Public Order and the Internal Security Apparatus: Affective Tension Monitoring as Police Epistemology 1

Introduction

In the early 1980s, in order to more effectively suppress public order threats from racialised minorities and organised labour, the English police developed a new taxonomy. The aim of this analysis was deceptively simple, the police wanted to predict where disorder would break out. To do this they identified the affective mood of ‘tension’ which they saw as sitting between the states of ‘peace’ and ‘disorder’. An atmosphere of tension was a generative state, they thought, where disorder could suddenly swirl out of their control. Peace or normalcy needed to be surveilled in order to identify tension, and tense communities needed to be more aggressively monitored and policed to prevent disorder. For the last 40 years, this police taxonomy of normalcy/tension/disorder has become operationalised in a ‘tensions monitoring system’ which has grown exponentially since its first inception. As part of this Special Issue on Lindsay Farmer's concept of ‘civil order’, this article follows the development and extension of the British ‘tension monitoring’ system. It aims to show the way in which Farmer's sense of ‘civil order’ can be understood for fields beyond (or beside) criminal law. The concept of the civil order can help us to describe the affective component of public order – the feeling of ‘peace’ or ‘normalcy’ – which the police seek to secure.

This article will trace the shifting techniques and analyses of tension monitoring. It argues that tension monitoring develops a way of knowing the affective life of the populace, and a way of securing (and by extension, fashioning) the civil order. In other words, tension monitoring is a key type of police epistemology, which is deployed as part of the internal security apparatus. I want to suggest that this particular police epistemology has a number key features. (a) It is less interested in veracity than standard police/security intelligence or investigation which aims at criminal prosecution. Instead it is structured around an understanding of the way information might resonate with the populace. (b) Its primary focus is on ‘community’ as the site of organic and unpredictable disorder. This sense of community is filled in with changing classed and racialised meanings as the political situation changes between the 1980s and 2000s. And (c) the epistemology makes a different collective subject to appear for the police. This is ‘the populace’ which is an aggregation of the different monitored communities. Playing a similar function to ‘the population’ in governmentality (Foucault, 2009), the populace functions as an aggregated totality of the levels of simmering affective discontent. As such, ‘tension monitoring’ helps us to see the crucial epistemic core of the internal security apparatus, at least as it relates to the sovereign peace.

Civil Order and the Sovereign Peace

At the core of this article is an engagement with Lindsay Farmer's concept of the ‘civil order’. In Making the Modern Criminal Law, Farmer deploys the concept of ‘civil order’ to help explain the emergence of the modern criminal law (2016). He gives civil order a threefold meaning of civil peace, life under stable institutions, and a sense of civility or civilisation. Civil order, in this sense, has a particular meaning, which is to say that it has a particular history. As Farmer notes, it is an order that has to be made and remade in various eras. This is important, because there is a tendency to see the question of the ‘peace’ or ‘order’ within a society as a sort of natural category – something which is given, simply because there is not anarchy and chaos. Farmer insists that civil order is not something which just exists in a ‘civilised’ state, because peace, civility and civilisation are fashioned, through subjecting, disciplining and pacifying the populace (see also Neocleous, 2000). This helps Farmer to understand the criminal law in a double sense: it both relies upon a sense of civil order and seeks to secure it. However, I want suggest that it can also open out in a different direction, that is it helps us to understand the sense of ‘the public peace’, ‘sovereign peace’ or ‘King's peace’ as an iteration of ‘civil order’.

Farmer's denaturalisation of the civil order helps us to see ‘the peace’ as a particular ordering, whose changing contours can be traced as it changes across time. What remains difficult, then, is to see or feel the peace. It is remarkably difficult to understand the ambient affective order in which we persist. In various works across the last decade, I have tried to use different approaches to grasp the ways in which ‘the peace’ can be understood as a ‘background’ of law (Wall, 2023a, 2023b). I have tried to understand the law of crowds (Wall, 2016), atmospheric techniques developed by the police (Wall, 2019a), the negation of the peace through the state of unrest (Wall, 2021), and through legal orders of ‘turbulence’ (Wall, 2017, 2019b). In this paper, I aim to go one step further, and explicate the idea of civil order and the apparatus of the sovereign peace, by looking directly at a particular police epistemology that is developed to understand the affective life of the populace. This police epistemology is oriented around securing the docility of the population in general, and of its ‘suspect populations’ in particular. It is a key part of the broader affective apparatus, and in this sense it is essential to understand the way in which it traces affects, sentiments, moods and atmospheres of discontent and docility.

The term affective apparatus is a crucial theoretical backdrop here. At various stages Michel Foucault used the term dispositive, which tends to be translated as ‘apparatus’, to designate a functional set of connections between diverse components. It designates:

a dominant strategic function that intervenes into forces of relations in order to direct and stabilize them, to manipulate them so they can be assembled into an integrated apparatus or network that draws the most heterogenous elements – institutions, statements, discourses – into its force field, supported by the power/knowledge nexus. Its networking power has the extraordinary capacity, Agamben adds, ‘to manage, govern, control, and orient – in a way that purports to be useful – the behaviours, gestures, and thoughts of human beings. (Edmondson and Mladek, 2017: 17)

In security studies, security tends to be primarily understood as ‘external security’ – protecting the state from ‘threats’ that are ‘out there’ – whether they foreign extremists, other states or refugees. It is increasingly important to address what is sometimes called ‘internal security’, by which is meant the creation and maintenance of ‘the sovereign peace’ within the borders of a state. Insofar as this concerns the feeling of ‘peace’ or ‘disorder’, the apparatus of internal security is primarily an affective apparatus. By this, I simply mean that it aims to understand and intervene in the collective moods and feelings of the populace. It aims to build and monitor an ‘affective infrastructure’ within particular communities, and the populace as a whole (Bosworth, 2023). The internal security apparatus seeks to understand and change the affective mood of particular communities or the populace at large.

The Emergence of Tension Monitoring

The practice of tension monitoring appears to originate in Home Office circular 39 in 1982. However, the text of this circular has been ‘lost’. It is not held by the National Archives in Kew or any of the other major policing and security archives in the UK. The National Archive's collection of government circulars in 1982 simply skips the entry, moving from circular 38 (‘Fee for licences for removal of human remains’) to circular 40 (‘Senior attendance centres’). The index of circulars simply records the word ‘confidential’ (see King and Brearley, 1996: 87–92, for discussion of these circulars). While the exact contours of this first direction are opaque, in a confidential letter from the senior civil servant Hayden Phillips which introduced the newly created public order manual (1982), we learn that senior police officers were enthusiastic about ‘the systematic monitoring of the potential for disorder’ both during periods of ‘tension’ and when disorder had broken out. In the ideas circulating at the time, ‘tension’ was tied directly to the outbreak of disorder. Later police texts would theorise this as a ‘riot curve’ and then the ‘disorder model’, where a community moves through affective stages of normality, high tension, disorder, de-escalation and back to normality (King and Waddington, 2004: 121). Tensions should be monitored, it seemed, because they could be used to predict disorder. Becoming sensitive to the affective environment in particular communities could help the police to prevent them becoming too restive.

Initial ideas on tension monitoring found their way into the 1982 Metropolitan Police's Public Order: Notes of Guidance for Senior Officers. The Notes explain: ‘It is often the case that serious public disorder occurs in areas where there are adverse social conditions such as high unemployment, poor housing and inadequate educational facilities. These factors can create a climate of tension which often results in confrontation with police’ (Metropolitan Police, 1980: 1). Deprivation, poverty and inadequate social provision generates anger and resentiment. The police's job is to ensure that the social peace that has been fashioned through pacification (Neocleous, 2000), is held together. The Notes of Guidance identify the need to track any ‘pre disorder indications’ through ‘some form of objective assessment which indicates significant changes in the mood of the community leading to an increased potential for disorder’ (Metropolitan Police, 1980: 1). They do not trace the cause of disorder, but rather they identify trigger events which might increase tensions. This first draft of the now familiar ‘tension indicators’ contains many of the markers that would become familiar in more recent iterations of the system: Police officers becoming the targets of abuse or assaults; An increased unwillingness to assist police; Racial attacks and other inter (racialised or classed) community acts of violence, significant anniversaries or events elsewhere in the country that might be contagious, or rumours. But they also included some curious elements, like ‘pop concerts in sensitive areas’.

It is important to recognise at the outset, that these ‘indicators’ do not simply identify something which exists there in society. Tension is certainly an intense atmosphere (cf. Wall, 2024), but one of the key insights of the work on (affective) apparatuses is that the institutional epistemologies are generative rather than reflective. Ben Anderson observes that ‘it is from within apparatuses that affective life becomes an object-target. Apparatuses produce specific versions of what affect, emotion and other modalities are and do’ (Anderson, 2014: 33). Complex and disparate affective moments and conditions become legible as ‘tension-generating’. Ultimately, the affective life of the populace persists in a myriad of ways, which is to say that we all live in a shared affective soup, which is full of eddies and swirls, forces, energies, tendencies, structures, discourses, etc. This infinite plurality exists at the level of the population, but at least at the outset, the police apparatus was less concerned about the general populace. It began by targeting particular ‘communities’ who were feared to hide the ‘enemy within’, particularly urban racialised minorities and working class communities (Milne, 1994). These communities, insofar as they exist as discrete groupings, also have an intensely plural affective life which is complex and chaotic. Just like the general populace, it is impossible to reduce this to a singular affective register. It is therefore important to understand that when the police tension monitoring apparatus targets the affective life of these communities by reducing it to a single register (normality, tension, disorder), they produce something distinctive. The police imagine correspondence between the experience of these communities and their singular register, but in fact they produce something which did not exist previously, and which is ascribed the status of ‘reality’. It is not that ‘tension’ or ‘calm’ do not pervade various heavily-policed communities, but that the tension monitoring apparatus seeks to identify these atmospheres in particular ways, and in so doing it gives them a particular (and new) form. Thus we need to pay particular attention to the way in which the state of tension is being created as an object-target of intervention. Anderson ascribes this ambivalence to the affective nature of the object-target. He comments that ‘affect ‘exists’ in the sense that affects are ‘inscribed within reality’ through apparatuses that know, render actionable and intervene in life’ (Anderson, 2014: 37).

In this first phase of tension monitoring, which persists up until 1999, the Metropolitan police insisted that it was important to understand tensions in a situated way because ‘the level of tension in any particular area may be indicated by incidents or factors peculiar to that area’ (1982: 2). While their initial attempts to draft indicators was incomplete and imperfect, in 1982 they also began to appoint District Information Officers whose role was to synthesise on-the-ground observations and signal any worrying developments to the Central Information Unit in the Met's public order branch (Joyce, 2013: 224). The introduction of roles aimed specifically at developing the tensions indicators would prove important. While the missing Home Office circular of 1982 would likely help us to see exactly what was going to be measured and reported on, we can, nevertheless find evidence of the system and how it began to operate.

By the time of the miners’ strike, we can observe the presence of a comprehensive (although unsophisticated) national reporting system for gathering observations about the mood and atmosphere of sites of potential disorder, and aggregating and analysing these observations about their affective tension. During the strike, day-to-day police situation reports (sit-reps) included observations about the tension levels in each of the coal fields. The national reporting centre (which had within months established an office in the heart of the miner's strike) was responsible for aggregating and analysing this data. The Association of Chief Police Officers archive in Hull has stored a large amount of these reports, and they make for fascinating reading. So for instance, on the first of April, 1984, the pit at Thorpe Marsh, near Doncaster, had five pickets of local origin in a ‘peaceful’ mood; but on the same day a crowd of 200 gathered at ‘Helcock and Simpson’ near Rotherham, and were ‘unruly’ in their behaviour. The sheets were supposed to identify the three officially categorised moods. A movement of the mood from ‘peaceful’ to ‘hostile’ or to ‘violent’ signalled a linear escalation of tensions (National Reporting Centre, 1981, Appendix C). But with such an enormous operation, inevitably the observed categories quickly frayed and began to multiply. So alongside the officially sanctioned moods, police also recorded moods that were: ‘friendly’, ‘belligerent’, ‘quiet’, ‘noisy’, ‘angry’, ‘fairly hostile’, ‘good’, ‘agitated’, ‘rowdy’, ‘stable’, ‘boisterous’ and ‘less amicable’.

The identification of crowded moods and atmospheres was not new. Public order policing in the UK (up until around 2010) was based on fin-de-siecle crowd theory which focused upon the irrational moods and sentiments that overtook crowds as they became ‘deindividuated’. So for instance, an undated public order training manual from the Association of Chief of Police Officers in the 1970s explained:

Any large crowd contains the potentialities for panic and riot. Consequently all crowds require police attention. The action and behaviour of the police towards a crowd influences the development of its potentialities. The police, therefore, should analyse the crowd and the circumstances of its formation; they should learn its purpose, appraise its temper and estimate the likelihood, nature and extent of immediate disorder. Police action towards the crowd should be based on this analysis. ²

Cohesion and the Problem of the Normal

From Saturday the 23^rd to Monday the 25^th of June 2001 Burnley erupted into rioting and disorder. Burnley is a town near Manchester in the Northwest of England. It has been scarred by racial tensions for decades, and has a significant far-right constituency. The 2001 riots involved groups of Asian and white racist teenagers and young men, fighting with one another and targeting symbols and infrastructure. As Ash Amin explained: ‘young Asians living in the old, rundown textile mill towns of northern England went on the rampage to protest against a long history of economic deprivation and hopelessness, white racist threat and violence, police intrusion and incursion, public-sector neglect, and failed ethnic leadership’ (Amin, 2003: 461, see also Bagguley and Hussain, 2008). Despite their confidence in the tension monitoring system, the police apparatus had not picked up the build-up of tensions, and so had not sought to deescalate or intervene until after the disorder had broken out. Reviewing this police failure, King and Waddington suggest that one of the key problems for the police had been the problem of the ‘normal’ in the tension monitoring system. In the eighties and early nineties, the indicators measured tensions against a mono-affective idea of normality. Erlichman had commented in the mid-1990s, we ‘need some indicator of ‘normal levels of tension’. For example, if we have 4 officers assaulted, on average, per week in [a particular neighbourhood] and one week we have 7, does this represent a significant ‘blip’?’ (Quoted in King and Brearley, 1996). There are places where the ‘normal’ level of tension is high, but this normal high-tension is not generative of public disorder. In particular, areas with higher unemployment, deprivation, police repression, or greater prevalence of racist or inter-community frictions will likely have a higher level of ‘normal’ tensions as far as the police are concerned. Should the police measure the tension against what is normal for those spaces, or continue to think of it as a more generic measure of what is normal for urban spaces or on a national scale?

In 1999, Her Majesty's Inspectorate of Constabularies (HMICs) began to adopt a different approach. They sought to replace the ‘normal’ as a mono-affective, universal component – a singular feeling of calm or peace. Instead they suggest: ‘what is defined as normal will vary from one area to another’ (HMIC, 1999, Quoted in King and Waddington, 2004: 122). In this sense, ‘normal’ signifies the level of tension that any given community will bear, without the immanent possibility of disorder. What is interesting here is not simply that tensions might differ, but that the levels of deprivation and disadvantage that some will bear is much higher than others, because of structural, cultural, ideological and communicative reasons. King and Waddington have shown the way in which the conception of ‘normality’ effectively obscured the racism which grounded the Burnley riot. ‘In applying their conceptual model, the police submission defines the pre-conflict state of ‘Normality’ in Burnley as one in which racism did not present itself as a particularly serious problem’ (King and Waddington, 2004: 124). In contrast, the Burnley Task Force reported that ‘anti-Asian sentiment… had been gradually stoked up and become manifest in the prelude to the disorder’ (King and Waddington, 2004: 129, see also Amin, 2003: 460). The sense of a gradual build-up is important here. Because of its incremental shift in the affective ambience, it never steps out from the police sense of the ‘normal’. King and Waddington insist that the ‘intelligence and interpretive mechanisms utilized by the police lack[ed] the necessary sophistication required to reasonably anticipate a probability of disorder’ (2004: 130). Ultimately, they show that ‘in terms of their use of tensions indicators, Burnley police were over-focused on straightforward criminal indices of impending disorder’ (King and Waddington, 2004: 134). This meant that they could not identify the slowly inflating sense of racial animus. King and Waddington insist that: ‘By recontextualising the Burnley riot in terms of relevant structural, cultural, ideological and communicative variables, it becomes possible to appreciate how the Burnley police critically underestimate the magnitude of the racist threat to its minority ethnic populations’ (King and Waddington, 2004: 134).

While the ‘normal’ provides a base-line against which to measure tension, it does not provide an object-target for an affective apparatus to change. The work that the concept of ‘the normal’ does, is to help the police measure its ‘other’ – tension/disorder. Following the 2001 riots in the Northern towns, the New Labour government turned to the concept of ‘community cohesion’ as a diagnostic tool for understanding social problems (Donoghue, 2016: 262, see also Home Office, 2001). Cohesion became a way of differentiating the ‘normal’ and therefore intimately connected with the story of tension indicators. While ‘social cohesion’ was a term popularised in various Europe institutions in the 1990s as a way of addressing concerns about social stability after neoliberal structural adjustment, ‘community cohesion’ in the UK stems directly from this problem of the normal affective atmosphere of civil order. In the New Labour discourse, community tension and cohesion were directly correlated. For instance, the Institute for Community Cohesion describe how this tension might grow in spaces of good cohesion because of particular events which negate that cohesion, or how tension might arise ‘as a result of the absence of those factors which produce cohesive communities’ (Institute of Community Cohesion, 2010: 11). They define tension as ‘a state of community dynamics which may potentially lead to disorder, threaten the peace and stability of communities or raise the levels of fear and anxiety in the whole or a part of the local community’ (Institute of Community Cohesion, 2010: 11). Normalcy remains within the models, but cohesion helps to create a more complex picture where disorder can also arise out of the erosion of the conditions for civil order.

As part of the community cohesion agenda, tension monitoring became folded into a much broader governmental exercise, focusing on local responses to community unrest. In particular, the government shifted from a police-oriented model to a partnership approach. This involved a major extension of local government activities in the field, fashioning a multi-stakeholder structure which would be closer to the ebb and flow of local feeling. Either the local authority or the relevant police force was to chair these multi-agency tension monitoring groups. But the group itself should:

include key partners from the statutory sector (eg. Housing, community safety, education, fire service, health, probation/youth offending team, community workers, neighbourhood wardens and police community support officers, National Asylum Support Service) and the relevant representatives from the voluntary, community and faith sectors’. (Department for Communities and Local Government, 2008: 17)

In 2008, the Department for Communities and Local Government pointed to the possibility of calming tensions through a much more varied series of responses. Instead of flooding the tense area with police officers (which would often exacerbate tensions) (HMIC, 2009: 172), they suggested outreach to faith communities (for instance, through Friday prayers for Muslim communities), requests for other community groups to publicly express solidarity, ‘mediation resources’ like ‘Neighbourhood Renewal Advisers’ and ‘positive youth engagement’ (Department for Communities and Local Government, 2008: 12). These interventions would employ a methodology that focused on the tensions experienced by the community, the objective evidence, and the potential tensions that might arise due to forthcoming events or predicted police activity (Institute for Community Cohesion, 2010: 14). Each of these three components could then be given a numeric value (1. Imminent; 2. High; 3. Moderate (high); 4. Moderate (low); 5. Above Normal; 6. Normal), and a descriptor (Falling; Static; Raised). The system therefore allowed the state agencies to differentiate different forms and levels of tension. ‘For example, a significant difference between experienced and evidenced tensions might indicate the need for reassurance (where experienced is much higher than evidenced)’ (Institute for Community Cohesion, 2010: 15). Crucially, particular emphasis was to be placed on ‘knowing how communities feel’ (Institute for Community Cohesion, 2010: 14), rather than simply identifying objective measures of crime. Thus the Institute warns that while a numeric value will be assigned by the tension within a community, these should not be readily compared. ‘Different areas will have different levels of tolerance to tension and different experiences of what is ‘normal’… and that what is high tension locally may not be viewed the same in another area’ (Institute for Community Cohesion, 2010: 15).

Without wishing to state the obvious, this represents a significant sophistication in the tension monitoring epistemology. Instead of a singular scale, which places police forces at the heart of the determination, in the 2000s, tension monitoring became a massive operation in which the entire local government, welfare and police services were drawn into a form of affective surveillance. This suggested that there would be a significant suppleness to the system, with immense data points available for analysis, through a hierarchical but largely delegated system. The responsiveness of this new, sophisticated system would come under significant scrutiny following the London riots of 2011, and the crisis that followed those events led to the second fundamental reconfiguration of the tension monitoring system.

The Security Solution

On the 4^th of August 2011, Mark Duggan was shot dead by police officers during a ‘hard stop’ in Tottenham in London. The following day the local tension monitoring group met and graded the tension as 5 (on the 6-point scale) across the board on the ‘experiences, evidenced and potential’ scales (de Heyer, 2020: 133). It is worth remembering that 6 equates to ‘Normal’, and 5 is merely ‘Above Normal’. They did not think that the tensions had even reached the ‘Moderate (low)’ level. That evening the crowd that gathered outside Tottenham police station became increasingly disorderly and one of the most extensive moments of rioting in decades in the UK began to unfold. One of the key issues which emerged in the wake of the 2011 ‘London Riots’, was the failure of the multi-stakeholder model of tension monitoring. The police wanted to know whether this was a systematic failure or merely a local failure to adequately understand community dynamics. Ultimately, the police decided that it was a failure of systematicity. They felt that tensions indicators had not formed the basis of the policing response to the growing discontent in Tottenham in the run up to the riots. HMICs wrote: ‘National work to collate and analyse information on community tension was detached from other sources of information and, as far as we can establish, did not inform the police's picture of events as they unfolded’ (HMIC, 2011a: 6). In response, therefore, the Inspectorate recommended the development of an ‘all source, fully networked public order intelligence hub, using advanced software to analyse trends in community tension (including through social media monitoring)’ (HMIC, 2011a: 9). In this, they turned to the existing security apparatus to provide new solutions, and in particular, to a new cooperation with the institutions of Britain's counter-terror system.

This represents a certain resolution of a tension that had been present since the summer of 2001:

In the context of Britain's concern with social cohesion, the British Muslim population has been defined as an acute problem requiring specific remedy. This concern became all the more acute following the ‘riots’ in northern towns in 2001 where issues of urban decay and violence acquired distinctive new focus: the problem of ‘self-segregated’ Muslim communities. (Husband and Alam, 2011: 2)

One of the key findings following the London riots was the need to incorporate social media intelligence (SOCMINT) into the broader analysis of moods and atmospheres of dissent, and in particular to develop the capacity ‘to analyse social media in real time in order that impending unrest can be managed’ (Burnap et al., 2013: 106–107; referencing HMIC, 2011b). Omand, Bartlett and Miller explain the attraction of social media; ‘Measuring and understanding the visage of millions of people digitally arguing, talking, joking, condemning and applauding is of wide and tremendous value to many fields, interests and industries’ (2012: 804). HMIC presented it as ‘the cyber equivalent of ‘the word on the street’. This fast paced open source information space should be part of the street which they patrol’ (2011a: 36). They explained that SOCMINT can be deployed to understand the intensity of feelings and sentiments of protest groups and the broader community, giving real-time situational awareness, insight into the mood of groups and identification of specific criminal plans or intent (2011a: 36, see also Omand et al., 2012: 806). Using SOCMINT for tension monitoring appeared to get around a major problem of the post-Burnley system. There was a tendency to only hear ‘upstanding’ members of the surveilled communities. SOCMINT promised to get at those ‘hard to hear voices’, which could not be represented in the post-Burnley system of terrestrial tension monitoring.

The National Domestic Extremism Unit (NDEU) was given responsibility for coordinating digital and terrestrial tension monitoring, as part of their broader protest surveillance remit. This unit was originally founded to deal with animal rights direct action in the 1990s and 2000s, but was reformed in 2004 to include surveillance on other forms of militancy. It was placed under the Metropolitan Police's Counterterrorism Command in 2011, following the initial revelations about the wrongdoings of the undercover officer Mark Kennedy (Dencik et al., 2018: 1437). The NDEU was renamed the National Domestic Extremism and Disorder Intelligence Unit (NDEDIU) in 2013, and stripped of its powers to run undercover police. From the outset, this unit was designed ‘to collate, assess, develop and disseminate intelligence and intelligence products across all police forces of the UK in order to prevent and detect acts of domestic extremism and protest that cross over the legal threshold’ (HMIC, 2012: 90). And so, following the London riots it was perfectly placed to integrate social media surveillance into the broader terrestrial tension monitoring reporting system.

As a proof of concept in 2013, the Cardiff Online Social Media Observatory created a platform to show that it was possible for researchers to observe and visualise social tension, ‘providing the potential to forecast social disruption via deviations from ‘normal’ levels that are established over time (be that minutes, hours or days)’ (Burnap et al., 2013: 107, see also Williams, 2013). But while this sentiment analysis was possible, three years later senior officers reported that it appeared over-rated. Dencik, Hintz and Carey interviewed five senior officers at the NDEDIU in 2015. While these officers were negative about sentiment monitoring, they did consider SOCMINT particularly useful in the identification of threat words, risky individuals and influencers/organisers (Dencik et al., 2018: 1441). In particular, they found that the police saw SOCMINT as a useful tool of pre-emptive policing. It would inform risk assessment of operations and other aspects of planning (Dencik et al., 2018). Thus, despite the initial promise of digital tension monitoring, by 2015, it had not delivered the revolution it had promised. Rather it seemed to augment other terrestrial forms of predictive policing. That said, it is important to note that little research has been done in the last decade on the subject. And that is a significant issue, considering the significant advances that have been made in algorithmic surveillance and control.

A Police Epistemology

At core of this genealogy of the tension monitoring system that has developed in England and Wales over the last 40 years, is a desire to predict when and where disorder might break out. But to undertake this predictive work, it is first necessary for the state to understand the shifting affective currents that exist within the populace. The different ways in which the police, local authorities and security services have gone about designing this ‘work of knowing’ constitutes what I am calling a police epistemology. It is important to see that tension monitoring has a very different epistemic structure to other forms of police knowing. We might identify three distinct parts of the epistemic structure which marks this difference. To begin with, we can see that there is a very particular sort of information being collected. Tension monitoring is different from both evidence gathering and the more general category of intelligence (although it may produce both evidence and intelligence). As forms of knowledge, evidence oriented around the vagaries of the criminal justice process, and intelligence is oriented around the possibility of police action or intervention. In the police epistemology, veracity is the key question for both intelligence and evidence. Misleading, erroneous or false evidence and intelligence is a problem for the police. Tension monitoring, on the other hand is focused on assessing the ebbs and flows of the affective life of the community, and in this way it sidesteps the question of truth. Because the aim of tension monitoring is to predict where disorder is likely to break out, what matters is the way information works on people. As the Institute of Community Cohesion explains, the system is ‘only as useful as the quality of the information fed into it’ (2010: 29). But quality is not measured by veracity: ‘False information can be as useful as true information, especially in the sphere of tension monitoring, where myth and rumour are powerful forces’ (Institute for Community Cohesion, 2010: 29).

The system is interested in ‘resonance’, that is, the sense that some event or idea will chime with the particular community (Institute for Community Cohesion, 2010: 29). This was one of the key problems with the police-oriented system of the eighties and nineties. The multi-agency and hybridised digital models sought to move the measurement of resonance away from the police, a force that was often detached from the communities in question. Digital tension monitoring seemed to move even closer to the ability to measure actual resonances within the population by tracking the intensity of their sentiments online. But when the police came to try to understand this, they found that it required far more sophisticated contextual understanding of the community (Dencik et al., 2018). Crucially, resonance and veracity are radically different forms of knowledge. Resonance is a measurement of the potential for affective intensity (Mühlhoff, 2019: 189). In the context of tension monitoring, it is a measurement of the latent potential of a piece of information, a story or an event to draw a community to movement.

Secondly, tension monitoring is unique as an affective apparatus because unlike major surveillance projects undertaken by the British police in the same era, it does not aim directly at individuals or specific groups. Tension monitoring makes a collective subject present. In recent years, this is theorised explicitly as a ‘community’, which includes communities of identity (‘people who share a common culture’), communities of place, or communities of interest (which is ‘some shared activity, such as employment, political activity or leisure activity’) (The Welsh Government, 2011: 12). From the outset, the police were concerned with how to control the political, economic and social expression of racialised minorities, working class communities, political radicals and oppositional groups. For over a century they focused on networks of dissent, centring ‘agitators’ and more recently the spectre of ‘aggravated activists’ in their analysis (HMICFRS, 2021). ³

Shifting to communities represents a shift away from thinking about collective dissent as the result of agitation of a small number of influential activists. Instead of a passive recipient of information, it sees the community as an active and dynamic part of the production of disorder. To understand this active processing, tension monitoring gradually develops mechanisms to understand existing structures of feeling (Carvalho, 2023). To a liberal sensibility, this focus on the community may seem less intrusive, particularly when contrasted with the individual targeting of the ‘spycops’ regime (Schlembach, 2020; Woodman, 2018). By targeting a broad swathe of the populace, particular individuals are included only in a general sense. In this way, tension monitoring only has a marginal effect on the individuals’ human right to privacy. But it is very clear that the tension monitoring apparatus aims to understand the affective life of the community in order to better control their expressions of dissent. One of liberalism's key tenets is that targeting an individual is an unacceptable invasion or violation, unless there is some important countervailing value. A state invading an individual's privacy in order to control and manage their political feelings and ideas is something approach an ur-failure for liberal political theory. Tension monitoring presents the possibility that police might have the same level of control of feelings, without ever invading an individual right to privacy. But because it does not specifically target individuals, it appears to be distinctly more palatable to both the public and to the field of human rights. This does seem to pose an uncomfortable question of human rights, for instance, will tolerate high levels of control of a community's feelings, as long as individuals are not specifically targeted.

Finally, while ‘community’ has been the primary subject of the tension monitoring system, it is important to see that at a certain level, the police and the security services seek to go beyond particular communities, and produce a different collective subject. At each stage, from the early 1980s, through the analogue multi-agency era, and very explicitly in our contemporary hybridised digital period, the police have been keen to aggregate community-level data. In this aggregation, a different epistemic subject begins to emerge. I want to suggest that we think of this as the ‘populace’. There are three closely related meanings to the term ‘populace’: ‘the ordinary people; the inhabitants of a particular place (a synonym of population), or a multitude, crowd or throng. Populace and population are synonyms in the sense that both suggest the need for management/government. But into this mix, the populace also adds the pejorative sense of the rabble or the masses’ (Wall, 2021: 8). In other words, in an etymological sense, while the population is the subject of governmental intervention, the populace is a security problem. Thus, as a way of thinking about the aggregated tensions data, the populace captures that sense of an affectively unstable collection of communities. In this police epistemology, the populace emerges as a ‘national picture’ of the potential for disorder. This is important when we return to the question of civil order, because it equated directly to the affective life of the populace – at least as it concerns the question of state security.

Conclusion: Knowing the Civil Order

Exploring the shifts in the police epistemology in public order policing opens a window onto the shifting nature of the performance of sovereign orders (cf. Matthews and Wall, 2024). From the ancient Greek irenarch to the British magistrate, and from the colonial police to contemporary digital affect policing, the way that agents of public order seek to know the ‘peace’ is crucial to the way that sovereignty is made present for the populace. But across these shifting orders, there are also changing police epistemologies. The specific object-target of public order knowledge has been recast from the identification of specific leaders and organised groups who might encourage dissent (Lobban, 1990); to the understanding of crowds and their psychology (Schweingruber, 2000). Since the early twentieth century, there has been a growing interest in that investigating affective currents of dissent at the level of the population or populace (Owens, 2015). But where the analyses of morale (Anderson, 2014) or the early forms of tension monitoring generally relied on indirect or symptomatic readings of sentiments, social media tensions indexing seems to offer a paradigm shifting opportunity for the police (Grill, 2021). In particular, it seems to short-circuit individualistic protections of privacy, offering new (and largely unprotected) collective subjects that are capable of being controlled. While existing individual rights protections can be used to make sure specific people are not identifiable, they do little to prevent the police developing ever greater techniques to control political expressions of dissent.

Farmer's conception of civil order is remarkably capacious, in the sense that it provides the ground from which criminal law emerges, and which becomes the object-target of criminal law. One component of the civil order, as Farmer defines it, is the civil peace. But while this is an important component, it is largely undertheorised in The Making of the Modern Criminal Law (2016). Farmer briefly explores different ways of conceptualising the peace – as the opposite of war, as state security, police order or an aspect of social ordering (2016: 37–60). While it is clear that the peace, like the civil order, is something that needs to be produced, The Making of the Modern Criminal Law largely lacks a strong sense of ‘the peace’. This article has sought to address that lack, by focusing on one specific technique – tension monitoring – and showing how that form of police epistemology helps us to understand the way in which the internal security apparatus seeks to know and manage the affective life of the populace.

In closing, I want to suggest that this is important, because it helps us to understand something particular about what Melayna Kay Lamb calls the ‘logic of police power’ (2024: 2). In her superb philosophical history of police power, Lamb argues that there is an essential connection between the police power (which is broader than contemporary police forces) and the sovereign order. The police is ultimately concerned with the security of the state and its institutions. Tension monitoring gives the police (and security forces) a window into the affective life of the populace. This can help to control particular communities, who are more readily restive. But it is also an essential tool to manage the populace as a whole, thereby ensuring that threats to the sovereign order, even when they come in popular or ruptural democratic forms, are eliminated. In this sense, the ‘deep logic of police power’ as Lamb puts it, is to prevent movements gathering momentum. The police must watch carefully for resonance-building eddies, combining street-level affective observations into a sophisticated and nuanced aggregated picture.

The challenge of thinking about ‘the peace’ as a key component of the sovereign order or the civil order, is that it is difficult to find the material to analyse. This is because we struggle to grasp the power of the background. As I explained elsewhere, the background is

precisely not where ‘the action’ happens (foreground) nor is it a space of reason or synthesis (middle ground). The background is scenic in the sense that it ‘sets the scene’, but it does so in a way that is always displaced. Without the background the middle and foreground make little sense, but at the same time it is only by fading from attention that we can begin to grasp it as background. (2023b, 191)