Definition: practices as competent performances

Practices are competent performances. More precisely, practices are socially meaningful patterns of action, which, in being performed more or less competently, simultaneously embody, act out, and possibly reify background knowledge and discourse in and on the material world. Practices, such as marking a linear territorial boundary, deterring with nuclear weapons, or finance trading, are not merely descriptive ‘arrows’ that connect structure to agency and back, but rather the dynamic material and ideational processes that enable structures to be stable or to evolve, and agents to reproduce or transform structures. We will explore the social–theoretical implications of this definition in the second part of the article. Here, our objectives are: (i) to differentiate between behavior, action, and practice; and (ii) to unpack the notion of practice by looking into its main conceptual elements.

In common parlance, the concepts of behavior, action, and practice often are used interchangeably. Conceptually, however, they are not the same. An easy way to grasp their differences is to conceive of these notions as a gradation: actions are a specific type of behavior, and practices are a particular kind of action (see Cook and Brown, Reference Cook and Brown1999: 387).Footnote 2 In a nutshell, the concept of behavior evokes the material dimension of doing, as a deed performed in or on the world; then the notion of action adds an ideational layer, emphasizing the meaningfulness of the deed at both the subjective and intersubjective levels; and, finally, the term ‘practice’ tacks another layer on the edifice or, better put, makes it hang together as one coherent structure, by pointing out the patterned nature of deeds in socially organized contexts. The distinction between behavior and action is the easiest to grasp: action is behavior imbued with meaning. Running in the streets aimlessly is mere behavior, running after a thief is an action endowed with meaning. Practices, however, are patterned actions that are embedded in particular organized contexts and, as such, are articulated into specific types of action and are socially developed through learning and training (Corradi et al., Reference Corradi, Gherardi and Verzelloni2010).Footnote 3 Action is always a constitutive part of any practice, and yet the reverse is not necessarily true. Action is specific and located in time; practices are general classes of action, which, although situated in a social context, are not limited to any specific enacting. The act of police squads chasing down criminal gangs is a practice, because it is socially structured and reiterated. Similarly, an American carrier entering the straits of Hormuz is an action endowed with social meaning. The same action, however, when embedded in an organizational context, repeated over time and space, constituted by knowledge about the exploitation of potential force, and articulated as part of a complex set of other social performances, which may require learning and training, is part and parcel of the practice of coercive diplomacy.

By international practices, we denote socially organized activities that pertain to world politics, broadly construed. In so defining the scope of our article, we do not take a position in the many definitional debates that rage in the discipline, such as those between comparative vs. international politics, or global governance vs. international relations. Instead, we argue that one of the key epistemological consequences of taking international practices seriously is precisely to bring those scholarly debates ‘down’ to the ground of world politics in order to empirically scrutinize the processes whereby certain competent performances produce effects of a world political nature. Put differently, the scope of analysis – global, international, transnational, regional, organizational, substate, local, etc. – is itself a matter of practice: defining what counts as an international practice and what does not is best left to practitioners themselves in their actual performance of world politics.

Let us now unpack the notion of practice. First, a practice is a performance (Goffman, Reference Goffman1959; see also Butler, Reference Butler1990), that is, a process of doing something. Contrary to entities or substances that can be grasped in a reified way, practices have no existence other than in their unfolding or process (Jackson and Nexon, Reference Jackson and Nexon1999). The performance of practice goes with, and constitutes, the flow of history. As a form of action, practice differs from preferences or beliefs, which it expresses, and from discourse or institutions, which it instantiates. Second, practice tends to be patterned, in that it generally exhibits certain regularities over time and space. In a way reminiscent of routine, practices are repeated or at least reproduce similar behaviors with regular meanings. These patterns, as we explained above, are part of a socially organized context, which not only gives them meaning, but also structures interaction. This is not to say that practice is strictly iterative, however, as there is always wiggle room for agency even in repetition (De Certeau, Reference De Certeau1990; see also Goffman, Reference Goffman1959; Turner, Reference Turner1994 for a critique). As a general rule, though, iteration is a key characteristic of practices – and the condition of possibility for their social existence.

Third, practice is more or less competent in a socially meaningful and recognizable way. The structured dimension of practice stems not only from repetition but also, and in fact primarily, from the fact that groups of individuals tend to interpret its performance along similar standards (Goffman, Reference Goffman1959). Social recognition is thus a fundamental aspect of practice; its (in)competence is never inherent but attributed in and through social relations. The notion of performance implies that of a public, of an audience able to appraise the practice. As Barnes (Reference Barnes2001) notes, contrary to habit, which is performed on an individual scale (and is apprehended as such), a practice can be done correctly or incorrectly. The ascription of (in)competence is an eminently complex social process; for instance, in some contexts, incompetent practice might be more ‘successful’ in bringing results than a virtuoso performance. Fourth, practice rests on background knowledge, which it embodies, enacts, and reifies all at once. Knowledge not only precedes practice as do intentions, beliefs, etc. In addition, intersubjectivity is bound up in the performance and can only be expressed as such (Wittgenstein, Reference Wittgenstein1958; see also Taylor, Reference Taylor1985). Background knowledge is practical; it is oriented toward action and, as such, it often resembles skill much more than the type of knowledge that can be brandished or represented, such as norms or ideas (Bourdieu, Reference Bourdieu1990).

Fifth and finally, practice weaves together the discursive and material worlds. Without language, communication, and discourse, people could not tell the difference between behavior and practice. Not only is language the conduit of meaning, which turns practices into the location and engine of social action, but it is itself an enactment or doing in the form of ‘discursive practices’ (Foucault, Reference Foucault1980). By nature, practices represent the world in specific ways; they implicitly make the claim that ‘this is how things are’ (Swidler, Reference Swidler2001). At the same time, practices are mediated by material artifacts (Reckwitz, Reference Reckwitz2002; see also Latour, Reference Latour2005). Practice typically is enacted in and on the world, and thus can change the physical environment as well as the ideas that individually and collectively people hold about the world.

As a preliminary illustration, take the practice of international summitry – G8 annual summits, for example. These meetings of state officials constitute an international practice insofar as they conform to the five dimensions that we just laid out. First, G8 summits are performances; they consist of a number of actions and processes that unfold in real time, from the welcoming ceremony to the joint press conference through the official photography. Second, these performances are patterned from one year to the next. Although each meeting boasts its own particularities, there is much regularity in their staging, including the pecking order or the mixture of formal and informal discussions. Third, participating state officials generally exhibit a variable degree of competence as they attend the summit. The media and populations typically recognize the meaning of a clip featuring the British prime minister casually joking with the US president, for example. Fourth, much of the performance rests on a form of background knowledge that is bound up in practices. For instance, there is a very specific and skillful way for state officials to subtly take a little distance from the consensus forged for the official communiqué. Fifth and finally, G8 summits are both ideational and material. Participants spend a lot of time publicly and privately talking about their meetings in order to represent preferences and policies. To do so, they make use of a variety of materials – conference rooms, ceremonial artifacts, the Internet, note exchanges with sherpas, etc.

Conceptually, any given practice can be appraised through different levels of aggregation. For example, the practice of international summitry is an aggregate of several competent performances, including formal dining, press conference delivery, bilateral work meetings, etc. We suggest that the identification of the most appropriate level of aggregation should be based on two criteria. First, the research puzzle; should it deal with international summitry, then it is more appropriate to conceive of G8 summits as one aggregate practice; a study into intergovernmental rites, however, may want to zoom in at a lower level. Second, the practical experience of performers helps decide what the most appropriate level of aggregation is. In the case at hand, should state officials act out G8 summits as one whole, then it is a relevant starting point. Sherpas, however, may conceive of the formal multilateral meetings as ‘where the action is’. Methodologically speaking, sense making and situatedness are particularly important aspects of the study of international practices.

The study of international practices also faces the issue of corporate practices, that is, practices that are performed by collectives in unison. In world politics, most practices belong to this type; war, for example, is a socially meaningful pattern of action which, in being performed more or less competently, simultaneously embodies, reifies, and acts out background knowledge and discourse in and on the material world. In a very important sense, G8 summits are performed not only by singular heads of state but also by large teams of representatives. In fact, because of the background knowledge that is necessarily bound up in it, practice is always a ‘collective accomplishment’ (Barnes, Reference Barnes2001). Consequently, we explain corporate practices as being both structured and acted out by communities of practice, and by the diffusion of background knowledge across agents in these communities, which similarly disposes them to act in coordination. For example, through country-to-country discussions held at different levels (heads of state, sherpas, political advisors, experts groups, etc.), a given country mission seeks to grasp, in a very coordinated fashion, what the position of a foreign capital is on a particular issue and how flexible it could be. Such corporate practices are not the action of one corporate agent (a state) but that of a community of representatives whose members enter in patterned relations, within an organized social context, thanks to similar background dispositions.

Application: bargaining with Schelling

In this section, we illustrate our definition of international practices by revisiting Schelling's seminal theory of strategic bargaining. In many ways, one cannot find a more sophisticated theoretical discussion of a key international practice than Schelling's. His oeuvre (Schelling Reference Schelling1980 (1960), Reference Schelling1966, Reference Schelling1978) lends itself particularly well to a practice interpretation because it is both analytical and prescriptive. Most saliently, it is a general theory of bargaining, which applies as much to maneuvering in limited war as to jockeying in a traffic jam (Schelling, Reference Schelling1980: 53). Moreover, Schelling's work is about a constellation of practices, with the practice of bargaining, or the competent arrival at converging expectations, at the top, followed hierarchically by practices of deterrence and compellence, relying on threat and assurance practices, which in turn are made possible by brinkmanship and limited war practices.

Schelling is particularly concerned with the strategic element of bargaining, which is premised on the interdependence of decisions:

Building on this definition, we want to show how Schelling's treatment of bargaining touches on, more or less directly, each of the main dimensions of practice that we laid out in the preceding section.

To begin with, the practice of bargaining is more or less competent in achieving coordination. The crucial skill is one of communication – ‘Getting the right signal across’, as Schelling puts it (Schelling, Reference Schelling1966: 113). Throughout the bargaining process, players need to communicate threats, commitments, and more. Skill is also needed in order to impress on others the right incentives. For example, ‘skillful diplomacy’ on the part of a deterring player requires structuring the game so as to leave the ‘last clear chance’ to avert threatened consequences to the deterred player (Schelling, Reference Schelling1966: 101). Similarly, brinkmanship means ‘manipulating the shared risk of war’ without lapsing into it, while ‘tactics of erosion’, which seek to probe commitment, consist of impressing innocence or inadvertence as the player trespasses given limits (Schelling, Reference Schelling1966: 99 and 67, respectively). Likewise, competent compellers should communicate their resolve without positioning themselves in a way that they are forced to act (Schelling, Reference Schelling1966: 84). In sum, for Schelling, a ‘theory of deterrence’ is ‘a theory of the skillful nonuse of military forces’ (Schelling, Reference Schelling1980: 9). In order to communicate, estimate, and influence intentions, bargaining requires following specific standards of competence.

Second, Schelling's bargaining practice refers to all communication, including explicit verbal communication, as tacit; maneuvering or jockeying for position does the talking (Schelling, Reference Schelling1980: 53). Schelling is primarily interested in what he calls ‘tacit bargaining’ or bargaining through deeds. As he puts it: ‘In warfare the dialogue between adversaries is often confined to the restrictive language of action and a dictionary of common perceptions and precedents’ (Schelling, Reference Schelling1966: 141). Commitment is more easily signaled in deeds than in words because ‘significant actions usually incur some cost or risk, and carry some evidence of their own credibility’ (Schelling, Reference Schelling1966: 150; see also Fearon, Reference Fearon1997, on ‘costly signals’). As a consequence, Schelling repeatedly insists that talk is cheap. That said, he also notices that ‘enforceable threats, promises, commitments, and so forth [should] be analyzed under the heading of moves rather than communication’ (Schelling, Reference Schelling1980: 117). This remark acknowledges, in a limited way, the possibility of discursive practices; if bargaining is about the communication of intentions, the performativity of language should obviously be taken seriously.

Schelling calls the locus of convergent expectations, which rely on both competent performance and tacit communication, a ‘focal point’ – basically, a fallback position of the form: if not here, where? In his famous example:

Several implications follow from this quote. First, focal points comprise background knowledge. They are ‘so natural’, argues Schelling about certain military practices, ‘that we may not even be inclined to question the principle involved. Some of these responses are so ‘obvious’ that one is unaware that ‘obviousness’ constitutes a striking principle of interactions in diplomacy’ (Schelling, Reference Schelling1966: 147). In fact, the common trait of all focal points is ‘some kind of prominence or conspicuousness’ (Schelling, Reference Schelling1980: 57). From a practice perspective, this prominence is not inherent but contextually and socially defined (Schelling, Reference Schelling1980: 221). In other words, different players will bargain around different focal points. In their reflexive calculations and expectations, players build on shared background dispositions accumulated in and through practice.

Second, focal points are intersubjective in nature – they constitute ‘the meeting of minds’ (Schelling, Reference Schelling1980: 83). There cannot be bargaining outside meaningful relations: ‘There is generally a necessity for some social activity, however rudimentary or tacit it may be; and both players are dependent to some degree on the success of their social perception and interaction’ (Schelling, Reference Schelling1980: 163). For example, the success of American deterrence during the Cold War rested in a very significant way on Soviet expectations; and reciprocally (Schelling, Reference Schelling1966: 56). Bargaining is fundamentally communicative and it necessarily involves knowledge. As a performance, it entails the presence of an audience, which matters a lot in making commitments, for instance. Alluding to the arms race, Schelling notes that ‘both we and the Soviets play to an audience of third countries. Prestige of some sort is often at stake in weapon-development competition’ (Schelling, Reference Schelling1966: 276; see also 49–50). Competent bargainers always calibrate their signals toward the intended audience so as to make them socially recognizable.

Third and related, focal points, and bargaining more generally, are infused with power. In particular, Schelling (Reference Schelling1980: 114) speaks of ‘the power of suggestion that is able to bring expectations into convergence’. Performative and communicative, suggestive power allows one to impose a particular focal point, which further reinforces one's advantage in bargaining: ‘the ultimate focus of agreement did not just reflect the balance of bargaining powers but provided bargaining power to one side or the other’ (Schelling, Reference Schelling1980: 68). Schelling forays into the power dynamics of bargaining by asking a fundamental question: ‘How does one person make another believe something?’ (1980: 23) He hints at an answer by raising the skillful use of cultural commonplaces and practices. For instance, he argues that commitment problems would be much eased in a society where ‘cross my heart’ is considered an ‘absolutely binding’ practice (1980: 24–26).

Fourth, focal points stem from the patterned nature of social practices. It is the regularity of players’ moves that makes the convergence of expectations and coordination possible. For example, American military practices in North Vietnam were ‘unambiguous’ in signaling limited warfare, argues Schelling, because they ‘contained a pattern’ which made coordination possible (Reference Schelling1966: 145). Similarly, there typically exists an ‘idiom’ in practices of retaliation – ‘a tendency to keep things in the same currency, to respond in the same language, to make the punishment fit the character of the crime, to impose a coherent pattern on relations’ (Schelling, Reference Schelling1966: 147). Inside the recursive logic that characterizes a practice's ontology, focal points make coordinated practices possible, which in turn produce focal points.

Bargaining exemplifies the doubly prudential nature of practices. On the one hand, past practices explain current ones because the patterns they form create focal points. Take what Schelling calls the ‘phenomenon of thresholds’, that is, conventional dividing lines that are recognized as such in bargaining. Interestingly, these ‘arise by a historical process, even inadvertently or accidentally, and can acquire status just by coming to be recognized over a prolonged period’ (Schelling, Reference Schelling1966: 135). Bargaining at time t, in other words, is informed by bargaining at time t − 1. On the other hand, current practices extend into the future. For example, in the game of chicken, ‘what is in dispute is usually not the issue of the moment, but everyone's expectations about how a participant will behave in the future’ (Schelling, Reference Schelling1966: 118). Bargaining at time t influences bargaining at time t + 1. In sum, practices carry the past into the present; and the present into the future.

Schelling's stress on tradition and precedent is one of the most striking aspects of his works. Speaking of the tradition of nuclear non-use, he contends: ‘Tradition and precedent are important here. […] Any particular limitation will be the more expectable, the more recognizable, the more natural and obvious, the more people have got used to recognizing it in the past. The line between nuclear and high explosives was not only observed during the Korean War but reinforced’ (Schelling, Reference Schelling1966: 138). In reinforcing certain meanings, including expectations, practices create socially constructed traditions. For example, asking ‘what is so different about nuclear weapons?’ Schelling responds: ‘everybody knows the difference. The difference is not tactical; it is ‘conventional’, traditional, symbolic – a matter of what people will treat as different, of where they will draw the line. […] it is by convention – by an understanding, a tradition, a consensus, a shared willingness to see them as different – that they are different’ (1966: 133–134). In a nutshell, the practice of bargaining is constitutive of social reality: through iteration, it creates ‘rules of the game’ (Schelling, Reference Schelling1980: 107), ‘conventions’ (Schelling, Reference Schelling1966: 123), and ‘norms’ (Schelling, Reference Schelling1980: 168–169). ‘If we always treat China as though it is a Soviet California’, observes Schelling, ‘we tend to make it so’ (1966: 60).

Schelling contends that traditions congeal into institutionalized practices after they cross a ‘tipping point’. A tipping point is a special case of a broad class of critical mass phenomena, which may be related to a critical number, density, or ratio, that becomes self-sustaining ‘once the measure of that activity passes a certain minimum level’ (Schelling, Reference Schelling1978: 95). Schelling says: ‘Confidence and tradition take time. Stable expectations have to be constructed out of successful experience, not all at once out of intentions’ (Schelling, Reference Schelling1966: 123). Moreover, establishing a practice may require learning. Mentioning the case of David and Goliath, Schelling, quoting Yigael Yadin, refers to the dueling practice. When Goliath says to the Israelites: ‘If he [David] be able to fight with me, and to kill me, then will we be your servants, but if I prevail against him, and kill him, then shall you be our servants, and serve us’, he is ‘teaching’ the Israelites to adopt the practice of dueling as an alternative to war (Schelling, Reference Schelling1966: 144). At the same time, Schelling also argues that people use practices to signal resolve, deter, and manipulate the adversary's expectations. Brinkmanship, for instance, seeks to manipulate information and signal resolve, and it depends on the mutual and similar recognition of risks. Observing that the superpowers competently practice limited war as a deterrent to aggression, Schelling calls on states to ‘leave something to chance’, namely, signal that all-out war may occur should limited war escalate (1980: 190).

The bargaining practice also weaves together the material, meaningful, and discursive worlds. Material artifacts foster intersubjectivity, and vice versa. As Schelling observes:

Nuclear arms control also rested on attaching the abstract and theoretical logic of stable nuclear deterrence to the number and quality of the missiles permitted on both sides. Schelling's notions of tacit bargaining via ‘red lines’ also weave the material and ideational worlds in interesting ways. This is not only because of the mutual dependence between tacit communication and material ‘lines in the sand’, but primarily because, due to planning processes in defense bureaucracies, red lines tend to become competent performances, whose reliance on lethal military capabilities make war less possible but more deadly (Schelling, Reference Schelling1966: 159, fn. 17). As we further illustrate below, international bargaining, and strategic interaction in the field of international security more generally, rest on a combination of material and cultural artifacts.

Social–theoretical implications: overcoming dichotomies

The promise that a practice framework holds in surmounting conventional divides in social theory is particularly striking. A practice framework not only transcends, but also synthesizes, three different approaches to culture that have characterized recent strands in social theory (Reckwitz, Reference Reckwitz2002). At first, culture was conceptualized as a set of ideas carried in individuals’ heads. In this mentalist version, akin to psychology, the mind is the site of the social. In IR theory, this cultural approach relies on social and cognitive psychology to explain foreign policy by means of individuals’ ‘ideas’ and beliefs (Jervis, Reference Jervis1976; Goldstein and Keohane, Reference Goldstein and Keohane1993). Subsequently, heavily influenced by poststructuralism, according to which culture exists outside of agents in chains of signs and symbols, culture took a turn to language and an understanding of meanings as located in discourse. In IR theory, this turn led to critical and poststructuralist readings of IR ‘texts’ (Der Derian and Shapiro, Reference Der Derian and Shapiro1989). The third approach to culture, characteristic of constructivism, places meanings as part of the intersubjective structures that emerge out of social interaction. A constructivist perspective in IR theory converged ‘around an ontology that depicts the social world as intersubjectively and collectively meaningful structures and processes’ (Adler, Reference Adler2002: 100), and around reflexive agents who reproduce and change social structures (Wendt, Reference Wendt1999). As one step further in the theorization of culture, practice theory is an invitation to build on these three strands of social theory and conceive of the social as bundles of ideas and matter that are linguistically, materially, and intersubjectively mediated in the form of practices. Culture, in other words, is not only in people's minds, discourse, and interactions, it is also in the very performance of practices (Swidler, Reference Swidler1986). From that perspective, practices not only organize the world – they are also the raw materials that comprise it. Thanks to this ontology, the concept of practice promises to move beyond a number of entrenched dichotomies in social theorizing.

First, practices are both material and meaningful. On the one hand, practices are material insofar as they are doings enacted in and on the world. Practices engage with the environment and its artifacts, whether natural, cultural, or political. In addition, many practices often involve the use of ‘things’ as indispensably as that of minds or speech. In Latour's (Reference Latour2005) action-network theory, for instance, material objects, agents, and meanings interact continually. In IR, the intersubjective structure of nuclear deterrence is sustained by the existence of thousands of warheads; and reciprocally (Pouliot, Reference Pouliot2010b). The practice of finance trading, to take another example, requires not just human beings and their physical faculties and actions, but also computers and other technologies that help institutionalize background knowledge as finance trading, and that allow people to perform their performances competently. In addition, practice typically does something in the world, and thus can change the physical world as well as the ideas that individually and collectively people hold about the world. Contrary to thinking or reflecting about the world in a contemplative fashion, practices are directed toward the material world and thus exist only embodied in materials. Practices also are bodily in the sense that their enactment involves corporeal parts other than the brain (Polanyi, Reference Polanyi1983; Bourdieu, Reference Bourdieu1990). At the same time, practices also are shot through with meaning – they are ‘culture in action’ (Swidler, Reference Swidler1986). A deed performed in a social setting cannot be said to have an immanent meaning encapsulated in its materiality. Through social interaction, people attribute meanings to their activities and build on these to interact further. In order for practices to make sense, then, practitioners must establish (contest, negotiate, and communicate) their significance.Footnote 4

As a performance in and on the world, practice leans on language in two senses: weak and strong. In a weak sense, language sustains intersubjectivity and thus links agency, structure, and process in socially meaningful ways. Without language, communication, and discourse, people could not tell the difference between behavior and practice (Rorty, Reference Rorty1982). The constitution of competent performance, in other words, is fundamentally epistemic, insofar as accounts of lived practices are textually constituted.Footnote 5 In its strong sense, not only is language the conduit of meaning, which turns practices into the location and engine of social action, but it is also and primarily an enactment or doing (Foucault, Reference Foucault1980). Discursive practices, thus, are socially meaningful speech acts, according to which saying is doing (Searle, Reference Searle1969). Although practices still rely on knowledge and embody material objects, in a discursive strong sense, the competence of routinely doing something socially meaningful often relies on discourse. It is thus relevant to conceive of discourse as practice and to understand practice as discourse.

Second, practices are both individual (agential) and structural (Giddens, Reference Giddens1984; Bourdieu, Reference Bourdieu2001). When ‘disaggregated’, practices are ultimately performed by individual social beings and thus they clearly are what human agency is about. Collectively, however, we understand practices as structured and acted out by communities of practice, and by the diffusion of background knowledge across agents in these communities, which similarly disposes them to act in coordination. Practices are agential, however, not only because they are performed by individuals and communities of practice, but also because they frame actors, who, thanks to this framing, know who they are and how to act in an adequate and socially recognizable way (Rasche and Chia, Reference Rasche and Chia2009: 719; see also Goffman, Reference Goffman1977). Since social structure does not cohere on its own (Sewell, Reference Sewell1992), agency means the human capacity to do things that could be done differently (Giddens, Reference Giddens1984: 9). Recursively, in and through practice, agents lock in structural meaning in time and space. Agency also means doing things for reasons, many of which are structurally supplied. Practices translate structural background intersubjective knowledge into intentional acts and endow them with social meaning. Structure, in turn, shows up in practices in the form of standards of competence that are socially recognized. There is, then, a normative or rule-like dimension to practice, which is bound up in its application. While performed by individual human beings, practices are possessions of collectives insofar as their meanings belong to communities of practice. ‘Suspended’ between structures and agency, practices are simultaneously enacted (agency) and inserted within a social context or political order (structure). The advantage of taking practices as the main site of the social thus lies in enabling a superior formulation of the agent-structure conundrum, where agency and structure jointly constitute and enable practices. By implication, the methodological ‘bracketing’ that is sometimes advocated in IR – begin with agents (or structures), and then look at the other side of the co-constitution equationFootnote 6 – can only take us some distance in understanding world politics.

Third, from a practice perspective, knowledge is not only located ‘behind’ practice, in the form of intentions, beliefs, reasons, goals, etc. It is also ‘bound up’ in the very execution of the practice. For the seasoned practitioner, knowledge does not precede practice but is ‘enclosed’ in its execution (Ryle, Reference Ryle1984). Of course, people reflexively think about their practices. Not only does practice not trump reflexivity, judgment, and expectations, which are core features of social life, but it actually depends on individuals’ reflexive, normative, and instrumental judgments to remain effectively institutionalized. Strategic practice, for example, reflects political judgments, whose ground is not only empirical, but also practical (Williams, Reference Williams1991; see also Kratochwil, Reference Kratochwil1989). Thus, to confirm that something is indeed what it is through repeated rituals of practice (Swidler, Reference Swidler2001: 89) requires reflexivity and judgment. Professionalization similarly relies on rationalizing and self-examining deliberative processes. Needless to say, reflexivity and judgment are also at the foundation of practice transformation. Taking practice seriously, however, draws special attention to all those meanings that are weaved into practice and that, as such, often remain tacit and inarticulate. With Searle, we call this knowledge, partly tacit, partly reflexive, ‘the Background’,Footnote 7 a notion that is also akin to Bourdieu's ‘habitus’.Footnote 8

Background knowledge consists primarily of intersubjective expectations and dispositions, which can be grasped only as embedded in practice. Individuals and groups act, interact, reason, plan and judge, symbolically represent reality, and have expectations of the future within a dominant interpretive backdrop that sets the terms of interaction, defines a horizon of possibility, and provides the background knowledge of expectations, dispositions, skills, techniques, and rituals that are the basis for the constitution of practices and their boundaries. Background knowledge, however, does ‘not create uniformity of a group or community, but organize[s] their differences around pervasive understandings of reality’ (Adler and Bernstein, Reference Adler and Bernstein2005: 296). Similarly, habitus refers to this embodied stock of unspoken know-how, learned in and through practice, and from which deliberation and intentional action become possible (Pouliot, Reference Pouliot2008). Contrary to representations, which usually are verbal and intentional, dispositions are often tacit and inarticulate: knowledge that is forgotten as such – unless it is reflexively recovered.

Fourth, practices partake in both continuity and change in social and political life. On the one hand, practices are the vehicle of reproduction. Intersubjectivity lives on in and through practice. The performance of practices in socially recognizable ways is the source of ontological stability in social life. At the same time, however, it is also from practices that social change originates. For one thing, practice-qua-performance is a process; change, not stability, is the ordinary condition of social life. As March aptly put it: ‘Change takes place because most of the time most people in an organization do about what they are supposed to do; that is, they are intelligently attentive to their environments and their jobs’ (Reference March1981: 564). Stability, in other words, is an illusion created by the recursive nature of practice. For another, new ways of thinking or doing necessarily emerge from the contingent ‘play of practice’ (Doty, Reference Doty1997), in which meanings are never inherently fixed or stable.

The material/meaningful, structural/agential, reflexive/background, and stability/change attributes of practice acquire concrete and workable theoretical and empirical meaning in the concept of communities of practice (Wenger, Reference Wenger1998; in IR, see also Adler, Reference Adler2005: 15–27, Reference Adler2008). Practices develop, diffuse, and become institutionalized in such collectives. A community of practice is a configuration of a domain of knowledge that constitutes like-mindedness, a community of people that ‘creates the social fabric of learning’, and a shared practice that embodies ‘the knowledge the community develops, shares, and maintains’ (Wenger et al., Reference Wenger, McDermott and Snyder2002: 28–29). The knowledge domain endows practitioners with a sense of joint enterprise, which ‘brings the community together through the collective development of a shared practice’ and is constantly being renegotiated by its members. People function as a community through relationships of mutual engagement that bind ‘members together into a social entity’. Shared practices, in turn, are sustained by a repertoire of communal resources, such as routines, sensibilities, and discourse (Wenger, Reference Wenger1998: 72–85, 209). The community of practice concept encompasses not only the conscious and discursive dimensions, and the actual doing of social change, but also the social space where structure and agency overlap and where knowledge, power, and community intersect. Communities of practice are intersubjective social structures that constitute the normative and epistemic ground for action, but they also are agents, made up of real people, who – working via network channels, across national borders, across organizational divides, and in the halls of government – affect political, economic, and social events.

How security practices constitute strategic interaction

This section suggests a new theoretical framework for understanding strategic interaction from the practice perspective. Building on the practice-oriented reading of Schelling that we proposed in the first part of the article, we inquire into the Cold War and post-Cold War strategic interaction between Washington and Moscow, with particular attention to deterrence and arms control practices.

Our practice-oriented explanation of strategic bargaining departs from classic research programs that explain strategic interaction on the basis of material power and/or interests, and the maximization of payoff values within strategic games (e.g. Snyder and Diesing, Reference Snyder and Diesing1977; Snidal, Reference Snidal1985; Lake and Powell, Reference Lake and Powell1999). Game theory has become the tool of choice for rationalists studying interdependent decision-making situations; it is supposed to model and represent strategic interaction between corporate and individual actors in world politics where the outcome depends upon the choices or strategies they make. The actor needs to discover the choices available to him/her and each other player and he/she needs to assign utilities for each possible outcome for each player.

We also depart from traditional constructivist approaches that take strategic interaction as resulting from ideas, norms and, more generally, knowledge, which actors learn, diffuse to each other, persuade each other with, and socialize others into (Finnemore and Sikkink, Reference Finnemore and Sikkink1998; Risse, Reference Risse2000; Checkel, Reference Checkel2005). In addition, our approach resembles, but also partly departs from, constructivist approaches that take strategic interaction as a mechanism for the constitution of common knowledge among actors, and that, when understood from a cultural symbolic perspective, is also a mechanism of learning and cultural change (Barnett, Reference Barnett1998; Wendt, Reference Wendt1999; on symbolic interactionism, see also Mead, Reference Mead1964; Blumer, Reference Blumer1969). We are admittedly sympathetic to Goffman's (Reference Goffman1970) approach to strategic interaction, where actors strive for advantage over the opponents’ moves (on which one's own moves depend) by means of symbolic communication and the power of performance in face-to-face encounters.Footnote 9 But our approach to strategic interaction focuses less on agents and their preferences and narratives, as Barnett's does, than on practices and their constitutive effects on both agency and structure, that is, the game per se of strategic interaction. Put differently, we use Gofmann to illustrate the micro-foundations of the macro effects of practices on strategic interaction. In so doing, our approach is inherently ‘relational’, but rather than taking processes of interaction as the building block of theory (Jackson and Nexon, Reference Jackson and Nexon1999), we broaden IR ontology by taking practices as constitutive of interaction processes. Similarly, we not only build on, but also expand, the poststructuralist focus on discursive practices and their constitutive effects on the ‘games’ of world politics (Doty, Reference Doty1996; Jackson, Reference Jackson2006; Hansen, Reference Hansen2006).

We instead take agency and agents as emergent from, and being continually reproduced by, practices, which capture both structure and self, and discourse/knowledge and the material world. Since making practices our main focus of analysis and of ontological concern helps avoid dichotomies in the social sciences and IR, we can nevertheless fruitfully incorporate insights from rationalism, constructivism, and poststructuralism. While we consider that the added value of practices lies in being able to explain differently and better not only strategic interaction, but also, for example, institutionalization, the constituting of social orders, the performative side of power, and change amidst stability, our focus on strategic interaction will enable us to also capture some of the practices’ other important roles and effects.

We argue that practices stabilize social structures and fix ideas and subjectivities in people's minds (or determine the dominant ideas that corporate actors focus on at a given point in time), thus constructing agents and agency. Our theory, in a kernel, is that practices structure and congeal thought and language into regular patterns of performance and turn contexts or structures into (individual and corporate) agents’ dispositions and expectations. This is why we take Goffman-like frames as the micro-foundations of the macro effects of practices on strategic interaction. Background knowledge is Janus-faced because, in addition to being intersubjective knowledge embedded in practices, it also constitutes subjective representations of intersubjectivity – mainly expectations, dispositions, and pre-intentional capacities – that make intentional states possible (Adler, Reference Adler2008: 202; Pouliot, Reference Pouliot2008: 274).

When states face each other due to a myriad number of reasons, their strategic interaction is affected, indeed constituted, not only by the cost–benefit analyses leaders make, the ideas and knowledge people carry in their heads, and the discourse they use to communicate. Rather, what states do vs. other states, the moves they make, the signals they give, and the language they speak are constituted by the practices they share. Think if you will of practices as focal points, which allow actors in strategic interaction to focus on similar scripts, or frames, and which affect actors’ dispositions and expectations. Practices, obviously, do not necessarily lead to commonality of interests, although they might; in cases of international conflict, which are prevalent in world politics, when interests diverge markedly, practices serve as structural, discursive, and epistemic focal points that make possible common knowledge and enable actors to play the international game according to similar rules, or at least in a way that is mutually recognizable.

Take, for example, strategic interaction between the United States and the Soviet Union during the Cold War. For most of the Cold War, strategic interaction between these actors was constituted by the practices of nuclear deterrence and arms control (Schelling, 1966; Morgan, Reference Morgan2003; Freedman, Reference Freedman2004). As time went on, and the background knowledge on how to prevent nuclear war via stable deterrence and arms control was more or less adopted by both superpowers – which became entrenched in the national security bureaucracies and the halls of government – the superpowers’ strategic interaction was constituted by, and causally related to, the practice of deterrence and arms control. The US and Soviet decisions of military hardware acquisition, the technologies they developed and brought to bear on the military field, the negotiations they undertook, and the social technologies they jointly developed to prevent nuclear war, such as the 1963 ‘hot line’, the moves they made vs. each other, the language they spoke, for example, about ‘counterforce’, ‘countervalue’, and ‘crisis stability’, were all constituted by the practices of deterrence and arms control they increasingly came to share.

The United States and the Soviet Union did not constitute a community of deterrence and arms control practice at the beginning of the Cold War. It was practice that turned the superpowers into players of a nuclear deterrence and arms control game. In time, however, the superpowers adopted identities that were associated with a community of deterrence and arms control practice and learned to competently perform the moves required to deter each other and thus to prevent nuclear war. Of course, a number of practices – Khrushchev's brinkmanship over the Cuban missile crisis being a case in point – were not recognized as competent by the other side, temporarily shaking the terms of strategic interaction (see Neumann and Pouliot, Reference Neumann and Pouliotforthcoming). As a general trend, though, with every new US administration and Soviet Politburo, the theories that first constituted the background knowledge of stable deterrence and arms control (which Thomas Schelling and other Cold War strategic experts first developed) became increasingly established in government circles and national security bureaucracies, setting the frame of mutually recognizable competent performance (Adler, Reference Adler1992).

The practices of stable deterrence and arms control, therefore, constituted strategic interaction, rather than the other way around. Practices helped construct common knowledge and a common understanding of what rationality means in practice when deterring nuclear war. It was neither an a priori rationality, nor ideas jumping from mind to mind, which helped to create strategic stability between the superpowers during the Cold War. Rather, it was doing or performing deterrence and arms control that practically taught the superpowers how to think alike when it came to the basics of deterrence and arms control. This process, whereby both social structures and social subjectivities were constituted by practices, did not come easily. The interactive process took time; it involved the diffusion not only of theories, but also of the practical ways of how to apply deterrence theories. First and foremost, it was a learning process of how to practice the prevention of nuclear war. This learning process was based not only on an exchange of ideas, such as in the Cuban Missile Crisis, but it also resulted from practical experience of how to get away from the brink of nuclear war. Learning by doing means more than the acquisition of new ideas – primarily, it consists of gaining ‘the ability to act in the world in socially recognized ways’ (Brown and Duguid, Reference Brown and Duguid2001: 200).

It also follows that the practices of nuclear deterrence and arms control were not the result of socialization but rather the opposite. Because of the practices of deterrence and arms control, actors on both sides of the divide were able to learn and socialize with each other about how to keep nuclear stability. Moreover, Cold War ‘nukespeak’ made sense only in the context of the practical meaning, which missiles, bombs, airplanes, and submarines were endowed with, and the competent performances surrounding nuclear stable deterrence and arms control. As the Cold War literature shows (Liklider, Reference Liklider1971; Kaplan, Reference Kaplan1983; Adler, Reference Adler1992; Evangelista, Reference Evangelista1999), it was mainly American experts who first developed the theories which were later diffused to and adopted by the Soviet Union, and which constituted the practices of deterrence and arms control. The Soviets, however, contributed to the development of deterrence and arms control practices by taking a more political approach than the Americans, which the Americans borrowed from. In this way, a social structure, consisting of shared background knowledge, was congealed and became stabilized, which became the source of meaning for what actors did. Actors’ subjectivities about how to engage in strategic interaction, and more precisely dispositions and expectations about what it is to competently perform stable deterrence and arms control, were therefore simultaneously fixed and selected in decision-makers and national and international bureaucrats’ minds. Superpower practitioners could have been focusing, for example, on conflict resolution-type diplomatic intercourse, or on counterinsurgency. They did not, however, because practice, more precisely a community of deterrence and arms control practice, both informally and formally as well as contextually bound them by a shared interest in learning and applying deterrence and arms control practices to the task of preventing nuclear war. In spite of conflicting political and military interests, deterrence and arms control's background knowledge endowed superpower practitioners with a sense of interconnectivity and joint enterprise, which constantly was renegotiated in light of the practices. Shared practices, in turn, were sustained by a repertoire of communal resources, such as routines, scripts, social technologies, and discourse, and by how (and how well) practitioners performed on the ground.

It is important to emphasize that not all US and Soviet practitioners and experts, however, accepted nuclear deterrence as the only, or the more efficacious, means of projecting military power to achieve political goals. Common knowledge also was not monolithic when it came to deterrence's modalities, scope, and reach. Within the United States and the Soviet Union, the convergence around deterrence practice was often constrained by domestic politics, ideological disagreements, and civilian-military disputes about the role of nuclear weapons. Still, deterrence and arms control strategy and practice allowed Cold War adversaries to share expectations of proper action and rationally to weigh policy options according to common knowledge of the situation (Adler, Reference Adler2009: 93).

To sum up how practices structure strategic interaction: (1) Interdependencies arise among people engaged in the same practices. Practices create agents and give meaning to agency. (2) Individual and corporate actors, who join in enacting a practice, or a community of practice, are socialized into the background knowledge on which the practice is based and learn both the community's identity and the competence that comes along with the performance. (3) Agents do make choices, as rationalists assert, but they do this in the context of situated actions in institutional settings, and as part of patterns made up of a variety of actions, which together make practices possible. Actually, practices help construct a practical and mediated understanding of what is rational in given situations, and how practitioners’ distinctive and often conflicting interests can be advanced by means of strategic interaction. (4) Contrary to classical constructivism, socialization, learning, and persuasion follow rather than precede practice; at best, they co-evolve. (5) In contrast to poststructuralism, practices are intimately related to subjects who are not passive performers of discursive scripts or texts, but are active agents of both stabilization and change. (6) The discursive side of nuclear deterrence and arms control practices cannot be entirely understood without the missiles, bombs and organizational resources, which over time sustained its existence and importance. (7) Practices turn common knowledge into deeds; the latter becomes part not only of what communities of practice know, but also do so together. (8) Practices set the rules and boundaries of strategic games, and determine not only the nature, motives, and payoffs of strategic bargaining, but also what signals are needed and for what purpose; they also help create focal points that can be used either in the present or in future encounters to achieve coordination.

A practice explanation of strategic interaction has important implications for future research on strategy and international security more broadly. First, we need to map those practices that are constitutive of strategic interaction and uncover the constitutive mechanisms at work. One may inquire, for example, about the effect of counterinsurgency practices on strategic interaction in the post-Cold War era (Adler, Reference Adler2009). In this case, we would be looking not only at whether counterinsurgency practices improve or actually worsen conflict management and control, but also at how the absence of structural common knowledge and background knowledge among state and non-state actors are impeding the development of shared practices that could help mitigate the worst consequences of asymmetrical warfare, such as the use of non-conventional weapons. Equally important would be to inquire about what roles practices come to play in the strategic interaction between so-called ‘rogue states’ – which, such as North Korea have, or such as Iran are actively seeking, nuclear weapons – and established nuclear powers of the West, such as the United States, or countries thought to have a sizable nuclear force, such as Israel.

Second, one cannot fully explain strategic interaction from a practice perspective without first asking where practices come from and how they became established. Thus, we may look at the evolution of a specific practice, that is, its contingent processes of transformation, namely, how and why certain practices acquire prominence in time and space. The focus here is on a practice's lifecycle (see Adler, Reference Adler1991; Finnemore and Sikkink, Reference Finnemore and Sikkink1998), which includes, in its ideal-typical form, the generation, diffusion, institutionalization, and fading of a particular competent performance. To be sure, not all practices go through these four phases; it is an empirical matter to determine the ebb and flow of a specific practice in history – whether it ever diffused or institutionalized, for instance. Some contemporary practices may fall short of meeting the third phase, and we can as well imagine ‘non-practices’ that have failed to emerge in the first place despite favorable conditions and supportive agency. The lifecycle of practice, in other words, is not a teleological framework, but a genealogy of the development, even if arrested, of a socially organized and meaningful activity.

Emphasis, for example, may be put on generative relationships, that is, instances or episodes of formative interactions, which, due to either material or ideational reasons, or both, facilitate the emergence of a new practice. To return to our illustration, the joint US–Soviet seminars on nuclear arms control of the 1960s; the Helsinki Final Act negotiations in the early 1970s, which led to the development of the Conference on Security and Cooperation in Europe, which played a role in bringing an end to the Cold War; and the negotiations that preceded Stockholm's first and seminal global environmental conference in 1972, from which the practice of sustainable development sprang, are examples of generative relationships. Alternatively, one may look into the diffusion of a practice – how intersubjective knowledge becomes more widely established in communities of practitioners and thus how the latter expand – or its institutionalization, whereby intersubjective knowledge becomes established as social structures.

‘Non-practices’, at the same time, may be as important to explaining strategic interaction as established practices, such as nuclear deterrence and arms control. This, for example, would be the case of nuclear (non-) proliferation practices between North Korea and Iran, on the one hand, and Western nuclear powers, on the other. In the case of Iran and the West, strategic interaction is mostly governed by either classic diplomatic or coercive diplomatic practices, as well as by economic and intelligence practices, which so far seem to be failing to prevent nuclear proliferation. Moreover, if and when Iran manages to develop and deploy nuclear weapons, it is questionable whether Cold War nuclear practices, such as stable nuclear deterrence and arms control, can be readily applicable to the Middle East and help prevent nuclear proliferation to neighboring states and also nuclear war. From a practice perspective, it will be important to know whether Iran and the West have or are developing common knowledge to manage nuclear weapons and nuclear crisis situations. As far as we know, there has been little effort to mutually learn about each other's attitudes toward religion, nationalism, victory, death, honor, and pride. There is a need, therefore, for diplomatic generative relationships between the parties aimed expressively at creating common knowledge. Western countries, for example, should become more attuned to Iran's tacit and explicit rationality rules and how Iranian leaders assess risks. Iranians, on the other hand, should learn more about how rational choice and cost-benefit analysis enter into the calculations of Western leaders. Iranians may also come to appreciate better the knowledge and practices that the United States and Russia developed during the Cold War. While Westerners may find it in their interest to transfer to the Iranians technologies that help prevent accidental nuclear war, Iranians may want to learn which deterrence moves, in the form of sticks and carrots, are intended to prevent mutual suicide (Adler, Reference Adler2009: 104).

Third, studying strategic interaction from a practice perspective raises questions about how practices generate transformation in world politics, or paraphrasing Sewell, how the ordinary unfolding of practice generates transformations (Sewell, Reference Sewell1992). In our above illustration, we took strategic interaction as intervening between practical causes and world political transformation effects. Generally speaking, practices help explain how the world is governed at particular points in time. First, transformation is the ordinary accomplishment of social life; stability is the result of the work of practice. Second, practice is the accomplishment of agency and more specifically of political contestation, from which transformations necessarily flow. Third, the engine of practice transformation is not only agentic but also structural because, locally, practices interact with one another as part of constellations.

We argue that transformations of social order are mediated by strategic interaction. When, for example, at the end of the Cold War deterrence and arms control practices were at least partly circumvented and replaced by ‘cooperative security practices’ (Adler, Reference Adler1998, Reference Adler2008; Adler and Greve, Reference Adler and Greve2009; Pouliot, Reference Pouliot2010a: 150–161), the nature of strategic interaction between the North Atlantic Treaty Organization (NATO) states and former Warsaw Treaty states changed dramatically. Military power began to be conceived differently, mostly as peacekeeping forces, and former enemies became ‘partners for peace’ (Gheciu, Reference Gheciu2005; Adler, Reference Adler2008). At issue was, first, the enlargement of NATO to the East, and second, the development of a new European security architecture built around not only established institutions, such as NATO, and new institutions, such as the Euro-Mediterranean Partnership, but also new practices, such as partnership, region-building, and the use of joint military training for the sake of socialization toward Western values of democracy, the rule of law and human rights. These practices helped constitute strategic relations between the East and West during the 1990s, especially until Russia begun at least partly abandoning them and returning to practices reminiscent of the Cold War game (Pouliot, Reference Pouliot2010a: 161–191). Agency played a key role in both the diffusion and contestation of old and new practices. For example, practices of ‘seminar diplomacy’ (Adler, Reference Adler1998), enacted by the Organization for Security and Cooperation in Europe and NATO, became a means for teaching democracy and the rule of law and human rights to Eastern countries (Gheciu, Reference Gheciu2005), creating new subjectivities and partly replacing deterrence practices as well as arms-control and disarmament conference meetings. On the other hand, the Balkan conflict in the 1990s and the concomitant recourse to warring practices and coercive diplomacy reminded everyone that transformation in strategic relations, most prominently between NATO and Russia, let alone transformation of the European security order, was not only occurring slowly and haphazardly, but was affected by the overlap of old and new practices and their contestation (Adler and Greve, Reference Adler and Greve2009; Pouliot, Reference Pouliot2010a, Reference Pouliotb).

In fact, NATO's success in imposing what it deemed the new practices of the post-Cold War game of international security to its former Russian enemy was relatively short-lived. By the mid-1990s, Moscow was already linking enlargement not to cooperative security practices, as per the Alliance's discourse, but to Realpolitik and Cold War containment. Despite a very proactive diplomacy, including a discourse of democratic peace and an unprecedented spending of resources, both material and symbolic, NATO was largely unable to set new terms of strategic interaction with Russia. In playing the game of international security, each side came to bring to the table increasingly distinct bodies of dispositions and expectations, as the Balkan conflicts demonstrate. The contemporary result is an awkward mixture of security practices in which the terms of strategic interaction combine strange bedfellows like spheres of influence and indivisible security, or common threats and mutual deterrence. What this story helps illuminate, among other things, is that the politics of practice can be grasped in the ways in which agents struggle to endow certain practices with political validity and legitimacy. New practices emerge out of authoritative definitions of truth and morality as promoted by certain segments of society; but this is a hard work of reification and power struggle.

Practices also interact with one another, insofar as the world is made up of various ‘constellations of practices’ (Wenger, Reference Wenger1998) or assemblages of communities and their practices that interact, overlap, and evolve. Practices in a constellation are interconnected – they share an epoch, a geographical place, a common object, a similar disposition; they react to the same conditions or perform the same functions, etc. It is the permanent state of connectivity and tension inside a constellation of practices that fuels transformation. For instance, not only are practices of cooperative security an assemblage of military, diplomatic, political, economic, and social practices, but they also consist of constellations of communities of practice, some of whose performances may be in the realm of peacekeeping, while others exist in the realm of economic integration. Looking at NATO's practices in the 1990s gives us an excellent picture of how practices come in bundles and constellations. Partnership for Peace, for example, emphasized military integration and cooperation and it eyed the prospect enlargement of the alliance. However, NATO practices of public diplomacy and seminar diplomacy, such as cooperative initiatives with the Mediterranean and Middle East countries, were aimed lower, namely, at achieving stability in, and exporting Western values to, those areas. In a similar fashion, most Western security practices toward Russia, in particular the development of a form of special relationship, brought together communities of practice that developed around deterrence and arms control, on the one hand, and cooperative security, on the other. As a result, the constellation of NATO-Russia security practices is infused with a potentially transformative tension between nuclear deterrence talk and everyday seminar diplomacy (Pouliot Reference Pouliot2010a, Reference Pouliotb).