Figurational thinking and first image theorizing in IR

Transcription

1 Paper for presentation at the ECPR Joint Sessions 2011 Workshop 14 'Bringing the Individual Back In' International Relations and the First Image First draft - Please do not quote Abstract The problem of theorizing the individual actor is strongly tied to the agencystructure problem. I propose that none of the major theoretical solutions to this problem have successfully conceptualized agency, due to the ontological assumptions of the initial problem formulation. If we examine scientific realist accounts, we find that they are based on a Cartesian subject. Here a core of agency escapes socialization, irreconcilably separating agency from structure. If we examine structuration theory, we find that the concept of 'knowledgeability' is rooted in a bodily and unconscious search for ontological security at an individual level. Both schools of thought remain trapped in an agency-structure dichotomy, and thereby squander the opportunity to fruitfully theorize agency and the role of individuals in shaping social reality. Post-structuralist perspectives, although they theorize the individual in non-essentialist terms are not particularly well suited to underscore the importance of individual action in IR either, as they focus on discourse broadly understood. I consequently suggest to side-step the unproductive agency-structure debate by taking a processual-relational approach that places 'figurations' at the center of inquiry. The proposed shift implies that political scientists move beyond thinking about how the (conceptually) isolated individual shapes international relations. Hence, the model of agential individuals, developed in this paper, suggests to think about actors in terms of figurationally embedded individuals (themselves understood as ongoing processes). This conceptual shift has implications for the way we think about first and second image theorizing, because a figurational logic blurs the boundary between part and whole underlying this categorization. Bernd Bucher Rosenbergstrasse St. Gallen bernd.bucher 'at' unisg.ch

2 1. Introduction - the lack of acting persons in social constructivist thinking Social constructivists have had a difficult time in incorporating first image theorizing. 1 But while the eclipse of agency has been widely documented (Checkel, 1998; Palan, 2000; Shannon, 2000; Jacobsen, 2003), the theoretical preconditions of the difficulties of incorporating agency in social constructivist scholarship have not received extended attention. I will argue in the following that social constructivists have not yet formulated a model agency to successfully underscore the importance of acting persons in International Relations (IR). This is so because both scientific realist and structurationist approaches (although they strongly differ in general outlook) take a substantialist starting point in the agency-structure dichotomy. Such a starting point encourages the introduction of a presocial (essentialist) subject at some point in the theory building process. This, on my reading, is problematic from a social constructivist perspective. Unfortunately poststructuralist approaches do not provide the sought-after theoretical concepts either. I will consequently focus on an alternative approach that promises to be able to more strongly focus on the actions of individuals in IR. In short, I propose to pay tribute to the importance of actually acting persons by modeling persons as processes within a figurational approach. In order to spell this notion out, I will draw upon the sociology of Norbert Elias, especially the shift from homo clausus to homines aperti. Figurational approaches consider the individual and society to be two aspects of one and the same process. Neither persons, nor the situations in which they act can be entirely separated from each other. Although processual-relational / figurational approaches focus on constitutive interdependencies, they are also able to conceptually grasp what persons do when they act. The individual in this sense does not entirely disappear from analysis. Quite to the contrary, figurational thinking is able to account for the importance of individauls in world politics without essentializing them. In order to formulate a heuristic with which to grasp what agents do when they act, I will very briefly introduce a slightly modified version of Mustafa Emirbayer and Ann Mische's concept of agency. The argumentation of this paper has implications for the notion of (first, second, third) image theorizing in IR (see Waltz, 1959; Singer, 1961), because it blurs the boundaries of the image logical underlying the analytic distinction between first and second image thinking. As such, figurational thinking can incorporate the importance of individuals because they are not conceptualized as a counterpart to structure. Scholars do not have to decide between first or second image theorizing, because both levels are intimately connected. I will develop my general argument in four major steps. Because the argument is based on the premise that social constructivism needs to formulate a non-essentialist notion of 1 Parts of this paper have been presented at the 52 nd annual ISA Conference in Montreal,

3 agency, I will in a first step explicate this statement. I will keep these remarks to a minimum in order to create space in which to engage the more pressing questions regarding the importance of individuals in IR. In a second step I will outline why scientific realist and structurationist approaches (formulated in the agency-structure debate) conceptually tend to posit pre-social and essentialized agents. 2 Both approaches are faced with choosing between the primacy of first or second image theorizing (Singer, 1961). In both instances agency and structure are closely related. I will therefore briefly comment on their respective understandings of structure, before more narrowly focusing on their conceptualizations of agency. Before moving on, I will discuss post-structural thinking in regard to agency. I will do so in passing, not because post-structuralist thinking does not have a concept of agency worth considering, but rather because the practical focus on discourse makes post-structural approaches an unlikely candidate to more strongly focus on the importance of individuals in IR. Following this discussion I will (in the third section of this paper) outline how the importance of acting persons can be taken seriously in social constructivist IR theory. In short, I propose to utilize a figurational approach developed by Norbert Elias. This shift in perspective enables scholars to incorporate the actions of persons in a non-essentialist fashion. Such a move, at the same time, suggests that the dichotomy between first and second image theorizing in IR is not as stark as is usually assumed. While figurationally embedded persons (in the plural), understood as processes, lie at the heart of the theoretical outline of this paper, it is also important to be able to grasp individual actions. Building on a figurational approach, I suggest to make the acts of persons visible through employing the concept of agency developed by Emirbayer and Mische (Emirbayer and Mische, 1998), in the fourth step of this analysis. In closing this paper, I briefly summarize my argument and reflect on its implications for the question at hand. 2. A very brief sketch of a non-essentialist reading of social constructivist thinking Although the importance of first image theorizing is deeply rooted in the notion that agents are involved in the ongoing construction of social reality, it has not been possible to remedy the lack of agential theorizing identified in the late 1990's (see Checkel, 1998). The inability of social constructivist approaches to broadly account for the role of individuals in 2 A logic of appropriateness approach amounts to a structural explanation of agency, and as such "it is untenable as a theory of individual action" (Sending, 2002: 443; see Checkel, 2001: 559; Goldmann, 2005, for a dissenting view see Mueller, 2004). I omit the analysis of the logic of appropriateness from the present analysis. 3

4 the ongoing development of international relations is especially troubling because the importance of agency is an outflow of basic ontological assumptions (see for instance: Ashley, 1986; Guzzini, 2000; Kratochwil, 1989; Kratochwil, 2008; Kratochwil and Friedrichs, 2009, Onuf, 1989; Wendt, 1987; Wendt, 1992). The practice of privileging structure over agency, and often eclipsing the latter altogether (see Jacobsen, 2003; Checkel, 1998; Risse, 2007; Shannon, 2000), has deprived social constructivist theorizing of the ability to understand the ongoing development of the social world, because the "agents who help create and change [social settings] in the first place" (Checkel, 1998: 325) drop out of the picture. The importance of underscoring the role of individuals and their actions is indeed a pressing and important issue for social constructivists (see for instance Checkel, 1998: 340; Shannon, 2000: 311). But clearly not any account of agency will do. Quite clearly, any formulation of agency must be compatible with the basic underlying (ontological) assumptions of the research approach. This, on my reading, implies that agency must be conceptualized in a non-essentialist way, because positing the self as a self-evident or secure foundation of knowledge production reproduces the problems associated with rationalist philosophy. 3 Following the linguistic turn, I maintain that the plausibility of taking the individual as a secure foundation from which to start thinking about the social world, has steadily decreased. From a social constructivist perspective language, practice and "agency [matter] in social life" (Kratochwil, 2008: 86), for both the construction of meanings/knowledge and social reality (Guzzini, 2000: 149). As such it is not possible to burden the mind-independent world with adjudicating scholarly disputes. To be concise, the social world and hence international relations, is a 'world entirely of our (unintentional) making' (see Onuf, 1989). This leaves no room for 'naturalized' starting points of inquiry. Social constructivism should therefore be understood in non-essentialist terms. This finding, in practice, creates problems for the attempt to move individuals closer to the center-stage of social constructivist theorizing. This is so because the main alternative conceptualizations of agency rely on a pre-social and essentialized agent in theorizing the relationship between agency and structure. In the following I will discuss to what degree the most prominent attempts to take agency seriously in social constructivist research fail to serve as a model with which to underscore the importance of individuals in IR. I will begin with a discussion of a scientific realist account, before turning to an analysis of agency in structuration theory. In a final step I will briefly comment on post-structuralist understandings of agency and outline why this line of thinking is not optimally suited to tackle the problem at hand. 3 This understanding of social constructivism is one among many alternatives. Other conceptualizations are not only possible, but even highly welcome (for a survey of constructivist thinking see: Ruggie, 1998; Adler, 2002, Guzzini, 2000; Hopf, 1998). 4

5 3. Agency in scientific realism, structuration theory and post-structuralist thinking a) Agency in Colin Wight's scientific realism Colin Wight (Wight, 2006) has recently provided an impressive conceptualization of agency. Although I consider this contribution to be very valuable, I argue nonetheless that Wight's understanding of agency is deeply problematic from a non-essentialist perspective, because it rests on a particular version of the Cartesian cogito. As such, Wight's agency model is not uncritically suited to serve as an underlying framework to emphasize the importance of individuals in IR. In order to substantiate this claim I will first briefly comment on Wight's conceptualization of structure, before discussing his notion of agency in more detail. This is an important preliminary step in that it makes visible how agency and structure are conceptually connected in the first place. Structures in this view are conceptualized as 'unobservables' (Wight, 2006: 121), which are only discernable through their independent causal effects. Structures are ontologically accounted for through (material) effects which are independent from agential instantiation or interpretation (see Wight, 2006: 159). In a nutshell, they are independent from agency. As such, there is "an ontological hiatus between society and people" (Bhaskar, 1979: 44) that requires us to distinguish between the "genesis of human actions, lying in the reasons, intentions and plans of people, on the one hand, and the structures governing the reproduction and transformation of social activities, on the other" (Bhaskar, 1979: 45). 4 But this ontological distinctness does not imply that agency and structure are entirely unrelated. Quite to the contrary, scientific realists conceptualize the simultaneous separateness and connectedness of agency and structure in terms of emergent properties. 5 A more comprehensive discussion would certainly have to take aim at this conceptual link. But as the argumentation at hand aims to pin-point problematic aspects in the articulation of agency, it has to suffice here to have hinted at the concept of emergence connecting distinct ontological levels. Returning to the conceptualization of structure, and making things short, structure is understood as "causal factors in the social world independent of agential understanding" (Wight, 2006: 154). In Wight's view, actors' understandings do not figure prominently into the notion of structure. Actors' interpretations are of secondary importance, as the (given) materiality of the social world places limits on the interpretations persons can reasonably and successfully entertain about the world. On this account material 4 This (ontological) separateness is decisive, because it serves as the condition of possibility to theorize the causal relationship between agency and structure. 5 "Emergence means that although the more complex levels of reality, for example societies, presuppose the more basic or less complex levels, for example, people, explanations of them are not reducible to the other" (Wight, 2006: 37). 5

6 referent(s) tie "differential meanings into a complex whole" (Wight, 2006: 159). 6 Wight's perspective stresses the centrality of actual relations in contrast to 'mere' interpretations In short, Wight underscores the existence of objective social conditions which are grounded in the materiality of the social world. 7 The understanding of structure outlined above emphasizes that structure and agency are two ontologically distinct things which stand apart from each other. Given this contextualizing outline of structure, I will now turn to a critical discussion of Wight's model of agency. In a nutshell, Wight differentiates between three levels of agency (agency 1/2/3 ), and I will in the following briefly outline what these different levels signal. I will primarily focus on agency 1 not only because it is central to the theoretical enterprise under scrutiny, but also because it is the most problematic level of agency in the narrative. 8 Concisely put, agency 1 refers to the things that people are capable (or incapable) of doing qua being human (see Archer, 1995: 255). It is, at base, conceptualized as "the 'self' which [stands] in relationship to the world by which it is [continually being] constructed" (Wight, 2006: 210). Agency 1, in this sense, is a logical prerequisite for socialization, not an empirically discernable aspect of agency. Wight here refers to the physical constitution of human beings as a condition of possibility for human actions and sociability. In the final analysis agency 1 (understood in terms of 'species being') is grounded in the materiality of the human body and serves to 'earth social life' (see Bhaskar, 1993: 164). 9 Wight at the same time stresses that the ability to act into the social world depends on one's social position (Wight, 2006: 212). The capacity to act 'is a function' both of specie's attributes and social positioning. The second level of agency (agency 2 ) then "refers to the way in which agency 1 becomes an agent of the socio-cultural system" (Wight, 2006: 213). Agency 2, thus understood, refers to how an agent can be an 'agent of something' (see Archer, 1995: 257). Agency 3 finally refers to the different positions within social structures that can be occupied by agents 1. These 'positioned-practice places' (such as peasant or party candidate) exist apart from the specific persons occupying them (see Wight, 2006: 6 This argument implies that in order to be able to reasonably speak about differing interpretations (of something), these differing interpretations must indeed refer to the same referent (for a critique see Hollis and Smith, 1991: 407). 7 To be fair, Wight does not reduce structure to materiality, but rather attempts to "incorporate the important dialectic between material and ideational factors" (Wight, 2006: 174). Wight further disaggregates material and ideational factors into four interrelated planes of activity. Roughly: material transactions with nature, inter-intra-subjective actions, social relations like class, group identity, or production, the subjectivity of the agent. 8 The concept of agency presented by Wight predominantly draws on Margaret Archer's "stratified model of people" (Archer, 1995). 9 Wight defines agency 1 as "embodied, intentional causality, or praxis" (Bhaskar, 1994: 100). 6

7 213). Positioned practices provide specific possibilities to actualize potentialities. "Agency 3 [then] refers to those 'roles' that agents 1 play for agency 2 " (Wight, 2006: 213). Within this account, Wight presents agents as structured entities. At the same time, the agent is based on monadic center (see Wight, 2006: 206), which serves "as an a prioristic anchorage" (Archer, 1995: 281). From the social constructivist perspective sketched-out earlier, this is a problematic theoretical conceptualization, because the pre-social, essentialized, human being "remains the alpha and omega" (Archer, 1995: 281) of Wight's model of agency. In other words, Wight's social actor is conceptually dependent on a type of pre-social Cartesian self. 10 Wight stresses that despite the importance of socialization, classification and signification, these processes "all-too-obvious[ly] [ ] still presuppose someone" (Wight, 2006: 207). But this argument is has its difficulties, if only, because signification does not presuppose a specific ontology of the signified. 11 Phrased differently, signification here is equated with subjectivation and this equivocation is not prima facie necessary, arguably not even plausible. To be precise, Wight does not negate socialization. He stresses that the person 'as a social agent' is structured by language and embedded in culture (agency 2/3 ). Wight freely admits that agents are not simply autonomous subjects but that they "navigate [their] social field through a set of structural formations and a network of heterogeneous discourses, which are not of [their] own creation" (Wight, 1999: 131/32). But the same is not also true in the case of agency 1. Wight, expanding on the statement quoted above, continues to argue: Yet, still, I think to myself, who is it who is cognizant of this? Is this belief in my constructed status itself a constructed effect? And is the belief that I should question the constructed certainty of my constructed status itself a constructed effect. I am now trapped in a viscous regress. In order to think the impossibility of my subjectivity I must presume the possibility of my subjectivity. In effect, there is a 'self' who is thinking about the relationship between the discourse and that same fragmented 'subject' (Wight, 1999: 131/32, emphasis added). Wight's entire conception rests on the possibility of subjectivity that is prior to language and culture. As such, he continually underscores that all social processes which involve agents are conditioned on the logical prerequisite of an 'I' that is prior to these processes (see Wight, 2006: 209). Conversely, the social world is outside of the 'I' - not empirically, but 10 The Cartesian ego is a claim to secure and trans-historic knowledge, and as such it functions in analogy to Wight's agency 1. But, to be fair, there are important differences at the same time. The Cartesian ego is not an embodied 'I'. Wight therefore does not commit to a distinction between res cogitans and res extensa. Wight hence does not opt for a dualistic position argued by Descartes. 11 The signification of Santa Clause, for instance, does not presuppose the materiality of the signified (see Kratochwil, 2007; see also Michel, 2009). 7

8 conceptually. 12 The 'I' at the center of Wight's account then supposedly escapes social construction. 13 Such a move opens the door for a foundationalist account of the social world. But the "claim, as in the Cartesian cogito, that the 'I' resists radical doubt because it is present to itself in the act of thinking or doubting is [an] appeal to presence" (Culler, 1983: 94). As such Wight's account of agency is subject to post-structuralist critiques of rationalist philosophy. At a general level, Wight tends to neglect the insights of contemporary language philosophy. This is visible in his equivocation of grammatical habits and ontological necessity. 14 Such a theoretical move presupposes that language is a neutral descriptive instrument which correctly depicts the world. But to the degree that social constructivists do not accept the language independence of knowledge, and take the performative dimension of language seriously, his argument does not hold. Once language becomes the object of critical reflection, the certainty with which the 'I' is postulated in this account begins to dissolve (also see Jackson, 2011). Additionally, Wight also disregards Meads insight that the self can alternatively be understood as continually arising social emergents (Mead, 2002: ). To be concise, in Meads view, the self emerges out of a conversation of specific gestures termed 'significant symbols'. 15 In such an exchange (conversation) he or she can take the role of the other. Consequently one is able to look back at oneself, and respond to oneself, from the point of view of the other. Or to put it differently, one can 'become an object to oneself' (see Morris, 1966: xxiv). 16 As such, symbolic interactionists view the self as continually arising out of a relational communicative process. They consequently do not posit a pre-social self which serves as a condition of possibility of socialization. 12 This is a necessary outflow of the notion that agency and structure stand in a causal relationship. 13 Wight argues against the social construction of the subject 'all the way down', also see Wendt's 'rump materialism' (Wendt, 1999). 14 Friedrich Nietzsche's criticism in 1886 already pointed to the central role of grammar and its implications for conceptualizing the social world, in arguing: "that [contrary to Cartesian philosophy] a thought comes when 'it' wants, and not when 'I' want. It is, therefore, a falsification of the facts to say that the subject 'I' is the condition of the predicate 'think.' It thinks: but to say the 'it' is just the famous old 'I' - well that is just an assumption or opinion, to put it mildly, and by no means an 'immediate certainty.' In fact, there is already too much packed into the 'it thinks': even the 'it' contains an interpretation of the process, and does not belong to the process itself. People are following grammatical habits here in drawing conclusions, reasoning that 'thinking is an activity, behind every activity something is active, therefore -.'Following the same basic scheme, the older atomism looked behind every 'force' that produces effects for that little lump of matter in which the force resides, and out of which the effects are produced, which is to say: the atom" (Nietzsche, 2007: 17-18). 15 Significant symbols are symbols which an individual not only sends, but also responds to respectively. 16 This is taken to be the defining characteristic of the 'self'. 8

9 To make things short, both language philosophy and symbolic interactionism strongly draw the persuasiveness of Wight's account into doubt. It, in the final analysis, centers on an essentialist notion of agency and is therefore not easily compatible with non-essentialist social constructivism. It thus can hardly function as a model of agency to highlight the importance of acting individuals in social constructivist thinking. b) Agency in Anthony Giddens structuration theory Having now very briefly discussed the major difficulties of Wight's account of agency, I turn to Anthony Giddens' structuration theory in which he pictures agency and structure as two sides of the same coin (mutual constitution). I will proceed by explicating the relationship between systems and structures, before examining the stratification model of agency in more detail. In a final step, I will critically reflect upon the approach and argue that structuration theory also has problems that warrant a search for other conceptions. Social Systems are thought of as "the stable patterns that give [ ] order to interactions" (Calhoun, Gerteis and Moody, 2007: 221). They are simply visible interaction patterns, or institutionalized features (see Giddens, 1984: 24, see also Bryant and Jary, 1991: 7). In this sense, social systems (which display structural properties) are external to the agent and (possibly) confront agents as objective states of affairs. 17 Giddens' notion of social system is therefore not unlike classical understandings of social structure. 18 Structures, clearly in contrast to Wight's account, are conceptualized as internal to the agent. Conceived of as rules and resources, they serve as common interpretive schemes within specific societies. Structures therefore "exist only as memory traces [which are] instantiated in action" (Giddens, 1984: 377), and have no independent ontological status apart from agents and their practices (see Giddens, 1984: 26). 19 To the degree then that structures are internal to agents, one can reasonably speak of their 'virtual existence' (Giddens, 1984: 25). Structure in this sense cannot be said to condition action. If this were the case, structure would have to be external to the agent to start with. This is decisively not the case. Rather agency and structure are inseparably tied together in that "agency is given meaningful form only through the 'generative schemes' of structure" (Calhoun, Gerteis and Moody, 2007: 17 Structural properties include political, economic, and legal (symbolic) orders, the most basic of which give a society its characteristic form. 18 Some critics consequently have stressed that the difficulties encountered in conceptualizing agency and structure are not overcome in Giddens account, but simply deferred to the relationship between agency and system (see Hay, 1995). 19 Memory traces are in this conception constitutive of the knowledgeability of agents, although memory traces do not entirely saturate the constitution of agents. 9

10 221). 20 Although structures are internal to agents, this does not imply that agents live in virtual worlds. The social world in this theoretical construction is not reducible to the 'knowledgeability' of agents. For one, structural properties of systems (symbolic orders in their political, economic, and legal dimensions) confront the individual in constraining and enabling ways as quasi-objective facts. 21 At the same time these structural properties are continually dependent on the actions of persons. The structural properties one finds in social systems are consequently both the "medium and the outcome" (Giddens, 1984: 25) of practices. Agents are the 'moving parts' in structuration theory. The term structuration, in a nutshell, attempts to capture the "modes in which [ ] social systems are produced and reproduced in interaction" (Giddens, 1984: 25). 22 Having outlined the central concepts of social systems and structures in Giddens account, I now turn to a discussion of agency in structuration theory. Agency, generally, refers to the capacity to act (not merely to have intentions of doing something). Giddens' shorthand for agency, namely that persons "could at any phase in a given sequence of conduct, have acted differently" (Giddens, 1984: 9), is not separable from structure. It does not intend to signal a voluntarist agent that stands free of structure. Structures as rules and resources are internal to agents in that they exist as memory traces and are instantiated in action. Agency and social acts are (at the most basic level) situated within "the temporality of day-to-day conduct" (Giddens, 1979). Agency, thus situated, becomes manifest, through interventions, into a "potentially malleable object-world" (Giddens, 1979). Giddens disaggregates agency into three dimensions, namely practical knowledge, discursive knowledge and unconscious motivation and I will briefly examine these dimensions of agency in the following. Giddens views the acts performed by persons to be primarily based on their practical knowledge of day-to-day situations (knowing 'how to go on', see Kaesler, 2003: 320). "Routinized practices are the prime expression of the duality of structure in respect to the continuity of social life" (Giddens, 1984: 282). This view allocates discursive knowledge to a position of secondary importance within structuration theory. How to go on in day-to-day life is not primarily based on the ability to verbally account for what one is doing. While 20 Consequently, structure cannot be thought of as a constraining normative order. Instead of thinking of structures as internalized norms and values, structures are conceptualized as generative schemes (rules) and material resources that function as the condition of possibility of (knowledgeable) action. 21 This is so because structural properties of institutions (potentially) hold "in suspension reflexively monitored conduct" (Kilminster, 1991: 96). 22 "To enquire into the structuration of social practices is to seek to explain how it comes about that structures are constituted through action, and reciprocally how action is constituted structurally" (Giddens, 1976: 161). 10

11 most actors will be able to give reasons for what they are doing most of the time (Giddens, 1984: 6), the need to explicate actions arises primarily in situations that are at odds with normal expectations. The second dimension of agency, namely discursive consciousness (knowing that) and the ability to verbalize one's actions are in this sense most important in situations which problematize day-to-day activity. Clearly the distinction between practical and discoursive knowledge is an analytical one, but one that underscores continuity and routine, rather than novelty and creativity. 23 But more importantly, Giddens also differentiates between practical knowledge and unconscious motivation. Here Giddens (metaphorically) assumes a 'bar' that separates the conscious and the unconscious. The latter enters into the knowledgeability of the agent only as modes of recall "to which the agent does not have direct access" (Giddens, 1984: 49). 24 The status of the unconscious is important and problematic for Giddens' stratification model of agency and I will, in the following attempt to pin-point why this is the case. Concisely put, a dominant role of the unconscious would jeopardize the conception of agency. Giddens cannot allow for a strong role of the unconscious, because to the degree that unconscious motivation would explain action, agents would become structurally determined. Since agents and structures are mutually constitutive of each other, actors would be reduced to 'structural dopes' if the unconscious actually structured conscious life (Boyne, 1991: 69). This is a conclusion which needs to be avoided in order to stabilize Giddens' ambitious theoretical construction. Giddens is hence moved to reject any conceptualization of the unconscious "that would have the effect of reproducing the primacy of social structure in the very constitution of the social agent" (Boyne, 1991: 53). In order to escape this view, Giddens argues that "the unconscious is, effectively, produced and maintained by [the] social actors themselves" (Boyne, 1991: 69). This in turn implies that unconscious motivation of agents must largely be subject to agents' capacities. As a solution to the looming difficulties of theory construction, these are conceptualized as organically rooted capacities. The (theory immanent) need to minimize the role of the unconscious is, in the final analysis, achieved by introducing "a primitive level of management of tensions rooted in organic needs" (Giddens, 1976: 117). This capacity to maintain ontological security, can only be understood in terms of an "ongoing accomplishment" (Giddens, 1976: 117) of agents which is not intertwined with structure. 25 The maintenance of ontological security, which plays a decisive part in the routinized 23 This focus makes Giddens' model less suited for IR where discursive knowledge is in high demand. 24 For an in depth perspective on the role of the unconscious see Boyne, 1991: It is therefore necessary that this capacity to maintain ontological security is fully present and active even in new-born children. 11

12 character of day-to-day activities, is thus rooted in a socially skilled agent whose agential capacities are pre-social. This amounts to the introduction of a skilled agent that is prior to, and apart from, structure. Giddens hence attributes agential capacities to the 'it', in arguing the need of a "theory of the subject (a theory of the processes through which the 'I' rises) [while also claiming] that even before the arrival of the 'I', there is still an 'it' that thinks, which is to say that there is still agency" (Boyne, 1991: 71). In short Giddens introduces an anemic but nonetheless substantial theory of the pre-social subject 'through the back-door'. His conception of agency consequently also carries with it a residuum of the Cartesian ego (see Boyne, 1991). Giddens conceptualization of unconscious motivation and the notion of ontological security posit an 'original' agency that escapes structuration. Such a notion of 'structure independent' agency violates Giddens' own notion of duality of structure/mutual constitution, and it is problematic from a non-essentialist social constructivist perspective. As such structuration theory, akin to scientific realist approaches, cannot be used to strengthen first-image theorizing in social constructivism. I maintain that the inability to completely resolve the tension between agency, structure and system is not to be found in Giddens', or Wight's lack of ingenuity. Rather it is to be found in the insolvability of the problem itself. The problem of the agent 'behind' the agent, of having to reconcile two separate social entities arises from a substantialist perspective (see Dewey and Bentley, 1949). 26 Something is required that is socialized, something is required that stands apart from structure in order to avoid structural determination. It is not possible to imagine that the 'I' can be socialized if the 'I' is not already given, consequently it must be there prior to socialization. The agent must in some way stand free of structure in order not to be reducible to structure. The part must be separated from the whole in an apodictic fashion. This starting point of inquiry tends to reproduce the necessity of an essentialized agent. As such the agency-structure dichotomy itself seems to be a stumbling block in the attempt to theoretically formulate a non-essentialist concept of agency with which to emphasize the role of individuals in IR. c) A few remarks on agency in Roxanne Doty's post-structuralist approach But starting with the agency-structure dichotomy does not make the conceptualization of a non-essentialist agent entirely impossible. Roxanne Doty, for example, offers an alternative account of the agency-structure problem that emphasizes practice (Doty, 1997: 375), and does not require an essentialized agent. Practices in a post-structuralist view are always embedded in discourses that make meanings possible. But the signification of practice is 26 For a discussion of mutual constitution and similar associated problems see Jackson and Nexon, 1999:

13 never unambiguous. Practices, like meanings, are indeterminate. Doty in effect argues for a decentering of practice and thereby eschewes "attempts to locate the source and meaning of practices in some determinable center, e.g. an unproblematically given subject or generative structural principles" (Doty, 1997: 376). Meanings, although open in principle, have to be continually arrested. These determinations of meanings are never absolute, but rather converge around a center that itself is structured. The center then is an effect of ongoing practice and it becomes decisive to investigate how the "structuring of structure(s), the effecting of a center, takes place" (Doty, 1997: 378). Post-structuralist research consequently focuses on the discursive practices "that work to exclude the potentially infinite chain of meanings and thereby move from excess to [temporarily] 'closed' structure" (Doty, 1997: 379). 27 Likewise, agency is "contingent, unstable and discourse specific" (Doty, 1997: 379). I agree with Doty, that this view is part of a social constructivist agenda, but is not taken seriously enough by most social constructivists, who fall back on a pre-discursive subject. The need to reserve a space for making choices and reflecting upon discourse as the criteria for agency in these accounts originates from the view that practices, as well as structures are fixed or closed (see Wight, 2006). If agents were fully constituted by practices, we would be left with a fully structured subject. "Subjects would be mere dupes acting according to what particular sets of practices dictated" (Doty, 1997: 380). The necessity to open space for agency and escape structural determinism, in Doty's view, is a function of the notion of 'closed structure'. If structures are conceptualized as "fully constituted objective wholes" (Torfing, 1999:148), one either thinks subjects as the mere effects of structure, or makes room for them to somehow escape or resist structure on the basis of pre-discursive qualities. Doty in her conception of indeterminate practices offers an alternative by opening space in analogy to the play of meaning found in Derrida's conception of language. "Agency is not understood as an inherent quality of individual human beings [ ], but rather as a positioning of subjects that occurs through practices, practices which are inherently discursive and ultimately undecidable" (Doty, 1997: ). Drawing on Laclau and Mouffe (Laclau and Mouffe, 1989: 111), she introduces the notion of subject-positions, which are created within discourse. The idea is that subjects are defined as a position within a discourse. It almost goes without saying that subject positions are "accorded varying degrees of agency" (Doty, 1997: 384). Agency in this perspective does not originate in a pre-discursive subject. Rather "agency results from a complex weaving together of the 27 The construction of meanings, in this view, involves the silencing of alternatives. In the enactement of one (malleable) center, other possible centers are excluded. This moves the question of power relations to the fore. 13

14 subject-positions and meanings that are available within multiple and overlapping discourses" (Davies and Harré, 1990: 59). But one could go one step further here and follow Ernesto Laclau who argues that the 'playfulness of structure' does not alone account for the impossibility of full determination. What might be added to Doty's approach is the notion of 'dislocation'. "Dislocation refers to the emergence of an event, or a set of events, that cannot be represented, symbolized, or in other ways domesticated by the discursive structure - which is therefore disrupted" (Torfing, 1999: 148). In Laclau's account dislocation prevents that structure is ever entirely determined. At the same time the subject cannot have an identity that is entirely structured by structure. But it is to some degree. It therefore becomes possible to speak of a "failed structural identity" (Laclau, 1990: 44). "The incompleteness of the structural identity constitutes the subject as the locus of a decision about how to establish itself as a concrete subjectivity with a fully achieved identity" (Torfing, 1999: 149). Identity then can never be fully achieved on the grounds of the undecidability of structure. The subject in this sense cannot be reduced to a signifier that would signify it completely. There cannot, in other words, be a relationship of identity between subject and signifier. The, 'subject before subjectivation' is consequently not understood as a positive presence, but as a constitutive lack (see Laclau and Mouffe, 1989; Torfing 1999: 52). The process of becoming somebody is an ongoing attempt to fill the ever present and necessary lack of complete identification. "The subject of the signifier is a retroactive effect of the failure of its own representation" (Žižek, 1989: 175). The subject or the agents in the approaches criticized by post-structuralists run the risk of postulating a positive center, or positive presence of agency. Doty, along with Laclau and Mouffe, attempts to conceptualize agency (as well as structure) on decisively nonessentialist terms. While conceptualizing the subject as a lack is an intriguing theoretical move, it suggests a focus on discourse, broadly conceived, when empirically engaging societies (which of course always also means society of states here). Individuals are placed within discourses and there is no objection to analyzing them in principle (see Gordon, 1999; Bevir, 1999). But the tendency has been to eclipse actors and the highly contingent unintended consequences of human agents acting into the world. The tendency to obnubliate the subject is the outflow of the need to avoid any 'original presence' in the form of an autonomous subject. The same must also be true for any notion of structure, which will not do as an original presence either (as the post-structuralist abandonment of Saussure's understanding of meaning exemplifies). The turn then to shifting discourses is highly plausible. At the same time, post-structuralist approaches seem to be less suited to focus on the importance of individuals in IR. Fortunately the subject, understood as a 14

15 constitutive lack, is not the only possibility to conceptualize the subject in non-essentialist terms. 28 While I consider the difficulties discussed so far to be intimately linked to taking the agency-structure dichotomy as a starting-point of theorizing, I will not expand on this aspect of the argumentation in any length. Suffice it to say that many of the difficulties associated with scientific realist, structurationist and post-structuralist accounts largely disappear if we think of persons (in the plural) in terms of relational-processes themselves (not just as being involved in process). In the following I would like to move to a discussion of figurational sociology, which conceptualizes persons in societies without taking refuge in the agency-structure dichotomy in the first place. 29 From a figurational perspective, it is possible to view acting persons (instead of abstract agency) and their (broadly understood) social ties (rather than structure) as two dimensions of a single overarching process, rather than two distinct entities. 30 In doing so, it permeates the boundaries between first and second image theorizing. The parts-whole dichotomy is thereby replaced by a figurational model in which persons are constitutively interdependent with each other. Bellow, I will outline such a figurational perspective in order to substantiate this claim (the irony notwithstanding) The importance of acting persons in figurational sociology Taking a figurational perspective makes an especially important difference in regard to understanding agency in the social world. In sharp contrast to (substantive) notions of 'core of agency', processual-relational thinking views persons as one type of processes among others. There is in this sense no 'I' apart from doing, not even in self-experience. The self is not understood as a unified (given) thing, a substratum underlying agency. The self is seen 28 Indeed post-structuralism embraces an aporia (Doty, 1997) between agency and structure that could also be avoided by taking a different point of departure. 29 For an outline of the substantialism and processualism see Dewey and Bentley, 1949; Rescher, 1996; Rescher, 2000; Jackson and Nexon, 1999; Bucher, 2011). 30 It is potentially problematic to view the social world in terms of agency and structure understood as entities. To be very clear in this regard: I am not making an ontological statement here. The problematic status of substantialism is not considered to be an outflow of lacking correspondence to a mind-external world. The problematic 'nature' of substantialism arises solemly in regard to specific questions being ask and the lack of utility such an approach provides in answering these questions. In short, I consider processual-relational approaches to be scientific ontologies (see Jackson, 2011). 31 Processual-relational thinking has been introduced by Patrick T. Jackson and Daniel Nexon (Jackson and Nexon, 1999) in IRT, and a number of processual-relational thinkers can be found in sociology (see for instance Abbott, 2001; Elias, 1978; Emirbayer, 1997; Kilminster, 1998; Mennel, 1992; van Krieken, 2007). 15

16 as a "unity of functioning" (Rescher, 2000: 16). Rather than searching for an operator, as in Descartes cogito, figurational thinking emphasizes operations (Rescher, 2000). The self in this view is conceptualized as a continuously integrated process. What is being disputed by processists is the view of persons as entities which exist in separation from their actions and activities in the form of some substratum of core agency. In short, the 'self' is conceptualized as a "unified manifold of actual and potential processes of action and capacities, tendencies, and dispositions to action" (Rescher, 2000: 15). The focus of theorizing is thus shifted from a "unity of hardware, of physical machinery, to a processual unity of software" (Rescher, 1996: 108). The question of what agency actually 'is', is consequently replaced by the question of what persons/agents 'do' (see Emirbayer and Mische, 1998), and I will spell this out after discussing figurational thinking more closely. Keeping the processual character of persons in mind, it is important to also underscore the relational dimensions of persons in order to articulate the importance of acting persons in social constructivist narratives of international relations in a non-essentialist fashion. In the following, I will briefly outline Norbert Elias' central concept of figuration, which underscores the basic relatedness and the constitutive ties of (interdependent) acting persons (Elias, 1969; Elias, 1978; Elias, 1991; Elias, 2000; Elias and Scotson, 1990). 32 From a sociological perspective, the concept of figuration aims to replace the concepts of agency and structure. In doing so it promises to be able to emphasize the importance of acting persons. In short, figurations are defined as dynamic social networks of constitutively interdependent persons. Figurations aim "to grasp the processual link between actors and society in a way that renders 'the individual' and 'society' 'two different but inseparable levels of the human world'" (Bauman, 1989: 39). Persons in this conception are (constitutively) connected to other persons by social ties. 33 Individuals from this perspective, are not understood as secluded, as given independently (homo clausus), but as fundamentally interdependent beings (homines aperti). Consequently, individuals are what they are only in light of the interdependencies in which they (as processes) are continuously developing. 34 'Society' is in turn understood as the ongoing medium and outcome of social practices. Rather than starting with agents and their environment, figurations signal that agents are always constituted in environments which are (largely) characterized by the functional interdependencies among persons. The relations that bond individuals together in a 32 For a discussion of Elias' sociology in IRT see Linklater, 2002; Linklater It is possible to apply this reasoning to states, as well as to human beings (see Jackson and Nexon, 1999; Bucher, 2011). 34 It is possible to draw on the insights of symbolic interactionists, hinted at above, to underpin this relational view of persons (see Morris, 1966: XIV). 16

17 'continuum of changes', in this conception, are "as real as the 'individuals' themselves" (Mennell, 1992: 256). As such it is unnecessary to resort to a pre-social agent that stands free of structure. The importance of individual action is central to this theoretical perspective. But the individual itself is not a monadic entity. Rather any individual is constitutively embedded in relational interdependencies. In order to further illustrate what the concept of figuration entails, I will briefly draw on a hypothetical and a historical example. Imagine that two persons (say Emile and Max, or for the sake of simplicity A and B) live secluded and without outside contacts. In such a setting, almost all communication will take place between these two persons. At a practical level, it is likely that they will be highly dependent on each other. At the same time, A and B will not only be interrelated at a practical and material level. Rather A and B need to be understood as constitutively interdependent as persons. The claim of figurational sociology is simply that if we want to say anything (sociologically) interesting about A, we need to take into account A's relationship with B. Given the setting of this hypothetical example, it would be surprising if the relation(s) between these two actors could be left unmentioned in studying either of them. The claim here is indeed that neither A nor B are given prior to their relation. What either of these two persons can do and, in a sense, 'what they are continually becoming' can only be determined on the basis of their relations toward each other. This is not to argue that the selves/identities of A or B are entirely 'saturated' by this specific relation. They can also be related towards abstract concepts, or other meaningful signs. 35 Ontologically speaking, A and B 'are' what they are (in the singular, as well as in the plural) only within figurations understood as a (constitutive) functional nexus. Emphasizing the processual character of this approach, it is important to underscore that the relatedness of A and B is not of a static kind. Rather, it is characterized by a dynamic interplay through which these two persons continually renegotiate themselves (see also Mead, 1966: ). The processual-relational character of persons (and societies) is also intuitively graspable from a long-term historical perspective. Taking comparatively small associations of Homo sapiens of the Upper Paleolithic, for an example, it is possible to identify short chains of interdependence and close ties among the individuals and the group as a whole (see Kaufmann, 2002). These associations were small in the sense that persons could have personal relations with all the other members of their society (see Elias, 1978: 137). People living in the Upper Paleolithic would have had very bad chances to survive (or strive) in 35 Persons, more generally, might be joined in functional nexuses through a (common) reference to mythological ancestors or shared deities. They might be joined (or separated) by a teleological narrative giving meaning to the future, or some notion of 'the good life'. Non-personal or nonimmediate 'ties' among persons are often made visible through the use of symbols such as flags, statues, monuments or specific historical narratives (see Elias, 1978: 137). 17