Security and the costs and benefi ts of manipulating analytical boundaries : Constructivist debates within European Critical Security Studies

In Germany, debates within European critical security studies are taken up belatedly. Yet the debates on identity and security, the interrelation of these two concepts, and the dilemma of writing security, raise fundamental issues with the theory of IR in general, as well as with security scholars’ self-perception. Shaped by the return of ideas, culture, and identity to IR and by the linguistic turn in the social sciences, the new European security theory challenges the tendencies to simply incorporate the new concerns as additional variables into positivist frameworks in order to explain changes in world politics. In developing a conceptual alternative, critical security studies point to innovative avenues of research.


The return of ideas, culture and identity:
A theoretical challenge T he 1990s were marked by debates about a return of ideas, identity and culture to International Relations and to Security Studies, this »last bastion of neorealist orthodoxy« (Krause 1998: 298).The debates were accompanied by disciplining practices of scholars representing dominant rationalist approaches.A number of these scholars attempted to incorporate concerns formerly absent from neorealist scholarship, such as identity or nationalism, into their studies in order to reduce striking defi ciencies in the explanation of many developments, especially in the post-Cold War era.As a result, in these studies identity groups were treated according to the realist ontology as given unitary rational actors and rising nationalism was explained merely as a tactical choice.Concerns like identity and nationalism were simply plugged into dominant frameworks.Despite this practice and the theoretical narrowness of neorealism, neorealist scholars claimed that constructivism was superfl uous and divorced from the real world.

Linkages between Classical Realism and Constructivism
Judging from the severity of the neorealist critique of constructivism, it is apparent that many neorealists are not suffi ciently aware of earlier work within classical realism.For this work not only offers the possibility for dialogue across analytical traditions, such as realism and constructivism, but also has in fact the potential to »speak directly« (Michael Williams) to controversies within constructivist theory.Despite the narrowness of Morgenthau's realist concept of power and the ahistorical quality that his »interest defi ned as power« has assumed in realism, he was actually aware that »interest determining political action in a particular period of history depends upon the political and cultural context« (Morgenthau 1967: 5).His understanding of power and interest as fl exible and indeterminate concepts and his emphasis on context dependency are based on similar premises as the constructivist defi nition of actors as products of complex historical processes encompassing social, political, material and ideational dimensions.Morgenthau's limited defi nition of the sphere of politics is due to his conviction that the introduction of certain issues into political research entails ethical questions, and that analytical neutrality can contribute to political irresponsibility, if consequences of social scientifi c research are not addressed.These concerns reappear in the debates within critical security studies about the political costs of a widened security agenda and the role of security analysts.
The fi rst part of the article presents a constructivist critique of the so-called »ideas« literature of the 1990's, focusing on the neo-positivist conception of ideas, as well as the disregard of the linguistic turn.In this part, key demands of a constructivist approach are introduced that form the basis of debates within European critical security studies.The second part examines some of these debates more closely.

The interconnection of ideas and interests
The so-called »ideas« literature, i.e. the renewed analysis of the role of ideas in foreign policy in the 1990s -including Katzenstein's volume »The Culture of National Security« and Sikkink's study »The Power of Principled Ideas« -has been criticised for being rather a completion of the positivist approach than a full-fl edged alternative (Cf.Laffey / Weldes 1997, Huysmans 2002: 43-44).Some of the criticised scholars bring out very clearly the limits of certain positivist theoretical approaches.Emphasizing the interconnection of interests and ideas, they claim: »To conceive of ideas as intellectual justifi cations of actions that people wanted to take anyway is to obscure the role of ideas in helping people grasp, formu-S+F (25.Jg.) 4/2007 | 203 late, and communicate social realities« (Sikkink 1991: 5).Yet, in practice, due to an analytical separation, the distinction between ideas and interests is retained with the effect that the social construction of ideas is disavowed because it is taken for granted that interests are given and can be determined in isolation from ideas.Furthermore, because of this distinction, ideas appear exactly as these »intellectual justifi cations of actions that people wanted to take anyway«.
The studies are characterised by a neo-positivist conception of ideas that focuses nearly exclusively on the more concrete level of ideas, the »beliefs about cause-effect relationships« and the »principled beliefs or normative ideas«, and tends to exclude the broader type of ideas, »world views« or ideas which provide »conceptions of possibility« (Laffey / Weldes 1997: 198).Yet, according to a constructivist perspective, something like sovereignty that in the »ideas« literature is seen as part of an external and objective reality can only be comprehended through a defi nition of the »universe of possibilities for action« that enable actors to acknowledge the legitimacy of that action, by that actor, in that context.From a constructivist perspective the meaningful constitution of social reality is thus also central to the more concrete causal beliefs (Cf.Hopf 1998: 178-178).
In the »ideas« literature, however, ideas are understood as additional variables that are used to explain changes in foreign policy that rationalism is unable to explain because rationalism focuses on material power and egoistic interests in the context of power realities, and not on the character of the ideas that people have (Cf.Laffey / Weldes 1997: 197).According to the neo-positivist approach of the »ideas« literature, causation can thus be inferred only if there is an observable change in a policy (unexplainable by rationalism), which can plausibly be traced to a co-variation between the policy change and the »ideas« of the policy-makers.

The signifi cance of language
Laffey and Weldes point out that the metaphors used by a number of scholars to explain the role of ideas have the effect that »ideas« are conceptualised as objects.This is odd because there is a well-developed theoretical framework dealing with the »articulation« of discursive elements (e.g.Ernesto Laclau, Stuart Hall, Lawrence Grossberg) (Ibid: 203).The critical security studies scholar Huysmans (2002: 44) criticises Katzenstein's »The Culture of National Security« for examining the causal work of norms and the importance of identity questions in security policies without a refl ection of the significance of language in social relations.The so-called linguistic turn for social theory is ignored.If, instead, ideas are seen as part of a broader set of linguistic and symbolic practices it becomes possible to »rethink ›ideas‹ as intersubjectively constituted forms of social action« (Laffey / Weldes 1997: 209).The notion of ideas as symbolic technologies reveals their constitutive nature.

The issue of power: material and discursive
The power of social practices is an important element of the constructivist approach.According to constructivists, practices »reproduce the intersubjective meanings that constitute social structures and actors alike« (Hopf 1998: 178).Subjects are constituted and reconstituted through political practices creating shared social understandings.During this process the subjects develop identities and interests.Absent interests can be seen as produced absences because »social practices that constitute an identity cannot imply interests that are not consistent with the practices and structure that constitute that identity« (Ibid: 176).According to Hopf, the ultimate power of practices is to »reproduce and police an intersubjective reality« (Ibid: 179).
In addition to the power to control intersubjective understanding, however, Hopf stresses the importance of »having resources that allow oneself to deploy discursive power -the economic and military wherewithal to sustain institutions necessary for the formalized reproduction of social practices« (Ibid).This combination of material and discursive power is signifi cant in order to refute the verdict of the mainstream scholarship that »issues of war and peace are too important for […] [a] discourse that is divorced from the real world« (Walt 1991: 223) and to counter the claim that the whole fi eld of constructivism lacks a theory of power.Constructivism is not only concerned with presumably »soft« issues such as ideas, symbols or discourse.Furthermore, the claim that power is both material and discursive helps to fully understand the role of bureaucracies and security professionals, an aspect that will be discussed in the second part.

Differences in European and American approaches
As this section will focus on the critical approaches to security studies in Europe there will only be a few remarks on the different developments of European and American constructivist approaches.Unlike most constructivist approaches developed in the United States, the critical approaches to security studies in Europe focus on the above-mentioned interaction of material and ideational / discursive power and do not see them as opposites (Cf.Büger / Stritzel 2005: 439).According to Jacobsen (2003: 40ff.),Continental debates about international relations theory are not well known to international relations scholars in the United States.Debates in British international studies journals usually appear in the adjacent fi elds of comparative politics, historical sociology and public policy.The Confl ict and Peace Research Institute in Copenhagen has worked intensely on the theme of »Non-military aspects of European Security« and its fi ve major books published in the period between 1987 and 1998 constitute possibly the most thorough and continuous exploration of the signifi cance and the implications of a widening security agenda for security studies in Europe.The school was shaped by specifi c European security experiences and questions.Studying security concepts in the divided Europe it understood that the security dynamic was not just driven by the two superpowers but had also a more internal European character; its approach partly refl ects how the European Peace Movement and the German Ostpolitik approached the East-West divide.Furthermore, the concept of societal security concerned with identity as the object to be secured relates specifi cally to the intensifying politicisation of migration from a security perspective and to negative reactions to the integration process in some European countries after the Maastricht Treaty.Huysmans reports about the diffi culties that Ole Waever experienced with this concept in the US as it was »not always simple to argue the relevance of the ethnic-cultural identity theme which is central to the concept and which builds upon a European historical-cultural understanding of the nation« (Huysmans 1998b: 484).On the other hand, because the security studies agenda of the Copenhagen school is interested in the security dynamic within the European region as a whole, it stresses a collective security problematic instead of a national security one like US security studies.

McSweeney and the Copenhagen controversy -Reifying society and identity?
In the so-called Copenhagen controversy, the book »Identity, Migration and the New Security Agenda in Europe« (Buzan et al.1993) was the centre of attention due to the societal security concept.In this book it is argued that a broader understanding of security is needed and the societal dimension is given a new status so that there is now »a duality of state security and societal security, the former having sovereignty as its ultimate criterion, and the latter being held together by concerns about identity« (Buzan et al. 1993: 25).In order to avoid methodological individualism and an »individualist, aggregate view of security«, Buzan and Waever (1997: 245) focus on the level of state and society and reject an additional focus on individuals and social groups.According to Buzan and Waever, societal security concerns »the ability of a society to persist in its essential character under changing conditions« (Buzan et al 1993: 23).Bill McSweeney has criticised Buzan and Waever for adopting a reifi ed vision of society and identity.Furthermore, he has reproached them for rejecting analysis of security and identity at the sub-state and sub-societal level and for separating the collective and the individual, equating the structural with the former and the atomistic with the latter (McSweeney 1998: 140).
According to McSweeney, the debate is not between methodological individualism and holism but about a certain version of holism to which Buzan and Waever subscribe and about its implications for determining the proper unit level of analysis.
Buzan and Waever refer to Durkheim's conception of society whereby we must treat society as a »reality of its own«, »not to be reduced to the individual level« (Buzan et al 1993: 18).A collective concept is always »more than the sum of its parts« (Buzan / Waever 1997: 243).McSweeney presents his critique against the background of a constructivist key concept -the mutual constitution of actors and structures: The characteristics of a collective concept are not more than its individual parts in the sense that they exist separately from them, external to them.A collectivity is not a social fact in the sense of a thing existing independently of the individuals who comprise it.A collective concept focuses on the structural properties of action that are inherent in every instance of individual interaction.The anarchy of the international order makes sense only in so far as we understand its place in structuring the actions of individuals, and, through them, states (1998: 139).
Here McSweeney refers to a complex problem.Buzan and Waever claim: »In our securitization perspective, identity is not a ›value‹ (i.e. the individual's), it is an intersubjectively constituted social factor« (1997: 245).Because of this focus on collectively held, intersubjective understandings they draw the conclusion that sub-state groups (let alone individuals) must not be the referent objects of societal security and thereby risk reifying a holistic vision of »society« as the only non-individualistic counter-referent to the state (Cf.Williams 1998: 436).Because of this they also become vulnerable to the criticism levelled against Hopf and the conventional constructivists: They keep scrutiny fi xed at state unit level at the expense of the many identities -bearing diverse interests and projects -competing within each state for power (Cf.Jacobsen 2003: 53).
Furthermore, McSweeney questioned Buzan and Waever's approach for singling out identity among the many objects susceptible to threat.Though they briefl y acknowledge that economic threats can also affect the security of a society as a whole, they simply assert: A society's survival is a matter reduce our conception of society to »its most ephemeral and empirically contentious component«, identity, while ignoring other elements.This leads us to the debate about the construction of identity and to the charge that Buzan and Waever's conception of society loses all touch with fl uidity and process, resulting in a near-positivist conception of identity (Cf. McSweeney 1996: 83).According to McSweeney, we »cannot decide the status, or even the relevance, of identity a priori.Where it is relevant, it is not necessarily the cause of a security problem [as Buzan and Waever assume].It is just as likely to be its effect« (Ibid: 85).
McSweeney (1998) stresses the changing, contingent nature of identity, it »leaves no ›sediment‹: it cannot petrify« ( 138), and contrasts it with the state: »Identity, unlike the state, has no empirical referent other than the process of constructing it« (137).In a similar way he contrasts identity with security: »A critical difference [between identity and security] appears […] when we consider that the perception and fear of threats to security can, in principle, be checked by observing and evaluating the facts external to the subject.[…] There is no court of appeal that can perform the same scholarly task for our sense of identity, personal or collective« (McSweeney 1996: 87).On these contrasts Buzan and Waever base their counterattack in which they return the accusation of not being constructivist enough.In their reply to McSweeney they claim: »For McSweeney there are constructed things -identity -and real things -the state, security!«(Buzan / Waever 1997: 243).
McSweeney sees identity as the most ephemeral component, a narrating, a storytelling or an active process on the part of individuals, which can only be grasped as process.Buzan and Waever treat security in a similar way.

Security, the Speech Act
Infl uenced by Waever's securitization approach (Waever 1995), the concept of security used by Buzan and Waever is subjected to a change.Between the publication of »Identity, Migration and the New Agenda for Security in Europe« (1993) and of the more pronouncedly constructivist »Security: A New Framework for Analysis« (1998), security changes from a perception to a speech act.As there is no objective reference that something is in and of itself a security problem, security is no longer a perception referring to something real existing independently of this perception.By regarding security as a speech act, »security is not of interest as a sign that refers to something more real; the utterance itself is the act.By saying it, something is done (as in betting, giving a promise, naming a ship).By uttering ›security‹, a state representative moves a particular development into a specifi c area« (Waever 1995: 55).When successfully performed, uttering »security« transforms an issue from, e.g.being an economic question, into being a security problem.
In response to McSweeney's criticism that they have adopted a reifi ed and immobilized vision of identity, Buzan and Waever point out that it is not the Copenhagen school that immobilizes identity.The school just shows the artifi cial immobiliza-tion of identity in the act of securitization.Yet a number of problems remain that are connected with the approach of the school: Their constructivism is unevenly distributed -they use a concept of security that is radically constructivist while their interpretation of social relations in general is not.They keep a deeply sedimented vision of society, identity, the state and anarchy.This unevenly distributed constructivism makes them vulnerable to the accusation of committing »ontological gerrymandering« by manipulating analytical boundaries.It also leads to an inability to analyse the mutual infl uence of the processes of securitization and identifi cation (Cf.Huysmans 1998: 493-494).In many regions where there is ethnic confl ict the process of identity formation -including the one at the sub-societal level -needs to be deconstructed in order to analyse in what way processes of securitization that take place are linked to constructions of identity.

Observers or advocates? The debate about the role of security analysts
Johan Eriksson (1999) has delivered the initial intervention in a symposium about the political role of security analysts and the dilemma of writing security.Eriksson briefl y characterises the perspectives of traditional security studies, the Copenhagen school and »critical security studies« (the Wales school) on the theme.While traditionalists are convinced that a distinction can be made between political and scientifi c scholarship and that they belong to the latter category, Buzan and Waever acknowledge the political role of security analysts in a totally different way.Traditionalists assign political scholarship to others arguing that a widening of the security agenda renders the concept analytically useless and is often done in order to legitimise political advocacy, especially on the part of peace researchers.Buzan and Waever, on the other hand, have gradually adopted a wider, more inclusive and more radically constructivist conception of security while retaining a relatively objectivist conception of the objects of security and social relations in general.They point to the »inherently political nature of any designation of security issues« (Waever 1999: 334).As they draw on constructivist language theory and see securitization as being about the power politics of a concept, they argue that the task for the analyst is to determine how, by whom, under what circumstances, and with what consequences certain issues are classifi ed as existential threats.

The Dilemma of Writing Security
The view that securitization is about the power politics of a concept is based on a performative understanding of language and leads directly to the problem of unwanted political consequences that a broadening of the security agenda might entail.Consequently, Buzan and Waever are criticised for treating security in terms of identity.McSweeney (1996: 91) argues that such a concept gives academic support for a renationalization of EU policy, as well as for anti-immigration policies.Buzan and Waever themselves concede that their societal security approach entails risks of legitimising non-state security policy and various self-declared »voices of society«, including fascists.
In their book »The European Security Order Recast« of 1990 they argue, »the security of human collectivities is affected by factors in fi ve sectors« (Buzan et al. 1990: 4): the military, the economic, the political, the societal and the environmental.In »Security: A new Framework« of 1998 these sectors appear again.Critics argue that this multisectoralism contributes to a proliferation of securitization because it objectifi es security and spreads the negative connotations of threats and enemies to new issue areas (Cf.Eriksson 1999: 316).
Buzan and Waever react to this in a similar way as they did earlier when they were criticised for their societal security approach -they claim that their approach is simply embedded in the empirical world: Talk about »wider concepts of security« and »new forms of security […] has been going on for at least two decades.
[…] [Our approach] is a way of turning the analytical question into an empirical one.Just as our approach is either state-centric or not, it is not in its set-up widening or not.We try conceptually and defi nitionally to be open, to create a formal concept, and to let the world be state-centric or not, widening or not.Therefore there is no contradiction between a sectoral approach and a wish to avoid securitization« (Waever 1999: 335).
Here Buzan and Waever differ markedly from the critical security scholars from the Wales school and a more normative perspective according to which the choice of the units of security (the answer to the question »whose security?«) is never just an empirical question but has always ethico-political implications.
Buzan and Waever acknowledge elsewhere that the academic debate about how to constitute security studies cannot responsibly proceed in isolation from the real world yet they trust that the »constructivism [of their multisectoral approach] delivers the means for questioning and politicising each specifi c instance« (Buzan et al. 1998: 212).They argue that the whole theory of securitization sharpens the eye for an already implicit logic within security discourse and can be helpful »in an almost diskursethisch [sic] sense to ask practitioners to be more explicit in explaining why their alleged ›threats‹ and ›security problems‹ should be lifted out of ›normal politics‹ into the realm of ›security‹« (Waever 1999: 37).
Williams agrees with this line of defence and stresses that, as a speech act, securitization »is located with the realm of political argument and discursive legitimation, and security practices are thus susceptible to criticism and transformation« (2003: 512).In this way, the theory of securitization is linked directly to explorations of the role of arguing processes, ethics and validity claims in constructivist approaches and critical theory (drawing on insights from theoretical debates within the German-speaking international relations community and the Frankfurt school and Jürgen Habermas).The »critical security studies« scholar Williams sees the Copenhagen school as largely immune from criticism in this respect and stresses that the bases of securitization theory are located within the context of classical realism and constructivist ethics.
Yet »critical security studies« scholars disagree on the political costs and benefi ts of a widened security agenda.Some scholars draw on the peace research tradition and count on the capacity of security language to prioritise issues and to mobilize people.They are thus convinced that one may employ security language in order to give human rights questions a higher visibility, for example.Their aim is to replace the realist meaning of security with a positive one that defi nes liberation from oppression as a positive good that should be secured (Cf. Eriksson 1999: 318ff. Huysmans 2002: 59).Referring to this attempt to perform securitization with an emancipatory interest, Huysmans objects that alternative constructions do not exist in a vacuum or sheltered space because they are part of a complex political game.Alternative constructions are thus embedded in relations of power that structure and restructure the social exchanges.Didier Bigo (1996: 55), a leading scholar from the Paris school, argues that opposing tactics do not necessarily radically challenge established politicisations as they often share the same concept of security and diverge only in their solutions.
Scholars from the Paris school of security studies -infl uenced by the work of Foucault and Bourdieu -focus their research on the complexity of the political game.Bigo analyses the process of professionalization within modern West European societies.This process gives security professionals in the bureaucracy a central role in the construction of security fi elds.
Bigo's research reduces a severe defi ciency of Buzan and Waever's securitization approach.Buzan and Waver presuppose that statesmen perform the key role in the securitization process.By introducing the concept of societal security they run into problems because the question »who is in a powerful position to speak security?«can no longer be answered simply by pointing to statesmen.That means that the institutionalisation of security practices remains undertheorized (Cf.Huysmans 2002: 54ff.).Bigo, on the other hand, shifts the focus from statesmen to the bureaucracy that he sees at the centre of this institutionalisation.The institutional position of security professionals lends them transformative capacity via »credible« technical knowledge.As an answer to the question »who is in a powerful position to speak security?« the Paris school thus convincingly points to institutionalised patterns of practices that simultaneously empower and constrain agents in their capacity to speak security (Cf.Bigo 1996Bigo , 2000)).
In view of the complexity of the political game, the Paris school also criticises that Buzan and Waever's concept of securitization as a speech act stresses only language and thus »omits all that is of semiotic interest, such as gestures, manoeuvres, the rituals of demonstration of force which are of course fundamental in the economy of securitisation« (Bigo 2000: 194).Williams (2003: 512) claims likewise that securitization theory needs a broader understanding of the mediums, structures, and institutions of contemporary political communication as this communication increasingly relies on the production and transmission of visual images.To a large extent, these last claims still remain desiderata that point to the future of security studies.

Conclusion
Summarising, I would like to stress the following points regarding questions of theory and the political role of critical security studies: Morgenthau's realism, whose complexity nearly fell victim to neorealist streamlining, and Buzan and Waever's securitization theory, that »manipulates« analytical boundaries, are key contributions to their fi elds.Morgenthau's realism is proof to the fact that disciplining moves can block productive debates on both positions and possibilities within the fi eld of international relations / security studies.The debates about the securitization approach confi rm that not all theoretical tensions or even contradictions within an approach are necessarily weaknesses since they may very well point to innovative avenues of research.
Different critical approaches, such as the speech act concept or attempts to broaden the conception of security in order to prioritise human rights issues, demonstrate the diverse and rich threads of the critical security studies research agenda.These different critical approaches aim for a desecuritization.Their goal is to explicitly uncover dimensions of the security formation that rationalist approaches have left implicit.A theorization of power relations and the symbolic dimensions of the complex political game of the security formation can be critical in itself -a fact well understood by the new Danish government that in 2002 tried to close down the Copenhagen Peace Research Institute.

2 Critical security studies in Europe: Copenhagen, Wales, Paris Within
Europe there are great differences in the various countries.In Germany for instance, critical approaches to security studies have hardly been adopted.Within this restricted adoption Katzenstein's »The Culture of National Security« of 1996 and Adler and Barnett's »Security Communities« of 1998 are still authoritative (Cf.Büger / Stritzel 2005: 438).European debates focus on three centres: The Copenhagen school and its concept of securitization, represented by Ole Waever and Barry Buzan; the Wales school / the so-called »critical security studies«, represented by Ken Booth, Michael Williams, Keith Krause, Bill McSweeney and Richard Wynn Jones; the Paris school and its analysis of the role of security professionals and bureaucracies, represented by Didier Bigo who uses approaches by Michel Foucault and Pierre Bourdieu.
Jacobsen criticises that only those approaches gain admittance to centre stage debates that »can be absorbed into reigning research agendas with minimal disturbance« (Ibid: 40).Thus, according to Jacobsen, only a single form of constructivism, called »conventional constructivism«, represented by Alexander Wendt, Jeff Checkel, Emanuel Adler and others has been thoroughly discussed.204|S+F (25.Jg.) 4/2007 3.