Prof. Andreas Bieler and Prof. Adam D. Morton
Agency and Structure in International Relations
'The difficulty . . . is the oppositional logic within which the agent–structure problem has been articulated, i.e. the agent–structure problem has been formulated within a system of thought that defines structures and agents as two distinct, fully constituted and opposed entities each with essential properties, while the central feature of structures, as defined in the agent–structure problematique, makes problematic this very distinction'.
- Roxanne Lynn Doty, ‘Aporia: A Critical Exploration of the Agent–Structure Problematique in International Relations Theory’
The problem of agency and structure
What is the relationship between human agency and (social) structure? Do actors actually have a free choice in their strategies or are they completely determined by their surrounding structures? Does structure influence action or is it simply the result of actors’ interaction? Every approach within the social sciences implicitly or explicitly adopts a conceptualisation of the relation between structure and agency. Importantly, as Colin Wight reminds us, ‘the agent-structure problem cannot be solved in the sense of a puzzle with an answer, but rather represents competing visions of what the social world is and what it might become’ (Wight 2006: 4). In this Class, we focus on how the agency-structure debate has been played out in International Relations. In the next section, we first look at the positions on structuralism and intentionalism and identify IR theories, which follow these diametrically opposite lines. The third section, then evaluates constructivist and historical materialist perspectives and their attempts to combine a focus on both structure and agency, before we turn to poststructuralism and conclude the Class with its attempts at going beyond the agent-structure debate.
Because the agency and structure problem is a general conceptual theme within the social sciences more broadly, in each section we first present the general social science position, before referring to related, specific examples in IR theory.
THINK POINT:
Why is the agent-structure debate so important in the social sciences?
Structuralism and intentionalism
In structuralism, human agency is assumed to be completely determined by the surrounding structure. There is no space for independent human agency. One key example of a structuralist approach in IR theory is neo-realism. Due to the international system characterised by anarchy and the supposed condition of a war of all against all, states are assumed to be functionally similar in that they strive to maximise their power in order to ensure their security and survival. The way individual states operate is then determined by the structure and here in particular the distribution of power capabilities between states, which is considered to be a system variable by Kenneth Waltz (see Class 2). In short, neo-realism is characterised by an outside-in causality, in which structure determines agency. Another structuralist approach, not covered in this resource, is world-systems analysis. Here, the position of individual states determines their behaviour. They are expected to operate according to whether they are in the “core”, the “semi-periphery” or the “periphery” of the world-system (Wallerstein, 1974: 134). Again there is no space for independent action by states.
Intentionalism is the direct opposite of structuralism. Here, agents are regarded as completely free in their strategies with the surrounding structure having no impact. A good example in IR theory is realism. Unlike neo-realism, realism explains international events through reference to the strategies of states or individual state leaders as main actors. In other words, international developments are explained through exclusive focus on the actions and interaction of states. Equally, a range of liberal approaches, concentrating on the interaction of actors at the international level, are part of intentionalism. Their ontology is different from realism in that they also assess agents such as transnational corporations, international organisations or regimes. Nevertheless, again, international developments are explained through a focus on key agents and their strategies and interaction. Of course, a lot of journalism follows this kind of explanation. Outcomes are simply considered to be the result of agents, be they states, policy-makers or institutions. Structure is not conceptualised in such accounts.
In sum, structuralism and intentionalism are two social science approaches that adopt a monocausal view on the agent-structure debate. While the former denies human beings any autonomy in their actions, the latter completely ignores the structures within which agents’ activities are located. Both are ultimately unsatisfactory. As Walter Carlsnaes points out (1992: 250), ‘as long as actions are explained with reference to structure, or vice versa, the independent variable in each case remains unavailable for problematisation in its own right’. Somehow, both agency and structure need to be conceptualized, when analysing international relations.
Combining structure and agency
Within the social sciences, there are a number of different ways of how the relations between structure and agency can be conceptualised. First, Anthony Giddens developed structuration theory. Central to his argument is the notion of duality, in which structure and agency are two sides of the same coin. Thus, structures are constituted by human agency and are at the same time the medium of this constitution. In Giddens’ own words (1984: 25):
'The constitution of agents and structures are not two independently given sets of phenomena, a dualism, but represent a duality. According to the notion of the duality of structure, the structural properties of social systems are both medium and outcome of the practices they recursively organise.'
As positive as this approach may look at first sight, it is not without problems. First, Giddens’ definition of structure and the closely related notion of duality are problematic. Structure as medium and outcome of social practices enjoys no existence on its own independent of the knowledge agents have of it. Nevertheless, an elaborated structure has properties which cannot be reduced to social practices made up of rules and resources instantiated via human interaction in the present. Second, Giddens overlooks the various depths of structural properties, which leads to a prioritisation of agency. He does not comprehend that ‘at any given time some properties are more resilient or engender more resistance to change than others’ (Archer 1990: 78). This lack of differentiation makes Giddens exaggerate voluntarism, i.e. the independent capacity of agents to do what they want irrespective of structure. Overall, as Hay shows, Giddens’ redefinition of ‘structure’ as rules and resources implemented in interaction opens up a new dualism, this time between agency and systems (Hay 1995: 198).
Within IR theory, Alexander Wendt has associated his constructivist account of agency and structure (see Class 4) closely with Gidden’s structuration theory. In order to account for both structure and agency, Wendt combines a structural approach with an historical analysis. ‘The core of this agenda is the use of structural analysis to theorise the conditions of existence of state agents, and the use of historical analysis to explain the genesis and reproduction of social structures’ (Wendt 1987: 365). Both methodologies within a structurationist investigation are combined through dialectical analysis, conceptualising agents and structures as mutually constituted or co-determined (Wendt 1987: 350). Structural research starts with concrete historical events and then abstracts to the social and internal organisational structures that have provided the conditions for the events to occur. Hence, it is questioned how the underlying social conditions, which provided particular structural limits and opportunities, came about. Then historical research follows, which involves questioning why certain events unfolded in the way they did. A combination of ‘structural-historical’ analysis thus provides explanation by structurally theorising causal powers, practices, interests and historically tracing the significant choices and interactions that led to specific events and the reproduction (or transformation) of social structures (Wendt 1987: 364).
Nevertheless, in the concrete execution of his approach, Wendt mixes up structuration theory with Roy Bhaskar’s transformation model (see below). In constructivism more generally, structures consist of material and ideological properties, which constrain and enable agency. In turn, collective human agency, over time, may bring about structural change. Hence, anarchy, the structure, is what states make of it considering that their interests and identities are endogenous to their interaction within the international system (see Class 4). Wendt’s ‘structural-historical’ analysis, in line with an analytical dualist approach, engages with both structure and agency at different time intervals. The way the two are, however, related to each other is perceived in purely external terms. For example, Wendt analyses structure consisting of material capabilities as well as intersubjectively constituted identities and interests while focusing on the agency of states. There is, however, no understanding of how structure and agents are internally related. The constitution of the international structure and the existence of states as main agents are simply assumed, adopted from neo-realist analysis. There is no attempt made at understanding their internal relation. Agents and structure are held as ontologically exterior, or always-already separate spheres, that are then combined as interacting realms. The problem of ontological exteriority is that the inner connection, or internal relationship, of entities is overlooked if they are held as separate and then combined. The question of why certain sets of ideas and not others have become part of the overall structure at one specific moment in time cannot be answered as a result (see also Class 10). Equally, due to this focus on the external relation between structure and agency it is only the ideational part of structure, which is considered to be changeable as a result of states’ changing interests and identities in processes of interaction, while the material aspects of structure remain the same.
Margaret Archer (1995) bases her morphogenesis approach on analytical dualism of structure and agency. Thus, she incorporates time to comprehend the agent-structure problem through analytical dualism. In her conceptualisation, there are no social structures without people, but specific structures at a particular time may be the result of actions in the past, not of actions in the present. Hence, some structures instantiated by human beings in the past, may confront people in the present as objective obstacles. Even here, though, we identify problems. First, does every action lead to an elaboration of structure? Similarly to Giddens, she does not acknowledge the different depth levels of structural properties. Equally, how can an empirical research programme be conducted, following the morphogenesis approach? In practice, it is very difficult to determine when structural elaboration by agency stops so that we can assess structure and its influence on agency, before turning back again to an analysis of agency, leading to further elaboration of structure. In other words, the cut off points of when to shift in the analysis from agency to structure and back are difficult to determine from within morphogenesis. Carlsnaes’ (1992) attempt at incorporating the morphogenesis approach into IR illustrates these difficulties well.
Roy Bhaskar (1997) offers yet another way of conceptualising agency and structure within his transformational model of social activity. It is based on scientific realism as a social philosophy of science. Structures themselves are deemed unobservable. Only their effects can be experienced. Social structures are concept dependent, activity dependent and time-space dependent and can, therefore, be transformed by human agency (Wight 1999: 119). Structures, being both constraining and enabling, do not determine human action, but define the potential range of alternative strategies. Hence, within each particular structural setting, agents have not unlimited, but nonetheless a number of different strategies available from which to choose their particular course of action.
A historical materialist perspective offers an example of a conceptualisation of agency-structure, based on Bhaskar’s transformational model. First, an historical materialist ontology understands the social relations of production as the generator of both agency and structure, thereby facilitating a focus on the internal relations. The social relations of production are the generator of structure in the way that the private ownership of the means of production and wage labour results in a number of concrete structuring conditions. It is, simultaneously, the generator of agency, because this particular way of how the social relations of production are organised results in class fractions of capital and labour as the main collective agents opposing each other.
This historical materialist ontology is combined with a historicist, post-positivist epistemology, emphasising how the past shapes the present. Hence, structures, which confront agency in the present as objective conditions, are understood as having been established by human agency in the past and are, therefore, changeable. Second, this epistemology allows us to understand the relationship between objective-subjective elements, drawing directly on Antonio Gramsci’s own reflections (see Bieler and Morton, 2001: 19-20). As Gramsci asserts (Gramsci, 1971: 445–6, emphasis added):
'It might seem that there can exist an extra-historical and extra-human objectivity. But who is the judge of such objectivity? Who is able to put himself [sic] in this kind of ‘standpoint of the cosmos in itself’ and what could such a standpoint mean? . . . Objective always means ‘humanly objective’ which can be held to correspond exactly to ‘historically subjective’: in other words, objective would mean ‘universal subjective’ . . . We know reality only in relation to man, and since man is historical becoming, knowledge and reality are also a becoming and so is objectivity.'
The geographical positions of North–South, as ‘historically subjective’ human constructions, would thus become accepted as ‘universal’ subjective definitions, or intersubjective understandings, of the social world while retaining a ‘humanly objective’ sense because of the direct physical impact they would have on people’s lives. Hence, there is an identity between subject and object. The historicist method does not entail collapsing agency into ideational practices. Ideas of the social world arise from, as well as construct, the material conditions in which social groups exist. The world is to some extent given, in the first instance, by the material conditions of the past which then set certain limits upon social practices and the conditions of existence for thought and action in the present and future. It is in this move from ‘historical subjective’ to ‘universal’ subjective and, thus, ‘humanly objective’ definitions that the internal relations between class agency and the social relations of production, which precisely include intersubjective understandings of institutions and norms, are revealed.
With an emphasis on the internal relations between structure and agency, an historical materialist perspective allows us to assess the structuring conditions of capitalism with a focus on the role of class agency. As discussed in Class 6, we can identify three structuring conditions of global capitalism: (1) competitiveness; (2) crisis tendency; and (3) outward expansion along the lines of uneven and combined development. The focus on competitiveness makes capitalism enormously productive and innovative, but also leads to regular moments of crisis, which are then partially overcome through outward expansion in the search for new markets and cheaper labour.
Class agency, in turn, is closely related to both the inner constitution and outward expansion of capitalism. The inner constitution and outward expansion of capitalism is a structuring condition resulting from the way structure is generated by the capitalist social relations of production, but the way this occurs is conditioned and shaped by the class agency of resistance. Hence, class agency is constrained by structure, but not determined. There are always several strategies available for class fractions, engendered by the social relations of production as main collective actors. Which strategy is chosen depends on class struggle. ‘In short, class struggle is the moment when agency meets structure, when labour meets the structuring conditions of the capitalist social relations of production. Class struggle is the process in which labour identities are formed and transformed. It is the moment when structuring conditions are being confirmed or changed’ (Bieler 2014: 122). Analytically, therefore, the emphasis has to be on class struggle in analyzing the interplay between capital and labour.
THINK POINT:
In what way does Giddens’ structuration theory establish a new dualism, this time between agency and system?
Conclusion: Going beyond structure and agency?
Due to its rejection of any foundationalism, poststructuralism argues that neither structure nor agency can be permanently established, as the epigraph from Doty (1997: 371-2) details. Hence, rather than overcoming the dualism of agency and structure, the approaches would have to prioritise at one point either agency or structure in their explanations. That is, they either revert to ‘a structural determinism or alternatively to an understanding of agency which presumes pregiven, autonomous individuals’ (Doty 1997: 366). As an alternative to the discussion of agency and structure as such, Doty recommends a concentration on practice. Provided practice is decentred and its indeterminacy acknowledged, she claims that the dichotomy of agency and structure can be overcome. ‘Practices, because of their inextricable link with meaning, have an autonomy which cannot be reduced to either the intentions, will, motivations, or interpretations of choice-making subjects or to the constraining and enabling mechanisms of objective but socially constructed structures’ (Doty 1997: 377). In short, both agency and structure are the effect of indeterminate, decentred practices. Nevertheless, Wight points to a core problem associated with post-structuralist approaches. ‘What in Doty’s account enables practices? What are the conditions of possibility for practices? What are the causal powers and processes that produce practices? These questions are never fully addressed’ (Wight 2006: 82-3). In other words, it is not clear what practices mean for Doty, what enables them. Ultimately, practices can only be explained by reference to further practices. Structure and agency may have been overcome, but only through an emphasis on indeterminacy and a decentred nature of practices.
In conclusion, unless we adopt a poststructuralist perspective, the problem of agency and structure will continue to be a conceptual challenge to IR theories. As always, rather than pretending that it does not exist, it is best to be open about one’s own position on this issue and be clear about the advantages of this position, but also its limitations.
References
Archer, Margaret (1990) ‘Human Agency and Social Structure: A Critique of Giddens’, in Jon Clark, Celia Modgil and Sohan Modgil (eds) Anthony Giddens: Consensus and Controversy. London: Falmer Press, pp. 73-84.
Archer, Margaret (1995) Realist Social Theory: The Morphogenetic Approach. Cambridge: Cambridge University Press.
Bhaskar, Roy (1997) A Realist Theory of Science, Second edition. London: Verso.
Bieler, Andreas (2014) ‘Transnational Labour Solidarity in (the) Crisis’, Global Labour Journal, 5(2): 114-33.
Bieler, Andreas and Adam David Morton (2001) ‘The Gordian Knot of Agency-Structure in International Relations: A neo-Gramscian Perspective’, European Journal of International Relations, 7(1): 5-35.
Carlsnaes, Walter (1992) ‘The Agency–Structure Problem in Foreign Policy Analysis’, International Studies Quarterly, 36(3): 245–70.
Doty, Roxanne Lynn (1997) ‘Aporia: A Critical Exploration of the Agent–Structure Problematique in International Relations Theory’, European Journal of International Relations 3(3): 365–92.
Giddens, Anthony (1984) The Constitution of Society: Outline of the Theory of Structuration. Cambridge: Polity Press.
Gramsci, Antonio (1971) Selections from the Prison Notebooks, ed. and trans. Quintin Hoare and Geoffrey-Nowell Smith. London: Lawrence and Wishart.
Hay, Colin (1995) ‘Structure and Agency’, in David Marsh and Gerry Stoker (eds) Theory and Methods in Political Science. London: Macmillan Press, pp. 189-206.
Wallerstein, Immanuel (1974a) The Modern World-System I: Capitalist Agriculture and the Origins of the European World-Economy in the Sixteenth Century, London: Academic Press.
Wendt, Alexander (1987) ‘The Agent–Structure Problem in International Relations’, International Organization, 41(3): 335–70.
Wight, Colin (1999) ‘They Shoot Dead Horses Don’t They? Locating Agency in the Agent–Structure Problematique’, European Journal of International Relations, 5(1): 109–42.
Wight, Colin (2006) Agents, Structures and International Relations: Politics as Ontology. Cambridge: Cambridge University Press.