Chapter 1: Origins, criticisms and the consociational response

Chapter 1: Origins, criticisms and the consociational response Tracing the origins, and development of a contested model Almost half a century after its first formulation by Arend Lijphart (1968), the consociational model of democracy remains deeply contested and fiercely criticised. Undoubtedly, this span of time has been crucial for the development of a fruitful scholarly debate that revealed its merits and shortcomings, stressed its pitfalls and rectified part of its weaknesses. This debate is still reverberating not only within the field of Comparative Politics but also among other disciplines such as Political Theory, International Relations and Conflict Resolution, revealing the model’s protean nature. Ongoing publications on consociationalism, critical or not, epitomise strikingly this continued interest (e.g Lijphart 2008, Taylor 2009, O’Leary and McEvoy 2013 McCulloch 2014). Since its first formulation, the model has been characterised by methodological contradictions, conceptual confusion and oscillating scholarly aspirations, which obscured its empirical and normative content. These contradictions impacted upon the model’s explanatory and heuristic capacities, the conditions that ushered in its emergence and perpetuation, its democratic credentials, and utterly its potential contribution into the field of conflict resolution and peace-building. The origins of the model Demystifying the architecture and the inner workings of consociational democracy admittedly begins with a systematic exploration of the ideas propounded by its most consistent defender, Arend Lijphart. Lijphart elaborated considerably on the strengths and weaknesses of the model, albeit not being its sole pioneer; a new generation of consociationalists has emerged, that represents the so called “new wave of consociations” (Taylor 2009, 7) within the genealogy of consociationalism. Lijphart attributes the terminological origins of the word “consociational” to David Apter and his study on Uganda (Lijphart 1968, 20 and 2008, 3). Indeed, the latter study appears in a footnote in Lijphart’s seminal article, Consociational Democracy (Lijphart 1969, 211) next to a reference on Johannes Althusius’s writings on consociatio –which quintessentially connotes a society of societies- (Princeton Encyclopedia of Self-Determination). In addition, earlier variations of consociational democracy were to be found –Lijphart claims- in the works of Gerhard Lehmbruch (1967), the Austro-Marxists, -such as Otto Bauer and Karl Renner- (Lijphart 1977, 43) but most preponderantly in the Politics of West Africa (1965) authored by Sir Arthur Lewis. To quote Lijphart, “political scientists like Gerhard Lehmbruch” and himself “discovered consociationalism a few years later”, but “Lewis invented it by trying to think what would be the logical solution to the problems in West Africa” (Lijphart 2008, 278). The purported long historicity of consociationalism and its alleged widespread application served a double purpose: firstly, it acted as a solid basis of support for the model’s normative potential, thus discrediting any allegations of Eurocentricity. Secondly, it enhanced the credentials of consociational democracy; the latter does not merely serve as a tool for the conciliation of divergent visions of the state but it utterly implies a certain kind of logic (Lijphart 2008), or a consociational esprit de corps. The first tentative effort to portray the contours of the consociational type of democracy took place in an article published in the journal of Comparative Political Studies (1968). In this article Lijphart strove to devise a new typology of democratic regimes distancing himself from Gabriel Almond’s prevalent typological scheme. According to Almond, political systems are distinguished according to the criteria of political culture and role structure (Lijphart 1968, 4). These criteria define roughly two political systems: the majoritarian Westminster model (Anglo-American type) and the multiparty proportional system endorsed in many European (Continental European type) countries. Whilst the Anglo-American type is characterised by a “homogeneous political culture” and a “highly differentiated role structure”, those categorised under the rubric of the Continental type are permeated by a sharply fragmented political culture and a plethora of secluded role structures or subsystems (Ibid). For Almond, fragmentation inexorably engenders segmentation and inevitably instability. He includes in this category the Third and Fourth French Republic, the Weimar Republic in interwar Germany but also postwar Italy (Ibid. 5). By contrast to the latter, the UK and the US exemplify two politically stable systems understood in terms of viability, which, according to Lijphart, expressed the dominant preoccupation of the comparativists of the era (Ibid. 8). In the words of Almond, the Continental type is prone to cause inertia (immobilisme) and fatally a ‘Caesaristic breakthrough’ gearing probably towards dictatorship. Instead of the cumbersome nature of the Continental type, the Anglo-American demonstrates a versatile capacity of absorbing external and internal loads on the political system (Ibid. 10) buttressed by a thriving separation-of-powers role culture (that includes not only the three branches of power but forms of civic association within the broader public sphere). Lijphart exposes his theoretical objections on precisely this distinction which he considers misguided. Henceforth, he will try to elaborate his typology by illustrating that some deviant cases disprove Almond’s scheme. He questions the validity of Almond’s typology by exploring the political culture and social structure of smaller European democracies such as The Netherlands, Austria, Switzerland, Belgium and the Scandinavian countries. Whilst political culture is homogeneous (Ibid. 16) in Scandinavia, this property does not equally apply on the other cases. In fact, Austria, The Netherlands, Switzerland and Belgium are essentially segmented societies with mutually reinforcing cleavages demarcating the populace along class, religious and ideological lines. Remarkably though, this group of countries fares satisfactorily in terms of systemic stability despite the conspicuous absence of overlapping membership in formal and informal civic associations (Ibid. 17). The lack of cross-cutting cleavages notwithstanding, the deliberate efforts put forward by the various sub-cultural elites compensate for their absence. This forms the main argument and the cardinal necessary variable for the consociational model to work. Given the dangers lurking in the absence of cooperation and overlapping membership, elites are prone to cooperate for the sake of the public good (Ibid.18). Elites choose therefore in specific critical junctures between constructive or destructive behaviour (Butchensøn 1987, 94). This is the self-denying hypothesis which explains, according to Lijphart why consociations are launched. Thus, “democracies with subcultural cleavages and with tendencies towards immobilism and instability which are deliberately turned into more stable systems by the leaders of the major subcultures may be called consociational democracies” (1968. 20). One year later, Lijphart published one of his most influential articles entitled Consociational Democracy (1969). In this article, he reiterates the aforementioned definition of consociationalism. Consociational democracies are those regimes that whilst highly fragmented in terms of cleavages, exhibit remarkable resilience. Paradoxically, the elites choose to altruistically cooperate to “counteract the immobilizing and unstabilizing effects of cultural fragmentation” (Lijphart 1969, 212). Yet, political stability ceases to be synonymous to viability. Quoting Claude Ake, Lijphart defines political stability in terms of degree, which is “quite out of proportion to its [society’s] social homogeneity” (Ibid.). Whilst viability implies longevity and it could be somehow measured by the temporal resilience of the cabinet, degrees of political stability connote obscurity leaving enough room for speculation. This is a problem that Lijphart acknowledges in his 1977 book Democracy in Plural Societies. As he remarks, political stability is a “difficult” and “ambiguous” term. Therefore, he seeks to employ it in a multidimensional fashion that denotes “system maintenance, civil order, legitimacy, and effectiveness”. Political stability entails foremost low levels of actual and potential civil violence under the veneer of democracy (Lijphart 1977, 4). According to Lijphart, there are four types of democratic regimes: the centrifugal, consociational, centripetal, and depoliticised democracy. Political culture (homogeneous/fragmented) and elite behaviour (coalescent/competitive) are the two defining criteria. The centrifugal type is the most problematic given the level of social fragmentation and the competitive behaviour of the elites which parallels that of the Continental systems in Almond’s typology. On the other hand, centripetalism features strong majoritarian elements: despite the competitive behaviour of the elites, homogeneity on the societal level and overlapping membership prevents any destabilising effects (Lijphart 1968 37-39). In The Politics of Accomodation: Pluralism and Democracy in the Netherlands Lijphart provides an in depth account of a deviant case which “approximates” (1975, 194) the model. But this time the model is named after “accommodation” and not consociation (Ibid). Significantly, the Dutch case amends the third pluralist proposition, that is the “idea of crosscutting affiliations” (ibid 7), thus representing a critical juncture in the lineage of the pluralist theory of democracy as developed by Aristotle, Alexis de Tocqueville, Rousseau and Madison -the first proposition refers to the implicit assumption that viable democracies face grave obstacles in “plural societies”, and the second, to the controversial implications that the existence of many secondary groups may entail- (Ibid. 3-5). In Democracy in Plural Societies (1977) Lijphart furnishes the most elaborate definition of consociational democracy by encapsulating two dimensions: a societal (segmented cleavages) and a behavioural (political cooperation of the segmented elites). According to Lorwin’s approach, “segmented pluralism” attempts to reconcile “religious and ideological” diversity leaving aside linguistic, class and ethnic differences (1971, 141). In this limited application of the model, elites within those segmented settings appear to have a rather nebulous role (Lijphart 1977, 5). By contrast, Lehmbruch lays emphasis on the bargaining side of the consociational spectrum and the altruistic tendency of the elites (their agency) to cooperate instead of adopting competitive strategies commonly bolstered in the majoritarian systems (1975, 378). Hence, Lijphart sketches his consociational model on an inclusionary fashion that embraces Lorwin’s segmented pluralism and Lehmbruch’s Konkordanzdemokratie. Yet, the genesis and development of consociational democracy spurs a number of logical tensions between a theory of democracy that amends previous propositions, and a Weberian ideal type that entails certain qualities. On the one hand, consociational democracy presumably explains stability and democracy under adverse conditions but on the other incorporates both, given its alleged capabilities. This is the tautological problem of consociationalism, which Brian Barry has eloquently stressed (1975b, 480) and which conflates problem and solution, cause and effect in a single definition. Those ambivalences are inevitably perpetuated, embedded as they are in the defining elements of consociationalism which form its backbone. Thus, if segmented pluralism and cooperative attitude on the elite level are the two features of an ideal type, a grand coalition, the mutual veto, proportionality and segmental autonomy are the ‘devices’ (Lijphart 1968, 21) or what Lijphart interchangeably terms as ‘elements’ and ‘characteristics’ of consociationalism (1977, 2002, 2003, 2004, 2008). In this early definition, a plural or divided society is essentially the model’s fifth element. Yet, Lijphart will later drop plural society from the definition. In fact, he reconsiders the validity of the initial fourfold typology, which “exaggerated the danger of democratic instability in the French Fourth Republic and Italy (the major examples of centrifugal type) as well as in the depoliticized type of democracy” (2000, 426). Within the context of the final formulation of consociationalism, the grand coalition prevails as the “prototypal consociational device” (Lijphart 1977, 31) through its manifold manifestations: it varies from formal schemes such as coalition cabinets (executive branch) to informal e.g. advisory councils (horizontal power-sharing or shared rule). Its essence lies in the legitimate representation of all major segments in the grand coalition and thus of all interests within its bounds. Whilst the grand coalition scheme can be crystallised in both formal and informal forms of representation, Lijphart posits in his 1968 article as a condition conducive to consociational democracy a popularly legitimated government by grand coalition (Lijphart 1968, 28). This probably implies that Lijphart had foremost in mind popular legitimation, enshrined in the formation of a collective cabinet. Yet, O’ Leary contests the idea of an all-inclusive consociation (complete consociation). Specifically, he identifies, besides the complete, all-encompassing executive of Lijphart, two other types of executive power-sharing: firstly, a concurrent executive, where each major segment is represented enjoying majority support from the communal pillar (50% plus). Secondly, a weak consociation, where each significant segment sends its democratically elected delegates in the executive but at least one of the communal leaders in it enjoys plurality support by its segmental followers in contradistinction to the majority support enjoyed by the other incumbents (2003, 12-13). In addition to the grand coalition, the mutual veto functions as a safety valve that allays the fears of minorities on issues deemed of vital importance. Its empirical manifestations are mirrored in forms of concurrent majorities in the legislature and/or the executive. Furthermore, the third element or device, proportionality, implies two aspects: firstly, the proportional allocation of civil service appointments or financial resources to the sub-cultural segments of the given society, and secondly, the adoption of the proportionality rule (PR) in elections. Finally, segmental autonomy can be both functional and territorial. In the first case, constituent communities are entitled to manage autonomously their educational and cultural affairs. By contrast, territorial autonomy refers to forms of self-rule based on carefully demarcated borders and competences (vertical power-sharing) such as federations, confederation or federacies (e.g. O’Leary 2008, Wolff 2008, 2009, 2010, O’Leary and McEvoy 2013). The importance of each of the four devices included in Lijphart’s theoretical scaffolding, the effects of their combined utilisation in different constellations or its universal endorsement remains uncertain. In the Typologies of Democratic Systems, Lijphart indicates that “the essential characteristic of consociational democracy is not so much any particular institutional arrangement as overarching cooperation at the elite level with the deliberate aim of counteracting disintegrative tendencies in the system” (1968, 21). Consequently then, the grand coalition is translated into the institutional embodiment of the altruistic attitude exhibited by the sub-cultural elites, whilst its inclusion into the toolkit of devices becomes at least dubious. Perhaps aware of those inconsistencies, Lijphart introduces a distinction between primary and secondary characteristics: “grand coalition and autonomy are the most crucial, whereas the other two are occupying a somewhat lower position of importance” (Lijphart 2004, 2008). This distinction notwithstanding, the indispensability of those characteristics remains ambiguous in the absence of the master variable of elite cooperation which theoretically explains stability in fragmented societies. The model’s normative and geographic expansion will further exacerbate those inconsistencies: yet, the four devices remain at the core of the theoretical scaffolding for a number of younger adherents of consociationalism (e.g. McGarry and O’Leary 2004, 2006, 2007, McCulloch 2014), whilst others speak about the diffusion of some “ahistorical principles” (Bogaards 2000, 409) or ultimately about a degenerated research programme (Lustick 1997). Critiques of consociationalism Consociational democracy has been brought under the spotlight of critical scrutiny on different grounds. Consociationalism, as underscored in the previous section, has suffered since its inception from conceptual and methodological shortcomings. Admittedly, this does not imply the model’s abrupt rejection or repudiation. Lijphart has progressively tried to rectify parts of those shortcomings by integrating into the model bits and pieces of the criticisms mounted against the theory. Critiques of the consociational model can be organised into three clusters. Firstly, a cluster of criticism pertains to the factors and conditions that buttress the model and that presumably enhance its predictive capacity. Secondly, a number of scholars denied the democratic character of consociations in terms of accountability, legitimacy, electoral contestation and political equality. Finally, a third debate, which is inextricably interwoven with the aforementioned ones, concerns the prescriptive capacity and adequacy of consociational democracy in plural (deeply divided societies) in contradistinction to majoritarian/centripetalists/ integrative-leaning models. Prerequisites, factors and conditions (1968-1977) Methodology is one of the main stumbling blocks in the theory of consociational democracy. Drawing inductively from specific case studies such as The Netherlands, Austria, Switzerland, Belgium, Lebanon and Malaysia, Arend Lijphart tried to extract some ‘factors’, ‘conditions’, and prerequisites that explain and predict the establishment of consociational democracy. Their subsequent universalisation raised eyebrows among the scholarly community regarding their status, validity, predictive capacity and significantly, their utility as deterministic panaceas. This confusion was further aggravated semantically by the interchangeable use of terms. Following an inductive approach, Lijphart offers a list of ‘prerequisites’ in the Typologies of Democratic Systems that explains why consociational democracies emerge and how they become sustained (1968, 22) if and only if rival sub-cultural elites internalise the norms engendered by the model. Lijphart identifies four attitudinal pre-requisites/attributes that he preserves in subsequent publications as largely compatible with the voluntaristic character of the model (1969, 1975, 1977): ability (probably refers to consciousness/ altruistic ethos) to recognize the dangers inherent in a fragmented system, commitment to system maintenance, ability to transcend subcultural cleavages at the elite level, ability to forge appropriate solutions for the demands of the sub-culture (Ibid. 22-24). To compensate for the predictive paucity of the prerequisites, which rely exclusively on the elitist volition, Lijphart enumerates six conditions (actually more) at the mass level that are “conducive to consociational democracy” (Ibid. 25): distinct lines of cleavage between subcultures, a multiple balance of power among the subcultures, popular attitudes favorable to government by grand coalition, external threats, moderate nationalism, a relatively low total load of the system. Whilst those are the conditions enumerated, Lijphart adds an additional condition in the last paragraph under the rubric of size. He indicates that what all cases share in common is a small surface. Crucially then, size emerges as a factor that has an effect on the last three (external threats, moderate nationalism, low load on the system). In Consociational Democracy (1969), Lijphart increases the number of the favourable factors (not termed conditions in this article) that generate and bolster elite cooperation (length of time a consociational democracy has been in operation, external threats, multiple balance of power among subcultures, relatively low total load on the decision-making apparatus, distinct lines of cleavage, internal political cohesion of the subcultures, adequate articulation of the interests of the subcultures, widespread approval of the government by elite cartel) (1969, 216-22). In the Politics of Accommodation, he discovers seven unwritten, implicit “rules of the game” that define the role culture in The Netherlands (1975, 122-123) that underpin perpetuation of stability in the political system. Yet, Lijphart does not substantiate their relation firstly, to the favourable factors and secondly, to the consociational devices used in The Netherlands, although he tends to include elements from both lists in the “rules of the game” (e.g. proportionality, deference of the mass to the subcultural leaders). Within the framework of The Netherlands, Hans Daalder contends Lijphart’s reading of events by remarking that the accommodative style that led to Dutch conciliation (1917) must be traced to the sixteenth, seventeenth and eighteenth century. Hence, the ensuing institutionalisation of consociationalism was not the effect of deep cleavages (as Lijphart argues) but the reason why those cleavages were ossified. Therefore, a culture of conciliation deeply rooted in history must be considered as a factor conducive to inter-elite cooperation (Daalder 1974, 616) especially under conditions of modernisation and mass mobilisation. Daalder’s scepticism reflects the problem of circularity inherent in the theory of consociational democracy. Lijphart elaborated in a more sophisticated fashion the factors that facilitate accommodation in Democracy in Plural Societies (1977). He embarks upon this effort by recognising that a prudent leadership plays a cardinal role in plural societies, not in appeasing the masses but the followers, that are defined as a distinct group of sub-elite political activists (1977, 53). In addition, he introduces a distinction between favourable and unfavourable factors. Yet this distinction loses much of its heuristic and predictive value, as the favourable factors are deemed hardly “indispensable” or “sufficient” (Ibid. 54). Among the factors identified, a multiple balance of power is the first. According to Lijphart, this factor ideally entails either a balance -or an approximate equilibrium- or the presence of three segments. Hence, all segments shall have the status of minorities when considered in relation to each other. When examined separately, segmentation should be internally translated into three or fours fragments. For Lijphart believes that negotiations are easier among fewer segments –ideally four- as this power constellation does not translate into zero-sum games of adversarial politics (Ibid. 56). Yet, realities in fragmented societies are usually unpredictably complex. For instance, Lebanon is officially divided into two segments (Christians and Muslims). This initial distinction notwithstanding, Christians are further subdivided in confessions such as Maronites, Greek Orthodox, Greek Catholics and ethnic groups such as Armenians (both Orthodox and Catholic). On the other hand, the Muslim community is chiefly divided between Sunnis and Shia in addition to the smaller Druze and Alawite sects. As De Juan notes “most ethnic groups are not internally homogeneous. Rather, they comprise various subgroups with different religions, languages, or customs. Such subgroup differences are often hidden by salient intergroup conflicts, yet they may also resurface when intergroup conflict has been settled” (2013, 79). A multiparty system is the second favourable factor by serving a duality of functions: ethnic parties represent segmental interests, whilst the communal electorate chooses their preferred delegates in the grand coalition (1977, 61-62). Multipartism with relatively few parties is according to Lijphart the ideal state of affairs (Ibid.) in accordance with the ideal number of segments mentioned above. Nevertheless, pure PR creates an opportunity structure that facilitates the establishment of new parties and thus new balances in the party system (McCulloch 2014, 134). Therefore Lijphart adds the proviso that moderate multipartism conveys its favourable effects in plural societies where all parties enjoy a minoritarian status and are of almost equal size (1977, 64). Size is a recurrent theme in Lijphart’s work. In Democracy in Plural Societies, Lijphart strives to explicate analytically the effects of size on consociational democracy. Implicitly though, he shifts his focus to Europe by arguing that “a striking characteristic of European consociational democracies is that they are small countries”(Ibid. 65). This statement contradicts parts of his previous work, where Lijphart included Lebanon among the iconic cases of consociational success. Probably, the collapse of the Lebanese state in 1975 acted as a major blow to the theory thus raising questions on the vicissitudes of state formation in the developing world. As regards size, Lijphart identifies four effects, -direct and indirect- (1977 65-70). The direct internal effect of smallness refers on the one hand, to the increased likelihood of elite familiarity which enhances accommodation. Despite its plausibility, closeness might instead exacerbate misperceptions and misunderstandings. Indeed, elite familiarity did not prevent collapse in Cyprus (1960-63). On the other hand, smallness limits decisively the reservoir of political talent which is crucial for the smooth functioning of a voluntarist model (Ibid. 66). Furthermore, the direct external effect assumes that an external threat will galvanise a trans-communal front for the sake of the state’s survival. Lijphart emphasizes the need of shared threat perception as, echoing Lehmbruch, internal replication of an external conflict hovers as the sword of Damocles over the society (Ibid. 67). Yet, perceptions of foes/friends differ markedly as the case of interwar Austria showcases. However, the indirect effects function in the benefit of the decision-making process, as smallness reduces the load exerted on the system. Theoretically, small size implies fewer preoccupations in the management of the internal and external affairs. In what regards the indirect internal aspect, “smaller countries are easier to govern than large ones because they tend to be less complex: the number and variety of groups and individuals whose interests and attitudes have to be taken into consideration are fewer”. But a small population does not automatically imply fewer demands for policymaking. Belgium, Cyprus (1960-1963) and Bosnia and Herzegovina were and still are extremely complex settings as evidenced in the establishment of multiple layers of power and the recurring crises of governance. Finally, Lijphart depicts a causal sequence in explaining the indirect external effect of smallness. A state’s limited power on the international scene leads to neutrality in foreign policy and consequently to minimal dilemmas. Nevertheless, neutrality does not deterministically imply a state’s survival in cases where regional actors value strategic calculations more than respect for sovereignty e.g. Lebanon, Bosnia and Herzegovina. Moreover, neutrality is an inherently political stance that may subsequently arise as the source of inter-communal frictions over the right course of (in)-action the state should follow. The next set of favourable factors relates to the cleavage structure (number, type and fragmentation) (Ibid 71-72). Within this context, Lijphart argues that “when there are two or more cleavages, one must examine how they relate to each other and, in particular, whether they tend to crosscut or to coincide” (Ibid. 75). In the case of cross-cutting cleavages, the mode of their intersection affects the number of segments and ultimately the balance of power between them. In addition, the cleavage structure impacts upon the inter-segmental “intensity of feelings”. This is a paradoxical inclusion, for consociational democracy excels by ‘producing’ political stability amidst mutually reinforcing cleavages. Lijphart deems therefore the contribution of the cleavage structure as desirable but “subsidiary” (Ibid 81). Thus, cleavages of equal intensity can have a positive effect when they lead to an all-minority state of fragmentation, and a negative, in cases of fractionalization (Ibid.). Still, unequal cleavages engender fewer and less homogeneous segments that under conditions of crisscrossing will yield moderate effects (Ibid). Furthermore, overarching loyalties enhance accommodation by moderating the conflictual potential of cleavages. Lijphart stresses the significance of some kind of nationalist spirit. Within the context of fragmentation, nationalism should be perhaps understood as a thin narrative that elicits a modicum of trans-segmental feelings of togetherness. Given the absence of such a unifying national narrative, Lijphart suggests segmental isolation and federalism. Lijphart argues that conditions in plural societies do not favour inter-communal exchanges on the mass level. Hence, communal seclusion appears preferable (Ibid 88) to forestall tensions in the ethnic mosaic, thus enhancing the prospects of inter-elite cooperation on the apex of the social pyramid. Whilst segmental isolation is to be found also as part of the consociational device of segmental autonomy, federalism is not. Federalism emerges as a favourable factor in the case of compact minorities which are geographically concentrated in homogeneous regions (Ibid.). Homogeneity is a permissive condition for a spatially-based organisation and management of communal affairs. Yet, Nordlinger contends that federalism invites secessionist trends when the federal region borders states inhabited by groups linked to it with relations of kinship (1972, 110). Last but not least, Lijphart seems to heed Daalder’s criticism on the origins of Dutch consociation. In contrast to his previous ventures, Lijphart not only includes a prior tradition of accommodation as a factor conducive to consociation but he also attributes it a higher value in comparison to the above mentioned ones (1977, 103). Yet, its addition into the list of the favourable factors makes only a marginal contribution to the problem of circularity: it is unclear whether a tradition of accommodation enhances the prospects of consociational democracy or rather it is consociationalism that entrenches this tradition. Factors and conditions (1985-2008) Democracy in Plural Societies has profoundly shaped Lijphart’s late work as he gradually started to settle on the architecture of consociationalism. Therefore, his aspiration to explore the potentialities of consociationalism in new cases comes perhaps as of no surprise. Yet, this shift to prescription carries with it part of the weaknesses left unaddressed during the formative years of the consociational model. This is more patently crystallised in some of the amendments put forward in the list of the favourable factors, which are modified to fit in the new cases. In this vein, Power-sharing in South Africa (1985) features conspicuously as an analytic endeavour to study the consociational prospects of this society. In this monograph, Lijphart alters the favourable conditions by reducing its number from nine to eight (1985, 119-128) trying to adjust the theory to the exigencies of the South African society. Firstly, Lijphart re-introduces the presence of external threats (1968 list) as a favourable and autonomous factor. Secondly, size and its outward manifestation, -smallness-, are replaced by a small population. It should be noted that Lijphart has already insinuated the importance of a small population in Democracy in Plural Societies, when explicating the direct internal effects of size (familiarity, personal relations/ poor reservoir of political talent). Thirdly, socioeconomic equality is conducive to a stable consociational regime. The inclusion of socioeconomic equality could possibly echo the lingering disparities between Europe and the developing world. Interestingly, a discussion on the role of socio-economic conditions has appeared already in a chapter on Consociational Democracy in the Third World (1977, 174) -not endowed with the status of ‘factor’ though-. Finally, Lijphart bestows the scientific veneer of positivism on the theory by quantifying the favourable and unfavourable conditions along a five-point scale. The results of this quantification were undoubtedly controversial: the measurement was based on flimsy assumptions and wishful thinking as evidenced in the prediction that the black community will remain fragmented along tribal lines, thus producing the necessary multiple balance of power (McGarry and Noel 1989, 7). Lijphart will finish the long peregrination on prerequisites, factors and conditions in an article on India (1996). As he confesses in the Introduction of Thinking about Democracy (2008), among the nine conditions found in the South African and Indian cases, those that matter the most are the absence of a solid ethnic and religious majority, and the absence of large socioeconomic disparities among the groups of a divided society (2008, 5). As can be inferred from the article on India, the following “background factors” represent the final ‘inventory’: absence of solid majorities, absence or presence of large socioeconomic difference among the communal groups, moderate number of segments, a multiple balance of power, a small population, external dangers, overarching loyalties, functional or territorial isolation, traditions of compromise (Ibid. , 51-52). Building on Lijphart’s prior work, Brendan O’Leary tries to investigate the prospects of a successful consolidation of consociationalism in Northern Ireland (1989, 2004) by shifting the level of analysis to the external setting. This is an important modification in the theory of consociational democracy as Lijphart’s work is primarily focused on endogenous factors. By exploring the status of the background conditions, O’Leary concludes that Northern Ireland scores negatively on almost all, with the sole exception of size and segmental isolation (2004, 113). Engineering some of the favourable conditions tends to be an implausible a solution too, unless elites are willing to fulfill the self-denying prophecy. He infers that given the conspicuous absence of inter-elite motivation, structured elite predominance and intra-segmental stability the prospects of a voluntary consociation are reduced to naught. Therefore, external actors should be invited to engage proactively in enforcing a consociational settlement. O’Leary underlines that a number of external actors -such as the UK and Ireland- should channel their efforts on two fronts: the external and the internal. On the external front, coercive consociationalism should entail an impartial support of consociation, whilst on the internal, the promotion of social and economic reforms that would redress past injustices (Ibid. 122-123). In the Foreword of Imposing Power-sharing (2005), O’Leary underscores the function of international norms as a factor conducive to consociations as reflected in the compromise of state sovereignty under the guise of international protectorates. He avoids nonetheless any reference to Lijphart’s list of conditions with the sole exception of size: yet, he tends to reject any direct or indirect effects derived from it. Nevertheless, a latent correlation between the establishment of consociational protectorates, small size and international norms (2005 xxxiii) might perhaps be established. Bosnia-Herzegovina and the Former Yugoslav Republic of Macedonia (FYROM) exemplify cases, where the USA and the European Union have exerted considerable coercion coupled with inducements to bolster consociational settlements (Ibid.). Whilst both approaches attempt to explain why consociations emerge and come into fruition, none of them elaborates how and why they are sustained. Lijphart has largely ignored this discussion by granting to the prerequisites both genetic and sustaining qualities (1968, 22). Conditions were simply relegated to the status of facilitating or favourable factors (2008, 5), whose utility looks in the bottom line, ambivalent. Adriano Pappalardo criticises Lijphart for precisely neglecting the role of conditions in ushering to consociational democracy (1981, 365). For Pappalardo, inter-subcultural stability and elite predominance over a politically deferential and organisationally encapsulated following are the only indiscriminately necessary and sufficient conditions for consociational democracy to succeed. In response to Pappalardo’s critique, it is overwhelming voluntarism that pervades Lijphart’s model translated into reliance on preponderantly attitudinal variables. Arguably, Ljphart was preoccupied with the model’s prescriptive expansion which may explain why factors and conditions were useful at the end of the day as mere indicators of success. An account on genetic and sustaining conditions is provided by Gerhard Lehmbruch who lays emphasis on the international environment and its impact on the perceptions of the elites (1975, 380). For Lehmbruch, genetic conditions involve the presence of some national symbols, a bitter, violent past, a tradition of accommodation, intense informal communication on the elite level, the absence of majorities (internal) and the presence of external threats (external). Consociationalism emerges when the internal favourable factors coexist with an imminent threat. Still, Lehmbruch identifies neutrality as the unique sustaining factor, which is strongly related to smallness and threat perceptions. However, Bogaards rejects Lehmbruch’s conditions as the birth of Dutch consociationalism did not witness any kind of external threats. Hence, the conditions that Lehmbruch posits should be viewed as requisites rather than as conditions (2000, 494). The discrepancy between genetic and sustaining conditions looms large in O’Leary’s theorisation of consociational engineering. Whilst crafting favourable conditions could be a solution, this artificial reality could wither away as soon as external actors withdraw or lose their interest in the process. Northern Ireland exemplifies a successful example of consociation (McCulloch 2014) without being a state though. Being part of the United Kingdom, Westminster has still a legitimate say in the internal proceedings in Belfast. Crucially, external engineering might be perceived by the society in question as another form of Western imperialism or standard of civilisation (Zaum 2007) within the confines of an unjust international society. Therefore, it is imperative to shed light on processes of constitutional engineering in order to trace the footprints, legacies and residues of exogenous interventions in the post-trusteeship era. By delving into cases such as Northern Ireland, Iraq or Lebanon, consociationalists can extract some informed hypotheses on the conditions of successful and failed settlements. Consociational democracy or consociational ‘democracy’? A second set of critiques revolves around the concept of democracy and its different facets within the consociational conceptual universe. Those discrepancies refer firstly, to the definition of democracy within the consociational literature, and secondly, to the tensions ravaging the model that view consociational democracy as an independent variable and a self-standing descriptive label at the same time. Democracy can be credibly defined as “a means of government in which… the people rule. It entails a political community in which there is a form of political equality among the people” (Held 2006, 13). On the other hand, consociational democracy straddles a variable that explains “political stability”, “democratic stability”, “peace and democracy”, (dependent variables that Lijphart has used interchangeably) over disparate communities coexisting within a society, and a descriptive category or an ideal model whose defining features entail a grand coalition, mutual veto, proportionality and segmental autonomy. In Democracy in Plural Societies, Lijphart eschews to denote democracy describing it merely as a concept that “defies definition” (1977, 4). He seems content to state that democracy will be henceforth understood in Dahl’s terms of ‘polyarchy’ (Ibid.). According to Robert Dahl, polyarchy essentially implies two dimensions: public contestation and inclusiveness (1971, 4). In an ideal democratic process, democracy should satisfy the following criteria: equality in voting, effective participation, enlightened understanding, final control over the agenda, inclusion (1982, 5). It is in this realm, that consociational democracy has been severely criticised. According to Hans Daalder, for instance, the rigidly hierarchical structures and the secretive brinkmanship that consociational democracy entails, promote “procedures of purposive depoliticization”, encourage “bargaining behind closed doors”, whilst undermining direct accountability (1974, 607-608). Indeed, the hierarchical structure of consociationalism has a difficulty in grappling with notions such as contestation, alternation in the executive, responsiveness and accountability. Firstly, contestation in the political field emerges as an inherently antithetical concept to the prerequisites of consociationalism. The materialisation of the messianic message of consociational democracy resides in the goodwill, the altruistic spirit and collegial ethos of the relevant elites. Inevitably, conciliation, decision-making by amicable agreement and consensus preclude any room for disagreement as this would in the long run lead to collapse. As delineated by Arend Lijphart (1968, 1969, 1977), consociationalism relies upon i) the elites’ ability to accommodate the divergent interests and demands of the subcultures, ii) the ability to transcend cleavages and to join in a common effort with the elites of the rival subcultures, iii) commitment to the maintenance of the system and improvement of its cohesion and stability and, iv) a shared understanding on the perils of political fragmentation (1969, 216). Therefore, structured elite predominance becomes imperative in order to appease oppositional reactions and enhance cohesion within the subcultural segment. Eric Nordlinger posits four conditions that facilitate the establishment of a secure and firm prevalence of the elites: apolitical quiescence and acquiescence to authority, patron-client relationships and mass parties (1972, 86) which are complemented by the practice of purposive depoliticisation. Deference is a trait propounded by Pappalardo’s account as well (1981, 365). In the Politics of Accomodation, Lijphart concurs with Nordlinger as he inductively identifies “depoliticization”, and “secrecy” as part and parcel of the rules of the game (1975 129-134); “deference” is posited as an extra factor that is conducive to conciliation (Ibid. 144- 162) on the inter-elite level. Depoliticisation in The Netherlands was achieved through the allocation of economic benefits, discussions on economic matters largely incomprehensible to the public and constitutional principles (Ibid. 129). Secrecy involved decision-making behind closed doors and limited dissemination of information (Ibid. 131) given the idiosyncrasy of the societal structure. Moreover, deference was defined as an “individual’s acceptance of his position both in social hierarchy and on the scale of political authority, accompanied by a low level of participation and interest in politics” (Ibid. 144-145). Under the conditions of reduced information, apathy, dearth of robust opposition, and responsiveness, contestation loses much of its dynamic. Lee Dutter asks whether The Netherlands can really be called a democracy whilst Daalder describes the system as an “oligarchy” (cited in Lustick 1997, 104). Contestation is further undermined by semantic shortcomings. In his early endeavours on consociational democracy, Lijphart framed the cabinet under the rubric of an “elite cartel” (1968, 1969). The connotations were remarkably deleterious as it ultimately implied an established cohort that rules in defiance of the electorate. Interestingly though, Lijphart repeals the latter concept by replacing it in Democracy in Plural Societies with that of the grand coalition. The grand coalition represents the main collegial body of decision-making, where adversarial politics are to be attenuated. But the concept of the grand coalition elicits certain dilemmas regarding the permissible degree of contestation. As noted above, the cabinet in the consociational model can be expressed in different schemes. Lijphart recognises as forms of grand coalitions among others, diachronic alliances (1977, 30) between two or more parties that tend to rule for an extensive time span, and rotating presidencies e.g. Colombia (Ibid. 33). Inexorably then, the merits of suffrage are reduced to participation in a symbolic procedure, that promotes clubism, whilst abandoning viable opposition politics (Brass cited in O’Leary 2005, 6). Van Schendelen castigates the conspicuous absence of opposition politics in consociationalism by arguing that in polyarchy, competition is the sine qua non in sharp contradistinction to consociationalism’s collaboration (own emphasis) (1984, 32). Opposition politics are not ruled out totally, according to Lijphart. For, as long as a parliament functions, a locus of criticism will be always available. Alternatively, demands for accountability could be raised in front of individual members of the grand coalition (Ibid. 48). Yet, Lijphart does not explain how criticisms mounted in the parliament can have an impact on the collegial executive without jeopardizing conciliation within the coalition. Importantly, the grand coalition comprises in principle all parties in the legislature. Moreover, holding accountable individual members of the cabinet depoliticises the conduct of politics not in the direction of moderation but in favour of personal rivalries and ‘vendettas’. Andeweg adopts a more realist stance by arguing that the lack of opposition in consociationalism notwithstanding, elections are held in a competitive fashion (2000, 530) or at least, in a more competitive in comparison to other arenas in the public sphere. O’Leary concurs that a “consociation does not eliminate democratic opposition within communities, but it does enable such divisions and oppositions as exist to flourish in conditions of generalized security” (2005, 11). . Contestation and opposition politics are feasible then under the form of concurrent and weak consociations (Ibid. 12-14). O’Leary presents three types of consociationalism, where opposition is possible. Firstly, a dominant segment rules on a consensual basis over its followers, whilst exerting supreme control over a minority. Israel epitomises an example of consociation coupled with control over the Arab segment of the Israeli society. Yet, it is doubtful that Israel can really express the consociational ethos as the systematic exclusion of Israeli Arabs violates both the model’s letter and the spirit as incarnated in inclusion. A second case is that of deliberate self-exclusion e.g. Northern Ireland (1976-1998), and a third, the introduction of thresholds in what regards participation into the executive e.g. South Africa, Northern Ireland. Yet, thresholds have often been manipulated by incumbents to constrain fair representation of minorities. e.g. Turkey (10% threshold). Therefore, introduction of thresholds is viewed with suspicion as impairing on the inclusionary aspect of consociationalism. It is precisely the inclusive character of consociationalism that Lijphart acclaims when arguing that it is “the best kind of democracy that can realistically be expected…” He explains that given the intractability of segmental cleavages, exclusion of “the minority segment or segments permanently from participation in the government” would be undemocratic (1977, 48). Hence, Lijphart attributes excessive significance to the inclusionary dimension of polyarchy. Indeed, Andweg concurs that consociationalism can on this ground “outperform adversarial democracies, since its very aim is to prevent the permanent exclusion from power of any social segment” (2000, 530). Yet, Lijphart appeared to initially have reservations on the model’s democratic quality (Bogaards 2000, 402). Responding to Bogaards criticism, Lijphart admits that he no longer believes that there is anything inherently undemocratic in the four consociational elements but he was merely influenced by the majoritarian bias of the time (Lijphart 2000, 426). The majoritarian bias is eminently tackled with inclusion and participation of all segments in the democratic process. Lijphart bases his conviction on the qualitative advantages of consociationalism in the Politics in West Africa, written by the economist Sir Arthur Lewis (1965). According to this book, the former British colonies of West Africa proved to be misguided by uncritically adopting the Westminster tradition of the former rulers. The reason lies in the one-party rule and the exclusion of major segments of the populace (Lijphart 1977 143-147). In Lewis’s thinking, majority rule should not be equated with bare majority rule. Instead, democracy finds its fullest expression in wider majorities, where citizens have the right to participate, directly or indirectly in the decision-making process (Lijphart 2008, 126). In the field of consociational democracy therefore, proportionality guarantees political equality through fair representation of all segments. In the words of Arend Lijphart, election by PR is the main self-determination method (Ibid. 6). In Power-Sharing in South Africa, Lijphart suggests that the most basic democratic principle is that of one man and one vote (1985, 81). He recognises though that equality through proportionality sometimes takes different forms as reflected in overrepresentation of minority segments and parity (1977, 41). Yet, it is not very clear why one-party rule should necessarily usher in a tyranny of the majority. Majoritarianism does not necessarily imply one-party rule as the example of the Conservatives-Liberal Democrats coalition in Britain explicitly demonstrates, -although not a plural or deeply divided society-. Power-sharing within a grand coalition is perhaps possible even under majoritarian electoral laws e.g. Lebanon. Besides, overrepresentation and parity could rightfully be blamed to violate political equality on the grounds of minority protection (Cyprus, Burundi) or favouritism (Lebanon) by external actors. The examples of Cyprus and Lebanon indicate that overrepresentation within a consociation can have the adverse effects by endorsing demands of fairness. However, contestation and inclusion epitomise the Janus-faced physiognomy of an ideal regime. Lijphart claims that it is either consociationalism or no democracy at all (1977, 238). The dilemmatic character of this proposition notwithstanding, it would be preferable to argue that as ideal type, democracy leaves ample space for creativity. In this sense, “If democracy is perceived as a device to keep political elites responsive and accountable to the masses, competition (and not just in the electoral arena) is essential, but if democracy is perceived as the avoidance of tyranny, inclusiveness is valid (Andeweg 2000, 531). Inevitably then, emphasis between the two dimensions of polyarchy depends on the preferences, mindset, and influence of constitutional engineers and political elites in a given formative critical juncture. Segmental autonomy and self-determination A sub-critique related to the democratic credentials of consociational democracy emanates from political theorists of the liberal and Marxist strand, who argue that, instead of freedom and self-determination, consociationalism reifies segmental identities (Kieve 1981, Brass 1991, Taylor 2001, 2009). This is most conspicuously reflected in the consociational device of segmental autonomy and the mutual veto. Whilst the grand coalition incarnates shared-rule, segmental autonomy champions self-rule on matters deemed vital for the ethnic segments (1977, 41-43). According to Lijphart, segmental autonomy “increases the plural nature of an already plural society” (Ibid. 42). This is culminated in the establishment of a plethora of voluntary communal associations and groups based on ascriptive pre-determination. Lorwin concedes that segmental pluralism is founded on those forms of association and its recognition (1971, 141). In Lijphart’s formulation, segmental autonomy echoes resoundingly Dahl’s conception of polyarchy. Yet, Robert Dahl warns that one of the defects of pluralist democracy is mirrored in the deformation of civic consciousness for aspects of the self tend to identify with the interest of the group instead of the public’s (1982, 43-44). This is highly expected to take place in a conflictual setting, where mistrust and perceptions of otherness rule supreme, and where organisations are structured on the basis of communal affiliations. Brass demurs at segmental autonomy, for communal stratification mistakenly perceives identities as objective factors, thus violating the rights of groups and the rights of individuals (cited in O’Leary 2003, 5-6). On a same note, David Held notes that the idea of democracy derives its power and significance from the idea of self-determination that is the freedom of citizens to choose their own association. Those choices should constitute the ultimate legitimation of the form and the direction of their polity (1995, 145). Rupert Taylor disapproves of the 1998 Belfast Agreement that grappled with the Northern Irish question on the grounds that it violated the individual right of self-determination. Instead, he propounds an approach that transcends ethnic boundaries and promotes social transformation through the engagement of movements (2001, 46-47). An interesting critique on the falsification of identities within the consociational realm comes from Marxist scholars. Kieve in particular, notes that political elites in The Netherlands perpetuated their rule and the existing rules of production through the manipulation of religious identities that cut across the class-based social movements (1981, 332). Indeed, Lijphart initially perceived identities as primordial and thus not potentially malleable (1977, 17). Thus, segmental autonomy arises as a necessary modality to allay mutual fears and suspicions. Nevertheless, as he confesses in The Rights of Minority Cultures, the case of South Africa exerted a decisive influence upon his thoughts on the capacity of political engineering in plural societies. Lijphart admits that in South Africa racial and ethnic identification artificially overlapped (1995, 280-281), creating bogus categories. He thus concludes that self-determination should be the default rule rather than pre-determination (Ibid. 282). This reconsideration notwithstanding, self-determination in South Africa was to be undermined by the proportional allocation of public offices based on ascriptive criteria in conjunction to the results of the general elections. Lijphart reiterates this preference for self-determination in Thinking About Democracy (2008), where he categorically states its superiority against pre-determination (2008, 4) and corporate forms of power-sharing. Preference for self-determination is echoed by scholars such as John McGarry and David O’Leary (e.g. 2004, 2007, 2009). They both propound a less rigid form of consociationalism that gives some breathing space to the individual in order to freely identify herself. Within the context of liberal consociationalism, self-determination is eminently manifested in voting on common electoral registers (O’Leary 2003), preferential balloting (McGarry and O’Leary 2001, 2004), and explicit protection of human rights and civil liberties. Towards a consociational universe? The third cluster of critiques pertains to the model’s expansion and evolution and shift from an empirical to a prescriptive type of regulating conflict within fragmented societies. As Andeweg underlines “consociationalism has expanded from an amendment to democratic theory intended to help explain the existence of democratic stability in a few small European countries, to a normative theory of consociational engineering in practically all deeply divided societies”(2000, 517). In this vein, numerous scholars have stressed the model’s frail foundations as reflected in its iconic case-studies. This evokes naturally questions regarding the applicability of consociationalism to other cases, bringing it into a ‘collision course’ with ‘competing’ models of power-sharing e.g. integration/centripetalism. At the epicentre of this controversy lies the notion of the plural society as one of the two indispensable components (the other being the grand coalition) upon which the accommodative edifice is grounded. On the notion of the plural society, Lijphart employs a definition conceptualised by Harry Eckstein as division along a number of “segmental cleavages” (1977, 3). Segmental cleavages may be of a religious, ideological, linguistic, regional, cultural, racial or ethnic nature, whilst civil associations are organised along those segmental lines (Ibid. 4). Daalder notes that consociationalists “assume without detailed political analysis that social divisions are automatically translated into political conflicts” (1974, 614), which implies preoccupation with a conflict’s intensity and thus the model’s indispensability. Within this framework, Brian Barry cautions against the blind implementation of consociationalism by comparing Northern Ireland to The Netherlands as a classical example of accommodation. Specifically, he indicates that comparing Northern Ireland to The Netherlands is not worthwhile given the different nature of the cleavage structure. Thus, segmentation is based on religion in the Dutch case, whilst the Northern Irish question evokes ethnic differences (1975a, 408) which render intra-segmental stability and accommodation implausible. Moreover, Barry rejects the classification of Switzerland as consociational due to the broad use of referenda in the decision-making process that inherently questions the adversarial nature of the Swiss society and the level of intersubcultural hostility (1975b, 488). Barry’s argument unequivocally challenges the consociational model per se; Lijphart argues that if any of the six archetypical examples of consociational democracy (Switzerland, The Netherlands, Belgium, Austria, Lebanon, Malaysia) are found not to be really plural societies, consociationalism can be potentially declared null and void (1985, 104-105). Therefore, he suggests that firstly, plurality is a matter of degree, “ranging from 0 percent pluralism in a completely homogenous society to 100 percent pluralism in the most extreme case of a plural society” (1981, 355). Secondly, degrees of pluralism are to be distinguished based on four criteria: In a completely plural society, it must be possible to identify exactly the segments into which the society is divided. It must also be possible to state exactly what the size of each segment is, that is, how many belong to each of the segments. In a completely plural society, there must be perfect correspondence between segmental boundaries and the boundaries between the political, social, and economic organizations. Political parties are one type of organization covered by the third criterion. The final test of a completely plural society is that, since party and segmental loyalties should coincide, there should be little or no change in the voting support of the different parties from election to election: in a perfectly plural society, an election is a segmental census. According to Lijphart, these criteria express only but an ideal type. No society fulfils all the above criteria and the degree of their deviation is indicative of their plurality (Ibid. 356). In practice though, operationalisation and measurement of pluralism in accordance to the aforementioned list looks troublesome. In fact, communal groups may be permeated by other religious, regional or linguistic sub-divisions. Lijphart, for instance, refers among others, to the vicissitudes of the secular segment in The Netherlands and its internal fragmentation into a Socialist and a Liberal strand (Ibid.). In addition, exact knowledge of a segment’s size is improbable. The plethora of identities an individual possesses notwithstanding, censuses are sometimes controversial if not threatening to stability in post-conflict settings. In Lebanon, a census was not held since 1932 because of considerations over the demographic balance of power (Faour 2007). Hence, the safest way of diagnosing a society’s degree of plurality over time becomes the electoral process, which, as a consequence of ethnic voting, reflects to a large extent the electorate’s preferences. Nevertheless, abstention from the electoral process can blur the real degree of pluralism. Jürg Steiner (1987), on other hand, shifts the unit of analysis from the societal level -at least for Switzerland- to the decision-making process in order to measure the degree of segmentation. According to Steiner, decision-making is the ideal locus to measure plurality as different cleavages play out differentially on specific issue-matters. Therefore, he identifies three behavioural patterns within the executive: unanimity, decisions by interpretation, and no decision at all (1987, 366). Yet, if information about the decision-making- both as process and output- might abound for Switzerland, this may not be the case for a number of developing countries. Still, Lijphart correctly points out that the aggregation of decision-making modes in different issue areas with decisions made by individual decision-makers would be difficult to translate into broader patterns that characterize entire political systems (1981, 359). In Power-sharing in South Africa (1985) Lijphart proposes an alternative classification which seems as the Gordian knot between precise measurement and total absence of classification. According to it, the universe of plural societies could be divided into three categories that of low, medium and high pluralism (1985, 130). It must be noted that the latter classification appears in his monograph on South Africa, which represents the first detailed effort on prescriptively universalising the model (Taylor 2009, 4). Yet, Lijphart falls short to explain how this classification emerges. Lijphart merely admits the insurmountable difficulties entailed in the task whilst reiterating the relevance of the four criteria (2002, 17). Therefore, scholars of comparative politics should skirt the problem of classification by intuitively following their personal judgment or what he has previously called “impressionistic judgment” (1981, 357). The intuitive impulse is presumably enhanced by the visible antithesis between plural and homogenous societies. The problems of measurement notwithstanding, recent works on power-sharing by consociationalists and non-consociationalists alike (e.g. Horowitz, 2000, Reilly 2001, Lijphart 2002 McGarry and McEvoy 2013, McCulloch 2014) tend to opt for the eminently qualitatively-oriented concept of severely or deeply divided societies. As deeply divided societies are understood those milieus where ethnic-group identities have a high degree of salience, exceeding that accorded to alternative identities (class, religion, region), whilst levels of antipathy between ethnic groups are high (Horowitz 2002, 18). This preference reflects also Lijphart’s theoretical shift from explanation to prescription but also the parallel geographical expansion of consociationalism from Europe to non-European societies (McGarry and Noel 1989, 4-5). Thus, Horowitz admits in The Architecture of Democracy (2002) that an encounter with consociational democracy and Arend Lijphart seems inevitable especially when grappling with conflict management and constitutional engineering in African and Asian societies (2002, 19). Horowitz represents perhaps the most consistent critic of the prescriptive validity of consociationalism, championing a majoritarian/centripetalist incentives-based approach for post-conflict societies (1991, 2000a, 2000b, 2008). The main centripetalist objections focus on the following points: firstly, the consociational approach is motivationally inadequate as it does not explain why statesmanship and good will should work under the painful strains of post-conflict recovery between inter-segmental elites of different size. Secondly, there are no guarantees that in the medium term, segmental elites will exhibit altruism by sacrificing the interests of their communal clientele. Thirdly, outflanking is a very credible scenario as extremists will seek to undermine compromising segmental elites. Fourthly, cultural autonomy and hence equality, are the products of conflict regulation and not eminently inherent in the prescriptions of consociationalism. Finally, PR ushers in post-electoral coalitions on the allocation of offices but not to pre-electoral ones which could be formed on the basis of mutually compromising ideological and political agendas (Horowitz 2000, 568-576, 2002, 20-22). Instead, centripetalists propose a different institutional arrangement that induces elites to cooperate by introducing built-in mechanisms within the cabinet. The aim is not consensus which is regarded as an extremely ambitious target, but accommodation. As Reilly notes, centripetalism seeks to encourage three distinct phenomena in divided societies i) electoral incentives that enhance inter-group coalitions ii) arenas of bargaining, that aim at vote-pooling and the reciprocal setting of agendas, iii) centrist political parties or coalitions (2001, 11). More concretely, centripetalism advocates preferential electoral laws of the majoritarian type such as the Alternative Vote (AV), which tend to reinforce vote-pooling among a society’s segments through the ordinal ranking of candidates until the absolute majority of preferences is reached (Ibid. 16). Yet, for vote-pooling to materialise, territorial engineering is required that cuts the state across into ethnically heterogeneous federal units. The inter-segmental moderation is further articulated into the election of a president through a system of geographically based electoral thresholds e.g. presidential elections in Nigeria require that the winning candidate garners plurality of support on the country level coupled with geographically distributed thresholds of 25% in 12 out 19 federal states (McCulloch 2014, 95). Centripetal claims are indeed plausible and do deserve thorough examination by constitutional engineers. Nevertheless, centripetalism faces a number of impediments too. First of all, the incentives-based approach lacks a reasonable pool of cases (Nigeria, Sri Lanka, Fiji Papua-New Guinea). Hence, generalisations based on the centripetalist strand of conflict management seem premature. The danger of generalisation and universal applicability is a feature that centripetalists share in common with consociationalism. Secondly, centripetalism suffers from some flawed assumptions. Specifically, centripetalists deem possible that a moderate centre will exist within a divided society, which under the influence of the AV, will attract similarly minded political forces. Ben Reilly argues that under the AV and its reliance on the electorate as the engine of moderation in divided societies, centripetalism may be particularly suitable for societies in which a moderate voice exists (2001, 178). To support his argument, Reilly cites the Good Friday agreement as an example of successful centripetal venture. Yet, it is extremely doubtful that moderation can prevail among the electorate especially in periods that follow deep animosity and ruthless violence. Therefore, timing is of crucial importance (Horowitz 2000b) as conditions might not be ripe for the endorsement of cooperative strategies. Moreover, moderation is always vulnerable to potential spoilers or outflanking from actors that are simply disinterested in participating in democratic procedures. Rabushka and Shepsle have criticised consociationalism on the same grounds, as outbidding tends to dissolve the elite cartel (Esman 2000, 101). Thus, the provision of incentives does not suffice to counter centrifugal tendencies. In Sri Lanka, the Tamils were doubly mistreated: firstly, by the unfair institutional framework that favoured the majority (the Sinhalese), and secondly, by the Tamil Tigers who intimidated any accommodation-seeking parties. On a similar note, the assumption that the electorate will be seeking moderation under preferential voting should be heeded with a grain of salt as it depends more on processes of social mobilisation and structured elite predominance. Finally, the Good Friday Agreement contains consociational rather than majoritarian/centripetalist features (Horowitz 2001, 89) as reflected in rights of cultural autonomy, concurrent majorities, a grand coalition and proportional voting. In addition, the centripetalist strategy of heterogeneous federal units elicits questions on districting. Centripetalists in fact, do not explain how political elites can mutually engage in the drawing of internal borders. To engage in gerrymandering in post-conflict settings requires either a pre-existing level of accommodation between the contending elites or the mediating/arbitrating role of external actors. Centripetalists posit Nigeria as a shining example of federalism, reflected in the demarcation of ethnically mixed entities (Horowitz 2000a, 2000b, Reilly 2001). In fact, it was the military that decisively intervened towards this direction (McCulloch 2014). More recent studies on conflict management try to prudently overcome the intellectual clash of models (Sisk 2013, McCulloch 2014) by laying emphasis on the subtleties that pertain to contextual factors. Conflict management is an inherently policy-oriented field, which calls for intellectual clarity and scientific knowledge. Besides, hybridity, eclecticism and partial adoption characterises the empirical world of conflict regulation as divergent preferences, trade-offs and negotiations preclude the sheer adoption of ready-made solutions. Timothy Sisk observes that “today the issue of short- versus long-term approaches to institutional design in protracted social conflicts …is a critical question for international mediators and as well as protagonists in conflict” (2013, 15). By temporally distinguishing between different phases along the post-conflict time scale, Sisk notes that, whereas consociationalism is the preferred conflict resolution approach in the short-term, centripetal devices shall be progressively introduced as more prone to long-term stability (Ibid 15-16). On the other hand, Allison McCulloch points to the “where-and when” of conflict management. She reaches the conclusion that the ability of power-sharing institutional designs to facilitate political stability depends on the nature of divisions in terms of intensity and demography. (2014, 143-144). A ‘consociational’ universe might be a misnomer. Yet, the prescriptive toolkit of consociationalism remains relevant albeit in its diverse manifestations. Therefore, in-depth analysis of particular case studies might prove of increasing importance to produce knowledge on an array of questions: firstly, why consociational settlements emerge and if possible, how they are sustained, which consociational devices are chosen, and how does the institutional framework constrain actors’ preferences. Sadly, intellectual controversies focus extensively on the potentialities of respective power-sharing models ignoring the subtleties embedded within divided settings. Delving into history might prove fruitful to understand the peculiarities, and capture the complexities endemic to divided polities.