Introduction

“Identity” has been one, if not the, conceptual shooting star in International Relations (IR) scholarship since the 1990s, at least among scholars seeking an alternative to the realist-rationalist vocabulary. And for its protagonists the concept is central to our understanding of international politics. David Campbell notes that “identity is an inescapable dimension of being. No body could be without it” (1992:9). Ted Hopf suggests that a world without identities would be a “world of chaos, a world of pervasive and irremediable uncertainty, a world much more dangerous than anarchy” (1998:175), and Anthony Burke claims that “there is […] no world politics without identity, no people, no states, no international system” (Burke 2006:394). According to these authors, “identities” manifest our ontology of the international and play a central role in politics. Bruce Cronin notes, “identities provide a frame of reference from which political leaders can initiate, maintain, and structure their relationships with other states” (Cronin 1999:18). These are strong claims, and reviewing them, that is, reviewing the place of “identity” in IR scholarship and assessing its analytical value, is not an easy task.

Crudely put, locating identity in IR can take two forms: one could try to abstract from different usages among scholars in the field and focus on what these usages have in common. Such an approach would attempt to describe the concept in a neutral way, geared toward presenting some sort of core meaning of identity in IR (for an attempt, see Fearon 1999; also Kowert 1999; Abdelal et al. 2006). Alternatively, one could maintain that such a neutral reading is impossible as different usages of “identity” are embedded in different/incompatible theoretical and normative positions, making it an essentially contested concept. Such an approach could do not much more than attempting to summarize the various, possibly contradictory, appearances and applications in the literature. For reasons which will become clear later on, this essay proceeds closer to the second approach, although by doing so it maintains that diverse applications do share important parameters. Overall, the essay will speak of an “identity perspective” to capture the different angles and suggests that a key function of “identity” in IR scholarship is that of an “eye opener” enabling scholars to see aspects of international politics they could/did not see before. This value added has an analytical and an ethical dimension.

Discoveries

My account of the migration of identity into the IR literature slightly differs from that of Lapid and Kratochwil (1996), who in an important publication suggested that the concept of “identity” (together with “culture”) had “returned” to IR in the early 1990s. A return implies that the concept was once present and was subsequently forgotten or ignored. The problem with this account is that, while the first generation of IR scholars discussed many of the phenomena now associated with studies of “identity,” like nationalism or norms, there was no substantial engagement with the concept itself.

While the nature of the Self and the meaning of personhood have occupied philosophers since Plato, the concept of identity entered the social sciences only in the 1950s with the work of Erik Erikson (Gleason 1983; Weigert 1983; Bloom 1990: ch. 2). And when in the 1960s it diffused from psychology into other disciplines, like sociology and anthropology, IR (at that time an “American discipline”) was taking the behavioral/scientific and, soon, systemic turn. This path took its main conceptual and methodological inspiration from micro-economic models and left little room for analysis of seemingly esoteric matters such as “identities” (let alone “identity crises,” the other term coined by Erikson). In retrospect, this neglect is ironic as the superpower/bloc confrontation seems to invite an analysis of foreign policy and international relations guided by images of “Self,” or community of Selves, and “Other(s).” To be sure, promising steps were made by Karl Deutsch in suggesting that transactions and communication may establish a “we feeling” through “partial identification in terms of self-images and interests” (Deutsch et al. 1957:36), with Kenneth Boulding’s notion of “national images” and Kal Holsti’s work on “role conceptions” held by foreign policy makers (Boulding 1959; Holsti 1970). Yet neither they, nor other prominent scholars like Robert Jervis, who highlighted the impact of perceptions and cognitive bias in decision making, made much use of the concept of identity (Jervis 1976).

Thus, when the term started to appear in the IR literature in the late 1980s and early 1990s it seemed fitting to speak of a “discovery” rather than a “return” (Enloe 1990; Bloom 1990; Wendt 1992; Campbell 1992; Dittmer and Kim 1993; see also Neumann 1999: ch. 1). This discovery was intrinsically linked to the rise of constructivism, whether in its “radical” or its “moderate” version (Hopf 1998). Although the research agendas pursued by these two strands, and, correspondingly, their understanding of what an identity perspective entails, differ and are often critical of each other, it is fair to say that identity is a constructivist concept if there ever was one. That constructivism is in the first instance a meta-theoretical commitment rather than a specific causal theory, as the term is generally understood, is important for understanding how an identity perspective has played out in IR scholarship (Guzzini 2000).

Identity/constructivism emerged in IR out of two overlapping trends: first, the postmodern Zeitgeist encouraged the questioning of accepted and “naturalized” categories associated with modernity. Embracing diversity and committed to an agenda of emancipation, postmodern thinking was to bring about the “death of meta-narratives” (Harvey 1989) and to unravel assumptions which had come to be taken for granted and justified with, for instance, the need for parsimony. In IR, this meant “fracturing and destabilizing the rationalist/positivist hegemony,” including its ontology of the international system, to establish a new perspective on world politics (Price and Reus-Smit 1998:263). The readiness to do so was aided, second, by the end of the Cold War and changing structures of governance. The dissolution of seemingly stable political entities such as the Soviet Union and Yugoslavia raised questions about the volatility of borders, loyalties, nationalism(s), and the ability to manipulate them. Simultaneously, the phenomenon of “globalization” and processes of European integration undermined the conception of the Westphalian state as the fixed/dominant entity in world politics. Against this backdrop, many IR scholars searching for new conceptual vocabulary turned to “identity” to highlight the socially constructed nature of the state and its interests, and to explain the causes of war and the conditions for peace.

Scholars applied an identity perspective to re-read the state (system) in various ways. William Bloom used the notion of collective identity to theoretically justify treating the state as an entity made up of individuals. He explored the phenomenon of “national identity” (another word for nationalism) to explain the identification of the “mass national public” with the state and the ability of the political leadership to represent/speak for the same (Bloom 1990). Alexander Wendt used it for systemic-level theorizing to offer a new reading of how states and the international system are constituted. Rather than viewing states as autonomous and competing power units, he argued that identities structure relations and vice versa, turning anarchy into a dynamic international society (Wendt 1992, 1994, 1999; also Dunne 1995). A number of scholars offered a mix of these two perspectives and used the concept of national/collective identity to deconstruct the “Westphalian model,” that is, the notion of the sovereign state as a fixed entity defined by a bounded territorial space. They pointed to its multifaceted and contingent nature, product of a history of shifting collective identities and social conventions (Ruggie 1993; Waever et al. 1993; Bartelson 1995; Cederman 1997; Ferguson and Mansbach 1996; Biersteker and Weber 1996; Reus-Smit 1999; Hall 1999). The emphasis on the social construction of borders also allowed the problematization of the political construction of “inside/outside” (Walker 1993) and the highlighting of the relevance of “cognitive regions” stretching beyond Westphalian borders (Adler 1997; see also Adamson and Demetriou 2007).

An identity perspective also promised an alternative to rational choice assumptions about interests (understood through the lens of “utility maximization”). In the attempt to endogenize interests, i.e., making their definition part of the theory, constructivists suggested that interests are dependent on conceptions of identity. Indeed, claims such as “identities are the basis of interests” (Wendt 1992:398) and “identities both generate and shape interests” (Jepperson et al. 1996:60) became a necessary routine in constructivist writings. Behind such claims was the argument that an actor needs to be aware of itself as a subject before it is able to formulate and pursue its interests (“act”). Said differently, “actors cannot decide what their interests are until they know what they are representing” (ibid.), or, simply put, “an actor cannot know what it wants unless it knows what it is” (Wendt 1999:231). With this assumption, proponents inevitably suggested that an identity perspective is superior to utilitarianism as it can explain all interests and the behavior said to follow from it, including that which appears “irrational” from a utilitarian perspective.

One area where the new perspective was put to use was the study of conflict. This ranged from Erik Ringmar’s argument that Sweden entered the Thirty Years War “in defense of its national identity” (Ringmar 1996:14), to Jutta Weldes’s account of how US decision makers framed the events known as the Cuban Missile Crisis in terms which allowed them “to re-enact U.S. state identity” (Weldes 1999:53; see also Campbell 1992). And when conflicts broke out in the early 1990s in former Yugoslavia, Chechnya, Somalia, Rwanda, and elsewhere, commentators and policy makers alike pointed to clashing collective identities embedded in claims of self-determination. Reflecting the debate between “primordialists” and “instrumentalists” among scholars of nationalism, they diagnosed either ancient ethnic rivalries or the political manipulation of identity narratives, frames which also featured prominently in Western debates over possible (military) intervention (Weber 1995; Crawford and Lipschutz 1997; Laitin 1998; Fearon and Laitin 2000; Hansen 2006).

The other major use of an identity perspective was to suggest that collective identities generate peaceful relations between states. Indeed, it was a dual use because it also saw the formation of such benign collective identities as an effect of cooperation. The two most prominent examples were Adler and Barnett’s (1998:48 ff.) adaptation of Deutsch’s argument to suggest that interaction may generate a “we feeling” and sense of community. Similarly, Wendt (1992, 1999) argued that states share collective identities which structure anarchy and that these identities may change and move relations from a situation where states identify each other as enemies toward one where they identify each other as friends (see also Cronin 1999). The argument has also been used to account for the “democratic peace” by suggesting that shared liberal norms and values generate a collective identity (Risse-Kappen 1995, 1996; Kahl 1998). Finally, scholars of European integration see evidence for processes of “Europeanization” and the emergence of a shared European identity among (representatives of) EU member states (Waever 1998; Green Cowles et al. 2001; Herrmann et al. 2004).

Disillusionments

From this brief overview it appears that an identity perspective offers significant analytical value. Upon a closer look the picture is less clear, however. While it spurred renewed attention in phenomena of “(trans)nationalism” and cultural parameters such as norms and ideas, this was not necessarily accompanied by a substantial engagement with the concept of identity. Indeed, the majority of the above-mentioned studies do not offer such an engagement. After a decade of using it, the concept remained vague and its analytical role difficult to grasp, prompting some disillusionment among IR scholars. Nicholas Onuf, who is credited with having introduced the notion of constructivism into the field, noted that identity was “one of the most fashionable concepts” but also “one of the murkiest – so difficult to fathom that […] I have been reluctant to use it” (Onuf 2003:26). Among rational choice proponents, James Fearon complained that the fact that there is no “short and adequate summary statement that captures the range of its present meanings […] amounts almost to a scandal” (Fearon 1999:2).

The lack of conceptual clarity can be attributed to the problem of identifying identities. How do we recognize an identity when we see one? The difficulty of giving a simple answer to what identity “is” translates into ambiguity about what it “does.” One could try and read the former through the latter. Analytically, this is exemplified in the seemingly trivial notion of identity as a “basis” or “origin” of interests noted earlier, which may tempt scholars to see research on identity simply as going one step back in the causal chain and explaining interests. This is problematic for a number of reasons. To begin with, while the claim may sound intuitively plausible, it is generally mentioned in passing and without any further elaboration on the relationship between identities and interests. Without such an elaboration, however, the two collapse into each other analytically. Constructivists will hold against this that the relationship cannot be grasped in familiar causal terms, that is, the two cannot be thought of as discrete variables that can be placed in a neat temporal sequence. Rather, certain interests are said to be intrinsic to certain identities; as Wendt (1998) puts it, identities “constitute” interests. While constructivists are not always clear what is meant by this term, it highlights that the explanatory logic of identity based arguments is different from the causal logic employed by positivists, with important methodological implications.

Even so, there remains ample room for confusion about how to conceptualize the relationship between identities and interests, including whether it makes sense to conceive of identities as a “basis” at all. Consider a simple example: A person claiming the identity of “professor” presumably wants to teach and/or do research. Yet here we already have two different and, according to most academics, conflicting interests. Moreover, it is likely that interests in teaching and/or research precede gaining such academic status and, instead, lead to becoming a professor. Thus, we are caught in circular reasoning and have to concede that such interests are held not only among those with the recognized status of “professor.” Then there is the question whether all interests are tied to an identity or whether some can be conceived of outside an identity framework, such as economic interests (“wealth”) or security interests (“survival”). Scholars like Wendt (1999) suggest the latter, which seems to severely compromise the “identity constitutes interests” logic by implying that identity does not go all the way down but is merely as a property of the state alongside others, such as territory or natural resources.

The view of identity as a “basis” is complicated further by holding that identities are not naturally given but constructed through ideas, norms, values, symbols, discourses, and practices, often subsumed under the label “culture.” This is not the place to examine the nature of these cultural parameters, each of which is difficult to grasp on its own. It suffices to say that they are generally seen as carriers and/or expressions of meaning structures through which a sense of “identity” is obtained and maintained. They provide frames which allow individuals and collectives to situate themselves in space and time and, in that sense, are orientation devices, a popular reading that IR scholars take from cognitive psychology (Jepperson et al. 1996; Hopf 2002).

Yet trying to shed light on this function and sorting the relationship between culture and identity raises a number of difficult questions. To take one prominent cultural parameter among constructivist scholars: which norms matter and why? When and how do they “constitute” identities? Does the identity-constituting (or “shaping”) function of norms sit alongside their regulative (or “constraining”) function? How do we tell one from the other? The fact that these questions appear more about the role of norms than identities is telling and poses the broader question of how processes of identity formation are tied up with the effect identities have on interests (and, by extension, policies). Put in simple terms, if the constitutive logic runs: culture → identity → interests, what role does “identity” play here? Is it a funnel, a catalyst, or a transformer? In positivist jargon, is it an intermediate or intervening variable? Again, constructivists will reply that any attempt to read identity as a variable in a causal chain is misplaced as culture, identities, and interests are mutually constituted. Yet unless this process is specified, the impression remains that much of the analytical lifting is done by “culture,” which raises the question of what “identity” adds to the picture. If identity functions as “a crucial link between environmental structures and interests” (Jepperson et al. 1996:59), with environmental structures taken here to mean “culture,” what is the nature of this link and what makes it crucial? The attempt by Jepperson and colleagues to clarify the link is curtailed by a graph showing multiple links and processes, exemplifying the confusion surrounding them (Jepperson et al. 1996:53).

Another major problem is that constructivist language often masks an essentialist reading of “culture” and, hence, of “identity,” a problem which has been pointed out and criticized by radical constructivists (Zehfuss 2001). This critique is not simply motivated by dogmatic anti-foundationalism and a commitment to an evolutionary ontology; it is informed by the observation that identities, or the cultural parameters they rely on, are contested within societies. Shifting attention to the phenomenon of “identity politics” reminds us that identity formation is an active/ongoing process and, if put in drastic terms, that identities are contingent constructs. This brings in agency and guards from an overly structural understanding of identity formation popular among those swimming in the stream of Durkheim and Parsons. Yet it also raises difficult questions about how stable or contingent identities are, whether they change gradually or suddenly, under what conditions such change occurs and, indeed, how to conceptualize identity change in the first place.

An overarching problem facing identity research is that the concept is used to make sense of seemingly opposite/contradicting phenomena. The brief overview given earlier shows that IR scholarship uses an identity perspective to view the state as both a coherent and a contingent entity by suggesting that collective identity can both constitute and undermine states; it is used to explain conflict with deeply rooted antagonisms or manipulative politicians, and to give an account of what creates peaceful relations. It is not surprising that critics find this multipurpose function of “identity” suspect. In a much noted article, Rogers Brubaker and Frederick Cooper (2000) summarize this critique by complaining that “identity” is used to describe both sameness and difference across time and space, designating both a foundational, essential element of human action and a fragmented, fluctuating Self. As they note, “it tends to mean too much (when used in a strong sense), too little (when understood in a weak sense), or nothing at all (because of its sheer ambiguity)” (Brubaker and Cooper 2000:1). If identity is what you make of it, it risks becoming an analytically empty and even “dangerous” concept (Narr 1999).

Although Brubaker and Cooper comment on the use of the term in sociology, it is not difficult to agree with them that “identity” has been used too loosely by those advocating its importance in IR as well, acknowledged even by scholars sympathetic to the concept (Chafetz et al. 1999; Weller 2000; Lapid 2001; Hopf 2002). Yet does this mean we should conclude that “identity is too ambiguous […] to serve well the demands of social analysis” and stop using the term (Brubaker and Cooper 2000:2)? While it is necessary to remind enthusiasts to think more carefully about what “identity” is doing in/for their analysis, moving “beyond” seems a classic case of throwing the baby out with the bathwater. If vagueness and multiple meanings disqualify a concept, then IR scholars also must ban terms like power, interest, or the state from their vocabulary. And apart from the fact that “demands of social analysis” are open to negotiation, it cannot be denied that research applying the term offered new insights into phenomena of international politics. Thus, as one scholar puts it, “it would not be fruitful to reject studies of identity […] merely by demonstrating that [our] understanding of the phenomenon is ridden with contradictions. There is ‘something’ that is being investigated” (Wagner 2002:48). And so, instead of discarding the term, IR scholars realized that a more careful engagement is necessary, whether by linking it more directly with the study of borders and orders (Albert et al. 2001) or by suggesting that endogenizing corporate identity was the “next step in constructivist IR theory” (Cedermann and Daase 2003).

What an Identity Perspective Has to Offer

Perhaps the most basic way to approach the concept is to say that the identity of X is the answer to the question: “Who/what is X?” Yet answers may differ depending on who is asked. We might decide that the answer which matters most is the one given by X itself (if it has a voice). Leaving aside for a moment that one may still get multiple answers, IR scholars seem to agree that, in the most basic reading, having an identity means having a sense of Self.

To get a better grasp of the analytical value of studying senses of Self in IR, it is useful to take another look at the suggestion mentioned earlier that some interests can/must be conceived of outside an identity framework. An identity perspective “all the way down” must refute this suggestion. Against the view that economic interests can be separated from identity claims, it would hold that the desire to accumulate wealth is tied to a conception of value, the meaning of which arguably is a social construct and embedded in images of the good life and of status one identifies with. The analytical purchase of an identity perspective is even more visible with regard to the alleged “core interest” of survival. Constructivists have made significant inroads into security studies re-examining the meaning and study of security (Katzenstein 1996; Krause and Williams 1997; Adler and Barnett 1998; McSweeney 1999; Farrell 2002). One common denominator of this literature is the claim that “the issue of identity […] is inseparable from security” (Booth 1997:88) or that “national security depends on national identity” (Kowert 1999:1). This claim is made on the level of ontology and is embedded in the reconceptualization of “the state” mentioned earlier, with the argument that the referent object or, rather, subject for security policy is the “identity” of society/the individual (Waever et al. 1993; McSweeney 1999). Rather than focusing on material properties of the state and concerns over “physical security” (or survival), it emphasizes cultural properties and the desire to maintain a stable sense of Self, or “ontological security” (Steele 2005; Mitzen 2006). In other words, an identity perspective generates its own conception of what it means for X to be (secure) and, consequently, what it means for X to survive.

This angle opens up a different way for thinking about the relationship between identity and interests. Instead of asking how the former constitutes the latter, it emphasizes the interest in safeguarding a certain identity, that is, it shifts the focus toward the “will-to-manifest-identity” (Hall 1999:6; also Bloom 1990). Assuming such a “will” suggests that an identity perspective relies on assumptions about human nature. In the literature the reading of identity as a stable sense of Self is variably linked to the human desire for cognitive stability (Kowert 1999; Hopf 2002), a moral compass (Reus-Smit 1999), self-esteem, and an “inherent sociability” (Mercer 1995:245) based on the desire for belonging and recognition (Ringmar 1996; Greenhill 2008), and expressed in the need for identification with or against “Others” (see below). It implies a human need for boundaries and bonds, and it confirms the importance of goals such as “prestige” or “honor” noted by classical realists (Lebow 2008a). Such assumptions may not always be openly voiced, in particular if the author subscribes to an anti-foundationalist stance, yet they seem to inform “identity” research in one way or another.

All this allows us to discard the misconception that an identity perspective is a “gap filler” useful only to explain “irrational” behavior. If rational action means action geared toward achieving a desired goal, then any attempt to manifest identity is rational and could even be read in cost/benefit terms. Thus, the aim of an identity perspective is not to explain altruistic behavior, an impression one may get from Wendt’s point that states sharing a collective identity will act out of “other-help” instead of “self-help.” This can be misleading as an identity perspective does not discard the idea of self-interested (or self-help) behavior. Rather, it shifts attention to the meaning of the “Self” in front. Such a shift does not reduce identity research to an “underlaborer function” of explaining interests but moves to the forefront the question how the “will-to-manifest-identity,” the desire for a stable sense of Self, is satisfied. It opens up new questions of how to deal with the tension between the assumed desire for stability, on the one hand, and the conception of the Self as evolutionary (always in the making) and multiple (varying with context), on the other. While few IR scholars have explored this tension, it informs work looking at active processes of identification and stressing the “performative” aspect of identity politics (Campbell 1992; Weldes 1999; Laffey 2000). It suggests that identities require constant affirmation and can be only temporarily stabilized. Identity claims are enhanced or undermined by certain practices, that is, what one “is” (or wants to be) is sustained by what one “does,” thereby blurring the means/ends distinction for understanding motivational aspects behind the will to identity.

Focusing on processes of identification also offers a useful way to distinguish between “culture” and “identity.” Whereas culture can be seen as a rich yet sufficiently vague web/depository of various structural parameters in a society (ideas, norms, symbols, discourses, etc.), identity claims are usually more sharp in attempting to define/clarify one’s position in a certain situation or relationship: “Identities simplify […] by making the unfamiliar familiar” (Hopf 2002:6). Understood as an active attempt to reduce complexity, their sharpness rests on a selected interpretation of certain cultural parameters, which makes identities (or identity claims) inevitably thinner than “culture.” While successful constructions of identity may render them social facts, such claims and the interpretations they rest on remain vulnerable precisely because they are thin and selective. Hence, they can be and often are contested, which generates identity politics (Reese-Schaefer 1999:18 ff.; Zehfuss 2001). Questions arising here concern when contestation turns into conflict, when conflict is destabilizing and generates an identity crisis, and what kind of politics we can witness in an identity crisis. IR scholarship has yet to fully address such questions (for first steps see Bially Mattern 2005; Steele 2005).

Arguably the most important facet of the will-to-manifest-identity is that it structures relationships. The widely accepted view among IR scholars, taken from social psychology, is that a sense of Self is socially constructed with or against certain “Others.” This opens up the conceptual terrain that all research taking an identity perspective must cross and come to terms with, namely the dual sources of identity (internal and external) and its seemingly dichotomous character (Self and Other).

The Internal/External Question

IR scholars generally adopt the view put forward by George Herbert Mead and Erik Erikson that identities are specific to relationships and defined in interactions, and that they have an “internal” (or individual, personal) and an “external” (or collective, social) dimension. Yet while the “internal/external” interface can be easily translated into the familiar relationship between the “domestic” and the “international,” specifying this relationship is anything but easy.

How to weigh the relative impact of intrinsic attributes and the socializing influence of the environment is the classic question underpinning the “nature/nurture” or “primordial/instrumental” debate. In social psychology it is captured in the interplay between personal and social identity and captured in Mead’s famous distinction between “I” and “Me” (Mead 1934). Mead argues that Self-fulfillment takes place through an evolutionary process in which the acting Self, the “I” driven by intuition and instinct, gradually comes to internalize the “attitude of the whole community,” devising a sense of “Me” by adapting to a social group (Mead 1934). Notably, the environment is reduced here to “Others.” This reading is followed here, although it is worth keeping in mind that conceptions of Self may be influenced in encounters with various environmental parameters, not just “Others.” To think about the relationships between Self and Other(s) requires unpacking the notion that identity refers to “mutually constructed and evolving images of self and other” (Jepperson et al. 1996:59) and that “individual and collective identity are co-constituted” (Hall 1999:36).

For IR scholars the task of unpacking such statements is not made any easier by first having to justify assigning an “individual” identity to aggregates such as nations/states which themselves are the product of a “collective” identity and, as noted earlier, often contested. And they have to account for how this national/state identity is embedded in a broader collective identity defined with external Others. The definition of identity as “the images of individuality and distinctiveness (‘selfhood’) held and projected by an actor and formed […] through relations with significant ‘others’” (Jepperson et al. 1996:59) and the suggestion that “individual identity acquires social significance only with reference to the identities of others” (Hall 1999:34 f.) overlap, yet the processes of projecting images and acquiring social significance can be read in different ways, depending on how one answers two simple yet crucial questions.

First, are images of Self and Other primarily defined internally or are they imposed from the outside? This question requires researchers to assess to what extent a state (or any other actor) gains its “identity” by (i) projecting domestically defined images outward into a collective, perhaps even defining the latter, or (ii) adapting to a certain conception of collective identity defined by external others. A sophisticated answer will not focus on one or the other but would suggest that a significant aspect of international politics consists of negotiating between these two forces. Second, what is the content of these images and of what kind is that reference to Others? There is no single answer to this question either, as Wendt’s suggestion of a spectrum from “negative” to “positive” identification exemplifies (Wendt 1999). The reason for this is that the very notion of “identity” contains an insolvable tension in meaning both similarity and distinctiveness (Heidegger 1957/1969). This tension bedevils all identity based arguments and invites the incoherent and often contradictory use of the term. The answers IR scholars have given to these two questions fall on either side of the spectrum, which can be illustrated by reviewing how identity perspectives have been applied at different levels of analysis.

Sorting through Selves and Others

Scholars like Wendt, who pitch their argument at the systemic level, champion neo-Durkheimean structural reasoning and emphasize that identity emerges out of interaction and is shaped by processes of socialization. If states only interact long enough they will gradually “internalize” collective norms around which social/collective identities revolve (Wendt 1999: chs. 6 and 7). With this argument proponents try to escape the notion that states are cultural dupes who take on a designated role according to a given script; instead, they suggest that because scripts emerge out of and can be changed through interactions, roles can be redefined/recast. Thus, Wendt argues the international system can evolve from a “Hobbesian” to a “Kantian” one where formerly antagonistic Others may become more like Selves by sharing a collective identity in which states recognize each other as “friends” (Wendt 1992, 1999). This is a liberal/cosmopolitan argument stressing the possibility of progress and has close affinities with some English School writings on international society (Dunne 1995). The limit of this approach is that it integrates “identity” into a neorealist ontology by treating it as yet another property of the state. It tends to give the upper hand to the external dimension and chooses to ignore the domestic formulation/contestation of Self/Other images, suggesting they are a function of the system. Such a systemic perspective makes it difficult to assess the interplay between Self(Selves) and Other(s), and to offer substantial and sophisticated insights about processes of identity formation and its political implications (see, however, Wendt 2003; also Cronin 1999; Rousseau 2006).

Scholars have applied the identity perspective to the community/group level in two ways. The account first draws on Social Identity Theory (SIT) and describes collective identity formation through an “in-group” versus “out-group” dynamic. While the initial reason for categorization may vary, once an “in-group” is delineated its members become biased in support of fellow members and consistently discriminate against nonmembers. Thus, there are always two types of “Others,” those identified as insiders who share a collective identity and are more “alike,” and those identified as “outsiders,” who appear substantially different, possibly sharing a collective identity of their own. Moreover, according to SIT, differentiation grows “the more we identify with our group” (Mercer 1995:245). Consequently, this approach is critical of the progressive and (potentially) universalist reading noted above. Although SIT does not suggest in-group–out-group relations necessarily becoming conflictual, it has been used to buttress a neorealist ontology by applying an “in-group” perspective to the level of the nation/state (Mercer 1995; see also Rousseau 2006). Its logic also echoes in accounts of how groups of states have defined their “shared culture” against another group, such as in the conceptions of Europe/the West versus the Orient/East (Said 1979; Neumann 1999) and Huntington’s infamous image of clashing civilizations (Huntington 1993).

The second account combines the integrationist/socialization logic with a regionalism lens to argue that processes constructing collective identities are primarily internally oriented. Expanding integrationist theories by Haas and Deutsch, it suggests that collective identities form through interaction, possibly within certain institutions, which strengthens mutual understanding and trust among states without requiring negative Others as a constitutive force. This argument can be found in the work on security communities (Adler and Barnett 1998), processes of “Europeanization” among EU member states, and the formation of a regional identity among members of the Association of Southeast Asian Nations (ASEAN) (Acharya 2000). Research on Europe, perhaps the most carefully examined phenomenon of “regionalization,” also highlighted the pluralist reading of collective identities by showing how national identities contain different readings of “Europe” (Herrmann et al. 2004; Risse 2005; for more skeptical readings see Smith 1992; Strath 2002). While treating Others as potential candidates for integration allows avoiding the “Huntington trap,” it also risks downplaying their role and pushing them off the analytical radar. That said, scholars have begun to think more creatively about the construction of different kinds of Otherness and presented more nuanced accounts of Self–Other relations (Diez 2004; Rumelili 2004). One particularly important insight is that processes of differentiation do not necessarily take place in space but may also occur in time, as in Ole Waever’s argument that a European security community is formed against memories of Europe’s conflictual past (Waever 1998).

A third and more recent body of scholarship takes a closer look at what makes a “significant” Other by pointing to phenomena of collective identity formation in special relationships. This perspective adopts the assumption that states seek recognition but asks whose recognition matters most. It suggests that notions of “community” and “group membership” are too vague and do not capture that demands for recognition are often targeted at certain/very specific “Others.” For instance, Adler and Barnett emphasize the importance of a “core state” serving as a “magnet”; although, apart from noting that such a state exercises power via “positive images” and expectations about “benefits,” they do not elaborate on the nature of this magnet and how it works (Adler and Barnett 1998:33–45). Other work suggests that collective identity is formed and anchored in more intimate social settings allowing states to maintain their (claim to) authenticity. This perspective both allows for and requires a close analysis of relational processes within which a collective (transnational) identity is developed, whether these are particular friendships (Bially Mattern 2005; Berenskoetter 2007) or bilateral processes of rivalry (Mitzen 2006).

Finally, a growing body of work focuses on how conceptions of national identity are constructed at the level of the state, specifically on how images of Self and Other influence foreign policy. It highlights identity politics within rather than between states to show how political elites/governments use a certain reading of national Selfhood, including the stereotypical images of “Others,” to justify certain policies. By tracing the instrumental use of identity-constituting narratives and the consequences, this perspective most clearly shows the agency exercised by political leaders. At the same time, it points to the fact that national identities and the cultural parameters within which they are embedded are contested within societies; yet it also reveals that participants in this debate cannot make arbitrary claims/interpretations if they wish to be successful in persuading their constituency. Valuable studies have been done on a variety of countries, including the US (Campbell 1992; Weldes 1999; Nau 2002); Russia (Prizel 1998; Hopf 2002); Germany (Berger 1996; Duffield 1998), and China (Dittmer and Kim 1993; Suzuki 2007). While these works offer insightful analyses of processes of domestic identity formation, the perhaps inevitable downside is a neglect of the external dimension and with it the risk of reducing the Other to a mere image, projected outwards and imposed upon others with little agency.

Of course, much of international politics occurs in spaces which cannot be captured by those four levels of analysis (systemic, regional, bilateral, state) but cuts across them. These spaces are filled and created by a variety of actors (communities/groups) besides (below, above, and in between) states involved in processes of collective identity formation. For instance, scholars have traced how NGOs diffuse norms and understandings of appropriate behavior around which identities form (Keck and Sikkink 1998; Risse et al. 1999), analyzed the political mobilization of collective identities among diasporas (Adamson and Demetriou 2007; also Laitin 1998), and debated the possible formation of a global civil society (Weller 2000; Chandler 2004). Constructivists also tend to highlight the role of (international) institutions in shaping the identities and interests of members, either by embodying the cultural parameters through which images of Self and Other are mobilized, or by providing sites of interaction where social learning/socialization occurs (Risse-Kappen 1996; Checkel 1999). These institutional functions can also be read into any of the four perspectives outlined above. The same goes for the literature inspired by Foucault and Bourdieu, which points to international networks and processes of governance manifesting certain conceptions of normality, thereby also defining categories of subjectivity for those (involuntarily) participating in this order. Scholars have shown such processes to be at work across policy realms, including cases of state building and the securitization of immigration (Merlingen and Ostrauskaite 2005; Huysmans 2006).

Whatever the angle, the challenge remains to take a closer look at the intersubjective dimension of “identity” by unpacking the processes by which Self/Other images are constituted within certain relationships. It is not sufficient any more to suggest that actors will play a certain role according to the rules of “group membership,” or to focus only on what actors perceive their role to be. Instead, in line with the emphasis on “processes” (Jackson and Nexon 1999; Lapid 2001; Guillaume 2007), analysts need to assess carefully what goes on in the “inter” (Kratochwil 2007). Constructivist research has come some way in doing so by investigating the microdynamics of identity formation and their political nature, whether this entails bordering or bonding, recognition or socialization. Scholars studying the discursive representation of events have shown how these must resonate with certain narratives of Self and Other to enable certain policies (Weldes 1999; Hansen 2006), whereas those following Giddens and pragmatist philosophy trace the constitutive function of shared practices (Mitzen 2006; Adler 2008). Such research allows both the refining and the refuting of some macro-level arguments. For instance, studies have gone beyond the assertion that “identity requires difference” (Connolly 1991) and become more nuanced in carving out the plurality of Self/Other conceptions structuring social relations (Odysseos 2007). And research has shown that actors adhere to (collective) norms because these are embedded in historically loaded structures of meaning, suggesting that identity claims do not simply rest on norms as such (Hopf 2002). More recently, scholars have pointed to the emotional dimension of identity politics and the “affective energies” contained therein (Ross 2006), and discussed phenomena of (soft) power exercised between actors sharing a collective identity (Bially Mattern 2005; Berenskoetter 2007).

A Note on Methods and the Normative Dimension

There is no single method for doing research from an identity angle. IR scholars have found themselves drawing on investigative tools and techniques from sociology, psychology, geography, or anthropology, and have used approaches ranging from discourse analysis to computational modeling (Hopf 2002; Herrmann et al. 2004; Bially Mattern 2005; Hansen 2006; Rousseau 2006; Abdelal et al. 2006). While this is not the place to review them all, it should be clear that an examination of how (the making of) frames of Self and Other guide policies faces methodological challenges for which some approaches are better suited than others.

IR scholars applying an identity perspective must sort through the difficult question of who the relevant carriers of Self/Other frames are, and who possesses agency in the construction process. An engagement with anything but individuals – whether states, institutions, or networks – must be prepared to address the charge of anthropomorphism, that is, of attributing human needs and characteristics to collectives. The infamous “as if” assumption certainly does not carry very far if one wishes to conduct empirical research. A better strategy may be to treat certain individuals, their discourses and practices, as representatives for the entity (or entities) being studied and to justify this step both logically and empirically. In any case, it seems unlikely that an understanding of the construction of Self/Other images and their political impact can be gained from the “outside.” There will always be scholars using positivist terminology by calling identity a variable and some may even attempt to treat it as such (Abdelal et al. 2006; also Hopf 2002). Yet to the extent that a study of conceptions of Self (or Selves) involves the study of representations, meaning(s), and even emotions, researchers will not get around an interpretative approach allowing them to gain sufficient knowledge of the salient cultural parameters mobilized in the process of identity construction and to grasp their political force.

Beyond the analytical purchase and the methodological challenges, it must not be forgotten that an identity perspective also has an important normative dimension. Broadly speaking, the study of why/how actors come to “hold” and “project” certain images of Self and Other necessarily brings with it a critique of the illusion of a “true” identity. And it requires paying attention to the ethical consequences of drawing boundaries around conceptions of Self and of designating Others.

This dimension was of course central to the first group of scholars advocating an identity perspective. It motivated their critique of and resistance to mainstream (read: rationalist/neorealist) scholarship in IR, which was accused not only of assuming fixed identities in international relations but also of defining “inside” and “outside” within the discipline itself by drawing boundaries of what was considered acceptable scholarship, excluding and discriminating against those who did not subscribe to the same canon (Connolly 1991). Feminists were among those sensitive to the construction of stereotypes and the unreflexive treatment of difference (Enloe 1990), an agenda which overlapped with postcolonial scholarship and its critique of Western categories (Paolini 1999; Inayatullah and Blaney 2004; Agathangelou and Ling 2009). Their call for problematizing “identities” thus went hand in hand with a critique of ontological and epistemological orthodoxy in IR scholarship and with the agenda of inserting voices that had been absent from the construction of knowledge (Tickner 1996:151). Inserting the study of identity into IR was thus understood as a power struggle about the possibility for change and the openness of the future.

One does not need to be a feminist or postcolonial scholar to appreciate the importance of this move, that is, to appreciate that denaturalizing identity claims and recognizing the multiplicity of “Selves” and “Others” opens intellectual and political spaces in which the formulation of alternative identities becomes possible. Ned Lebow (2008b) draws on the ancient Greeks to argue that identity does not require negative stereotyping; and even the progressive arguments of moderate constructivists like Wendt open up new ways of how (collective) identities could be thought and, consequently, lived. And pointing toward new possibilities will always be necessary. IR scholars taking an identity perspective cannot ignore that fixing images of Self and Other is a practice very well understood and carried out by intellectuals and policy makers alike. Samuel Huntington may be the most prominent example of someone who consistently advocated stable frames for a certain particularistic conception of “the West” and of “America” (Huntington 1993, 1996, 1997, 2004). Such frames may provide comforting clarity in times of perceived flux, yet they also do harm: whether in the discourses of al-Qaeda, the Manichean worldview of the former US administration of George W. Bush, or immigration policies in Europe – cultural parameters serving as seedbeds for collective identities continue to be mobilized by political elites to justify violent practices (Sen 2006). To the extent that IR scholars have an ethical responsibility to be reflexive about the impact of their work, a critical stance vis-à-vis practices of defining Self/Other images should be common sense.

Conclusion

The concept of “identity” has come a long way in International Relations scholarship, and despite all its vagueness it is here to stay. If we accept that there is a “will to manifest identity,” then the processes by which boundaries are drawn around conceptions of Self, such as a national identity, must be considered a central dimension of international politics. And so, while the decade of discovery starting with the end of the Cold War was followed by a moment of disillusionment, deflating some of the enthusiasm for “identity,” we have since seen a period of refinement in the attempt to understand world politics from an identity perspective. For most IR scholars, this does not entail adding the concept to the existing (main)stream of research by integrating it in traditional analytical schemes. Rather, it seems that an identity perspective has motivated them to assemble a new conceptual toolkit, which enables us to see and make sense of phenomena we did not see before. Thus, instead of trying to grasp it as a variable the concept of identity is perhaps better understood as an “eye opener,” which reshuffles the way we think and tell stories about international politics.