Resumé

The article discusses the validity of consociational theory, an influential trend in political science research in the 1970'5. The theory, as formulated by Arend Lijphart, both seeks to explain the conditions for political stability in fragmented societies, and to present a model for democratic government in such societies. The first part of the article gives an interpretation of the theory. The second part presents the main critics of the consociational paradigm. The article concludes that the theoretical arguments against consociationalism carry considerable weight, but the normative challenge of consociational theory remains when confronted with the quest for conflict management in deeply divided societies.

The Field of Research

Analyzing the conditions for achieving and maintaining stable democratic government is a basic concern in contemporary political science. Since World War II a wide variety of approaches, methods, and concpts have been applied in the search for a deeper understanding of this field of inquiry: What are the basic characteristics of democracy? How do we explain that democracy has. taken roots in some societies but not in others? What is the explanatory relationship between socio-economic development and democratic government? Is democracy the outcome of certain nation-building processes? How does elite control versus mass participation vouch for democratic government?l

The purpose of this article is to discuss the basic assumptions of a trend in this research, the so-called theory of consociational democracy (or shortly consociational theory), '... that has been among the most influential contributions to comparative politics the last decade' (Powell 1979). This approach focuses on socio-cultural divisions and modes of elite co-operation as the explanatory variables accounting for the actual or potential existence of stable democracy in what is called 'plural' (or 'segmented', 'fragmented') societies.

---

* I would like to thank Brian Barry, Bernt Hagtvet, and Øyvind Østerud for comments and encouraging discussions in 1982 during the preparation of this article.

The consociational theory tries to answer the question of how to manage inter-group conflicts within the framework of democratic institutions, given a deeply divided society — that is: given extremely unfavourable conditions for a stable democratic government. On a very general level the answer is implied in the definition of 'consociational democracy' as formulated by Arend Lijpart (1969, 216), usually regarded as the 'founding father' of the theory:

Consociational democracy means government by elite cartel designed to turn a democracy
with a fragmented political culture into a stable democracy.2

Thus, the term 'consociational democracy' not only encapsulates a theory but also a programme: it suggests the imposition in plural societies of a certain kind of government based on the principle of elite accomodation as an intermediate stage in the political development towards stable democracy. The theoretical implications of this suggestion, as spelled out in detail by Arend Lijphart (1977), is clearly of relevance not only for democratic theory but also for theories of political development and conflict resolution. Before we proceed with a discussion of the theory itself, a few words should be said about the basic concept of 'consociation'.

According to Brian (1975a, 478), ' ... 'consociation' is simply a more or less obsolete word meaning much the same as 'association' ...'. Why then was this most uncommon word introduced in the first place when a common one was adequate? Obviously, the selection of this concept had scholarly reasons. 'Association' has connotations making it equivocal. 'Consociation' was picked probably because it is not commonly used, giving the scholars freedom to define it and restrict its usage.3 Lijphart uses the word exclusively to denote a category of democratic systems. He defines the term 'consociational democracy', not the concept 'consociation'. 'Consociationalism' is the set of principles legitimizing consociational democracy. In Lijphart's terminology, 'consociation' logically implies both an act of elite association (or political accommodation) and the normative rules of the game. Furthermore, the concept as developed by Lijphart is closely connected to the historical experience of certain small West European democracies, notably the Netherlands from 1917 to 1967. As Brian Barry points out (1975a, 477) it is similar to notions like 'appeasement' (in relation to the Nazis) and 'detente' (between East and West in our times). A critical question is therefore whether the theory of consociational democracy is of relevance for societies other than the classical European examples (which also includes Belgium, Austria, and Switzerland).⁴ Much of the debate in the last years among political scientists concerns the empirical applicability and theoretical potentialities of consociationalism. Lijphart (1977, chs. 1 and 5) argues that consociational theory provides an answer to the problems of stability

and conflict management in those Third World countries where socio-cultural fragmentation limits the possibility for competitive democracy. He does not think, however, that non-competitive democracy is necessarily the best conceivable kind of democratic government, but that it represents the only possible option in certain societies, the alternative being no democracy at all.

The categories in Lijphart's typology of democratic regimes (Table 1) are labelled in terms of stability/instability. Lijphart ranks the centrifugal type of democracy (combining a fragmented society with adversarial elite behaviour) as the most unstable (the French Third and Fourth Republics, the Weimar Republic, the Austrian First Republic, the short-lived Spanish Republic of the early 1930'5, and post-war Italy), whereas the centripetal type (homogenous society and adversarial behaviour) and consociational democracy are the most stable (the Anglo-American and Nordic countries, and the Low countries, Austria, and Switzerland, respectively). The depoliticized democracy (combining the stabilizing features of both the centripetal and consociational types) is given an in-between position because it is believed that a democracy without competitive behaviour will degenerate into a democratic Leviathan. In such a system, characterized by bureaucracy and politics of bargaining and compromise among established elites, young people and intellectuals will feel alienated and frustrated. The student revolts in many Western countries in the late 1960's could on the basis of this reasoning be explained as a consequence of the lack of internal conflicts and alternative channels of voice in these countries since World War 11.

As a model for democratic government, consociationalism has been developed in the tradition of the liberal-democratic concept of democracy.s For Lijphart, Gabriel A. Almond's typology of political systems from 1956 is the point of departure and basis for his own typology.6 Although it has been argued with specific reference to Almond's typology, that consociational theory '... presents a formidable challenge to existing typologies of democratic regimes, and to widely held beliefs on the conditions of effective and stable democratic rule' (Daalder 1974, 605), no doubt the theory accepts the basic importance and validity of Almond's explanatory variables and concept of democracy. In fact, Lijphart (1977, 6) says that '... the consociational model

represents a constructive effort to refine and build onto Almond's influential typology'. The challenge to Almond's typology is that it does not explain the political stability of the consociational democracies (combining stability with a fragmented political culture). But this does not necessarily represent a threat to its theoretical foundations. Lijphart treats the consociational countries as 'deviant cases'; cases that must be explained in spite of the general validity of Almond's typology. He solves the problem by introducing a new independent variable:

The political stability of the consociational democracies must be explained in terms of an additional factor — cooperation by the leaders of the different groups which transcends the segmental or subcultural cleavages at the mass level — rather than by assigning them to an in-between position on the explanatory variable of political culture (Lijphart 1977, 16).

Characteristics of the Consociational Model

As have already been indicated, the consociational model can be described in terms of three basic variables: Degree of socio-cultural plurality (independent variable); modes of elite behaviour (independent variable); and political stability (dependent variable).

A Plural Society

Democracy is usually associated with the principle of majority decisions within a system of free opposition. This is believed to guarantee political stability because the opposition will raise to power when the majority of the voters no longer accepts the government in office. But such a kind of democracy is only possible in societies with a relatively homogenous political culture — that is: the different cleavages in the soeciety must cross one another, moderating the potential load of conflict inherent in each cleavage. Thus, if a class devision is reinforced by a religious division a more unstable political system is expected than in a society where these two cleavages cross one another. This is a familiar proposition in pluralist theory.7 Reinforcing cleavages create political subcultures, dividing society into what has been termed 'social fragmentation', 'vertical pluralism', 'segmented pluralism', 'verzuiling' (a Dutch metaphor). Thus, a plural society is characterized by cultural fragmentation which tends to dominate political life and represent a continual challenge to national unity. The organized political subcultures resulting from and expressing such fragmentation are often given country-specific terms such as the Dutch zuilen ('pillars'), the Austrian Lager ('camps'), and the Bel^n families spirituelles.

In Lijphart's typology of democracies, political culture represents a dichotomous
variable (plural/non-plural). The 'plural societies' constitute the possible
universe of cases to which the consociational theory should apply according to

its own assumptions. But there is the problem, as Jiirg Steiner (1981) has made convincingly clear, of arriving at a distinct definition of this universe; of drawing the line of demarcation between societies that should be included in the theory and those which should be left out. Brian Barry, for example, argues (1975a, 501) that Switzerland — one of the 'classicaF consociational democracies — does not qualify as a plural society after all, because the social cleavages are not clearly represented on the political level (which indicates cross-cutting cleavages). In his reply to Steiner, Lijphart (1981, 356) suggests that the variable be made continuous, measuring the most important dimensions of political culture as degrees of pluralism, ranging from 0 percent pluralism in a completely homogenous society to 100 percent pluralism in the most extreme case of a plural society. He defines 'a completely plural society' as follows:

1. In a completely plural society, it must be possible to identify exactly the segments into
which the society is divided.

2. It must also be possible to state exactly what the size of each segment is, that is, how many
people belong to each of the segments.

3. In a completely plural society, there must be perfect correspondence between segmental boundaries and the boundaries between the political, social, and economic organizations. 4. Political parties are one type of organizations covered by the third criterion. The final test of a completely plural society is that, since party and segmental loyalties should coincide, there chould 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.

But the basic problem remains: How to devise quantitative indices of pluralism for cross-national comparative purposes, and — even if that problem was solved — how to decide what degree of pluralism should qualify a society for inclusion in the universe of cases covered by the consociational theory. Lijphart admits these problems (at least the problem of quantification). Lijphart seems to have no real solutions to suggest. In his latest response to this critique, Lijphart (1981, 357) states that 'if we want to treat the degree of societal pluralism as a variable at the macro-level, it is better to make well-informed impressionistic judgements based largely on qualitative data than to make no judgment at all.'

In my opinion, the difficulties of arriving at a clear definition of plural society — leaving the exact empirical universe of consociational theory open for varying interpretations — is not a problem threatening the validity of the theory as such. It is rather a challenge for the scholars of the consociational school to agree on a Popperian convention determining the criteria for the empirical applicability of consociational theory.

Coalescent Elite Behaviour

The distinguishing variable of consociational theory is the prevalent pattern of
elite accommodation in a country. This represents the second dimension in

Lijphart's typology. Only a coalescent kind of elite behaviour makes a plural society a consociational democracy, while an adversarial behaviour creates a 'centrifugal democracy'. The underlying assumption is that deeply divided societies can be stabilized by a conscious effort on the part of political elites. Provided there is a genuine feeling among the elites of the necessity to keep the country together, they can deliberately seek to counteract the immobilizing and destabilizing effects of cultural fragmentation.

Consociational democracy can be defined in terms of 'rules of the game' or modes of political decision-making. According to Lijphart, this can be expressed in the form of four basic 'consociational devices': Grand coalition, mutual veto, proportionality, and segmental autonomy:

(1) Grand coalition is the most important of the consociational principles. Elites in plural societies must deviate from competitive practices of political decision-making on the national level (as is the accepted norm of Western democracies), and regulate their inter-communal relations by forming some kind of elite cartel. This can assume different forms, based as it must be on the specific conditions in each society. Consociational democracy results from a unique formal or informal 'historical compromise' between the elites of the most salient segments in the plural society.

The essential characteristic of the grand coalition is not so much any particular institutional
arrangement as the participation by the leaders of all significant segments in governing a
plural society (Lijphart 1977, 31).

Of the four countries most frequently discussed in the consociationalist literature, only Austria and Switzerland offer examples of grand coalitions institutionalized in the national executive. The Dutch zuilen and the Belgian families spirituelles have never joined in grand coalition cabinets, but by and large both countries have been ruled by shifting party coalitions larger than the minimum winning size. And complementing these coalitions, the principle of grand coalition has been implemented on other levels in the decision-making system. Influential councils and committees (like the Dutch Social and Economic Council) have been set up as devices for power-sharing, thus de-emphasizing the role of the cabinet as the pivotal forum of national politics. Minorities and opposition groups will more easily accept not to take part in the national executive, provided there are alternative channels available for influencing the outcomes of the political processes. This mechanism is also illustrated in a wellknown article by Stein Rokkan covering the case of Norway (Rokkan 1966).

(2) Mutual veto, proportionality, and segmental autonomy. These three instruments of consociationalism (all of them deviating from the majority principle) are secondary and complementary in relation to the grand coalition. They are applied in order to secure basic minority interests and a system of real

powersharing between the segments. Mutual veto means that no important decision can be made against the expressed opposition of one segment. Proportionality is a method of allocating civil service jobs and financial resources among the different segments. Segmental autonomy means 'rule by the minority over itself in the area of the minority's exclusive concern' (Lijphart 1977, 41).

Such autonomy is possible only if each segment has a well organized internal structure and established procedures for decision-making. Federalism can be seen as a special form of segmental autonomy in a society where the main cleavages are territorial. The important thing is to secure border maintenance between each segment, allowing a policy for 'good political neighbourhood'. On the other hand, this way of organizing decision procedures in the society tends to reinforce the degree and intensity of segmentation and motivate political mobilisation on a mass level in each segment. Commenting on this, Hans Daalder (1974, 607) says:

Consociational democracy ... tends to show a curious mixture of ideological intransigence on the one hand and pragmatic political bargaining on the other. Separatism makes for a dogmatic, expressive style of politics within ideological families. But relations among subcultures are settled by a process of careful and businesslike adjustments.

Brian Barry (1975b) has warned that the consociational model will have a reasonable chance of success only in plural societies with a low degree of conflict potentialities on the intersegmental level. If the relations between the segments are strongly conflict laden, the consociational model will rather promote conditions for civil war than for political stability. A similar way of reasoning is reflected by Eric A. Nordlinger — a proponant of consociationalism — when he explicitly excludes the dimension of segmental autonomy from his general model of conflict management in deeply divided societies (Nordlinger 1972).

This raises an interesting question also discussed by Hans Daalder: What is the specific relationship between the nature of cleavages in a society and political conflicts in that society? Daalder (1974, 612-614) criticizes most of the consociationalist scholars for being too preoccupied with the mechanic structure of divisions (how many cleavages, how large are they, which kind of cleavagecoalitions could form what kinds of government, etc.) rather than qualitative aspects of the cleavage-formation (class, religion, language, ethnicity, etc.).

What distinguishes the kind of political conflicts resulting from class cleavage and ethnic cleavage respectively? Are all social cleavages necessarily transformed into political tensions? If not, why are some cleavages more salient in this respect than others? If a consociational kind of government is established in a society divided along ethnic lines, would not this be a starting point for the eventual

partition of the country?B When the divisions are of a class or religious nature, the basic question confronting the segments of the society is: How should we, the people of this country, manage our internal problems? Barry (1975a, 503) continues:

Naturally, a question so posed lends itself to consociational methods of resolution. But where the basis of division is ethnic the question may not be how the country is to be run but whether it should be a country at all. Should one group secede and become an independent state? Should they merge with an ethnically-cognate neighbour state? Or should the whole of the existing state absorb or be absorbed by another? These are questions which allow little room for negotiation because at least one party may not want a solution within the existing framework.

But again, this critique is not necessarily a threat to the consociational theory
as such, but an important argument for restricting its applicability.

Political Stability

Stability is the phenomenon to be explained and the aim to be achieved within the consociationalist framework. In the literature on consociational democracy, stability is usually defined in terms of societal attributes. A democratic system is stable given the presence of indicators like legitimacy, flexibility, governmental longevity, constitutional continuity, and effective decision-making, and unstable given the degree of violence and existence of political movements with revolutionary aims.₉

In other words, it can be argued that 'political stability' is a value-laden term, an expression of the Western liberal-democratic Utopia developed during the 1950's under the headline The End of Ideology'.lo The political stability of the Western democracies, especially the United States and Great Britain, was seen as resting on a basic harmony enveloping the society: a 'positive peace' crosscutting both the horizontal (socio-economic) and vertical (cultural) cleavages (Galtung 1969). In such societies rational and effective decision-making by competent and representative elites is possible, leaving 'ideological politics' and mass mobilization behind as vestiges from a more conflict-ridden stage of development. This idealized view of the Western societies having reached the final stages towards 'real' democracy was badly shaken, however, in the late 1960's and early 1970's with the widespread unrest in most of these countries — upheavals which more than anything else were ideological in nature. The new critics maintained that the established theory of Western democracy in itself was 'ideological', distorting the social facts and legitimizing the existing powerrelations is society. The actual political stability in the Western world after World War II was rather explained in terms of economic growth and control of the masses through ideological hegemony and repression by the ruling elites.¹¹ In other words, this is a conception of political stability stripped of its positive connotations and described as some kind of 'negative peace'.

This focus on control in explaining political stability is also reflected in an article by lan Lustick (1979). He argues that the consociational theory is insufficient for the purpose of analyzing political stability in deeply divided societies. By excluding the control perspective, the consociational scholars, according to Lustick, ignore the possibility that stability might result from coercion and repression of subordinate groups in society. Like Brian Barry, he criticizes the consociational scholars for disregarding the analytical implications of 'the anti-democratic manipulative nature of many consociational "techniques"' (Lustick 1979, 330). While the proponents of consociationalism are preoccupied with the puzzle of how to reach institutional arrangements for facilitating elite agreement, Lustick directs our attention to the repressive measures necessary to keep the carefully balanced system together. This also provides us with an analytical approach for evaluating the degree of 'structural violence' in the political system. In my opinion, such evaluation is especially important in the case of consociationalism because it explicitly .deviates from traditional democratic practices. Under which conditions are such deviations justified? It is not difficult to agree with Lustick that an empirical and conceptual analysis of systems of control in deeply divided societies is essential for a comprehensive understanding and evaluation of politics in these societies. Given unjust systems of control, such analysis makes it possible, says Lustick (1979, 344), 'to identify typical weaknesses, patterns of breakdown, and appropriate strategies for resistance. (It) thus represents a necessary part of any struggle to dismantle systems of control that now exist or those that may emerge in the future.'

Critics of the Consociational Theory

We have already pointed out some difficulties inherent in the conceptual framework of consociational theory. Part of these problems can be discussed and solved without questioning the basic assumptions of the theory, such as the problem of arriving at clear and operationalizable definitions. But the theory has also been met with more serious challenges. What especially has been questioned is the relevance of elite behaviour as an independent variable accounting for political stability.

Basically, this touches the question of the theoretical status of political
science as such. As Brian Barry (1975a, 494) puts it:

In a broader setting the 'consociational democracy' thesis can be seen as part of the movement among political scientists in recent years towards a reassertion of politics as the 'master science' in reaction to the socio-economic reductionism implicit in the explanatory claims to political sociology.

Of significant interest in our context is how the consociational theory explains
the transformation of conflict immanent in the socio-cultural and socioeconomic

structure in society to the level of political behaviour. Why does the theory only concentrate on the political elites? Are the elites free to choose the course of their own political behaviour? Why do they behave coalescent in some instances and adversarial in others?

The focus on elites can be explained in terms of Lijphart's 'self-denying hypothesis'; outbursts of potential vulcaneous conflicts in vertical fragmented societies can be prevented only by acts of cooperation by elites representing the conflicting interests. This is what keeps the nation together: an institutionalized superstructure constituting the arena for formal and informal bargaining between segmented elites, based on a fundamental willingness of the actors to follow the rules of the game. Vertical segmentation means that each segment is clearly insulated from one another, with its own school system, cultural institutions, and political organizations. In a completely segmented society, there is practically no horizontal interaction between the segments at the mass level. Now, Lijphart, Nordlinger, and others are not neglecting the structural and ideological surroundings of the elite's political playground. On the contrary, a strong emphasis is put on the conditions for consociational practices: a multiple balance of power among the segments; small size of the country involved; overarching loyalties — both within segments (for example religion) and across segments (nationalism); segmental isolation; prior traditions of elite accommodation; a relatively low load on decision-making processes; political and economic stakes of all elite groups in the political functioning of the system; a low degree of mobilization and/or a high degree of encapsulation of the masses; and an awareness of clear and present dangers to the system as a whole.

What remains unclear, however, is the nature of the accommodative behaviour of the elites. Is it based on a free choice or is it socio-structurally determined, leaving the elites with practically no other choice than accommodation — given the totality of the situation?¹² Hans Daalder (1974, 613) remarks:

Paradoxically, this trend of thought combines a somewhat deterministic belief in the explosive
potential of social cleavages and in the ability of autonomous elite groups to thwart their
effect.

Lijphart seems to think that the elites are free to choose in the sense that different courses of actions seem open to them. The situation is not unequivocal; alternative strategies must be analyzed. Furthermore, Lijphart concentrates on critical junctures in the history of the society under consideration, when intersegmental conflicts have reached a level which threatens the system as such: that is the moment for the elites to choose between constructive or destructive behaviour; that is their chance to agree on a formula for conflict management and decision-making as a basis for a future stable democratic government.

The Nation-Building Perspective

Stein Rokkan, Hans Daalder, and other nation-building theorists are more concerned with the multidimensional historical process of cleavage formation in accounting for the emergence of different kinds of elite behaviour and institutional arrangements.

Rokkan committed most of his efforts as a social scientist to construct a functional-dynamic 'conceptual map' of Europe, consisting not only of a set of basic dimensions and variables explaining political development, but also defining the exact relationship between these dimensions and variables at the different stages of development. He thus analysed centre- and state-formation processes and modes of political mobilization as conditioned by geo-political location combined with the emergence of economic, cultural, and social organizations and networks (Rokkan 1975, 1980). Relating to the three consociational systems of the Netherlands, Switzerland, and Belgium, he says (1973, 19):

It was no accident that the three consociational systems had to be developed along the broad 'buffer zone' between northern Protestantism and southern counter-Reformation in Europe. Nor was it an accident that they had emerged out of the strong networks of independent cities at the margins of the large continental systems of territorial control. Consociation represented a constructive solution to an inherited configuration of cultural and economic forces: a stronger drive towards administrative centralization. A greater investment of resources in efforts of cultural unification might have disrupted the balance in each system.

In his article on verzuiling from 1977, Rokkan more specifically developed a typology of organizational responses to cleavage crystallization in its different stages. He distinguishes between responses through an electoral and a corporate channel, and defines verzuiling as 'the degree of interlocking between cleavagespecific organizations active in the corporate channel and party organizations mobilizing for electoral action' (p. 565).

Rokkan's perspective provides the theoretical basis for Hans Daalder's critique of the consociational theory. Daalder (1974, 609) critizes the theory because 'it does not answer the vitally important question why and how such consociational systems developed.'° There has been some controversy between Daalder and Lijphart over the interpretation of one specific historical experience, the Dutch Pacification of 1917 — a prototype of consociational behaviour in Lijphart's works. Daalder claims that the Pacification was made possible as a consequence of specific traditions of elite accommodation developed in the Dutch polity since the 16th century, and was in many ways a prelude to the Dutch verzuiling, rather than a response to the challenge of the threatening segmentation (as is Lijphart's interpretation). Daalder (1974, 616) elaborates:

Only after 1917 did the various Dutch groups develop their strong network of subcultural

interest organizations. Although such organizations were conducive towards isolating Dutchmen of different ideological origins, none of these segments ever was an effective threat to the Dutch state as such. Consociationalism in the Netherlands should therefore not be regarded as a response to the perils of subcultural splits (as Lijphart argues), but as the underlying reason why slowly developing subcultural divisions never did become perilous when social modernization led to mass mobilization.

For his own part, Lijphart does not see a contradiction between himself and Daalder. Rather, he adds a historical dimension to his model, making prior accommodative behaviour one of many 'favourable conditions' for consociational democracy:

A predemocratic historical tendency toward moderation and compromise can indeed be an independent factor that can appreciably strengthen the chances of consociational democracy. Switzerland and the Netherlands never experienced any substantial period of absolutism, and pre-absolutist traditions of diffuse and dispersed power relationships remained strong. Straight majority rule — the democratic equivalent of royal absolutism — therefore never had a strong appeal (Lijphart 1977, 100).

But Lijphart concludes the discussion with his critics on this point by reiterating his stress on 'the voluntary, rational, purposive, and contractual elements of consociational democracy'. To him the consociational model as a normative model 'presupposes that consociationalism is an example that can be freely and deliberately followed.' This is the presupposition on which Lijphart's concept of 'consociational engineering' rests: 'that leaders in plural societies are capable of creative and constructive acts of free will.' And furthermore:

This book's message to the political leaders of plural societies is to encourage them to
engage in a form of political engineering: if they wish to establish or strengthen democratic
institutions in their countries, they must become consociational engineers (Lijphart 1977, 223).

In other words, what most clearly distinguishes Lijphart's consociational theory from the nation-building theories is this programmatic aspect of his model. This is probably also a reason why he so unequivocally rejects the assertion, most clearly put forward by nation-building theorists, that 'consociational democracy' as a theoretical construction is indissolubly related to specific historical experiences of certain smaller European countries. He wants to make his consociational model universal in space and time, freeing it from a theoretical landlock within some kind of 'conceptual map' of Europe.

The Marxist Interpretation

The most critical statement on the consociational theory has come, not
surprisingly, from a Marxist point of view. This position is made clear by
Ronald A. Kieve in an article where he discusses Lijphart's analysis of the

political development in Holland. Kieve's main argument is that the politics resulting from the socio-cultural divisions in the society can be understood in its right context only if related to 'those divisions that are most basic to modern societies, that is, class conflict' (Kieve 1981, 314). The 'free will' of the group elites can not be attributed any independent explanatory status. With references to historical cases, Kieve argues that in times of politicized class conflict 'the second-order divisions (ethnic, religious, cultural, racial, and linguistic cleavages) are superseded precisely because the specific political form that they take is determined by the class structure of that society.' (p. 314).

In short, Kieve introduces class structure as the basic explanatory variable determining the relationships between all other variables in the system: Class relations determine both how socio-cultural structures are reflected on the level of elite behaviour and how socio-cultural structures and elite behaviour effect political stability. Thus, within this conceptual framework, class relations are likely to be regarded as a basic mediating force in society.l4 This means thai Kieve's explanatory model is much more complicated than Lijphart's because class structure is singled out as an explanatory level on its own. The problem is to define exactly how class structure determines relationships on other levels in the Marxist explanatory model.

With an obvious risk of oversimplifying, the difference between a consociational
and a Marxist model explaining political stability in plural societies
can be represented as in Figure 1.

Kieve sums up his critique of the consociational model as follows (p. 322):

The accommodationalist model, by stressing the procedural element of elite interaction,
neglects the social determinants of these procedures. The model mistakes the cause for its

effect: it assumes that the cause of political stability or social peace is the internatilization by the relevant elites of accommodationalist attitudes, rather than that the acceptance at any given time of accomodationalist procedures is the outcome of social and political conflict that reflects significant changes in the balance of class forces. Thus, the proponents of the model seize upon the results of conflict — which might take the form of compromise on particular issues — and assume that very result to be the mechanism or motive force of political relations.

In the Marxist interpretation, the political struggle between the leaderships of the most salient zuilen in modern societies is a struggle for the strongest possible position in the system, i.e. a strongest possible influence over state apparatus, constitutional organs, and resource allocations. This struggle is not one that threatens the basic societal order of the system as such (that is to say, the capitalist mode of production).

Rather, for each elite group it is a question of conquering positions within the system, securing both segmental interests and a power-base for a continuous elite position within one's own segment. This interpretation is supported by the fact (also evident from Lijphart's work)¹⁵ that consociationalism tends to reinforce the segmentation of society. This means that as a superstructure of cooperative institutions is established across the segmental cleavages on the elite level, cross-cutting interaction on the mass level declines. Following consociational arrangements, each segment in the society will usually be more thoroughly organized with its own cultural, social, and economic institutions. Consequently, while the elite groups in the society (in the Dutch case, a capitalist society) can accommodate their interests qua elites, the masses are split in different zuilen with limited possibilities of coordinating their basic interests qua masses. Boundary maintenance between segments is thus not only a prerequisite for consociationalism, it also serves the interests of the ruling classes, thus exacerbating class conflicts in the society. An example mentioned by Kieve is that the powerful Dutch Catholic hierarchy through most of the consociational period of the country banned Catholic workers from cooperating with workers from other sections of society, especially secular socialists (Kieve 1981, 318-319). According to Kieve, then, consociationalism in a plural capitalist society functions as a way of arriving at orderly procedures for solving political problems within the ruling classes and as a system for dividing and ruling the working class.

Consociationalism: Against 'the Laws of History'?

Many nation-building theorists, Marxists, and integration theorists would agree on a meta-theoretical point of critique: consociationalism is against the nature of history — it is in effect a programme for institutionalizing, or 'freezing', conflicts in the society which result from social divisions that are continuously changing, creating new structural configurations. Modern society is inevitably

moving towards greater unity, due to the socio-economic integration and the rapid development of communications.¹⁶ Keeping the historically determined socio-cultural divisions artificially alive by building them in as part of the very fabric of societal institutions will necessarily create instability. The carefully balanced consociational institutions will not be easily changed because there will always be one party preventing such changes by use of veto. Therefore, relatively soon after their establishment, the consociational institutions no longer reflect social realities adequately, and thus in themselves create tensions. Lebanon (surprisingly described by Lijphart (1977, 5) in 1977 as a 'relatively successful example of consociational democracy in the Third World') can serve as an example. The institutionalized majority position in the constitutional organs of the declining Christian section of the population is part of the picture that explains the outbreak of the civil war in 1975. In fact, a census has not been conducted officially in the country since 1932 because the Christians know that it would reveal a Muslim majority and thus jeopardize the legitimacy of their own constitutional privileges.

In other words, the core of this argument is that the building of institutions should follow 'the logic of history', not preserve structures that are historically doomed. Modernization without social integration is practically unthinkable. If the configuration of societal forces at a given time in history creates tensions that threatens the system as such, it might be better in the long run that the old system is allowed to break down (revolution, secession, partition, etc.) as a basis for reordering the social system, than to prolong the life of a sick organism.

Conclusion

The arguments against the consociational theory undoubtedly carry considerable weight. Even if we accept the basic assumptions of the theory, we are confronted with important problems at the conceptual level when applying the model: What is the exact meaning of the key concepts 'plurality', 'stability', 'coalescent behaviour', and — especially if applied outside the Western liberaldemocratic cultural sphere — 'democracy'? On the other hand, problems of definition and operationalization are well known in all theoretical constructions in the social sciences. Escaping conceptual ambiguity at one theoretical level or the other seems almost impossible. If the proponents of the consociationalist approach continue their internal debate and refine their definitions of the key concepts, there is no reason why the consociational theory should be less applicable for empirical research than alternative and more established theories.

The discussion about the validity of the basic assumptions of consociationalism
may in part reflect different Weltanschauungen in political science
research. The question demarcating the positions is this: Can politics be

analyzed as a non-reducible cognitive attribute of political elites; that is, is the political process in a significant way formed by 'the free will' of political actors, or are the directions of political acts more or less determined by historical traditions and class interests, leaving 'the free will' of political leaders with a theoretical status as no more than appearance? As we have seen, the consociationalists are ambiguous but tend to agree with the former position; the nationbuilding theorists concentrate on historical determinants (the latter position) but accept the salience of purposive elite behaviour in forming political coalitions and creating conditions for national integration; the Marxists are clearly defining politics as a secondary phenomenon in relation to socioeconomic structures and class relationships (the latter position).

It seems reasonable to conclude, after having reviewed a considerable part of the literature on consociational democracy, that the main problem confronting the proponents of the consociational model is connected to the theoretical claim that the model could be attributed universal validity in explaining the existence or non-existence of political stability in plural societies. Sticking to the two basic independent variables, socio-cultural plurality and modes of elite behaviour do not seem to be a sufficient answer to the critics of the model. On the other hand, a universal concept of 'consociational engineering' is exactly what makes the consociational theory interesting and provoking. Whatever the theoretical critique might say of the explanatory power of the model, the normative challenge remains: That human beings, as rational and calculating beings, should be able to manage conflicts between themselves in spite of unfavourable historical and socio-structural conditions.

This emphasis on rationality and choice means that the consociational theory is probably of greater relevance for general conflict theory and game theory than for theories of political development and stability. More specifically, consociational theory relates to a situation with two or more collective actors that cannot be integrated (for one reason or another) to form one unified actor, but who realize (for one reason or another) that they have to coexist within and delegate some sovereignty to a common institutional framework. The consociational theory underlines that successful conflict management in such situations is dependent on a mutual respect for the autonomy and integrity of each actor and on deviations from the majority principle in the decision-making procedures. This in itself is in my opinion a major contribution. Reformulated in such terms, the theoretical contributions of the consociationalist scholars could be of relevance for the analysis of possible procedures for coexistence in such deeply divided countries as Palestine (Palestinian Arabs and Israeli Jews) and South Africa (blacks and whites).

---

NOTES

1. Highlights of the literature on democracy with relevance for this paper, are: Kariel 1970, Almond & Powell 1966, Dahl 1966, Tilly 1975, Moore 1967.

2. First presented at the 1967 World Congress of the International Political Science Association in Brussels, and first printed in World Politics, vol. XXI, January 1969, p. 207-225. Independently of Lijphart, Gerhard Lehmbruch presented his 'A Non-Competitive Pattern of Conflict Management in Liberal Democracies: The Case of Switzerland, Austria, and Lebanon', printed in Mcßae 1974.

3. In modern social science, the word 'consociational' was first applied by David Apter (1961).

4. Studies with relevance for consociational theory of these countries are: Lorwin 1966, Bluhm 1973, Steiner 1974.

5. C.B. Macpherson (1967) distinguishes between a liberal-democratic, a communist, and a Third World concept of democracy.

6. Gabriel A. Almond (1956) distinguishes four political systems according to the criteria of political culture (homogeneous/fragmented) and role structure (high degree/low degree of differentiation): Anglo-American (homogeneous, highly differentiated); Continental European (fragmented with role systems embedded in the political subcultures); Preindustrial (fragmented, low degree of differentiation). What basically distinguishes the Continental European from the Anglo-American system (which together constitute the democratic systems in the typology) is differences in political culture. Due to such differences, according to Almond, the Anglo-American systems are more stable than the Continental European (with the Scandinavian and Low countries somewhere in between).

7. See, for example, Lipset 1960, Truman 1951, Bentley 1955.

8. There is a huge literature on ethnicity and politics. Connected to our discussion of consociationalism, the following works are of special interest: Kuper & Smith 1969, Enloe 1973, Katznelson 1972, D. Rothschild 1970, J. Rothschild 1981, Leifer 1981.

9. For different concepts of political stability, see Hurwitz 1973.

10. See the introduction in Kariel 1970.

11. Intellectually, this critique has derived its inspiration to a considerable degree from the Frankfurt School of philosophy, especially from the works of Herbert Macuse, Theodore W. Adorno, and Jiirgen Habermas.

12. Zuckerman (1975) presents an interesting discussion of different approaches t^ this problem.

13. See also Daalder 1971, 1981.

14. Methodologically, this point is derived from Wright 1978, 15-30.

15. See for instance Lijphart 1977, 42.

16. This is an underlying 'optimistic' assumption in most of the studies of integration, both national, regional, and international, in the 1950's and 1960'5. See Deutsch 1953, 1966, Haas 1958, Nettl & Robertson 1968, Levy 1960. R.J. Harrison says that analysis such as those mentioned above 'take twentieth century conditions as an obvious starting point, assume their existence in the explanatory hypotheses, and often appear to derive a personal commitment to integration from a conviction that it is an answer to some of the problems of this century. It is an answer that has, at least, prima facie plausibility' (Harrison 1975, 11).

REFERENCES

(A systematic bibliography on consociationalism appears in Acta Politico, vol. XIX, January
1984).

Almond, G.A. 1956. 'Comparative Political Systems', Journal of Politics, vol. XVIII, no. 3.

Almond, G.A. & Powell, G.B. 1966. Comparative Politics. A Developmental Approach. Boston:
Little, Brown.

Apter D. 1961. The Political Kingdom of Uganda: A Study in Bureaucratic Nationalism.
Princeton: Princeton University Press.

Bentley, A.F. 1955. The Process of Government: A Study of Social Pressures, 4th ed. Evanston:
Principia Press of Illinois.

Barry B, 1975a. 'Review Article: Political Accomodation and Consociational Democracy',
British Journal of Political Science, vol. V., October 1975.

Barry B, 1975b. 'The Consociational Model and Its Dangers', European Journal of Political
Research, vol. 111, no. 4.

Bluhm, W.T. 1973. Building an Austrian Nation: The Political Integration of a Western Nation.
New Haven and London: Yale University Press.

Daalder, H. 1971. 'On Building Consociational Nations: The Cases of the Netherlands and
Switzerland', International Social Science Journal, vol. XXIII, no. 3.

Daalder, H. 1974. 'The Consociational Democracy Theme', World Politics, vol. XXVI, no. 4.

Daalder, H. 1981. 'Consociationalism, Center and Periphery in the Netherlands' in Torsvik,
P., ed., Mobilization, Center-Periphery, Structures, and Nation-Building. Oslo: Universitetsforlaget.

Dahl, R.A., ed., 1966. Political Opposition in Western Democracies. New Haven and London:
Yale University Press.

Deutsch, K.W. (1953) 1966. Nationalism and Social Communication. Cambridge, Mass.:
MIT Press.

Enloe C. 1973. Ethnic Conflict and Political Development. Boston: Little, Brown.

Haas, E.B. 1958. The Uniting of Europe: Political, Social, and Economic Forces 1950-57.
Stanford: Stanford University Press.

Harrison, R.J. 1974. Europe in Question. Theories of Regional International Integration.
London:'George Allen & Unwin.

Hurwitz, L. 1973. 'Contemporary Approaches to Political Stability', Comparative Politics,
vol. V, no. 3.

Kariel, H.S. 1970. Frontiers of Democratic Theory. New York: Random House.

Katznelson, I. 1972. 'Comparative Studies of Race and Ethnicity', Comparative Politics, vol. V.,
October 1972.

Kuper, L. & Smith, M.G. 1969. Pluralism in Africa. Berkeley and Los Angeles: University of
California Press.

Kieve, R.A. 1981. 'Pillars of Sand. A Marxist Critique of Consociational Democracy in the
Netherlands', Comparative Politics, Vol. XIII, April 1981.

Levy, M.J. 1960. The Structure of Society. Princeton: Princeton University Press.

Leifer, E.L. 1981. 'Competing Models of Political Mobilization: The Role of Ethnic Ties',
American Journal of Sociology, vol. LXXXVII, no. 1.

Lijphart, A. 1969. 'Consociational Democracy', World Politics, vol. XXI, January 1969.

Lijphart, A. (ed). 1969. Politics in Europe. Comparisons and Interpretations. Englewood
Cliffs: Prentice-Hall.

Lijphart, A. 1974. 'Consociational Democracy' in Mcßae, K., (ed.) Consociational Democracy
Accomodation in Segmented Societies. Toronto: McClelland & Stewart.

Lijphart, A. 1977. Democracy in Plural Societies: A Comparative Exploration. New Haven
and London: Yale University Press.

Lijphart, A. 1981. 'Consociational Theory. Problems and Prospects', Comparative Politics,
vol. XIII.

Lipset, S.M. 1960. Political Man: The Social Bases of Politics. New York: Doubleday.

Lorwin, V.R. 1971. 'Segmented Pluralism: Ideological Cleavages and Political Cohesion in the
Smaller European Democracies', Comparative Politics, vol. 3, no. 2.

Lustick, I. 1979. 'Stability in Deeply Divided Societies: Consociationalism versus Control', World
Politics, Vol. XXXI, No. 3.

Macpherson, C.B. 1967. The Real World of Democracy. Oxford: Oxford University Press.

Nettl, J.P. & Robertson, R. 1968. International Systems and the Modernization of Societies.
London: Faber & Faber.

Nordlinger, E.A. 1972. Conflict Regulation in Divided Societies. Occasional Papers No. 29.
Cambridge, Mass.: Center for International Affairs, Harvard University.

Powell, G.B. 1979. 'Review of Arend Lijphart: Democracy in Plural Societies', American Political
Science Review, vol. LXXIII, March 1979.

Rokkan, S. 1966. 'Numerical Democracy and Corporate Pluralism' in Dahl, R.A. op.cit.

Rokkan, S. 1973. 'Centre-Formation, Nation-Building, and Cultural Diversity: Report on a
UNESCO Programme', in Eisenstadt, S.N. & Rokkan, S., eds., Building States and Nations.
Vol. I. Beverly Hills and London: Sage Publications.

Rokkan, S. 1975. 'Dimensions of State Formation and Nation-Building: A Possible Paradigm
for Research Variations within Europe', in Tilly 1975.

Rokkan, S. 1977. 'Towards a Generalized Concept of Verzuiling', Political Studies, vol. XXV, no. 4.

Rokkan, S. 1980. 'Territories, Centre, and Peripheries: Toward a Geoethnic-Geoeconomic-Geopolitical
Model of Differentiation within Western Europe' in Gottmann, J., ed., Centre and
Periphery. Spatial Variations in Politics. Beverly Hills and London: Sage Publications.

Rothschild, D. 1970. 'Ethnicity and Conflict Resolution', World Politics, vol. XXII, July 1970.

Rothschuild, J. 1981. Ethnopo/itics. A Conceptual Framework. New York: Columbia University
Press.

Steiner, J. 1974. Amicable Agreement versus Majority Rule: Conflict Resolution in Switzerland.
Chapel Hill: University of North Carolina Press.

Steiner, J. 1981. 'The Consociational Theory and Beyond', Comparative Politics, vol. XIII,
April 1981.

Tilly, C, ed., 1975. The Formation of National States in Western Europe. Princeton: Princeton
University Press.

Truman, D.B. 1951. The Governmental Process: Political Interests and Public Opinion. New
York: Alfred A- Knopf.

Wright, E.O. 1978. Class, Crises, and the State. New York: Schocken Books.

Zuckerman, A. 1975. 'Political Cleavage: A Conceptual and Theoretical Analysis', British
Journal of Political Science, vol. V., No. 2.