Localising ‘radicalisation’: Risk assessment practices in Greece and the United Kingdom

This article juxtaposes anti-radicalisation policy in the United Kingdom, one of the pioneers in the field, with Greece, one of the latecomers. Drawing on localisation theory, our aim is to understand how ‘common knowledge’ of radicalisation and counter-radicalisation has materialised in the United Kingdom and Greece by exploring the development and use of radicalisation-related risk and vulnerability assessment tools. We argue that the radicalisation ‘knowledge’ was localised more seamlessly in the United Kingdom, which can be attributed to the country’s ‘norm producer’ status on the field of European counter-radicalisation. By contrast, the ‘knowledge’ was subjected to significant ‘re-framing’ and ‘stretching’ to fit with the Greek context. This is associated with the country’s ‘norm adopter’ status on the field of European counter-radicalisation, as well as with a ‘spill-over effect’ from a national context of deeply polarising and contentious counter-terrorism policies. We maintain that these localisation processes reveal two distinct assemblages of governing radicalisation.


Introduction
Several critical approaches to counter-terrorism have focused on the genealogy of the muchdiscussed and much-disputed concept of 'radicalisation'. Besides problematising the diffuse meaning and uncertain analytical value of the concept (Kundnani, 2012;Neumann, 2013), these have enhanced our understanding of terrorism-and radicalisation-related policies as a mode of government (De Goede and Simon, 2013;Ragazzi, 2016); how such policies construct social and political actors as terrorist threats (Heath-Kelly, 2013;Heath-Kelly et al., 2014;Jackson et al., 2009;Kundnani, 2014;Pantazis and Pemberton, 2009); the relations between counter-terrorism, counter-radicalisation and social policy (Mythen et al., 2017;Ragazzi, 2017); and various discriminatory effects of the logics of suspicion and anticipation involved in counter-terrorism and counter-radicalisation policies (Amoore and De Goede, 2008;Aradau et al., 2008;Aradau and Van Munster, 2009;Mythen and Walklate, 2008). More recently, scholars have engaged with the development of indicators-based approaches to counter-terrorism and counter-radicalisation that aim to identify and act upon individuals who might resort to politically and religiously motivated violence (e.g. Augestad Knudsen, 2020;Monaghan and Molnar, 2016).
So far, the majority of these works has been single-case studies, most of which have focused on the United Kingdom (e.g. Bigo and Guittet, 2011;Heath-Kelly, 2013;Mythen et al., 2017;Pantazis and Pemberton, 2009). Within that context, studies have explored the ways in which counter-terrorism and counter-radicalisation policies are developed and justified (Bentley, 2018;Edwards, 2016;Heath-Kelly, 2013;Mythen et al., 2017). Others have analysed the broader repercussions of such policies for Muslim communities and other 'suspect populations' (Heath-Kelly, 2012;Kundnani, 2015;Pantazis and Pemberton, 2009;Thomas, 2010). Possible implications of the involvement of educational institutions -such as schools and universities - (Chadderton, 2012;Cram, 2012) as well as the National Health Service (NHS) in such policies have been explored too (Heath-Kelly, 2017;Summerfield, 2016). Recent studies have been attentive to the implications and dilemmas of developing and using indicators-based approaches to counterterrorism and counter-radicalisation (Augestad Knudsen, 2020). While a few cross-country comparisons exist (Haverig, 2013;Poynting and Perry, 2007;Spalek and Alia, 2007), these have tended to analyse countries at the 'forefront' of counter-terrorism and counter-radicalisation policies. 'Late adopters' of such policies are considerably underrepresented, leaving unexplained how countries of this kind have chosen to follow the trendsetting countries, as well as which parts of the forerunners' radicalisation 'knowledge' they have implemented in policy, and how. Attempting to correct this imbalance, this article employs the diverse case-selection strategy, which has as its main objective the achievement of maximum variance along relevant dimensions (see Seawright and Gerring, 2008). In this regard, we juxtapose anti-radicalisation policy in the United Kingdom, a typical/representative case of the pioneers in the field, with Greece, a typical/representative case of the latecomers, and a country on which no critical radicalisation research has been conducted to date. Indeed, the diverse case-selection strategy is particularly useful when the investigation is exploratory. Moreover, it demonstrates stronger claims to representativeness than any other small-N case study, and the additional advantage of introducing variation (i.e. 'pioneers' vs 'late adopters') on the key variable of interest (i.e. the adoption of radicalisation-related risk and vulnerability assessment tools) (see Seawright and Gerring, 2008).
Drawing on the research critically engaged with indicators-based approaches to counter-radicalisation, we also place our article in the literature framing counter-radicalisation policies as a mode of government. Following De Goede and Simon (2013: 316) in particular, we conceptualise radicalisation-related risk assessment tools as part of an 'assemblage' of governing radicalisation; an assemblage comprised of 'particular threat representations, knowledge practices, training programmes and strategies for intervention'. Like they do, we acknowledge that 'countering radicalisation does not necessarily materialise in the same ways across different localities' (De Goede and Simon, 2013: 327); indeed, the focus of this article is on processes of localising (counter-)radicalisation in the two locales of the United Kingdom and Greece.
Specifically, the article explores how 'common knowledge' of radicalisation and counter-radicalisation has materialised in the United Kingdom and Greece -countries both drastically different, and equally familiar with terrorism -by looking at the development and use of radicalisation-related risk and vulnerability assessment tools: the Extremism Risk Guidance (ERG22+), the Extremism Risk Screen (ERS), and the Vulnerability Assessment Framework (VAF) in the United Kingdom, and the 'Observable Indicators Manual' (OIM) in Greece. The centrality of the tools in UK counter-terrorism can be illustrated by the fact that the ERG22+ and ERS are used to assess incarcerated offenders of concern, and the VAF is used to assess non-incarcerated individuals vulnerable to possibly becoming 'radicalised'. In the case of Greece, OIM constitutes the only radicalisation-related risk and vulnerability assessment tool available, making it a valuable source in the study of how 'common knowledge' has materialised in the country. Our main theoretical framework is Acharya's theory of localisation, involving 'the active construction [. . .] of foreign ideas by local actors, which results in the former developing significant congruence with local beliefs and practices' (Acharya, 2004: 245). Localisation theory has already proven useful to critical approaches to security (see Capie, 2008;Stritzel, 2011aStritzel, , 2011bStritzel, , 2014, but has so far not been applied to radicalisation.
We argue that radicalisation 'knowledge' has been localised in both the United Kingdom and Greece, with a more seamless congruence-building process in the United Kingdom -an outcome directly related to the country's 'norm producer' status on the field of European counter-radicalisation. By contrast, the 'knowledge' underwent considerable 're-framing' and 'stretching' in the Greek context. This is directly associated with the country's 'norm adopter' status on the field of European counter-radicalisation, as well as with a 'spill-over effect' from a national context of deeply polarising and contentious counter-terrorism policies. This, we argue, reveals two distinct assemblages of governing radicalisation: a 'complex' one in the United Kingdom and a 'rudimentary' one in Greece.
After briefly outlining localisation theory and its empirical application to critical approaches to security, we distil the Europe-wide 'common knowledge' of radicalisation that was localised in the United Knowledge and Greece. We proceed with analysing how these basic 'truths' have been localised in each case study, before turning to the assemblages of governing radicalisation these localisation processes resulted in. Finally, we conclude by reflecting on the driving forces, actors, and catalysts behind these localisation processes.

Localisation theory
Localisation theory explains how transnational norms or ideas/beliefs are incorporated into new contexts. As initially conceptualised by Acharya, localisation involves 'the active construction (through discourse, framing, grafting, and cultural selection) of foreign ideas by local actors, which results in the former developing significant congruence with local beliefs and practices' (Acharya, 2004: 245). This framework has been used within Security Studies to understand the diffusion in national and regional contexts of small arms norms (Capie, 2008), as well as interpretations of organised crime and rogue states (Stritzel, 2011a(Stritzel, , 2011b(Stritzel, , 2014. Localisation may start with a reinterpretation and re-representation of the 'foreign' norm/idea/belief through framing and grafting, but may entail more complex processes of reconstitution to make the 'foreign' congruent with a pre-existing local normative order or belief structure (Acharya, 2004: 244). What distinguish localisation from related frameworks are its emphasis on a dynamic congruence-building process -rather than dichotomous acceptance/rejection -and its insistence on the agency of local actors -rather than portraying them as passive recipients and learners of norms or ideas/beliefs. Localisation theory focuses on the role of local actors as 'insider proponents'. These can be individuals, epistemic communities, or non-governmental organisations (NGOs), usually physically present within the new locale, with reasonably direct access to policymakers (Acharya, 2004: 248). They can be part of the new locale's government, broader policymaking elite, or civil society. Regarding the localisation of threat interpretations, Stritzel (2011a) has highlighted the role of local 'security professionals', broadly understood as anyone with 'security expertise', from police or intelligence agents to politicians, journalists, and academics (also Bigo, 2002). With reference to such actors, Stritzel (2014) has usefully analysed how collective interpretations of threats can change with their incorporation into new locales, and how such localisation processes are key to understanding how global narratives on the 'nature of threats' evolve (Stritzel, 2014). Localisation theory acknowledges that local actors may localise a 'foreign' norm/idea/ belief on their own initiative if they consider their existing beliefs and approaches as inadequate and in need of broadening and strengthening (Acharya, 2004: 247). Another motivation may be to enhance their own legitimacy, status, authority and/or resources (Acharya, 2004: 245-246), a desire sometimes 'catalysed' by a major security or economic crisis, systemic change, domestic political alterations, or the regional/international 'presence' of an issue.
The actual process of localisation of a norm/idea/belief usually includes local actors employing various strategies of matching and/or addressing factual or political incompatibilities to enhance compatibility between content and context. These strategies include (re)framing, selection, grafting, and linking (see Acharya, 2004;Barnett, 1999;Stritzel, 2011aStritzel, , 2014, as well as 'strategies of stretching' and even altering of the new norm/idea/belief (see Farrell, 2005;Stritzel, 2011aStritzel, , 2014. The overarching goal of these strategies is to make a pre-existing but 'foreign' norm/idea/belief seem local and appealing to local audiences. The process of localisation is deemed successful when an existing institution responds to a 'foreign' idea by expanding its functional scope, or by creating new policy instruments to pursue new goals, without replacing its pre-existing ones (Acharya, 2004: 253).

Radicalisation 'truths'
The United Kingdom's and Greece's implementation of radicalisation-related risk assessment tools represent just two of many manifestations of how Europe-wide 'common knowledge' about radicalisation has become localised nationally. It also shows how radicalisation-related risk assessment tools have come to constitute one aspect of an assemblage of governing radicalisation (De Goede and Simon, 2013: 316). Before turning to how the processes of localisation unfolded, however, the key 'truths' (see Bigo, 2006) of the 'common knowledge' about radicalisation need to be discerned: what, exactly, was localised?
The mid-2000s saw the establishment within European counter-terrorism of certain assumptions about the nature of radicalisation and how it could explain terrorist attacks. Ideas about radicalisation's explanatory force emerged in proper after the 9/11 attacks but became consolidated among European policy makers and security practitioners with the later attacks in The Netherlands, Madrid, and London. Before this, policy makers and practitioners in Europe had tended to view terrorism as an international threat caused by international groups and developments. But in the absence of obvious links between all of these attacks and a central Al Qaeda command, practitioners increasingly identified the terrorist threat with 'home-grown' as well as 'self-radicalising' persons (Europol, 2010;Intelligence and Security Committee, 2006). Radicalisation came to be seen as the process through which an 'ordinary citizen' becomes a terrorist by gradually accepting and endorsing the use of violence (see Kundnani, 2015). European policy documents started to define radicalisation along these lines; a process involving the increasing acceptance of the use of violence to achieve political, ideological, and/or religious aims (Council of the EU, 2010: 6;Schmid, 2013: 12).
Throughout Europe, a basic 'truth' about radicalisation emerged: that radicalisation is a risk that can lead to terrorism. Consequently, 'countering radicalisation' became a measure of counter-terrorism (see also De Goede and Simon, 2013: 319), based on the assumption of a second 'truth': that terrorism can be stopped by rerouting or reversing radicalisation at an early stage (see also De Goede and Simon, 2013: 321). The understanding of radicalisation as a process or continuum made intervening upon it possible; it also enabled the possibility of anticipating threats arising from radicalisation (De Goede and Simon, 2013;Heath-Kelly, 2013;Kundnani, 2014;Martin, 2014). Arising from these assumptions was thus a third 'truth', grounded in the perceived need to detect and measure the warning signs of radicalisation in order to stop terrorist violence: that radicalisation is measurable. From around 2006 onwards, the logic of indicators-based risk assessment was then transferred to the domain of European counter-terrorism and counter-radicalisation (see also Monaghan and Molnar, 2016: 394).
These highly compressed three 'truths' comprised the backbone of the 'knowledge' about radicalisation upon which the assemblage of governing its risk became based in Europe. Together capturing the essence of this assemblage, they render radicalisation governable through the mobilisation of a logic of 'precautionary risk' and enable the deployment of a 'calculus of risk' (De Goede and Simon, 2013: 328) which emphasises the presence of 'catastrophic and radically contingent elements' (Aradau and Van Munster, 2007: 103; see also Bigo, 2006;Mythen et al., 2017). The empirical application of these logics of suspicion, anticipation, and prediction to the field of counter-terrorism and counter-radicalisation (Amoore and De Goede, 2008;Aradau et al., 2008;Aradau and Van Munster, 2009;Mythen and Walklate, 2008) can best be exemplified with the development and use of terrorism-and radicalisation-related risk assessment tools.

Context and early localisation
In the United Kingdom, the development and use of radicalisation-related risk assessment tools represented the localisation of 'truths' that were not altogether foreign. Indeed, the United Kingdom could be described as a 'norm producer' on the field of European counter-radicalisation policies (Directorate-General for Internal Policies, 2014: 13). This is not only due to its long experience with domestic counter-terrorism -until the late 20th century, largely centred on the threat from (Irish Republican Army) IRA-affiliated groups -but also in part because of the United Kingdom's (at least pre-Brexit) significant status within the European security environment overall. It was under the UK presidency during the second half of 2005 and shortly after the London bombings that the very concept of radicalisation went mainstream and became institutionally significant within the EU (Coolsaet, 2016: 17). Against this background, the process of localising the radicalisation 'truths' in the United Kingdom represented both their materialisation in this specific locale and the strengthening of the perceived policy legitimacy of this 'knowledge' Europe-wide.
The clearest institutional and policy materialisation of the first radicalisation 'truth' in the United Kingdom happened in 2006, when the Labour Government presented 'Prevent' as a central strand of the United Kingdom's counter-terrorism strategy CONTEST (HM Government, 2006; CONTEST was also revised in 2009 and 2011). 1 The year before, ideas of radicalisation as a cause of terrorism had gained ground in the United Kingdom when it was discovered that three of the four Al Qaeda affiliated suicide bombers in London had been British citizens residing in Leeds. Counter-terrorist attention then turned to the role of radicalisation in seemingly turning locals into terrorists (see Bakker, 2015;Pantucci, 2015). Apart from its Prevent strand, CONTEST in 2006 consisted of -as it still does -the 'work streams' Pursue, Protect, and Prepare (HM Government, 2011a; also Brown, 2009;Gregory, 2009;Home Office, 2017). Prevent, as the name suggests, seeks to 'stop people becoming terrorists or supporting terrorism' (HM Government, 2011a) and is concerned with 'tackling the radicalisation of individuals' (HM Government, 2006: 1). While the 2006 strategy did not define radicalisation, it referred to individuals that 'become radicalised, to the extent of turning to violence to resolve perceived grievances', and to 'radicalisation which can lead to terrorism' (HM Government, 2006: 9, 5). Acknowledging that 'only a tiny minority of radicalised individuals actually cross over to become terrorists', it also described radicalisation as a 'risk' (HM Government, 2006: 10, 6). The 2006 CONTEST strategy hence localised the idea that radicalisation exists as a terrorism-related risk.
In 2011, a new version of the Prevent strategy entrenched this 'truth' by defining radicalisation as 'the process by which a person comes to support terrorism and forms of extremism leading to terrorism' (HM Government, 2011b: 108), and by increasing the centrality of Prevent within CONTEST as a way to 'address radicalisation' (HM Government, 2011a: 12). Whereas earlier versions of CONTEST had distinguished between international and domestic terrorism, the 2011 version abandoned this distinction (e.g. HM Government, 2011a: 17) and implicitly posited individual radicalisation as the link between international terrorism and domestic risk. In 2015, Prevent was turned into a statutory duty obliging a range of authorities such as the NHS, universities, and schools to refer their concerns over the possible radicalisation of specific individuals (HM Government, 2015a, 2015b). 2 The second radicalisation 'truth' was institutionally localised in the United Kingdom in 2012, when the Channel process became part of Prevent and central to UK counter-terrorism (after having been piloted in 2007). Defined as a 'police-coordinated, multi-agency partnership that evaluates referrals of individuals at risk of being drawn into terrorism' (HM Government, 2011b: 57), Channel became the operational mechanism by which those referred to Prevent over radicalisation concerns were assessed and directed (or not) towards interventions intended to stop or reverse this process (Coppock and McGovern, 2014;HM Government, 2012: 2). In practice, Channel involves multi-agency panels chaired by a local authority conducting assessments on 'how best to safeguard (the individuals referred) and support their vulnerability (to radicalisation) through a support package tailored to individual needs' (Association of Chief Police Officers, 2012). Channel is meant for any age group, but those referred are generally young; 57% of those referred in 2016-2017 were aged 20 or under (Home Office, 2018: 4). Around 61% of Prevent referrals during the same period concerned Islamist extremism, while 16% regarded right-wing extremism; of those receiving Channel follow-up and interventions, respective percentages were 55% and 37% (Home Office, 2018: 4). The Channel process is voluntary; a person will only become part of the programme if having agreed to take part.
Clearly, Channel was founded on the idea that radicalisation can be stopped and/or reversed. Like the first 'truth', then, the second one was institutionally localised into UK counter-terrorism policy directly and 'wholesale', without any stretching or amendments. The United Kingdom's smooth institutional and policy localisation of these first two radicalisation 'truths' set the framework for localising the third 'truth', of radicalisation being measurable, through the insertion of radicalisation-related risk and vulnerability assessment tools at different parts of the UK counter-terrorism apparatus.

Measuring and countering radicalisation
The most direct prompt for the development of radicalisation-related risk assessment tools in the United Kingdom was the country's 2006 update of its terrorism legislation. 3 That year, the UK terrorism act (TACT) listed as terrorist offences several crimes that did not entail direct violence, but instead involved, for instance, 'encouragement of terrorism', 'dissemination of terrorist publications' (deliberately or recklessly), 'preparation of terrorist acts', 'training for terrorism', 'attendance at a place used for terrorist training', and the 'making and possession of devices of materials'. 4 Consequently, most individuals sentenced for terrorism in the United Kingdom after 2006 had not themselves committed acts of terrorist violence. This led offender risk assessors in United Kingdom prisons to call for a new tool that could measure the risk of these offenders. The risk now at stake was not primarily that offenders could act violently towards others, but rather -to simplify only slightly using the terminology of the time -that they had been 'radicalised'. The language of 'radicalisation', however, was absent from the ERG22+ tool launched in 2011, with its authors later acknowledging the term's imprecise and dispute-ridden nature (Lloyd and Dean, 2015).
Based on 22 indicators categorised by 'engagement', 'intent', and 'capability', and allowing those assessing to add factors they deemed relevant, the ERG22+ was originally intended for use on TACT offenders in England and Wales' prison system. 5 An ERG22+ was, and still is, first carried out upon a prisoner's entry into the prison, to get an idea of the individual risk profile and the role each indicator played in the original offence. Subsequently, ERG22+ assessments are carried out at regular intervals and added to the individual's file, allowing for comparisons and measuring of development. ERG22+ assessments influence decisions regarding a person's security categorisation, release, and referral to a Healthy Identities Intervention programme (Dean, 2014). Although the ERG22+ was developed based on work with British, Al Qaeda-influenced TACT offenders at the time of the wars in Afghanistan and Iraq, the tool is being used also on 'extreme right-wing offenders, animal rights activists, female extremists, and gang members who's offending is based on joint enterprise' (Lloyd and Dean, 2015: 50).
Although the language of radicalisation, as mentioned, is absent from the ERG22+, two of its 'spin-off tools' show its organisation and indicators' close association with the radicalisation 'truths' and their localisation in the United Kingdom context. The first of these tools is simply a shorter version of the ERG22+, called the ERS. The ERS is used in United Kingdom prisons to assess violent and non-violent non-TACT offenders in cases where someone has concerns over their 'possible involvement or interest in extremist groups, causes, or ideas' (Lloyd and Dean, 2015: 40). Given the proximity of this language with the official UK understanding of radicalisation, another way of putting this is that the ERS is used to screen the possible radicalisation of prisoners not (yet) sentenced for terrorist offences.
The second assessment tool developed based on the ERG22+ is explicitly aimed at assessing individuals where there is concern over their possible radicalisation. In 2012, the VAF became part of the Channel programme in England and Wales (HM Government, 2012: 2-3). It contains the same 22 indicators as the ERG22+ and identically groups these into the categories of 'engagement', 'intent', and 'capability'. A VAF assessment is meant to provide guidance for appropriate follow-up and interventions if an assessor records findings under several, or key, indicators which suggest that an individual is 'vulnerable to' or 'at risk from radicalisation' (HM Government, 2012: 2). Unlike the ERG22+ and the ERS, the VAF is used on individuals who may well be non-offenders, and overwhelmingly on persons below the age of 20.
Neither of the ERG22+, ERS, or VAF tools is 'actuarial'; they do not produce a number which alone is meant to reflect an individual's risk. Nonetheless, they do record indicators in a manner offering practitioners a sense that they are in fact measuring radicalisation -within a single assessment, and by allowing for consecutive assessments to be measured against each other. This is particularly the case with the VAF as used in Channel: while the ERG22+ is primarily being carried out by forensic psychologists trained in the advantages and pitfalls of score-based risk assessment, the VAF is being conducted by professional groups without such training or expertise. The ERG22+ has recently been reviewed by the United Kingdom's Prison and Probation authorities for continued use, while two additional assessment tools have recently been developed by UK police to supplement the use of the VAF in cases where radicalisation or violent extremism might be at issue.
Enabled and conditioned by the United Kingdom's post-7/7 2005 policy, institutional and legal localisation of the first two radicalisation 'truths', the 2011-2012 embrace of the ERG22+, ERS, and VAF tools hence showed the United Kingdom's localisation of the third 'truth': that radicalisation is measurable. In fact, the United Kingdom's uses of the ERS and the VAF in particular -with their respective implicit and explicit targeting of radicalisation -appear as manifestations of the United Kingdom's localisation of all three 'truths'. The United Kingdom's development and implementation of the three tools seem to have happened seamlessly, with broad involvement from various parts of the United Kingdom's counter-terrorism apparatus, including the police, prison service, and local authorities. Criticisms of the tools from activists and academia only surfaced much later and have so far seemed to have limited impact (e.g. Qureshi, 2016

Pre-localisation context
Greek legislation does not include any provisions pertaining to radicalisation, but does provide for the punishment of terrorism. Greece has a rich history of politically motivated -far-left/far-right -violence and terrorism (see Anagnostou and Skleparis, 2015;Kassimeris, 2013). International terrorism that made its appearance in the country at the end of the 1960s also bore the imprint of politically rather than religiously motivated violence (see Kaminaris, 1999). Significantly, issues of political violence and terrorism have always been highly contentious in the country, and the localisation of counterterrorism policies has faced strong resistance (see Karyotis, 2007). This has slowed down and partly hindered Greece's localisation of the Europe-wide 'knowledge' about radicalisation.
Counter-terrorism policies became highly politicised already in the aftermath of the civil war (1949) in Greece. An elaborate surveillance and security apparatus was established and directed against the 'internal enemy', namely communists and those suspected of leftist leanings. Further solidified during the junta years (1967)(1968)(1969)(1970)(1971)(1972)(1973)(1974), this apparatus deepened the chasm between conservative patriots and the rest -communists, leftists, and their sympathisers (see Samatas, 2014), and ensured that counterterrorism remained a source of friction in the post-junta period between the two main political parties. Indeed, until the early 2000s, Greek political elites failed to agree on the significance of terrorism and the threat that it posed (Karyotis, 2007). A new law adopted in 2001 -despite fierce opposition by the minority parties of the left -constituted the first serious state attempt to tighten the relevant legislation, and marked the securitisation of terrorism by acknowledging it as a cardinal security threat to the Greek state and society (Karyotis, 2007: 276). Building on this as well as the momentum provided by the preparations for the 2004 Athens Olympic Games (the first post-9/11 games), Greece introduced unprecedented counter-terrorism measures, later described as the largest security operation in peacetime Europe (Samatas, 2007). However, the end of the Olympic Games saw Greece drastically reducing its counter-terrorism infrastructure (Borgeas, 2013).
At a time when terrorism concerns were peaking and 'knowledge' about radicalisation was increasingly localised in Europe, as manifested by public funding of counter-and de-radicalisation programmes, terrorism was relegated to an issue of secondary importance in Greece. Greek counter-terrorism policies had always been event-driven and polarising, and with the disbandment of the 'first generation of terrorists' in early 2000s and the conclusion of the Olympics, with threat anticipation at a low ebb, terrorism was no longer on top of the agenda. Even in the mid-2000s, 'radicalisation' was a rarely heard term in policy circles in Greece.
This changed with the December 2008 events, which paved the way for the localisation of the 'common knowledge' about radicalisation. The killing of a teenager by a police officer sparked demonstrations in Athens, which quickly spread across Greece. They lasted a couple of days and were accompanied by riots, acts of looting, and widespread violence. Thousands of young people -from affluent backgrounds, from anarchist and far-leftist movements, from poor and marginalised districts of Athens, hooligans, as well as destitute immigrants (see Economides and Monastiriotis, 2009) -participated in extensive acts of violence, raiding stores, and setting banks on fire. The events deeply affected the militant far-left scene, as they marked a new wave of guerrilla-style terrorist attacks directed against the police, businesses, political offices and personnel, and other institutions. Arguably, the events gave rise to Greece's 'second generation of terrorists' (see Kassimeris, 2011. If the focus of counter-terrorism in pre-2008 era was on mesoand macro-level factors that likely cause terrorism, the December 2008 events called attention to micro-level determinants that can drive young people from diverse backgrounds to upgrade their Molotov cocktails and stones to improvised explosive devices and Kalashnikov rifles.

Measuring and countering radicalisation
Within this context, the Centre for Security Studies (CSS) 6 became increasingly active in the field of counter-radicalisation. In 2012, it established a research team specifically dedicated to the study of radicalisation, terrorism, and organised crime, while in 2015, it started cooperating closely with the EU's Radicalisation Awareness Network (RAN) in the implementation of 'counter-radicalisation' activities. In September 2016, the CSS and the Directorate of State Security of the Ministry of Public Order and Citizen Protection organised a series of training seminars and 'train the trainer' workshops in five Greek cities for the purposes of the 'Counter-Radicalisation and Extremism' project. This constituted the first official attempt to localise 'knowledge' of the concept of radicalisation in the country. The workshops aimed to raise awareness of radicalisation, train frontline professionals in the use of radicalisation-related risk assessment tools, and create a pool of trainers since 'tackling radicalisation is a sine qua non of the fight against terrorism and violent extremism'. 7 The link between radicalisation and terrorism -the first 'truth' about radicalisation -was hence included in the stated aims of the initiative. A training guide and two pocket manuals were produced -one operational tailored for frontline professionals titled 'Observable Indicators Manual' (OIM), and a radicalisation awareness one targeted to the general public. Both the format of the seminars and workshops, as well as the project's other deliverables drew upon similar initiatives from other EU memberstates (see RAN, 2017).
OIM is a 31-page long classified document, authored by a team of social scientists under the guidance of the CSS and the Directorate of State Security of the Ministry of Public Order and Citizen Protection. It is structured around eight themes: (1) operational use of the manual; (2) outline of the phenomenon of violent radicalisation; (3) context of engagement for frontline professionals; (4) recognition and categorisation of indicators; (5) a list of observable indicators; (6) a list of indicators of preparation of a terrorist action; (7) overview of politically (far-left/anarchist, far-right) and religiously motivated radicalisation; and (8) presentation of symbols, emblems, and logos, as well as tattoo themes related to key extremist and terrorist organisations. The inclusion of three sections of radicalisation indicators (comprising almost half the manual), as well as a section on the 'signs' of key extremist and terrorist organisations manifest the localisation of the second and third 'truths' about radicalisation: that terrorism can be stopped by rerouting or reversing radicalisation, and that radicalisation is measurable. According to its authors, OIM was created in order to 'provide frontline professionals in Greece with a practical and useful "tool" that can be used by them in their daily work'. Intended end users include law enforcement, as well as professionals who work in the healthcare sector and the penitentiary system.
The preface of the manual identifies the target groups of the OIM as 'the thousands of foreign fighters that departed to the battlefields of the Middle East and North Africa', referencing a Europe-wide concern. Yet the authors stress in the next sentence that countering 'ideologically motivated extremism and terrorism (far-left/anarchist, far-right) constitutes the main preoccupation of the Greek security authorities'. The manual defines 'violent radicalisation', as 'a process through which an individual comes to accept the use of violence for the accomplishment of political, ideological or religious aims'.
The OIM visualises the radicalisation process as a six-level pyramid based on Moghaddam's (2005) 'staircase' to the terrorist act and the 'staircase model' of radicalisation used in the Community Policing and the Prevention of Radicalisation (CoPPRa) training programme for frontline practitioners (originally developed in Belgium): at the bottom of the pyramid lies 'perceived unfairness', consecutively followed by 'identification with a radical ideology', 'participation in an organisation/group', 'legitimisation of violence/taking up action', 'violent extremism,' and 'terrorism'. This representation holds the promise of anticipating the terrorist threat by attending to radicalisation, and hence shows the localisation of the 'truth' of radicalisation existing as a precursor to terrorism. Yet, although radicalisation is presented as a linear process, the authors stress that 'radicalisation does not necessarily lead to violent extremism or terrorism'; it 'can be either a slow and gradual process, or an abrupt and instantaneous one'. It usually takes 2-3 years, but can last up to a decade, while the final 'step' from violent extremism to terrorism can be accelerated. According to the OIM, the 'radicalisation process resembles the "snakes and ladders" board game'. This presentation of radicalisation as a continuum opens for the possibility of stopping or reversing it through early intervention, and manifests the localisation of the second 'truth'.
According to its authors, the OIM is intended to be a prognostic tool for use by frontline professionals to prevent the risk of violent extremism and terrorism by 'recording and categorising visible indications of ideologically and religiously motivated radicalisation and violent extremism'. In other words, it draws upon the 'science of indicators' (Monaghan and Molnar, 2016: 397), based on the assumption that radicalisation is measurable, manifesting localisation of the third key 'truth' about radicalisation. The 'observable indicators' in the manual have been largely borrowed from CoPPRa and similar radicalisation-related risk assessment tools (see RAN, 2017). In the words of the manual, 'each separate indication, or even a combination of a few, does not constitute alone full and adequate evidence, but a sign that a specific individual is undergoing a process of radicalisation'. While the authors warn end users that OIM should not be used as a 'checklist of symptoms' determining whether a person has been radicalised or not, they also state that 'the more indications there are, the higher the probability is that a process of radicalisation has started'. The authors stress, however, that 'recognising and identifying indications should not result in stigmatisation, discrimination and xenophobia against certain groups'.
In sum, all three key 'truths' of the Europe-wide 'common knowledge' about radicalisation have been localised with the OIM. The manual establishes a link between radicalisation and terrorism, suggests that timely intervention upon -or 'control' of -radicalisation can break this link, and includes 43 'observable indicators' reflecting the belief that radicalisation is measurable. Yet the 31-page-long document lacks practical guidance on appropriate counter-and/or de-radicalisation interventions, and on how to capture the radicalisation risk apart from counting the number of indications. It is, moreover, notable that Greece's localisation of the 'common knowledge' of radicalisation includes one local 'twist': The authors of the OIM state that 'radical behaviours are not essentially problematic' and call for tolerance towards (certain) 'radical behaviours' as well as 'youth spontaneous unlawful conduct'. According to the OIM, insensitive or disproportionate interventions may be counterproductive, and could exacerbate radicalisation and strengthen its link with terrorism. The manual implies that interventions should be moderate and take place between level five ('violent extremism') and six ('terrorism') of the 'radicalisation pyramid'. It does not provide, however, any direction as to how to distinguish between 'spontaneous' and acceptable radical behaviours and unacceptable ones.
The OIM's call for tolerance towards (certain) 'radical behaviours' should not be taken as a sign of unsuccessful localisation of the radicalisation truths in Greece, but rather as a manifestation of the important role of local actors in localisation processes (see Acharya, 2004;Stritzel, 2011aStritzel, , 2014, and the heterogeneity of the assemblages of governing radicalisation (see De Goede and Simon, 2013). The 'radicalisation knowledge' was localised in Greece at a time when a coalition government between the left-wing SYRIZA ('The Coalition of the Radical Left') and the right-wing ANEL ('Independent Greeks') was in power. The two parties were positioned at opposite sides of the aforementioned contentious and polarising issue of political violence; SYRIZA's approach to political violence was much more lax compared to that of right-wing political parties. Shortly before taking office, SYRIZA was portrayed by other politicians and media as 'flirt(ing) with violence if not fomenting ultra-leftist terrorism' (Stavrakakis and Katsambekis, 2014: 134). This sharply contrasted the right-wing political parties' and security professionals' rigidly rejecting political violence, in effect especially left-leaning political violence. In localising the 'radicalisation knowledge' in Greece, it seems that this polarisation spilled over into the OIM's call for tolerance towards (certain) 'radical behaviours' and 'youth spontaneous unlawful conduct', manifesting both 'truth stretching' and gathering of the disparate elements that make up the local assemblage of governing radicalisation.

A complex and a rudimentary assemblage of governing radicalisation
The previous sections demonstrated how the Europe-wide 'common knowledge' about radicalisation was localised in the UK and Greece through the production and adoption of new radicalisation-related risk assessment tools. Unsurprisingly, the three 'truths' were more congruently localised in the United Kingdom, due to the initially higher compatibility between the 'common knowledge' and the local context, given the United Kingdom's 'truth/norm producer' status. By contrast, in the case of Greece, the three 'truths' went through 'stretching' in attempts to make the content fit the local context and bring together the heterogeneous elements that make up the local assemblage of governing radicalisation. The localisation of the Europe-wide 'knowledge' about radicalisation -distilled in the form of the three 'truths' -in the United Kingdom's and Greece's terrorism-and radicalisation-related risk assessment tools hence reveals, respectively, complex and rudimentary assemblages of governing radicalisation.
To offer some further detail, complex assemblages are heterogeneous dynamic networks made up by the non-hierarchical self-organisation of autonomous diverse elements (see Levkoe and Wakefield, 2014). Each element is connected to others in countless ways, blurring the boundaries between them, and rendering them mere delineations of convenience. Through a complex network structure with multiple and overlapping points of contact, the diverse elements that comprise it converge, but are never perfectly aligned. Indeed, the dynamic character of these complex assemblages is informed by the constant negotiation and renegotiation of the connection points between their elements. The United Kingdom represents a typical case of a complex assemblage of governing radicalisation. A wide range of actors and agents -security professionals, civil servants, civil society organisations, health and social workers, school and university teachers -are drawn together and set up a number of diverse and dynamic governing connection points between them through policy, institutional arrangements, and law. Rudimentary assemblages, on the other hand, are also heterogeneous and dynamic, yet much less non-hierarchical and self-organised, and comprised of far fewer diverse elements. Greece, as a typical case of a rudimentary assemblage, showcases the leading role played by the CSS and the Ministry of Public Order and Citizen Protection in the making of this assemblage by bringing together a diverse, albeit limited, set of security professionals, loosely connected with each other. This structure of rudimentary assemblages allows for tracing the process through which diverse elements come together to form them, and for understanding their heterogeneous and dynamic properties by aggregating their individual parts. Rudimentary assemblages are potentially complex assemblages 'in the making'; studying them can shed light on the inner workings of the latter.
The processes of localising the 'knowledge' about radicalisation in the United Kingdom and Greece were in both cases sparked by a security crisis. In the United Kingdom, the initial inability of security professionals to quickly establish a clear link between the 7/7 London terrorist attacks and central Al Qaeda shifted attention onto the dangers associated with so-called 'home-grown' terrorism in the mid-2000s. In the emerging threat environment, existing models and tools came to be seen as lacking. In Greece, it was the 2008 events in Athens, which marked an unprecedented new wave of youth-driven guerrilla-style terrorist attacks. This threat could not be addressed with existing tools and knowledge. These respective security crises functioned as catalysts for the localisation of the 'knowledge' about radicalisation in each of the two countries.
Local context -including each country's threat representations, knowledge practices, and strategies for intervention -determined the ways in which the three 'truths' about radicalisation were localised. Regarding localisation of the first 'truth' in the United Kingdom, the shadow of the 7/7 terrorist attack, led to an initial focus on radical Islamism (HM Government, 2006), which was later stretched to include the full spectrum of extremist ideologies, as well as a blurring of the distinction between international and domestic terrorism. By contrast, this divide was retained in Greece. Deeply ingrained threat representations based on past and present violence and terrorism ascribed domestic political (i.e. far-left/right) characteristics to the radicalisation risk.
Similarly, local context shaped the ways in which the second and third 'truths' about radicalisation were localised in each country. The 2006 amendments to the United Kingdom's terrorism legislation were introduced at the same time as the Prevent part of the UK counter-terrorism strategy was increasingly prioritised. Radicalisation-related risk assessment tools and indicators became a way to build practitioners' confidence in their ability to accurately determine who to select for interventions aimed at halting or reversing terrorism-related radicalisation. The indicators' importance to UK counter-terrorism was reflected in the expansion of their uses: while the ERG22+ was originally intended for use on sentenced TACT offenders, the shorter ERS version was used on a wider categories of prisoners -and the VAF was used explicitly on (as of yet) non-terrorist individuals deemed vulnerable to 'radicalisation'. The localisation of the second and third 'truths' in the United Kingdom manifested a complex assemblage of governing radicalisation in that it entailed a non-hierarchical self-organisation of security professionals, drawn together and forging alignments in the name of countering radicalisation.
Conversely, the ways in which the second and third 'truths' about radicalisation were localised in Greece represent a rudimentary assemblage of governing radicalisation. Unlike in the United Kingdom, Greece's localisation process did not generate any specific counter-and de-radicalisation programmes or referral mechanisms. No statutory duty to refer or prevent radicalisation has been assigned to frontline professionals, and civil society organisations have not been given any roles in the procedure. In short, the localisation of the second 'truth' that radicalisation can be stopped or reversed has not led to any institutional changes or law amendments so far. The same can be said about the localisation of the third 'truth' of radicalisation's measurability. The OIM's set of 43 indicators was directly translated from the CoPPRa without testing or validation for the Greek context. Furthermore, while the OIM's target population is broad and includes both offenders and non-offenders, it does not foresee any specific method of assessing and recording individuals' radicalisation-related risk, nor the systematic recording of reported suspicions.
These are all characteristics of a rudimentary assemblage of governing radicalisation, consisting of a far less non-hierarchical self-organisation of far fewer diverse elements. The dominant role was played by the CSS and the Ministry of Public Order and Citizen Protection, which 'stretched' the 'truths' about radicalisation and amalgamated heterogeneous attitudes towards political violence and terrorism in an attempt to make this content fit with the Greek context. Unlike the United Kingdom's TACT, Greece's terrorism legislation contains only a few terrorist crimes that do not entail direct violent behaviour against other people. This limits the potential elements that can pull together in the assemblage and the potential points of contact between them to security practitioners only, and delays the earliest stage of intervention upon radicalisation. It also postpones the intervention stage and, by extension, limits the number of elements and their potential connection points through the OIM's proclaimed call for tolerance towards (certain) 'radical behaviours' and 'youth spontaneous unlawful conduct'. The forces that could contribute to gather disparate elements, such as civil society organisations, health workers, and school and university teachers, to an assemblage allowing for early interventions are extremely weak in the Greek context. While basic convergence exists on the three 'truths' about radicalisation, what is at stake in Greece is rather the very meaning of the concept and what it entails in a highly politically charged context.

Conclusion
In this article we have attempted to explore how certain Europe-wide 'truths' about radicalisation were localised in the United Kingdom and Greece through the development and implementation of radicalisation-related risk assessment tools. We found that in both countries, the 'truths'-that radicalisation is a risk that could lead to terrorism; that terrorism can be stopped by rerouting, even reversing radicalisation at the earliest possible stage; and that radicalisation is measurable -have been localised. However, the congruence-building process has not been equally smooth in both cases.
In the case of the United Kingdom, the localisation process was more seamless, which appears related to the country's 'norm producer' status on the field of European counterradicalisation. An 'official' link between radicalisation and terrorism was established in 2006 when Prevent became central to UK counter-terrorism, with the aim to stop terrorism through tackling the radicalisation of individuals. The concept was initially ill-defined though associated with Islamist terrorism, but was later defined and stretched to include other ideologies. Legislative amendments of the UK Terrorism Act in 2006, which included several non-violent crimes, as well as the 2012 introduction of Channel as Prevent's operational mechanism, marked the localisation of the second 'truth' about radicalisation and the possibilities of intervening against the process. These developments enabled the assessment of individuals 'at risk' of radicalisation and their direction towards interventions. The creation and use of risk and vulnerability assessment toolsthe ERG22+, ERS, and VAF -exemplify the localisation of the third 'truth' about radicalisation being measurable.
By contrast, the localisation process in Greece has been less smooth, as the radicalisation concept was subjected to considerable stretching and re-framing in order to fit there. This seems directly related not only to the country's 'norm adopter' status on the field of European counter-radicalisation, but also to the 'spillover effect' from national counterterrorism policies; a diachronic issue of polarisation and contentiousness in Greece. Only in 2016 did radicalisation and terrorism become linked, through the 'Counter-Radicalisation and Extremism' project, which produced one guide, two manuals, and trainings for security practitioners to 'fight against terrorism and violent extremism' through 'tackling radicalisation'. Despite initial references to returning Jihadists, radicalisation was soon re-framed and associated with political/ideological violence. Moreover, the concept was significantly stretched -as manifested in the call for tolerance towards (certain) 'radical behaviours' and 'youth spontaneous unlawful conduct' -apparently an attempt to strike a balance between divergent attitudes towards political violence. Yet, the link between radicalisation and terrorism remained primarily a discursive one and has not yet resulted in any institutional or legislative changes. It has, however, led to the introduction of a radicalisation-related risk assessment instrument, the OIM, though yet unaccompanied by context-specific indicators, precise methods of assessing and recording individual levels of radicalisation-related risk, or targeted intervention programmes.
In the case of the United Kingdom, the assemblage of governing radicalisation consisted of the non-hierarchical self-organisation of numerous autonomous diverse elements, apparently driven by a broadly rooted need to make sense of and act on the post-9/11 and post-7/7 threat situation and maintain the United Kingdom's forerunner status on the field of European counter-terrorism. A wide range of authorities were involved in the assemblage and set up a number of diverse potential governing connection points through policy, institutional arrangements, and law. By contrast, in Greece, the governing assemblage consisted of a hierarchical self-organisation of fewer elements. The CSS and the Ministry of Public Order and Citizen Protection played key roles; the former in an attempt to enhance its own status and resources, 8 and the latter in an effort to increase the government's legitimacy. Indeed, the Greek government was still trying to rebuild its credibility within the EU, after allegations that the country's borders were left purposely uncontrolled, and amid threats from certain Greek officials to instrumentalise the refugee situation and the EU's fears of returning foreign fighters to gain leverage in the bailout negotiations (see Skleparis, 2018).
Our analysis has highlighted the importance of context and agency in the localisation of a security threat (see also Capie, 2008;Stritzel, 2011aStritzel, , 2011bStritzel, , 2014. Similar to De Goede and Simon (2013), our case studies have also demonstrated that understanding and countering radicalisation do not materialise in the same ways across different national locales. Localisation processes can lead to the formation of both complex and rudimentary context-specific assemblages. Indeed, as radicalisation 'truths' keep spreading, the number of rudimentary assemblages in particular will keep rising. Critical comparative research on such constellations is needed if we are to better understand the inner workings, driving forces, and disparate elements that make up the more elaborate ones, since rudimentary assemblages of governing radicalisation are potential complex assemblages 'in the making'.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.