Abstract
A growing wave of scholarship suggests that ideology has demonstrable effects on various forms of armed conflict. But ideology remains a relative theoretical newcomer in conflict research, and scholars lack developed microfoundations for analyzing ideologies and their effects. Typically, existing research has primarily presented ideology as either an instrumental tool for conflict actors or a source of sincere political and normative commitments. But neither approach captures the diverse ways in which contemporary social science theorizes the causal connection between ideas and action, and both struggle to reconcile the apparently strong effects of ideology on conflict at the collective level with the relative rarity of ‘true believers’ at the individual level. This article addresses such problems by providing key microfoundations for conceptualizing ideologies, analyzing ideological change, and explaining ideologies’ influence over conflict behavior. I emphasize that ideology overlaps with other drivers of conflict such as strategic interests and group identities, show how ideologies can affect conflict behavior through four distinct mechanisms – commitment, adoption, conformity, and instrumentalization – and clarify the role of both conflict pressures and pre-existing ideological conditions in ideological change. These microfoundational claims integrate existing empirical findings and offer a foundation for building deeper explanations and middle-range theories of ideology’s role in armed conflict.
Does ideology matter in armed conflict? A glance at the most influential theories of conflict would suggest not, since few accord ideology much significance (Gutiérrez Sanín & Wood, 2014: 213–214). Yet in recent years, a growing wave of research has begun to link ideology with demonstrable effects on multiple forms of organized violence, including interstate conflict (Haas, 2005; Owen, 2010), civil wars (Balcells, 2017; Costalli & Ruggeri, 2015; Gutiérrez Sanín & Wood, 2014; Hafez, forthcoming; Oppenheim et al., 2015; Thaler, 2012; Walter, 2017), terrorism (Asal & Rethemeyer, 2008; Ron, 2001), state repression (Scharpf, 2018) and genocides and mass killings (Harff, 2003; Kim, 2018; Straus, 2015). Ideology’s role can certainly be overstated, since superficial ideological cleavages sometimes bear little resemblance to the underlying drivers of violence (Kalyvas, 2003, 2009). But contentions that ‘ideologies are an important basis of conflict’ (Cohrs, 2012: 53), and that ‘neglecting ideology would leave major war-related phenomena unexplained’ (Gutiérrez Sanín & Wood, 2014: 214) are increasingly common. 1
Ideology remains, however, a relative theoretical newcomer in recent conflict research. 2 By comparison with ideational phenomena like identity, norms or ethnicity, conflict scholars do not possess a well-developed theoretical literature to inform analysis of ideology, and rarely reference specialist work from other fields, such as political psychology, social movement research, and intellectual history. 3 In consequence, debates in conflict research are characterized by considerable uncertainty over the microfoundations of political ideology: exactly what ideologies are and how they can influence political outcomes (see also Kertzer, 2017). This has not stopped recent work from investigating ideological aspects of conflict, but it impedes integrated theoretical advances and renders explanatory appeals to ideology incomplete – limiting many discussions to individual empirical findings or ad hoc observations about particular cases.
Two main problems result. First, recent scholarship has struggled to clarify the underlying causal logics through which ideologies might shape conflict. Typically, theorists have primarily portrayed ideology as either an instrumental tool for conflict actors or a source of sincere political and normative commitments – what Gutiérrez Sanín & Wood (2014) term the ‘weak’ and ‘strong’ programs of ideology research, respectively. But this division is unsatisfying. There is no compelling reason to limit ideology to only one of these roles, and even taken together they fail to capture the diverse range of ways in which contemporary social science has theorized the causal connection between ideas and action. Most seriously, this ambiguity over ideology’s causal logic is at the root of an unresolved macro–micro paradox: the simultaneous presence in most conflicts of salient ideological patterns at the collective level but highly mixed underlying motives and beliefs at the individual level (Kalyvas, 2009: 592–594). In other words, while relatively few conflict participants appear to be devoted ‘ideologues’, recent research nevertheless identifies significant ideological effects on conflict. So where do these effects come from? The weak program’s suggestion that ideological appeals are instrumental tools does not resolve this paradox. One still needs to explain how ideologies can elicit instrumentally useful behavior if few people really believe in them (Fearon & Laitin, 2000: 846; Gutiérrez Sanín & Wood, 2014: 222).
Second, most recent conflict research neglects ideological change, focusing on ideology as an independent variable and attempting to identify its effects on conflict. Since actors’ ideologies are generally formed through complex context-specific histories, this stance is understandable. Yet it leaves ideology’s explanatory significance uncertain. As several scholars observe, ideologies often change during conflict under the influence of strategic and material incentives (Schubiger & Zelina, 2017: 350). But if these incentives are overriding – if groups simply select their ideological positions according to tactical convenience – then ideology is weakened as an independent explanatory variable, since the group’s strategic situation largely explains behavior (Weinstein, 2007: 21–22). If, on the other hand, path-dependencies can imbue ideologies with enduring influence independent of conflict incentives, then ideological effects may be persistent and profound. At present, however, the interaction of conflict incentives and ideological path-dependencies has received little attention.
This article addresses these two problems, contributing to recent efforts to theoretically refine ideology and establish it as a central component of the analytical toolkit for conflict scholars (Gutiérrez Sanín & Wood, 2014; Ugarriza, 2009). By drawing together recent conflict research and specialist work on ideology from other disciplines, I develop core microfoundations that explain ideology’s capacity to shape conflict behavior and clarify its explanatory relevance. This does not produce an all-encompassing ‘master-theory’ of ideology, but given the diversity of ideologies, their complex relationship with conflict, and the interdisciplinarity of research on that relationship, such a theory is probably unviable. These microfoundations do, however, resolve core puzzles that have troubled existing debates over ideology and contribute the groundwork needed to build middle-range theories and focused explanations of the difference it makes in conflict. 4
I proceed in three stages. The next section briefly addresses the conceptualization of ideology, summarizing a definitional convergence in recent scholarship before resolving some lingering ambiguities. The second section then addresses the causal processes linking ideology to conflict behavior. I propose a unified account of ideology’s microfoundations that stresses ideology’s dual causality – ideologies provide conflict actors with sincerely internalized worldviews and are constitutive of the social structures and environments in which those actors operate – and show how this account can solve key puzzles concerning ideology’s role. The final section of the article then links these microfoundations to the problem of ideological change. I affirm that conflict processes do shape ideological change but highlight how ideological path-dependencies and agency endow ideologies with independent explanatory relevance.
Conceptualizing ideology
‘Ideology’ is infamous for its superfluity of meanings. But recent work, both in conflict research and more specialist ideological analysis, has increasingly converged on definitions that cast ideology broadly, to denote the distinctive political worldviews of individuals, groups, and organizations. In a leading study, Freeden (1996: 3) defines ideologies as: ‘those systems of political thinking, loose or rigid, deliberate or unintended, through which individuals and groups construct an understanding of the political world they, or those who preoccupy their thoughts, inhabit, and then act on that understanding’. In work on armed conflict, Ugarriza & Craig (2013: 450) similarly define ideology as ‘a set of political beliefs that promotes a particular way of understanding the world and shapes relations between members of a group and outsiders, and among members themselves’, while Gutiérrez Sanín & Wood (2014: 215) present it as ‘a more or less systematic set of ideas that includes the identification of a referent group […] an enunciation of the grievances or challenges that the group confronts, the identification of objectives on behalf of that group […] and a (perhaps vaguely defined) program of action’. In contrast to narrower conceptualizations that present ideologies as highly systematic, idealistic or fanatical, ideologies in this conceptual tradition are profoundly ordinary and ubiquitous features of politics (Norval, 2000: 316). As Gutiérrez Sanín & Wood (2014: 214) argue: ‘all armed groups engaged in political violence – including ethnic separatist groups – do so on the basis of an ideology […] a set of ideas that include preferences (possibly including means toward realizing those preferences) and beliefs’. 5
Such definitional convergence is a step forward, but it does not fully clarify ideology’s conceptual relationship to other aspects of conflict – raising longstanding worries that broad definitions are so all-encompassing that they banally guarantee that ideology matters and lack substantive implications (Mullins, 1972: 498). These worries are mistaken: a broad definition does not make everything ideology, and the fact that political actors all have ideologies does not guarantee that ideologies actually matter. After all, no one would deny that all humans have personalities, yet personality-centered theories have achieved limited traction in political science. Ideologies might rarely make much difference, and many conflict dynamics prove remarkably consistent despite the diverse ideologies of conflict actors.
However, broad conceptualizations of ideology do emphasize that the common tendency to separate ideological factors from ‘pragmatic’ or ‘strategic’ concerns with security and power involves a false dichotomy. Ideologies are not simply idealistic political programs pursued with a blind disregard for strategic interests, but shape actors’ understandings of security, strategy, and power politics. Indeed, famous ideologues such as Vladimir Lenin, Mao Zedong or Osama bin Laden devoted tremendous intellectual attention to strategy – such thinking was not a ‘sacrifice’ of their ideological principles but a core part of their political worldviews. American liberals and conservatives, likewise, differ in their views of national security and disagree over how it should be pursued (Gries, 2014). This link between ideology and strategic concerns is often at the heart of ideology’s impact on conflict. Staniland (2015: 782), for example, suggests that the strategies chosen by the Indian and Pakistani governments for dealing with different militia organizations are rooted, in part, in ‘different ideological projects at the heart of Indian and Pakistani elite understandings of nationalism’. Scharpf (2018: 7) likewise emphasizes that Argentina’s 1975–81 state terror was orientated around ‘an ideological macro-program that advocated the elimination of all elements linked to the international communist conspiracy’. Presenting debates over ideology’s relevance as revolving around the importance of ‘ideological motives’ versus purportedly ‘unideological’ material, instrumental or strategic interests is thus unhelpful and empirically misleading. The real issue is whether we need to examine actors’ distinctive ideologies to explain how they understand their interests and the ways they pursue them (Straus, 2015: 11–12).
Similarly, broad conceptualizations recognize that ideologies exist in mutually constitutive relationships with other ideational phenomena such as identities, norms, and frames, rather than standing in explanatory competition with them (Cohrs, 2012: 54–56; Ugarriza, 2009: 84). It is not the case, for example, that conflicts are either about ideology or about identity (contra Huntington, 2002: 21; Kaldor, 2012: 7–8). Identities have determinate impacts on conflict in light of how they are activated and mobilized within particular ideological programs and narratives; simultaneously, ideologies often constitute conflict identities (Balcells, 2017; Graham, 2007: 241; Gutiérrez Sanín & Wood, 2014: 215; Hammack, 2008; Malešević, 2006). Political actors of all stripes – from jihadists to communists, and from liberals to ethnonationalists – act under the mutual influence of both. Indeed, important conflict identities, such as the ‘kulak’ peasants targeted by Stalinist mass violence or the Islamic State’s classification of infidels, often reflect ideological constructs that bear little resemblance to traditional self-understandings or demographic categories. It is therefore unsurprising that ‘objective’ measures of identity fractionalization often fail to predict violence (Gartzke & Gleditsch, 2006), yet subjectively understood identities do generate important fault lines in conflicts (Hammack, 2008: 238–239; Tezcür, 2016: 253–254). The same can be said of other ideational phenomena: frames may resonate or look absurd according to actors’ broader ideological understandings, and norms may need to be supported, ignored, or torn down according to actors’ ideological beliefs and preferences (Adamson, 2005; Bakke, 2014: 158–160; Snow & Benford, 1988).
A broad definition of ideology tells us nothing about the scale of aggregation at which specific ideologies are conceptualized. The common tendency is to focus on familiar ‘big isms’ such as ‘liberalism’ or ‘Islamist fundamentalism’, but these are often overaggregated, since such labels (let alone the broadest ideological categories like ‘nationalist’, ‘left’ or ‘right’) encompass numerous distinct political worldviews (Ahmad, 2016; Gries, 2014: ch.2; Morgan & Wisneski, 2017; Schubiger & Zelina, 2017). ‘Big isms’ have their place, but their familiarity is no reason to expect that they actually identify the sets of ideas which shape particular conflict behaviors, and they can obscure unconventional, hybrid or contextually specific ideologies (Wood & Thomas, 2017: 33, 44). Research on armed conflict often needs to work with more precise articulations of the ideologies of actors under study and treat the appropriate scale of aggregation as an open question (Cohrs, 2012: 56–66; Zaller, 1992: 27). Often, this involves attending to key ideological heterogeneities within collective actors, rather than assuming that the superficial ideology of a conflict faction uniformly characterizes its actual coalition of human agents (Kalyvas, 2003; Shesterinina, 2016). Indeed, successful armed groups often translate leadership ideologies into more vernacular versions for grass-roots members or public consumption, with Eck (2010) identifying this practice as critical to the recruiting successes of the Maoist insurgency in Nepal. Since ideological content changes in translation, this distinction between elite and vernacular ideology is often significant in clarifying the overlaps and differences in the motives and understandings of the leaders, rank-and-file members, and broader constituencies of conflict factions (Gutiérrez Sanín & Wood, 2014: 215; Schubiger & Zelina, 2017).
From ideology to conflict behavior
Recent research has identified numerous effects that ideologies have on conflict processes. For example, ideologies influence actors’ threat perceptions. Actors appear more likely to see each other as threatening as ideological distances between them increase (Gause, 2003; Haas, 2005), 6 and extreme ideological constructions of certain groups as irreconcilably hostile provide critical rationales for mass violence or genocide (Harff, 2003; McDoom, 2012; Straus, 2015). Ideologies also appear to shape actors’ strategic propensities (Wood & Thomas, 2017: 33). Ideologies affect the willingness to use violence in the first place: Asal et al. (2013) find, for example, that Middle Eastern political organizations with gender-inclusive ideologies are strongly inclined towards nonviolence. But ideologies also shape actors’ propensity to employ specific violent strategies like civilian targeting or gender-based violence: many Marxist groups, for example, appear ideologically averse to large-scale attacks on civilian communities (Asal & Rethemeyer, 2008; Kalyvas & Balcells, 2010: 421, 426; Oppenheim & Weintraub, 2017; Thaler, 2012). Ideologies also affect actors’ conflict capacities, often proving central to the initial mobilization of recruits and the power groups can marshal thereafter (Ahmad, 2016; Costalli & Ruggeri, 2015; Eck, 2010; Snow & Benford, 1988; Tezcür, 2016; Walter, 2017). 7
Ideology’s principal cognitive mechanisms
Such ideological effects result from various sorts of behavior: from leaders’ strategic choices to the trained combat behavior of fighters to citizens’ decisions to tacitly support a conflict faction. They all, however, raise an essential microfoundational question: how exactly are ideologies able to influence such behavior? Existing research rarely answers this question explicitly. Typically, scholars rely on one of two presumptive microfoundational approaches: ‘weak program’ accounts, where ideologies have influence because people use them as instrumental tools to, for example, mobilize supporters or solicit patron support (Gutiérrez Sanín & Wood, 2014: 217–220; Snyder, 2000; Walter, 2017); or ‘strong program’ accounts, where ideologies have influence because people sincerely believe in and are committed to them (Gutiérrez Sanín & Wood, 2014: 220–222).
Yet this divide between weak/instrumental and strong/commitment-based accounts of ideology is problematic. Weak/instrumental accounts struggle to explain how ideological appeals can be useful if no one really cares about the ideology in question (Gutiérrez Sanín & Wood, 2014: 222). Strong/commitment-based accounts can in principle explain such usefulness by focusing on the sincere ideological resonance of instrumental appeals: extremist groups like ISIS can use ideology to recruit zealous fighters, for example, because those fighters are themselves sincerely committed to the ideology (Walter, 2017: 11–12). But this explanation often implies levels of widespread ideological commitment that are at odds with the uneven and heterogenous state of sincere belief actually found among rank-and-file conflict participants. We are still left, in short, with the macro–micro paradox: apparent aggregate ideological effects on conflict often seem disproportionate to underlying levels of deep commitment.
These problems arise because neither the strong nor weak program does justice to modern social scientific understandings of the diverse causal processes through which ideas may influence action. This section therefore proposes a microfoundational account of the mechanisms linking ideology to behavior that transcends the weak-program/strong-program debate, drawing on ideas familiar in leading social, economic, and psychological theory but which have rarely been systematically applied to ideology. At its core is the claim that ideologies matter through two basic causal pathways: first, by providing conflict actors with sincerely internalized political worldviews, and second, as a structural feature of those actors’ social environments, manifested in political norms, institutions, and policy paradigms. These two pathways are not mutually exclusive and define a continuum encompassing various forms of ideological influence. But I abstract four principal cognitive mechanisms from that continuum – ‘commitment’, ‘adoption’, ‘conformity’, and ‘instrumentalization’ – arguing that, while there is some overlap between these mechanisms at the margins, they identify the major causal links between ideologies and actors’ decisionmaking. This account, summarized in Table I, subsumes but goes beyond the processes highlighted by the weak and strong programs, and avoids the explanatory gaps and ambiguities both programs face.
Internalized mechanisms
The most familiar and obvious way in which ideologies can influence behavior is through some degree of internalization. Internalized ideologies are defined by individuals’ sincere beliefs, values, schemas, and preferences, which influence reflective and unreflective cognitive processes of both perception and decisionmaking. Such internalized ideological elements therefore shape how individuals make intellectual and emotional sense of political situations and how they evaluate actions or policies as desirable, efficacious, and legitimate, influencing their likely behavior. 8 This causal power of internalized ideologies can, however, arise through two main kinds of cognitive relationship to the ideas in question.
Commitment
Most obviously, ideological components can shape decisionmaking when individuals feel some degree of direct and relatively stable commitment to the ideas involved (Gutiérrez Sanín & Wood, 2014: 220–222). This need not involve reflective, systematic or fanatical internalization – commitments could be relatively inchoate yet still powerfully affect behavior (Cohrs, 2012: 56; Staniland, 2015: 778). Indeed, while individuals with highly elaborated ideologies are relatively rare, political psychologists have found that most ordinary citizens possess committed values and beliefs that are patterned along ideological lines (Gries, 2014; Haidt, 2012; Jost, Federico & Napier, 2009; Jost & Major, 2001; Morgan & Wisneski, 2017; Zaller, 1992). Nor need ideological commitments provide primary motives for conflict behavior – they could, for example, involve beliefs about matters of fact or legitimating frames for certain courses of action that still shape decisionmaking (Jost & Major, 2001). But sincere ideological commitments carry intrinsic resonance for individuals, and consequently bear directly on processes of perception and decisionmaking. Contrary to the assumption, for example, that Soviet leaders simply used communist ideology as rhetoric for public consumption, the opening of the Soviet archives has revealed how, for most of the Cold War, ‘officials and leaders, in forums never intended for public scrutiny, took ideology very seriously’, making decisions in light of communist principles and frameworks to which they appeared sincerely committed (Gould-Davies, 1999: 92). Although ideological commitments obviously need not be irrational, they are often most visible (and explanatorily crucial) when they elicit behavior that seems hard to explain in conventional instrumentally rational terms. Islamic State’s brutality, unwillingness to cooperate with other groups, and determination to impose its strict interpretations of Islamic law over conquered territories have often proved highly counterproductive, for example. Many scholars suggest that the organization’s persistence with such tactics reflects, in part, sincere ideological commitments among its membership (Byman, 2016: 132–139, 150–153).
Adoption
Political psychologists, sociologists, and political communications theorists have long recognized, however, that individuals often sincerely accept ideological positions even though they do not feel any intrinsic commitment to the ideas involved. Individuals nevertheless adopt those ideas, typically because they have become linked to deeper beliefs, values, and concerns or because they fill gaps in their political worldview left unaddressed by their intrinsic commitments (Jost, Federico & Napier, 2009; Zaller, 1992). As Neumann (2013: 882) notes, for example, many recruits to jihadist terrorist organizations do not express longstanding religious or political commitments but join the organization out of a core underlying political grievance or particular sense of Islamic identity, on the basis of which they then adopt the whole ideology of jihadism (see also Kelman & Hamilton, 1989: 105). Adoption is frequently rooted in ‘identification’ – individuals adopt ideas simply because they are associated with identities or organizational roles they feel genuinely committed to (Kelman & Hamilton, 1989: ch. 5). This process, familiar in domestic politics, applies equally to conflict. Just as partisan citizens often sincerely accept the rightness of a certain policy simply because it is advocated by ‘their’ political party (Zaller, 1992), so might combatants sincerely adopt certain goals or standards of behavior because they are socialized to associate such behavior with their membership of an armed group (see also Checkel, 2017). Identification may also blend with self-interest. Describing the perpetrators of Nazi atrocities, for example, Allen (2002: 114–115) observes how many ‘identified their individual interest so strongly with those of “the German people” or other grand entities beyond themselves [that] they readily developed genuine attachments to the ideals of those organizations which promoted their careers’. Interpersonal relationships are also often crucial – again, research on terrorist radicalization has emphasized that many individuals adopt the ideologies of extremist movements more out of attachments to friends who have joined the movement than because of the intrinsic resonance of extremist ideas themselves (Atran, 2008: 6). 9
Crucially, though, adopted ideological notions are still internalized – sincerely accepted by individuals and potentially offering critical roadmaps to action, which often provide the key link between individuals’ private grievances and collective political behavior (Costalli & Ruggeri, 2017: 924–925). Consequently, ideological adoption may be essential in explaining which course of action individuals take: disaffected rebels who adopt a revolutionary Marxist ideology may be discouraged from abusing local civilian populations, for example, whereas a sectarian ethnonationalist ideology might lead them to willingly participate in massacres (see Hoover Green, 2016; Oppenheim & Weintraub, 2017). But individuals do not act out of the intrinsic appeal of adopted ideological ideas – so those ideas may not provide an individual’s deepest motives and the individual’s broader personal history may show little longstanding attachment to them. It is likely that individuals who adopt ideological components may therefore display distinctive behavioral properties, proving more tolerant of ideological compromises or deviations than individuals who feel real commitment, for example, or displaying weaker ideological loyalty in the face of personal frustration or conflict failures (Oppenheim et al., 2015). Such contrasting effects of ideological commitment and adoption are little studied, however, making this an important area for future research.
Structural mechanisms
Focusing only on internalization, however, significantly underestimates ideologies’ potential impact on human behavior. Individuals are influenced in their choices not just by their own sincere ideological beliefs, but by their perceptions of the ideological character of their social environment. The apparently dominant ideologies of groups, organizations, and societies – visible in political norms, policy paradigms, and institutions – exert social influence on individuals, generating structural opportunities, constraints, and incentives that encourage individuals to comply with ideologies irrespective of their own underlying views.
Where does such structural influence come from? The mutual constitution of structures and agents is widely recognized, and ideological structures are partly rooted in the internalized ideologies of powerful agents or networks of agents in a social system (Costalli & Ruggeri, 2017; Owen, 2010). The Cold War era, for example, was characterized by a powerful capitalist-communist ideological structure, which induced many groups not deeply committed to those ideologies to nevertheless frame their conflicts in such terms. The most important foundation of this ideological structure was clearly the two superpowers’ willingness to offer material and political support to those who claimed to act in the name of their ideology (Gonzalez, 1968; Gould-Davies, 1999; Kalyvas & Balcells, 2010: 420–421; Owen, 2010: ch. 6). Similarly, Salafi-Jihadism has become an attractive ideological framework for armed groups in part because it allows them to call upon the support of powerful transnational networks of jihadist activists and sympathizers (Adamson, 2005; Bakke, 2014; Hegghammer, 2010/11; Owen, 2010: ch. 7; Walter, 2017).
Yet, like most social structures, ideological structures are not reducible to the sincere convictions of powerful actors or constituencies. 10 Even in the absence of such foundations, convergent expectations about ideologies constitute ideological structures. If people expect that other people will follow an ideology, this can incentivize them to do likewise, creating a self-reinforcing dynamic that reproduces such expectations and sustains the ideological structure (Legro, 2005; Owen, 2010: 51–52; Wendt, 1999: chs 1, 3, 4). The theoretically crucial result is that ideological structures often have a force that is disproportionate to the levels of sincere internalization that underlie them (Hardin, 2002: 16; Kuran, 1989; Noelle-Neumann, 1974). For example, by the last decade of the Soviet Union, very few individuals appear to have sincerely believed in official communist ideology. Yet the official ideology continued to powerfully shape mass behavior for several years due to widespread expectations that it would do so, and due to the collective action problems involved in overturning those expectations. Non-believers remained ideologically entrapped, in other words, within existing communist norms and institutions. Only a sustained political struggle between Gorbachev and his opponents eventually triggered a cascade of defections from the official ideology and collapse of the structure, as expectations in its enduring relevance were eroded (English, 2002; Haas, 2005: ch. 6; Kuran, 1991).
As with internalization, ideological structures operate through two principal cognitive mechanisms.
Conformity
First, a group, organizational or societal ideology can shape individual behavior through conformity effects: the widely researched tendency of individuals to, often unreflectively, comply with expectations of behavior generated by peer pressure, orders from authorities, organizational routines or similar social influences (Kelman & Hamilton, 1989; Milgram, 2010; Zimbardo, 2007). As social psychologists emphasize, this tendency of individuals to simply ‘go along’ with social expectations reduces the power of sincere personal preferences, but it renders the ideologies that influence social expectations of behavior crucial (Milgram, 2010: 143–147; Zimbardo, 2007: 226–227). Many combatants may simply conform to an armed group’s operational principles rather than internalizing them, for example, but if such principles reflect ideological blueprints, then the content of the ideology will still determine the resulting behavior. While likely relevant across forms of armed conflict, social conformity has been especially emphasized in research on genocide and mass killing: Browning (1992/2001) and Chandler (2000), for example, highlight such dynamics in Nazi and Khmer Rouge atrocities, respectively. In both cases, many perpetrators displayed limited deep conviction in the ideological rationales for violence, yet the institutionalization of those rationales in perpetrating organizations generated intense conformity pressures to participate in mass murder, even in the absence of personal gain or coercion.
Instrumentalization
In addition, ideological structures are the key foundation for the second structural mechanism: the calculated instrumentalization of ideology. Scholars frequently emphasize such instrumental usage in armed conflicts – ethnonationalist leaders like Slobodan Milošević, for example, are often characterized as ideological opportunists, using ethnonationalist rhetoric to mobilize political support (Silber & Little, 1997). But incentives to instrumentally use an ideology only exist if those targeted by ideological appeals appear likely to respond in ways that the instrumentalizing actor finds beneficial. As Fearon & Laitin (2000: 846) observe, scholars who appeal to instrumental usage often fail to fully explain this ideological responsiveness, as I call it, of target audiences.
Explication of the microfoundations of ideology suggests two main explanations. First, targeted individuals might, as Gutiérrez Sanín & Wood (2014: 222) argue, have sincerely internalized the ideological elements in question. Milošević’s nationalist claims genuinely resonated with some ordinary Serbs, and with key Bosnian-Serb nationalists like Radovan Karadžić and Ratko Mladić, whose support Milošević sought. But such internal resonance is not a necessary condition for ideological responsiveness, and often seems to be absent among many of those mobilized by ideological appeals. Evidence of mass nationalist extremism in early 1990s Yugoslavia, for example, is relatively weak (Gagnon, 2004: 2–3; Malešević, 2006: ch. 7; Mueller, 2000; Oberschall, 2000: 988). Alternatively, then, audience members may themselves be responding to structural pressure – their convergent expectations that certain ideological appeals will work encourage them to respond positively to such appeals, creating a self-fulfilling prophecy (Wendt, 1999: 42). Perceiving an apparently dominant nationalist mood, for example, most Serbs may have responded to nationalist rhetoric or policies out of conformity pressures or their own instrumental incentives (see, in general, Gagnon, 2004; Oberschall, 2000: 989–998; Silber & Little, 1997). These explanations are not mutually exclusive. Instrumentalization is typically powerful precisely because the two dynamics work in tandem: some audience members respond due to sincere sympathies, others due to structural pressure, and many due to both.
Instrumentalized use of ideology need not be ‘top-down’. Political elites deploy ideological claims to mobilize supporters, but followers may also instrumentalize leadership ideologies to advance their careers or other private agendas, and clients may exploit patrons’ ideologies as a mechanism of soliciting support – a familiar feature of Cold War international politics. Kaufmann (1996: 143) suggests, for example, that the Hmong minority of Laos and Thailand, while engaged in a political struggle with the central Laotian and Thai governments, used ideological affiliations as a purely tactical means for seeking external support. The Hmong struggling against the communist Pathet Lao appealed for help from the United States by portraying themselves as anti-communists, while those fighting a US-backed Thai government across the border aligned with the Communist Party of Thailand. Instrumental considerations also constrain as well as empower actors, most notably through the need to legitimate one’s behavior to minimize costly counteractions and the loss of external and internal support (Drevon, 2017; Jost & Major, 2001; Oppenheim et al., 2015).
This all creates the potential for considerable complexity in instrumental usage of ideology, involving multilink chains of instrumentalized performances. Soviet leaders were not, for example, always enthusiastic about supplying costly support in response to instrumental left-wing appeals from revolutionaries around the globe, yet felt intense pressures to do so because they in turn were instrumentally reliant on communist ideology for national and international legitimacy (see Gonzalez, 1968). Often, multiple coexisting ideological structures impose cross-cutting instrumental costs and benefits generated by networks of numerous relevant ideological audiences. Drevon (2017), for example, analyses the ambiguous and contextually variant consequences of armed groups publicly aligning with Salafi jihadism in the contemporary global ideological environment. Jihadist claims and policies can mobilize networks of extremist fighters, yet simultaneously alienate moderates and draw in opposition from powerful states (see also Bakke, 2014; Walter, 2017: 34–35). Groups therefore typically face strategic dilemmas in navigating different instrumental costs and benefits of ideological expressions, while also having to weigh the trade-offs these create for their genuinely internalized ideological preferences.
Interactions between internalization and structure
Individuals act under multiple motives and will generally relate to available ideologies through varying mixtures of commitment, adoption, conformity, and instrumentalization. Moreover, individuals’ relationships to ideologies are dynamic. Individuals who instrumentalize or adopt certain ideological components may eventually develop sincere commitment, or become disillusioned but have placed themselves in situations where they still face strong conformity pressures or instrumental incentives (see also Checkel, 2017: 596). It wasn’t until the summer after the January 1959 Cuban revolution, for example, that Fidel Castro coupled his regime to communist ideology – in large part due to the instrumental incentives of Cold War ideological structures (Gonzalez, 1968). This choice, however, ultimately led to communism becoming deeply entwined into the political DNA of Cuban political institutions and internalized with considerable sincerity by Cuban leaders.
Internalized and structural mechanisms also interact. As suggested above, actors often face trade-offs between their internalized ideologies and ideological structures. Structural incentives constrain actors’ freedom to follow their sincere ideological preferences, but equally, sincerely internalized ideologies can shape actors’ willingness to comply with a structurally powerful ideology – those with strong private antipathy may refuse to do so (Granovetter, 1978: 1435–1438; Noelle-Neumann, 1974: 48–49). In addition, the structural dominance of ideologies can encourage internalization – individuals are often sincerely persuaded by ideas that appear widely endorsed or which they are socialized into by major institutions (Checkel, 2017; Nyhan & Reifler, 2010: 308). Many may have been drawn to Marxist-Leninist ideas in the Cold War or Islamic State’s brand of jihadism since 2011, for example, in large part due to their apparent ideological momentum (see also Byman, 2016; Kalyvas & Balcells, 2010). Ideological effects often arise, therefore, from networked interdependencies of different sorts of actors guided by different mechanisms, with the largest scale effects emerging from mutually reinforcing internalized and structural dynamics. For example, neoconservative justifications of the Iraq War – as an exercise in rapid democracy-promotion which would positively transform Middle Eastern regional security – were, in many respects, dramatic breaks from previous US policy assumptions and appear puzzling and dangerous from conventional strategic perspectives (Flibbert, 2006: 310–311; Gilpin, 2005: 5–6, 17). These justifications proved so consequential, however, because they were simultaneously longstanding commitments for key members of the Bush administration, provided a plausible roadmap of action for broader sympathetic constituencies after 9/11, were successfully institutionalized within the administration (as critics of the war were sidelined) in ways that created strong pressure for officials to support an emerging ideological consensus, and were instrumentally effective in mobilizing public support and legitimating the administration’s priorities (Flibbert, 2006).
This account of the internalized and structural mechanisms through which ideologies can influence behavior offers solutions to several key puzzles. Most importantly, it resolves the macro–micro paradox. Contrary to the assumptions of many ideology-sceptics, large numbers of fervent ‘true believers’ are not necessary for ideology to matter. More limited internalized commitments, ideological adoption, and conformity to or instrumentalization of ideological structures all allow ideologies to shape the behavior of conflict participants. Ideologies are powerful because they can bind diverse individuals into programs of collective action via these multiple interacting mechanisms and generate emergent structural effects that may be disproportionate to levels of highly committed belief. Significant ideological effects in collective action are therefore compatible with high degrees of ideological heterogeneity at the individual level, and it is to be expected that different ideological elements will play different roles for different participants in conflict (Browder, 2003; Cohrs, 2012: 56). This account also explains how ideological minorities, such as religious extremists, ethnonationalist demagogues, or revolutionary activists, can ‘capture’ the politics of a group without either mass conversion to the ideology or immense coercion. Control of propaganda or greater stridency in political discourse can make minorities appear dominant – generating ideological structures that incentivize moderate but fractured majorities to stay silent and comply with the minority ideology (Hardin, 2002; Kuran, 1989; Noelle-Neumann, 1974).
Ideological change and explanatory relevance
Ideologies are not static features of individuals, groups, organizations or societies, but change before, during, and after conflict. The consequences of such change can be profound: Hegghammer (2010/11), for example, suggests that ideological changes within transnational Islamist networks are crucial in explaining the rise of Muslim Foreign Fighters from the 1980s onwards (see also Bakke, 2014), while the collapse of the Soviet Union and the Cold War’s global capitalist-communist ideological structure initiated profound transformations in conflicts across the planet (Kalyvas & Balcells, 2010). Yet few existing studies explain why certain ideological changes occur and others do not, instead tending to take the ideologies of groups as exogenously given starting points for analysis. 11
One major exception is the recognition that actors’ ideologies change endogenously to conflict – reconfiguring under the influence of strategic/material conditions and often coevolving in dynamic competition with rival actors’ ideologies. Ron (2001), for example, shows how Sendero Luminoso’s increasing radicalism and escalation of violence in the early 1980s was provoked in part by a fear of marginalization by other left-wing groups in a period of democratization. Thaler (2012) observes how Frelimo’s and the MPLA’s initial ideological aversion to targeting civilians was eroded after independence, with the increasing recruitment of opportunistic members to repress counterinsurgencies and the waning of leadership ideological commitments under domestic and international pressures. Many Marxist movements, moreover, have abandoned purely class-based ideological platforms due to the incentives to co-opt powerful ethnic and nationalist sentiments that prove advantageous in mobilizing supporters (Graham, 2007: 242; Ugarriza, 2009: 93–96). Consistent with this perspective, scholars frequently emphasize how sudden ideological and behavioral changes, including radicalization and mobilization towards conflict, are provoked by structural, material, or political ‘shocks’ at the social or individual level (Costalli & Ruggeri, 2015: 125–126; Flibbert, 2006: 328 fn. 358; Legro, 2005).
As noted in the introduction, however, this emphasis of strategic incentives, political rivalry and shocks in explaining ideological change could diminish ideology’s apparent explanatory relevance. If actors tend to adopt whatever ideological positions are most strategically useful, one might think that ideologies are little more than an intervening variable – underlying material, political or conflict conditions, and the incentives they generate, largely determine outcomes (Fearon & Laitin, 2000: 846). Brooks & Wohlforth (2000/01), for example, recognize how Gorbachev’s ideological changes facilitated the end of the Cold War, but contend that these changes were essentially mandated by the economic and strategic unsustainability of orthodox communism. Weinstein (2007), similarly, emphasizes the difference between ideological and opportunistic armed groups in explaining contrasting behavior towards civilians, but presents a group’s ideological or opportunistic character as largely determined by material factors: groups reliant on natural resources or external sponsorship tend to attract opportunists, while groups without such resources need to recruit ideologically committed members.
Such underlying strategic and material conditions undoubtedly matter. But specialist research on ideology provides little basis for thinking that ideologies are reducible to them – in particular, because pre-existing ideological conditions also powerfully affect the direction of subsequent ideological change. ‘Collective ideas,’ as Legro (2005: 13) observes, ‘fundamentally shape their own continuity or transformation.’ Two sources of path-dependency are key.
First, new experiences, information, and ideas are perceived by conflict actors via their existing internalized ideological frameworks (Zaller, 1992: 22–28). This produces both boundedly rational path-dependencies, where ideologies provide ‘cognitive priors’ that shape the interpretation of new information (Checkel, 2017: 597; Jervis, 1976: 191–192), and non-rational path-dependencies, where individuals engage in motivated reasoning to interpret new information in ways consistent with their existing ideologies (Jost, Federico & Napier, 2009; Nyhan & Reifler, 2010). There was ample evidence in 2003, for example, that US military intervention in Iraq carried poor prospects of bringing improved regional security and democracy to the Middle East. Yet, longstanding ideological convictions of key administration officials about the efficacy of military force and Saddam Hussein’s contribution to regional insecurities led them to disregard such evidence and retain a range of optimistic rationales for war (Flibbert, 2006; Gilpin, 2005).
Second, ideological changes are (dis)incentivized according to existing ideological structures. Groups may, for example, stick with existing ideologies out of fear of membership defection, loss of public legitimacy and credibility, or the withdrawal of patron support (Drevon, 2017; Gutiérrez Sanín & Wood, 2014: 220). Even as sincere faith in orthodox communist ideology declined among Soviet elites in the 1980s, for example, ‘hardliners’ feared that abandoning the ideological struggle against global capitalism would weaken the militarized party–state apparatus, and so bitterly opposed reforms (English, 2002: 72–78, 83–87).
Political and military organizations or networks, such as armed groups, state agencies, and social movements, typically intensify both sources of path-dependency and are therefore often crucial sites of sustained ideological influence. For a start, organizations employ a wide range of socialization processes that actively reproduce existing internalized ideologies and ideological structures among their members (Checkel, 2017). Organizations and networks also typically involve sustained informational dependencies, in which broader memberships gain most of their political information from leaders or other figures of influence and may disregard contrasting information from outsiders (Hardin, 2002). Moreover, as patterns of organizational activity that reflect ideological precepts accumulate, increasingly complex amalgams of policies and behavior depend upon the ideology, raising transitional costs to changing or abandoning it. Within organizations a wide range of interests – not only political, but also careerist, bureaucratic, and identity-related – typically become bound to the maintenance of the ideology, strengthening instrumental disincentives for change (Allen, 2002).
So, while ideologies are sometimes abandoned or rapidly changed, these path-dependencies can explain their considerable durability even in the face of material and strategic incentives for change (Goldgeier & Tetlock, 2001: 72–73; Pierson, 2004: 38–41). Indeed, ideologies shape how actors perceive material and strategic incentives for change in the first place, and some ideologies are so entrenched that even the strongest apparent incentives are resisted or disregarded. Nazism or Khmer Rouge communism, for example, drove their states to military devastation and internal collapse without any sign of ideological repentance. Paradoxically, moreover, armed conflict often strengthens rather than overrides internalized and structural path dependencies, because the increased stress, information asymmetries, and uncertainty of conflict settings encourages actors to lean more heavily on their internalized ideologies and intensifies the need for support gained through established structural relationships (Hoover Green, 2016: 621; Shesterinina, 2016: 411–412).
At the same time, internalized and structural path-dependencies also help explain key moments of ideological emergence or transition. Sudden and radical changes often look puzzling from perspectives too exclusively focused on strong ideological commitments. Why, for example, would extreme ethnonationalism suddenly gain mass appeal among citizens of the Yugoslavian republics who had previously declared significant support for shared Yugoslav identity (Oberschall, 2000)? Appreciating that observable ideological manifestations are often rooted in structural dynamics constituted by convergent expectations renders this less surprising – because expectations about others’ behavior can collapse or transform more easily and rapidly than sincerely internalized beliefs and attitudes, and they do so according to powerful threshold effects (see also Granovetter, 1978; Kuran, 1989; Pierson, 2004: 82–90). Consequently, small shifts in social expectations about ideology can sometimes produce massive changes in behavior.
Internalized and structural path-dependencies also explain how ideological agency at critical junctures can have long-lasting effects. Leaders and activists play a major role in establishing the ideological blueprints for organizations or broader societies – and these can bring broader combatant behavior in line with leaders’ ideological priorities (Gutiérrez Sanín & Wood, 2014: 218–220, 222; Hoover Green, 2016; Oppenheim & Weintraub, 2017; Thaler, 2012: 549) while also creating profound and possibly unintended effects down the timeline (Pierson, 2004: 10–13). The particular ideological choices of revolutionary actors appear, for example, to powerfully shape subsequent ideological restraints on mass killing in societies they win control of (Kim, 2018; Nyseth Brehm, 2016; Straus, 2015), while a few key Marxist revolutionary thinkers have shaped the political aims and strategic doctrines of rebel movements across the globe (Kalyvas & Balcells, 2010: 420–421, 425–426). Such leaders are not unconstrained, of course: ideological innovations are conditioned by strategic and material incentives and existing ideological structures and generally need some resonance with broader constituencies to take hold (Bakke, 2014; Gutiérrez Sanín & Wood, 2014: 222; Snow & Benford, 1988). But such constraints still leave latitude for a range of specific ideological forms, and supporters can, through adoption, internalize considerable ideological content that lacks specific resonance.
Much more research is needed on this interaction of conflict-incentives, ideological path-dependencies, and agency, requiring deeper engagement with relevant specialist literatures – such as those on norms, contentious politics, social movements, and ideology itself. But ideological change need not be treated as exogenous or epiphenomenal, and may be an important factor – among others – in shaping conflict occurrence and character.
Conclusion
It is increasingly clear that the study of ideologies can make important explanatory contributions to the study of armed conflict. Real world conflict actors rely on distinctive internalized ideological worldviews that influence their priorities, values, and operating assumptions in conflict, while also existing in social environments structured by ideological cleavages, norms, and institutions. Consequently, ideologies often explain variation that eludes models purely oriented around strategic and/or economic incentives. Regimes and organizations in similar strategic positions do not target civilians uniformly, civilian populations in similar conditions of economic and social deprivation do not all mobilize for armed rebellion, and states in inferior positions of relative power do not always see this as a source of worrisome insecurity. Though never the only factor, the ideological character of political actors and their environments are often key sources of such variation.
Research has made significant advances in addressing such issues, but this article has suggested several ways to deepen theorization of the ideological aspects of armed conflict. Debates need to move beyond the true but increasingly trite observation that few conflict actors are ideological fanatics, and recognize that for ideology to be relevant, broad and deep belief in a single dominant ideology is not required. Instead, ideological effects are rooted in complex interactions between varying degrees of sincere ideological internalization and the pressures and incentives of ideological structures. This should be the theoretical starting point of future research on ideology and armed conflict. Since different elements of ideologies matter in different ways for different individuals, scholars need to develop richer and more disaggregated pictures of the elite and vernacular ideologies involved in different conflicts. Since chains of internalized and structural influence emerge from complex heterogenous networks of actors internal and external to the conflict, scholars should map such networks and engage in cross-scale analysis of how organizational, societal, and global ideological structures interact. Future research also needs to devote far greater attention to processes of ideological change, and their roots in strategic incentives, ideological agency, and relevant ideological path-dependencies.
Politics is always about more than ideology, and its centrality and explanatory power varies across cases and aspects of armed conflict. But when individuals and groups engage in organized violence, there is almost always an ideological dimension in play. Having been assumed for too long to be largely epiphenomenal, that ideological dimension ought to be a central focus of study in our ongoing efforts to understand armed conflict.
Footnotes
Acknowledgements
For invaluable feedback I thank Aisha Ahmad, Victor Asal, Kristin Bakke, Alex Bellamy, Elizabeth Frazer, Mohammed M Hafez, Thomas Homer-Dixon, Helena Ivanov, Theo McLauchlin, Rich Nielsen, Sam Plapinger, Andrea Ruggeri, Lee Seymour, Anastasia Shesterinina, Kai Thaler, Timothy Williams, Katrin Wittig, Ali Zeren, the Journal of Peace Research editors, and three anonymous reviewers.