Policing Violent Extremism: How the Global War on Terror Meandered Into Local Municipal Policies in Sweden

Introduction

There has been a continuing discussion within academia over the past decade concerning how the war-on-terror discourse (Hodges, 2011) and its counterpart, prevention of violent extremism, affect traditional pedagogical and social work, the so-called first-line professions. One focal point in that discussion has concerned the basis for and outcome of instructing first-line professionals to look for signs of radicalization and detect youngsters vulnerable to becoming terrorists (Davies, 2008; Hodges, 2011; Husband & Alam, 2011; Kundnani, 2014). It is also clear that the discussion is ongoing; examples from several different countries, for instance, the United Kingdom, the Netherlands, France, Denmark, Norway, and Sweden, show that they are adopting various measures meant to prevent violent extremism (PVE). One measure is to install action plans to PVE, often referred to as PVE work or CVE (countering violent extremism) work (Davies, 2008; Heath-Kelly, 2013). Sweden had its first national action plan for this purpose in 2011 (Regeringens skrivelse, 2011); that plan was followed by a national investigation on how to provide efficient measures against violent extremism (Statens offentliga utredningar [SOU], 2013) as well as the installation of a National Co-ordinator (NC) against violent extremism (Kommittédirektiv, 2014). The NC has underlined the responsibility of local municipalities to serve as gatekeepers in preventing recruitment to violent extremism and terror (SOU, 2016). Until April 30, 2017, 134 Swedish municipalities, out of 290, had adopted local action plans enabling them to provide PVE services. This often involves new tasks for teachers, social workers, and youth workers in detecting high-risk individuals and communities. It is, therefore, crucial to understand how these local action plans are constructed and adopted; how they play out against the traditional duties of teachers, social workers, and youth workers; and how they are related to international approaches to PVE work.

The 290 Swedish municipalities are the prime agencies in providing education and social services for Swedish citizens. Preventing criminality and providing security is the duty of the police, which is a national authority. To understand how Swedish employees at the municipal level, that is, teachers, social workers, and youth workers, are assigned to perform work traditionally considered the responsibility of the state, we must seek to understand how the PVE discourse has undergone a process of policy transference on an international arena. PVE policies are transferred from the international arena to a national Swedish arena and thereafter they are re-contextualized in various Swedish municipalities. We will also observe what the consequences of this might be.

At the beginning of 2016, the NC reported that only 7% of Swedish municipalities had local PVE action plans (Nationella samordnaren mot våldsbejakande extremism, 2016); this was criticized by the NC and gained significant attention in the media (Delling, 2016; Juhlin & Burström, 2016, TT, 2016). This obviously led to a rapid process, given that only 1 year later, more than 40% of municipalities had adopted such plans. The NC to some extent paved the way for this development not only by arguing for the necessity of these plans but also by providing a 12-point program (Nationella samordnaren mot våldsbejakande extremism, 2015a) as a guideline for how to write a PVE action plan as well as a knowledge resource on how to detect and understand violent extremism called the Conversation Compass (Nationella samordnaren mot våldsbejakande extremism, 2015b). These recommendations and knowledge resource specifically point at the important role of the educational system and teachers in PVE work. The present article will focus on how municipalities are managing their responsibility to PVE and in particular on how this is affecting the teaching profession. Critical discourse analysis (CDA) has been deployed in investigating the process of policy borrowing when these first 127 local action plans were produced in 134 municipalities, some of which have adopted action plans together. By the assistance of CDA, the study will help to clarify the discursive practices in relation to the growing Swedish PVE work.

Aim and Research Questions

The aim of the article is to investigate how the Swedish local PVE action plans are constructed, with particular focus on the role of the educational system and teachers. The article will investigate how teachers are sought to be deployed in PVE work and how these concepts are the result of vertical policy transference from the national and international policy production levels as well as how various concepts are horizontally transferred between various Swedish municipalities. Of particular interest is teachers’ specific role within pedagogical PVE work. Policy borrowing, as a theoretically informed methodology, is used in the present article to understand this process, but the theoretical basis is CDA. The following research questions (RQ) have guided the investigation:

The article is constructed as follows: I first provide some background on how the transfer of PVE polices is undertaken from the international arena to the national Swedish arena and then an overview of relevant previous research within the scope of the article. Afterward, I discuss the theoretical and methodological starting point, before presenting the data, analysis, and concluding discussion.

The Origin of the Local Action Plans

Local Swedish PVE action plans were adopted over a few months. Before April 2016, there was just a small number of local action plans; half a year later, the vast majority of the 127 plans in 134 municipalities, which provide the data for the article, had been adopted. Some background is required to understand what enabled this process to occur so rapidly and where the policy ideas came from. It may seem obvious that the policy ideas were transferred, vertically, from the European Union (EU) to the national Swedish level. However, Balls (2008) pointed out that policy borrowing is seldom a one-way process: it involves networks of policy producers. The rock bottom of this process, according to Ball (2008), is the notion that the state is no longer the single predominant actor for producing welfare-related policies, but is now challenged by a market-oriented understanding of welfare production that involves networks of policy producers operating on an international arena. This also seems to be relevant to understanding the policy transfer process concerning PVE work and can hardly be overlooked if we wish to understand how local Swedish PVE action plans are constructed. The EU considers itself to be an actor involved in transferring PVE policies to (vertically) and between (horizontally) member countries. This goes back to the EU counterterrorism strategy from 2005 (Council of the European Union, 2005), which includes aspirations to PVE by assisting development of national and local action plans.

We learn, by reading the strategy document, that the EU has installed a particular working group, Radicalisation Awareness Network (RAN, 2010), whose duty is to support member countries’ efforts to prevent radicalization. According to RAN’s presentation of its work, we learn that

The RAN CoE acts as a hub for connecting, developing and disseminating expertise. It supports and coordinates RAN, and fosters an inclusive dialogue between practitioners, policy-makers and academics. (RAN, 2016a, p. 9)

What we see is a hub where various professionals, experts, interest groups, and policymakers mingle within a framework that has the authority to distinguish between relevant and irrelevant PVE strategies, measures, and policies. This is in good accordance with Ball’s (2008) findings on how policies are created within networks rather than in a top-down or bottom-up vertical process. This makes the RAN hub particularly interesting as regarding studying who is providing whom with policy ideas and for what reason. One important outcome of RAN is its collection of various methods sampled from the member countries (Mattsson, Hammarén, & Odenbring, 2016), 98 of them in total. These then form the basis for various in-service training courses for so-called first-line practitioners, but RAN also produces workshops for policymakers (RAN, 2016b). Within RAN, we can see a cycle in which nongovernmental organizations (NGOs) sign up to provide methods and experiences to the RAN network that are then transformed into content for various forms of recommendations and returned to the member countries as directions for polices. In other words, we have a body that is capable of extracting local experiences relevant to their preset understanding of preventing violent extremism and then sending it back as a basis for policies.

This leads to the next question: What is transferred? Obviously, one transfer consists of the collation mentioned above as well as the various workshops. But we also learn that RAN stresses the importance of locally grounded strategies:

The early detection and prevention of radicalisation is most effective at the local, community level. The RAN brings together local prevention coordinators and has developed advice on how to create local prevention frameworks and interventions. (RAN, 2016c, p. 2)

What we see is that RAN is urging its member countries to build their capacity on a local level and, naturally, to make use of the directions and methods it provides. So it is not a prewritten policy that is transferred, but rather lessons learned and certain ways of understanding violent extremism.

Within the Swedish context, this has been followed up by firm recommendations from the Government, that is, the NC to PVE, to the municipalities, encouraging them to adopt local PVE action plans (SOU, 2016). The NC has developed, as already mentioned, specific guidelines with 12 recommendations for a local action plan (Nationella samordnaren mot våldsbejakande extremism, 2015a); the purpose of the guidelines is to support the municipalities in their efforts to develop these plans. The advice given in this recommendation is quite general and includes having preventions at different levels, co-operating within established infrastructure, and analyzing the local needs. But the guidelines do indicate that schools, social welfare services, and the police are indispensable partners as well as recommend use of a particular method/policy doctrine called the Conversation Compass (Nationella samordnaren mot våldsbejakande extremism, 2015b) as a basis for educating policymakers about violent extremism.

Survey of the Field

There is little research on how policy borrowing contributes to establishing the spread of PVE work between and within countries. But there is one important piece by Keohane (2008), who pointed out the structuring process between the EU, which needs its members to contribute to the overall goals of the Union’s counter-terror strategy, and the member countries, which need independence in terms of each country’s own national strategy. This process seems to unify the national strategies in accordance with EU-level goals. This is one way to understand how so-called forceful measures to prevent terror within the counter-terror strategies have come to include and distribute ideas about soft measures against terror, that is, family support, pedagogical, and social work, which have then surfaced in the form of prevention of violent extremism, the PVE discourse, in various countries (Coolsaet, 2011).

There is considerably more research on how the PVE discourse affects pedagogical work by merging terms, concepts, and theories from various security arenas and bringing them to the classroom. Coppock and McGovern (2014) investigated how a discourse on child protection was mixed with the concept of psychological vulnerability, which, according to Coppock and McGovern, stems from the radicalization discourse and comes together within the Prevent program, which is the national PVE program in the United Kingdom (HM Government, 2015a). They showed how Prevent aspires to give guidance to teachers to detect pupils, particularly male students of the Muslim faith, by looking for warning signals that could indicate these pupils are in a “would-be terrorist stage.” The risk signals are provided by various teaching resources and are exemplified in the article using excerpts from one of these resources, which suggests that teachers

. . . May begin with a search for answers to questions about identity, faith and belonging . . . May be driven by the desire for “adventure” and excitement. . . . May be driven by a desire to enhance the selfesteem of the individual and promote their street cred. (Coppock & McGovern, p. 249)

Now this is, as Coppock and McGovern (2014) argued, not a very credible basis for detecting individuals who might join terrorist ranks in the future. But it becomes meaningful because this aspiration is sustained by a much larger discourse, namely the one that constructs childhood as vulnerable. This discourse relies on a shifting focus from the social contextual factors that surround the child to individual psychological weaknesses within the child. Coppock and McGovern claimed that this way of understanding the so-called vulnerable child as not having had a “normal” childhood has led to an overfocus on psychological interventions meant to protect the child. This is done at the expense of recognizing both the social context surrounding the child, which is unlikely to change after a psychosocial intervention, and the child’s right to autonomy. The vulnerable childhood discourse is now at hand when future terrorists are to be found among British pupils.

O’Donnell (2016) followed a phenomenon closely connected to the one of combining perceived vulnerability with radicalization, namely, what O’Donnell refers to as the creation of thought crime within the PVE discourse. She argued that, despite the lack of any evidence that would withstand scholarly criticism, the PVE discourse constructs vulnerable students who are to be monitored in the classroom without any crimes having been committed or planned. She wrote

Despite an extensive literature challenging claims and policies that presuppose linear (and non-linear) relationships between radicalisation, extremism, violent extremism and terrorism . . . Empirical research does not show a link between radicalisation and terrorism, or even between extremism and violence or terrorism . . . (p. 57)

Baker and Ludwig (2018) investigated how disaster management contributes to the securitization of society through the use of disaster and war rhetoric in the policy process. They showed how a new sociocultural post-9/11 discourse emerged as a response to the sense of a life-threatening condition. What seems rational in the immediate responses to terror attacks alters the overall sociocultural political system in the long run when war rhetoric is deployed in new arenas of public life:

The increasingly securitized approach to disaster management that employs war-like metaphors might not seem troubling on a superficial level, but there are grave consequences to such allusions in real world situations. (p. 3)

What we see in previous research are new expressions of how pedagogical work is articulated when terms, ideologies, and traditions from various arenas merge in the fight against terrorism in the classroom. Concretely, this seems to require viewing the classroom as an arena for prevention of violent extremism and terrorism and to be a way of articulating, in acceptable terms, the child as simultaneously vulnerable, and thereby entitled to protection, and a threat, and thereby a legitimate target of various interventions.

Method

The central theme of the study is to analyze the construction and content of the Swedish local PVE action plans as well as how these plans may affect the educational system and the role of teacher. As we have learned, the plans studied here were all produced under a very limited time span, so it is important to understand the relation between the local needs identified in these municipalities and the national need to encourage the creation of local action plans. According to Ball (2008), policing processes can be understood as top-down or bottom-up, but cannot in either case be reduced to simple linear processes. The process is also intermingled with the participation of policing networks, interpretations at various levels, and various forms of resistance. For this reason, the policy texts, in the local action plans, cannot be taken at face value in the sense that they represent needs and measures that are well adapted to local circumstances. To decode the policy structure of these local action plans, CDA was used as the theoretical framework and concrete methodology.

According to Fairclough (1992), CDA depicts three layers: the text, the discursive practice, and the social practice. The text, both the verbal and nonverbal, is the coded result, or action, of the discursive practice in which texts are produced and consumed. CDA understands, from a critical realism perspective, a discourse as simultaneously constitutionalizing and constitutionalized. The discourse is preceded by a constitutionalizing social practice, but is then also constitutionalizing for the social practice that follows. The production and consumption of texts are part of a continual process that is always interacting with the social practice: the two cannot be fully separated from each other. The social practice provides limits for what texts can be produced and how they can be consumed because it is a predisposed constitutionalized condition. This condition is then understood as a particular discursive order. But there is always the possibility for creativity within a discursive order to bring parts from different discursive orders together within the discursive practice. This is then understood as recontextualization, which may be able to bring about a change within the social practice as well, that is, to be constitutionalizing for a new social practice (Fairclough, 2010). To investigate the presence of recontextualization, it is vital to look for intertextuality, that is, to discover how different texts are related to each other. Manifest intertextuality that merges into a particular discursive order from another discursive order is a clear sign of recontextualization. This process also demands that the analyses strive to detect hegemonic discursive orders, that is, the influential discursive orders that drive recontextualization. The central analytic terms that were deployed are manifest textual content, hegemonic discursive orders, and recontextualization.

In concrete terms, first the focus turned on the textual content of the Swedish local action plans in an attempt to detect their manifest content. This was done by searching for the objects and subjects in the local plans as well as the intentions behind the local plans. In other words, the attention was brought to search for how the municipalities understand their role and duty to PVE and how this is relevant to the traditional tasks of a municipality.

After establishing the manifest textual content, the analyses focus on finding the discursive practices and hegemonic discursive orders involved in policy production and transference. This is done by looking at how particular teachers and their position in preventing violent extremism are talked about and understood in these local action plans. Thereafter, the focus turned to how the plans legitimize various forms of pedagogical interventions.

And finally, the analyses have enabled to point out how discursive practices may bring about recontextualization and constitute new preconditions for social practices among teachers. In more concrete terms, in total, 127 plans from 134 municipalities that were adopted between the April 2015 and April 2017 were analyzed. The average plan was nine-page long. Each action plan was initially read once and all objects and subjects as well as objectives were marked and categorized. After this process, it was clear what sort of categories that were dominate as well as which were deviant. This turned the focus to the discursive practices in the dominant categories. By rereading all the plans with dominant categories again, but now focusing on how they articulate the reasons behind addressing objects, subjects, and objectives, the discursive practices came clear. At this stage, it was possible to point at hegemonic discursive orders that had been in operation in the development of these plans and also thereby concluding the data processing.

Presenting the Results

The result will be presented by addressing each of the RQ under each headline. It will thereby follow the methodological structure described above in such way that the first RQ will provide answers about the manifest textual content, the second RQ will provide answers about the discursive practices, and the third RQ will provide answers regarding recontextualization.

Analyses

What we have seen in these local action plans is the predominance of a discursive order that stems from the security policy. The main interest is in the earliest possible detection of at-risk individuals, seemingly even before these individuals consider themselves to be members of violent extremist organizations and surely before any laws have been broken. This is in good accordance with what Coppock and McGovern (2014) and O’Donnell (2016) have shown in their research on the tendency of PVE discourses to connect vulnerable youngsters with risk signals and potential future risk before any laws have been violated.

To achieve the goal of identifying vulnerable youngsters, I find that all plans, with the exception of 16, talk about the importance of having in-service employee training, particularly for teachers, on the topic of violent extremism. At the same time, we cannot find more than a few cases of in-service training on how to sustain or foster democracy. This is of importance in understanding the recontextualization in these plans. In 111 out of the 127 action plans under study, there are clear references to the NC and/or the Conversation Compass, but we only find a few references to preexisting policies regulating traditional schooling and social work, that is, interventions to promote the interest and well-being of the individual. This could also be understood in relation to Baker and Ludwig’s (2018) research on how disaster management contributes to the securitization of society. The fear of terrorist attacks, as understandable as it is, paves the way for making far-reaching connections between terrorists and presumed signs among youngsters—connections that may predict criminal behavior long before it even appears in the consciousness of the suspected future perpetrators. What it does is to make perpetrators out of teenagers on the basis of policy documents that are not rooted in local preconditions or needs, but instead in a discursive practice that is relevant to the war-on-terror discourse—despite the fact that there is no scientific evidence of any kind to support the notion that it is possible to profile future terrorists (Rae, 2012). We may also say that the vulnerable individuals, and the teachers who are supposed to detect them and report them to the police, who appear in these policy documents, are all seen as disconnected from the surrounding society and no efforts are made to try to discover the reasons for their claimed behavior. They are, according to Coppock and McGovern (2014), said to be vulnerable to ideological messages and seemingly without any agency of their own to act in one way or another. This is something that resonates in the vast majority of the plans in the present study. The solution for this claimed vulnerability, which is put forward in the plans, is that teachers should keep an eye on easily persuaded youngsters. In other words, these young individuals are portrayed as mere objects influenced by seductive messages; the saviors are teachers, social workers, the NGOs, and religious leaders who must co-operate and if need be report the youngsters to the security police. It could be argued that the security-driven discursive order, which dominates these action plans, makes recontextualization possible by bringing together schooling and social work with the police authorities under the banner of preventing future terror and violent extremism. This then paves the way for some of the municipalities to take the discourse one step further by imposing extrajudicial measures that are then spread horizontally.

To sum up and using Fairclough’s terminologies, we can see that, in the majority of these plans, the manifest textual content is rooted in the security-driven discourse, with very few exceptions. Terms, subjects, and objects are seen through the lens of the security of society and not through that of the needs of pupils. Pedagogical work and the duty of the teachers are decontextualized within these plans by coding the understanding of violent extremism, risk individuals, and relevant preventive measures as primarily promoting the safety of society at the expense of pupils’ constitutional rights. If such had not been the case, we might have expected concrete discussions about how to promote democracy, efforts to prevent racism, and discussions on human rights, including the lived experiences of pupils and so on. The concrete information now found in these local PVE action plans consists of various suggestions on what to report to the security police.

We see a recontextualization similar to the one pointed out by Coppock and McGovern (2014) and O’Donnell (2016), that is, combining concepts about vulnerability with preventing recruitment to terrorist organizations. On the contrary, that recontextualization is, as they have shown, already part of the PVE discourse and is now being played out in Swedish municipalities among teachers and social workers. This becomes particularly clear when the plans are compared on a horizontal level, revealing what the municipalities are transferring between each other. From the NC, thus vertically, they incorporate ideas about risk signs, co-operation with various actors, definitions of violent extremism, and so on. But horizontally, we see that the municipalities borrow more concrete measures from each other, such as how to co-operate with the police, including reporting events or individuals who have not committed crimes according to Swedish law, but who display what is understood as risk behavior within the PVE discourse.

Discussion

The present study has contributed to our understanding of the development of the Swedish PVE discourse and how this is related to an international PVE discourse that has trickled down from the national Swedish level to the municipalities. We could also see that there is a horizontal transference of measures between municipalities. The manifest textual content is closely related to a security-driven discourse, and a discursive practice relevant to the war on terror is played out in the vast majority of municipal action plans. It is quite clear that these plans contribute to the securitization of society and that they seldom try to connect suggested actions to identified needs in the local municipalities themselves—even less so to the needs of supposedly radicalized youngsters.

It is far too early to predict anything about the future development of the Swedish PVE discourse or to talk about how these plans may affect the social practices of teachers and social workers. This would require other kinds of empirical data than the policy documents presented here. But it is reasonable to assume that forceful incorporation of the discursive practice relevant to the emerging discourse of these plans into pedagogical and social work may very well bring about a shift in the social practices of these professional arenas. According to Fairclough (2010), social practice constitutes discursive practice. This could very well be understood as one reason why we talk about claimed potential future terrorists as vulnerable individuals when we are addressing the duties of teachers and social workers in the war on terror. After all, it is important that these professionals safeguard vulnerable children. But changes in discursive practices also constitute future social practices. We should ask what might happen if teachers and social workers were to start identifying youngsters as vulnerable to radicalization and reporting them to the security police, without any criminal allegations, the goal being to prevent future terror attacks. By tradition, teachers are instructed to safeguard their pupils’ human rights. It is not difficult to see how this mission and reporting noncriminal events to the police might conflict. Among the basic human rights are freedom of speech and freedom of assembly, which might very well be compromised if teachers comply to the letter with what is written in some of these plans and reported here. It is not the task of the present article to evaluate whether this shift in social practice is happening or, if it is, what the result might be. But asking municipal staff, particularly teachers, to look for risk signals and report them to the police—rather than pointing out how teachers can teach to uphold democracy—is tantamount to asking them to control how their pupils think rather than asking them to develop pupils’ ability to think and reason about the world they live in. It could also very well be argued that even without any practical measures taken by teachers and social workers, the language in these plans is enough to undermine the trustful relations that are the very basis of pedagogical work.