citizens elections parties

Transcription

1 Series Editors: Dario Castiglione (University of Exeter) Vincent Hoffmann-Martinot (Sciences Po Bordeaux) citizens elections parties approaches to the comparative study of the processes of development Stein Rokkan with Angus Campbell, Per Torsvik, and Henry Valen

2 Stein Rokkan First published in 1970 by Universitetsforlaget, Oslo First published by the ECPR Press in 2009 The ECPR Press is the publishing imprint of the European Consortium for Political Research (ECPR), a scholarly association, which supports and encourages the training, research and cross-national cooperation of political scientists in institutions throughout Europe and beyond. The ECPR s Central Services are located at the University of Essex, Wivenhoe Park, Colchester, CO4 3SQ, UK All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Typeset by the ECPR Press Printed and bound by Lightning Source British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Paperback ISBN: Hardback ISBN:

3 Other titles in the ECPR Classics series: Identity, Competition and Electoral Availability: The Stabilisation of European Electorates People, States and Fear: An agenda for International Security Studies in the Post-Cold War era Stefano Bartolini and Peter Mair Barry Buzan Elite and Specialized Interviewing Lewis A. Dexter System and Process in International Politics Morton A. Kaplan Democracy Jack Lively Individualism Steven Lukes Political Elites Geraint Parry Parties and Party Systems: A Framework For Analysis Giovanni Sartori Territory and Power in the UK James Bulpitt Beyond the Nation State Ernst Haas Electoral Change: Responses to Evolving Structures in Western Countries Party Identifi cation and Beyond: Representations of Voting and Party Competition The State Tradition in Western Europe: A Study of an Idea and Institution Mark Franklin, Thomas Mackie and Henry Valen Ian Budge, Ivor Crewe and Dennis Farlie Kenneth Dyson Titles in the ECPR Monographs series: The Return of the State of War: A Theoretical Analysis of Dario Battistella Operation Iraqi Freedom Gender and the Vote in Britain: Beyond the Gender Gap? Rosie Campbell Paying for Democracy: Political Finance and State Kevin Casas-Zamora Funding for Parties The Politics of Income Taxation: A Comparative Analysis Steffen Ganghof Joining Political Organisations: Institutions, Laura Morales Mobilisation and Participation in Western Democracies Citizenship: The History of an Idea Paul Magnette Representing Women? Female Legislators in West European Parliaments Deliberation Behind the Closed Doors: Transparency and Lobbying in the European Union Mercedes Mateo Diaz Daniel Naurin Globalisation: An Overview Danilo Zolo Political Confl ict and Political Preferences Claudia Landwehr

4

5 introduction to ecpr classics edition Alan Renwick, University of Oxford 1 Stein Rokkan s origins lay on the periphery of a periphery; yet he became one of the central figures of European comparative politics and political sociology in the post-war decades. He was born in 1921 on the Lofoten archipelago in the far north of Norway and raised in the nearby town of Narvik. From these unlikely beginnings, he went on to become president of the International Political Science Association (IPSA), vice-president of the International Sociological Association (ISA), and founding chairman of the European Consortium for Political Research (ECPR). 2 No less a figure than Seymour Martin Lipset described him shortly after his death in 1979 as the preeminent political sociologist of his generation (Lipset, 1981: xxi). He remains widely cited in diverse literatures today. Rokkan wrote numerous papers, articles, and chapters. 3 Citizens, Elections, Parties (hereinafter CEP) was, however, his only single-authored book, and even it is in fact a collection of papers whose original publication dates range between 1955 and As several authors, including Flora (1999) and Mjøset (2000) have pointed out, 1970, the year in which CEP was published, marked something of a turning point in Rokkan s research. Until that time, his overriding concern was to understand the political behaviour particularly the voting behaviour of citizens. As I shall discuss, the approach he took to exploring this goal had shifted substantially over time, but the core purpose had remained the same. After 1970, Rokkan did not abandon that goal; but he did extend beyond it. He sought to understand more broadly the historical development of European states and nations, developing his model and conceptual map of Europe. That later work has been anthologized in another volume supported by the ECPR, painstakingly compiled by Peter Flora (Rokkan, 1999). There is some overlap between that volume and CEP in the seminal works that Rokkan produced during the late 1960s. Nevertheless, CEP remains the most complete guide to Rokkan s work up to 1970, and, as Flora himself acknowledges (Flora, 1999: 17), it is for this work that Rokkan is most widely known today. CEP thus remains an important source that deserves to be read widely. As I have indicated, the core question at the heart of CEP is that of what explains the political behaviour of citizens. In his earliest work, Rokkan approached this question through survey research that illuminated features of individual citizens that shaped their political behaviour. He quickly became interested, however, primarily in variations in citizen behaviour across countries particularly the countries of Western Europe and in how these differences were generated by structural factors. He focused on two sets of structural factors in particular: insti-

6 ii citizens, elections, parties tutions and cleavages. It is for his work on these structural factors that Rokkan is most famous, and he remains widely cited in the literatures on the sociological origins of voting, the determinants of the choice of electoral systems, the shapers of party systems (particularly the question of whether party systems are frozen ), and the nature and effects of path dependence. My purposes in this introduction are twofold: I present an overview of Rokkan s core thinking as represented in CEP; and I seek to show how reading CEP can enhance our understanding of Rokkan s contributions to the contemporary literatures just cited. Many authors have cited just a small range of Rokkan s writings, but wider reading of the papers in CEP offers a more rounded conception of his perspective. I begin by considering how Rokkan came to focus upon the structural underpinnings of citizen behaviour. The Shaping of Citizens Behaviour: From Micro to Macro Factors A person familiar only with Rokkan s later work might be surprised to discover that, during his earliest professional years, he was a leading exponent of research into political behaviour based on sample surveys. The paper reproduced here as Chapter 12 was, indeed, co-authored with one of the pioneers of survey research in the United States, Angus Campbell, while Chapter 14 betrays the strong influence of Campbell s great rival, Paul Lazarsfeld. Chapter 11 reports findings from national surveys conducted before and after the Norwegian elections of From the earliest stages of his career, Rokkan was particularly interested in comparative research: Chapter 12, written with Campbell, compares the Norwegian survey data with findings from similar studies of the American elections of 1956; and the earliest papers in the book Chapters 8 and 9 report the findings of a sevencountry study of the political attitudes of teachers. What interested Rokkan above all in these comparative studies were the differences he observed between countries: particular relationships between variables could be observed in some countries that were not present in others (see, e.g., p. 323). His core concern came to be to understand these differences: why do voters in one country behave markedly differently from those in another country, even in response to similar stimuli? Rokkan sought to answer this question by switching from micro-micro studies studies that used individual-level variables to account for individual-level citizen behaviour to macro-micro studies which sought explanations for citizen behaviour in the structures within which that behaviour took place. This switch is explicitly expressed in Chapter 12, where differences between Norway and the United States are interpreted as reflecting contrasts in the institutional settings for citizen decisions about politics in our two countries. The authors add, We thus become concerned with macro comparisons (p. 394). Rokkan theorizes the distinction between micro and macro approaches further in the paper that serves as the introduction to the book (Chapter 1), written in

7 introduction to ecpr classics edition iii The move from single-country to comparative analysis permits and demands greater attention to the structural contexts of the individual reactions to politics (p. 15). As in the section from Chapter 12 just quoted, Rokkan appears at this stage to conceive of these structural contexts primarily in terms of institutions: he argues it is crucial to understand the series of decisions which set the formal conditions for the political mobilization of the masses of inarticulate subjects within each territory (p. 30); and he defines structurally set restraints as the rules of procedure and the enforcement practices, the number of alternatives and the difference between them, the methods used in aggregating the choices and determining the outcomes, the probabilities of pay-off for choices for each of the given alternatives (p. 18). The other dimension of structural constraints that later gains great prominence in Rokkan s work cleavage structures is mentioned in this chapter (pp. 21, 24), but receives much less attention. The core of Rokkan s analysis of the structural determinants of citizen behaviour is provided by Chapters 3 and 4. Chapter 3 is the only chapter in the book not to have been published previously in the same form. Nevertheless, it is an amalgam of two earlier studies: the famous chapter Rokkan wrote with Seymour Martin Lipset in 1967 concerning the historical development of party systems (Lipset and Rokkan, 1967); and a single-authored piece by Rokkan first published in Chapter 4, meanwhile, contains Rokkan s most detailed analysis of electoral institutions and was first published in In these two chapters, both aspects of the structural context institutions and cleavages gain considerable attention. In order to understand the development of party systems, Rokkan argues that we clearly have to concentrate our initial efforts on the construction of two distinct typologies of sequences: one for institutional outputs, another for sociocultural inputs (p. 79). In fact, as I shall argue, cleavages tend in these chapters to accrue a more fundamental role than do institutions. I discuss Rokkan s thinking regarding institutions and cleavages in the following three sections. Given their prominence in Rokkan s work, I begin with cleavages and then turn to the impact of institutions. Finally, I consider the origins of institutions: while Rokkan does grant institutions a role in shaping the structure of competition and therefore the nature of citizen behaviour, he also argues that those institutions are, in part at least, shaped by underlying cleavage structures. It is primarily for this reason that cleavages come to dominate his approach. Cleavages and the Formation of the Party System Cleavages are, for Rokkan, key conflict lines within a polity (p. 102). They are fundamentally sociological phenomena. They do not necessarily relate straightforwardly to the party system that prevails at any given time: famously, a party system may become frozen around particular cleavages even after the underlying cleavage structure has moved on. Rather, cleavages are the lines of opposition

8 iv citizens, elections, parties within society; they do have a strong tendency to translate into the political arena; but precisely how this occurs depends on institutions. Rokkan identifies four primary lines of cleavage that, he contends, have shaped the development of party systems in Western Europe. These are: the conflict between the central nation-building culture and the increasing resistance of the ethnically, linguistically, or religiously distinct subject populations in the provinces and the peripheries (p. 102), which Rokkan most commonly refers to as the centre-periphery cleavage; the conflict between the centralizing, standardizing, and mobilizing Nation- State and the historically established corporate privileges of the Church (p. 102) the state-church cleavage; the conflict between the landed interests and the rising class of industrial entrepreneurs (p. 102) the land-industry (e.g., p. 131) or urban-rural (e.g., p. 134) cleavage; the conflict between owners and employers on the one side and tenants, labourers, and workers on the other (p. 102) the owner-worker cleavage. Rokkan initially presents these cleavages within the framework of Talcott Parsons s fourfold classification of the functions of a social system. For most contemporary readers, less familiar with Parsons s work than were Rokkan s generation, this can lend the discussion a rather forbidding air. Yet full understanding of Parsonian theory is not necessary for grasping the point of Rokkan s analysis: as Lipset and Rokkan themselves acknowledged, We might no doubt have come up with a very similar paradigm without recourse to the Parsonian core model (Lipset and Rokkan, 1967: 25). The main point at this stage is simply to understand the four key cleavages. These cleavages originate, Rokkan argues, in two revolutionary episodes: first, the National Revolution or Democratic Revolution, which began in France in 1789 and which pitted supporters of the strong nation-state against opponents in either the peripheries or the church; second, the Industrial Revolution, originating in Britain, which spawned conflict between old landed interests and the rising entrepreneurs and between workers and owners (pp ). As ever, however, Rokkan is interested principally not in the cleavages themselves, but rather in differences between the party systems of different countries that create different structures for the political behaviour of individual citizens. Thus, he is concerned with differences among the cleavage structures of Western Europe. These derive from different experiences of the two revolutions, shaped, in turn, by different prior conditions in each country. These differences relate, primarily, to the first three cleavages: the fourth emerged in similar terms in all West European countries (in contrast to the United States) and therefore tended to reduce cross-national variation (pp ). Rokkan sees the particular forms taken by the first three cleavages in each

9 introduction to ecpr classics edition v country as shaped by three critical junctures. The second and third of these critical junctures are the national and industrial revolutions already described. The first critical juncture, meanwhile, defines the prior conditions in each country and consists in the outcome of the Reformation: where the forces of the Reformation prevailed, the state came to control the national church; where, by contrast, the Counter-Reformation won out, the state was allied to the Roman Catholic Church. By dichotomizing the outcomes of each of the three critical junctures, Rokkan defines eight possible paths generating eight distinct cleavage structures (see p. 116). The outcome of the third critical juncture, like the first, is straightforward: the core actors who control the machinery of the state ally at the time of the industrial revolution either with landed or with urban interests. The former case is illustrated by Britain, where landed interests were on the whole prosperous, facilitating the alliance between rural and urban elites represented by the Conservative Party. In the latter case, however, the split between urban and rural interests is more intense, leading to the development of separate parties of agrarian defence. The outcome of the second critical juncture the national or democratic revolution is more complex, and Rokkan s theorizing perhaps a little more ad hoc. Where the Counter- Reformation prevailed, the outcome at the time of national revolution can be either the continuation of state-church alliance, as in the Habsburg Empire, or the victory of secularizing forces, as in France, Italy, and Spain. Where the Reformation succeeded, however, the national revolution is much less decisive; the contrast that Rokkan draws is between those cases in which the national church s dominance is unchallenged and those with a strong Catholic minority. The original article by Lipset and Rokkan raises the possibility of adding a fourth critical juncture the Russian or International Revolution (Lipset and Rokkan, 1967: 47 50) and Rokkan takes this further in Chapter 3 of CEP. The Russian Revolution, he argues, accentuated existing lines of cleavage within the working class movement. He suggests that it was most likely to lead to radicalization within the working class and a split between socialist and communist currents where nation-building was recent and less fully settled, or where church-state conflicts persisted (p. 137). This fourth critical juncture is thus accompanied by a fifth cleavage line pitting the strategy of integration into [the] national polity against commitment to [the] international revolutionary movement (p. 131), which Rokkan characterizes variously as the Communism vs. Socialism cleavage (p. 133) and as the International-National cleavage (p. 134). That Rokkan views this cleavage as belonging to the core set is indicated by the concluding section of Chapter 3, which refers to the five Cleavage lines as if they form a settled menu. Indeed, at one stage (Table 6, on p. 133), Rokkan appears to distinguish seven cleavages, but nowhere else does the number rise above five. Thus, while the Lipset-Rokkan set of four cleavages is almost universally cited, Rokkan may ultimately have preferred to think in terms of five.

10 vi citizens, elections, parties The Impact of Institutions As we have seen, Rokkan became interested in the impact of institutions as early as the 1950s, and questions about variations and changes in the institutional rules of the game (p. 78) lay at the heart of the papers that became Chapters 3 and 4 of CEP. Given his general focus upon the determinants of citizens political behaviour, Rokkan was most concerned to understand the institutions that shaped citizens participation in the polity. He identified four types of institution characterized in terms of four thresholds of citizen participation: the threshold of legitimation: whether opposition to the ruling elite is allowed; whether rights of assembly, expression, and publication are protected (p. 79); the threshold of incorporation: whether the potential supporters of an opposition movement have formal rights of participation above all, whether the franchise has been extended to the mass of the population (p. 79); the threshold of representation: whether barriers to gaining representation in elected bodies are high or low; this is determined, above all, by whether the electoral system is majoritarian or proportional; the threshold of executive power (referred to in Lipset and Rokkan [1967: 27] as the threshold of majority power ): whether access to executive power is determined by the results of elections, and whether such access is awarded to the group able to secure the parliamentary majority or according to the Proporz system of consociationalism. Rokkan does at times discuss other institutions notably secret voting (pp ). But the four thresholds form the core of his analysis, and I shall restrict the discussion here to them. The first two thresholds are clearly fundamental to the process of democratization. As Rokkan writes (pp. 79, 82): The first two thresholds control the development of competitive mass politics. Once the threshold of legitimation is lowered there is a significant change in the character of politics: conspirational elite conflicts and repressive measures against dissidents tend to give way to public debate and open competition for support. Once the suffrage threshold is lowered the potential audience for such debate and the potential market for such competitive efforts increases by leaps and bounds: the result will almost invariably be a rush to develop organizations for the recruitment of support and for the consolidation of political identities. Passing the first threshold is necessary for the development of any meaningful sort of party system: if no oppositions are permitted, cleavages cannot manifest themselves in stable competition for power. Similarly, passing the second threshold is required before the full cleavage system can be expressed in the partisan arena: without enfranchisement of the mass population, neither the owner-worker cleavage nor the communist-socialist cleavage can emerge in the party system. Thus, these first two institutional thresholds are fundamental to the story of party

11 introduction to ecpr classics edition vii system development that Rokkan tells and fundamental to the structural context determining how (indeed, whether) citizens can participate in the polity. The fourth threshold is also, in part, essential to democratization. This threshold really involves two elements: first, whether executive power is subject to democratic control; second, assuming democratic control, whether the executive is constituted on majoritarian or consociational grounds. Until the first of these thresholds is passed, clearly, we cannot speak of democracy. Rokkan s discussion of this threshold is, however, brief, and is largely restricted to a description of the varying circumstances in different countries at different times (pp ). Presumably this is because it does not have a direct bearing upon how ordinary citizens behave. As ever, Rokkan is principally concerned to explain variations in the opportunities available to citizens across countries, and for this purpose it is the third threshold that is most salient. While differences in timing regarding the first two thresholds may have left some legacies, by the time Rokkan wrote they had been passed in all the countries that interested him except Switzerland, where women could not vote until Rather, it was in terms of the third threshold that significant variation remained: while most countries in Western Europe adopted some form of proportional representation, a few Britain and (except for brief interludes) France did not. Understanding the effects of different electoral systems is therefore essential for Rokkan s approach. Given this, Rokkan spends remarkably little time exploring those effects. He notes in Chapter 3 of CEP that the stakes of the game will vary with the rules of representation and the rules of access to executive power but does not elaborate (p. 87). He also suggests that The introduction of PR in the final phase of mass mobilization helped stabilize, if not ossify, the structure of partisan alternatives in the central, more differentiated regions of each country (p. 90) a surprising claim that I analyse further when discussing the freezing hypothesis below, but one that does not get us far in understanding the different effects of proportional and majoritarian systems. One reason for this lacuna becomes manifest in Chapter 4: Rokkan does not believe that any clear effects of different electoral systems can be identified. Discussing various attempts to pin down such effects, he concludes, It turned out to be simply impossible to formulate any singlevariable statements about the political consequences of plurality as opposed to those of PR. A variety of contextual conditions had to be brought into the analysis: the character of the national cleavage system; the cultural conditions for the legitimation of representatives; the burdens of government and the leeway for legislative versus executive action (pp ).

12 viii citizens, elections, parties Yet to claim that the difference between majoritarian and proportional electoral systems has no effects at all would be extreme, and this is surely not a claim that Rokkan would make: he presumably mentions the threshold of representation in the first place because it matters because it makes a difference to who gains access to the representative organs of the state. In order to understand Rokkan s failure to examine the effects of different electoral systems in any detail, we must therefore look elsewhere. In fact, it appears that the main reason is that Rokkan sees the electoral system as itself the product of deeper forces above all, of the cleavage structure. The independent causal power of electoral institutions themselves is therefore limited. I examine Rokkan s account of the choice of institutions in the following section. The Choice of Institutions Regarding the determinants of the speed with which countries pass through the first two of the four thresholds, Rokkan offers four hypotheses (pp ), which qualitative evidence tends to support (pp ). Regarding the fourth threshold, Rokkan says nothing about why some countries democratized their executives before others. He suggests that consociational arrangements are more likely to be adopted in deeply segmented polities (p. 92) and adds several further hypotheses regarding the likelihood of minority participation in the Executive (pp ). It is, however, to the determinants of change in the electoral system the third threshold that Rokkan devotes most sustained attention. He begins by noting that, before the wave of proportionalization in the early twentieth century, three different majority systems were in operation among the countries of Western Europe: single-round plurality systems; two-round systems in which a simple plurality was all that was required in the second round; and two-round systems in which access to the second round was restricted to the top two candidates in the first round, thereby necessitating an absolute majority. 4 He explains these differing starting points simply in terms of long-established tradition: The first of the three procedures, he writes, had been established in England since the Middle Ages, whereas the method of repeated ballots had a long tradition in the Roman Catholic Church and was retained in most countries subject to strong Catholic influence (p. 156). In seeking to understand whether and when proportional systems were adopted, Rokkan explores a number of avenues. He probes the possibility that adoption of proportional representation may be easier in smaller than in larger countries, and he proposes several mechanisms that might underpin such a pattern (pp ). He devotes most attention, however, to the proposition that electoral system choice relates to the cleavage structure of the society that: The pressures for PR will increase with the ethnic and/or religious heterogeneity of the citizenry and, even in ethnically/religiously

13 introduction to ecpr classics edition ix homogeneous electorates, with the increased economic differentiation generated through urbanization and the monetization of transitions (p. 88). He finds that, indeed, the earliest moves towards proportional representation (PR) came in the ethnically most heterogeneous European countries: Denmark in 1855; the Swiss cantons in 1891; Belgium in 1899; Moravia in 1905; Finland in 1906 (p. 157). Furthermore, later moves towards proportionality in ethnically more homogeneous countries occurred where the cleavages generated by industrialization proved particularly divisive. Where the property-owning classes were able to make common cause against the rising working-class movements, as in Britain, their interests were served by and they had the institutional power to ensure the maintenance of majoritarianism. Where, by contrast, the cleavages between the rural and the urban interests went too deep, as in the Nordic countries, a broad anti-socialist front could not emerge and the fragmented establishment forces preferred to protect their interests through proportional representation (p. 158). Thus, the greater the number of strong cross-cutting cleavage lines in a polity, the more likely is it to adopt a proportional electoral system. Any effect that system might have in permitting proliferation of parties is endogenized in the process that creates it. This implies that the electoral system cannot be used to engineer the party system. As Rokkan says of all those who argue for electoral reform as a means towards systemic change, They tend to express the same naive belief in the possibilities of electoral engineering, and they show little awareness of the cultural and the organizational conditions for the acceptance of different systems of representation (p. 167). He does not argue that the electoral system is wholly determined by the structure of society: in considering the decisions in Sweden, Norway, and Denmark in the 1950s to replace the d Hondt form of proportional representation with a modified form of Sainte-Laguë, for example, he appeals to specific details of the party systems themselves, not just to the broad structure of cleavages (pp ). But such space appears to be open only at the level of details; in respect of the broad choice of electoral system types, partisan interests are essentially determined by deeper factors. Thus, while the institutions represented by the first, second, and fourth thresholds do retain independent causal power in Rokkan s analysis, very little space remains for the electoral system the third threshold to alter the structure of opportunities available to voters. The most widely cited aspect of Rokkan s work the so-called freezing hypotheses can be placed in this context. This hypothesis posits that party systems froze in most European countries after the 1920s, despite huge change at the societal level. And the explanation for this freezing lies in part in the second threshold: in the effects of the wave of mass enfranchisement that immediately preceded it. I discuss the freezing hypothesis in detail in the next section.

14 x citizens, elections, parties The Freezing Hypothesis The words in which the freezing hypothesis is introduced by Lipset and Rokkan (1967: 50) are very well known: the party systems of the 1960 s refl ect, with few but signifi cant exceptions, the cleavage structures of the 1920 s. This is a crucial characteristic of Western competitive politics in the age of high mass consumption : the party alternatives, and in remarkably many cases the party organizations, are older than the majorities of the national electorates. Unusually, Chapter 3 of CEP, based heavily on the Lipset and Rokkan chapter, does not simply integrate these words into the text: rather, it places them in quotation marks and adds, This organizational lag has rarely been given the attention it deserves in comparative work on contemporary electoral data (p. 90). Three main questions can be asked and have been asked often of the freezing hypothesis: first, what is it that freezes?; second, what does freezing mean here?; third, what produces this freezing? I consider these questions in turn. The original statement by Lipset and Rokkan quoted above identifies two objects of freezing: party alternatives and party organizations. Bartolini and Mair (1990: 63 65) among others have argued that the distinction between these objects is important. Party organizations are particular political parties: freezing at the level of party organizations requires that the same parties remain important within the party system over time. Party alternatives, by contrast, refer to familles spirituelles to parties or blocs of parties representing particular positions on particular cleavages. Freezing of party alternatives requires that major parties occupy broadly the same positions in respect of the key cleavages, even if the identities of the parties occupying those positions change. Some authors (e.g., Rose and Urwin, 1970; Maguire, 1983; Shamir, 1984) have analysed the freezing hypothesis in terms of the first of these interpretations and have tested it using various measures. Bartolini and Mair (1990), by contrast, emphasize the second, and therefore focus on inter-bloc volatility. Yet there appears to be no basis for asserting such priorities in Lipset and Rokkan s original text: they tend to mix talk of organizations and of alternatives, and there is no evidence that they clearly conceptualized the difference between them. Nor is the matter clarified by the comments Rokkan adds in CEP: even in the quotation above, Rokkan slides from organizations at the beginning to alternatives at the end. Ambiguity surrounds the meaning of freezing too. It seems to imply that once freezing has occurred no change can take place at least until some further event triggers a thaw. Yet the term organizational lag that Rokkan employs in CEP implies a different dynamic: one in which changes in underlying factors can be expected to feed through to the party system, but only gradually. Even in their original statement, Lipset and Rokkan noted that the increasing irrelevance of the

15 introduction to ecpr classics edition xi frozen party alternatives was generating frustration, alienation, and protestation within the organizationally least committed sections of the community, the young and, quite particularly, the students (Lipset and Rokkan, 1967: 54), but added, The probability that such resentments will coalesce into movements broad enough to form viable new parties is on the whole low (Lipset and Rokkan, 1967: 55). In the chapter prepared for CEP in 1970, Rokkan repeats the first of these quotations, but perhaps because of the momentous events of the intervening three years he does not repeat the second. And, indeed, we know that Rokkan referred some years later to the unfreezing of party systems and voter alignments in the 1960s or 1970s (Rokkan, 1977: 36, quoted in Flora, 1999: 18). These various hints suggest that Rokkan did not view freezing as absolute, but preferred to think in terms of a lagged response in the party system to underlying sociological change. In this he may have differed from his co-author: even late in his life, Lipset continued to argue that the old cleavages remained dominant (Lipset, 2001: 7). Greater clarity regarding the dynamics of freezing may come from clearer understanding of the mechanisms that underlie it. I have already quoted one sentence from CEP that relates to this issue: The introduction of PR in the final phase of mass mobilization helps to stabilize, if not ossify, the structure of partisan alternatives in the central, more differentiated regions of each country (p. 90). Rokkan makes a similar claim regarding the impact of PR at p. 77, yet at neither point does he explain it, and in fact the claim is a very odd one. In Rokkan s own terms, the adoption of proportional representation involves a lowering of the threshold of representation, making it easier for new groups to enter the legislature. Similarly, for Sartori (1968: 278), pure PR is a no effect electoral system: one that allows the forces in society to translate unmediated into the party system. For these reasons, the adoption of PR should make freezing less rather than more likely. Fortunately, however, other statements do imply a plausible mechanism underpinning freezing. Lipset and Rokkan (1967: 51) write, It is difficult to see any significant exceptions to the rule that the parties which were able to establish mass organizations and entrench themselves in the local government structures before the final drive toward maximal mobilization have proved the most viable. The narrowing of the support market brought about through the growth of mass parties during this final thrust toward full-suffrage democracy clearly left very few openings for new movements. That is, freezing is underpinned by a mechanism of market capture: once a segment of the population has been mobilized behind a particular party organization or alternative it becomes difficult to remobilize it in an alternative direction. Freezing generally occurs in the 1920s because this is when universal suffrage is achieved in most countries: with no further population segments to mobilize, the support market is saturated.

16 xii citizens, elections, parties This mechanism was strongly prefigured in Lipset s earlier work. In an extended discussion of continuities in voting behaviour in the United States, Lipset (1960: 290) argues that once a party has captured the support market in a particular area, voters allegiances may become intertwined with other aspects of their lives, making any change of allegiance costly. Yet we should not suppose that these ideas came solely from Lipset: Rokkan too had identified similar mechanisms, in relation not to the party system, but to the newspaper market. In Chapter 13 of CEP (first published in 1960), Rokkan finds that many more Norwegians vote for the Labour party than buy Labour-supporting newspapers, while the reverse is true for Liberals and Conservatives. His explanation lies in the fact that the Liberal and Conservative newspapers came first; they were able to capture local newspaper markets by providing extensive coverage of local matters and de-emphasizing their partisan affiliations. New Labour-supporting newspapers could not compete with long-established rivals on this basis and therefore targeted instead the core of Labour s most committed supporters. Thus, an equilibrium was reached in which newspaper circulation numbers did not match up to the support of the related parties. Rokkan also adds flesh on the bones of the mechanism of market capture in relation to the party system. We should look once more at the sentence, already quoted twice, in which Rokkan suggests that the adoption of PR aided freezing: he proposes that freezing occurs specifically in the central, more differentiated regions of each country (p. 90). Similarly, he argues that the disillusionment of the 1960s characterizes primarily the central areas and the cities, while in many of the rural peripheries the old parties are still eagerly mobilizing further support (p. 91). These remarks relate to another strand in Rokkan s work, represented in CEP in Chapters 6 and 7, concerning the differences between the politics of the centre and of the periphery. Rokkan may have moved at the heart of European political science, but he did not forget his roots, and the analysis of peripheral regions was an abiding interest that carried through also to his later work in the 1970s. In Chapters 6 and 7 he argues, in essence, that voters do not always mobilize immediately after enfranchisement (or, at least, not in a manner reflecting their underlying interests), and that these lags are longer in peripheral than in central areas because of the greater hold of tradition. Thus, for example, women s turnout was initially below that of men, and this lag was more pronounced in peripheral than in central areas (p ). If freezing is generated by market capture, therefore, it should happen following enfranchisement more quickly in central than in peripheral areas, and this is indeed the proposition that Rokkan affirms in the passages quoted. The mechanism that Rokkan posits as underlying freezing is thus clear. But does this mechanism generate genuine freezing: does it establish an equilibrium that survives until punctuated by some external shock disturbs the system? Or is

17 introduction to ecpr classics edition xiii it able merely to sustain an organizational lag, in which the party system adjusts to changes in the underlying social structure only slowly? On the one hand, the evidence seems to favour the lag interpretation. Freezing is sustained only for so long as the existing alternatives capture the support market; they must continue to do so even as one generation is replaced by another. Yet Lipset and Rokkan s original chapter acknowledges that, even by the 1960s, this inter-generational transfer had begun to break down among those students least likely to follow their forebears, and Rokkan s work on peripheries suggests that such patterns should be expected gradually to suffuse through the wider population. On the other hand, however, this does not reckon with two facts. First, political parties, like Norwegian newspapers, are capable of adapting. They are adept at incorporating new issues such as environmental politics or the politics of terrorism, security, and immigration into their programmes, thereby enabling their survival. This may permit freezing at the level of party organizations, though whether it is still meaningful to speak of the freezing of party alternatives may be open to debate. Second, where citizens become so disillusioned as to withdraw entirely from the political sphere, they may cease to have any influence on the party system: political parties compete for the votes of those likely to turn out on polling day. Full analysis of these propositions would require careful empirical enquiry for which this short introductory chapter is not the place. Suffice now to say that, while the chapters of CEP provide no conclusive answer as to what Rokkan would have predicted, they do offer valuable materials for understanding Rokkan s thinking in respect of the freezing hypothesis beyond what is available in the original statement in the Lipset-Rokkan chapter. ROKKAN THE METHODOLOGIST I have focused so far on Rokkan s substantive contributions to our understanding of the political world. Yet it would be surprising if a scholar who began his professional career conducting sample surveys and ended it constructing heroic models of pan-european history from the early Middle Ages to the present told us nothing about the methods of generating knowledge. His use of survey methodology was not notably innovative: though he was one of the pioneers of comparative survey research, he was hardly alone in this regard. Rather, it is primarily as an early advocate of triangulation that Rokkan the methodologist remains notable today. Before highlighting this, a brief review of Rokkan s methods is in order. Rokkan s use of sample survey evidence is represented in CEP by Chapters 9 and 10 reporting on the seven-country teacher study and Chapters 11 and 12, which report findings from the 1957 Norwegian election surveys, compared in the latter case with evidence from the United States. In addition, Chapter 8, written in 1969, provides a history of survey research and discusses in some detail three