Karsten Friis


The article discusses the rise of societal and ethnic conflicts. It focuses on some of the analytical tools designed to illuminate these processes. More specifically it builds on the framework of the so-called "Copenhagen School of Security Studies" but seeks to develop a more flexible approach on the concept of identity. To understand how an identity gains strength, Roland Barthes’ theory of myths is examined and combined with Michel Foucault's concept of power. Finally a concept of agency is added, namely what has been labeled a “conflict entrepreneur”. It is argued that by combining these theoretical insights the process of constructing Self and Other, and the subsequent alienation of those who not fit these categories, will be better understood.



            Why is it that people are willing to die and kill for their communities? How can an identity be so strong that it is not worth living without? What makes “ethnic” or other identity-based wars often more brutal and violent than traditional wars?

            To illuminate these questions, this article will discuss the concept of societal security as used by the “Copenhagen School of Security Studies” (McSweeney, 1996). Through a brief discussion of different approaches to national identities, the relational and boundary drawing aspects of all identities will be underlined. To understand how an identity gains strength, how it becomes an “identity trump”, Roland Barthes’ (1993) theory of myths will be examined. According to him, a myth is a depoliticized speech; it turns history into nature and makes the ambiguous ordered. This way an identity can be associated with a persistent core of values or culture to which a people identifies. Further, this will be combined with a theory of identity building based on the Foucauldian concept of discourse, which underlines the complexity of an identity process and its decentralized power.

            Even though the discursive field is important for identity building, some sort of agency must be introduced to the securitizing process. A “conflict entrepreneur” has been suggested for this purpose (Eide 1997). Such an agent provides both a representation of a world of chaos and a necessary solution to this. This is done through “natural” and depoliticized myths, which provide legitimacy for the entrepreneur and his policy. This may mean suppression or expatriation, making the in-betweeners, the liminars, match the map of only Self and Others.


The Copenhagen School of Security Studies

            Over the last couple of years, the framework of the so-called Copenhagen School of security studies has inspired a growing number of studies and triggered several theoretical debates. The growing importance of the Copenhagen approach is probably due to its wide applicability on almost any empirical matter as well to a theoretical approach, which highlights the procedural aspects of conceiving a threat and taking actions against it. Let me begin with a short summary of some of the Copenhagen School’s main insights.

   In Ole Wæver’s (1993:8) definition, “‘[s]ecurity’ signifies a situation marked by the presence of a security problem and some measure against it.” Security is therefore a reaction to a perceived threat. Contrary to most other theoretical approaches to the study of security, the Copenhagen School deliberately avoids the ontological debate of what is “most threatening” (military, environment, poverty, and so forth). Instead they have developed a framework that can be used on all fields, by focusing on the process of labeling a threat, the process of securitization: “They have to be staged as existential threats to a referent object by a securitizing actor who thereby generates endorsement for emergency measures beyond rules that would otherwise bind” (Buzan 1997:13). This process is the same in all sectors, military as well as non-military, and is the core of security as a social process. The Copenhagen School thereby rejects security as something objectively “given”, but regards it rather as a social process applicable to any perceived value, any chosen referent object. A referent object is thus what is considered to be existentially threatened by the securitizing actor, traditionally the state. But anything can be made into a referent object. Security is a social construct and must be analyzed as such.

The central point in the Copenhagen School framework is therefore the process of securitization, by a securitizing actor with a referent object. The action of securitizing is labeled

“…a speech act, it is not interesting as a sign referring to something more real, it is the utterance in itself that is the act: by saying it something is done (like betting, giving a promise, naming a ship). By saying ‘security’ a state-representative moves the particular case into a specific area; claiming a special right to use the means necessary to block this development” (Wæver 1993:7). 

The subjective and social construct of security is thereby underlined. This way the Copenhagen School opens for a constructivist approach to security studies, dismissing any “objective threats” and regarding security and securitization as an active social process. 

Societal Security

To apply this theory to communities other than states, such as ethnic groups, the Copenhagen School introduced the concept “societal security” (Wæver et al. 1993). The basic argument is that whereas a state needs to secure its sovereignty in order to survive, a society will secure its identity. According to Wæver et al. (1993:21), a society is defined as “...having a high degree of social inertia, a continuity often across generations and a strong infrastructure of norms, values and ‘institutions’ in a wider sense.” The referent object in a securitization process will be society (Wæver et al. 1993:26), represented by the collective identity of this society. Thus, “[s]ocietal security is about situations when societies perceive a threat in identity terms” (Wæver et al. 1993:23). The collective identity is therefore what the society considers vital for them in order to remain a society. If something threatens the core symbol of “us”, if there would be no “us” without it, then a call for protection would be expected. Thus, whatever the members of a society perceive to be core values or symbols can become a referent object in a securitization process.

This seems promising, but there are some complicated aspects associated with such an approach. A major question is how we analyze this process; how we deal with such a vague concept as a collective identity. What is identity, and how is it articulated? How can it be securitized? Is it something that is “there” prior to a securitization, or is the securitization also an identity building process? I will argue the latter (see also Albert 1998, Huysmans 1998), and try to show this by discussing the process of identity building more closely, and argue in favor of a position that regards identity as an on-going process.

The Ontological and Epistemological Questions of Identity

            How should the concept of identity be treated analytically? A constructivist ontology would emphasize the mutual constitution of agent and structures (Hopf 1998:172), thus focusing on the ongoing process of identity formation. Given their emphasis on the socially constructed character of security, one would expect the Copenhagen School to also regard identity in constructivist terms. But a critique has been risen by Bill McSweeney (1996) who claims that the Copenhagen School has an excessively rigid view of collective identity. This critique is mostly based on the book by Wæver et al. (1993), which also presents the Copenhagen School’s deepest analysis of societal security to date. McSweeney seems to have a point. There are tendencies toward a rigid and objectivist view of identity in the book. In phrases like “Societal security concerns the ability of a society to persist in its essential character under changing conditions and possible or actual threats” (Wæver et al. 1993:23, my emphasis), the Copenhagen School seems to emphasize a very traditional view of identity. To consider a society as having an “essential character” reflects a positivist ontology. Of course, it could simply be a badly formulated sentence, or that “essential” still leaves room for evolution. This seems likely since they later have claimed that they are well aware of the constructed character of identities and thus have rejected McSweeney’s critique.

            Still, their position is a little vague and unclear. In his article, Barry Buzan (1997:19) explicitly argues that “...even the socially constituted often gets sedimented as structure and becomes so relatively stable as practice that one has to do analysis also on the basis that it constitutes...” Similarly, in their answer to McSweeney, Buzan and Wæver (1997:243) argue that, “...when an identity is thus constructed, and becomes socially sedimented, it becomes a possible referent object for security.” Also, in their latest book they (Buzan et al. 1998:121) present what they consider “...the most common issues that have been viewed as threats to societal security.” These are all based on that “people X” are being challenged by this and that. “People X” already exists; their identity and values are there before the threat begins. At another point, Buzan and Wæver (1997:243) argue that “[t]his does not imply that identities do not change, only that we should not expect everything to change all the time: certain things stay the same throughout the period relevant for analysis.” They somehow seem obsessed with a need to “fix” an identity when they conduct their analysis of securitization. This implies overlooking the process of mutual constitution, in this case between the securitization process and the identity referent. Thus, their epistemology seems to be at odds with their constructivist ontology. This could lead to a position that holds an identity to be independent of the securitization.

            As McSweeney (1996:84) points out, the Copenhagen School uses phrases like “independent variable” (see Wæver et al. 1993:185) to describe a society in a securitization process. This also is consistent with a positivist epistemology. Following the logic of a constructivist ontology (mutual constitution), there can never be such thing as an “independent variable” in social science. All social constructs are always in-the-making and have impact on each other. Nothing is ever independent. The securitization process will also always feed back on society. Identity is not just there; it is constantly represented and struggled over, and securitization will in most cases be a part of that process. Identities are social processes and must be analyzed as such, not simply as objective givens that can trigger securitization. Thus, to study identities one must try to grasp how they emerge and change. To simply describe them at a given point would be to miss that. In Espen Barth Eide’s (Eide 1998:71) words: “...we may choose to think of securitisation as a process which simultaneously constitutes the group and securitises it towards other groups” (original emphasis).[2] Identity or society is not a constant. It is not something objectively given, and because it is struggled over it simply cannot function as an objective foundation from which securitization derives. This way of thinking may lead back to the rigid positivist and neo-realist stance from which the Copenhagen School seeks to distance itself.[3]

            Further, even if one for analytical reasons accepts identity as “fixed”, for instance to be able to concentrate on other aspects (like a securitization process), this approach obviously loses valuable insights and aspects of both the identity building and the process of securitization. Of course, the empirical investigation required to be able to cover the whole process of change and mutual constitution is vast; “[c]onstructivism is no shortcut” (Hopf 1998:198). Yet, by combining the two processes a richer picture of the whole process will be obtained.

            The problem is that the Copenhagen School also argues in constructivist terms. For instance Buzan et al. (1997:120) write the following passage: 

“Threats to identity are thus always a question of the construction of something as threatening some ‘we’ – and often thereby actually contributing to the construction or reproduction of ‘us’. Any we identity can be constructed in many different ways, and often the main issue that decides whether security conflicts will emerge is whether one or another self-definition wins out in a society”

         This seems well in line with constructivist ontology on identity and with their own claims in the debate with McSweeney, but it is at odds with some of their other statements. It is unclear what the Copenhagen School position really is. However, regarding national identity, their position seems clear. They build on Anthony Smith’s (1983, 1991) theories. Below I will try to show the limitations of this approach and instead argue in favor of a position based on Benedict Anderson’s (1991) theory of imagined communities.

 National Identity

            There is no doubt that nations are historical constructs. They are results of a special European movement and some active nation building elite. Phrases like “the invention of tradition” (Hobsbawm and Ranger 1983) underline this fact. At the same time, nationalist movements have been trying to establish and argue that there is a history, a culture and a tradition from which the nations derive. This ambiguity can also be found in the theorizing of nationalism. The question is to what extent are the nations rooted in a historical and cultural base, or whether these links are merely created by the nation builders.

            Anthony Smith’s (1983, 1991) theories of nationalism are based on a so-called ethnie, which denotes the “core” of a nation. This perception of a common name, ancestry, culture, history and homeland, together with some sort of solidarity, is a necessary condition for a nation according to Smith. Thus, there is something essential in the culture, something that hardly changes or evolves. Smith acknowledges that these ethnies have to be engaged and given significance; they are not just “there”. However, they are nevertheless a precondition for a national identity.

            The Copenhagen School’s approach to national identity is largely based on Smith. It is in accordance with their view of identity as historically constructed. Simultaneously the approach emphasizes some core values (the ethnie) which are most likely to be securitized in a society. This seems likely enough. The problem with this approach, however, is that it cannot help us understand changes. It can illuminate how we got nations, and how they remain by relying on the historical background. But what if they do not remain? What if new identities emerge instead of today’s nations? Or what if the major symbols of a nation change? That can hardly be because of the articulation of an ethnie.

            If a new ethnic consciousness rises somewhere, its origins can be traced. Historical links and symbols as well as narratives and symbolic places can be found. But that does not explain why it suddenly emerged as it did, at that particular time. Attempting to do so on the basis of historical trajectories will run the risk of ending up with a reductionistic explanation of necessity. The rise of the group had to happen due to the historical circumstances. This, of course, is both dangerous and misleading. This is a reminder of one of the important ethical dimensions of poststructuralist theorizing. All so-called “objective” explanations are in fact subjective (and thus political), and all “objective variables” leading to arguments like “Y had to happen because of X” are dangerous, as they tend to disentangle the responsibility of a social action. If something happened because it “had to”, then who is to blame? If the historical background is used as an explanation, then what else could be done? Popular explanations of the war in the former Yugoslavia have, for instance, sometimes been based on a “lid-theory” saying that with the end of communism “ancient hatreds” bloomed as if someone had removed the lid of a boiling pot. This implies a dangerous historical necessity that should be avoided.

            If we stick to the former Yugoslavia, could the rise of the ethnic identities there be explained through a Smithean framework? Of course, the three ethnic groups had a knowledge of their ethnicity before the war. There existed some symbols, some historical narratives and the like. By searching, some sort of ethnies would be discovered. But that does not explain why and how these suddenly became so important, nor why they became securitized. This shows the limits of the focus of “essential characters” or an ethnie in the studies of securitization. Instead, we need to find out how something becomes considered essential, and that requires another approach.

            Actually, the Copenhagen School raises a similar question: “How is an identity forgotten and/or reinvented?” (Wæver et al. 1993:37). To them, Smith’s approach is best fitted to answer that question: “It might well be that Anderson and maybe Gellner have the superior theories of the origins of nationalism, but ... Smith ... probably has more to offer on the workings of an existing national community” (Wæver et al. 1993: 37). But this claim is never really followed up and justified in the book.

            I would therefore rather claim an opposite argument. Benedict Anderson’s (1991) theory is superior to Smith’s, especially “on the workings of an existing community”, because it is concerned with identity as a process. It is not restricted to one special historical happening (the rise of nations), but it tries to give a more general picture of how a group of people continuously experiences its common identity. His expression “imagined communities” underlies one important feature, namely the on-going process. People have to imagine a community, a society, or a common identity, for it to be real. It does not exist unless people think so, unless they feel some sort of solidarity with people they never have met.

            In Christopher J. Ullock’s (1996) words, Anderson’s approach is based on “a metaphysics of becoming” (as opposed to a “metaphysics of being”), something that makes his theory dynamic. It focuses on the process of identity formations and is not limited to its historical origins. This makes the theory more generally applicable, both to kinds of collective identities other than states and to nationalist movements outside Europe which have a totally different historical background. It can also be useful for understanding developments of new communities with new collective identities. A non­territorial global identity community, like the Hells Angels for instance, cannot be understood on basis of any ethnie, but it must be studied on the basis on their on-going self perception.

 Identity as Process

            Identity is not something that just is; it can rather be described as a process. Collective identities are never stable and objective but always in-the-making. They are subjectively experienced and expressed, a result of social communication and perceptions of Self and Others. A basic insight in all identity theory is the importance of an Other.[4] Without an Other there would be no Self. What constitutes an identity is therefore more the difference from the Other than the sameness of the Self. It is the definition of the differences that establishes the Self. In Anne Norton’s (1988:3) words: “Meaning is made out of difference”.

            This is a classic insight from Fredrik Barth (1969:10) who argued that ethnic identities were a result of communication with other groups: “...ethnic distinctions do not depend on an absence of social interaction and acceptance, but are quite to the contrary often the very foundations on which embracing social systems are built.” Identities are always a result of communication; they are relational.

            Further, the symbols that turns out to be of importance are randomly chosen: “The features that are taken into account are not the sum of ‘objective’ differences, but only those which the actors themselves regard as significant” (Barth 1969:14). You cannot tell in advance which features will be significant for a group’s identity; it is a result of the relationship and interaction with other groups. Normally, this will imply choosing symbols that are different from your neighbors’, thereby drawing a border and establishing the categories Self and Other.

            In short, identity building is a boundary-drawing process, an attempt to define something, to give it a meaning, and to establish an order of knowledge. A boundary in identity terms is not geographical but social. It is a way of defining the symbols, values and meanings that are central for a group. It is a deciding of the inside and the outside. The outsider can be geographically outside, as in an ideal nation, but also inside. Geographical borders are results of these social borders, not the other way around.[5]

            Those who do not fit into the definitions of the Self are different and may also be excluded, estranged or alienated. They may even be securitized and become enemies. This way the securitization also functions as an identity building process. The Others are securitized because they are considered to represent symbols and values that are incommensurable or threatening to those of the Self-group. I will come back to the relation to securitization.

            This approach does not neglect historical traits (like language, culture and so on). But as long as these have to be interpreted and represented through subjects and through subjective experiences and language, they are subjective social constructs. There is no direct correlation, no necessity between the ethnie and the national identity. Collective identity cannot be reduced to the ethnie. Even though traditions, language, myths, and other features existed, there were also plenty of other myths and traditions that were never turned into major symbols for the nation, and also plenty of potential symbols that transgressed the borders of what later became a nation. Thus, the representation of an identity, the common imagination, is of more interest than its historical background. Nations (and other collective identities) do have a history. They have risen in a special historical context, but that does not explain how they evolve and transform today. Collective identities are not here because of history but because of the way we use history. It is the re-cycling of historical narratives, symbols and myths that keeps a nation alive. This re-cycling also implies constantly adapting to a new context and a constant re-interpretation of the symbols. The meaning of these might change, even though the symbols as such remain.

            Still, this alone does not help us understand securitization. Of course, for an identity to be securitized, it must be considered important. What needs to be investigated is how an identity gains strength and is considered essential; how it becomes what Craig Calhoun (1994:11) has called a “trump” -- an identity that is stronger than “...all other forms of identity, including those of community, family, political preference, and alternative ethnic allegiances”. In Mathias Albert's (1998:27) words: “If we want to gain an understanding of ‘securitization’ that does not freeze one reference point, i.e., (societal) identity, but can account for identity’s processual character too, we have to be able to explain why some issues are more ‘prone’ to being securitized than others...” 

Trumps and Myths

            It is necessary to get a better grip on this rather vague concept of identity. How can it be operationalized for analysis? If identity building is about drawing boundaries, then it seems necessary to investigate this further. For instance, what is it that is protected by these boundaries, or what is it that is established as signs of the identity through this boundary drawing? Along which dimensions is the identity manifested? In other words, how is an identity articulated and expressed?

            The answer is normally through symbols around which everyone can cluster. However, anything can be a symbol, so that does not really help. The symbols must also represent a meaning and thus some sort of value. What is important is that they are not excluding anyone within the Self­group. They must be considered all-embracing, essential, and context-traversing. How is that achieved?

            One way it happens is by lifting the symbols out of the realm of politics -- by depolitization. A depoliticized speech (meaning any utterance, meaning or expression) is what Roland Barthes describes as a myth. Barthes’ (1993) classical work on this phenomena is still valuable. According to Barthes, a myth is not a false representation of reality, as it is often defined in daily usage of the word. Rather, it is a special way of representing reality. It is, in semiotic terms, a meta-language, or a second-order semiological system, where meanings are redefined and given a new content.[6] For example, histories about wars can function as a myth for those who hear or read it afterwards. The history itself can very well be true and accurate, but when it is read as a myth, it gives a specific cultural and political impact on society that may not follow logically from the original history.

            A central quality of a myth according to Barthes (1993:125) is that it “ speech stolen and restored”. The restored meaning is different from the original; it is simplified and ambiguity is made into order. A myth transforms history into nature, and makes speech innocent in the sense that it is naturalized (Barthes 1993:129, 131). In other words, a myth gives “...a historical intention a natural justification, making contingency appear eternal”, and thus a “...myth is depoliticized speech (Barthes 1993:142, 143, original emphasis). This is what gives it power. Barthes exemplifies this with a picture of a black man in French uniform, saluting the French flag. The picture speaks against colonialism as oppression and gives the impression of a Great Empire consisting of enthusiastic and patriotic citizens. Thus, the myth gets rid of the 

“...contingent, historical, ... fabricated quality of colonialism. Myth does not deny things, on the contrary, its function is to talk about them; simply, it purifies them, it makes them innocent, it gives them a natural and eternal justification, it gives them a clarity which is not that of an explanation but that of a statement of fact ... [I]t abolishes the complexity of human acts, it gives them the simplicity of essences...” (Barthes 1993:143)

            The effect is powerful when something social is made essential, when history is turned into nature, and when politics is depoliticized. If an identity is built upon such myths, it too will be essentialized, naturalized and considered beyond politics. It becomes a trump. If a community’s historical narratives are simplified and purified, perhaps represented as an eternal fight between Good and Evil, its identity building effect is enormous. Then, the Self is not only representing the Good; it is the Good. This way myths also have ethical aspects, as they can legitimize political actions that are based on simplified mythical interpretations of the past. For example, such representations (we are Good, they are Evil) leave no room for in-betweeners, for so-called liminars, who inhabit both the realm of the Self and the Other (Norton 1988). Liminality will always exist because all boundary drawing is limited, in the sense that all categories will contain some ambiguity. No nation, no gender, no meaning is ever “pure”. There are always some cross-overs who by their simple existence represent a political challenge to the order of things: “The mingling of identity and difference in the liminal challenges the integrity of the paranoid observer” (Norton 1988:55). Thus, a strong myth can have severe political implications because of its natural appearance and its apparent innocence. It can, for instance, serve to legitimize ethnic cleansing. I will come back to this later.

            This leads us to the question of power. Who controls or constitutes myths and identities? Is the meaning established through powerful actors and seductive ideology, or is the power less traceable in terms of locus? 

Identity as Discourse

            One problem with Barthes’ approach is his indication that it is the bourgeois who somehow controls the myths. He (1993:142) sees it as “ ...the most appropriate instrument for the ideological inversion which defines this society...” and part of the “bourgeois ideology”. The power of the myth is connected to this bourgeois ideology. Such Marxian views of power have been criticized, among other things for implicitly suggesting that there is an objective, hidden truth “out there” that the bourgeois is trying to mask by ideology and use of power. The idea of such a “truth” not only connotes a problematic objectivist ontology (“the real truth”), but it also turns power into a question of legitimacy (Hindess 1996). By disclosing the illegitimate bourgeois power of oppression, this implicitly means that there is another form of power that is more legitimate -- one which is emancipated from class-interests. Even though strong actors are often important, as we also shall discuss below, there is no doubt that collective identities can not be reduced to ideology or “false consciousness”. If people conceive it as real, an ideological conviction must be possessed to claim that they are “really” wrong. It also implies a very elitist self-perception of the scientist, since s/he is the only one who can expose this “truth”.

            Instead of regarding depoliticized myths as part of a bourgeois ideology, Barthes insights can be used in a fashion that neither places power within a (illegitimate) class perspective nor assumes a “hidden reality”. This can be achieved by analyzing myths, and identity process in general, as a discourse in Michel Foucault’s (1972) definition of the word. Then we do not have to rely on some strong powerful agents or on an all-embracing ideology that seduces the masses. Truth in a discourse is not based on powerful actors or objective realities but on a horizontal field that accepts or rejects utterances. This field is based on what has already been said and done -- not only through language but also through institutions, practices and so forth. In this way meaning and truth are established as well as objects of knowledge. Madness, for example, is a category, an object and a field of knowledge that is the result of a historical discursive development. Not only science but also institutions, therapists, and others participated in this discursive field. Thus, to Foucault, discourse is not simply a linguistic term; it also encapsulates the material realm. In this way, an identity is not simply a narrative but involves also institutions, practices and the like, which participate in the creation of signs, meanings or objects.

            With Foucault, power can be regarded as something productive and discursive, without a clear locus. It is producing categories, traditions and normality. Foucault (1973, 1977) focused on social categories such as madness and criminals as well as institutions like hospitals, prisons and assembly halls. However, the ordering of people into Selves and Others can easily be understood in the same way. The strong “normality” of national identity in many states and the often widespread feeling of centuries of national consistency are examples. Collective identities are constantly being built -- in sports arenas, in history books, in culture and in politics.

            This discursive normalization process of Foucault also carries some similarities to Barthes’ depolitization in the way that both are concerned with showing how something apparently essential is in fact elusive or a historical and social construct. It is the entire discursive field that moves something into the natural and depoliticized realm, which normalizes the myths and other symbols of which constitute an identity. The power in a successful depolitization or naturalization should not be underestimated.

            Collective identity is an on-going discourse, consisting of several actors, institutions, theories, and rituals. This discourse creates symbols and gives them meanings and values. In this way both the historical emergence of the categories Self and Other can be studied. For example this may be through a genealogy which is very conscious on avoiding historical “presentism”[7] and anachronisms -- “the fallacy of confusing rational functions for origins” (Price 1995:85) as well as the on-going dynamics of an identity today.

            It is sufficient here to note that this further develops Anderson’s dynamic “metaphysics of becoming,” as it gives us a better understanding of how identities can become “identity trumps,” and how they become depoliticized and “natural”. The “imagined community” possesses a referent object, something to imagine such as myths. Let me, instead, return to the question of securitization and see how such a process fits into an identity building discourse. 

Societal Securitization: Actors

            At its extreme, people are willing to die and kill for their identity. For this to happen, the identity must be considered existential. It must be considered something as basic as life itself, something of which you cannot live without. People will rarely kill for their companies, their choice of fashion, or some other “identity light”. They fight for the identity trumps. And as we have seen, trumps are trumps because they are depoliticized and “natural” within the group. I have argued this is due to the discursive field in which the identity is embedded. From this point of view, the naming of crucial values, and thus of Self and Others, is a discursive practice. It is not something which is done primarily by powerful agents but by a mix of these and historical traits, myths, values and traditions. Then what about securitizing, the process of moving something into a state of alert? Can it also best be analyzed as a discursive practice?

            If so, it seems to be at odds with the Copenhagen School’s emphasis on securitizing actors and “the speech act”. These concepts seem to point towards an agency oriented approach.[8] No doubt, securitization is often a very concrete political move, seemingly done by discernable actors at a given time. It is an observable process. In terms of societal security, it has been argued over and over that securitization is a process of making difference into radical otherness,[9] or in Copenhagen terms, moving the representation of the Other from a politicized to a securitized position. That does not happen by itself. Espen Barth Eide (1998:74) has launched the concept “conflict entrepreneurs” to grasp why difference sometimes becomes conflictual. By doing this he (1998:74) is “bringing agency and strategy back in”, because “...organised conflict in one way or another needs an element of active ignition, of agency”.

            Thus, a conflict entrepreneur is “ individual who takes the necessary and deliberate steps to ignite a violent conflict by utilising a specific situation or in order to gain something through the exploitation of new power relationships” (Eide 1998:75). This is done not necessarily for personal reasons but also for collective reasons. The point is that he is able to enter the ongoing identity process and manipulate and influence it. He thus has power to interpret and represent history and symbols in his own way, and represent the Other as an existential threat to them. Control over media is often important in this respect. There is little doubt that in cases like Bosnia, such entrepreneurs were important triggers to the conflict, as Eide and others (e.g., Majostorovic 1997) have demonstrated.

            Does this mean that the discursive approach to identity building is misleading? Not necessarily. The entrepreneur is also dependent upon a context which gives opportunities and limitations. For sure, securitization cannot be done by anyone at anytime with any referent object they like. In Eide's (1998:58) words, “[t]he decision to accept the securitization of a particular topic rests with the audience, not with the actor who speaks”.

            There is no need to reduce this to an agency-structure-debate, if that implies a need to decide, and an inherent either/or ontology (Doty 1997:374f). Following a constructivist ontology it is the process of mutual constitution which is of interest. Foucault's discourses might not leave sufficient room for political entrepreneurs, but introducing agency does not necessarily imply a rejection of the entire Foucauldian framework. Erik Ringmar (1996) has for instance introduced the concept “formative moment” to describe periods where there is more room for agency. According to Ringmar (1996:91) such a moment can be described as “...a period when new metaphors were launched, when individuals and groups told new stories about themselves, and when new sets of rules emerged through which identities were classified”. This is a moment of opportunity of which some entrepreneurs gain more influence and power than they otherwise would.

            There is power in the discourse that provides such opportunities. For instance, a king was no doubt a powerful person some centuries ago, but he was nevertheless dependent upon the political system that allowed a king and a kingdom. Thus, we cannot locate a locus of power once and for all; any actor acts in a context that never can be left out of the analysis

            Besides, it is not unusual that the securitizing discourse also includes (what comes to be) the Other. A conflict entrepreneur will gain a lot of help if someone else is arguing similarly in the other camp. Samuel Huntington's (1993) attempted securitization of “the Rest” (versus “the West”), is said to be popular reading among Islamic fundamentalists. They tend to agree from the opposite position.

            Let us therefore assume that a societal securitization process often includes some entrepreneurs who, due to the situation and discursive network, can maneuver themselves as representatives of the community, and thus define its boundaries. The entrepreneur might use some myths and/or contribute to its continued existence. What also must be done, however, is to provide an order or a world view of which the Self is situated. This means moving politics abroad, out on the borders, through what Hegel (1952) has called a “dialectic of conflict”. A part of the logic of delivering an order to a given group on the inside is that anarchy simultaneously is moved out into the spatial outside (Bartelson 1995:210ff). This creates two ontological rooms, one of order and one of anarchy and danger (Walker 1993). But how is this done? 

From Liminars to Others

            The problem with societal securitization is one of representation. It is rarely clear in advance who it is that speaks for a community. There is no system of representation as in a state. Since literately anyone can stand up as representatives, there is room for entrepreneurs. It is not surprising if we experience a struggle between different representatives and also their different representations of the society. What they do share, however, is a conviction that they are best at providing (a new) order. If they can do this convincingly, they gain legitimacy. What must be done is to make the uncertain certain and make the unknown an object of knowledge. To present a discernable Other is a way of doing this. The Other is represented as an Other -- as an unified single actor with a similar unquestionable set of core values (i.e. the capital “O”). They are objectified, made into an object of knowledge, by re-presentation of their identity and values. In other words, the representation of the Other is depoliticized in the sense that its inner qualities are treated as given and non-negotiable.

            In Jef Huysmans (1998:241) words, there is both a need for a mediation of chaos as well as of threat. A mediation of chaos is more basic than a mediation of threat, as it implies making chaos into a meaningful order by a convincing representation of the Self and its surroundings. It is a mediation of “ontological security”, which means “...a strategy of managing the limits of reflexivity ... by fixing social relations into a symbolic and institutional order” (Huysmans 1998:242). As he and others (like Hansen 1998:240) have pointed out, the importance of a threat construction for political identification, is often overstated. The mediation of chaos, of being the provider of order in general, is just as important. This may imply naming an Other but not necessarily as a threat.

            Such a dichotomization implies a necessity to get rid of all the liminars (what Huysmans calls “strangers”). This is because they “...connote a challenge to categorizing practices through the impossibility of being categorized”, and does not threaten the community, “...but the possibility of ordering itself” (Huysmans 1998:241). They are a challenge to the entrepreneur by their very existence. They confuse the dichotomy of Self and Other and thereby the entrepreneur’s mediation of chaos. As mentioned, a liminar can for instance be people of mixed ethnical ancestry but also representations of competing world-pictures. As Eide (1998:76) notes: “Over and over again we see that the “liberals” within a group undergoing a mobilisation process for group conflict are the first ones to go”.

            The liminars threaten the ontological order of the entrepreneur by challenging his representation of Self and Other and his mediation of chaos, which ultimately undermines the legitimacy of his policy. The liminars may be securitized by some sort of disciplination, from suppression of cultural symbols to ethnic cleansing and expatriation. This is a threat to the ontological order of the entrepreneur, stemming from inside and thus repoliticizing the inside/outside dichotomy. Therefore the liminar must disappear. It must be made into a Self, as several minority groups throughout the world have experienced, or it must be forced out of the territory. A liminar may also become an Other, as its connection to the Self is cut and their former common culture is renounced and made insignificant. In Anne Norton’s (1988:55) words, “The presence of difference in the ambiguous other leads to its classification as wholly unlike and identifies it unqualifiedly with the archetypal other, denying the resemblance to the self.”

            Then the liminar is no longer an ontological danger (chaos), but what Huysmans (1998:242) calls a mediation of  “daily security”. This is not challenging the order or the system as such but has become a visible, clear-cut Other. In places like Bosnia, this naming and replacement of an Other, has been regarded by the securitizing actors as the solution to the ontological problem they have posed. Securitization was not considered a political move, in the sense that there were any choices. It was a necessity: Securitization was a solution based on a depoliticized ontology.[10]

            This way the world-picture of the securitizing actor is not only a representation but also made into reality. The mythical second-order language is made into first-order language, and its “innocent” reality is forced upon the world. To the entrepreneurs and other actors involved it has become a “natural” necessity with a need to make order, even if it implies making the world match the map. Maybe that is why war against liminars are so often total; it attempts a total expatriation or a total “solution” (like the Holocaust) and not only a victory on the battlefield. If the enemy is not even considered a legitimate Other, the door may be more open to a kind of violence that is way beyond any war conventions, any jus in bello.

            This way, securitizing is legitimized: The entrepreneur has succeeded both in launching his world-view and in prescribing the necessary measures taken against it. This is possible by using the myths, by speaking on behalf of the natural and eternal, where truth is never questioned. 


            I have addressed the question of how an identity can become securitized and have argued in favor of a flexible and dynamic view of identity, based upon a metaphysics of becoming. This should help us avoid some theoretical pitfalls which could lead to a rigid objectivist and causal analysis. Through a brief discussion of different approaches to national identities, the processural, relational and boundary drawing aspects of all identities were underlined. To understand how an identity gains strength and how it becomes a “trump”, Barthes’ (1993) theory of myths was examined. As a myth is a depoliticized speech, it turns history into nature and makes the ambiguous ordered. It is thus easy to securitize. However, myths and identities cannot be established by strong agents alone. Identity can better be seen as a Foucauldian discourse, a productive field of order, truth and knowledge. On the other hand, when it comes to the securitizing process, some sort of agency must be introduced. The “conflict entrepreneur” may function as such an agent, even though still embedded in the discursive field. For him to succeed, he must provide both a world-view and a solution to the threats this view contains. He will have to make the uncertain certain by naming liminars and Others. The liminars represent by their very existence a threat to the entrepreneur’s order of things. They must therefore be “ordered” -- forced to become either a Self or an Other. The reason why this can be done legitimately is through a depoliticized innocent representation of the world, a representation based on myths. 


Thanks to all the participants in the workshop on “the Copenhagen School of Security Studies” at the Norwegian Institute of International Affairs, 2 October 1998, especially Lene Hansen and Iver B. Neumann. Special thanks also to Kari M. Osland and Jeffrey T. Checkel for comments. 


Albert, Mathias. 1998. “Security as Boundary Function: Changing Identities and ‘Securitization’ in World Politics.” International Journal of Peace Studies, Vol. 3, No.1, pp. 23-46.

Anderson, Benedict. 1991. Imagined Communities. London: Verso.

Bartelson, Jens. 1995. A Genealogy of Sovereignty. Cambridge: Cambridge University Press.

Barth, Fredrik. 1969. Ethnic Groups and Boundaries: The Social Organization of Culture Difference.  Bergen: Scandinavian University Press.

Barthes, Roland. 1993 [1957]. Mythologies. London: Vintage.

Buzan, Barry. 1997. “Rethinking Security After the Cold War.” Cooperation and Conflict, Vol. 32, No. 1, pp. 5-28.

Buzan, Barry, Ole Wæver and Jaap de Wilde. 1998. Security. A New Framework for Analysis. London: Lynne Rienner.

Buzan, Barry and Wæver, Ole. 1997. “Slippery? Contradictory? Sociologically Untenable? The Copenhagen School Replies.” Review of International Studies, Vol. 23, pp. 241-250.

Calhoun Craig. 1994. “National Traditions: Created or Primordial?” In Øystein S?rensen ed., Nasjonal identitet - et kunstprodukt? KULTs skriftserie nr. 30/NASJONAL IDENTITET nr. 5. Oslo: Norwegian Research Fund.

Campbell, David. 1993. Writing Security: United States Foreign Policy and the Politics of Identity. Minneapolis: University of Minnesota Press.

Checkel, Jeffrey T. 1998. “The Constructivist Turn in International Relations Theory (A Review Essay).” World Politics, Vol. 50, No. 2, pp 324-348.

Checkel, Jeffrey T. 1997. “International Norms and Domestic Politics: Bridging the Rationalist-Constructivist Divide.” European Journal of International Relations, Vol. 3, No. 4, pp. 473-495.

Connolly, William E. 1991. Identity\Difference. Ithaca NY: Cornell University Press.

Doty, Roxanne Lynn. 1997. “A Critical Exploration of the Agent-Structure Problematique in International Relations Theory.” European Journal of International Relations, Vol. 3, No. 3, pp. 365-392.

Dreyfus, Hubert L. and Rabinow, Paul. 1982. Michel Foucault. Beyond Structuralism and Hermeneutics. London: Harvester.

Eide, Espen Barth. 1998. Between Rationalism and Reflectivism – Constructivist Security Theory and the Collapse of Yugoslavia. Hovedfag-thesis, Institute of Political Science, Oslo: University of Oslo.

Eide, Espen Barth. 1997. “‘Conflict Entrepreneurship’: On the ‘Art’ of Waging Civil War”, in Anthony McDermott, ed., Humanitarian Force. PRIO Report 4/97, Oslo: International Peace Research Institute.

Finnemore, Martha. 1996. National Interests in International Society. Ithaca, NY: Cornell University Press.

Foucault, Michel. 1977. Discipline and Punish. The Birth of thePrison. London: Penguin Books.

Foucault, Michel. 1973. Madness and Civilisation: A History of Insanity in the Age of Reason. New York: Vintage/Random House.

Foucault, Michel. 1972. The Archelogy of Knowledge. New York: Pentheon Books.

Hansen,  Lene. 1998. Western Villains Or Balkan Barbarism? Representations and Responsibility in the Debate over Bosnia. Ph.D. Dissertation, University of Copenhagen, Institute of Political Science.

Hegel, Georg Wilhelm Fridrich. 1952. The Philosophy of Right: The Philosophy of History. Chicago: William Benton.

Hindess, Barry. 1996. Discourses of Power: From Hobbes to Foucault. Oxford: Blackwell.

Hobsbawm, E. J. and Terence Ranger. 1983. The Invention of Tradition. Cambridge: Cambridge University Press.

Hopf, Ted. 1998. “The Promise of Constructivism in International Relations Theory.” International Security, Vol. 23, No. 1, pp. 171-200.

Huntingtion, Samuel P. 1993. “The Clash of Civilisations?” Foreign Affairs, Vol. 72, No. 3, pp: 22-49.

Huysmans, Jef. 1998. “Security! What Do You Mean? From Concept to Thick Signifier.”  European Journal of International Relations, Vol. 4, No. 2, pp. 226-255.

Katzenstein, Peter, ed. 1996. The Culture of National Security: Norms and Identity in World Politics.  New York: Columbia University Press.

Klotz, Audie. 1995. Norms in International Relations: The Struggle Against Apartheid.  Ithaca, NY: Cornell University Press.

Majostorovic, Steven. 1997. “Ancient Hatred or Elite Manipulation – Memory and Politics in the Former Yugoslavia.” Foreign Affairs, Vol. 159, No. 4, pp. 170-182.

McSweeney, Bill. 1996. “Identity and Security: Buzan and the Copenhagen School.” Review of International Studies, Vol. 22, pp. 81-93.

Neumann, Iver B. 1996. “Self and Other in International Relations.” European Journal of International Relations, Vol. 2, No. 1, pp. 139-174.

Norton, Anne. 1988. Reflections on Political Identity. Baltimore: Johns Hopkins University Press.

Price, Richard. 1995. “A Genealogy of the Chemical Weapons Taboo.” International Organization, Vol. 49, No.1, pp. 73-103.

Ringmar, Erik. 1996. Identity, Interest and Action. A Cultural Explanation of Sweden’s Intervention in the Thirty Years’ War. Cambridge: Cambridge University Press.

Smith, Anthony D. 1991. National identity. London: Penguin.

Smith, Anthony D. 1983. Theories of Nationalism. New York: Holmes and Meier.

Szakolczai, Arpád. 1993. Nietzsche’s Genealogical Method: Presentation and Application. Florence: European University Institute, Working Paper SPS 93/7.

Ullock, Christopher J. 1996. “Imagining Community: A Metaphysics of Being or Becoming?” Millennium: Journal of International Studies, Vol. 25, No. 2, pp. 425- 441.

Visker, Rudi. 1995. Michel Foucault. Genealogy as Crituque. London: Verso.

Walker R.B.J. 1993. Inside/Outside: International Relations as Political Theory. Cambridge: Cambridge University Press.

Wæver, Ole. 1993. Securitization and Desecuritization. Copenhagen: Copenhagen Peace Research Institute, Arbejdspapir 5.

Wæver, Ole, Barry Buzan, Morten Kelstrup and Pierre Lemaitre, eds. 1993. Identity, Migration, and the New Security Agenda in Europe.  London: Pinter. 


[1] This article is based on a paper presented on a workshop on “the Copenhagen School of Security Studies” at the Norwegian Institute of International Affairs, 2 October 1998.

[2] Parts of Eides work is published. See Eide (1997).

[3] To be fair the Copenhagen School is not alone. In the last couple of years a growing bulk of empirical work based on constructivist ontology suffers from these epistemological limitations. See Checkel’s (1998) review of Finnmore (1996) Katzenstein (1996) and Klotz (1995). This problem is due to an explicit urge to still speak in terms of dependent and independent variables, causation and explanations. This goes especially for the US branch of International Relations constructivists who obviously wish to “bridge the divide” (Checkel 1997) between the “neo-neo rationalists” and the poststructuralists and thus still use some positivist concepts and methods. The price to pay for this is inconsistency between their ontological stance and epistemological approach.

[4] See for instance Connolly (1991), Norton (1988) and Neumann (1996). Theories of nationalism, on the other hand, tend to be so focused on the creation of a core-identity that they overlook the relationship to other collectives. This also goes for Anderson (1991).

[5] This means that geographic identity boundaries depend on social boundaries. State boundaries can sometimes be almost the same, but hardly ever fully. There is no necessity between a political unit and an identity unit. Of course, when there already is a distinct political unit, like a sovereign state as in Europe, it is of little surprise that the identity which emerged followed the state boarders. On the other hand, sometimes pan-movements proved stronger, like in Germany, and sometimes not, like the pan-Scandinavian movement around 1850. Political-juridical borders are by no means constitutive of identity boundaries.

[6] Even though Barthes (1993:109) analyzes myths through semiology, it is not just linguistic expressions that can function as myths. Anything can be a myth, both verbal and visual expressions, like for instance photography or a car.

[7] Presentism means writing “history of the past in terms of the present” (Bartelson 1995:55). This means to consider contemporary analytical tools and interpretations as objective or unproblematic, and not taking into consideration that they are historically situated. For discussions of genealogy in general, see for instance Bartelson (1995), Dreyfus and Rabinow (1982), Szakolczai (1993), Price (1995) and Visker (1995).

[8] They never really discuss or develop the question of agency.

[9] See also for instance Connolly (1991), Campbell (1993), Neumann (1996), and Hansen (1998).

[10] The Copenhagen School’s stepwise illustration of a securitization process may be confusing. They say that an identity can go from a non-politicized to politicized to securitized mode (Wæver 1993). What should be underlined is that this politization is always directed at the boundaries. Within the group the referent object is necessarily de­politicized in the sense that it encapsulates the entire group. The point with securitization is that the politization/securitization is about placing the political challenge or the threat outside the group.