Initially I posed the question of how nineteenth-century liberals in two European regions went about managing political association. From a certain point of view this issue might seem too singular to attract any serious attention, were it not for the wider ramifications of the problem.
Firstly, political association was without doubt one of the major challenges faced by modernizing societies on the threshold of democracy. Modes of organization and instruments for collective action, which we today take for granted, had to be invented when modern political movements formed. In the wake of the Napoleonic Wars, regimes across the whole of Europe faced the challenges of social upheaval, demands for constitutional reform, and the need for economic reform. All of this required new ways of conceptualizing the relationship between subject/citizen and ruler/state. Political parties filled the role of mediating agents, but only eventually. New forms of civic and political association had first to be tested in real-life politics. Therefore it was not merely an expression of wishful thinking that made Erik Gustaf Geijer hail voluntary association as emblematic of a new time and a new society in the 1840s. The kind of ubiquitous organizations he had in mind, and which have been depicted here in chapter 2, really were a societal innovation. ‘Politics’–always a flammable topic–became a critical factor in that perspective. Early forms of civic and political association were certainly elitist. And ‘politics’ as such remained exclusive. Since the Renaissance it had, by definition, been closely related to and, in fact, more or less conflated with ‘reasons of state’ (cf. Viroli 1992); in our context Bismarck’s reign in imperial Germany represents the clearest expression of such an conception. Quite naturally these circumstances made it difficult to establish parties as a legitimate form of association and as legitimate representatives of mass-based, social, and economic interests. The impression of parties being mere coteries rather than a viable solution for competition between interests and the selection of new governing elites (Gunther & Diamond 2001) lingered on throughout the nineteenth century.
Secondly, liberalism as such is of particular interest in regard to party systems. As diverse in composition as the opposition against l’ancien regime was in different countries, this feature was reflected in the making of ‘Liberalism’. Yet it it crystallized into one of two major, radical alternatives for modern society, and into a driving force behind political modernization and democracy. From a Marxist point of view, as well as in various strands of postmodern philosophy, it is opportune to declare the demise of liberalism. Perhaps the critics of liberalism are even right on this matter. Liberalism may be but one of several possible strategies according to which mankind has tried to manage modernization and the tensions between the individual and the collective. But it is a very important one, and from a historical perspective a basic fact remains: regardless of the obvious drawbacks of ‘economic individualism’, Western society, and the manner in which we manage our polities, is to a considerable extent a result of ‘political individualism’.
What concepts such as ‘liberty’ and ‘individual freedom’ may lack in clarity in everyday use, they make up for in terms of flexibility and applicability when put into practice in different social and cultural contexts. The two case-study regions researched in this book are excellent illustrations of this. And still, despite all obvious differences between Värmland and Schleswig-Holstein, there were also certain important similarities. Indeed, as my analysis demonstrates, liberalism, although technically speaking rational and anti-clerical, could quite easily merge with various expressions of religious values and activities on a grass-roots level. It is difficult to ascertain the extent to which the patterns of ‘frisinne’ in Värmland and ‘linksliberalismus’ in Schleswig-Holstein correspond to that typical of English popular liberalism in the era of Gladstone and Dickens. For instance, Biagini points out that nonconformism infused liberalism to the extent that it actually became an expression of ‘applied Christianity’ (Biagini 1992 [2004]: 31–41, quote at 34). However, evidence suggests this to be the case, particularly as far as Sweden as a whole and Värmland in particular are concerned (chapter 3).
Yet alignments such as these also had inbuilt tensions between different strands of individualism and different modes of association. The result was by all means a compromise if viewed from an organizational perspective on liberal party formation. With respect to the models proposed first by Maurice Duverger (1967 [1954]) and later Angelo Panebianco (1988), what the notion of ‘organization by means of proxy’ really conveys is the importance of ideas and ideology to party formation. Ideology as crystallized in early nineteenth-century forms of civic association was certainly a crucial factor in the original design of liberal parties, whether in Sweden or in Germany. In relation to emerging mass-politics by the end of the century, however, old and new forms of association clearly became locked in conflict with each other.
I have described, at some length, the tensions between the Liberal Party and its national umbrella organization in Sweden (i.e. ‘Frisinnade landsföreningen’), as well as the juxtaposition of ‘Bewegung’ and ‘Partei’ in the German context. This dimension of organizational dynamics embodied the struggle between middle-class elite politics by proxies, and more extensive grass-roots mobilization by means of mass-organizations and regular parties. These were options which, I argue, were of a strategic nature only secondarily but ideological primarily. Thus we also approach what was arguably one of the main reasons why liberalism ultimately failed in terms of an organized political movement and party. In line with Barrington Moore’s classic study (1967), however, the outcome with regard to democracy differed immensely, depending on the socio-economic setting of the process. Sweden and Germany are both critical cases in that respect.
Beginning with Sweden, the country displays a pattern of gradual and negotiated democratization. Liberalism provided an important link between early, popular radicalism and what later, in the inter-war period, became the nucleus of a modern welfarestate derived from a Social Democratic–Agrarian Party alignment in the 1930s; this line of argument has been forwarded most notably by Hurd (2000). Swedish liberalism compared to German liberalism was in that sense more inclusive, and prepared nineteenth-century society for more extensive grass-roots participation in politics. In a more specific sense the failure of liberalism, therefore, was due to it lacking the means–in ideological but also organizational terms–to maintain its initially extensive electoral support during the crucial years following the First World War. From that point of view it is interesting to look at some of the old arguments on the issue of the roots of the ‘Swedish Model’. Most notably Therborn (1989) has suggested that two of the main reasons for the Swedish pattern of negotiated democratization under the social democratic aegis were the more benign nature of the state, and the inherent weakness and fragmentation of the middle classes. This, in turn, paved the way for the social democratic labour movement to conquer and subsequently dominate the political arena, and to gradually permeate the state bureaucracy and make it an instrument of welfare policies (see also Esping-Andersen 1990).
However, this interpretation holds up to closer scrutiny only insofar as we accept two important premises: firstly, that the liberal and conservative parties depended, if not solely then in any event heavily, on the support of what was a numerically weak middle-class electorate; and, secondly, that ‘left’ and ‘right’ were really the only two political orientations in nineteenth-century society. With respect to the first assumption, it simply lacks support in empirical evidence concerning the Liberal Party/‘Frisinnade landsföreningen’. The average liberal voter in 1900 was, as I have shown in the previous chapters, rural and more often than not agrarian (see also Nilson 2010). As far as the peasants are concerned, it is difficult to posit them as unanimously ‘left’ or ‘right’, at least up to the First World War, viz. the formative stage of the modern Swedish party system. Regarding the second assumption it is, in brief, anachronistic from a historical point of view. Although most notably the tariffs debates in the 1880s promoted a left-right cleavage of the classic (‘normal’) kind, it is only at a later stage, beginning with the step-by-step extension of the right to vote from 1909 onwards, that a conception of political life in terms of ‘left’ and ‘right’ becomes entirely meaningful in analytical terms.
The liberals’ claims to represent a non-biased, cross-class, middle-of-the-road politics might appear naïve in hindsight. It might also have lacked credibility among contemporary opponents and, increasingly, a majority of voters. The fact remains, however, that ‘left’ and ‘right’ as we perceive these orientations were only part of the picture. The role of the farmers as a core group among the liberal voters and, increasingly, the challenge of organized Agrarianism, provide apt illustrations of the problem. Yes, Agrarianism was certainly in many ways conservative in outlook but, no, not every facet of it can so easily be translated into a matter of left vs. right (Eriksson 2010). Because of the diverse composition of the liberal partisans, the prospects of alternative paths to political modernization were always there. That none of these alternative paths were successfully explored is another matter.
The post-1870 German case is more complicated to assess, not least because of the greater regional diversity in terms of political culture. What held true for Prussia did not necessary hold true for her more recently acquired provinces, let alone the south German states. Despite the shortcomings of the ‘Sonderweg hypothesis’, though, any standard assessment of German political modernization must still account for the specific nature of the state; for the domestic social and economic tensions in the Weimar era; for the almost hypnotic fear of ‘Socialism’ among the German elite and the middle classes; and, importantly, for the xenophobic and militant outlook of German nationalism. In particular Kurlander (2006a, 2006b) stresses that an overwhelming majority of the German liberals were ethnic nationalists by the time of the First World War. At the same time, however, there were strategically important exceptions to this pattern, as I have tried to demonstrate in chapter 4. Kurlander’s succinct argument is nevertheless compelling. This is not least the case if we take stock of what traditionally was a core group of liberal voters in Schleswig-Holstein, the farmers. In a sense Kurlander’s explanation touches upon that of Barrington Moore (1967) and Tilton (1975), i.e. that the German farmers never completely modernized in terms of values and beliefs. But if so, we need also to ask how and why the liberal mindset developed in this direction. How, more precisely, were societal and political value systems created and diffused, and how were political ideologies translated to the grass-roots level?
I would not hold that liberal ideology in itself, at least not left-liberalism, was generically infested by ethnic nationalism as this dimension increased in importance in the 1890s. Indeed, as late as 1898, for instance, the Kieler Zeitung had stressed that ‘parochial anti-Semitism’ and liberalism were opposites.1 What should be noted was a gradual shift in this direction, much due to tactical necessity; this is, if anything, illustrated by the intricacies of liberal election campaigning during the period studied in this book. Rather, ideology was certainly important, but for more reasons than the dissemination of ethnic nationalism. What changed with the appearance of the ‘Bund der Landwirte’ and the ‘Deutsch-sozialen Reformpartei’ was the competition; more specifically, the number of available options to the electorate.
Continued left-liberal electoral success in the early twentieth century was therefore dependent on a combination of ideological adaptation to changed conditions and, importantly, the safety catch provided by the election system with its second ballots. This was crucial in a social environment characterized by peasant individualism and self-determination; features which could just as easily be incorporated with extreme, right-wing politics as with liberalism. This was particularly the case in a situation where organizational tenets of political action were lacking. Yet at the same time the liberals’ emphasis on individuality seems to have been counterproductive precisely in establishing an extensive, centralized party organization. Although the ‘Reformpartei’ failed to make its position in the party system permanent, the inroads made into the electorate if anything demonstrated the drawbacks of liberal organization by proxy at large and left-liberal organization in particular. Put simply: the need for ideological adaptation as well as the dependency on the institutional underpinnings of the electoral system were triggered by the organizational weakness of the party apparatus. The latter, in turn, reflected the traditional liberal ambiguity towards civic and political association.
Although the organizational factor has been highlighted most recently by Thompson (2000), it is important to discuss the implications more comprehensively. Following Gunther and Diamond’s (2001) blueprint of party functions, political parties also serve as instruments of ‘societal representation’ and ‘social integration’, not only electoral mobilization. Insofar as they succeed in doing this, even modern mass parties may well be considered as large-scale examples of social cooperation. Social cooperation, in turn, implies shared values and beliefs on the basis of which mutual trust may evolve (see chapter 1). Consequently, parties are also important instruments for the formation and dissemination of political values and beliefs in society. If proper institutional arrangements, which facilitate the building of trust and shared beliefs, fail to develop, so also do cooperation and the formation of, for example, democratic values. This, I believe, is the ultimate lesson to be learnt from liberal organization in the case of Schleswig-Holstein.
The difference between left-liberal party organization in Schleswig-Holstein and that of ‘Frisinnade landsföreningen’ in Värmland is, in terms of principle, only a difference of degree when viewed in this perspective. In both cases proxies, not outright, mass-based organizations, patterned liberal mobilization, although Swedish proxies drew on much more extensive grass-roots networks. Differences such as these between the regions lead us on to the more general problem of making historical comparisons. Heinz-Gerhard Haupt is one of a long succession of scholars who stress that ‘all historians compare’ (Haupt 2007: 697). Indeed we do. And his claim that localities and regions may be more appropriate units of analysis than countries is also a reasonable one. Presumably the latter holds true not only in regard to the study of social practices and demographic patterns–two examples advanced by Haupt (2007: 711)–but also in relation to political organization and mobilization. This is so since modern party systems owe their origins as much to regional processes and conflicts as they do to national initiatives. The Netherlands and Germany are but two classic examples (Lipset & Rokkan 1967); Sweden a perhaps less obvious but still third illustration of this pattern. For instance, if nothing else the tariffs elections in the 1880s and the different regional attachments of protectionists and free traders demonstrate this. Yet this does of course still leave us with the question of which sub-national regions and localities to compare.
In this case two dissimilar regions have been analysed in the search for an explanation for one common feature, i.e. relatively speaking successful political mobilization brought about by emerging liberal parties. In social science terminology the approach is basically a ‘most different systems design’ (MDSD). Compared to comparative research informed by ‘most similar systems design’ (MSSD), it does not aim to mimimize the number of experimental variables (Przeworski & Teune 1970: 32). Rather, the principal task in … [MDSD] is to find relationships among variables that can survive being transported across a range of very different regions’ (Peters 1998: 41). Note however, firstly, that ‘regions’ in the context of Peters do not necessarily imply sub-national regions but, more often, groups of countries and hence a type of macro-level analysis which is different from the kind of perspective I offer here. Secondly, I have refrained from using the above terminology until now largely because both strands of comparative research are more often than not associated with statistical probing and testing of hypotheses. Neither are possible in the present case, nor do I believe them appropriate considering the nature of the problem at hand. Nevertheless, the scientific logic underpinning MDSD and MSSD raises important issues with respect to the number of cases appropriate for qualitative analysis.
On the one hand, for instance, Peters warns against two-country studies since these impose limits as to how far the results of empirical observation may be stretched and, at the same time, make it difficult to control for extraneous variation. Indeed, from that perspective MDSD may be problematic since its logic is ultimately Popperian: the ambition is to falsify any number of possible explanations in the search for those factors which may explain an observed similarity. MDSD executed on a small number of cases would therefore seem more suitable for formulating hypotheses rather than for testing them (Peters 1998: 37–41, 58–69; see also Lijphart 1971). On the other hand, any in-depth approach as well as the need for extensive historical contextualization does, of course, limit the number of cases likely to be successfully managed. This is certainly true, although at the same time I find Peter’s analogy to Clifford Gertz’s famous method of ‘thick description’ (Peters 1998: 5–6) slightly misleading. One need not be excessively post-structuralist in outlook to realize that the historian’s task is never strictly descriptive. Nor can it provide a fully ‘neutral’ context to the cases being compared. Since historians cannot create their own primary data for experimental purposes, there is and always will be a tension between the need to rely on available primary sources, such as they are, and the equally important lack of trust in just about any claim made by these sources. Membership statistics from contemporary civic associations are a case in point (chapters 3 and 4).
Because of this, the possible number of cases for comparative purposes remains limited. ‘Normally, it is not possible to integrate more than two or three realities in a research project’ (Haupt 2007: 703). We then approach the reasons for including Värmland and Schleswig-Holstein in this particular study. To begin with, there is a large body of literature ostensibly indicating that both regions were atypical with respect to liberalism. Reiterating my initial argument, however, liberalism could just as easily become successful in rural environments as in urbanized and industrialized ones. Urbanization and industrialism were two of the outcomes of modernization, but these processes posed new challenges with political implications for rural societies, too. This certainly goes for Scandinavia and it holds true for parts of Germany as well, such as Schleswig-Holstein. In fact, writes Haupt, participants in the old Sonderweg debate all too carelessly let themselves become infused by ‘an idealized vision of the British situation’, and too easily overlooked German regional differentiation (Haupt 2007: 707). Rather, as any attempts at unearthing anything like ‘average liberalism’, or an ‘average path’ to political modernization are likely to fail, the historian’s task must be to identify and explain variation in terms of this process. That political modernization is not necessarily sequential does not thereby rule out certain crucial commonalities between seemingly different countries and regions. This feature is amply illustrated by the cases of Värmland and Schleswig-Holstein.
In relation to the crucial variables researched here–strong liberal turnout in a basically rural environment–Värmland and Schleswig-Holstein therefore turn out to be quite similar to each other regarding the modus operandi in terms of organization. This is the case regardless of the orientation of liberal policies in the main in Sweden and Germany (one set of policies gradually being more oriented towards progressivism, the other becoming more closely bound up with the intricacies of economic liberalism, deregulation, and free trade). This also holds good regardless, for instance, of the position taken in relation to the respective electoral systems and their design. In the early 1890s, local liberals in Värmland pointed to the example of Germany and the system of direct elections based on the rule of absolute majority, as a means of engendering grass-roots participation and, consequently, the liberal turnout. Only gradually did proportionalism, as later promulgated by the 1909 election reforms, gain ground.2 Absolute majority and the system of second ballots were, however, used by German liberals to effectively block unwanted grass-roots competition from the socialists.
Finally, considering the rural framework of liberalism and, indeed, of all political movements in these two regions, it is worthwhile noting the importance of geography. Party formation processes and political mobilization were, in one sense, all about distance. From this perspective rural regions, particularly in Scandinavia, become extreme from a historical point of view. Put simply, organizing political parties and mobilizing one’s partisans is obviously easier in densely populated areas than in regions where much arduous and time-consuming travel is required. Albert Hänel’s criss-crossing of Schleswig-Holstein during the 1884 elections is one example of this (chapter 4). Swedish campaign workers had to overcome worse obstacles than that.
Indeed, Värmland was far from being the worst of the boondocks. Conditions in the north were even more extreme: in connection to the 1911 elections one of the liberal campaigners in the province of Västerbotten covered 1,634 kilometers by train, by horse, and by bicycle (chapter 3, n. 63). The same distance travelled could have taken him down to Cracow, almost the whole way across to Iceland, or to the far side of Moscow, should he have so wished. People like him might have been particularly hard-nosed. We do not know. It is safe to say, though, that feats such as these make the successes of rural liberalism and its organization, haphazard as they many times were, all the more remarkable. It was, perhaps, precisely because of factors such as these that organization by means of proxy worked, at least up to a point, or–even–became a necessary means during the initial stages of party formation. Compared to a homogeneous and hierarchic mass-organization, which depended on continuous directions from the national executive, a political movement based on more or less autonomous bodies was more flexible during an age in which the technical means of communication were poorly developed. This might be somewhat of a paradox. For all its drawbacks–one of which facilitated the rise of political right-wing alternatives in Schleswig-Holstein–organization by means of proxy also served a purpose, when viewed in relation to the practical constraints on political association in turn-of-the-century Europe.
We have so far spoken of parties rather than party systems, the latter a more important analytical category from the point of view of political scientists. Still, even studies of single parties and political movements may be enlightening from a systems perspective. Thus, from a systemic perspective the main problem for contemporary liberals in both Sweden and Germany–that of managing emerging mass-politics–was a general feature of party systems development in countries across the Continent as well as in the US. Although the precise sequence of events as well as outcomes differed between Europe and the US, ‘the development of the contemporary party systems … coincided with extensions of the franchise and the mobilization of a peripheral electorate’ (Petrocik & Brown 1999: 43). This, in turn, unleashed the kind of challenges and crises faced by contemporary parties and factions when the expanding electorate gradually realigned according to the diverse patterns typical of our two regions. Self-evident as this observation might seem, it is nevertheless important for our historical understanding of party dynamics. In the words of Petrocik and Brown, ‘electoral mobilization is not likely to be a major contributor to future party system changes among the developed democracies’ (ibid.). In terms of causality, therefore, the eventual failure of organized liberalism in Sweden and Germany was as much a reflection of more general trends in Western political modernization as it was due to any specific historical ‘necessity’ inherent to Swedish and German liberalism. Ultimately the matter of ‘sequentialism’ versus ‘non-sequentialism’ does, of course, remain more complex than this. But at least with respect to the rise and fall of liberalism, these theoretical models fail to capture the intricacies of the story.