Abstract
A broad array of authors and schools have influenced Barry Buzan and Ole Wæver’s formulation of securitization theory, including John L. Austin, Jacques Derrida and Carl Schmitt. This article draws attention to and strengthens the post-structuralist elements in the writings of Buzan and Wæver, as this part of the theory has received less attention than those attributable to Schmitt and Austin. Starting from securitization theory as developed by Buzan and Wæver and engaging with later expansions of the theory, I suggest a post-structuralist framework built around three questions: Through which discursive structures are cases and phenomena represented and incorporated into a larger discursive field? What is the epistemic terrain through which phenomena are known? And, what are the substantial modalities that define what kind of an issue a security problem is? The last part of the article brings this framework to bear on the ‘Muhammad cartoon crisis’ that began with the publication of 12 cartoons depicting the Prophet Muhammad in the Danish newspaper Jyllands-Posten in 2005.
Introduction
The political-ontological assumptions of the Copenhagen School have been the subject of debate ever since the ‘Copenhagen School’ designation was first used by McSweeney (1996). McSweeney accused the theory of societal security developed by Wæver, Buzan and their associates of adopting a fixed understanding of society and identity that ran contrary to the fact that the latter was constituted in discourse by ‘political leaders, intellectuals and countless others’ (McSweeney, 1996: 90). Buzan and Wæver replied that one might study the implications of discursive fixity without assuming that society was ontologically given. They also fired back that while McSweeney called for a deconstructionist theory, theirs, with its understanding of ‘powerful symbols, labels and the discursive structure of political moves’, was fully so – a claim that was backed up by references to Laclau, Mouffe and Zizek (Buzan and Wæver, 1997: 244). In the 15 years that have passed since this exchange, the speech-act approach to security – which was one of several elements of the theoretical corpus of the Copenhagen School – has branched out and consolidated under the rubric of securitization theory, and a series of critical engagements have followed in the trail broken by McSweeney. Like Balzacq (2005: 173–4), many have taken the view that the weaknesses – and strengths – of securitization theory ‘arise from unspecified yet deeply held philosophical views’. Tracing such views, other theorists have deepened our understanding of securitization theory’s roots and advocated revisions and conceptual expansions of the theory. Huysmans (1998) and Williams (2003) have traced the impact of Schmitt; Balzacq (2005), Vuori (2008) and Floyd (2010) have drawn on Austin and Searle and presented new speech-act-based securitization theories; Stritzel (2007) has discussed the role of Derrida and Butler; and Salter (2008) has introduced Goffmann in a microsociological approach to the settings within which securitizations take place.
The first aim of this article is to add to this body of literature by exploring in more detail the post-structuralist influence on securitization theory. Stritzel (2007) and Floyd (2010) have aptly noted that there is an indebtedness to Derrida and Butler, one that I believe is worthy of further pursuit. 1 Following Campbell (2007: 206), I understand post-structuralism as ‘a critical attitude, approach, or ethos which calls attention to the importance of representation, the relationship of power and knowledge, and the politics of identity in the production and understanding of global affairs’. While there are differences in the positions of post-structuralist theorists such as Derrida and Foucault, the main approach of post-structuralists in international relations has been to highlight the ways in which these positions can be combined. The reason for (re)discovering securitization theory’s post-structuralist roots is not merely that of providing an archival corrective that catalogues the influence of post-structuralist texts on Buzan and Wæver. The impetus behind the article is rather that a more (explicitly) post-structuralist securitization theory may enable a more critical engagement with the politics of security – that is, the way in which security discourses form the objects of which they speak (Foucault, 1974: 40–9; Campbell, 1992), as well as how they draw upon and provide particular subjects with authority and power while closing off others.
The second aim of the article is to construct an outline of a securitization theory that is based on post-structuralism but targeted to securitization analysis and analytically applicable. Based on Foucault’s concept of discursive formation, I therefore suggest three specific concepts: structural incorporation (the substantial signs that make up discursive formations), epistemic terrain (the episteme asserted or assumed), and substantial modality (the type of logic through which ‘security’ is constituted).
The third aim of the article is to show the distinctiveness of post-structuralist securitization theory in an illustrative case-study of the Muhammad cartoon crisis. Except for mostly shorter analyses in Danish, there has been virtually no security studies engagement with the cartoon crisis (Hansen, 2011), while the crisis has been embraced by adjacent academic fields such as political theory, media studies, geopolitics and anthropology (Lindekilde et al., 2009). Thus, the fourth aim of this article is to show which specific insights post-structuralist securitization theory can bring to the existing body of works on the cartoon crisis.
The article proceeds as follows. The first section traces securitization theory’s post-structuralist genealogy through a reading of Buzan and Wæver’s foundational texts and how these have been read by later securitization theorists Balzacq (2005) and Stritzel (2007). This reading allows me to clarify basic post-structuralist understandings of materiality, speech and ‘reality’. The second section introduces in more detail the concepts of discursive formation, genealogy and deconstruction, and suggests the three conceptual expansions – structural incorporation, epistemic terrain and substantial modality – that bring the concept of discursive formation closer to the study of security. This section also explains where and how my post-structuralist revision departs from or concurs with other attempts to expand securitization theory. 2 The third section is devoted to the analysis of the Muhammad cartoon crisis.
A post-structuralist reading of securitization theory
As Floyd (2010: 9–42) has shown, the intellectual genealogy of Buzan and Wæver’s securitization theory is complex, and the fusion of (neo)realism and post-structuralism not without slips and ambiguities – a ‘real existential threat’ surfaces, for instance, in the presentation of security’s self-referentiality in Buzan et al. (1998: 24). No reading can therefore redeem – or doom – Buzan and Wæver’s position as one that is inherently, only, always and securely post-structuralist. Nor should the criticism, my own included (Hansen, 2000), that parts of Buzan and Wæver’s theoretical corpus are insufficiently post-structuralist – for instance, in relation to the stance on silence – be retracted. The reading below is thus neither an exoneration nor a canonization, but a (re)constructive, encouraging reading that identifies the post-structuralist gist of securitization theory in order to push that theory further.
A post-structuralist understanding of discourse is at the centre of Buzan and Wæver’s theorization of security as a speech act and ‘a self-referential practice’ that constitutes something as a security issue; of their ‘definition and criteria of securitization’ as being ‘the intersubjective establishment of an existential threat with a saliency sufficient to have substantial political effects’; and of their epistemological and methodological stance that securitization should be studied in discourse as it does not need extra-discursive ‘indicators’ (Buzan et al., 1998: 24–6). 3 Securitization theorists, however, have argued that these passages indicate an ambiguity in Buzan and Wæver’s theory. According to Balzacq (2005: 179–81, 195n16), there is a fundamental difference between understanding security as ‘a self-referential activity’ and as ‘an intersubjective process’. Buzan and Wæver lean, holds Balzacq (2005: 181), toward the former understanding, which is a post-structuralist, internalist one that understands concepts as shaping ‘their’ contexts, in principle to the point where ‘what is “out there” is [thus] irrelevant’. Balzacq (2005: 181–2) calls for an externalist position that acknowledges the existence of ‘external or brute threats’, and for an ontological view of language as not constructing reality but ‘at best’ shaping our perception of it. Securitization theory, advises Balzacq (2005: 176, 178), should conceptualize context as an ‘explanatory variable’ that influences the ability of securitizing actors to persuade an audience, and audience and context should be accorded ‘causal weight’. Stritzel (2007: 378n6) holds that Balzacq devotes too much attention to the procedural side of speech-act philosophy, and he discusses more thoroughly the post-structuralist impact on Buzan and Wæver’s writings (see also Vuori, 2008: 74). Yet, Stritzel (2007: 369) suggests a distinction between an internalist, more post-structural position indebted to Derrida and Butler and an externalist one concerned with how security is impacted by contexts, the sociolinguistic (‘the network of constitutive rules and narratives that surround a single linguistic act’) and the sociopolitical (‘sedimented social and political structures that put actors in positions of power to influence the process of constructing meaning’).
If one comes to securitization theory through post-structuralism, though, it is not clear that Buzan and Wæver face the problems or tensions that Balzacq and Stritzel attribute to them. First, the self-referentiality of security does not imply that it is disconnected from intersubjective processes or, as Balzacq (2005: 179) seems to suggest, that it negates the audience. Rather, self-referentiality, as conceptualized by Wæver and Buzan, identifies the absence of an extra-discursive criterion upon which security can be grounded. This ontological view of security leads the security analyst to take an epistemic view of discourse as intersubjective processes through which securitizing actors make claims about their ability to identify ‘security as it really is’, usually by constituting ‘the meaning of context’. Balzacq appears, in other words, to conflate Buzan and Wæver’s ontological conception of security with their epistemological theorization of security as a discursive practice.
Second, Balzacq’s claim that Buzan and Wæver take an internalist position that neglects context – ‘external or brute threats’ – reinstalls the materialism–idealism distinction that post-structuralists have laboured to deconstruct. As in the frequently quoted passage from Laclau and Mouffe (1985: 108), post-structuralists ‘affirm the material character of every discursive structure. To argue the opposite is to accept the very classical dichotomy between an objective field constituted outside of any discursive intervention, and a discourse consisting of the pure expression of thought’. As Campbell (1992: 6) points out, this does not imply that all representations are equally powerful or capable of manifesting themselves as speaking on behalf of ‘external reality’, but rather that ‘the idea of “external reality” has a particular currency that is internal to discourse’. As post-structuralism – along with Buzan and Wæver – fails to accept an extra-discursive context, it is misleading to hold that Wæver sees the speech act as ‘not related to its context’ (Stritzel, 2007: 361), as this presupposes context being recognized and then rejected. It also presupposes a decoupling of the speech act from the discursive structures through which ‘context’ is constituted.
Third, Stritzel’s reading of Buzan and Wæver’s theory as devoid of context rests upon a problematic understanding of Butler’s notion of performativity, a concept taken up both by Wæver and by post-structuralists in international relations (Campbell, 1992; Weber, 1998; Hansen, 2006). Stritzel (2007: 362) holds that performativity implies a view of actors and their social relations as constituted ‘retroactively’. Were this the case, one might see why Stritzel would conclude that Buzan and Wæver assign language supreme power over ‘context’. But, Butler’s point is that the subject appears not after the speech act but in it, as ‘identity is performatively constituted by the very “expressions” that are said to be its results’ (Butler, 1990: 25). As Weber (1998: 79–81) underscores, because identity is produced through ongoing performative reiterations, there is not a ‘post-performative’ moment where the subject – individual or state – retroactively looks back.
A clarification of the concept of performativity may also explicate the concept of audience. Given that audience acceptance – or, at a minimum, toleration (Buzan et al., 1998: 25) – constitutes the difference between an attempted and a successful securitization, this concept has rightly been singled out as in need of further discussion (Balzacq, 2005). Post-structuralists like Butler generally use the concept of subject rather than audience: the former is seen as constituted in discourse, whereas most usages of ‘audience’ presume this to be an entity that exists prior to speech. The latter position is illustrated by Balzacq’s (2005: 184) description of an audience whose feelings, needs, interests and experiences a securitizing actor should identify with. The Copenhagen School’s use of the term ‘audience’ might therefore provide the theory with a concept that sits uneasily with post-structuralism. Yet, Wæver’s (2002: 26–7) stance that what is of interest is not what a ‘population’ believes ‘but which codes are used when actors relate to each other’ supports a ‘post-structuralization’ of the concept of audience. This ‘audience’ appears in discourse as a constituted subject whose ‘opinion’ might be detected through surveys, polls or elections. This does not mean that responses to securitizing discourses are irrelevant, but these are important precisely because they enter the discursive, political sphere, not because they stem from an extra-discursive ‘audience’. Thus, when ‘members of the audience’ speak – invoking the subjectivity of ‘audience’ – they speak for that discursively constituted subject, not from it.
A post-structuralist deepening of securitization theory
The advantage of the concept of discourse is thus that it provides a way to theorize the constitution of ‘materiality’ and the impact of ‘external reality’ without recourse to the objective–subjective, material–ideational dichotomies. 4 Discourse in post-structuralist theory also entails an understanding of language – and thus of the subjectivities that are invoked in the legitimation of security policies – as simultaneously structured and unstable. The structural aspect is stressed in Foucault’s formulation of the discursive formation as ‘a system of dispersion … between objects, types of statement, concepts, or thematic choices … [that] define a regularity’ (Foucault, 1974: 38; emphasis added). Discursive formations are subject to particular rules of formation, but these rules are not settled. As Foucault’s studies of academic disciplines and what he defines as their episteme show, ‘the discursive practices that give rise to epistemological figures, sciences, and possibly formalized systems’ are situated within ‘a given period’ (Foucault, 1974: 191). Foucault (1978: 115) also stressed that history may proceed through ruptures, but that it is often told as though it were a seamless process where such ruptures and ‘deviant’ subjectivities and assumptions have been forgotten. As a ‘history of the present’, genealogy is positioned by Foucault in opposition to such a ‘metahistorical’ teleology, and it ‘must be sensitive to their [i.e. events’] recurrence, not in order to trace the gradual curve of their evolution, but to isolate the different scenes where they engaged in different roles’ (Foucault, 1984: 76–7). Genealogy must also ‘define even those instances when they are absent, the moment when they remain unrealized’ (Foucault, 1984: 76).
The understanding of language as structured, yet inherently unstable, is also at the core of the work of the other main post-structuralist to have influenced post-structuralist security studies, including Wæver’s work, namely, Derrida. As Derrida’s famous notion of deconstruction stressed, language is made up of dichotomous pairs, where one side of each dichotomy is deemed superior to the other. Yet, by working through these dichotomies, one can show how such superior–inferior relationships depend upon a series of unstable distinctions that ultimately fail to give the text the ‘security’ it seeks; there is, in other words, an inherent undecidability in any text or discourse (Norris, 1987: 34–5). As many of Derrida’s deconstructions focused on dichotomies said to govern political thought from the classics onwards, such as writing–speech, Derrida also took dichotomies to have a historical status that linked their use across specific texts. While there are certainly differences between Foucault and Derrida, they converge on providing post-structuralist concepts that facilitate critical engagement with the ways in which political discourse and text works. 5
Moving to Buzan and Wæver’s securitization theory, the influence of Foucault is perhaps most apparent in their conceptualization of security as a particular form of discourse or modality (Wæver, 1995: 70). Security is a discursive formation that identifies a particular political system of dispersion that includes a type of statement (threat, defence, emergency), objects (referent objects) and thematic choices (survival). The securitized is in turn part of the larger meta-political discursive formation that also includes the politicized and the non-politicized (Buzan et al., 1998: 23). Within the discursive formation of the securitized, Buzan and Wæver’s theory identifies as specific discursive forms five sectors – military, environmental, economic, societal and political – each with specific referent objects and actor-threat-vulnerability dynamics. This reading of the influence of Foucault’s (at least partially structural) concept of discursive formation leads me to disagree with Stritzel’s (2007: 374) suggestion that ‘Structural elements of the Copenhagen School are rather weak’. However, I do concur that securitization theory might be further specified in three ways that link post-structuralism, predominantly through Foucault, to the study of security.
The first suggestion is to further highlight the specific content of the discursive formations that strive to assign ‘security’ to a given event, case or process, through structural incorporation. Buzan and Wæver note that ‘the security speech act is not defined by uttering the word security’ (Buzan et al., 1998: 27). The specific sign ‘security’ might not invoke the political modality of security, and the security modality might be invoked through words other than ‘security’ or ‘threat’. 6 This, as Vuori (2008) shows, might constitute a particular challenge in non-democratic settings, but the challenge is also present in democratic ones. We may therefore supplement Buzan and Wæver’s theory with post-structuralist theory – and analysis – that traces the structuring significance of the terms through which ‘security’ is ‘made’ or ‘unmade’, either by looking to the way in which ‘security’ is linked to other signs within specific texts and discourses (Trombetta, 2008: 587) or by examining genealogically how it resides within concepts that no longer explicitly articulate ‘security’. 7 One example of such an analysis is Trombetta’s (2008) powerful study of how climate and environmental security discourse shows signs of breaking with the political modality of ‘security’ as described by Buzan and Wæver.
In terms of supplementary theorization of the ‘sign significance’ of security, one option is Wæver’s (2002) theory of foreign policy as ‘layered discursive structures’, where foreign and security policies towards ‘Europe’ are understood through concepts of state and nation. Another possibility is to go through the recently launched concept of macrosecuritization – that is, those securitizations ‘that speak to referent objects higher than those at the middle level (for example, “universal” religions or political ideologies; one or more of the primary institutions of international society) and which aim to incorporate and coordinate multiple lower level securitisations’ (Buzan and Wæver, 2009: 257). In terms of structural incorporation, the second part of the definition describes the ability of macrosecuritizing discourses to structure other security discourses. Thus, to articulate signs that resonate with those of an established macrosecuritization is to bring in that wider spectre of universality and global range. A third possibility is my own theory of ‘basic discourses’ derived from the representations and policy options that are available within a particular political field – such as, for instance, ‘Balkan’ and ‘genocide’ in the case of the Bosnian War (Hansen, 2006). Crucially, here, the representations around which basic discourses are structured are not necessarily the most empirically prevalent terms. Representations might also recede from explicit use but still structure political debate if their meaning is transferred to other terms or deeply institutionalized. This in turn illustrates the difference between post-structuralist discourse analysis and content or framing analysis focusing on the explicit frequency of terms (Vultee, 2010).
The second suggestion is to highlight the epistemic terrain upon which securitizations are constituted. Buzan et al. (1998: 32–3) hold that successful speech acts require a compliance with the internal conditions of the speech act – that is, securitizing actors need to follow the right grammar and plot structure as well as ‘the particular dialects of the different sectors’. Among the external conditions are the social capital and authority of the securitizing actor. Securitization theorists like Balzacq (2005: 190–1) and Vuori (2008: 70–1) have, as noted above, drawn further attention to the significance of authority and knowledge for whether securitizing actors succeed in convincing audiences, mostly through Bourdieu’s concept of capital. Yet, the epistemic terrain draws attention to the slightly different question how an episteme – in Foucault’s (1974: 191) terminology, ‘the total set of relations that unite, at a given period, the discursive practices that give rise to epistemological figures, sciences, and possibly formalized systems’ – is asserted or assumed within debates on security. The shift, in other words, is from actors to regimes or discourses of knowledge. This draws upon a longer post-structuralist attention to the significance of genre for knowledge production (Shapiro, 1988), an attention that resonates with Salter’s (2008: 328) argument in favour of seeing genre as important for how securitizations are made. Following general post-structuralist thought, the epistemic terrain is understood as inherently unstable and conflictual; thus, the analytical question becomes how understandings of epistemic certainty are mobilized and thought stabilized within security discourses.
The third suggestion for how to foreground post-structuralism within securitization theory is through an engagement with substantial modality, that is, the societal sphere a security issue is said to be situated within. Substantial modalities have epistemic aspects in that what constitutes acceptable knowledge shifts from say a legal to a religious modality, but the concept of structural modality is closer to Buzan and Wæver’s sector in that it identifies substantial logics through which objects and subjects are constituted. This implies in turn that solutions to security problems vary across modalities. A legal modality situates the judiciary as the discourse and institution through which arbitration should be sought, whereas a religious modality implies a different set of discursive and political options. Buzan et al. (1998: 7–8) hold that sectors must be aggregated to get a full analysis, but the claims that each sector has a distinct referent object–threat dynamic and that sectors have been introduced to account for new phenomena on the political agenda (Wæver, 1999: 335; Laustsen and Wæver, 2000: 705) lend support to the view of sectors as ‘neatly separated’ (Stritzel, 2007: 366; Trombetta, 2008: 589). Some securitizations might have implications for several sectors, but their relationship can then be sorted out analytically. Substantial modality by contrast takes a more critical view of the logics through which referent objects, threats and securitizing actors are linked. Rather than securitizations belonging to specific sectors, it assumes that there is an undecidability in terms of which referent objects are invoked and therefore for whom and how ‘security’ should be achieved (Hansen, 2000). Even in those cases where referent objects appear straightforwardly and stably articulated, and the difference between an analytics of sector and one of substantial modality rather minimal, this implies a more critical, deconstructivist approach to security discourses. As in Buzan and Wæver’s securitization theory, an analysis of substantial modality implies a concern with the wider public sphere within which securitizations – and their substantial modalities – are made and compete. Salter’s (2008) Goffmannian theorization of settings offers by comparison a microsociological focus on the specificity of different epistemic communities, political locations and media.
A post-structuralist securitization analysis
Turning now to how a post-structuralist securitization theory can be applied in an empirical analysis of the Muhammad cartoon crisis, given space constraints, what follows is an indicative illustration with a particular focus on bringing out the distinctiveness of the three concepts laid out above. From a securitization perspective, the first question is of course whether the cartoon crisis was a case of securitization at all. Here, the concept of structural incorporation provides with us a good place to begin an answer.
Structural incorporation: ‘The clash’ as a macrosecuritization
From the first publication of the cartoons in the Danish newspaper Jyllands-Posten, the editor acted as a securitizing actor. His editorial – tellingly entitled ‘The threat from the dark’ – explained that ‘the Western world’ is confronting ‘threatening voices from the dark Middle Ages’, those of ‘Islamic spiritual leaders’ who are so ‘sickly over-sensitive’ to any ‘disagreement’ that incidents like the murder of Theo Van Gogh take place (Jyllands-Posten, 2005). In the vocabulary of securitization theory, the threatened referent object is ‘the West’ as the embodiment of ‘Enlightenment’ values, and the threat stems from ‘Muslim’ imams, as well as from a wider ‘Islamic worldview’. As freedom of speech constitutes an – if not the – identity of ‘the West’, its endangerment equals an existential threat. Given this level of insecurity, Jyllands-Posten demands that ‘Western society’ and ‘Muslims’ resist. The securitizing status of Jyllands-Posten’s text becomes even clearer if one traces the way in which it is structurally incorporated within the macrosecuritizing discourse of ‘the clash of civilizations’, a discourse that was already part of the discursive formation that made up security discourse in Denmark and beyond. Even if not explicitly articulated, a conflict between ‘Muslims’ and non-Muslims is also evident in the first major ‘counter-securitization’, a letter sent to Danish Prime Minister Anders Fogh Rasmussen in October 2005, signed by eleven ambassadors from ‘Muslim’ countries. 8 Claiming to speak on behalf of the ‘Muslim world’ (the referent object), the diplomats work as securitizing actors, pointing to Danish Minister of Culture Brian Mikkelsen’s ‘statement on war against Muslims’ and the ‘demeaning caricatures of Holy Prophet Mohammad’, which the diplomats ‘strongly feel’ went ‘against the spirit of Danish values of tolerance and civil society’ (existential threats). The diplomats do not issue an explicit threat to ‘Denmark’, which would have been unusual in a diplomatic request, but they nevertheless point out that this ‘discriminatory tendency’ can ‘cause reactions in Muslim countries and among Muslim communities in Europe’. ‘Given the sensitive nature of the matter’, the diplomats conclude by requesting ‘an urgent meeting’ with Prime Minister Fogh Rasmussen.
Fogh Rasmussen refused the meeting, and events escalated over the following months, reaching a peak in January–February 2006, when extraordinary measures adopted by governments and non-state actors included the withdrawal of ambassadors from Copenhagen, a massive boycott of Danish products in the Muslim world, and large-scale riots across the streets of the Middle East, Asia and Africa. Reportedly up to 250 people were killed, mostly in Nigeria, Libya and Afghanistan (Klausen, 2009: 107), and authorities – most conspicuously in Damascus – were unwilling or unable to prevent the torching of embassies, another token of acceptance of the extraordinary. Nonviolent events, such as a demonstration in London organized by 650 mosques and religious organizations that brought out ten thousand people, showed that the cartoons mobilized audiences and actors beyond Denmark and the Muslim world (Klausen, 2009: 193). By March 2006, demonstrations had dwindled, but the cartoons failed to disappear, although they were significant mainly for debates within Denmark and for Denmark’s relationship to ‘the Muslim world’ (see Klausen, 2009 for a thorough account).
On one hand, from a securitization perspective, whether Jyllands-Posten and the ‘Muslim’ diplomats securitized the cartoons – and how – is of course an important question. On the other hand, what is more important still is the discursive constitution of ‘the crisis’ as one of civilizational clash, even if such a representation was ‘forced on people’ (Sacco and Spiegelman, 2006) or ‘imagined’ (Eide et al., 2008; see also Lindekilde et al., 2009: 296). The fact that the cartoon crisis was constituted not as a freestanding event with its own specific, independent discursive dynamics, but as an instance of the macrosecuritization of ‘the clash of civilizations’, one of whose subdiscourses was the ‘war on terror’, had several implications. First, as observers like Asad (2009: 20) pointed out, as the crisis unfolded, the link between the cartoon crisis and the War on Terror became increasingly pronounced and such linkages radicalized and securitized subjects on each side of ‘the war’. The bundling character of ‘the clash of civilizations’ also had a more indirect effect: even when the ‘war on terror’ was not explicitly invoked, it still lurked as an ambiguous potential connection. Second, the constitution of ‘the crisis’ as one of clashing civilizations implied that expressions, decisions and choices became constituted as signs of a civilizational identity (Asad, 2009: 21). The claim, for instance, that the cartoons were problematic on the ground that they were disrespectful of ‘Muslims’ became a sign of ‘the anti-democratic’ ‘Muslim’ Other, rather than a specific, legitimate engagement in public debate (Rostbøll, 2010: 412). Third, ‘the clash of civilizations’ made ‘religion’ the subject category through which the crisis should be understood. Yet, ‘the Muslim’ was not only a religious subject, but one that crystallized race, culture, gender and politics. Politically, this Muslim subject was constituted as sceptical of or radically opposed to democracy, free speech, liberty and individual rights; culturally, its practices, customs and aesthetics were different from those of ‘the West’ (Levey and Modood, 2009: 442). Crucially, this ‘religious stereotyping’ also involved a racialization (Levey and Modood, 2009: 441–3; Lindekilde et al., 2009: 303). The incorporation of the religious, the racial, the political and the cultural within one ‘Muslim’ subject not only homogenizes difference; it also produces an almost impenetrable subject whose ‘entire way of life’, history and culture is condemnable (Rostbøll, 2010: 406).
The epistemic terrain: Visual representation and speech
In terms of the epistemic dynamics, a central issue was whether the cartoons violated a ban on depicting Muhammad or whether they were to be judged in terms of how they portrayed him. Several analysts pointed out that the prohibition on depicting Muhammad applied to non-Muslims only (Levey and Modood, 2009: 430), and that not all Muslims agreed with a general taboo (Klausen, 2009: 137–9). For those Muslims who held the latter view, what was offensive about the cartoons, said Klausen (2009: 139), was not visual representation as such, but the fact that it happened ‘in a disrespectful manner’. Moving from ‘general Muslim’ readings of the cartoons to those of academic analysts themselves, there was considerable disagreement as to whether the cartoons were offensive or even hateful, and thus as to whether they ‘spoke security’ (Post, 2007: 84; Levey and Modood, 2009: 441, 443). Disagreement also reached to the level of epistemic strategy, as some analysts treated each cartoon separately while others discussed them as a homogeneous group exemplified by Kurt Westergaard’s ‘bomb cartoon’ (Spiegelman, 2006; Klausen, 2009: 20–7).
The specific question of whether ‘the’ cartoons were offensive opened up a wider epistemic issue, namely, that of certainty – more specifically the certainty with which the observer could determine what the cartoons ‘uttered’. Levey and Modood (2009: 429, 437, 439, 445n4) held that Westergaard’s cartoon ‘was intended or may reasonably be read’ as commenting on the faith of ‘Islam’ as well as on ‘Muslims’, and that while the cartoon might be seen as ‘ambiguous’, it was the responsibility of the cartoonist or the editors at Jyllands-Posten to rectify this. Others, including Spiegelman (2006: 48–9), argued that Westergaard’s claim that his cartoon represented not ‘Muslims’ or ‘Islam’ but terrorist and murderous fundamentalism should be taken into account. The degree to which the cartoons could be said to speak offensively is important, not least because it raises questions of agency and subjectivity. If the cartoons were offensive, the relation between them and an offended Muslim audience is simply one of publication and reception. If the cartoons were to be constituted as offensive, the relation is one where agency and interpretation are required. Audiences, Muslim as well as non-Muslim, have to be made aware of what the cartoons signify – and what ‘the’ proper response is. Thus, rather than reaching ‘a’ Muslim subject or audience, the cartoon crisis becomes a process through which ‘Muslims’ are constituted as such.
The cartoons were not just images; they were also situated within the realm of the satirical, and this in turn raised the question whether that particular political-epistemic realm was different from the realm of ‘honest’, ‘literal’ or ‘factual’ representations. Most analysts were sceptical of Jyllands-Posten’s claim that to ridicule ‘Muslims’ was to treat them as equal citizens or that the invocation of satire would automatically exempt Jyllands-Posten or the cartoonists from charges of offensiveness. That said, some also acknowledged the situated character of caricature, irony and humour, and that the cartoons had been produced for a Danish audience whose sense of permissibility might differ from non-Danish ones.
Substantial modality: Law, religion and reason
What comes out strikingly when we look at the cartoon crisis through the analytical lens of structural modality is a battle over which modality and hence which referent object should ‘govern’ the crisis. Most commentators agreed that the legal realm and the question whether the cartoons were protected under ‘free speech’ or whether they were blasphemous and/or instances of hate speech was too narrow. Asad (2009) held that the invocation of law itself rested upon a particular liberal view of religion and the political subject. The liberal assumption, explained Asad, is that belief cannot be coerced; it resides within the individual mind and falls, therefore, outside of the field of political regulation. The Muslim position, by contrast, sees religious belonging as something that can be forced or seduced, ‘but it is illegitimate to do so’ (Asad, 2009: 45). For Muslims, what happens when blasphemous images, such as the cartoons, are published is thus not a matter of ‘free speech’ or a legal transgression, but a challenge to ‘a living relationship’ (Asad, 2009: 46). Or, in the words of Mahmood (2009: 78), a perception that ‘one’s being, grounded as it is in a relationship of dependency with the Prophet, has been shaken’. 9 Accordingly, to refer to the legal system as that which should arbitrate whether the cartoons were offensive or not might through its privileging of the Western, liberal position marginalize those who call for a critical engagement with – although not necessarily a ban on – the publication of the cartoons. And, if the referral of the cartoons to the legal modality is one that simultaneously takes ‘the crisis’ outside of the political, then the issue is fully depoliticized (Hansen, 2011: 302–3).
Yet, while ‘religion’ was clearly at the centre of the crisis, one might also caution against seeing the crisis as belonging squarely to the substantial modality of religion. Muslims ‘demanded that their religion not be mocked or ridiculed’, but Muslims also held that the cartoons were emblematic of wider anti-immigration and anti-Muslim policies that would situate the crisis as one over political rights and respect rather than (solely) religion (Rostbøll, 2010: 407, 415). The question whether the crisis is constituted as political or religious in terms of ‘the Muslim’ subject is important because it charts two different political and ontological terrains. If the conflict is religious and over ‘faith’, then the ‘Muslim’ subject is one who is concerned with the extent to which his or her religion is violated or respected; if the conflict is political, then the ‘Muslim’ subject is one who challenges the way in which he or she is treated in terms of rights and political respect. If ‘the Muslim’ is political, he or she meets ‘the West’ on a shared political terrain; if ‘the Muslim’ is religious, one has two potentially antagonist terrains. The constitution of ‘the Muslim’ response as a securitization of religion rather than political status also informed the notion that this was a response motivated by ‘religious feelings’. This might be empirically accurate in some – perhaps even most – cases, but this representation also reinforced the dichotomy between a reasonable, liberal West and an emotional, religious Muslim world. The linking of reason–West and feelings–Muslims implied that liberalism was constituted as devoid of feelings, or that feelings and religion were inevitably intertwined.
Conclusion
This article has trekked one of the ancestral trails that has remained virtually non-travelled in the past decade’s concern with expanding the range and depth of securitization theory, namely, that of post-structuralism. Rather than taking a chronological route, I opted for a more genealogical approach, starting with how recent securitization theorists have analysed the impact of post-structuralist philosophy on Buzan and Wæver’s canonical writings. My critique claimed that while Balzacq and Stritzel set out to rescue securitization theory from its alleged internalist reliance on an idealized notion of language and its ‘vague and undertheorized terminology’ (Stritzel, 2007: 358), Buzan and Wæver are less in need of rescue than Balzacq and Stritzel suggest. Both theorists offer, in turn, revisions that break with Buzan and Wæver’s discursive, post-structuralist assumptions. 10 Whether one wants to go down the route indicated by Balzacq and Stritzel or to take the more post-structuralist approach that I then laid out in the second part of the article is up to each securitization scholar to decide.
The three conceptual elaborations of securitization theory facilitated an analysis of the cartoon crisis that emphasized points that previous studies have either downplayed or overlooked: first, the implications of the structural incorporation of ‘the crisis’ within the macrosecuritization of ‘the clash of civilizations’ for the subjects that were constituted; second, the way in which debates over interpretation and epistemic certainty were part of ‘the crisis’; and, third, the extent to which the crisis was one over political logics and referent objects. The three conceptual elaborations constitute one possible expansion of more general post-structuralist ideas. I think it thus quite likely that other post-structuralist attempts at deepening securitization might follow, in a similar way to that in which the engagement with Searle and Austin has produced a distinct body of second-generation securitization theories.