Contingency and modernity in the thought of J.P. Arnason

This article argues that Arnason’s writings succeed in pushing civilizational analysis — most prominently developed by the late Shmuel N. Eisenstadt — in a much-needed direction. Coming from an action-theoretical background in which the creativity of actors is strongly emphasized, Arnason is critical of approaches within civilizational analysis that tend to downplay contingency within historical processes. Especially by focusing on the role of political power and imperial encounters, Arnason demonstrates how civilizational analysis can be further developed in ways that do not automatically assume the linearity and long-term persistence of civilizational paths.

theory of the 1950s and early 1960s insofar as the 'original' and highly questionable dichotomy between tradition and modernity has now been replaced by a kind of crude process-thinking that expects an ever more cosmopolitan world. Quite a few remnants of modernization theory can also be detected in John Meyer's world polity approach, which enormously simplifies Talcott Parsons's theoretical insights by focusing almost exclusively on processes of cultural diffusion. One could continue in this manner, but a further screening of the market of contemporary macro-theories would not lead to a significantly different result: most of these approaches will not present a very convincing picture of modernity, its dynamics and conflicts (cf. Knöbl, 2007: 21-60).
From this perspective, there is some hope that civilizational analysis might provide a way out of the impasse since it explicitly does not rely on disputable premises such as those of the paradigms just mentioned. Whatever the weaknesses of this approach -and the following pages will deal with some of them -it is clear that practitioners of civilizational analysis, by focusing very strongly on religious and cultural patterns of civilizations, are able to avoid economic reductionism. Civilizational analysis can also leave behind the problematic dichotomy between tradition and modernity by emphasizing the very different dynamics of different civilizations. And last but not least: theorists within the civilizational paradigm for good empirical reasons do not expect a homogenized world but rather a very heterogeneous one, not one modernity, one world polity, one world society, etc., but multiple modernities according to the still recognizable effects of civilizational patterns. Thus, if we want to advance the status of macro-sociological theory, there is no better way than to use civilizational analysis as a starting point. But, as we will see, civilizational analysis needs to be sensitive to some major theoretical and methodological problems -problems that can be dealt with only by categories and instruments that can grasp the phenomenon of historical contingency. And it will be my contention, that Johann P. Arnason's way of doing civilizational analysis is the most promising in this respect. Proving such a claim, however, is not an easy task since the debate between the main discussants within civilizational discourse is oftentimes rather complicated. For the most part, my purposes are best served by contrasting the approaches of Eisenstadt and Arnason. But sometimes it is also necessary to argue with Arnason against himself in order to demonstrate the full extent of his originality. Let me begin with the latter approach, arguing with Arnason against himself.

I
In a recent article on Iceland -where he was born and grew up -Arnason began his historical-comparative analysis of the region with a theoretical statement that seems doubtful to me. There he claims that 'civilizational approaches have proved particularly instructive in regard to historical breakthroughs and turning-points, and the theoretical perspectives of scholars in the field have to some extent been influenced by their choice of paradigmatic cases' (Arnason, 2007: 2-3). One would certainly not like to argue with the second part of this quote. The difficult task of analyzing whole civilizations and mastering long periods of history requires the individual researcher to closely focus on his or her field of specialization. There are few (Eisenstadt and Arnason among them) who can really claim to know the literature on more than one or two civilizations. Therefore, it is understandable that most authors who try to generalize their theoretical insights are necessarily biased by the cases they know best. The first part of the quote above is more problematic, however. And this is the point where it is necessary to argue with Arnason against his own statement. What do I mean by that? First of all, it is certainly true that the writings of Eisenstadt, of Arnason himself and a few others on civilizations have had and still have a huge impact on the social sciences and the humanities and have forced various disciplines to focus their research much more than in the past on a period which Karl Jaspers has termed the 'Axial Age'. Although the contours of this age are highly contested (cf. Arnason et al., 2005), few would doubt that the 'Axial Age' was a time of enormous cultural creativity and thus a major turning point in world history, one worthy of being studied intensively. Therefore, Arnason's claim that civilizational analysis is instructive in 'regard to historical breakthroughs and turning-points' seems plausible. But only at first sight because the counter-argument: 'civilizational analyses are in danger of neglecting turning points and thus historical contingency', is at least as convincing. Why? According to mainstream civilizational analysis, it was the axial breakthrough which decisively shaped the patterns and dynamics of the then emerging civilizations. Such arguments almost necessarily tend to downplay the importance of all later events since they will not and cannot seriously change the contours and developmental path of the particular civilizations formed by this turning point. Defining a civilization by the specific form of its axial breakthrough almost always leads to a devaluation of its subsequent history.
One can clearly see this point in the work of Shmuel Eisenstadt: I think that it is more than just by chance and more than an odd functionalist remnant when Eisenstadt speaks of 'cultural programmes' of civilizations or a 'cultural and political programme of modernity' ('kulturelles und politisches Programm der Moderne') (cf. Eisenstadt, 2000: 15ff.), thus somehow assuming the existence of a kind of software that was developed at a very early phase of a civilization and that somehow determines all the processes still to come. Arnason himself is highly critical of this -as he calls it -functionalist language in Eisenstadt's work. What he does not see so clearly or, at least, what he seems to downplay, is the fact that this is not Eisenstadt's problem alone, but one of civilizational analysis in general. The tension between structure and process, between fixed patterns and contingent events -a tension to be detected in all categories that try to define social wholes -seems to be particularly strong in the case of 'civilizations' since their long existence stretching oftentimes over many centuries makes it difficult to give events and thus historical contingencies their due respect. Events -so the kernel of the argument goes -tend to play a rather marginal role in all civilizational narratives since they threaten the theoretically assumed identity of any particular civilization. The longue dure´e of civilizations seems somewhat incompatible with the idea of contingent events, with turning points and cultural breakthroughs after the 'Axial Age'.
My claim therefore is that civilizational approaches have not only failed to come to terms with -as Arnason (2003a: 107) calls it -Spengler's problem, namely 'the idea of cultures as closed monads', but also with the problem of contingency since, contrary to Arnason's statement above, civilizational analysis is clearly not 'instructive in regard to historical breakthroughs and turning-points' per se. The seduction of constructing all too linear historical processes always looms large within civilizational approaches.
However, if there is one author who really tries to get contingency into the theoretical framework of civilizational analysis, it is Johann P. Arnason. But by doing this, he is certainly not using mainstream sociological instruments and concepts. On the contrary, his somewhat unique ability to resist this seduction of marginalizing contingency depends on conceptual tools he has developed over a long time period in which he must have often felt quite uncomfortable with the theoretical developments that were going on within sociology. An appropriate intellectual heritage that he could use for his theoretical purposes was difficult to find within the discipline. Certain currents within philosophy were of much greater help. To be more concrete at this point: Arnason's life-long engagement with Marx's term 'praxis', his hermeneutical insights into the neverending conflict of interpretations within cultures (and civilizations) and his attempts to reformulate in a macro-sociologically meaningful way Cornelius Castoriadis's description of the possible emergence of radically new cultural fantasies or 'magmas' allowed him to develop a strictly anti-evolutionist and anti-functionalist research programme that promises to overcome the major problems of civilizational analysis.

II
If one wanted to summarize Arnason's theoretical work during the 1970s and 1980s, one would probably not be too far off the mark to claim that he attempted to build a theory of action and a theory of culture sensitive enough to come to terms with human creativity. His first major monographs (Von Marcuse zu Marx. Prolegomena zu einer dialektischen Anthropologie, 1971, and Zwischen Natur und Gesellschaft. Studien zu einer kritischen Theorie des Subjekts, 1976) -although clearly written from a Marxist perspectiveattempted to go beyond the assumptions of orthodox Marxian theory. As Arnason saw it, the term 'praxis' as developed in the early writings of Marx contained the seeds of a theory of creativity which was not further developed in Marx's later works and which was more or less neglected by the different schools of Marxism that flourished during the 1960s and 1970s. Although a student of Jürgen Habermas, Arnason criticized his teacher for such neglect, since in Habermas's famous distinction between 'work ('Arbeit') and 'interaction' ('Interaktion') the former Marxian notion of 'praxis' was reduced to 'work' and was thus understood merely as purposeful or rational action. Although Habermas's equal emphasis on work and interaction kept his overall theoretical approach far away from economic reductionism, his very restrictive definition of work, which excluded many of the elements that the Marxian term 'praxis' had originally contained, produced its own problems. According to Arnason, such a theoretical move tends to downplay the fact that various moments in 'work' cannot just be subsumed under the logic of purposeful rational action since work in itself creates new needs and meanings that were not there before the work process started. To put it simply, in work, there is always some kind of a surplus of meaning which does not fit the purposeful, rational or instrumental model of action. The means-ends scheme is not sufficient to grasp all of the aspects of man's relation with nature. It is not only 'interaction' -as always very strongly emphasized by Habermas -that has to be understood as a non-teleological process, but also work. This was Arnason's early insight that led him to prefer the term 'praxis' over 'work'. Thus, from the very beginning, Arnason insisted on the creative dimension of all human actions. This is certainly not a minor point and not one relevant only to a (micro-) theory of action. As Arnason immediately made clear, the idea of the creative potentials of work, i.e. of 'praxis', precludes any teleological assumptions concerning the process of human history. 1 The reason here is not difficult to see: if the appropriation of nature is not a process of pure technological or rational control, then a comprehensive and especially linear process of evolution and rationalization cannot be expected. Worldviews contain, among other things, specific and often quite different interpretations of the relationship between humankind and nature, and these interpretations cannot be forced into a developmental frame -as many sociological theorists, including Habermas, seem to have assumed. Of course, culturally diverse world-views can be rationalized internally but, given the different relations between humankind and nature, not all world-views can be transformed into one another. There is no universally valid developmental logic of world-views or cultures, as Habermas and many others would like us to believe -not even in the area of humankind's dealing with nature. Thus, Arnason defended a very strict anti-teleological and anti-evolutionist position that made him an outsider in the field of macro-sociology in which evolutionist constructions were a kind of boom industry in the 1970s and 1980s, especially in Germany.
Arnason's emphasis on the interpretative dimensions of 'work' allowed him to criticize and connect the insights of Karl Marx and Max Weber at the same time. According to Arnason, it was due to self-misunderstanding that Marx abandoned the idea of 'praxis' that has been so ingeniously developed in his early works. Drawing too uncritically on the utilitarian tradition of political economy in his middle and later works, Marx came to interpret human history exclusively in terms of progress in the forces of production. The result was a rather linear model of history in which revolutions affect fundamental changes in the relations of production, but not fundamental (cultural) transformations to work itself. Such a transformation could not even be dealt with within Marx's theoretical framework so that his philosophy of history in the end reproduced the type of economic reductionism that he had inherited from the utilitarian premises of political economy. Hence, Marx could not grasp the enormous cultural transformation that was the origin and the effect of the advent of modern capitalism. Although Marx emphasized the revolutionary role of capitalism more than anybody else, he -according to Arnasonironically did not fully realize that capitalism was above all a new cultural project and thus one of the decisive moments of discontinuity in human history -a point missed not only by Marx but by the majority of social scientists of the nineteenth and twentieth centuries.
Max Weber in his 'Protestant Ethic', in contrast, clearly saw the enormous cultural shift that accompanied the rise of modern capitalism. But Weber did not make much use of the potentials of this insight either, since he tried to understand human history through his master concept of rationalization, based on his famous typology of action in which purposive rational action was considered to be of utmost importance. This obstructed conceptualization of those creative aspects of human action that Arnason has always stressed. Thus it is not surprising that Weber's concept of rationalization was as sharply criticized by Arnason as the slightly different one used by Habermas. Such concepts, as Arnason correctly sees it, blind us to the discontinuities and contingencies of human history that result from the creative potentials of humanity. 2 Arnason's insistence on these aspects of discontinuity and contingency brought him into close contact with Cornelius Castoriadis, whose understanding of the imaginary content of all symbols made him a radical critic of the deterministic and reductionist approaches so common within the social sciences. If it is true that symbols are institutionalized and that all institutions have symbolic aspects not deducible from nature or from social structures, then -according to Castoriadis -it is clear that each form of institutionalization has creative aspects and that all institutions inherited from the past result from imaginary processes. They are expressions of the imaginary potentials of individual and collective actors. Speaking of 'The imaginary institution of society' (L'institution imaginaire de la socie´te´) -the title of one of Castoriadis's major worksalso implies a more or less discontinuous interpretation of history. If fantasies and collective 'magmas' shape processes of institutionalization and these imaginaries are in many respects radically new, then one not only has to reckon with newly emerging 'magmas' in the future; one must also be sensitive towards the (symbolic and thus institutional) ruptures within history which should not be disguised by teleological frames of interpretation. Arnason emphatically supported these ideas: they underlined his early insights about the deficiencies of the mainstream sociological theorizing that had neglected human creativity.
In the 1980s, Arnason continued and broadened his research as he became increasingly dissatisfied with Habermas's concept of 'interaction'. In a similar move to the one he had made a decade earlier when he had deconstructed and reconstructed the term 'work', Arnason now began to theorize culture as well as the various tools and concepts that are necessary to deal with this difficult term. His starting point, again, was Jürgen Habermas's theory, especially his term 'interaction'. As is well known, Habermas understands interaction as a non-teleological process since its end-result is not predetermined: Although understanding ('Verständigung') might be seen as the telos of communicative action and interaction, the very content of that understanding remains open. As Arnason makes clear, however, it is doubtful whether there is an in-built telos of communication, one that aims or even leads to mutual understanding. Although 'Verständigung' is used in a very formal way by Habermas in order to avoid the accusation of teleological reasoning, Arnason's position is very different: he just does not accept that communication is intrinsically aimed at understanding. He therefore sought to distance himself from those Habermasian concepts that he had previously used or at least accepted (Arnason, 1988). Again radically emphasizing the openness and creativity of human existence and action, from this point on he preferred the terms 'experience' and 'interpretation' in order to indicate a definitive break with over-rationalized and teleological concepts of action and interaction. As he sees it, the term 'interpretation' retains the hermeneutical insight of the never-ending conflict of interpretations, an insight which became absolutely central when Arnason began his turn towards historical sociology and civilizational analysis in the 1990s. This insight does not allow civilizations or cultures to be interpreted as closed and harmonious value systems since one always has to be aware of the existence of conflicting interpretations. And, drawing on Castoriadis's arguments mentioned above, one must also reckon with radically new meanings which might emerge at some point in history and which might therefore change cultural and civilizational patterns in rather surprising ways.
Closely related to Arnason's attempts to develop categories for an open, but nevertheless sociologically meaningful concept of culture are his reflections on power. Although one needs more than a theory of culture in order to explain social order, and although power relationships have been and remain most important in all social contexts, sociology has often failed miserably in theorizing this phenomenon. According to Arnason, one of the main problems was always that social scientists have misinterpreted power by 'naturalizing' it. Power, however, is not a 'brute fact'. It is, on the contrary, dependent on cultural resources and backgrounds. Arnason's reflections on power resemble those on 'praxis' and 'work'. For the same reason that one should not naturalize 'praxis' by defining it as a culturally neutral form of appropriating nature, one should also not naturalize power. The ideas and uses of power are deeply structured by cultural premises that differ significantly between societies and civilizations. Therefore, only by taking into consideration the possibility of different visions of power, are social scientists able to come to terms with the nation-state, with different forms of imperialism, etc.
-which are certainly not exclusively the products (or unintended consequences) of the actions of rationally calculating agents. Furthermore, emphasizing the cultural dimensions of power makes one sensitive towards the discontinuities and contingencies in the course of human history -a point Arnason has repeatedly stressed throughout his work.
Summarizing the various theoretical moves surveyed above, it is clear that Arnason's reflections culminate in a multidimensional, context-sensitive and non-reductionist conceptual framework in which 'work'/'praxis' and the production of wealth, 'interpretation' and the production of meaning and culture, and, last but not least, 'power' were decisive elements: The concepts of wealth and meaning refer to ways of appropriating and articulating the world, intertwined with but irreducible to the exercise of control over it; power is, by contrast, defined as a more strictly social category, tied to the field of interconnected actions, and although its social dynamics also translate into control and conquest of the natural environment, this extra-social side can be theorized without making power synonymous with the very capacity to intervene in the course of events. (Arnason, 2003a: 202-3) With this conceptual framework, Arnason began to work in historical sociology at the end of the 1980s and, in a very short period of time, established himself as one of the most interesting and innovative figures within this field. Space does not permit an overview of the many books Arnason has published since the early 1990s or to adequately situate him in this field. My aim in the following paragraphs is a more modest one: to argue that Arnason's writings in the context of civilizational analysis can be interpreted as a continuing dialogue with Shmuel Eisenstadt. This dialogue has forced Eisenstadt to change his original approach quite significantly due to the strength of Arnason's arguments, arguments in which Arnason's emphasis on the role of historical contingency played a decisive role. To make this point, however, it is necessary to return once again to the peculiarities of civilizational analysis.

III
Although scholars in the debate about the so-called 'Axial Age' undoubtedly have very different interests. Whereas, for example, theologians are much more concerned with the specific contours of the then created transcendent realm than with the effects of this cultural creativity on later phases of human history, social scientists are particularly interested in these cultural effects. It is also clear that the strong and rather new interest in the concept of civilization is significantly motivated by sociological ambitions to achieve an appropriate diagnosis of our own time. One studies civilizations on the assumption that inter-civilizational comparisons might provide hints and insights into the dynamics of the contemporary world. As Arnason himself puts it: 'Scholarly work on the comparative analysis of civilizations has shown that the religious-political nexus ('le théologico-politique,' as some French authors have called it) is a particularly rewarding starting-point for strategies of comparison' (Arnason, 2006b: 108). And although some comparisons do not aim for conclusions concerning contemporary affairs, it is also clear that the debate about the 'Axial Age' has some connections with the old 'problématique' of the 'Rise of the West'. Here the question is: why were some civilizations more dynamic than others and therefore headed towards scientific, political or cultural breakthroughs in later periods -not least towards the breakthrough to modernity? Whether one likes it or not, this kind of reasoning -and I think Eisenstadt's entire body of work can be interpreted as a kind of transformation of Weber's specific question concerning the occidental 'Sonderweg' -is based on arguments of the path-dependent type. Is there a 'founding moment' in the history of a civilization that created a particular 'religious-political nexus' which laid the tracks for the future development of this civilization? I think this is the point from which to begin if one wants to understand the ongoing dialogue between Eisenstadt and Arnason. It is here that one can most clearly see the differences between -as Arnason (2002: 13) puts it -his own 'historicist' and Eisenstadt's 'culturalist' interpretation of civilizations.
If one takes seriously this ambition for path-dependent explanations within civilizational analysis, quite a few theoretical and methodological problems arise -all related to the problem of contingency (cf. Knöbl, 2007: 92-107). These have to do with the specific form of path-dependence arguments: a contingent 'founding moment' is at the centre of the theoretical construction since an event (or 'turning point') initiates a path-dependent, often highly deterministic, development. To quote James Mahoney, one of the leading authors in the debate on path dependence: The identification of path dependence ... involves both tracing a given outcome back to a particular set of historical events, and showing how these events are themselves contingent occurrences that cannot be explained on the basis of prior historical conditions. Because the presence or absence of contingency cannot be established independent of theory, the specification of path dependence is always a theory-laden process. (2000: 507-8) Note, that Mahoney is not making a claim about the contingent or determined character of reality. His claim is not an ontological one. His argument is rather that social scientists construct theories with which some facts -the contingent event that triggers the trajectory -cannot be explained by the theory being used to explain the path-dependent process. It is not the event itself that is contingent; it is the event being theorized within a particular paradigm that appears to be contingent. Thus, a contingent event or a 'turning point' is created by the theoretical paradigm being used. Or, as I would put it: discussing 'contingent events', 'path-dependent developments', or 'turning points' is just an element of a (theoretical) narrative, a 'theoretical' story. If that is correct, at least two specific features of narratives and stories have to be taken seriously.
First, every narrative strategy requires the selection of events and structures for setting up a convincing 'plot'. These events and structures must be brought into a temporal order so that the audience -as David Maines (1993: 21) has pointed out -gets some insight into the 'tempo, duration, and pace' of the process under consideration. This seems to be trivial, but in fact has enormous methodological consequences which are often neglected in historical-sociological research. Since the 'contingent event', the 'founding moment', the 'turning point' is often considered as the most important element of a (theoretical) story; scholars often implicitly ascribe these events and moments some kind of a 'genetic code'. As Michel Dobry (2001) puts it: the 'contingent event' or 'turning point' is exclusively considered from the perspective of the end result of the pathdependent process, as if the contingent event already contains the seeds of the developments yet to come. Thus a strong teleological element enters the argument. When this is the case, it often does not seem necessary to describe the sequence of the events in detail. Since the 'plot' is exclusively created from the point of the end-result of a particular process, some kind of a deterministic development is implied that obviously makes a serious analysis of 'tempo ... and pace' superfluous.
I would argue, however, that a detailed analysis of the tempo and pace of a pathdependent process is not only necessary, but also very hard to deal with, especially if researchers want to bridge long time periods. The shorter the time span of a story, the easier it is to provide details of the tempo and pace of the path-dependent process. This is therefore a real challenge for all scholars working within the civilizational paradigm, which focuses so firmly on long periods of time.
Second, path-dependent approaches often analyze the beginning of a development rather exhaustively. There are some good reasons for this, of course. But it must not be forgotten that the end of such a development also has to be justified by theoretical means. The end of a path-dependent process -like its starting-point -is not 'objectively real' either, but is also part of the narrative construction. Only a precise definition of the explanandum of the research question can provide plausible reasons for identifying 'this particular' end-point of a social process. The majority of path-dependent analyses, however, often consider only two 'turning points': one with which the path begins and one with which it ends and thus a new and different path emerges. All too rarely is the question asked if in between these two 'turning points' there might be other events of interest, even other possible turning points which could make it possible to tell a completely different theoretical story, to see a completely different path. And all too rarely is yet another question dealt with, namely whether the originally discovered 'turning' and 'end-points' might be considered as merely intermediate steps of a development that in fact bridges a much longer time period, as the originally formulated path-dependent story was intended to do (Haydu, 1998: 353). Thus, what might at one moment look like some kind of a starting point of a new trajectory could -using another theoretical framework -also be regarded as a continuation of the old path. If that is true, then it is a methodological necessity to question one's own narrative constructions by continuously considering alternative plots concerning different possible paths.
If one takes these two problems seriously, the focus of the dialogue between Eisenstadt and Arnason on civilizational analysis becomes quite clear: it is Arnason who continuously urges Eisenstadt to be much more sensitive towards the phenomena of contingency and warns him of the problems and dangers of arguments of path dependence. At least four such warnings or critical comments can be detected in this dialogue.
First, Arnason goes along with Eisenstadt in the attempt to define civilizations by using religious-political criteria. He therefore has no problems in giving the so-called 'Axial Age' as it was formulated in Eisenstadt's original approach a central position in theorizing civilizations. It is typical of Arnason, however, that from the beginning he began to qualify some of Eisenstadt's premises and assumptions. First of all, Arnason is quite clear that the focus on an 'Axial Age' must not seduce historical sociologists to assume that this period of enormous human creativity was the starting point of an ever increasing reflexivity (Arnason et al., 2006: 11). Such assumptions carry in themselves the danger of a teleological view of history and thus tend to neglect the possibility of a reversal of trends by contingent events, to be more precise: of a suppression of reflexivity which was always a possibility in the course of human history as well. Then, Arnason immediately warns those who use the term 'Axial Age', and especially Eisenstadt, that the creation of new political-religious horizons in different parts of the world should not lead to the assumption that civilizations are completely defined by these peculiar creative moments of the 'Axial Age': The general category of 'axial civilizations' . . . seems to involve quite strong and debatable assumptions about cultural orientations embodied in and constitutive of whole civilizational complexes. A more limited conception of axial patterns, centred on the relationships between intellectual and political elites and their role in historical transformations, would have to allow for contextual determinants that vary from case to case. (Arnason et al., 2006; see also Arnason, 2005b: 124) Although Eisenstadt certainly cannot be accused of neglecting power struggles between different kinds of elites during the early phases of a civilization, Arnason, for good reasons, hints at the fact that in most of his writings Eisenstadt seems to freeze the results of these early power struggles as if the once negotiated 'religious-political nexus' will determine the fate of a civilization for hundreds or even thousands of years to come. In short, the analysis of the 'founding moment' of a civilization, i.e. a particular axial breakthrough, must not theoretically predetermine the internal unity of a civilization. The degree of unity has to be analyzed empirically. 3 This argument is clearly based on Arnason's understanding that culture is not a fixed entity, but a constellation in which the creativity of actors and the impact of contingent events always allow the emergence of new collective fantasies or 'magmas' (Castoriadis) that might considerably change the trajectory of a civilization. In this respect, Arnason emphasizes the creativity of civilizations and thus the surprising turns in their paths much more than does Eisenstadt. And he is more precise on this point: it is often at the margins of a civilization that new creative meanings might emerge insofar as 'the societies least affected by the civilizing process and most distant from its main centres are for that very reason most capable of religious breakthroughs that can give a new meaning to the whole process' (Arnason and Stauth, 2004: 34). This gives Arnason a much stronger position from which to make the persistence of civilizations an empirical question. Although he does not deny the fact that there are civilizations whose existence stretches over many centuries, from the beginning of his work on civilizations Arnason (2004: 110) was very sensitive about the possibility of 'multi-civilizational sequences'; about the fact that different civilizations came into being in a relatively short period of time. Arnason was therefore always very interested in that period of Late Antiquity when the Roman Empire was succeeded by three different civilizations, Western and Eastern Christendom and, later, Islam -civilizations which in themselves each transformed rather rapidly.
Second, if the persistence and stability of cultural traits of a particular civilization are in fact an empirical problem rather than simply a theoretical premise, then the following question arises: by what means are cultural traits preserved? This is an important question -even Arnason accepts that there is at least some stability! It is Arnason's conviction that Eisenstadt does not give power relationships their due respect in this regard. Although he admits that in Eisenstadt's overall theoretical framework 'le théologicopolitique' is central to the conceptualization of civilizations (Arnason, 2002: 10), he at the same time claims that Eisenstadt's attempt to theorize civilizational modes of reproduction often pushes the 'politique' of 'le théologico-politique' aside in favour of theological or religious aspects (Arnason, 2002: 71). 4 Even when analyzing the founding moment of a civilization, Eisenstadt's writings are sometimes not sufficiently elaborated. For example, his focus on the interactions and conflicts between elites during the 'Axial Age' -conflicts often brilliantly analyzed by Eisenstadt -cannot encompass all of the dynamics of the political because the political is also defined by cultural visions of power! Thus state-and empire-building as specific concretizations of such visions were insufficiently considered in civilizational analysis (Arnason, 2005a: 47). This is important not only for an adequate historical reconstruction of the 'Axial Age' itself, but also for the later history of civilizations, since 'power' not only has the tendency to stabilize social processes and thus to prevent contingency -it also frequently creates moments of contingency through the unintended consequences it produces. This leads on to the next point.
Third, if civilizations are defined by visions of power, then imperialism and empirebuilding are possible options for particular rulers. If this is the case, then (violent) encounters between different civilizations will almost necessarily happen, and events and processes such as wars of conquest, migration, etc. will follow. Arnason very early began to criticize (cf. for example, Arnason, 1993 and Eisenstadt's tendency to paint a much too isolated picture of civilizations and thus his neglect of their entanglements. In an important essay entitled 'Understanding intercivilizational encounters', Arnason began the difficult task of developing a 'phenomenology of encounters ' (2006a: 39) between civilizations (see also Arnason, 2003b: 308). There he demonstrated how important such a typology is since imperial encounters and their contingent effects often have increased but also (via a strengthening of indigenous civilizational traditions) suppressed reflexivity. Again Arnason's message is: the contingent emergence of visions of power and thus the encounter of different civilizations are triggers for rather surprising processes. If it is true, as has been argued above, that many practitioners of path-dependent analyses (including Eisenstadt) overemphasize the importance of the founding moment of a trajectory, then Arnason's emphasis on the contingent effects of imperial encounters is the appropriate means to resist the seduction of teleological constructions of history in which events can only play a minor role: Arnason's theoretical tools -his emphasis on the role of particular cultural visions of power and on the creativity of civilizations during their entire history -enable him to present a type of civilizational narrative that fulfills the aforementioned methodological requirements.
Fourth, this leads to the final critique in Arnason's dialogue with Eisenstadt. Although Eisenstadt's view of modernity is already multi-dimensional and characterized by many conflicts -Jacobinism and religious fundamentalism are certainly at the centre of Eisenstadt's writings since the 1990s -Eisenstadt's analyses of these phenomena nevertheless remained on a rather abstract level. This is not the case with Arnason's writings. His books on the Soviet experience (1993) and his comparative essays on Russia and modern China (cf. 2003b) made clear that he, even more strongly than Eisenstadt, refused to paint an 'integrative' picture of modernity (Arnason, 1997: 354). Counter-visions towards Western modernity were and are always around, visions that emerge and often become reality by contingent events -events that, for example, are caused by imperial encounters. Arnason's theoretical tools are sophisticated enough to not assume that these counter-visions are settled and fixed from the beginning. Again, his understanding of the creative potentials of cultures gives him the opportunity to see how visions of power often develop their own dynamics in the very process of their realization, something which cannot be grasped by a pure means-ends scheme. The Leninist and Stalinist project -not much different from the one envisioned by Mao -was the result of such an unpredictable dynamic process, one in which rather contingent circumstances caused a nightmare for millions of people -a moment of cultural creativity, to be sure, but a most horrible one.

IV
The dialogue between Arnason and the work of Eisenstadt will certainly continue since the problems of civilizational analysis have not all been settled as yet. Nobody probably can predict how Arnason will elaborate on the insights first formulated by Eisenstadt who has passed away much too soon. There are too many contingencies! But one thing, at least, seems certain and is thus not affected by contingency: one cannot ignore Arnason's brilliant insights in comparative-historical analysis.
Notes relationships considering that he began his international career in the early 1960s with a masterpiece on historical empires (Eisenstadt, 1963).