A ‘good job’ in difficult conditions: Detectives’ reflections, decisions and discriminations in the context of ‘joint enterprise’

‘Joint enterprise’ is described as a ‘dragnet’ that draws disproportionate numbers of black and minority ethnic young men into the criminal justice system in England and Wales. While stereotyping by the police and prosecution has been blamed for this distributive injustice, empirical research on joint enterprise is limited. This article presents the findings from a study of homicide and ‘gang’ detectives in London in which they rebut accusations of racial stereotyping when investigating multi-handed crimes. Instead, they claim that the disproportionality reflects the involvement of larger numbers of primarily black men in violent crime. Using Margaret Archer’s social realist theory, detectives frame their actions as being driven by their ‘ultimate concerns’ to do a good job of protecting the public and obtaining justice for victims within the difficult social and cultural context in which they operate. However, in this article, we expose the racialized notions of risk in detectives’ narratives and argue that such ‘colour-blind racism’ is likely to contribute to young black and mixed-race men being overrepresented in cases that draw on the principles of joint enterprise dragging those on the periphery of group violence into the criminal justice net.


Introduction
Joint enterprise is the term used to describe a complex set of legal principles in English and Welsh law that allows more than one individual to be convicted of a single criminal offence. For example, a person can be prosecuted for a crime committed by another if it can be proven that they 'intended to assist or encourage' the principal offender in the substantive offence (this is the principle of secondary liability) (Crown Prosecution Service, 2018). Joint enterprise allows the criminal justice net to be cast widely capturing individuals who do not perpetrate the substantive offence. For example, multiple individuals have been charged and convicted with homicide, despite some not being present at the scene (see R v Walters and others [2017] and R v Wright and others [2017]). 1 As a legal doctrine, joint enterprise has come under a significant amount of criticism from practitioners, academics, grassroots campaigners and those convicted of 'joint enterprise murder' (Hulley et al., 2019). In particular, this critique has focused on the apparent disproportionate application of the doctrine in cases of serious violence involving black and minority ethnic young men (BAME) (Crewe et al., 2014;Hulley et al., 2019;Williams and Clarke, 2016), which in part has been blamed on racial stereotyping by the police (Williams and Clarke, 2016). While research highlighting discriminatory attitudes and actions among police officers towards BAME groups is well established (Bowling and Phillips, 2007;Hall et al., 1978), to date, no research has been undertaken of police investigations that draw on the principles of joint enterprise. In this article, we provide a nuanced analysis of the work of police detectives in this context to explore the ways in which their investigative practices may contribute to the disproportionate impact of joint enterprise on black and mixed-race men.
Using empirical data from interviews with police detectives, conducted as part of a study of multi-handed serious violence among young people, we report on officers' rebuttals of accusations of racial stereotyping. Drawing on Archer's (2003Archer's ( , 2010Archer's ( , 2012 social realist theory, we explore how detectives explain disproportionality by reflecting on their attempts to achieve their 'ultimate concerns' in the 'difficult' structural and cultural conditions that they face. Specifically, we highlight detectives' claims that they are attempting to 'protect the public' from the risk of violence, in the context of rising levels of serious violence driven by group offending, which they argue disproportionately involves black men as perpetrators. However, we also show that entrenched in their discussions are racialized notions of risk, supported by a 'gang' narrative, which provides a clear and consistent narrative to frame multi-handed violence as primarily involving black men. Rather pessimistically, we predict that, given this construction of multi-handed violence, the observed disproportionality is likely to continue-due to the 'contextual continuity'

Joint enterprise and disproportionality
The legal doctrine of 'joint enterprise' allows parties to a crime to be held equally accountable if the crime is considered to have been committed as part of a collective action. While no data are systematically collected on the extent to which joint enterprise is applied in practice, figures published by the Crown Prosecution Service (2019) over two years (2012)(2013) showed that, on average, around 130 cases of homicide involved multiple individuals per year, with around 3.5 defendants per case. During this period there were on average 504 homicides per year (ONS, 2019), suggesting that multihanded cases represented over a quarter of all homicide incidents. The Bureau of Investigative Journalism (2014) estimated that during this time and in the years before (between 2006 and 2014), nearly 2000 people were prosecuted for homicide using the principles of joint enterprise.
Since these figures were collected there has been a change to the doctrine. In R v Jogee and Ruddock [2016] UKSC 8, the Supreme Court ruled that the law had taken a 'wrong turn' by misinterpreting the doctrine for more than 30 years. Prior to 2016, the standard of proof required to convict a secondary party was 'foresight', so that to be convicted of the substantive offence the secondary party must have foreseen the possibility that the principal offender 'might act as he did' (R v Jogee and Ruddock [2016] UKSC 8). Therefore, the threshold for guilt was lower for the secondary party than for the principal party who, to be found guilty of murder, would have had to 'intend' to cause serious harm or death. Under these terms, joint enterprise was branded a 'lazy law' (Krebs, 2010), as it allowed the police and prosecution to focus on proving association between the parties involved rather than culpability for the offence itself (Green and McGourlay, 2015). In R v Jogee the route to conviction using the principle of foresight was abolished. While foresight alone is no longer sufficient to convict a secondary party of the principal offence, secondary parties can still be convicted of the principal offence if they 'intentionally assist or encourage' the principal offender to commit the criminal act, and foresight can still be used as evidence of 'intention to encourage or assist'. Despite the changes in the law, criticisms of joint enterprise continue, focusing on the legal legitimacy (i.e. the fairness of the law itself) (Hulley et al., 2019) and, particularly, the distributive injustice of the doctrine. 2 While few research studies exist, those that have been conducted indicate a disproportionate number of BAME men are serving prison sentences for offences using the joint enterprise doctrine. In a study of prisoners serving long life-sentences (received when they were young), three times as many 'joint enterprise prisoners' were black/black British or mixed race, compared to the general prison population; and 11 times as many prisoners were black/black British compared to the general population of England and Wales (Crewe et al., 2014). Similarly, a survey of men convicted under the principles of joint enterprise found that over half (53.1%) self-identified as belonging to a BAME group (compared to 45.6% white) (Williams and Clarke, 2016). Significantly, BAME prisoners were more likely than their white counterparts to report that the term 'gang'or 'gang-associated' 'cues and signifiers' (e.g. street names)-was evoked by the prosecution during their trial (Williams and Clarke, 2016: 15). 3 This led Williams and Clarke (2016: 3) to conclude that the disproportionate application of joint enterprise principles to cases involving BAME men was the result of 'racialized stereotypes that [. . .] underpin contemporary policing and prosecution strategies'.

Theorizing the structural and cultural context of detective work
Detective work is a case-focused enterprise (Bacon, 2016) in which officers are motivated to gather irrefutable evidence for use by the prosecution in efforts to 'clear up' those cases for which they are responsible. As McConville et al. (1991) note, prosecution cases are to be understood as social constructions wherein the police are involved not only in the collection of evidence, but also the creation of evidence for use in court. In the construction of a case the 'detective is in the process of transforming an individual event into categories which have a character of permanence and exactness', in which the role of the police is to distil complex information into simple artefacts of instrumental use by the prosecution (McConville et al., 1991: 13). These constructions, presented as 'realities' and 'truths', are difficult to rebut, especially if formed from enduring and negative stereotypes. In theorizing the work of police detectives, and specifically in efforts to understand how discrimination influences investigations into serious violence, it is essential to consider not only how the police construct cases but also the structural conditions within which detective work takes place as police actions 'cannot be divorced from the context in which they are situated' (Foster, 2008;Innes, 2003: 15). Margaret Archer's social realist theory allows for a nuanced analysis of the impact of structure and culture on police officers' agency. Archer (2010) describes how the structural and cultural context conditions individual action. Rather than structural conditions being objectively enabling or constraining, Archer (2010: 282) argues that actors reflect on the extent to which the conditions they operate under facilitate or prevent their quest to achieve their personal goals or 'ultimate concerns'. Thus, the conditioning influence of the structural/cultural context [. . .] works through shaping the situations-from the accessibility of resources to the prevalence of beliefs-in which agents find themselves, such that some courses of action would be impeded and discouraged while others would be facilitated and encouraged. (Archer, 2010: 277, emphasis added) To illustrate this point, Archer (2003) highlights how a reduction in income, caused by a structural rise in inflation, impacts upon an individual's spending power. In this situation, a person is potentially 'worse off' so they enter into intensive deliberation on their objective concerns (i.e. how to make ends meet) in line with the shifting structural conditions (i.e. loss of employment) and act accordingly. To fully comprehend the individual's actions, an appreciation of 'the agent's project in relation to her social context ' (2003: 131) is crucial. Consideration must be given to the reflexive deliberations of the individual when determining what matters most (their 'ultimate concerns') and the structural and cultural circumstances they confront 'involuntarily ' (2003: 131).
According to Archer (2003), the relationship between structure and agency is cyclical. Structural and cultural properties condition individual action which is, in turn, transformed or preserved. Where the structural properties shape an agents' actions in a similar way to previously, morphostasis results, characterized by 'repetitive situations, stable expectations, and durable relations-and with them, habitual action' (Archer, 2012: 60). Where individual action is modified in response to the structural conditions, morphogenesis (structural and cultural change) results. In this context, 'in place of habitual guidelines, subjects become increasingly dependent upon their own personal concerns as their only guides to action' (Archer, 2012: 65).

The structure, culture and police 'ultimate concerns'
The criminal justice system in England and Wales has undergone significant transformation in the movement towards neo-liberalism. For the police, this shift has introduced changes in officers' material existence and their cultural 'order', requiring the prioritization of actuarial strategies that focus on the identification and management of 'risky' populations (Feeley and Simon, 1992) and the achievement of specified targets (de Maillard and Savage, 2018). This cultural change has created work pressures that have been further exacerbated by a 20% reduction in police funding since 2010 (HMIC, 2013) at a time when violent crime has increased by 132% (in the five years between 2013 and 2018) (ONS, 2019). Incidences of knife crime in London are 20.5% higher than in 2016 (Greater London Authority, 2017: 11) and homicide is at its highest level since 2008 (with 726 homicides recorded in the year ending March 2018) (ONS, 2019).
In this context, policing in the broadest sense can be considered as a 'forceful reaction to conspicuous signs of disorder whether or not of a criminal nature', and necessary to protect and serve the citizenry through the control, deterrence and punishment of identified offenders (Bowling et al., 2019: 24). And although some scholars have argued there is no monolithic or homogenous 'cop culture' (Waddington, 1999), the constant vigilance, social isolation and repeated exposure to dangerous experiences generates a police personality that is said to be suspicious, cynical, thrill seeking and prejudicial; one that labels suspects as 'respectable' or 'rough' (Reiner, 1992: 117), as assholes (Van Maanen, 1978), scumbags (Ilan, 2016) or 'dross' (Choongh, 1998). In the current context of rising rates of murder, and acting on a heightened sense of mission to 'get the bad guys' and protect the public from 'evil forces' (Bacon, 2016: 119), 'gangs' and those identified as 'gang members' represent the so-called folk devil par excellence (Hallsworth and Young, 2004).
While gang-related violence constitutes a small proportion of violent offences it has been identified as 'the driving force behind some of the most serious violence' in London (MOPAC, 2018: 4). Gang members are purportedly responsible for half of all knife crime with injury in the capital (Centre for Social Justice, 2018)-of which 57% are very serious or fatal-60% of all shootings with injury and 30% of all homicides (MOPAC, 2018: 10). In light of the explicit association made between serious violence and 'gangs', official responses have been targeted towards eradicating 'gang violence'. For example, gang injunctions introduced in the Policing and Crime Act (2009) impose geographic and relational restrictions on persons involved in 'gang-related' activities and since 2011 suspected gang members and associates (including victims of serious violence) have been placed on the Gangs Matrix database and subject to surveillance (Amnesty International, 2018; Williams and Clarke, 2016). 4 Similarly, it has been argued that joint enterprise, a 300-year-old common purpose doctrine, has been resurrected 'as a matter of policy in the course of dealing with the 1990s gang panic' (Squires, 2016: 942).

'Gangs', ethnicity, risk and joint enterprise
The almost symbiotic association between gangs, violence and ethnicity that emerged from the late 1980s onwards has, according to Smithson et al. (2013), compounded the view of the BAME population as 'dangerous' and a particularly 'high risk' to the public. Just like the 'black mugger' in the 1970s, illustrated in Hall et al.'s (1978) seminal work Policing the Crisis, the manifestation of 'gang, gun and knife crime' in the contemporary context has served to justify the punitive policing of black and ethnic minority groups in an attempt to 'do something' about the proliferation of street violence. Official statistics have been 'provided [to add] scientific credibility to beliefs that were once anecdotal' (Van Cleve and Mayes, 2015: 420; see also Hall et al., 1978). Indeed, official statistics have recently reported large increases in black males aged 16-24 as victims of knife enabled homicide (ONS, 2019) and disproportionate numbers of young black men have been identified as 'offenders' of violent crime in London (MOPAC, 2018). While such disproportionality might simply reflect the greater 'availability' of young black men in the population of offenders (Waddington et al., 2004) it also gives weight to the "'black-on-black violence" stereotype (label)' (Gabbidon, 2010: 108) that has proliferated through time. Political calls to recognize that 'black kids are killing black kids' and calls for the 'black community to mobilise in [the] denunciation of a gang culture that is killing innocent young black kids' (Blair, 2007: 10) have done much to (re)ignite entrenched stereotypes of the associations between particular types of criminality and black young men, in particular. A recent example of the continuation of this political rhetoric linking gangs to serious violence can be found in David Cameron's assessment of the English riots in 2011: At the heart of all the violence sits the issue of the street gangs. Territorial, hierarchical and incredibly violent, they are mostly composed of young boys, mainly from dysfunctional homes. They earn money through crime, particularly drugs, and are bound together by an imposed loyalty to an authoritarian gang leader. They have blighted life on their estates with gang-ongang murders and unprovoked attacks on innocent bystanders.
That the gang was incorrectly identified (see LSE and the Guardian, 2011) as orchestrating these riots was not coincidental; it was in keeping with, and followed, a political agenda to tackle serious gang violence outlined in the 2007 Tacking Gang Action Programme (Home Office, 2008). Alongside the 'new penology ' (Feeley and Simon, 1992) with its focus on risk, the gang narrative 'forms an irrefutable discourse that provokes crime control responses that allow for the identification and management of problematic behaviours and communities' (Smithson et al., 2013: 117). 5 Nowhere is this sentiment more clearly expressed than in a statement by Lord Falconer in support of joint enterprise convictions for 'gang-related' offences: The message the law is sending out is that we are very willing to see people convicted if they are 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 that killed?' And I think broadly the view of reasonable people is that you probably do need a draconian law in that respect. (Lord Falconer, cited in Jacobson et al., 2016: 15) The reference to public opinion in Lord Falconer's statement betrays a utilitarian philosophy that foregrounds the protection of society over that of the individual and negates due process in favour of a crime control model of justice in which the incarceration of the secondary party, in a gang-related offence, is simply collateral damage. Central to this argument is the theory of shared intention of the 'plural subject' (Gilbert, 2000, cited in Amatrudo, 2016 or what Bratman (1984) might call 'we action': simply, that a joint venture imposes a moral commitment on the event suggesting that people are obligated and fully committed to the collective action (Gilbert, 2000, cited in Amatrudo, 2016. Together, the embedded nature of 'gang talk' in criminal justice discourse and public policy (Fraser et al., 2018;Hallsworth andYoung, 2008, 2010) and the term 'gang' not being race neutral facilitates storylines of collectivity that potentially criminalize all people of colour (Alexander, 2008). Thus 'the uncritical and arbitrary employment of the gang label, particularly within a risk paradigm' (Smithson et al., 2013: 113) pathologizes and criminalizes the urban black male in wider social culture (Stuart, 2019) leading to the incorrect assessment of black men as criminal (Gabbidon, 2010) all in an effort to 'protect the public' (Williams and Clarke, 2018: 1).
The framing of serious violence as gang-related violence could be viewed as a tactical political ploy initiated by the police under pressure from central government to fix the problem of violence (Gilroy, 2002;Reiner, 2010) and as a response to public scrutiny in the face of rising rates of serious violent crime. In what remains of this article we examine the narratives of homicide and 'gang' police detectives and interrogate their reflections, decisions and discriminations when investigating collective violence of a serious nature. Here we consider the extent to which detectives' perception of risk frames and shapes the way in which they work towards achieving their ultimate concerns in the structural and cultural context of policing violent crime in London.

The study: A note on method
This article draws on the findings of a larger study examining serious youth violence, friendship and legal consciousness in the context of joint enterprise. The study involved semi-structured interviews with four groups, although for this article the focus is on those with police officers only. 6 Nineteen detectives from the Metropolitan Police Service (MPS) were interviewed, drawn from Homicide Crime Command (12) and Trident Gang Crime Command (7)-the latter of whom were responsible for tackling 'gang-related' crime, including non-fatal shootings.
Access to this group was negotiated through senior MPS staff. 7 A senior officer in each of the Homicide and Trident units sent an email to their teams and officers volunteered to participate. This technique of purposive sampling generated 20 volunteersalthough when contacted, one officer did not respond. Therefore, 19 officers were interviewed, of whom 17 were male and two were female. Most officers self-identified as 'white' (16); two as 'mixed ethnicity' and one participant's ethnicity was not recorded. At the time of their interview, all but two of the participants were serving officers 8 and occupied the ranks of Detective Constable (three); Detective Sergeant (five); Detective Inspector (six); Detective Chief Inspector (two); and Detective Superintendent (three). 9 Nearly all had served longer than 10 years (17)-the remaining two had served between seven and 10 years.
Interviews lasted between one and three hours; all were audio-recorded and professionally transcribed verbatim. The interviews included questions related to officers' professional histories; the process of investigating cases of serious 'group-related' violence involving young people; understandings and perceptions of joint enterprise and the law related to serious violence; insights into the relationship between young people, violence and friendship; and perceptions of young people's understandings of the law. The data were analysed on a computer software package (NVivo) using an iterative approach. This is a reflexive process, whereby the data are repeatedly revisited to develop and refine meaning and understanding (Srivastava and Hopwood, 2009). Interestingly, and fitting for our work, this task has been described as a process of detective work (Patton, 2002).

Disproportionality by default, not design
Despite scholarly arguments to the contrary, detectives in the study argued that research indicating disproportionality in the numbers of black and mixed-race prisoners convicted using the principles of joint enterprise was not due to 'racialized stereotypes' in policing and prosecutorial practice (Williams and Clarke, 2016), but was the result of higher numbers of BAME young men involved in serious violent crime. Indeed, detectives stated that the MPS was more racially sensitive than previously, primarily due to lessons learnt from the Stephen Lawrence Inquiry (Macpherson, 1999): '[t]he Met is a better place than when I joined it. We've had lessons taught to us as a result of things like the Lawrence Inquiry, the Stephens' report and that sort of thing and, yes, we've evolved' (Eric, Detective Superintendent). Therefore, while detectives recognized that prior to the publication of the Stephen Lawrence Inquiry, there were 'major issues' with racism and discrimination in the service, it was their view that, as an institution, the police were now 'generally better professionals than 20 years ago' (Connor, Detective Sergeant) and did not 'pick on people because of the colour of their skin' (Jeremy, Detective Constable). The extent to which such narratives were due to a legislative requirement to be 'sensitive to issues of diversity' (following the Stephen Lawrence Inquiry) or a cultural shift in attitudes towards BAME communities is unclear. 10 Rather, detectives consistently framed violent crime as disproportionately involving black men, in particular, as victims and perpetrators: Sadly, the majority of the victims and the suspects involved in our crime are from the black community.
(Eric, Detective Superintendent) We're talking about young black men, are statistically more likely to be both involved in and victims of knife crime. (Keith, Detective Chief Inspector) Reflecting Left Realist arguments, which locate the causes of crime in social exclusion and relative deprivation (Lea and Young, 1984), detectives reasoned that such offending and victimization was rooted in marginalization. Officers noted that it was 'an unmistakable fact that violent crime, certainly gun crime, and a lot of knife crime, disproportionately affects the poorest communities in London many of whom [are] the black community' (Eric, Detective Superintendent). Recognizing the political sensitivity of their claims, a small number of detectives called for the police, politicians and the public to 'face up' to the 'fact' that much of the violence seen over the last decade involved young black men, and to 'deal with it in an honest, transparent, and justified way' (Davinson, Detective Chief Inspector): Let's have the conversation on who's being killed, then, through knives and that sort of stuff. Because if it disproportionately affects one group, do we not need to target that group? Is that not using the intelligence and the information that you have at your disposal? I remember Paul Condon saying back in the 90s that most robberies in London were committed by young black men. And it was a fact, but he was absolutely castigated. That's where I'm saying we need to have grown-up conversations about criminality. So, if you look at paedophiles, they're generally white middle-aged men wearing glasses. If you look at burglary, again, quite a white youth . . . But if you look at knife crime . . . If that affects a particular proportion of the community then we have to talk about it, we have to have that conversation.
(Vincent, Detective Inspector) While detectives primarily drew on their experience, police statistics on serious kniferelated violence and homicide in London show half (49%) of knife crime victims were BAME individuals (MOPAC, 2017: 90) and, according to the statistics on homicide in England and Wales in 2018 (ONS, 2019), one quarter (25% = 70) were black, of whom 41 (58%) were young (aged 16-24 years old). The majority of these incidences (71% or 50) were recorded by the MPS (ONS, 2019: 14). 11 Although these statistics appear to corroborate the officers' views, the limitations of such data must be recognized given the imperfections of recording and reporting practices, arguments regarding the prejudicial nature of the criminal justice system (Lammy, 2017) and the relative paucity of data correlating suspect characteristics and crime type. 12 It is also noteworthy that Williams and Clarke's (2016) analysis of Mayor's Office for Policing and Crime (MOPAC) data illustrated that 50% of individuals accused of serious youth violence in London were white.
Additional problems with data make robust deductions about the current level of violence and its (dis)proportionate impact on ethnic populations difficult (including outdated population data 13 and the underreporting of violent crime). It therefore prohibits our ability to conclude, with certainty, whether officers' assertions about violent crime in London are an accurate reflection of the situation. In the context of their experience, detectives resolutely argued that youth violence involving young BAME men was what 'us as police officers are trying to catch up with' (Connor, Detective Sergeant). Drawing on discourses of risk and protection, detectives claimed that they were working towards their ultimate concerns to protect young black men by attempting to prevent the violence that disproportionately affected them. For the detectives in the study, the disproportionate impact of joint enterprise on young black men was reflective of the contextual reality rather than racism.

Police detectives' ultimate concerns and the subjective project
According to MOPAC (2017), the highest priority for police is to ensure the safety and security of citizens, while previous research suggests that homicide detectives are concerned 'to crack the case' and 'solv[e] the murder' (Foster, 2008: 100). Detectives in the current study reported their 'ultimate concerns' (Archer, 2003) as twofold: 'justice for victims' and 'public protection':

What would you see as the most important aspect [of your work]?
This sounds really cheesy, but it is the service I guess, the service of two-fold: removing baddies from the streets and providing some sort of justice, closure, whatever you want to call it to the victims or, in our case, the victims' families being, yes, the main driver I suspect.
(Anne, Detective Inspector) Justice for the victim meant getting 'everybody that's involved [. . .] to accept responsibility for the part they played in the offence' and ensuring 'all those responsible' were 'convicted appropriately' (Marcus, Detective Inspector). Officers reasoned that victims (in cases of non-fatal violence) and victims' families wanted, nay deserved, a thorough investigation, convictions and appropriate prison sentences: [The victims' families justice goals]: I think they just want to know that someone is not walking the street, allowed to have an ordinary life, to go to the chicken shop, to go have a drink, to sit with their parents, whatever it is; they just want to know that something has been done.
(Jeremy, Detective Constable) Getting justice for the victim was also relevant to the second (and, for many, the principal) goal of public protection. Fundamentally, We want to take these bad guys off the street, don't we? We want to get the guns off the street, and we want to get people that are handling the guns off the street. We want to get people that are stabbing . . . it doesn't matter who you are stabbing, it doesn't matter whether it's a gang member or not, the fact is that you are stabbing another person and it's violence and it's against the law, we want to take you off the street. (Louise, Detective Sergeant) The doctrine of joint enterprise offered significant legal support to police officers mandated to protect the public from the risks posed by the multiple 'bad guys' involved in 'group' violence. Certainly, detectives, like Vincent, tended to agree with its basic premise that secondary parties should be held accountable for the substantive offence: We always look at the joint enterprise as where they are doing something, which actually is more than mere . . . [presence . . .] but we take the view that if you're there, you know what's going on, and if you are involved in it-so if you're surrounding them, if you are going towards them-then there is an aspect . . . you're part of that joint enterprise. You're part of that.
(Vincent, Detective Superintendent) The position espoused above was normative and, arguably, was invoked to attribute liability. Many detectives conceptualized suspects in multi-handed violence as the 'plural subject' acting with 'shared intention' (Gilbert, 2000, cited in Amatrudo, 2016 and with the requisite mens rea. Take Rob's statement as an example, if you get 10 people surrounding someone with knives, no, it doesn't matter who inflicted what wound. As long as they were there, part of that group, they're in it. I think the important thing is putting people there and painting the picture of what they did as whole rather than individually, because, at the end of the day you'll never be able to say [who did it]. It's impossible; joint enterprise is there to cover the police in that respect. (Rob, Detective Sergeant) Underpinning this view is the assumption that all individuals are cognizant of the group's intentions, all are responsible for the criminal behaviour of affiliates and all are complicit in the harm and wrongdoing. This approach fits with broader policing research which shows that the police conceptualize themselves as regulators and moral arbiters for the state 'perpetually engaged in a struggle with those who disobey, disrupt, do harm, agitate or otherwise upset the just order of the regime' (Van Maanen, 1978: 308; see also Loftus, 2010;Reiner, 2010). In this sense, detectives argued that the ethnicity of suspects was irrelevant when chasing the 'bad guys', but they were aware of how it would be (mis)construed: I don't think that ethnicity plays a big role in the direction of the investigation, nor indeed how people think about them, and you probably won't believe me, or a lot of people won't believe me, sometimes I forget the ethnicity of the people that we're looking at. (Austin, Detective Inspector) Echoing previous research, the core mission for these detectives was to be the 'thin blue line', protecting the weak and vulnerable from the dangerous (Reiner, 2010). In the context of their work 'the dangerous' were best summarized as 'groups running around [. . .] committing these offences' (Darnell, Detective Constable).

Collective violence: Serious violence as gang-related
Detectives in the study commonly framed inter-personal street violence as 'gangrelated'. Indeed, 'gang' offences were cited as the principal driver of the recent increase in violence that the officers reported: 14 'I would say that the gang-I use the term 'gang'-gang or group fighting, violence, is on the up' (Darnell, Detective Constable). The construction of urban violence as 'gang-related' reflected general political rhetoric and criminal justice policy (HM Government, 2018), as seen here in a quote from Davinson (Detective Chief Inspector): for many, many years it [gang violence] has been an issue, it has been a problem. We're seeing an increase in that gang violence recently; and it clearly is. I think there appears to be some kind of an escalation in the mindset of, 'I don't care if I kill somebody; I don't care if I get killed, this is the road I've taken in my life and my eyes are open to the potential ramifications of it.' I think that side of it has potentially increased over the past few years.
In his assessment, Davinson, draws attention to the seemingly rationalized mindset and the fatalistic nihilism of individuals involved in gangs. Such attitudes reflect a perception about suspects, and street gang culture, that scholars have argued allows for the erosion of innocence and a presumption of guilt which, according to Goff et al. (2014), impacts most negatively on black boys who are, within the context of criminal justice, more likely to be viewed as guilty (Gabbidon, 2010) by law-enforcement officials. Arguably, if the presupposition is that 'gang-involved' individuals have a predetermined 'mindset to kill' it not only removes the protections afforded by due process (i.e. the presumption of innocence) and renders an individual(s) capable of inflicting serious harm, but also implies intention and culpability. As Davinson goes on to confirm: As I say a lot of our work is young people involved in gangs, opposing gangs, and for a lot of them life is cheap. And the way they obtain reputation, credibility, respect, is by resorting to violence and extreme violence, and they really don't care if they kill somebody. Unfortunately, that's the scenario, especially in London. That's the scenario we're faced with.
(Davinson, Detective Chief Inspector) Therefore, in attempts to protect the public from the increasing risk of serious violence, including homicide, detectives reasoned that they needed to shield them from collective gang-related violence which, to their reckoning, was more likely involved young black men.

The racialized nature of 'gang' talk
When reflecting on their working practices, detectives repeatedly engaged in 'gang talk' (Hallsworth and Young, 2008) reducing complex situations to simple constructions of 'gang violence' in their pursuit of potential suspects. While some officers cautiously avoided conflating 'gangs' with ethnicity, most invoked the term 'gang' to explain rising levels of crime and drew on racialized stereotypes which categorized and positioned young black men as risky, dangerous and enculturated into a criminogenic 'black' street subculture. Reminiscent of the subculture of violence argument, posited by Wolfgang andFerracuti (1982 [1967]), a small number of detectives explicitly described the existence of a pro-violent black subculture, in which individuals condoned the use of violence: You do have this whole 'young black male'. That is a lot of what we are dealing with. Both sides of it as well: the victim and the suspect groups.

Do you get any sense of why that is?
I think it's . . . In certain communities some things are more acceptable. And I think it's fashionable. I wouldn't say it is exclusively black, because there are white guys who come into it; I would say it's a street culture or urban culture. There are white guys involved in it, but it is predominantly a black man's or a black kid's world. I think it's cool to be a criminal.
(Patrick, Detective Sergeant) It was not possible to ascertain whether such views were those of a few 'rotten apples', 'dissonant' officers' (Foster, forthcoming), or illustrative of an institutionally racist police culture. However, the concept of 'black criminal subculture' as identified by Patrick, exposes the readiness of some officers to associate violence committed by young black men with 'cultural attributes' that criminalize a whole community (Wilson, 2005, cited in Gabbidon, 2010. Such labels have 'collateral consequences' (Wilson, 2005, cited in Gabbidon, 2010 including, in the context of this study, the potential disproportionate application of joint enterprise principles to convict multiples of young black men for serious violence (see Williams and Clarke, 2016).
Constructing multi-handed violence as 'gang-related' provides a coherent narrative, enabling the drawing in of multiple people as 'secondary parties' for a single incident, particularly where identifying the main protagonist in a murder is difficult. It enables officers to cast the net widely early on in their investigations in line with their conception of risk, experiential knowledge and the 'early arrest' strategy described post-Stephen Lawrence Inquiry: 15 If you're just looking at who's responsible for [an offence], and if it's a group offence, who are the members of the group that were responsible for this, if we can prove they were there, arrest them. [. . .] At the arrest stage it's simply, get them in and let's see what evidence we can get then. (Keith, Detective Chief Inspector) Over-representation of BAME individuals at arrest is significant as it ensures that they disproportionately appear at subsequent stages of the criminal justice process (Lammy, 2017). 16 However, for detectives in the current study, recriminations of disproportionality due to stereotyping at the arrest stage were further negated by the perception that the Crown Prosecution Service (CPS) acted as a 'buffer' in the charging process, providing a safeguard against accusations of racism. Moreover, the 'good success rate' (Toby, Detective Superintendent) in past cases of multi-handed serious violence, had a 'conditioning influence' (Archer, 2010: 277) on officers. It encouraged them-that the way in which they had processed such cases to date had been appropriate, despite accusations of stereotyping.
The lack of successful appeals (since the changes to the doctrine) was also 'reassuring' to detectives and supported the continuation of doing things the way they had always been done. And though detectives believed the 2016 Supreme Court ruling (in R v Jogee) had 'tightened the law a little' (Davinson, Detective Chief Inspector) and had made the CPS 'a little less keen' (Louise, Detective Sergeant) or more cautious when charging suspects in multi-handed cases, most officers reflected that it had not changed the investigative process or altered the way in which suspects, (including secondary parties) were drawn into, or eliminated from, an investigation. This was because the police role was viewed as one of 'fact finder and nothing more' (Ray, Detective Sergeant). Thus, changes to the doctrine had 'not affected us too badly' (Eric, Detective Superintendent).

Concluding thoughts
The few criminological studies that have provided data on the use and experience of joint enterprise have added much to our understanding of this complex doctrine, most significantly, drawing attention to the disproportionate number of black and mixed-race young men in cases where it has been applied. The available research has drawn exclusively on the subjective experience of prisoners (Crewe et al., 2014;Hulley et al., 2019;Williams and Clarke, 2016) and has explained the disproportionality in terms of the 'stereotyping' invoked in policing and prosecution strategies (Williams and Clarke, 2016). In an effort to triangulate our knowledge about the application of joint enterprise, in this article we have offered a nuanced analysis of the disproportionate outcomes for black and mixedrace men using the testimonies of Homicide and Trident detectives investigating cases of serious group violence in London.
It is significant that in the current study we were not confronted with explicitly racist attitudes towards BAME men providing clear and unequivocal evidence that this was the basis of disproportionate outcomes. Rather, police detectives claimed that they faced disproportionate numbers of young black men as victims and perpetrators, and, in this context, were attempting to protect the public, including the black community, from serious violence using the tools available to them. Thus, in ways that corresponded with Archer's thesis, the detectives reasoned that their actions were driven by reflecting on their ultimate concerns, within the objective conditions they faced, which were beyond their control.
Still, in their narratives was evidence of 'color-blind racism', which 'is like racism lite (Bonilla-Silva 2006, 3) as it "otherizes softly" in the subtlety of language and coded meaning' (Van Cleve and Mayes, 2015: 412). An explicit feature of such 'color-blind racism' is the avoidance of explicitly racist language in the context of 'racial norms [that] disallow the open expression of direct racial views and positions' (Bonilla-Silva, 2002: 46)-here, it is argued, in the era post the Stephen Lawrence Inquiry. 17 Instead, the language of risk, central to the new penological movement of recent decades, enables the 'punitive governance of communities of color while denying that racism has a significant impact on disproportionality' (Van Cleve and Mayes, 2015: 412).
In the current study, detectives consistently associated young black men with risk and danger, reduced their offending to 'gang talk' and 'continued the use of unproblematized concepts of race and racial differences in the investigation of serious violent crime' (Covington, 1995: 548). The common use of the gang narrative, as part of the cultural and political context in which detectives did their job, is likely to contribute to the disproportionate number of black young men convicted of serious violence, particularly as multiple individuals can be drawn in as secondary parties to a single offence. The framing of street-based violence as collective in this way overshadows the interpersonal violence that occurs between individuals and recourse to the 'gang' allows for the re(interpretation) of street-based violence as acutely dangerous and more threatening to the 'moral majority'-the public who the police are trying to protect.
While supporting previous work highlighting the problematic nature of the gang discourse and the real risks it poses to young BAME men's lived futures, we want to highlight three additional points. First, the significant influence of the law and politics in directing the efforts of the police (Hallsworth, 2006). The doctrine of joint enterprise allows for everyone associated with the crime to be under suspicion as secondary parties. It supports the detectives' dragnet strategy to 'get them all in', to not have to distinguish between the person 'who put the knife in' and the others present, and 'see what emerges' from the prosecution's case. As such, the doctrine encourages the wide casting of the criminal justice net, and the police to scoop up the 'by-catch', the 'small fry' on the periphery drawing them in to the prosecution process. 18 Dragging secondary parties into an investigation is problematic for many critics of joint enterprise. In doing so this undermines the detective's direct claims that they are doing a 'good job' (in difficult circumstances) if one considers 'good enough policing' to be fair and effective, and exercised in ways that make sense to the outside observer (Bowling, 2007).
Second, the lack of robust data on the perpetration of serious violence enables the police to deflect accusations of discrimination. In the absence of clear and consistent data on the application of joint enterprise principles in practice, on the demographics of those involved in the full range of serious violent crime (e.g. murder, manslaughter and grievous bodily harm), 19 including the extent to which particular incidents involve more than one individual, and up-to-date population data by geographical area, the detectives' reflective deliberations cannot be sufficiently verified or discredited and ultimately the causes of disproportionality cannot be confidently asserted. We cannot emphasize enough the need for such data and for further qualitative exploration of the work of police detectives that examines 'front' and 'backstage' attitudes (see Foster, forthcoming), with a sensitivity to the subtleness of 'color-blind racism' (see Bonilla-Silva, 2002 for details). Similarly, the lack of data on the use of joint enterprise principles in practice makes tracing disproportionality of BAME men at each stage of the criminal justice process impossible. It is significant that, since 2016 when changes were made to the doctrine of joint enterprise, detectives asserted that the CPS had 'tightened' up on charging decisions and was demanding more robust evidence of a secondary party's 'intention'. However, in the absence of official data, the current study found little evidence that these changes had impacted significantly on the investigative process suggesting microlevel morphostasis (Archer, 2003(Archer, , 2010(Archer, , 2012 and the continuance of the ongoing issue with disproportionality.
As Reiner (1992: 157) points out, the prejudicial and biased attitudes of police may not translate into discrimination and differentiation if legal, ethical, organizational or situational constraints preclude this. It is our interpretation that what is needed to deal with disproportionality is: a recognition of the features of colour-blind racist attitudes and practices (Bonilla-Silva, 2002); publicly available robust data-detailing the application of joint enterprise by offence, sex, ethnicity and crime type plotted against population data delineated by age and race (and geography where relevant); a change in the law in line with fair labelling (to hold individuals accountable for the criminal acts they committed); and, crucially, changes in prosecutorial practice to reflect such fair labelling principles. These actions are likely to facilitate a morphogenic shift (Archer, 2003(Archer, , 2010(Archer, , 2012 leading to a reduction in the number of individuals drawn into the joint enterprise dragnet. self-identified as being convicted under joint enterprise principles (n = 37) and young people in the community (aged 25 and under) with a range of experiences of serious violence (n = 58). 7. Ethical clearance for this part of the research was granted by the Metropolitan Police Service and the University of Cambridge where the research was based. 8. Two officers had retired from duty; one a week prior to the interviews and the other had been retired for three years. 9. For the two retired officers, their most recent rank prior to retirement is reported. 10. Foster's work (forthcoming) draws attention to the importance of presenting a public facing persona that does not bring the police service into disrepute. Few detectives in Foster's study overtly expressed racist views. 11. Seven hundred and twenty-six (726) homicides were recorded in the year ending March 2018.
Over one-third (285 or 39%) of cases involved a knife or sharp weapon representing the highest since 1946 (see ONS, 2019). 12. For example, while 'prosecution and conviction' data from the Ministry of Justice (2018) provide data on 'violence against the person' by ethnicity, they do not break down the data by offence or by age. It is also noteworthy that neither the Crime Survey of England and Wales (2019) nor Homicide in England and Wales (2019) (ONS, 2019) cite the ethnicity of suspects. 13. It is difficult to obtain specific data showing current population rates broken down by age, ethnicity and geography. This makes ascertaining 'baseline figures' for disproportionality in prison (national) and policing (regional) a particular (and important) problem. Current available data show that 40% of people in London identified themselves as BAME in the 2011 census, with 13% identified as black (ONS, 2013). Within the black and mixed-race population in London, data also show higher proportions were aged 25 and under, compared to other ethnic groups in London (GLA Intelligence, 2013). In the context of rising population rates across London (London data store: https://data.london.gov.uk/) and rising rates of BAME population, it is likely that our baseline figures are somewhat outdated. 14. Note that for the Trident detectives this is likely to be affected by the fact that their remit was 'gang-related violence'. 15. Under the Police and Criminal Evidence Act (PACE, 1984, s. 24) officers can arrest anyone they have reasonable grounds for suspicion of being guilty of a crime. 16. The three other points of disproportionality are trial in Crown Court, sentencing and adjudications in prison discipline. 17. It is relevant that in the follow-up report to the Stephen Lawrence Inquiry, the Home Affairs Select Committee noted there to be 'far less unpleasant language used on the basis of race amongst police officers' (House of Commons, 2009: 19). This reflects Bonilla-Silva's (2002: 42) observations-in relation to race-that 'the normative climate of what can be said in public has changed dramatically' and is thus likely to influence what is disclosed to researchers in the current study. 18. Thanks to the anonymous reviewer for encouraging this analogy. 19. Not just that caused by knives, but across the range of injuries.