(Re)producing Guilt in Suspect Communities: The Centrality of Racialisation in Joint Enterprise Prosecutions

Joint enterprise (JE) is an extraordinary legal device deployed to punish and (re)produce those who are frequently presented as threatening the normative boundaries of the British state. In acknowledging the global relevance of over-representation and the use of collective punishment, this paper presents the accounts of prisoners who have been convicted under JE laws across England and Wales. Analysis reveals a particular process of criminalisation through police and crown prosecution teams’ construction of the ‘gang’ narrative in courtrooms to drive the disproportionate punishment of members of negatively racialised communities. Of concern, the findings reveal that young Black men in particular are at risk of being convicted and punished for offences they did not commit. This paper empirically demonstrates how such racial injustice originates from a series of targeted and criminalising policies and practices.


Introduction
The question of racial disproportionality within the criminal justice system (CJS) remains a vexed issue in England and Wales amid the ongoing contention that negatively racialised groups have a quantifiably different experience of the CJS (Ministry of Justice [MOJ] 2017b). The 'extreme over-representation' of racialised and indigenous people (Cunneen and Rowe 2014) in other countries with colonial histories presents an urgent focus for comparative criminology (Tubex 2013). Contemporary political, criminological and policy discourses seeking to explain such differences in treatment remain inconclusive, and yet paradoxically serve to affirm the imagery of negatively racialised communities as being disproportionately involved in the committal of serious forms of offending behaviour. This paper begins by rehearsing the salience of racial disproportionality throughout the CJS of England and Wales. Yet, rather than a simple regurgitation of the 'statistics on race in the CJS' (MOJ 2017a), we reflect upon how these repeatedly produced but unexplained associations legitimise the application of distinct penal apparatus into negatively racialised communities. To this end, we draw upon the findings to emerge from a collaborative research study into the contemporary use of joint enterprise (JE) a common law doctrine that legitimises the collective punishment of multiple people for a single crime.
Prior to discussing the findings, this article establishes the re-emergence of JE in England and acknowledges the enduring nature of this common law doctrine as a feature of British colonial rule of law. A review of the existing contemporary criminological and legal literature on JE reveals evidence of racial disproportionality and the potential significance of the 'gang', yet relative silence on the significance of processes of racialisation in JE prosecutions of serious violent offences.
Our findings challenge this silence by revealing how criminalising associations of the 'gang' as a guiltproducing device relies on the reproduction of racialised associations to criminality, resulting in a significantly high number of ethnic minority groups serving lengthy custodial sentences for offences they did not commit. Such police and law enforcement processes and practices are contingent upon the specific construction of young Black men as culturally predisposed to gang-enabled violence, rendering them as legitimate objects for policing and collective forms of punishment. The findings developed throughout this paper facilitate the (re)conceptualisation of the causal mechanisms that drive institutionalised forms of oppressive policing that results in racialised CJ outcomes. Therefore, this paper marks a significant contribution to a growing body of critical and progressive criminological research seeking to expose and disrupt harmful and disproportionate processes of criminalisation.

Reporting and Explaining Over-Representation: 'Racial Disparity', 'Bias' and 'Possible Discrimination'
If you're Black, you're more likely to be in a prison cell than studying at a top university. And if you're Black, it seems you're more likely to be sentenced to custody for a crime than if you're white. We should investigate why this is and how we can end this possible discrimination.

Former Prime Minister David Cameron announcing a review of racial bias (UK Government 2015)
The most recent publication of Statistics on Race in the Criminal Justice System (RCJS) in England and Wales reveals that Black, Mixed, Asian and other minority ethnic (BAME) people are not only eight times more likely to be stopped and searched by police compared to their white counterparts, but also three and a half times as likely to be arrested (MOJ 2017a). Evidence from the same report suggests that while 15 people per 10,000 of the 'white' population are serving a custodial sentence, this figure increases to 52 per 10,000 for the 'mixed' group, 75 per 10,000 population for the 'Black' group and yet is 13 per 10,000 for the 'Chinese and other' group. Alongside these findings, Her Majesty's Inspectorate of Prisons analysis reveals that over half of children and young people in prison or the secure estate are members of minority www.crimejusticejournal.com IJCJ&SD 118 2020 9(3) ethnic groups (Green 2019). Despite these problematic statistics, there is a paucity of empirical or theoretical knowledge to understand the enduring differential experiences of minority ethnic groups within the CJS of England and Wales.
While the RCJS states that 'no causative links can be drawn from these summary statistics' (MOJ 2017a: 7), this biannual publication advertently infers a propensity to criminality for those who are captured within the analysis, as notes on the complementary infographics repeatedly suggest that differences are 'possibly associated with variance in offence types between ethnicities'. The publication can arguably then be (mis)read as a proxy for offending, rather than the cumulative effects of individual interactions and encounters with CJS practitioners and agencies (Tufail and Poynting 2013;Williams 2018). The reticence of government to interpret the statistical data further serves to conceal the processes and practices that drive the 'possible discrimination', inferred by Cameron above.
While RCJS states that the 'identification of differences should not be equated with discrimination' (MOJ 2017a: 11), the tone of the now-published Lammy Review affirms that ethnic 'disparity' is driven by 'racial bias' within the system (MOJ 2017b). The Lammy report refers to forms of 'implicit' and 'explicit' bias, yet often fails to explain the causal processes that drive 'explicit' bias (citing only the empirical data from the project examined in this article when referring to such 'explicit bias'). Problematically stopping short of calling out the multiple forms of racism(s) embedded within the system and its processes (Fekete 2017). The analysis commissioned for the Lammy Review confirms that offending behaviour patterns of BAME and white people are broadly similar, and in some instances lower for BAME groups, yet the treatment is disproportionate (Uhrig 2016). Of further significance, while the Police Service was excluded from the scope of the Lammy Review, with the terms of reference focused upon the MOJ, 1 Uhrig (2016) highlights the intersections between police arrest and subsequent racial disparities throughout the CJS by consistently signposting interactions with police as the salient 'juncture' from which ethnic disproportionality emerges.
Alongside this, the question of racial disparity has been subject to scrutiny by other recent governmentsponsored reviews. The Young Review, commissioned in 2014 to 'improve outcomes for young black and Muslim men in the [CJS] in England and Wales', and the Angiolini Review, published in 2017 on deaths and serious incidents in police custody, both concluded that criminal justice responses are driven by an array of stereotypical, racialised assumptions. For example, Angiolini (2017: 87) noted, 'the stereotyping of young Black men as "dangerous, violent and volatile" is a longstanding trope that is ingrained in the minds of many in our society'.
For Phillips and Bowling (2007: 429), 'old' and 'new' images of Black communities as prone to violence, drugs and disorderly behaviour have become 'entrenched in the public consciousness'. Similarly, Gutzmore (1983) noted some 35 years ago that there has long existed a 'special social category' attributed to young Black men, which serves to legitimise the development and application of an increasingly complex and punishing penal apparatus. We argue here that the emergent use of JE represents the latest manifestation of the state's punitive reaction to an imposed stigmatic 'special social category' upon Black, mixed-race and Asian people, which has resulted in the reproduction of longstanding racialised and criminalising assumptions.

The (Re)emergence of Joint Enterprise
JE is a doctrine of common law that permits the prosecution of more than one person for the same offence. The doctrine can apply where suspects have played different roles, and where a suspect was not in the proximity of the offence committed. Intrinsic to its application is the principle of 'common purpose', in which it is alleged that individuals have planned to commit a crime together. Where such a 'common purpose' is demonstrable, peripherally associated individuals, or 'secondary parties', may be held liable for crimes committed by a member of a group, even though they may not have participated in the crime or intended to commit it (Crown Prosecution Service 2018). Therefore, convictions under JE are contingent upon police and prosecution practitioners establishing common purpose, thus, building associations to www.crimejusticejournal.com IJCJ&SD 119 2020 9(3) affirm that said group shares a 'belief and contemplation' that the principal 'offender' might commit an offence and, thereby, themselves are culpable and guilty. Given the significant power of JE to criminalise and convict individuals based upon their association to an offender or a crime, it is alarming that there continues to be no official record of its use (Jacobson, Kirby and Hunter 2016).
One of the first attempts to explore the prevalence of JE prosecutions across England and Wales was undertaken not by criminology or legal scholars, but by the Bureau of Investigative Journalism. Here, it was estimated that between 2005 and 2013, 'at least 1,800 people and up to 4,590' have been prosecuted for JE homicide (McClenaghan, McFadyean and Stevenson 2014). Since then, and following a sustained legal campaign against the use of JE, the Supreme Court recognised, in overturning the conviction of Ameen Jogee in 2016, that the law had taken a 'wrong turn' in the 1980s, confirming the realistic possibility that miscarriages of justice have occurred in the prosecution of JE cases. 2 It is pertinent that the forces shaping the problematic use of these laws over the last four decades have been both legal and political (Green and McGourlay 2015). Indeed, while the Supreme Court acknowledged that 'the law took a "wrong turn" ' (Squires 2016), a central driver in the reincarnation of JE has been political will, driven by a law-and-order rhetoric. The support for JE was most clearly expressed by Lord Falconer in 2010, who argued: the message that the law is sending out is that we are very willing to see people convicted if they are a part of gang violence-and that violence ends in somebody's death. Is it unfair? Well, what you've got to decide is not, 'Does the system lead to people being wrongly convicted?' I think the real question is: 'Do you want a law as draconian as our law is, which says juries can convict even if you are quite a peripheral member of the gang which killed?' And I think broadly the view of reasonable people is that you probably do need a quite draconian law in that respect. (BBC4 Today Programme 2010) A central contention of critical social research is that the law is used to reflect existing determining structural relations (Clarke, Chadwick and Williams 2017), and that criminal justice policy and practice reproduce neo-colonial power relations (Cunneen and Rowe 2014). Reflecting on the work of Barbara Hudson, Scott (2015: 14) reminds us that 'rather than being neutral, the law reflects existing discriminatory power relations: the presuppositions of law were male, white and middle class and reflected their material and property interests'. Historically, then, JE endures as a feature of British law, with similar legal principles and histories of collective punishment extending to those overseas territories with colonial ties to the United Kingdom (UK) (Nijjar 2018). The second case featured in the landmark Supreme Court ruling referenced above, Ruddock v the Queen [2016] UKSC 8, originated from Jamaica following the murder of a man in Montego Bay. In 2016 in Zimbabwe, there were reports that JE laws were used to convict three Movement for Democratic Change activists for murdering a police officer (Nyawanza 2016). In another case in South Africa, 270 striking miners were collectively punished for the murder of their co-workers following the fatal shooting by police of 34 striking miners in Marikana (Dixon 2015).
Within these examples, we can detect the use of legal principles of JE to manage groups connoted with a 'special categorisation' through the criminalisation of non-criminal behaviours, both to manage political dissent and to punish collectively those who threaten the political and/or social order. Elsewhere in the world, other forms of collective punishments are being utilised to punish racialised communities, such as the Law of Parties and Racketeer Influenced and Corrupt Organizations Act in the United States (US) (Woods 2012). Similarly, collective border legislation in Palestine denies or restricts access to resources and social goods to those deemed as outsiders (Gostoli 2016).
Critically, what arises from examining these international cases is a political utility of collective punishments, and specifically JE, which serves to demarcate, criminalise and regulate the Other (Spalek 2008). Within this context, JE is an extraordinary device selectively deployed to produce and regulate those presented as threatening the normative boundaries of the state (Aliverti 2016;Spalek 2008).
Only recently have concerns about the use of JE begun to permeate academic debate, with many contributions highlighting its utility to convict gang members (Green and McGourlay 2015;Squires 2016 four types of possible misapplications. The authors draw on a number of hypothetical 'gang' scenarios to demonstrate the: legitimate concern that prosecution policy is casting the net too wide to catch as many gang members as possible, making the task of the judge and the jury even more difficult. (3) In a valuable contribution to the debate, Green and McGourlay (2015) argue that police and prosecution teams can view this legal 'incoherence' as a virtue of JE cases, enabling the lowering of the standard of proof, by which police interpretation of associations can become the central argument in establishing guilt. The matter of gang identification is taken up in an earlier paper by Pitts (2014: 50), for whom 'the central question in the case of JE concerns how the police and prosecution establish who is, and who is not, a "gang member" '. While these contributions seek to understand the relationship between JE and gangs, none examine the significance of defendants' race or ethnicity, or how racialised assumptions are integral to prosecution narratives centring the 'gang'.
We know the vast majority (80-90%) of police-defined gang members throughout England and Wales are racialised as Black (Bridges 2015;MOPAC 2016;Williams and Clarke 2016). While insightful, and pointing to the significance of 'gang crime', these debates have to date been relatively silent on theorising the process of racialisation, overlooking (or not seeing) its significance in mediating the 'gang' narrative in the prosecution of serious violent offences. As with the statistics on race discussed at the outset of the paper, by establishing that racial disparity exists, without explaining how or why this occurs, they risk reproducing the enduring and seemingly infallible construction of the violent Black offender.

The Research Study
In criticising the assumed objective interrogation of official statistics presented herein, a critical interventionist approach was used to guide this empirical study (Clarke, Chadwick and Williams 2018;Hillyard et al. 2004). This method to social research builds upon the principles of being there and further acknowledges positionality, specifically the importance of foregrounding concealed or minority perspectives (Phillips and Bowling 2003).
The findings are primarily based on a semi-structured survey of prisoners who were serving custodial JE sentences. The survey design used an adapted version of a questionnaire developed by Eady et al. (2013), including mainly open-ended questions to understand individuals' characteristics and their experiences of the criminal justice process, with particular focus on prosecution narratives. The survey asked questions about JE prisoners' perceptions of police and prosecution strategies in their case and, specifically, about any reference to gang involvement. It also collected data on personal characteristics (open-ended questions on ethnicity and age at charge and current age, as well as the convicted offence and the geographic location where it was committed).
JE is currently a concealed practice, insomuch as its use is not routinely recorded at any stage of the criminal justice process. Consequently, it is impossible to know how many people are today serving prison sentences as a result of a JE prosecution. Participants in this study were identified with support from the Joint Enterprise Not Guilty by Association (JENGbA), 3 drawing on the list of 'inside campaigners' held on its case systems by February 2015. Paper copies of the survey were posted to approximately 500 prisoners, with stamped addressed envelopes included for return. On receipt, the survey responses were inputted in full and coded and analysed through an iterative process. The data were exported using Statistical Package for the Social Sciences software to support further thematic and statistical analysis. In recognising the particular challenges of engaging in critical social research (Clarke, Chadwick and Williams 2018), care was taken with the data to ensure confidentiality and anonymity, both when handling the data and during the dissemination process.
In the absence of any official recording of JE prosecution, this study is not without limitations. First, due to lack of official data, we were unable to ascertain the extent to which the sample represented the wider JE www.crimejusticejournal.com IJCJ&SD 121 2020 9(3) prisoner population. However, a range of other sources (Crewe et al. 2015;Crewe, Hulley and Wright 2014;Eady et al. 2013;Jacobson, Kirby and Hunter 2016), using different sampling approaches, found similar age, 'race' and ethnicity profiles of JE prisoners. Second, while all efforts were made to contact 'known' JE prisoners, we recognise the potential bias affecting both the initial list (derived from those prisoners in contact with JENGbA) and the response rate. Yet, this research demonstrates the value of such partnerships between researchers and grassroots organisations. While presenting particular challenges to research, such partnerships are necessary for a progressive criminology working towards social justice (Clarke, Chadwick and Williams 2018;Goddard, Myers and Robison 2014).
To the best of our knowledge, this effort represents the largest study of JE prisoners undertaken to date. In total, 241 completed surveys were returned, representing a response rate of 48%-an excellent return for a targeted postal survey.

Findings
The average age of respondents was 27, with almost half (48%) being below 25. This is a significantly younger profile than is found in the wider prison population, with just 18% of all prisoners under the age of 25-a proportion that is decreasing each year (MOJ 2017a). The JE prisoners aged between 18 and 25 were serving sentences with an average length of 20 years. It is noteworthy that 53 prisoners were sentenced to more than 20 years, including 10 young adult prisoners serving JE sentences of over 30 years and one child sentenced to 26 years in total.
In terms of race and ethnicity, rather than imposing the traditional 16+1 census categories, we asked respondents to self-define their ethnicity. This approach resulted in over 60 distinct categories, illustrating how self-definition defies official categorisation and the producing tendency of ethnic categories, as found within official governmental processes and outputs. Collectively, 53% of the sample self-defined as belonging to a Black, mixed-race, Asian or other minority ethnic group, with 46% self-defining as 'white'.
In this paper, we identify those non-white prisoners as the negatively racialised group because, despite forming the majority of our sample, it is their negative racialisation which is significant in the process of criminalisation and punishment.
The majority of prisoners belonging to negatively racialised minority groups were under 25 years old (62% v. 41% for the white group), with all 21 child prisoners (i.e., those under 17 years of age) selfidentifying with this group. In addition, these prisoners were serving on average longer custodial sentences (22.3 years) compared to the white prisoner group (19.6 years).
All but 10 respondents (4%) were serving sentences for murder, with the small remainder imprisoned for Section 18 violence and other violent offences. Self-reported offending history varied with just under a quarter (23%) of respondents having 'no previous convictions' prior to the JE prosecution. Related to this, respondents were asked to indicate the number of co-defendants (Co-D's) who were prosecuted as part of their court case. On average, respondents noted four Co-D's within their case, which, even using a conservative approach, estimates upwards of 660 individuals implicated in the court cases of participants in our sample. Of significance, 100 (41%) respondents disclosed that they were 'not at the scene' of the crime for which they had been imprisoned. Of those 'not at the scene', the majority (70%) also reported 'no contact' with the co-accused ahead of or during the event for which they were serving a custodial sentence. Where respondents claimed not to have been at the scene, it becomes incumbent upon police and prosecution teams to demonstrate how individuals who were not present during or peripheral to the offence committed share a common purpose.
This developed as the central line of inquiry for the study in seeking to reveal the strategies through which the respondents were (co)located or (virtually) placed at the scene, and then prosecuted for the offence. What emerged from the JE prisoner accounts was the centrality of two interconnected and powerful policing and prosecution resources-a criminalising guilt-producing 'gang' narrative, deeply dependent on a range of reproductive racialised constructs of the Black criminal Other.

Constructive Criminalising Associations: The 'Gang' as Signifying Intent and Common Purpose
If a jury sees five young black friends in the dock, they don't see five individuals, but one black gang. The jury might be influenced by racist media stereotypes, and because of this they may find it difficult to separate, apportion responsibility, [and] determine culpability. Seven black kids in the dock immediately becomes a dangerous black gang, because that is exactly the way members of the jury are taught to see black boys by the media.

-Gloria Morrison (Fekete 2012)
Morrison's statement uncompromisingly situates the racialised 'gang' construct as a potential device open for (mis)use by police and prosecution teams in JE cases. In advancing this critical line of inquiry, the study explored the extent to which constructing 'gang' identity was utilised in prosecuting JE cases. To this end, one-third (31%) of respondents stated that references to gangs were not used in their prosecution's argument. A further 8% of prisoners 'did not know', were 'unsure' or 'could not remember'. That prosecution teams evoked the term within the majority of JE cases illustrates the pervasiveness of using this remarkably vague and fluid resource to build a narrative. In fact, we suggest that it is precisely this lack of definition-the malleable nature of the 'gang'-that marks its use as a prosecution strategy. By relying on 'gang talk' (Hallsworth and Young 2008)-a recognisable set of linguistic cues echoed within policy, media and academic discourse, which deftly tie issues of serious youth violence to the unreliable construct of 'the gang'-the prosecution draws on a ready-made narrative to construct the primary association necessary to infer collective intent.
For the majority of respondents (97%), the deliberate and strategic imposition of the gang label was contested, dismissed as untrue or a 'made-up' feature of the prosecution narrative: I have never been in a gang. I was a family man who had a good job. No, I have never been in a gang and I have no previous convictions of being in a gang, and there is no proof that I am in a gang. It's all made up. I don't agree with the prosecution constantly using the word 'gang' because we were not a gang. One was a friend and the other my customer. I disagree with this description. This is because only [two] people out of the group of nine I was in were gang members. I have never classed myself as a gang member. One of my [codefendants] was an active 'gang member' but I was not. I was a friend of a gang member, so I was also judged to be a gang member. They said we were a drugs gang, but I only knew one of my [codefendants]. I didn't even know the alleged shooter before my arrest. No link to him whatsoever. We knew each other from school and two of my [codefendants] I'd never met.
Five respondents indicated that they were or are 'gang members'. However, even here, while retrospectively applied, the respondent deemed his gang status as historical and unrelated to the offence for which he was serving his current custodial sentence: I went to prison at 18 and got out at 23 for robbery, and during that time I realised the socalled gang wasn't what I thought it to be. You find out who your real friends when you go jail. So when I got out I was no longer involved.
While many refute or resist the imposition of the 'gang' label and its relevance to individuals' present custodial sentence, respondents' statements converge as a counter-narrative seeking to explain, or rationalise, how the prosecution may have been able to evoke the 'gang' in their court case. For some, explanations centred on one's proximity to family and friends who had previously been defined (by police) as a 'gang' suspect. These associations consequently emerge through reconstruction of non-criminal www.crimejusticejournal.com IJCJ&SD 123 2020 9(3) behaviours as 'gang related' due to the proximity or relationship defendants share with those who are 'suspect' by the police and their partners. This apparent vagueness in how gang associations are established in courtrooms reflects the imprecise nature of wider policy, practice and academic definitions of 'gangs' (Fraser and Atkinson 2014;Katz, Webb and Schaefer 2000)-one that statutory documents trace as increasingly (and intentionally) permissive and 'flexible' (GOV.UK 2015). Given this situation, deploying the 'gang' construct within a court as evidence, or more appropriately as a guilt-producing strategy, might be viewed as a highly contentious and controversial practice. In their analysis sponsored by the Home Office, Gunnell, Hillier and Blakeborough (2016) highlight the gang-producing tendencies intrinsic within police attempts to build 'networks' of gang nominals (suspect individuals). Developed from Greater Manchester Police 'intelligence' on five 'gang members', a social network analysis web was produced, reflecting an 'urban street gang network' of 137 nodes or nominals. Importantly, of the 137 individuals, a staggering 115 (84%) were connected not through criminal but 'other social links' referred to as 'family', 'friendship' and 'romantic' links. Given the report acknowledged no 'criminal', 'drugs' or 'gang' links, they still received police (and subsequently academic) attention due to their proximity to the (social) network of the five 'gang' members. Clearly, policing responses to the 'gang' based on such methodologies will drive social regulation, not only of those who are perceived to be gang members, but also the family, friends and unsuspecting community members of so-defined individuals: We were just friends, normal working teenagers. I was brought up with the same group of people through school to holidays with family; we were very close and always together so the prosecution found it easy to call us gang members. I don't agree, as we're just neighbours! I have a little bond with the brothers, as they helped me out in school from bullies, etc., and one was going out with my cousin for a short period.
These multiple accounts represent different JE cases prosecuted in different cities and towns across England, with 33 different police force areas cited by respondents. Yet, there is striking convergence in the similarity of the language and strategies deployed within the court arena. Respondents' accounts begin to illustrate how everyday, non-criminal encounters and associations with friends, family within local neighbourhoods, and communities become manipulated and criminalised through the construct of the 'gang'.
By utilising the gang as the primary form of association, prosecutors serve to symbolically communicate to the sentencer and, more importantly, the jury a story-one that consistently bears the hallmarks of the myriad contemporary mediated construct that 'urban gangs' are responsible for serious violent offending (Williams and Clarke 2018a). This is regardless of the empirical evidence from Manchester and London that demonstrates how little of the serious youth violence recorded in these cities is 'gang-related' (Williams and Clarke 2016). Analysis of police data in London reveals that less than 5% of serious violence in the country's capital is recorded as 'gang related', a figure that has remained stable for at least three years (London Authority Police and Crime Committee 2016).
Importantly, our analyses revealed that the prosecution was more likely to use 'gang talk' where the defendant was from a negatively racialised minority group. In this regard, over three-quarters of nonwhite prisoners (79%) reported that the 'gang' narrative was a feature of the prosecution's argument in their court case. Comparatively, less than 39% of the white group noted use of the 'gang' in this context. The study found that, statistically, prosecution teams are more likely to use the term in JE cases involving young Black and mixed-race people. 4 By further examining the prosecution strategies used in courtrooms to connect the policing of gangs in JE cases, intrinsic to this endeavour is the deliberate use of racialised signifiers that littered respondents' accounts and experiences. Next, we identify these as reproductive associations, acting to produce and impose criminalised traits to justify the application of JE laws. Such racialised associations include www.crimejusticejournal.com IJCJ&SD 124 2020 9(3) linguistic cues, place-based euphemisms and non-criminal cultural artefacts, delineated as intrinsically criminogenic and unceremoniously communicated within the court arena.

(Re)producing Racialised Associations: Exposing the Effects of Racialisation as Signifying Criminality
In support of the criminalising associations constructed through the broad yet effective strategy of evoking 'gang talk', many respondents disclosed a number of specific strategies through which prosecution lawyers communicate the gang narrative to a judge and jury. Interlinking these processes of criminalisation with the historically significant resource of racialised constructs of criminality creates a particularly powerful narrative in courtrooms. In some cases, defendants reflected how the presiding judge actively dismissed unsubstantiated references to gangs, yet once implied the gang association develops using a range of other racialised signifiers.
The accounts reveal a predominance of linguistic cues or 'dog whistles' often particular to the local area in which a trial was underway. Here, we capture two central cues-reference to neighbourhood and use of historic and mediated gang names, extending to the dehumanisation of groups of defendants by applying animal references to them as a collective. Alongside this is a range of police intelligence sources that also enter the courtroom to support these narrative devices. Increasingly, these draw on new technologies and social media. Yet again, their disproportionate use with negatively racialised defendants reveals further the reproductive nature of the particular surveillance and policing of suspect and racialised communities that drives the collection of police intelligence, and emerges as a rich prosecution resource in the courtroom.

Neighbourhood and Spatial Euphemism: Reproducing Suspect Communities
Prisoners reported that references to the 'area' or community were a key feature of the criminalising gang narrative used in JE cases, whereby proximity seemingly affirmed their gang-suspect status. As one respondent argued, 'just because we are from the same area and are of a certain colour does not make us a gang'.
Often, police and prosecution teams made repeated references to imagined, ethnically heterogeneous neighbourhoods (e.g., Moss Side in Manchester or St Ann's in Nottingham) to misleadingly make inferences of criminal tendencies: 'St Ann's, where we are from, has this reputation. The term 'St Ann's' was used to group us together'.
In their work on the effects of racialisation in the co-construction of crime and poverty narratives in the US, Loyd and Bonds (2018: 906) demonstrate how geographically located communities, defined nationally by zip codes, become a 'spatial euphemism … regularly invoked by news media, policy makers and residents alike as shorthand for racialized crime and poverty'. The significance of the historical and structural relations of racism is effectively erased from such narratives, with the stigma falling squarely on the community. In these cases, to be blemished and stigmatised as being present in, or a member of, the 'community' legitimates the imposition of a 'gang' association, driving over-policing and collective punishment, in turn.
In this context, residents' simple movement around the local area-their everyday encounters and interactions-can become part of the courtroom narrative to demonstrate collective intent. Pitts (2014: 51) similarly expresses concern for those individuals 'who are uninvolved in gang-related illegalities', yet are drawn into prosecution narratives in JE cases. For example, reporting how associations such as being on 'nodding terms' or 'playing Sunday league football' together can result in becoming codefendants on a JE murder charge: www.crimejusticejournal.com IJCJ&SD 125 2020 9(3) We are a group of young lads who smoke weed and fuck around, and we get labelled a gang! To me, a gang is a group of mates, but the prosecutor made it sound as if we are a gang walking around with weapons protecting the area.
Sitting alongside the strategic and linguistic use of 'area names' was reference to discourse that sought to infer the area claim-making of space for those accused in JE cases. The language of 'turf' or 'territory' was drawn upon to signify secondary associations to drug dealing, again cementing within courtrooms the popularised and stereotypical construction of 'gangs'.
New Chapter, Old Story: Associating to 'Gang' Folklore The sociopolitical construction of the 'urban street gang', echoed over time by a range of media sources, becomes a rich resource in the courtroom (Williams and Clarke 2018a). The layering of such discursive strategies included references to 'gang' names, often localised depending on where the trial takes place, and powerfully synonymous with particular criminalised (and racialised) communities: 'They said we was Gooch 5 [gang name], but I ain't no Gooch member and I wasn't even there.' Respondents who lived elsewhere in England disclosed references to other locally recognisable gang names-for example, the 'Burger Bar Boys' from Birmingham. More than one respondent from different cities noted metaphorical reference to the 'Kray Twins'. In naming a 'gang', regardless of any substantive evidence or clear commitment by police to confirm individuals' association to a so-called gang, the prosecution simplistically situates the accused within the continuum of gang stories and folklorenarratives shaped and amplified over time by local news media (Gutzmore 1983;Williams and Clarke 2018a). In addition, respondents revealed the appropriation of 'gang-speak' to elicit mediated and imagined constructs of violent collectives such as 'crew', 'hoodlums', 'click', 'soldiers', 'troops' or 'posse'. Finally, the ultimate metaphorical signifier used in six different cases was to dehumanise the accused group with reference to 'animals baying for blood' or 'a pack of wolves' or 'a pack of animals'. Such references to groups prosecuted under JE as being less than human, notably through repeated use of 'the wolf pack' to describe groups, have been highlighted in other reviews of legal cases (Green and McGourlay 2015). Collectively, the location-claiming, naming and dehumanising cues powerfully serve to shape a collective purpose-including those who are peripheral to events-in prosecution narratives that seek to collectively punish.

Police Intelligence: Objectifying the Subjective
Not all individuals who reported that 'gang' was invoked at trial were from negatively racialised communities, yet most (70%; 101 of 144 individuals) were. Thus, understanding the significance of this in JE prosecutions is crucial. Comparative analysis of these cases suggests that distinct strategies are employed for defendants racialised as Black. Self-identified white prisoners (32%) were nearly twice as likely to express that while the 'gang' had been part of the prosecution narrative at trial, the prosecuting team brought 'no evidence' to demonstrate their associations. This was the case for only 17% of non-white respondents.
The use of expert witnesses was a key strategy, reflecting how these voices in court are able to transpose and infer fact from what is in reality opaque and subjective local police intelligence. This included police officers taking the stand, as well as other experts in 'face-mapping' and 'cell-site' analysis. Inferred here, such 'witness' testimonies are intrinsically reliant upon the existence of particular 'crime-control' tools and strategies being available and in place in the local community, including CCTV cameras, mobile phone masts, an operational police 'gang list' or team, and higher levels of stop and search. In light of the racialised basis of the 'gang', as employed in England and Wales, and the government resources that accompany police-defined gangs (Amnesty International 2018), it is significantly more likely that such complex devices are made available within racialised communities to facilitate gang-making (Fraser and Atkinson 2014). These policing resources, repeatedly drawn on by the prosecution teams to (re)construct associations and locate peripheral individuals to a crime scene, were coded and classified in our study as www.crimejusticejournal.com IJCJ&SD 126 2020 9(3) 'police intelligence'. Such methods were present in over half of cases involving negatively racialised defendants, compared to use in one-quarter of white cases.
The shaping of such intelligence will inevitably benefit from the positioning of distinct police gang teams in racialised communities, such as Trident in London and Xcalibre in Manchester. 6 The over-policing of young people within these communities (Amnesty International 2018; Williams 2018)-regardless of lack of proven criminal involvement or, importantly, any independent factual verification of the intelligence sources presented in court-becomes a rich resource for the prosecution seeking to associate defendants with criminality and connect them to events of the case.

Association Through Imagery and Music: Criminalising Culture
A final, and further racialised, feature of the prosecution strategies reported by prisoners was the incorporation of images and videos of defendants engaging in non-criminal behaviours, which, through popular culture, become indicative of criminal involvement and affirm a presumption of criminality. This included photos in which individuals were holding hands in a particular form, revealing tattoos, dressed in a particular colour or referring to music groups. Playing certain music videos or reading out song lyrics in court, particularly from 'hip-hop', 'grime' and 'rap' genres, formed a resource for building criminalised associations against negatively racialised groups and individuals. Such strategies were evident in 11% of cases against these defendants, compared to less than 2% of cases involving white respondents. Significantly, where prosecution evidence included such police intelligence sources, including youth culture references to 'gang insignia', the individual was more likely to report 'not being at the scene'.
In recent years, a series of hysterical media outbursts occurred on both sides of the Atlantic, framing social media and forms of rap music as catalysts for violence (Kubrin and Nielson 2014). Such linking of culture to violence is inherently racialised, with the most recent discussion in the UK focused on 'drill' music. This resulted in video content being stored on police databases as intelligence and in some cases removed from YouTube (Dearden 2018). In the US, examination of the specific ways in which prosecutors use music lyrics as evidence of 'a defendant's knowledge, motive, or identity with respect to the alleged crime' (Kubrin and Nielson 2014: 186) reveals how devastating such practices can be. This analysis is particularly relevant in JE trials in which the onus is on prosecutors to establish collective foresight and defendant intent.
What has emerged alongside this debate is the investment of police resources to trawl social media for images and content. These 'fishing' activities generate vast catalogues-with the Metropolitan Police reported to have a database of 1,400 videos (Waterson 2018)-and evidence suggesting such content is not always gathered using appropriate legal protections for human rights (Amnesty International 2018). Critical analysis must examine the cumulative effect of these practices, where police databases of suspects and surveillance are unaccountable yet authorised to criminalise and punish through such powers as JE.

Conclusion: Disrupting Dangerous Associations
Upon analysing JE prisoners' voices, we detect the symbiotic (re)production of the 'criminal Black Other' (Williams and Clarke 2018b), criminalised and punished in JE prosecutions through the 'gang' construct. These strategies, through which the non-criminal identity of secondary parties becomes blemished, impose stigma by attributing racialised criminal markers upon those in the periphery (Howarth 2006). These descriptions of courtroom strategies have long been precipitated by defaming those communities within which many of the research respondents live (Gilroy 1982;Howarth 2006).
This article presented evidence of how racialised constructions of the 'gang' drive distinct and punitive responses that accelerate over-representation of young Black and mixed-race men in JE convictions, not offending behaviour. It is the array of criminalising associations that dangerously conspire to represent negatively racialised young people and their communities as a distinctive 'special category' posing particular 'risks' to be managed through intrusive criminal justice practices.