voluntary associations, social capital might “overcome the poverty and violence of South Central Los Angeles...or nurture the fledgling democracies of the former Soviet empire” (Putnam 1993). Fukuyama made similarly ambitious claims, stating social capital is “important to the efficient functioning of modern economies, and is the sine qua non of stable liberal democracy” (Fukuyama 2001). The concept also influenced thinking within the development policy community. The World Bank launched a “Social Capital Initiative” and one vice-president stated “social capital contributes significantly to sustainable development...Without social capital, society at large will collapse, and today’s world presents some very sad examples of this” (Grootaert 1998). Such thinking has since advanced and theory now more explicitly acknowledges the potentially dark side of social capital. Putnam refined the concept to distinguish between bridging capital that links heterogeneous groups and bonding capital that links individuals within homogenous groups. The latter is exclusionary of outsiders. Portes has explained how social capital can also constrain individual freedom and lead to excessive claims by group members as well as downward-leveling norms (Portes 1998). Ostrom and Ahn point out that the mafia, cartels, and criminal gangs all depend on the cohesive 5 power of social capital (Ostrom and Ahn 2009). Berman argued that Germany’s rich associational life during the Weimar Republic undermined the regime and facilitated the rise of Hitler (Berman 1997). Yet, notwithstanding the scholarly rebalancing of the concept, its positive perception persists in popular consciousness and in certain policy communities. Social Networks Although often conflated with social capital, social networks represent a distinct, expanding field of research. They merit analytical differentiation from social capital for two reasons. First, the distinction avoids the conceptual confusion between the sources and resources of social capital. Social networks are a source of social capital, but are not the actual resources that comprise social capital (Lin 1999). Second, a social network is a normatively more neutral term than social capital and can more readily be thought of having positive and negative consequences. As social organisms, almost without exception we each belong to some community and are embedded in some network structure. Social network theories acknowledge the empirical reality that our choices and actions are often interdependent (Marwell, Oliver et al. 1988; Gould 1993). We do not inhabit a theoretical universe of atomized individuals pursuing choices and actions independently of each other. What others choose to do influences what we choose to do. Atomistic approaches that emphasize individual attributes in accounting for individual actions overlook the importance of the social context within which individuals operate. Given the interdependence of many individual decisions, social networks feature in explanations of a wide variety of human phenomena, but their role is particularly prominent in theories of episodic collective behaviors. They have, for example, been identified as a determinant of participation in social movements (Snow, Zurcher et al. 1980; McAdam 1986; Passy and Giugni 2001), electoral mobilization campaigns (Huckfeldt and Sprague 1992), and political revolutions (Opp and Gern 1993). Certain scholars have argued more formally that social networks represent another solution to the collective action dilemma. Ostrom and Ahn argued that the trust generated through repeated interactions 6 between networked individuals facilitates collective action (Ostrom and Ahn 2009). Others pointed to structural characteristics of networks. Marwell et al. first showed through computer simulation that the prospects of collective action increase as the density and centralization of ties increases and the cost of communication decreases in a network (Marwell, Oliver et al. 1988). Gould demonstrated that the effects of network density and size on collective action were in fact non-linear and contingent on how centrally located volunteers are in the network (Gould 1993). Siegel argued participation in collective action is a function of network structure type small worlds, villages, opinion leaders, and hierarchical and the distribution of individual motivation levels in the network (Siegel 2009). A separate and growing area of research has recognized the sociological foundations to participation in various violent phenomena. Social networks demonstrably have a dark side. The study of terrorism has made important advances using social network analysis techniques (Perliger and Pedahzur 2011). The social ties that bring future terrorists together exist long before the individuals turn to violence (Sageman 2004), the number, strength, and type of ties shape the particular roles and influence individuals have in the group (Krebs 2002; Brams, Mutlu et al. 2006; Pedahzur and Perliger 2006), and network structures may adapt to counterterrorism strategies (Enders and Su 2007). In civil wars, voluntary recruitment in rebel movements is partly determined by an individual’s pre-existing friendship and family ties to other rebel group members (Humphreys and Weinstein 2008). The concentration of social ties in a community also mediates who is denounced and targeted at the local level in civil war violence (Kalyvas 2006). In genocide, preexisting social ties facilitate recruitment into the violence, as do group ties forged through participation (Fujii 2009). In rebellions, participants are drawn in through preexisting interpersonal ties and rebel solidarity is maintained through trust among connected participants (Petersen 2001). In communal violence, the weakness of interethnic ties and strength of intraethnic ties predispose communities to violent confrontation (Varshney 2001). In riots, participants are better connected and more involved in community life than non-participants (Scacco 2009).