The Contingency of Constructivism: On Norms, the Social, and the Third

Transcription

1 655879MIL / Millennium: Journal of International StudiesKessler research-article2016 Article The Contingency of Constructivism: On Norms, the Social, and the Third Millennium: Journal of International Studies 2016, Vol. 45(1) The Author(s) 2016 Reprints and permissions: sagepub.co.uk/journalspermissions.nav DOI: / mil.sagepub.com Oliver Kessler University of Erfurt, Germany Abstract This article argues that constructivism has not engaged with the concept of contingency sufficiently. While such noted constructivists as Onuf, Kratochwil, and Wendt often refer to double contingency, it is the concept of norms rather than contingency that is used to characterise constructivist theorising in International Relations (IR). In this article, I outline how moderate and radical constructivists differ in their take on norms and thereby establish how the problem of contingency is actually at the core of constructivist theorising. The discussion then shows how Kratochwil, Onuf, and Wendt have made use of double contingency while moderate constructivists have re-introduced the single actor to show how norms cause action. The third part moves beyond the double contingency framework. By differentiating the social from society, this section shows that a third position can be identified. The concept of triple contingency then could be a way ahead for the theoretical discussion on constructivism itself. Keywords constructivism, contingency, triple contingency Introduction Constructivism is a strange animal in International Relations (IR). 1 On the one hand, it is considered to be one of the most important theoretical movements whose study of 1. I use strange in the same way as Nicholas Onuf, Worlds of Our Making: the Strange Career of Constructivism in International Relations, in Visions of International Relations: Assessing an Academic Field, ed. Donald J. Puchala (Columbia: University of South Caroline Press, 2002), Corresponding author: Oliver Kessler, University of Erfurt, Nordhäuser Str. 63, Erfurt 99089, Germany. oliver.kessler@uni-erfurt.de

2 44 Millennium: Journal of International Studies 45(1) norms, intersubjectivity, and the social has changed the conceptual apparatus of IR profoundly. 2 On the other hand, as a field of study, it is characterised by its fragmentation and internal diversity. 3 Ted Hopf, for example, distinguishes between conventional and critical constructivism; 4 Karin Fierke separates consistent (critical) from in-consistent (conventional) constructivism; 5 Stefano Guzzini and Knud Erik Joergensen both identify an epistemological and an ontological approach; 6 while others refer to moderate and radical modes of constructivist theorising. 7 It seems that norms, intersubjectivity, and the social can mean utterly different things and so it is far from clear what constructivism stands for. Its success seems to be based on an increasingly ambiguous conceptual apparatus, and hence it is no surprise that the core of constructivism is now difficult to characterise. 8 Of course, there appears to be a broad consensus that the moderate-conventional constructivism is associated with both Alexander Wendt s rump materialism and the liberal norms-based analyses by Audie Klotz, Martha Finnemore, Kathryn Sikkink and Thomas Risse, 9 while the radical-critical-epistemological stream 2. For constructivism and norms see Karin M. Fierke, Constructivism, in International Relations Theories: Discipline and Diversity, eds. Tim Dunne, Milja Kurki and Steve Smith (Oxford: Oxford University Press, 2010), Joshua Goldstein and Jon Pevehouse, International Relations (Boston: Pearson, 2011). See also Cynthia Weber, International Relations Theory: A Critical Introduction (London: Routledge, 2009), The only exception is Jeffrey Checkel in The Constructivist Turn in International Relations Theory, World Politics 50, no. 2 (1998): See also Ted Hopf, The Promise of Constructivism in International Relations Theory, International Security 23, no. 1 (1998): K. M. Fierke Constructivism, in International Relations Theories: Discipline and Diversity, eds. Tim Dunne, Milja Kurki and Steve Smith (Oxford: Oxford University Press, 2010), Knud Erik Jørgensen, International Relations Theory: A New Introduction (New York: Palgrave Macmillan, 2010), 163. Stefano Guzzini, A Reconstruction of Constructivism, European Journal of International Relations 6, no. 2 (2000): See Oliver Kessler, World Society and the Problem of Practices, Millennium Journal of International Studies 44, no. 2 (2016): This ambiguity might be due to the fact that differences refer to different inter-disciplinary projects and are never confined to IR s boundaries like in Realism and Liberalism where differences can be dealt with internally. I thank Timo Walter for insisting on this point. 9. See for example Emanuel Adler, Constructivism and International Relations, in Handbook of International Relations, eds. Walter Carlsnaes, Thomas Risse and Beth A. Simmons (London: SAGE Publications, 2002), ; Jeffrey Checkel, Norms, Institutions, and National Identity in Contemporary Europe, International Studies Quarterly 43, no. 1 (1987): ; Martha Finnemore and Kathryn Sikkink, International Norm Dynamics and Political Change, International Organization 52, no. 4: ; Alexander Wendt, The Agent-Structure Problem in International Relations Theory, International Organization 41, no. 3 (1987): ; Alexander Wendt, Social Theory of International Politics (Cambridge: Cambridge University Press, 1999).

3 Kessler 45 takes the work of Friedrich Kratochwil 10 and Nicholas Onuf 11 as its vantage point. However, I think there is something relevant to Onuf s concern over the danger of constructivism becoming a fad, and to Stefano Guzzini s identification of the rather ambiguous relation between rational and constructivist approaches when the latter is conceived as the middle ground. 12 I am certainly not interested in rehearsing these differences and the narratives around them as we might have heard them too often already. Yet, I do think that some of these confusions are due to the constructivists rather casual use of their core concepts of norms, the social, and intersubjectivity. Constructivists too often use these terms without specifying what they actually mean and how these concepts are supposed to work. How often is the term social construction (or the fancier version of the discursive construction ) referenced without any discussion of what this social is and how it works? How often do theoretical discussions stop by simply pointing to norms and intersubjectivity as if these concepts were self-explanatory? And how often do we read of society without any specification of whether a national, international, or world society is intended? To step beyond these confines, I do not want to propose better definitions, but a change of perspective. In contrast to the common entry through norms, I suggest to approach constructivism through the concept of contingency. While contingency is often referred to, it is not treated as a theoretical category in its own right. 13 Still, contingency can help us to move beyond the simple use of the social, intersubjectivity, or norms by providing a way through which it becomes possible to distinguish different uses of these terms alongside diverse forms of contingency. 10. See in particular Friedrich Kratochwil, Rules, Norms and Decisions: On the Conditions of Practical and Legal Reasoning in International Relations and Domestic Affairs (Cambridge: Cambridge University Press, 1989); and Friedrich Kratochwil, How Do Norms Matter?, in The Role of Law in International Politics: Essays in International Relations and in International Law, ed. Michael Byers (Oxford: Oxford University Press, 2000), Nicholas Onuf, World of Our Making: Rules and Rule in Social Theory and International Relations (London: Routledge, 1989). 12. See Nicholas Onuf, Theory Talk: Nicholas Onuf on the Evolution of Social Constructivsm, Turns in IR, and a Discipline of Our Making. Available at: theory-talk-70.html. Last accessed May 20, 2016; Stefano Guzzini, A Reconstruction of Constructivism in International Relations, European Journal of International Relations 6, no. 2 (2000): For recent moves towards that direction see Ty Solomon, The Politics of Subjectivity in American Foreign Policy Discourses (Ann Arbor: University of Michigan Press, 2015); Andrew A.G. Ross, Mixed Emotions: Beyond Fear and Hatred in International Relations (Chicago: University of Chicago Press, 2013); Brent Steele, Alternative Accountabilities in Global Politics: The Scars of Violence (London: Routledge, 2012). I thank an anonymous reviewer for pointing out these allies. Due to space constraints, a discussion of these brilliant contributions must wait for another occasion.

4 46 Millennium: Journal of International Studies 45(1) By forms of contingency I mean the differentiation into forms of single, double, and triple contingency, 14 which describe the number of available discursive positions one assumes: single contingency refers to the classic subject-object distinction; double contingency is widely used in sociological theory and is often referred to as intersubjectivity or interaction between ego and alter; triple contingency takes into account the position of the third who observes ego and alter. I certainly do not claim that these forms are my invention and I recognise that instantly, several approaches could at this point legitimately point out that they already have taken contingency in its forms and guises into account. 15 Yet while we may legitimately say that various approaches make use of different discursive positions already, constructivists may not have explored them fully. The objective of this article is thus rather modest: to clarify the different uses for core concepts among different constructivists along these three forms of contingency. In this vein, this article hopes to stimulate a debate on the uses of these forms, the conceptualisations of the social, and the consequences for understanding politics. 16 By putting contingency upfront, I want to advance the discussion in three areas: first, I want to show that the reference to norms is in itself not the defining move for constructivism. Instead, norms are a solution to the problem of contingency and they work rather differently in the contexts of single, double, and triple contingency. The sentence constructivism is the study of norms in itself is thus either highly problematic or meaningless. Second, I want to point out that the distinction of moderate and radical constructivism is parallel to single and double contingency: moderate constructivists consider only one side of the double contingency problematique with the consequence that the intersubjective and social is not located in the continuous re-processing of the two self/other complexes, but is inscribed into the concept of norms itself. Norms are said to be intersubjectively made and formed, even though they then manifest themselves in a quasi-objective fashion in front of single actors. Third, I propose to differentiate the social and society along the lines of double and triple contingency. Too often have constructivists interchangeably used 14. I do think that contingency and not norms should have become the key reference term by which constructivists are differentiated from positive approaches. Instead, we even had to witness how the study of norms was transformed into an empirical research programme framed in the vocabulary of variables, hypothesis testing, and necessary conditions. Ted Hopf, Reconstructing the Cold War: The Early Years (Oxford: Oxford University Press, 2012), 27. For a discussion on this point, see also Patrick Thaddeus Jackson, Conduct of Inquiry in International Relations: Philosophy of Science and Its Implications for the Study of World Politics (Abingdon: Routledge, 2010), One anonymous reviewer rightfully pointed this out. Also post-structuralist approaches could legitimately point to their use of thirds. One could think of the excluded third in Derrida or the Parasite in Serres. This only points to the need for the long overdue debate between constructivism and post-structuralist approaches, in particular Foucault s concept of power/ knowledge, Butler s schemes of intelligibility and Spivak s notion of hegemony. I thank the editors for pointing this out to me. 16. Even though this is a rather abstract formulation, I hope that I can specify this point below.

5 Kessler 47 the concepts of social, intersubjective, and society without much reflection. Here, I propose that the social raises the question of how inter-subjectivity can actually be established while society needs to take the third into account. This latter point raises new issues such as the institutionalisation of norms and the way they are discursively stabilised. That said, I think it is time to acknowledge that the social or the social construction of xyz has done its work. It is time to examine what the social actually is and what it does. This article is structured in three steps. The first step outlines the difference between single and double contingency and traces some of their uses in constructivism. The second section elaborates on this point and shows how moderate constructivists substitute double for single contingency situations with the consequence that positivist criteria of science are re-introduced. The third step then elaborates on the distinction of double and triple contingency and proposes to use the social and society as concepts parallel to them, rather than sublimated within them. Double Contingency: the Problem of Inter-subjectivity and Norms Constructivism is commonly understood through the concepts of norms and intersubjectivity. Yet the attempt to link any theoretical approach to a single concept is bound to fail, since every concept may be analysed from a variety of angles or perspectives. Thus there is no single theoretical core or identity that implies a concept presupposes a specific approach. Therefore, something other than just norms or intersubjectivity needs to ground constructivism. This section proposes instead to approach constructivism through the concept of contingency. To be more specific: I want to show that constructivists take as their vantage point not norms per se, but the problem of double contingency. To begin, this discussion requires a clarification of what single and double contingency actually mean. Single contingency is familiar to anyone acquainted with the notion of rational choice. It refers to the way in which one actor has to make an informed decision in the face of a contingent reality. 17 For example, the microeconomic literature on decisions under uncertainty has convincingly shown with basic standards of rationality in place that there are equilibria and best practices. What is more important at this point, is that the framework of single contingency is based on what Benjamin Herborth has called the quest for certainty with the objective to find adequate and right representations to aid decision making. Single contingency frameworks emphasise therefore methods and research design that control and guide the space between the individual and its (purportedly) objectively given reality This holds true for the classic decision under uncertainty literature as defined by expected utility theory. For game theory, see below. 18. Benjamin Herborth, Theorising Theorising: Critical Realism and the Quest for Certainty, Review of International Studies 38, no. 1 (2012): See also Oliver Kessler, The Failure of Failure: On Constructivism, the Limits of Critique, and the Socio-Political Economy of Economics, Millennium Journal of International Studies 44, no. 3 (2016): 348.

6 48 Millennium: Journal of International Studies 45(1) In contrast, the concept of double contingency describes a situation where ego and alter, as two actors, experience and realise their mutual contingency in relation to one another: not only is the behaviour of alter contingent to ego, but ego realises that its own behaviour is contingent on ego s expectations about alter. 19 In other words, both recognise that they are each equally cognisant of their behaviour as being contingent upon one another. 20 Double contingency does not only take into account that ego and alter are two actors, but that there are two subject-other complexes with the consequence that both ego and alter become aware of the contingency of their own position. As Talcott Parsons has summarised this problem: In interaction ego and alter are each objects of orientation for the other. The basic differences from orientations to nonsocial objects are two. First, since the outcome of ego s action (e.g. success in the attainment of a goal) is contingent on alter s reaction to what ego does, ego becomes oriented not only to alter s probable overt behavior but also to what ego interprets to be alter s expectations relative to ego s behavior, since ego expects that alter s expectations will influence alter s behavior. Second, in an integrated system, this orientation to the expectations of the other is reciprocal or complementary. 21 This double contingency creates a productive indeterminancy of action and raises further problems of how these expectations are to be coordinated, how expectations are formed, and how they are performed. This problem of double contingency became important in constructivist social theory: for example, both Jürgen Habermas and Niklas Luhmann, as two key sources of constructivism in IR, have made extensive use of the theorem of double contingency and used their critique on Talcott Parsons as a stepping stone to develop their own particular conceptual toolkit. Whether it is in the form of factual consequences of counter-factual validity claims in Habermas s Theory of Communicative Action, 22 or the temporalization of communication argument found in Luhmann s Social Systems, double contingency plays a major role in conceptualising how actors 19. For a good discussion see also: Raf Vanderstraeten, Talcott Parsons Luhmann and the Theorem of Double Contingency, Journal of Classical Sociology 2, no. 2 (2002): Even though Parsons did coin this term, it was Max Weber who used the mutual taking into account of expectations to separate social from other forms of action and hence invented this concept of the social. 20. In the following, I will concentrate on how the concept of double contingency became influential for constructivism through Parsons, Habermas and Luhmann. Of course, it would be equally legitimate to discuss it through Wittgenstein s concept of rule following, language games and the public language argument which equally presuppose the second actor. This would also show how speech act theory was developed as a response to Wittgenstein. In addition, it could also strengthen the link between double contingency and necessity and contingency where logic is related to rule following and language. 21. Talcott Parsons and Niels Shils, Towards a General Theory of Social Action: Theoretical Foundations for the Social Sciences (Cambridge: Harvard University Press, 1951), 105, quoted also in Vanderstraten, Parsons, Luhmann, and the Theorem of Double Contingency, Jürgen Habermas, Theorie des Kommunikativen Handelns (Frankfurt am Main: Suhrkamp, 1981), in particular Part VII.

7 Kessler 49 may relate to one another. 23 I mention this to highlight not only that much more work needs to be done to understand how different conceptualisations of double contingency relate to questions of normativity, the social, and critique, 24 but also to point out that with George Herbert Mead (through Parsons), Habermas and Luhmann, key theorists for constructivists in IR are already in place: in the work of Onuf and Kratochwil, Habermas and Luhmann are key reference points and the importance of Mead for Wendt is also well documented. The point is not to argue that one version of double contingency is better than the other or which theorist is right or wrong, but to emphasise that the concept of contingency is indeed at the core of constructivism. Let us now consider three instances where this double contingency is discussed within constructivism. Double Contingency in Constructivism Friedrich Kratochwil in his Rules, Norms, and Decisions assumes that action is a rulegoverned activity and that our social world escapes the world of observational facts (i.e. single contingency). 25 To reduce explanations to brute facts assumes that we can explain action on the basis of propositions alone. Yet our social world requires us to analyse illocution and perlocution and hence double contingency: for a successful speech act, we need the actor who utters the words, the actor who understands, and the shared conventions that guide that interaction. 26 The same emphasis on double contingency is present in Nick Onuf s conceptual framework of rules and rule. 27 Contingency is used as a counter-concept to positivism insofar as he emphasises the contingent co-constitution of human beings and societies: people make societies and societies make people. For Onuf, the concept of rules intermediates between the two sides and allows us to trace the constructions of social relations, in particular through the performative power of language. Through the concept of rules, Onuf traces certain functions a social order has to fulfil: in this context, rules have to 23. Niklas Luhmann, Soziale Systeme: Grundriß einer allgemeinen Theorie (Berlin: Suhrkamp Verlag, 1985), Parsons, Habermas and Luhmann entertain rather different ideas of how the social relate to normativity and critique. See for example Niklas Luhmann, Die Gesellschaft der Gesellschaft (Berlin: Suhrkamp Verlag, 1997), Here, double contingency has the properties of an autocatalytic factor: without itself being consumed, it enables the construction of structures on a new level of ordering, which is regulated by that perspective on perspectives. Thereby and this is why one can speak of auto catalysis the problem of double contingency is itself a component of the system that it forms. Thereby, Luhmann criticises the emphasis of the normative dimension for the understanding social order. 25. See in particular Kratochwil, Rules, Norms and Decisions, Ibid., 32. As discussed above, Nicholas Onuf also refers to speech act theory to explain the contemporary social order. As the concept of double contingency is inscribed in speech act theory (illocution and perlocution), I do not discuss Onuf at this point in further detail. 27. The concept of contingency features prominently in his Tainted by Contingency : Retelling the Story of International Law, in International Legal Theory: Essays and Engagements , ed. Nicholas Onuf (London: Routledge, 2008), I will nevertheless concentrate on his general writings on constructivism instead of reiterating this essay.

8 50 Millennium: Journal of International Studies 45(1) coordinate (1) naming and relating; (2) enabling and disabling; (3) having and using. Rules that structure the first function we call norms, rules that structure the second function we call commands, and the third function we call rights. These concepts are further developed through speech act theory. Building on Searle, Onuf shows how commissive, assertives and expressive speech acts relate to norms, commands, and rights respectively. How these speech acts are institutionalised then also determines rules and hence the kind of political society we live in. 28 Likewise, Alexander Wendt refers to the problem of double contingency when he discusses the original position : for example when Montezuma and the Spanish meet for the first time, it is only through their interaction that identities are formed and cultures emerge. 29 Even though, Wendt argues, that in situation of double contingency: Ego and Alter are not blank slates, and what they bring to their interaction will affect its evolution. They bring two kinds of baggage, material in the form of bodies and associated needs, and representational in the form of some a priori ideas about who they are. 30 At this point, we can also see some differences: Kratochwil is interested in how intersubjective meaning is created and how the structure of the exchange of reason(ing) is formed; Onuf is interested in the interplay of rules and rule; Alexander Wendt links the problem of double contingency to his rump materialism as the possibility to err still points to causal powers of nature as Wendt explains: Had Montezuma adopted this alternative representation of what the Spanish were, he might have prevented this outcome because that representation would have corresponded more to reality [ ] The external world to which we ostensibly lack access, in other words, often frustrates or penalises representations. 31 Double Contingency and the Study of Norms This brings us to the first area to which this article aims to contribute: norms are not the starting point of constructivism, but a specific solution to the problem of contingency. To illustrate this point, Friedrich Kratochwil s Rules, Norms and Decisions is a good place to begin. Right from the outset, he makes clear that our conventional understanding of social action and of the norms governing them is defective because of a fundamental misunderstanding of the function of language in social interaction, and because of the a positivist epistemology that treats norms as causes. Communication is therefore reduced to issues of describing facts properly, i.e. to match of concepts and objects, and to the ascertainment of nomological regularities. Important aspects 28. Onuf, World of Our Making, Through the recourse to speech act theory, the contingency Onuf observes is not that between people and society, but how rules fulfil certain functions in the interaction between speaker and hearer or ego and alter. 29. Alexander Wendt, Social Theory of International Politics, Ibid., Ibid., 89.

9 Kessler 51 of social action such as advising, demanding, apologizing, asserting, promising, etc., cannot be adequately understood thereby. 32 The abandonment of the mirror image of language and its consequences for understanding social action provides the rationale for the later chapters. Here, we can also see that Kratochwil is not interested in the validity of norms per se, rather he uses norms to touch upon questions of communication and language. 33 With this link of norms to language and the theory of action, Kratochwil differentiates himself from behavioural and positivist approaches and thereby creates the space for constructivism. 34 As he points out, normative statements that contain words such as ought and must can neither match the outer world, nor can they be reduced to the psychology of the speaker. Hence, neither the world of intention (understanding), nor the world of facts (explanation) is of any help. 35 Instead we have to resort to the world of institutional facts that only exist in and through the use of language. Therefore, for Kratochwil, we have to analyse the discursive structure and the exchange of reasons that create meaning inter-subjectively. Here, it is the contingent connectivity of sentences that allow us to understand why and how certain opinions become authoritative, as well as how reasons are formed and supported. This contingency is reconstructed through specific styles or modes of reasoning. 36 This subtle emphasis on contingency also has methodological repercussions as practical reasoning, styles, and the connectivity of sentences never follow the ideal of necessity. Hence, any hope placed on induction or deduction as scientific ideals miss the point as both presuppose a hierarchical relationship between sentences (which is contrary to the exchange of reasons) and in the end aspire to produce necessary knowledge. Instead, contingency needs to be aligned with our methods and methodology, 37 which has led Kratochwil to embrace ideas of conceptual history 38 and pragmatism. 39 Similarly, Onuf is not primarily interested in whether a specific norm 32. Kratochwil, Rules, Norms and Decisions, See also Kratochwil, How Do Norms Matter?, 45; Karin M. Fierke, Links Across the Abyss: Language and Logic in International Relations, in International Studies Quarterly 46, no. 3 (2002): Kratochwil, Rules, Norms and Decisions, As I will elaborate in the next section is it interesting to note that Kratochwil distances himself from both explaining and understanding as both fall back to the individual actor and exclude intersubjectivity. 36. Kratochwil, Rules, Norms and Decisions, 34, See also Friedrich Kratochwil and John Gerard Ruggie, State of the Art on the Art of the State, International Organization 40, no. 4 (1986): Friedrich Kratochwil, Sovereignty as Dominium ; Is there a Right of Humanitarian Intervention?, in Beyond Westphalia? National Sovereignty and International Intervention, eds. Michael Mastanduno and Gene Lyons (Baltimore: Johns Hopkins University Press, 1995), See also Stefano Guzzini, The Concept of Power: A Constructivist Analysis, Millennium Journal of International Studies 33, no. 3 (2005): In particular Guzzini emphasises that power implies a could have been otherwise which is merely a different formula for contingency. 39. Friedrich Kratochwil, Of False Promises and Safe Bets: A Plea for a Pragmatic Perspective in Theory Building, in Journal of International Relations and Development 10, no. 1 (2007): 1 15.

10 52 Millennium: Journal of International Studies 45(1) is valid or not, but how norms, commands, and rules relate to specific speech acts and through their institutionalisation create rule. 40 Neither Kratochwil nor Onuf are primarily interested in norms in themselves, rather they use norms to open IR up for social and political theory. Hence, their aim is not to test the validity of norms empirically, but to ask how through a critical engagement with speech act theory norms relate to contingency and thus to rule. 41 The concept of norm (or rights for that matter) 42 is never used for its own sake, but as a critique on the level of action theory through which IR is separated from political theory, history, and social theory. 43 The Loss of Double Contingency and the Question of the Social The last section outlined how constructivists share an interest in double contingency. It also showed that the common reference to norms per se as defining rationale of constructivism is more part of the problem than part of the solution. In this section, I want to explore the second area, i.e. that moderate constructivists have replaced double contingency with single contingency considerations. Here, two moves are particularly relevant: the agent structure debate and the empirical applications of norms. 44 The Agent-Structure Debate Let us first look briefly at the agent-structure debate as it unfolded between Alexander Wendt on the one side and Martin Hollis and Steve Smith on the other. Alexander Wendt famously opened this debate by accusing classical approaches in IR as taking either the individual or the structure as ontologically primary. For him, constructivism takes as its vantage point the co-constitution of actors and structures. In particular Hollis and Smith 40. Onuf, World of Our Making, Yet it is also here where Onuf and Kratochwil differ: Kratochwil focused on the discursive structures in the pragmatic use of speech acts and tries to move beyond speech act theory itself through an emphasis on their perlocutionary effect. Onuf built on John Searle s discussion on different types of rules to show how they perform certain functions in a social order. How this different take on speech act theory gives rise to different understandings of practice, norms and rights is still a question IR has not yet taken up, even though the books were published over 20 years ago. 42. See Kratochwil, Rules, Norms and Decisions, Also Kratochwil, The Status of Law in World Society (Cambridge: Cambridge University Press, 2014), I have not dealt with Alexander Wendt in great detail here. However, I do think that through his interest in the cultures of anarchy, he shares the interest in the contingency of norms as even the Hobbesian culture of anarchy is first of all a culture based on a set of norms. 44. Hence, I would not argue that Wendt misread the literature or that he did not take this double contingency seriously. I think the opposite is true. Yet I think that Wendt simply argued his case when faced with criticism. The subsequent departure appears to me to result from the dynamics of the debate itself.

11 Kessler 53 have taken Wendt to task for treating this as a predominantly ontological problem and I do think they are right. Yet when Hollis and Smith proposed explaining and understanding as two general ways of considering philosophies of science and IR theories, in order to show that epistemology is prior to Wendt s ontology, they unintentionally opened the gateway to remove double contingency from consideration: both explaining and understanding are less interested in the formation of symbols, or of an inter-subjective meaning system on the basis of which ego and alter can bridge the space between them, than whether explaining and understanding are co-extensive with causal and constitutive questions, respectively. In other words, the fluidity of social relations and the contingency inherent within them, disappears. Wendt characteristically refers to the problem when he argues that Norms are causal insofar as they regulate behavior. Reasons are causes to the extent that they provide motivation and energy for action. And so on. All of these phenomena involve rules and self-understandings ( ideas ), but this does not preclude their having causal effects. 45 In this quote, symptomatically, the key reference point is not inter-subjectivity, but behaviour by individual actors. In the end, this whether reasons can be causes? debate was translated to IR from the analytic philosophy of mind. The reasons and causes considered are those of an individual actor and the question is whether reason (understanding) can be considered as a cause. In this sense, that Wendt s famous cultures of anarchy are formulated along different legal theories that equally focus on individual reasons/ causes to obey a norm might be a case in point: power (Hobbes), interests (Locke), or legitimacy (Kant), are each subtly grounded upon individual decision making. Although this change of terms was eventually legitimised by the attribution of explaining with third person and understanding with first person perspectives, what holds the third and the first person together remains the individual subject. The difference between first and third person then is only whether we explain behaviour from the outside or from the inside, yet the object of analysis, i.e. the individual act, is shared. What got lost on the way is the you. In the end, it is the agent (!) structure problem. Norms in Constructivist Research As a second move, the norms literature is illuminating at this point. In particular, prominent scholars such as Martha Finnemore, Audie Klotz, Kathryn Sikkink, and Thomas Risse have advanced the constructivist agenda on norms and have certainly had some impact on many of these discussions. Constitutive for their analyses is their attempt to build bridges and to produce a research programme that stands firm to Keohane s early critique on what he called reflectivism. 46 This literature above all sought to establish constructivism vis-à-vis realism and liberalism: whereas the other approaches 45. Wendt, Social Theory of International Politics. 46. Robert Keohane, International Institutions: Two Approaches, International Studies Quarterly 32, no. 4 (1988): See also Martha Finnemore and Kathryn Sikkink, Taking Stock: the Constructivist Research Program in International Relations and Comparative Politics, Annual Review of Political Science 4 (2001):

12 54 Millennium: Journal of International Studies 45(1) would point to interests and power, constructivists could point to the force of norms. Whereas the former focus on states, constructivists can highlight the importance of nonstate actors and so-called transnational advocacy networks that facilitate the diffusion and implementation of norms in different issue-areas. 47 These networks attempt to tie different actors (civil societies, states, international organisations) together at various levels (domestic, international transnational) 48 and thereby help to explain why states comply with norms. 49 Even though the literature evidently showed that norms matter, it also changed the analytical framework of research in three respects: first, the literature is ultimately interested in why states comply. With this focus on compliance, this literature departs from the double contingency problematique. Characteristically, the two models that flourish within constructivism, that is, the norm life-cylce model 50 and the spiral model, 51 both emphasise the socialisation of individual actors through which (individual) interests and identities are changed. In the end, the decision to obey or to deviate remains with the individual state. In this sense, it is interesting to note that in this literature Habermas s interest into the general rationality of action is transformed into individual rationalities of action, such as the logic of consequence and the logic of appropriateness. Both remain, however, firmly rooted within the confines of individual choice. Characteristically for this literature, Ian Hurd has also defined an apparent inter-subjective concept like legitimacy as the normative belief by an actor that a rule or institution ought to be obeyed. It is a subjective quality, relational between actor and institution, and defined by the actor s perception of the institution. The actor s perception may come from the substance of the rule or from the procedure or source by which it was constituted. Such a perception affects behavior because it is internalized by the actor and helps to define how the actor sees its interests. 52 This quote nicely shows that the social enters the conceptual framework only through endogenous preferences. At the same time, the mechanisms of the social works 47. See Filipe dos Reis and Oliver Kessler, Constructivism and the Politics of International Law, in The Oxford Handbook of the Theory of International Law, eds. Florian Hoffmann and Anne Orford (Oxford: Oxford University Press, 2016), for a longer.discussion of these models. Here, authors such as Martha Finnemore, Margaret Keck, Thomas Risse, Stephen C. Ropp, Kathryn Sikkink and others have done valuable research. 48. Margaret E. Keck and Kathryn Sikkink, Activists beyond Border. Advocacy Networks in International Politics (Ithaca: Cornell University Press, 1998). See also Sanjeev Khagram, James V. Riker and Kathryn Sikkink, eds., Restructuring World Politics: Transnational Social Movements, Networks, and Norms (Social Movements, Protest, and Contention) (Minneapolis: University of Minnesota Press, 2002). 49. Martha Finnemore and Kathrin Sikkink, International Norm Dynamics and Political Change ; Thomas Risse, Stephen C. Ropp and Kathryn Sikkink, eds., The Power of Human Rights. International Norms and Domestic Change, Introduction (Cambridge: Cambridge University Press, 1999). 50. Finnemore and Sikkink, International Norm Dynamics and Political Change, Risse, Ropp and Sikkink, The Power of Human Rights. 52. Ian Hurd, After Anarchy: Legitimacy and Power in the United Nations Security Council (Princeton: Princeton University Press, 2007), 8.

13 Kessler 55 differently now. Remember that in the context of double contingency, norms allow ego and alter to position themselves vis-à-vis each other and to cross the space between them. Norms are produced through the continuous interaction (or communication) between these actors and thereby also define who the actors are and what rights and obligations certain moves in the language game entail. Yet within the liberal norms literature, norms enter through the formation and change of individual preferences and identities. This leads, as Sterling-Folker has perceptively pointed out, to functionalist histories implying that individual norms are valid because they are now conformed to. Meanwhile, however, norms present themselves as a quasi-objective force in front of states, thereby forcing them to make a decision of some sort. 53 Third, and true to the single contingency framework, the literature embraces once again the quest for certainty with its focus on method and research design. The objective of these analyses (and increasingly so) is to produce necessary and scientific knowledge, i.e. to provide criteria and to identify mechanisms that are subject to falsifiable hypothesis tests. For example, Audie Klotz is convinced that the first generation of constructivists like Kratochwil and Onuf did not offer a specific theory to test. 54 Also Ted Hopf follows the scientific path with falsifiable hypotheses and variables that vary. 55 The transformation of norms into an empirically driven research programme embraces a positivist concept of science through which analyses are evaluated and judged. The methodological consequences that Kratochwil and Onuf drew from contingency, i.e. that contingency has to be mirrored and taken seriously even at the level of methodology are abandoned in the process. 56 In summary, moderate constructivists have, on the level of methodology, replaced contingency by necessity, and instead engaged in a quest to produce necessary knowledge. As a consequence, interest in the special kind of contingency that norms give rise to is not continued. This also has repercussions for the concept of politics and authority. For radical constructivists, politics and authority are linked to ways of world making, as Onuf s famous title highlights: 57 the production and fixation of signs through which reasons can be given, political projects formulated and inter-subjectivity fixed. Here, norms opened the gate to embracing a different form of social theory in IR where the international could be conceptualised in social terms, thereby bridging IR with social and political theory. To make visible how our world was made, (radical) constructivism unearthed the natural and the given, problematised its apparent reality, and then made visible again what was previously made invisible. For moderate constructivists, in contrast, politics is co-extensive with the validity of norms, which is automatically linked to compliance and thus to state behaviour. It tests individual norms empirically and captures them in terms of hypotheses, 53. Jennifer Sterlin-Folker, Competing Paradigms or Birds of a Feather? Constructivism and Neoliberal Institutionalism Compared, International Studies Quarterly 44, no. 1 (2000): Audie Klotz, Case Selection, in Qualitative Methods in International Relations: A Pluralist Guide, eds. Audie Klotz and Deepra Prakash (London: Palgrave, 2008), Ted Hopf, Reconstructing the Cold War. 56. On this point see also Patrick Thaddeus Jackson, Conduct of Inquiry in International Relations. 57. Onuf, World of Our Making.

14 56 Millennium: Journal of International Studies 45(1) linear causality, trigger points, and sufficient conditions all of which presume the generality and de-contextuality of scientific research. From the Social to Society: the Problem of Triple Contingency Thus far, it has been argued that the problem of double contingency is central for constructivism. Double contingency points to the pragmatic use of language where, as Kratochwil elaborates, [c]ommon understandings can be arrived at through the stabilization and evocation of certain generally shared expectations in a specific situation. 58 On the one hand, this raises the question of how claims to the validity of norms (and not the objective force of norms per se as moderate constructivists claim) is what actually guides the reasoning process between ego and alter. 59 At the same time, this quote by Kratochwil raises another conceptual issue that, to my mind, constructivists have not yet taken up: For instance, Kratochwil speaks of a given situation. 60 Apart from the reasoning processes involved, something else plays a role here the being in a situation. Later, Kratochwil links practical reasoning with topoi from the rhetorical tradition. Topoi are the commonplaces for practical reasoning, 61 the seats that not only determine the starting point of argumentation, but which are also decisive for attaining assent to choices. Yet, when Kratochwil points to topoi, I think he actually steps beyond the confines of just double contingency, into something else. As he later elaborates, This topos expresses some shared interpretation of actions on the basis of certain practical experiences. Such a topos is therefore a shared judgement in a society that enables the respective actors to back their choices by means of accepted beliefs, rules of preference, or general classification schemes [.] 62 Two issues are remarkable here: first, Kratochwil speaks about society. Leaving aside that it is not clear which kind of society we are talking about here, the use of society instead of the social (as in social action) highlights another issue here at stake, other than mere intersubjectivity. 63 Secondly, the quote mentions shared judgment and accepted beliefs. In the following pages, Kratochwil then discusses how judgement also relates to argumentative starting points, the structure of argumentation, the assembling of evidence, and the arranging of the material. Here, law serves as one example, but not as the grounds for his argument. 58. Kratochwil, Rules, Norms and Decisions, Ibid., Situation is also an often used reference in Kratochwil, Status of Law in World Society. It seems to me that this is an undertheorized category. 61. Kratochwil, Rules, Norms, and Decisions, Kratochwil, Rules, Norms and Decisions, Unfortunately, this holds also for Wendt s and Onuf s common conviction that man makes society and society makes man.

15 Kessler 57 I think the question of situation, general beliefs, shared understandings etc. presupposes institutionalisation without which it is simply not possible to move from ego and alter to the judge, the court, the defendant, and any other position. Indicatively, in his recent book, The Status of International Law in World Society, Kratochwil again points out that institutional trust cannot be derived from the interaction between actors. There is a general, institutional expectation that differs from the expectation vis-à-vis the other, as he confirms: If it is true that law differs from expectations emerging out of interactions, and that it differs even from expectations about expectations since differences among these secondary expectations must be adjudicated by a court it is also true that courts are again bound by (much looser) expectations of how such conflicts are to be settled. Clearly, the development of such tertiary expectations, which are supposed to set the tide of forum shopping or endless litigation, certain rules of regulating jurisdictional competences and discretion have to exist. 64 Similarly, when Onuf asks how institutionalization is possible or when he argues that norms need to be institutionalised in order to become legal, there is also another story involved: in order to become institutionalised, norms need to appear as stable, meaning that ego and alter cannot do as they please and substitute one norm for another. There is a distinction to be made here between expectations on institutions or institutional trust and the stabilisation of mere interactions. This raises the question of the third that the next section introduces. The Third and Social Imaginaries Even though I cannot provide a full account of institutionalisation at this point, it is hopefully sufficient to see that this question of expectation and interaction points to the necessity of a third position or the acknowledgement of the need for a triple contingency: that ego and alter are observed by a third person. The question that now arises is, how do the processes of structure formation differ when ego and alter realise that they are being observed by the third? The interaction of ego and alter does not take place in a void, but already in a situation defined by the third. From this perspective, the third position becomes relevant as soon as we move from the social (i.e. the two actors mutually taking each other into account, including questions of recognition and alterity) to institutionalisation and thus the creation of social orders or societal structures. The social and society thus point to different sets of problems where the social can be associated with the processes of double contingency as outlined above. Here, interactions are social facts indeed. However, something else takes place when we talk about society, international society, or world society, and hence there is a change from social interactions to the question of societal order-formation and/or institutionalisation through which specific situations, roles, and positions emerge. Unfortunately, it is also beyond the scope of this article to provide a full historical account on the third as a theoretical category. It is only surprising that the history of the 64. Kratochwil, Status of Law in World Society, 95.