Some writers have argued that “fractionalization,” understood as the division and fragmentation of society into a plurality of groups, leads to authoritarianism and violence. Others, investigating the causes of democratic regress, believe that “polarization,” understood as the division of society into two substantial groups, may be the problem. To be sure, the literature on democratization considers polarization (p) and fractionalization (f) to be alternative options. Papers listing the causes of democratic failure or regress detail a variety of causal stories, from ethnic or religious fragmentation to class conflict, to creeping authoritarianism and radical polarization, but the final shortlist seems to pin down f and p as the two most likely culprits.
Polarization accommodates a variety of determinants, like inequality or the lack of institutional mechanisms of mutual assurance between competing elites, whereas the focus on fractionalization feeds on a renewed interest on regionalism and politics. So the literature treats fractionalization and polarization as epistemic rivals in explaining democratic regress, although the results often diverge sharply. Gregory P. Casey and Ann L. Owen (2014) claim evidence that fractionalization explains institutional failure better than inequality and that “previous results indicating negative effects of inequality may be inadvertently capturing the impact of ethnic fractionalization.” M. Steven Fish and Matthew Kroenig (2006: 831) stress instead how “some writers have argued that fractionalization per se is not conducive to violence or authoritarianism; rather, polarization … may be the problem.”
The problem is that f and p have two different logics. F in fact, occurs in situations of ethnic and religious heterogeneity and is an objective condition amenable to empirical scrutiny, whereas p occurs when the two groups that fill the political spectrum push their partisan attitudes to the extreme. Confrontation may escalate into reciprocal loathing in a drift that, by all means, is the effect of deliberative attitudes. In this sense p is a deviation from a (normative) standard with possibly bad effects on either democracy or democratic progress. It is a departure from a “normal” situation that one could describe either in terms of a functioning organized opposition within representative institutions or of a normal political dialectic within democratic public spheres. This departure, though, is always the effect of people’s ability to side with either camp in political debates, to articulate and appreciate the concepts and keywords of public discourse, and, hence, to engage in distinctively normative practices.
When Alberto Alesina and colleagues (2003: 156) say that “simple measures of polarization reach their maximum,” they describe the peak of an intrinsically normative attitude that varies from rapprochement—for example, from the “normalizing rhetoric and attitudes” of centrist politics—to a confrontational stance verging on rancor, extremism, and, ultimately, paranoia. P typically occurs and reaches its maximum in the “Manichean endtimes” of the Trump era, when political divisions become social, and new, deepened social cleavages eventually turn psychological.1 This political climate has global dimensions, and moderate publics experience everywhere new formats of political communication, where increasingly partisan behavior is driven less by ideological considerations than emotional attitudes (Abramowitz and Webster 2016; Hetherington and Rudolph 2015; Iyengar et al. 2012). Polarization, in other words, is not a fact (like fractionalization) and does not generate democratic regress; rather, it is one of the epiphanies of that regress.
This article shows that the choice between f and p is not neutral or grounded on an empirical analysis of “clean” data, and the preference for f as the principal culprit in scenarios of democratic regress seems largely to depend on the structure of the argument. It is in fact the strictly empirical line of inquiry in democratization studies that tends to leave alone the notion of p.
I believe that when we shift our attention from fractionalization to polarization we are not simply moving along a continuum but rather make an epistemic leap from the realm of facts to the realm of normative problems. And we fail to register the slight gap that we cross when we go from one realm to the other.
The distinction between fact and status is captured in a line of Andrew Vincent’s book Nationalism and Particularity (2002: 4), where he points out that “the communal and tribal fragmentations of, for example, the Balkans, Northern Ireland, Indonesia and Rwanda are not preferable forms of social existence. They are rather stark realities which have to be mediated in some way.” The question we should ask, then, is whether the description of a state of affairs like ethnic fragmentation is per se a clue to the condition and strength of democracy. The continuum between a descriptive account of a state of affairs (the “stark reality” of ethnic divisions) and a normative status (like a good or bad or “preferable” form of social existence) is a fallacy that remains largely unaddressed in democratization studies. And it is this unexamined continuum that makes the data generated by the literature ultimately sterile. The literature, in fact, fails to give us a clear and coherent causal account of democratic regress, and the relative vagueness is due to biases and misconceptions at the framework level. The literature is fragmented in a variety of pieces and is typically dense with contributions in the subarea of democratization studies, a realm of concern that is typically monodisciplinary, with little or no inputs from the normative subfields of political science.
Stephan Haggard and Robert R. Kaufman (2016: 125) argue that in the field of democratization studies “generalizations supported by cross-national statistical work permit numerous anomalies.” Carsten Jensen and Svend-Erik Skaaning (2012) stress the inconclusiveness of surveys and analyses on the link between fractionalized societies and democracy. Hinnebusch (2006: 377) maintains that earlier failures led to a revision of modernization theory that “was based more on empirical studies” and less on deductions from the actual experiences of the case studies. Fish and Kroenig (2006: 829) found “no empirical evidence that social fractionalization influences democratization,” neither would there be evidence if one reversed the arrow of causation. This does not mean that fractionalization does not influence democratization but, rather, that the causal discourse linking the two terms is framed in ways that are spurious and irrelevant to establishing the point and that empirical strategies of collecting data are not per se a viable tool for dictating the terms of a (meaningful) causal narrative.2
There is one more issue that complicates the picture: that the simple observation of a nexus between increasing levels of fractionalization and democratic regress does not necessarily mean that the nexus is causal. It is a kind of post hoc ergo propter hoc fallacy—or effect—that seems to guide a large swath of the literature on democratization. Here, the conclusion is based solely on the order of facts—namely, on the mere consecutio of fragmentation and democratic drawback. It is therefore the framing of the issue that biases the outcome. Or, in other words, it seems that the moral to be drawn from theoretical narratives of low democratic progress (or incipient democratic regress) appears to be already packed in the premises of the picture.
One last dynamic that needs to be addressed is the distinction between facts and norms, the use of such terms as normative and normativity, and the idea that democracy is a normative status. My goal is to bring normative political theory to bear on the whole area of democratization and show that the ways we frame the issue determine the outcome of our epistemic endeavors.
Normative political theory may help enrich the discourse we use to articulate phenomena that would otherwise remain opaque and partial to our understanding. It has at its core a notion, normativity, concerning all those phenomena that in an astonishing variety of ways constrain human conduct. Normative constraints, though, are less physical constrictions than more or less binding ties and obligations that we make when we either undertake commitments ourselves or attribute entitlements to others. So, given that both commitments and entitlements are not facts of a physical sort, we need to specify the property that gives them their distinctive (normative) character. The response to this problem is that such constraints are normative and not physical because they occur within the “space of reasons” (see Sellars 1997).3 Reasons, to be sure, are not constrictions in a factual, physical sense nor do they have the power to bind us in a sort of juridical way. They “give us our bearings” as we face relevant facts or aspects of the world (Pettit 1993: 190). Normativity, therefore, concerns the capacity to act on reasons, or to make one’s conduct subject to them, or to adjust to or conform with either reasons or arrangements that prove to be sensitive to reasons.4 It is within this space of reasons that people, regardless of the factual divisions of society, articulate narratives of grief, contempt, and retribution that enables them to coalesce into rival blocs.
In the following I shall first engage with polarization as the normative cause of a normative problem. I will continue this line of analysis and open up the discussion on democratic regress to insights from normative theories. Then I will find a theoretical perspective on polarization and democracy by gearing the discourse on the distinction between monadic and dyadic attitudes in political science. And I shall bring to the fore a novel theoretical perspective, mimetic theory, that considers polarity the fundamental structure of politics. In the final section I will make a case for an interdisciplinary and open-ended dialogue among scholars so as to make the agenda of their discipline more open, self-aware, and inclusive.
A strand in the literature on democratization analyzes logics of polarization—macrosocial, political, and spatial—and elaborates on the impact of these logics on current trends of democratization. I believe that these logics generate distinctive patterns and that cross-disciplinary analysis is needed in order to understand the impact of these patterns on democratic planning and governance.
What I find vicious in the idea that f and p are rival causal stories is that these stories lack a defining turn, the juncture in which the fact of f impacts the set of normative statuses that define “democracy.” Let us consider a similar story: according to Alan Booth (1985: 153), “residing several stories above the ground” seems to cause marital stress (and increases the risk for divorce). Here the “epistemic leap” is between a fact (“housing type and household crowding”) and the quality (and resilience) of a normative status (e.g., marriage). Stories of, let’s say, “matrimonial regress” are normative tales in which factual premises are imbued with momentous normative qualities. Here, what may hold married couples together is not a number of unspecified factual elements but rather normative attitudes—for example, rules of thumb adopted after possibly careful deliberation. But the causal explanation Booth offers fails to expand on how facts affect attitudes.
So what I am not arguing is that p, more than f (ethnic, linguistic, and religious), curbs the chances of democracy’s success. Instead, I am saying that facts do not, in principle, transform themselves into problems, and highly fractionalized societies can still be democratic if framed within and constrained into viable democratic institutions. Facts are neutral unless they are so stringent and inescapable to force the actors to take specific attitudes.
There are circumstances in which primary ethnic designations are abandoned to embrace a more comprehensive collective identification. Here a collective attitude helps overcome factual divisions, as in the decades before the eventual breakup of Yugoslavia, where people had gradually abandoned age-old ethnic designations and were more prone to declare themselves as Yugoslavs (Sekulic et al. 1994: 85). And it happened, again, with varying intensity, with the idea of Europe (and the designation of European), when, for instance, nations from the former Yugoslavia decided “to detach themselves from the region and recast themselves as lost sheep of the West, returning to the European flock” (Steil and Woodward 1999: 95). It is a statement, a normative one, that people are occasionally willing to make.
Normative attitudes, though, are volatile and transient, and may fail to cope with issues that remain normatively intractable: facts, in other words, may backfire, leading—if we stick to the same line of examples—to “Bosnia’s theoretically and normatively insoluble problem—different ethno-national groups harboring incompatible notions of self-determination” (Bose 2010: 137).
To be sure, fractionalized societies are possibly more prone to democratic failure than homogeneous ones, but the principal distinction here is not the one between fractionalization and homogeneity. This distinction distracts attention from the underlying issue of polarization and its impact on the chances of democratic progress. And scholars may have credited fragmentation and fractionalization with (normative) effects brought about by a normative problem, such as strong polarization. As M Fish and Kroenig have pointed out, “polarization complicates democratization, even if simple fractionalization does not” (2006: 832). In other words, a normative problem may complicate a set of normative statuses, even if a simple fact or sequence of facts does not.
Democratic progress, when not complicated by a degree of ethnic fragmentation, is generally simple. Fractionalized societies, though, may allow for a degree of democratic success so long as fragmentation does not set off dynamics of polar radicalization. It is in fact polarization, more than a friction among different ethnic realities, that complicates democratization. Some scholars have considered situations in which ethnic fragmentation is upgraded to a normative problem: Sabine Carey (2002: 69) has suggested that ethnic polarization in the party system could be a key factor in explaining democratic failure. However, ethnic fragmentation does not easily escalate into a normative fact, for there is no continuum linking the fact with the possibly forthcoming status. And upgrading is the only term we could possibly use to describe this leap from fact to status.
There could be cases in which a degree of ethnic fragmentation may lead to dynamics of polarization. Fragmentation can become a serviceable rationale, and political leaders in multiethnic contexts have often played the ethnic card so as to exploit new friend-foe borderlines. What we observed in the ex-Yugoslavia in the early phases of its disintegration was a strong opinion alignment “along multiple lines of potential conflict.” And opinion radicalization may lead to situations in which structured cleavages compound themselves and create a cumulative (ethnic) disaster (Baldassarri and Gelman 2008: 409).5
But here, as I said before, we are not moving along a continuum, for we are making an epistemic leap from facts to statuses. The question we should ask then is whether the description of a state of affairs like ethnic fragmentation is per se a clue to the conditions of democracy. When we tell stories of democratic regress that begin with polarization we remain consistent throughout because the regression of a normative status can only be explained in terms of normative risk. And polarization is a high-risk dynamic inherent with democratic arrangements.
An escape from the aporiae of a vicious epistemic continuum may consist in adopting consistency throughout, thereby reducing the risk of epistemic gaps. Consider, for instance, the efforts of the empirically oriented literature to streamline and smooth out the noncontinuum facts-statuses into the epistemic continuum of facts-facts: Jensen and Skaaning have replaced the spurious and (epistemically) high-risk notion of “democratization” with the quasi-factual concept of “modernization,” arguing that fractionalization does not have a direct impact on democracy but, rather, that highly fractionalized societies are hardly likely to modernize their political systems. Whereas the positive effect of modernization increases with decreasing levels of ethnic fractionalization (Jensen and Skaaning 2012). On a similar vein, the seminally influential Alesina and colleagues examine “the effects of ethnic fragmentation” on the quality of institutions and policy (2003: 157). The authors do not fall in the epistemic trap I have envisioned above, as the “quality of government” is determined by measuring a number of factual “policy indicators,” which is different from democratic statuses. But ethnic fragmentation, according to the authors, is a competitor of polarization in their search for clues to democratic regress. And pushing the regress to the extreme, the authors claim that “whether societal conflict is the result of fractionalization or polarization is largely an unresolved question in theory, calling for empirical work.” Their indication is simple: empirical work is all that is needed to resolve a theoretical question. And accordingly, the authors plead for “economists to do more case histories of development, economic policy, and government quality in ethnically diverse places” (Alesina et al. 2003: 179). This is at odds with what Haggard and Kaufman (2016: 128) recommended: that empirical research on an event so relatively rare as democratization is plagued by so many anomalies that “complementary qualitative analysis” is needed in order to redress the obvious biases.
What still remains to be determined is whether in reducing the continuum to facts-facts one ends up losing normative insight and critique what one makes up for in epistemic consistency. In the following I shall address this seemingly tedious puzzle by taking issue with the apparent (and often self-aware) inconclusiveness and sterility of exclusively empirical accounts. I will argue then, in the third section, that this also depends on yet another bias in the ways and methods we use to make sense of political realities: the monadic bias.
Mistaking the Effect
The interest in polarization dynamics in US politics shows a pattern of its own.6 Analyses tend to unfold across disciplines, and scholars writing on polarization seem to agree on the fact that they are dealing with a problem, namely with the effect, normally checked by sound political institutions, of dynamics inherent to group behavior. Empirical analysis is used to describe situations in which polarization modifies the structure of the system. However, polarization remains a problem, with origins in the deliberative attitudes of both individuals and groups.
The framing of p as a genuinely normative problem can be found in Cass Sunstein’s analysis of like-mindedness in American politics, stating, “Group polarization is the typical pattern within deliberating groups” (2009: 3, italics mine). Sunstein extends the pattern to a number of situations across time and regions: “deliberative enclaves of like-minded people” are the ultimate realities in dynamics of group polarization, where the emphasis is placed (consistently) on the ability of groups, and people within them, to make reasoned evaluations and then side with like-minded individuals and groups. But polarization occurs, according to Sunstein, when groups engage with controversial issues—that is, when they deliberate on issues of ought or ought not.7
Sunstein explains failures in deliberation as drifts from normative truths that reasonable people would normally recognize as such. Or in terms of the radicalization of epistemic and normative mistakes. It all happens in people’s heads (and not, to be sure, in the world, where facts take place), not so much in a distant political reality but in their perception and interpretation of this reality.
Sunstein’s argument reflects the attitude of placing the topic of polarization at the crossroads of disciplines and discourses and to acknowledge the normative nature of the issues. Facts are relevant in his narrative so long as they are registered within people’s attitudes and discourse. In the realm of democratic risk, however, facts do not play a direct causal role.
Attempts to factualize a status and restore a continuum in place of an epistemic gap are instrumental to fabricating pleasing narratives of democratic success. According to Timothy Frye (2010: 5), the pace and progress of democratization is slower when executives have a more “difficult time in conducting economic and institutional reform,” and this would normally occur when the core supporters of the executive in power are separated by a deep (economic and ideological) ridge from the supporters of a potential incumbent. The government will be pursuing partisan policies aimed at delivering benefits to their core supporters, and these benefits will alienate people who do not benefit from those policies.
Here we have an account of democratic failure centered on polarization, even though polarization is described as a fact, not as a problem—namely, as “policy distance” between an incumbent and a prospective executive. It is interesting that Alesina and colleagues acknowledge the basic fact that “the degree of polarization should increase as the distance between groups increases,” but the authors also seem to grant that the story of democratic regress that begins with polarization and ends, ultimately, with social conflict may contain some spurious element that dent its epistemic value. They make in fact the curious admission that “the concept of distance is hard to capture with simple measures” (Alesina et al. 2003: 178).
All efforts to give an account of democratization based on the spurious continuum of facts-facts end up removing statuses from the picture and envision a world that fails to acknowledge normative qualities. Here, in fact, the relevant data are supplied by means of empirical scrutiny, not via normative considerations. Frye’s (2010) work on polarization in Eastern Europe belongs to a mainstream methodological orientation of the literature. The bulk of this literature seems to resort consistently to the notion of fragmentation to explain dynamics and phenomena that seem to be the effect of polarization. But this framing of issues and analyses does in fact bias the outcome and generates spurious results.
Studies that challenged the causal narrative that proceeds from fractionalization and ends with democratic regress argue either that democracy could work in fractionalized contexts or, alternatively, offer evidence for other culprits, such as corruption, detachment from political life, and the decline of democracy’s operative structures like political parties.
Polarization is a fact of European politics. Nikolas Busse (2015), in an article on the FAZ entitled “Die Mitte Verliert” (“The Center Loses”), describes an emerging phenomenon that complicates the pleasing causal narrative of democratic progress in full-fledged democracies.
In democracies changes in power are obvious [in Demokratien sind Machtwechsel natürlich]. But all the more often, new tribunes of the people tear apart their own societies and the European people. And the degradation of political language that goes with the ascent of those tribunes is precisely an indication that the moderate center, which is intent on balance and compromise, is losing ground. Maybe the era of centrist politics that shaped the Western world after the post-Cold War is coming to an end. And also in the United States a persistent polarization of political parties has become a fact of everyday life.(Busse 2015; author’s own translation)
Polarization is therefore a radicalization of political discourse, not the measure of a distance. However, the idea that polarization is not a mere fact of politics, amenable to data collection and data analysis, is not obvious. Considering polarization to be a fact or, alternatively, a complication of normative statuses has consequences on how we decide to repair the problem. One could generalize a bit for the sake of argument and say that empiricists favor incentives, believing that one could escape typical “polarization traps” by means of institutional tricks like, in Frye’s (2010) words, “benefits.” Conversely, normativists explain how one could escape these traps by focusing on people’s ability to think, judge, and form a degree of consensus on issues that have the power to polarize society.
Think of Jürgen Habermas’s hypernormative attitude: he draws a distinction between the practice of “reaching understanding” from that of “bargaining”: “in the former case, appeal is made to the consideration of norms and values; in the latter, to that of interest positions.” To Habermas, under the conditions of what he calls “norm-governed action” (e.g., the kind of action that engages the parties to reach an understanding), parties in a dispute “can reach a settlement by ascertaining, on the basis of an existing value-consensus, what one ought to do in the disputed case.” Bargaining, instead, applies in conditions of “interest-governed action,” when the parties “reach a settlement by achieving a balance among their interests—normally in the form of compensations for disadvantage—on the basis of their factual power position and the corresponding threat potentials” (Habermas 1998: 140).
We have seen how Timothy Frye (2010) shifted the discourse on polar dynamics from its elective milieus, the sites of deliberative speech where actors operate in consideration, to follow Habermas, “of norms and values,” to the realm of policy analysis. In this way Frye reaches undoubtedly a higher level of objectivity in his eminently empirical analysis, and the story he tells is not plagued by epistemic gaps or fallacies: it is, in fact, consistent throughout. However, his definition of polarization is highly elusive and does not account for dynamics of rapprochement elicited by means of discursive practices of “reaching understanding.”
It is clear that part of our problem, the construction and misconstruction of false continuities, depends on the available discourses in democratization theory. The construal of a continuum, in other words, depends on the ways the available logics frame the issue. But current trends in the literature on democratic regress make the logics available, and the existing biases and misconstructions will endure until we decide to challenge those trends by means of a rival theory that would enable us to frame issues of liability in democratic regress by disclosing the false continuities that our treasured orthodoxies contributed in creating. And it would be great (i.e., epistemically profitable) to find a theory that shows that democracy is a polarity standard whose structure is fundamentally dyadic and that this polarity may degenerate into dynamics of harsh polarization. And that if this happens, the whole democratic edifice is in jeopardy. If a theory like this exists, we need to look for it.
Redressing the Bias: Dyadic vs. Monadic
Many issues related with the maintenance and expansion of democratic institutions have been investigated by distinguishing between systemic and monadic levels of address. A systemic approach supports the realist claim that the structure of the international system, rather than state-type should be central to our understanding of state behavior. Accordingly, constructivist theories claim that the rationale for state behavior (notably, the process of, and commitment to, democratization) is always systemic and not contingent on state-based features. Monadic approaches, based on the behavior and agency of single actors (e.g., the states) seek to isolate loci of agency from the general context. A number of scholars, however, have analyzed the spread of democracy as a result of influence and emulation (Gunitsky 2014). As it has been argued,
much of the discussion of comparative politics is based on the analysis of individual countries, or components of countries. This approach remains valuable and important. That said, it is increasingly evident that those individual countries are functioning in a globalized environment and it is difficult if not impossible to understand one system in isolation. To some extent the shifts in national patterns are mimetic, with one system copying patterns in another that appear effective and efficient.(Caramani 2008: 55)
There is a new constellation of studies, lacking a univocal disciplinary basis, that seeks to define the mechanism of mimetic interaction in reciprocal terms, where imitation has no univocal direction of fit, where the actors are not trapped in a rigid dialectic of original and copy. Mimetic theory may provide valuable insights into dyadic patterns and dynamics, where polarization is (counterintuitively) the end result of mimetic attitudes.8
Sadly, political scientists are keen on attributing original intentions to political actors so that whatever is not original is said to be derivative and clichéd. As it has been pointed out, “in our post-romantic era, we still assign disproportionate value to originals” (Fuchs 2001: 164), so that we have all succumbed to the myth of the ontological primacy of originals over facsimiles (Fuchs 2001: 165). Failure to acknowledge the mimetic mirroring occurring in the reciprocal dealing among actors has the effect of biasing our perceptions and obscuring the dyadic dynamics that compound political processes.
Mimetic theory has tried to combine the efforts of different disciplines, and scholars engaging with this new constellation of studies have argued that systemic patterns in global politics are inherently dyadic.9 The overcoming of the monadic perspective is inscribed in the future of a discipline that acknowledges the fundamentally dyadic structure of reality and considers “jealousy of state,” imitation, and rivalry critical for understanding social psychology.10
Mimetic theory, to put it bluntly, stresses the symmetry of the actors over and beyond their attitude to imitate each other; it provides a factual basis to a type of discourse that seeks to explain how dynamics of polarization are fundamental and original and, therefore, precede the identitary dynamics that so many writers detect in situations of (high-risk) ethnic fractionalization. Mimetic theory constitutes a functional support to my analysis for the simple reason that it reveals with great strength that polar conflict is fundamentally antecedent to identitary fragmentation.
The formation of polar (mimetic) opposites precedes the emergence of identity cleavages in the sense that people “have a feeling” of their diversity and, thereby, form elementary notions about their identity after having experienced hostility toward a group that often shares the same language and morphology.11
Lene Hansen and Ole Wæver (2002) have challenged the view that conflicts in which identity is used to rationalize hostility are actually caused by identity issues.
How problematic this approach is becomes immediately visible in the case of the conflicts in the former Yugoslavia. The underlying assumption of much identity/culture based thinking about conflict is that the propensity for conflict correlates with differences and cultural distance. However, Croats, Serbs and Bosnians are quite close for most purposes, including language. And this closeness is actually part of the reason for the conflict.(Hansen and Wæver 2002: 43)
In other words, conflictive mimesis between individuals and groups is not the upshot of prior identity cleavages; rather, the rivals become identity aware as a consequence of mimetic impulses that cause them to conflict.
Conflicts have a mimetic core so long as people engage in dynamics of reciprocal imitation that culminate in a general convergence on the same objects. In other words, they become rivals because, having socialized in the same mimetic milieus, they end up desiring the same things.
In cases where a “mimetic strain” is at work, recourse to “social identity” is of little use to explain the conflict for the simple reason that the radicalization of identity is, again, the upshot of the mimetic crisis and not what sets it off. Ultimately, explanations based on mimetic theory and explanations based on identity radicalization repel each other. In a number of conflicts in which animosity is explained in terms of territorialization of identities, a mimetic dynamic is arguably at work, pushing the rivals against each other and forcing people who socialized in an environment with weak identity ties and possessing unclear identity marks (children of interethnic marriages and, in Bosnia, nonethnically Bosnian Muslims) to take on a clearly perceptible identity and thereby engage in the ongoing polarization. Claims of exclusive possession of a given portion of land often take place in environments in which identity issues become politically sensitive after individual and group rivalries set in. And both rivals and third parties use identity to rationalize the ensuing conflict, thereby contributing to conceal the mimetic activity that polarizes the scene.
Identity was deemed to be the cause of the myriad civil wars—fought at the microlevel in towns and villages with an ethnically mixed population—that punctuated the blurred map of what was left, in the early 1990s, of the former Yugoslavia.12 But the division and fragmentation of society into a plurality of groups per se did not lead to authoritarianism and violence; rather, the escalating divisions within each of these micro-contexts into two substantial groups brought about institutional failure and, ultimately, the Yugoslav war. And to place the fact of ethnic fragmentation and the subsequent polarization on a continuum would be a mistake.
In order to appreciate the elements of innovation implied in this perspective we need to clear the ground of several misconceptions and expose the biases and inconsistencies that a number of orthodoxies cause. We need to consider alternatives and test established routines, challenging monadic approaches based on the idea of a core locus of control, a quintessential nucleus of agency that directs the choices of single actors. And we will see the advantages of an analysis that takes into account dyadic directions and patterns in which each actor is examined in its constitutive relationship with its dyadic “other.”
Dyads and dyadic patterns are not new concepts in political science. Scholars have long been aware of the epistemic advantages of sampling pairs of items in order to measure the risk of, for one thing, interstate conflict. The problem is how to sample relevant (nonrandom) dyads in order to produce valid and nonbiased inference (Lemke and Reed 2001). Referring to dyads is relevant so long as dyads are the effect of polarization dynamics. In other words, dyadic patterns are the effect of repulsive dynamics that, in the long run, disaggregate international systems.
This attempt at reframing the logics of international politics along dyadic patterns is instrumental to solving the biases that undermine current efforts in democratic planning. My thesis is that the principal obstacle in the way to global democracy is not elite behavior, political culture, or state type, nor is it a systemic failure on the supply side—namely, the inability of such actors as the United States or the European Union to handle the third wave of democratization and bring it to successful completion. Rather, it is a more diffused imbalance that from the international system descends upon the single states and distorts the logics and process of democratization. And it is polarization the critical element in situations of democratic risk, when mimetic antagonists become polar opposites.
A Bigger Picture
There must be a “bigger picture” when it comes to dynamics of democratization, and this big picture must have empirical and normative features. But a combination of empirical and normative seems to be a little unpopular in these days, and scholars’ agenda is limited only to facts amenable to empirical scrutiny.
I believe that the puzzle addressed in this article reveals the slow erosion of a theoretical realm and the general attempt at renegotiating the borderline divide between subdisciplines within the general domain of political science.
Polarization, admittedly, is inherently elusive and generally hard to detect, but its contours become visible if one adopts normative lenses instead of empirical means of analysis. And democratization, in order to be viable, needs to be grounded on epistemically sound theoretical frameworks where normative lines of appraisal are integrated with empirical scrutiny. So, ultimately, questions of what “endangers democracy” or what are democracy’s “antagonists” need to be framed in a fundamentally normative fashion.
Normative statuses come to the fore when we understand that polarization is a glitch in the deliberative attitudes of a group, such as when the group falls apart and splits up in two blocks divided by what looks like an unbridgeable breach.13 It is a serious challenge for democracy, and the search for a policy solution along a causal continuum that goes backward from democratic failure all the way to the fact of fractionalization is wrong.
In order to redress the bias in our models of address, we need to shift our approach to studying democratization toward normative types of address. Facts of politics can certainly be clustered in normative grids and can generate variations on a normative continuum to illustrate the evolution of a normative status from “high-risk” or “endangered” democratic polities to full-fledged, typically “normal” democratic rule.
So, as I perceive it, the challenge is not so much to accomplish something, to add new knowledge and understanding to a solid and substantial basis of knowledge already acquired, but rather to determine how to effect a departure from that basis of knowledge and to undo the effects of older stipulations in the theoretical discourse on democratization.
Interestingly, a powerful analysis of the big picture of American democracy is put forth by legal theorists like Sunstein because it is the very subject itself to advise normative strategies of address. Scholars, and not only in America, seem to be aware of what they call “polarization traps,” but it is difficult to frame a meaningful discourse on the ways and strategies of escape from these gaps that is not normative. Sultan Tepe (2013) argues that it is ultimately individuals’ ability to engage in politics beyond the confines of partisan attitudes that presents an escape from these traps. But we observed that escape routes could be geared toward either the attempt of reaching (mutual) understanding (via normative means) or the fact of bargaining, where the parties are given incentives and benefits. A big picture with chances of epistemic success needs to integrate insights from these two realms, trying to avoid explanations grounded on a spurious continuum of facts-statuses.
So it is not that we need to replace the classical, formal analysis based on empirical inference with normative reasoning but rather that there are some situations in which polarization is inconsistent with empirical inference and we need to borrow insights, strategies, and epistemic mores from some cognate disciplines in order to integrate the ways and mores that structure empirical research.
Democracy and Interdisciplinarity
Group psychology, for one thing, offers interesting inputs into the effects of polarization, albeit on a different scale. But the case of group psychology shows how scholars can engage in an expedient division of labor with valuable research outcomes. Normative considerations can live on a par with empirical ones, although the case of polarization indicates clearly that if every problem has in it the seeds of its own solution, polarization remains intractable if one sticks to purely empirical means.
The primary aim of this article, though, was not to determine the real antagonist, the ultimate culprit; my concern was instead to pin down the stakes involved in finding the culprit. I believe that the stakes are not “pure,” meaning that the question of which is the real antagonist is not a purely scholarly one; rather, it has to do with a tension between subfields within the same discipline and the attempt to remove normative insights from political analysis. So the stakes happen to be located elsewhere; they are not theoretical but rather practical, for the problem extends way beyond academic scholarship and discourse. And speaking of stakes, sanctioned by a conventional usage in academia, the idea that f is a liability in stories of democratic progress and that a fact could be, in principle, a predictor of a status, may lead to a general and deeply biased understanding of the causes of democratic standstill. So the stakes are ultimately located outside, in public discourse, in the innumerable ways we look at the world and predict trends and developments on the basis of default causalities. And in order to reach a better vantage point and a clearer view of the stakes involved, we need to undo the layers of misconstruction that bias the issue of the real causes of democratic regress and failure.
Here maybe a “battle between disciplines for recognition” (Graham and Lepenies 1983: xvi) is being fought and the whole field of democratic theory is possibly complicated by a close-knit intersection of disciplines and discourses such as law, political science, philosophy, sociology, group psychology, and so on. To be sure, formal and empirical analyses show themselves to be incomplete as soon as we support a definition of polarization that psychologizes the context in which this phenomenon occurs. And whereas fragmentation was the elective culprit for democratic regress in such journals as, for example, the Journal of Economic Growth, the Journal of Comparative Economics, and the Journal of Development Economics, we see polarization at the center of papers featured in the area of communication studies.
Mimetic theory surfaces spasmodically in a number of disciplines and carries the biases and connotations of its practitioners, but its discourse remains open and inclusive, building its scope of action on an openness to renegotiating the boundaries between anthropology and the social sciences. To René Girard “we have entered an era in which anthropology will become a more suitable tool than political science” (2010: 2), and it may be true that current challenges have pushed political science past the limits of its canonical discourse. This is more than a mere expansion of competences, as it involves a redefinition of disciplinary objects. For the status and meaning of terms such as democracy, institutions, and borders change when exposed to cross-disciplinary scrutiny. And by the same token such terms as ritual, the sacred, and sacrifice, once forced to obsolescence by our progressive conceptions, gain new consideration and status (see Brighi and Cerella 2016; Juergensmeyer 2003; Wydra 2015).
But still, the anthropological takeover of political science is not yet academic reality. Students of political science begin their apprenticeship in the discipline by learning the distinction between empirical and normative and end up in a field such as democratization studies, in which normative expertise is declared of little use even though the concepts they use, like democracy and polarization, are packed with normative meaning.
The obvious way out of these dilemmas would be to increase perspective and reach a new vantage point from which biases and epistemic mistakes could be dispassionately assessed. But perspective can be gleaned only via interdisciplinary dialogue, by engaging with discourses that acknowledge the gap between facts and norms. But this is not a way to reclaim the “title” of normative political philosophy to engage (or reengage) in serious debates, after the fasts of the 1970s and 1980s, but rather to see that some important elements could be saved from a distant legacy, such as the distinction between facts and norms drawn by Jürgen Habermas (Habermas 1998). And our search could reach out to other shores, like philosophy, notably the philosophy of language (see for one Brandom 1994). Or even sociology of language and the collective work done in the field of conversation analysis (CA), where scholars have tried to combine normative attitudes with empirical approaches—formal coding, for example, belongs to “the view of the world as an aggregate of facts, while CA is involved in the view of the world as normative connections. These worldviews are entirely different and therefore nonrival, and they may be able to provide each other with well-grounded resources” (Nishizaka 2015: 26).
Another promising field, where methodological considerations about the status of legal reasoning and demonstration are duly debated among scholars of different methodological persuasion, is empirical legal analysis (Epstein and King 2002). And we may perhaps need to start from here, from a new alliance among disciplines and scholars with the objective of reintegrating normative considerations in the discussion on the chances for democracy to expand and develop into the kind of ideal universe that normative theorists since Plato have tried to envision. Curiously, we are inclined to talk only about facts and leave alone problems. But a purely empirical vision brings about not only a biased perspective on our subject but also a spurious and self-serving selection of the items in the scholarly agenda.
Other disciplines have been facing this very same puzzle: empirical legal studies value statistically significant correlations and data. However, “ascribing meaning to those correlations is a much more difficult task” (Edwards and Livermore 2009: 1907).
The notion of “space of reasons” was used by post-Sellarsian philosophers such as John McDowell and Robert Brandom. McDowell has expanded this space into a “space of justifications or warrants” (McDowell 1994: 7).
See Joseph Raz’s definition of normativity: “the normativity of all that is normative consists in the way it is, or provides, or is otherwise related to reasons” (Raz 1999: 67).
In “deeply divided societies” multiple lines of potential conflict can possibly neutralize conflict. Israeli and Turkish politics offer cases in which religious partisans’ political positions differ from their party leadership, and issue-based coalitions may form across different ideological groups. And, the argument continues, “given the tendency of the political elite to exacerbate divisions for political expediency, it is ultimately the ability of individuals to engage in politics beyond the confines of party politics that presents an escape from these polarization traps” (Tepe 2013: 831).
Interestingly, 2009 bore witness to an escalating interest in the United States for the issue of polarization in American politics. Marc Hetherington and Jonathan Weiler (2009), Sunstein (2009), and Bill Bishop (2009) were all published in the same year (Baldassarri and Gelman  was published a year before).
This includes questions like “should States allow same-sex couples to enter into civil unions? Should employers engage in ‘affirmative action’? … Should the United States sign an international treaty to combat global warming?” (Sunstein 2009: 5).
Mimetic theory is generally associated with the work of one author, the French anthropologist René Girard. However, Girard himself stated on several occasions that his work drew on a number of “revelatory” authors belonging to a marginal tradition that escapes classification who revealed the nature of social bonds and conflicts to be mimetic rather than egocentric. The mimetic perspective applies an anthropological insight (a biological predisposition to imitation that humans share with the higher primates) and aims to deconstruct the truisms on which canonical analyses are based.
Hobbes, in Chapter XIII of Leviathan, argued that “in all times, Kings, and Persons of Soveraigne authority, because of their Independency, are in continuall jealousies … having their weapons pointing, and their eyes fixed on one another; … which is a posture of war.” Hobbes here established the same nexus between reciprocal imitation and rivalry that constitutes the central tenet of mimetic theory. A powerful “Jealousy of State,” together with a more intrinsic impulse to provide for the welfare of its subjects, was the driving force in the continual warfare that characterized international politics in the early modern age.
“In general, it does not matter for our purposes whether ethnic differences reflect physical attributes of groups (skin color, facial features) or long-lasting social conventions (language, marriage within the group, cultural norms) or simple social definition (self-identification, identification by outsiders). When people persistently identify with a particular group, they form potential interest groups that can be manipulated by political leaders, who often choose to mobilize some coalition of ethnic groups (“us”) to the exclusion of others (“them”)” (Alesina et al. 2003: 176).
In Bosnia in the early 1990s towns and villages with a very high concentration (up to 40 percent) of ethnically and confessionally mixed marriages bore witness to a shocking escalation of violence between ethnic-based factions, and people caught in the mimetic thrall of the mounting civil war were understandably losing their bearings. And the many who had married across ethnic lines realized that the world that had once been indifferent to their marriage had “changed so profoundly that what had been accepted and even treasured only a few years ago has suddenly been rendered vestigial and taboo” (Weine 1999: 19).
Rawls’s overlapping consensus is meant to effect the transition from struggle to compromise. The reasonable consensus envisioned by Rawls would depend on people’s ability to engage in consensual practices of inclusion and commitment, operating on those comprehensive doctrines that allow for the realization of an ultimate goal, a well-ordered society (Rawls 2005). Ronald Dworkin has linked this inability to reach a reasonable consensus to dynamics of hyperpolarization in American politics: “Democracy can be healthy with no serious political argument if there is nevertheless a broad consensus about what is to be done. It can be healthy even if there is no consensus if it does have a culture of argument. But it cannot remain healthy with deep and bitter divisions and no real argument, because it then becomes only a tyranny of numbers” (Dworkin 2006: 6).
AbramowitzAlan I. and Steven Webster. 2016. “The Rise of Negative Partisanship and the Nationalization of U.S. Elections in the 21st Century.” Electoral Studies 41 (1): 12–22.
AlesinaAlbertoArnaud DevleeschauwerWilliam EasterlySergio Kurlat and Romain Wacziarg. 2003. “Fractionalization.” Journal of Economic Growth 8 (2): 155–194.
BaldassarriDelia and Andrew Gelman. 2008. “Partisans without Constraint: Political Polarization and Trends in American Public Opinion.” American Journal of Sociology 114 (2): 408–446.
BoothAlan. 1985. “Quality of Children’s Family Interaction in Relation to Residential Type and Household Crowding.” In Habitats for Children: Impacts of Density ed. John Wohlwilland and Willem VanVliet145–164. Hillsdale, NJ: Erlbaum.
CareySabine. 2002. “A Comparative Analysis of Political Parties in Kenya, Zambia and the Democratic Republic of Congo.” Democratization 9 (3): 53–71.
EdwardsHarry T. and Michael A. Livermore. 2009. “Pitfalls of Empirical Studies that Attempt to Understand the Factors Affecting Appellate Decisionmaking.” Duke Law Journal 58 (8): 1895–1989.
FishM. Steven and Matthew Kroenig. 2006. “Diversity, Conflict and Democracy: Some Evidence from Eurasia and East Europe.” Democratization 13 (5): 828–842.
GunitskySeva. 2014. “From Shocks to Waves: Hegemonic Transitions and Democratization in the Twentieth Century.” International Organization 68 (3): 561–597.
HinnebuschRaymond. 2006. “Authoritarian persistence, democratization theory and the Middle East: An overview and critique.” In Democratization13:3373–395.
JensenCarsten and Svend-Erik Skaaning. 2012. “Modernization, Ethnic Fractionalization, and Democracy.” Democratization 19 (6): 1117–1137.
MontalvoJosé Garcia and Marta Reynal-Querol. 2005. “Ethnic Polarization, Potential Conflict, and Civil Wars.” American Economic Review 95 (3): 796–816.
PluyePierreRoland GradAlissa Levine and Belinda Nicolau. 2009. “Understanding Divergence of Quantitative and Qualitative Data (or Results) in Mixed Methods Studies.” International Journal of Multiple Research Approaches 3 (1): 58–72.
RazJoseph. 1999. “Explaining Normativity: On Rationality and the Justification of Reason.” In Engaging Reason: On the Theory of Value and Action67–89. Oxford: Oxford University Press.
SekulicDuskoGarth Massey and Randy Hodson. 1994. “Who Were the Yugoslavs? Failed Sources of a Common Identity in the Former Yugoslavia.” American Sociological Review 59 (1): 83–97.
TepeSultan. 2013. “The Perils of Polarization and Religious Parties: The Democratic Challenges of Political Fragmentation in Israel and Turkey.” Democratization 20 (5): 831–856.
WagnerAlex. 2016. “Trump, Sanders and the Coming Apocalypse.” AtlanticJune 20. https://www.theatlantic.com/politics/archive/2016/06/trump-and-sanders-and-the-apocalypse/486428/.
WarnerBenjamin and Mitchell S. McKinney. 2013. “To Unite and Divide: The Polarizing Effect of Presidential Debates.” Communication Studies 64 (5): 508–527.