Slippery? contradictory? sociologically untenable? The Copenhagen school replies * B A R RY B U Z A N A N D O L E WÆ V E R In the January 1996 issue of the Review, Bill McSweeney argues that our 1993 boo k, Identity, Migration and the New Security Agenda in Europe (IMNSAE), ‘subverts ’ the analysis of Buzan’s People, States and Fear (PSF) ‘without enhancing our u nderstanding of the problem of security’ (p. 93).1 Of the many charges that McSw eeney brings to bear we will address three. First is that societal security is m erely a trendy response to current concerns about nationalism rather than a more theoretically considered move. Second—and this seems to be the core of his comp laint—is that the view we take of ‘identities’ is far too objectivist and not (d e)constructivist enough, and that our approach makes it impossible to consider t he process of identity formation as part of the politics of security. Third, he says that Buzan’s association with IMNSAE contradicts strong positions he develo ped in PSF and that his analysis has therefore become incoherent. McSweeney’s se cond and third points themselves seem contradictory: PSF is much more objectivis t than IMNSAE, and IMNSAE quite constructivist. Our next book2 is even more cons tructivist, and goes much further than IMNSAE towards opening up many more kinds of referent objects for security. This we believe to be defensible because we h ave developed a way of specifying security as an extreme form of politicization (in whatever sector) and thus of avoiding the proliferation of securitizations t hat has tended to accompany the wider security agenda. Given that development, a nd since McSweeney’s article raises, but does not satisfactorily answer, several issues central to security studies, we felt that it required an answer.3 * We are grateful to Lene Hansen, Eric Herring, Jef Huysmans, Richard Little, He ikki Patomäki and Michael C. Williams for comments on an earlier draft. 1 Bill M cSweeney, ‘Identity and Security: Buzan and the Copenhagen School’, Review of In ternational Studies, 22 (1996), pp. 81–93; O. Wæver, B. Buzan, Morten Kelstrup a nd Pierre Lemaitre with David Carlton et al., Identity, Migration and the New Se curity Agenda in Europe (London, 1993). 2 Barry Buzan, Ole Wæver and Jaap de Wil de, Security: A New Framework for Analysis (Boulder, CO, 1997). 3 Identity, Migr ation and especially its central concept of ‘societal security’ is controversial . We are aware of several good critiques: Didier Bigo, ‘The New Field of Securit y in Europe: Mixing Crime, Border and Identity Controls’, in Anne-Marie Le Gloan nec and Kerry McNamara (eds.), Le Désordre européen [working title] (forthcoming ); Ken Booth, book review in International Affairs, 70 (1994), p. 171; Lene Hans en, ‘The Conceptualization of Security in Post-structuralist IR Theory’ (MA thes is, University of Copenhagen, 1994); Jef Huysmans, ‘Migrants as a Security Probl em: Dangers of ‘‘Securitizing’’ Societal Issues’, in R. Miles and D. Thänhardt ( eds.), Migration and European Integration: The Dynamics of Inclusion and Exclusi on (London, 1995); Janus Mortensen, ‘Sikkerhed som talehandling: En kritisk genn emgang af Wævers sikkerhedsbegreb’ (Security as Speech Act: A Critical Examinati on of Wæver’s Concept of Security) (unpublished paper, Institute of Political Sc ience, Copenhagen); Yosef Lapid and Friedrich Kratochwil, ‘Revisiting the 241 242 Barry Buzan and Ole Wæver 1. The place of societal security in security theory McSweeney criticizes us for raising societal security as a response to ‘the pressure of events’ rather than as a result of theoretical considerations (p. 81). This is a charge that can be laid at the door of nearly all IR theory, from Idealism and Realism to Interdep endence and IPE, and we do not deny following that well-trodden path. But our mo ve was also a response to a theoretical challenge. Most ordinary people if asked about European security will start talking about nationalism, ethnic con ict in East-Central Europe and possibly about migration. And they would be greatly sur prised to learn that such phenomena have no place in classical security theory. This was a theoretical challenge, because these issues were not simply absent in the sense that classical security studies did not care; they were radically abs ent because they could not be represented in the classical state-centric theory. Rather than abandon existing theory and mainstream debate by taking the reducti onist path to individualbased security logic (on which more below), we saw it as a challenge to devise a theoretical conception of identity-related security iss ues that was at the unit level, and therefore interoperable with classical secur ity theory. As argued by Lapid and Kratochwil, some other neorealists assimilate d identity and nationalism into classical theory by simply treating nations as s tates, and identity as one more resource, thus avoiding any revision of the basi c theory.4 We tried instead to revise the basic, traditional conception of secur ity so that it could still say the old things but also include the new things in their own right. We tried to show how ‘societies’ de ned in terms of identity c ould be seen as the referent object for some cases of securitization, where that which could be lost was not sovereignty but identity. The two share the role of being the de nition of existential threat: for a state, sovereignty de nes when a threat is existential, because if a state is no longer sovereign, it is no lo nger a state; and similarly identity is the de ning point regarding existential threats for a society because it de nes whether ‘we’ are still us. In the tradit ion of security studies with its focus on the interaction of units and their con cern for others’ threat to their survival, it was crucial for us to be able to d e ne a new kind of unit in order to grasp the way other things than states had b ecome referent objects for security discourse. 2. Objectivist and (de)constructivist approaches to identity and societal securi ty McSweeney states that ‘The analysis of collective identity can be approached from a deconstructionist, sociological angle, which focuses on the processes and practices by which people and groups construct their self-image. Or it can be a pproached from the more common objectivist viewpoint’ (p. 82) Why does this choi ce have to be a hard either/or? If one studies only the processes by which ident ities are formed, then 4 ‘‘National’’: Toward an Identity Agenda in Neorealism?’, in Yosef Lapid and Frie drich Kratochwil (eds.), The Return of Culture and Identity in IR Theory (Boulde r, CO, 1996), pp. 105–26; Martin Shaw, Global Society and International Relation s: Sociological Concepts and Political Perspectives (Cambridge, 1994). We apprec iate the opportunity to take stock of critiques, re ect self-critically on our w ork, and put forward some further arguments at this later stage of the debate. L apid and Kratochwil, ‘Revisiting the “National”’. Discussion: the Copenhagen school replies 243 identity never becomes a ‘thing’ at all: there is never a product as such. And c onversely, if one studies the politics around the established identities (as we do), why does that have to mean positing identities as God-given, immutable, and intractable by sociological, ‘deconstructionist’ analysis? Why can one not thin k of identities as de nitely being constructed by people and groups through nume rous processes and practices, and that when an identity is thus constructed, and becomes socially sedimented, it becomes a possible referent object for security ? As we see it, one can choose to place the analytical emphasis on either end of the spectrum. Doing so produces different kinds of inquiry, probably for differ ent purposes. But there is no reason to picture either approach as unable to acc ept the existence of the other. The main weight of McSweeney’s accusation is tha t we impose a rigid, ‘objectivist’, ‘near-positivist’ view of identity on societ y (p. 83); but he ignores the explicitly constructivist approach to society (and even more to security) set out in chapter 2 of IMNSAE. To take something as bei ng more than the sum of its parts does not make it ‘immune to process inquiry’ o r make its values and vulnerabilities ‘objective’ in the positivist sense (p. 84 ). Because we talk of individuals actually identifying themselves as members of society, and because we talk about how societies re ect on threats to, and defen ce of, what they take to be their identity, McSweeney concludes that we project ‘ ‘‘society’’ and ‘‘identity’’ . . . as objective realities, out there to be dis covered and analyzed’ (p. 83).5 There are no statements to this effect in the bo ok, and a number directly to the contrary. McSweeney must therefore assume that, since we treat identity in some speci c situations as an object of security con cern (that which is to be defended), we think that identity is always a thing, a nd an immutable one at that. This is not a logical conclusion, nor is it a corre ct description of our position. To take identity as a possible object of securit ization, one has only to assume that it holds a social power that makes it ef ci ent to invoke it, and that it has a form which makes security discourse possible (i.e., it has a claim to survival as well as a clear image of what non-survival would mean). Usually this demands that the referent has become relatively stabi lized in social practice. This is our view. The state is not a constant either, yet there is a lot of security policy to defend it. France has changed over the centuries, but there is a French security policy. There are actors who mobilize security policy in the defence of something which is ‘thingish’ enough to be inv oked in this way. Identities too can be defended. This does not imply that ident ities 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 an analys is. A very big part of social science is about what to take as relatively more xed than what.6 5 6 On objectivist, subjectivist and intersubjectivist approaches to security, see, most thoroughly, Jef Huysmans, ‘Making/Unmaking European Disorder: Meta-theoreti cal, Theoretical and Empirical Questions of Military Stability after the Cold Wa r’ (Ph.D. thesis, Catholic University of Louvain, 1996), pp. 48–57, 84–6. For Mc Sweeney there are constructed things—identity—and real things—the state, securit y! Identity he argues is peculiarly uid and therefore not to be treated as an o bject. This he argues by various contrasts to how other things are tangible, mea surable and to be ‘challenged by evidence’. We prefer to take a social construct ivist position ‘all the way down’. However, identities as other social construct ions can petrify and become relatively constant elements to be reckoned with. Es pecially, we believe security studies could gain by a constructivism that focuse s on how the very security quality is always socially constructed: issues are no t security issues by themselves, but de ned as such as a result of political pro cesses. 244 Barry Buzan and Ole Wæver McSweeney tells us: ‘Identity is not a fact of society; it is a process of negot iation among people and interest groups. Being English, Irish, Danish is a conse quence of a political process’. We agree. He continues: ‘and it is that process, not the label symbolizing it, which constitutes the reality that needs explicat ion’ (p. 85). Maybe, but we doubt that this would be a very effective approach f or security studies. One could study process, just as one can study the historic al origins of a state to explain why it is there as object of security policy, r ather than studying its current security policy. But to understand an identity a s a possible referent object for security policy, one actually has to understand the label symbolizing it. McSweeney asks for a deconstructionist approach to id entity. That means precisely that one has to understand powerful symbols, labels and the discursive structure of political moves that surround them. Security di scourse always uses a symbol or a concept—as all other discourse, it is unable t o grasp the thing or people as such. A label surely can be securitized. McSweene y sees social identity as permanently mutable and unstable: ‘never more than a p rovisional and uid image of ourselves as we want to be’ (p. 90). We agree that identity is socially constructed, but see it as often solidly sedimented. Furthe rmore, the knowledge that an identity is never fully stabilized, that it is alwa ys problematic, should not lead us to just denounce the possibility of doing sec urity in its name. Quite the contrary, this lack is often the key to understandi ng its vulnerabilities, restlessness and security efforts. If we want to underst and the peculiarities of the branch of security policy that is conducted on beha lf of identity, it is indeed helpful to investigate the inherent paradoxes of ac ting in defence of an identity which is never simply constant in itself, but alw ays contains a longing for a desired self. Collective identities of this sort ca n never be more than a series of partially or temporarily successful, but ultima tely impossible, closures.7 Our rejection both of McSweeney’s characterization o f our position, and of his either/or choice about analytical method has two root s, one normative and the other ontological. Both issues are important to how sec urity studies is pursued, and are worth investigating a little more closely. Ontological issues Here the problem seems to be our preference for methodologica l collectivism versus McSweeney’s for methodological individualism. The issue is much bigger than methodology, the underlying question being whether to reduce s ecurity to the level of individuals. This is an ontological issue: is society th e sum of the individuals or does it have group qualities that go beyond the sum of its parts? Even Durkheim, who held that society had sui generis features that were to be located as attributes of society as a whole, actually studied societ al processes and vigorously defended individualism. Just because we are methodol ogical collectivists, does not bar our 7 Ernesto Laclau and Chantal Mouffe, Hegemony and Socialist Strategy: Toward a Rad ical Democratic Politics (London, 1985); Ernesto Laclau, New Re ections on the R evolution of Our Time (London, 1990); Slavoj Zizek, For They Know Not What They Do: Enjoyment as a Political Factor (London, 1991); Ole Wæver, ‘Insecurity and I dentity Unlimited’, in Le Gloannec and McNamara (eds.), Le Désordre européen. Discussion: the Copenhagen school replies 245 engagement with questions about how collective identities are shaped. The argume nt of McSweeney, as of many critical theorists and peace researchers, seems to b e that if one wants to open up to a world beyond the state, one has to take a bo ttom-up, individualist (and/or small-group) perspective. In this individualist p erspective, identity is indeed one ‘among the countless values which people are concerned about and which can be attributed to the collectivity of society’ (p. 84), in which case it seems problematic to single it out above the numerous othe r values. McSweeney prefers an analysis of all the individual values that can be threatened, and is here taking over formulations verbatim from peace researcher s such as Johan Galtung and Jan Øberg. He seems to want to de ne a priori that a ll security is reducible to individual security. As we argued in IMNSAE (pp. 20– 7), to move down to the individual level has severe consequences. It is possible to take the individualist, aggregate view of security, but as far as we can see , unless one is extremely careful, this becomes another mono-unit ontology, wher e all security is ultimately individual security and the security of the state h as to be measured and discussed on the basis of how it in uences the aggregate s ecurity of ‘its’ individuals. We resist this turn because the state cannot be re assembled from individual-level attributes; it has sui generis statelevel attrib utes and one has to see the state itself as a unit reality. Individual security can be studied from our perspective, because we are interested in all action tha t ful ls the criteria of being a security speech act. Doing this in the name of individuals is, however, much more dif cult than action in the name of limited c ollectivities or on behalf of general principles. In our securitization perspect ive, identity is not a ‘value’ (i.e. the individual’s), it is an intersubjective ly constituted social factor. To us it seems that the two approaches are complem entary: each can do things that the other cannot. The individualist approach is not able to grasp a lot of the securitization that takes place, which mostly has various limited collectivities—states, nations or, as we show in our next book, speci c principles at the international level—as referent objects. Neither is i t able to manage larger interactive formations—for instance, regional security c omplexes—as our more Realist and reactionary approach can. Conversely, we cannot answer critical and emancipatory questions about the ‘real’ security of margina lized groups who do not articulate security demands in any powerful way. This cr itical thrust in McSweeney’s enterprise underpins the normative problems that se parate our positions, and allows us to put the charge of objectivism back at his feet. Normative issues McSweeney more than hints that the purpose of studying process is to discredit as political manipulations at least some claimed identities. As an antidote to nationalist attempts to present identities as necessary or innoce nt, it is politically important to expose the tainted roots of all identities. B ut this neither sorts problematic or arti cial identities from ‘authentic’ ones, nor identities from more ‘real’ political referents. Is there any state whose c urrent existence does not depend on centuries of violence, selective memory and politically motivated identity politics? As Derrida 246 Barry Buzan and Ole Wæver argues: ‘You cannot object to a unity simply because it is the result of a proce ss of uni cation . . . [T]here are no natural unities, only more or less stable processes of uni cation, some of them solidly established over a long period of time’.8 Against our allegedly static, objectivist concept of identity, McSweeney rather disturbingly proposes that identity be ‘corrected’, and calmly goes on f rom there to call for someone to take on ‘the task of speaking ‘‘objectively’’ f or the society’ (p. 88). According to McSweeney (p. 87), ‘perception and fear of threats to security can, in principle, be checked by observing and evaluating t he facts external to the subject’. McSweeney feels able to assess security perce ptions for their objectivity. We do not, so we designed a security theory that i s much more radically constructivist. As against McSweeney’s traditional, critic al approach, with its ‘objective’ requirement for understanding of security, our approach has the advantage of insisting that any securitization always rests on a political choice. Security can never be based on the objective reference that something is in and of itself a security problem. That quality is always given to it in human communication. And when securitization is seen as a political cho ice, there is less chance that security gets idealized as the sought for conditi on, and more chance that the path to desecuritization—taking things back into no rmal politics—stands out more clearly.9 This is the starting-point that McSweene y missed in IMNSAE. He therefore imposed a false reading on all the rest, leadin g to the paradoxical accusations that we are too objectivist. But why would McSw eeney counter our approach with one ‘correcting’ identities, why expect security analysts to be able to arbitrate between competing identity claims (p. 88)? McS weeney’s scepticism towards societal security seems to stem from a concern that identity is often not the root cause of con icts but rather an instrument used b y (nationalist) elites (p. 86). Others have also spotted this problem with an id entity approach to security. As argued above, we are sceptical about attempts to judge which identities are authentic and which not, because all are constructed and all are shaped by politics. Once mobilized, identities have to be reckoned with as something people perceive that they belong to, and act upon as objective , given. The Israeli–Palestinian con ict is not solved by exposing the contingen t nature of both identity groups. It might be a part of con ict resolution to st imulate collective rede nition in each group to change the constellation away fr om complete incompatibility, but no solution is viable that denies either group a right to survive. There is a consistent, though in our view often unhelpful, a lternative to our approach in the stand taken by most post-structuralists (and s ome radical constructivists): to question all identities, celebrate contingency, and generally aim for weaker, more self-consciously fragmented identities: a ‘p olitics of disturbance’.10 McSweeney’s suggestion seems to be less radical than this but also less theoretically consistent. He wants ‘criteria for legitimizing decisions about identity’ (p. 90); he wants to be able to correct identity clai ms. To correct can either mean just to change as a result of debate, or it can m ean to righten, to approach to the true. We address these separately as correcti ons 1 and 2. 8 9 10 Jacques Derrida, ‘The Deconstruction of Actuality’ (interview), Radical Philosop hy, 68 (1994), p. 41. Ole Wæver, ‘Securitization and Desecuritization’, in Ronni e D. Lipschutz (ed.), On Security (New York, 1995), pp. 46–86. William E. Connol ly, Identity/Difference: Democratic Negotiations of Political Paradox (Ithaca, N Y, 1991), and The Ethos of Pluralization (Minneapolis, 1995). Discussion: the Copenhagen school replies 247 Correction 1 means to engage in debates over how ‘we’ de ne us, and there should be no doubt that this will often be a major issue in any speci c security con i ct involving identity. We cannot claim to be able to tell what is the ‘correct’ identity, but we hope to be able to predict some consequences of one or the othe r self-de nition due to the way different identities will interact in security m ode. This seems an appropriate task for the security expert. With Durkheim we co uld say: ‘Yet because what we propose to study is above all reality, it does not follow that we should give up the idea of improving it.’ 11 More speci cally, w e suggest in IMNSAE a focus on how constellations form with identities at differ ent levels and how to assist developments where these become mutually compatible (pp. 193ff. and the case-study of Europe, ch. 4). We think that one can and sho uld engage in critical debates over how communities construct their identities. But we think it is too optimistic to think this will solve all the problems. The re will still be the issues we were concerned with arising from security that ac tually gets articulated in the name of threatened identities. Correction 2 is so me form of reasoned intervention telling what is the right identity. We are scep tical of correction 2 because we are unable to follow McSweeney into ‘objective’ security. To be able to tell people that they are not what they think they are demands an objectivist conception of identity. McSweeney rightly criticizes us f or being relativist: ‘we are stuck with every other community’s account of its i dentity’ (p. 87). In the good classical Realist tradition, a major task for secu rity analysis is to help actors understand how others construct their conception of security. This should not be replaced by a demand that others think ‘correct ly’ in accordance with some scienti c theory of security (which usually means ho w we would like them to conceptualize themselves in ideological terms that suit us). He is worried that our position leads to something like: ‘We may not like w ho they are, but if they think that way, so be it’ (p. 87). Classical Realists a s well as post-structuralists will prefer this to the universalism and harmony-o f-interest assumptions necessary to avoid such situations. There will be others who are different; if we can’t live with that, we will certainly have security p roblems. McSweeney’s argument at this place is perplexing. He tells us at length that identity cannot be just read from polls, culture or some other form of his tory of the community, it ultimately involves a choice. This is exactly the view we presented. After our review of the literature on nationalism in IMNSAE (ch. 2), we conclude (like most others in the eld) that there can be various objecti ve markers at play— language, history, culture, race, political borders—but that ultimately national identity cannot be de ned in terms of any of these, only as the choice of identi cation made by individuals. All these conditions might str ongly in uence their choice, but none determines it. Because of McSweeney’s eith er/or move quoted above, he has constructed a Wæver et al. that say with the nat ionalists that identity is objective, given and necessary, it is what we are and have to be because of history. He then succeeds in refuting this ctitious posi tion. McSweeney’s task in this section, however, was to get to a way to ‘correct ’ identities—against our ‘relativism’—so just pointing to the element of choice does not help him much. McSweeney ends up with a strong call (p. 89) for a refer ee to settle identity questions authoritatively, but at this crucial point of hi s analysis 11 Emile Durkheim, The Division of Labor in Society (New York, 1984 [1893], preface to 1st edn, p. xxvi. 248 Barry Buzan and Ole Wæver (pp. 88–90) he becomes frustratingly unclear. His worst case seems to be that id entity remains unsettled and vulnerable to political manipulation; his best that it is somehow negotiated amongst the citizenry. He is not clear about what role security analysts are supposed to play in this process, and if they are to part icipate how they should separate the role of analysts from that of participant i n the securitization/ desecuritization process. Although he wavers on the questi on, the state lurks in the background as the authoritative imposer of ‘arbitrati on’ should negotiation fail and anarchy threaten. In this discussion McSweeney rst slips into reifying society as that which is contained by the state, and the n into the assumption that this ‘society’ must somehow nd, or have imposed on i t, a collective image of itself. Our startingpoint was that we found this essent ially Hobbesian position deeply unsatisfactory. We wanted to leave room for a co ncept of society detached from the state, and for circumstances in which identit y politics was about maintaining difference rather than nding a collective imag e. In those circumstances the question is not McSweeney’s one of who arbitrates, but whether there should be an arbitrator, and how de nitions of difference can be constructed in ways that exacerbate or mute insecurity. Our unease with the state is supported by a very long list of profoundly problematic state intervent ions on identity issues. To take just one: the Kurds in Turkey. Should the Turki sh state really be the nal judge on Kurdish identity and security? As we explai ned in IMNSAE (pp. 24–5), this was a major reason for our revision of PSF concep ts: when societal security was conceived of as one more form of state security—‘ the way states could be undermined or destabilized by ‘‘their’’ societies becomi ng threatened or weakened in terms of social cohesion and identity’—it had the p erverse effect that a state would feel most secure if some minority could just b e put down. If one wants to take this minority seriously and say societal securi ty is about their security, one has to open up to a more complex landscape of mu ltiple referent points for security. In sum, McSweeney seeks to cast our positio n much more narrowly than is in fact the case. Ironically, his attempt to formul ate a critical position pushes him towards an objectivism that he otherwise want s to reject, whose problems he has not resolved, and from which our constructivi st approach offers at least a partial escape. 3. Buzan and the Copenhagen school McSweeney mounts a vigorous attack on Buzan, arguing that the work of the Copenhagen school, particularly IMNSAE, does not re medy the shortcomings of his previous work, but rather guts the general state-ce ntric assumption that underpinned many of its most useful ideas (pp. 82, 91–3). McSweeney puts himself in a dif cult position. He seems to defend PSF ‘as the ca non and indispensable reference point for students of security’ (p. 81), while w anting to attack IMNSAE. Yet PSF should be the more objectionable to him for bot h its greater objectivism and its state-centricity, while IMNSAE actually moves towards his preferences in terms of both its subject focus and its constructivis t method. Fortunately, McSweeney’s argument that by signing on to IMNSAE Buzan h as collaborated ‘in the abandonment of state primacy’ (pp. 82, 92) is so oversta ted that Discussion: the Copenhagen school replies 249 most of the points he tries to hang on it fail by default. We argue that what is or is not prime in international security, including the state, depends on hist orical conditions. The particular case of 1990s European security is dif cult to grasp if seen simply as a constellation of nation-states. Much more of the dyna mics can be brought out by a constellation made up of at least three kinds of (n on-like) units: states, nations and the EU. We do not, as McSweeney would have i t, argue that societal identity has now become the core value in security (p. 82 ), only that it can become a referent object for security action. While McSweene y is right to point out that IMNSAE does raise questions about Buzan’s formulati on of weak and strong states, and security complexes, he is wrong to think that Buzan has therefore ‘to reformulate his entire theoretical framework’ (p. 92). I n one sense, the arguments in IMNSAE are simply an elaboration on the whole prob lematique of weak states. Giving societal security the status of a referent obje ct does not prevent the existence of strong states. Nothing in the idea says tha t collective identity has to be in opposition to the state, or even that societa l issues have to become securitized. But it does enable one to look more deeply into the problems of weak states, where societal insecurity is often a central i ssue. McSweeney worries unnecessarily about the impact of all this on the link b etween strong states and mature anarchy. Although an important logical and ideal ist component in PSF, that link was always highly quali ed: strong states were a necessary but not a suf cient condition for mature anarchy. PSF had little usef ul to say on how to solve the problem of weak states. IMNSAE does not solve the problem either, but it does offer better analytical tools for examining it. In t his context, the idea that the international system is not in any way allowed ‘t o determine shifts in the security position of the state’ (p. 92) has never been part of Buzan’s position and is a contradiction of the central tenets of all fo rms of structural realism. McSweeney is right to hint (pp. 91–2) that IMNSAE cre ates dif culties for security complex theory. Why was the concept of security co mplex not used more in IMNSAE? Why not either construct a ‘societal security com plex’, or integrate the new concept of societal security directly into classical security complex theory as presented in PSF, which was constructed from the pol itical and military sectors and was purely state-based? It is not obvious that s ecurity complex theory with its basic claim about a regional focus for security dynamics also holds for the new sectors of security, the environment, economic a nd societal security. To insert the security of societies into regional formatio ns de ned by the states, as we did in IMNSAE, was not an ideal solution, and it demands serious re ection whether security complex theory can be rearticulated f or a post-sovereign system where actors other than states are also players. This problem of how to reconcile the new sectors of security studies with security c omplex theory is a core theme of our next book. 4. Conclusion If space allowed, there are other points we could take issue with in McSweeney’s piece. But it is more pertinent to raise some general questions that stem from the nature of his review. Most worrying is McSweeney’s implicit argument that there is only one co rrect way to study security. We believe that there are many ways to 250 Barry Buzan and Ole Wæver understand security, and that each will have its merits and its drawbacks. Focus ing on any one element will always make some things clearer at the cost of obscu ring or distorting others. That is the nature of social theory, and there is no escape from it. We also found it odd that he (kindly) designates us as ‘the Cope nhagen school’, and then ignores what this might mean. There are enough institut ional barriers against collective writing, without the academic critique and deb ate also being unable to acknowledge collective works as collective. By focusing on Buzan, McSweeney virtually ignores Wæver, who made the main theoretical cont ribution to IMNSAE.12 This blindness seems to explain how McSweeney missed the s trong constructivist approach to societal security. He also missed the opportuni ty to consider all four of the works he lists, and thus to get some handles on t he nature of the school. How is it that the reactionary objectivist Buzan, and t he postmodern Realist Waever have been able to work together—with each other, as well as with liberals like Pierre Lemaitre, Morten Kelstrup and Jaap de Wilde—a nd what kind of synthesis have they created? There is also a certain implication that having written one classic position piece, Buzan should either shut up or go on repeating it endlessly. Even though we dispute much of McSweeney’s accusat ions of inconsistency, the fact that he makes it suggests that authors are not a llowed to develop or change their positions. PSF was valuable because it helped to start a debate about the concept of security. It was never intended to be the last word on the subject, and it has served as a springboard to help others, in cluding its author, to formulate alternative positions. The eld will develop as those positions (including the methodological individualist, critical one favou red by McSweeney and others) articulate themselves and compete to see how well t hey help us to understand and act upon the security problems of the day. 12 Ole Wæver, ‘Security, the Speech Act’ Working Paper, 1989/19, (Centre for Peace and Con ict Research, Copenhagen); ‘Securitization and Desecuritization’; ‘Europ ean Security Identities’, Journal of Common Market Studies, 34 (1996), pp. 103–3 2; ‘What is Security? The Securityness of Security’, in Birthe Hansen (ed.), Eur opean Security—2000 (Copenhagen, 1995), pp. 222–49; ‘Sikkerhedspolitik—nationals statens monopol?’ (The Concept of Security—A Monopoly of the Nation State?), Gru s, 46 (1995), pp. 43–70; ‘Insecurity and Identity’; ‘Societal Security—A Concept and its Consequences’, in Cooperation and Con ict (forthcoming).