Governing Through (An Exclusive) Community: Limitations of State Conceptualisations of ‘ the Community ’ in Domestic Violence Policies

In response to growing evidence that cultural values and behaviours are key drivers of men ’ s use of domestic violence against women, states across the globe are increasingly implementing prevention policies aimed at mobilising cultural change within the community. Through an examination of one Australian state ’ s recent and signi ﬁ cant domestic violence policy reform, we demonstrate that, although state-led efforts to change community culture hold merit, they can also be undermined by exclusive constructions of the community. As a result, efforts to change community culture exclude the very group whose values and behaviours are most problematic: men who perpetrate domestic violence. We argue that broadening conceptualisations of community is of critical importance for policies seeking to change community culture. Such conceptualisations must necessarily include men who perpetrate domestic violence, as theirs are the values and behaviours that most urgently require change.


Introduction
Leading multilateral organisations have long recognised the imperative for states to implement targeted and robust policies to address men's use of domestic violence against women (DV) (United Nations, 2006;World Health Organization, 2009;Council of Europe, 2011;The World Bank, 2018). Increasingly, these organisations are foregrounding the critical role that community culture-including norms, values, and expectations around gender and violence-plays in creating a social environment in which DV can prosper. No longer seen as a private or even family issue, DV is now firmly recognised as a societal problem that state policy must address.
Drawing on the understanding of DV as underpinned by community culture, states across the Western world are experimenting with social policy levers to reduce DV by mobilising cultural change within the community (see, for example, Commonwealth of Australia, 2019; Government of Canada, 2019; UK Home Office, 2019; US Congressional Research Service, 2019). To achieve this, states are increasingly employing strategies to 'govern through community' (Rose, 1999). The shift towards governing through community in the DV context reflects broader policy shifts within prevailingly neoliberal political climates that seek to foreground individual and community responsibility in addressing social issues (Dean, 2002). A well-developed body of scholarship examines this broader policy shift, demonstrating that the way states conceptualise and engage with communities in policy development and implementation processes has important implications for who is targeted and engaged in policymaking processes (Barnes et al., 2003;Connelly and Richardson, 2004;McCarry et al., 2018;Rolfe, 2018). However, as state-led efforts to change the community culture that underpins DV are relatively new, to date there has been limited research surrounding how states conceptualise and engage with the community in the development of DV policy.
In alignment with the broader international policy landscape, Australian state governments are implementing policies that seek to actively mobilise the community to create cultural change and reduce DV. The State of Queensland, in particular, has engaged in significant policy reform over the past five years, now foregrounding the importance of mobilising 'a significant shift in community attitudes and behaviours' (Queensland Government, 2016: 15). Drawing on Queensland as a case study, we contribute to addressing the aforementioned knowledge gap by examining how these reforms conceptualise and engage with the community as a site for change. Notably, we find that perpetrators and potential perpetrators of DV are excluded from conceptualisations of the community and are given little responsibility for enacting change. Drawing on these findings, we argue that there is little to be gained by targeting cultural-change strategies at community members who already hold the desired attitudes. Future national and international policies must actively work to include perpetrators of DV in conceptualisations of the community, not to condone their values and behaviour, but to help facilitate change.

Queensland policy context
In Australia, policies that recognise and seek to address the cultural facilitators of domestic violence represent a significant and positive shift away from previous generations of domestic violence policy, which have been criticised for insufficiently understanding and engaging with the highly gendered and cultural nature of the issue (Murray and Powell, 2011). Contemporary Queensland policy, in particular, foregrounds culture as a core contributor to domestic violence and targets the community as a site for change. This is underpinned by a view that community members who witness domestic violence as bystanders have a moral duty to intervene. Changing bystander behaviours is thus a core aspect of Queensland's current policy, as 'changing bystander behaviour goes to the very heart of cultural change' (Special Taskforce on Domestic and Family Violence in Queensland, 2015: 162).
Despite the progress that current culture-focused policies represent, however, leading scholarship identifies numerous cultural, structural, and individual factors that contribute to DV and require targeted interventions (Murray and Powell, 2011). The historical failure of Australian policies to recognise and respond to structural facilitators of DV, in particular, has been strongly criticised by scholars in the field (see, for example, Wall, 2014;Salter, 2016;Kuskoff and Parsell, 2020). While we recognise the significance of such critiques-and, indeed, the importance of addressing structural facilitators of DV-it is not the purpose of this article to contribute to them. Rather, we seek to understand how the Queensland Government is attempting to change community culture, and what this means for how community members are conceptualised and engaged through policy. Ella Kuskoff, Andrew Clarke, and Cameron Parsell Conceptualising the c ommunit y in state-based policy The notion of 'the community' has come to play a central role in policies informed by neoliberal rationalities in recent decades. This is due, at least in part, to the dominant influence of neoliberal rationalities that position the self-governance of political subjects as the optimal means of addressing social problems; and which problematise 'top-down' state intervention as both stifling of self-governance and unresponsive to the diverse and changing needs of citizens (Miller and Rose, 2008;Dean, 2010). Community is attractive from this perspective, as it is seen as a self-regulating sphere of social relationships that are built around the spontaneous bonds and shared moral commitments of its members. Political authorities thus seek to 'govern through community' by cultivating and facilitating its self-governing capacities, and harnessing them to political ends (Rose, 1999).
Importantly for this article, community is understood in academic and policy discourse as being underpinned by a shared 'culture' comprising mutually agreed upon norms, expectations, and moral commitments (Etzioni, 1997). It is this apparent culturalmoral unity/integration that makes community an attractive resource for governing, as it suggests that political authorities can leverage an already existing set of shared values and obligations in the pursuit of their objectives. Efforts to govern through community thus entail the construction in policy discourse of what community culture is or should be. Culture becomes both a target and a resource for governing: on the one hand, policy is increasingly authorised with reference to its alignment with ostensible community norms and values; on the other hand, the shaping of these norms and values is a means through which policy can harness the self-governing capacity of communities to the remediation of social problems (Rose, 1999). It is this latter process in particular that we see at play in contemporary policy, as we outline below.
While states often position their engagement with communities as a means of more effectively addressing social problems, scholarship demonstrates the limitations inherent in dominant political conceptualisations and enactments of community (Adams and Hess, 2001;Wilcox, 2006;Cooper, 2008;Taylor, 2011;Reynolds et al., 2015). A common critique is that the construction of community as a moral space defined by shared values and feelings of unity and support has divisive and exclusionary effects. The unity of communities and their culture is premised on the establishment of 'Us' versus 'Them' dichotomies, which entail processes of exclusion where those who fail to uphold the values shared by community members are deemed unworthy of inclusion. 'Us' group members who belong to the community are thus positioned as moral and good, in opposition to 'Them', the potentially dangerous others who fail to conform. As Tester (1992: 47) argues, 'it is precisely the identification of an abhorrent 'them' which makes 'us' possible'. Importantly, moralistic divisions between 'Us' and 'Them' help to legitimise the dominance and privilege of the 'Us' group and reinforce the exclusion of the 'Them' group. This highlights the necessity of interrogating how states conceptualise the community through their policies, including who they perceive as belonging to 'Us' and 'Them' groups, and what the implications of these perceptions might be.
An enduring body of scholarship examines states' conceptualisations of the community in a variety of policy contexts (Barnes et al., 2003;Newman et al., 2004;Wilcox, 2006;Narayanan et al., 2015;Reynolds et al., 2015). This research illuminates numerous negative implications that stem from states' tendencies to conceptualise the community as a homogenous group united by shared values and exclude those who do not conform to Governing Through (An Exclusive) Community the norm. Policy conceptualisations of community have been found to reflect social hierarchies, thereby reifying unequal power relationships and reinforcing the marginalisation of minority groups (Wilcox, 2006;Narayanan et al., 2015;Reynolds et al., 2015). This is particularly problematic in the context of DV. It is well demonstrated that women in minority groups experience DV in complex and compounding ways, and that cultural assumptions that underpin policy and practice can act as barriers to accessing support (Burman et al., 2004;Burman and Chantler, 2005). Moreover, varying cultural attitudes about what constitutes DV can hinder the capacity of perpetrators from some groups to even identify their behaviours as violent and thus participate in a process to facilitate change.
Significantly, the assumption that all community members share certain values and experiences can lead states to exclude already marginalised groups from policy discussions, instead favouring the input of so-called experts or community leaders. However, while the problematic reproduction of social hierarchies and inequalities that result from states' narrow conceptualisations of the community is generally understood, there is limited current knowledge regarding how this process plays out in DV policy that seeks to change community culture. The highly complex nature of DV and its roots in inequalities and community culture positions this as a significant gap that must be addressed if we are to understand and, if necessary, improve policy attempts to mobilise cultural change.

Me thods
Using the Australian state of Queensland as a case study, we begin to address this gap in knowledge by examining how the Queensland Government conceptualises and engages with the community in its current DV policy. Specifically, we address the question: How does Queensland's DV policy construct 'the community', and what are the implications for the policy's ability to achieve community cultural change? To answer this question, we analyse data both from current Queensland Government policies and members of the Queensland DV policy community, using a critical discourse analysis (CDA) framework.

Analysis of policy texts
To set the temporal dimensions of our study, we draw on Fairclough's (1992) concept of 'a moment of crisis'. According to Fairclough, moments of crisis offer useful starting points for defining temporal dimensions of critical policy studies, as they allow for the examination of new policy approaches, while simultaneously acknowledging the influence of historical policy processes. We take as our moment of crisis the Queensland Government's establishment of the Special Taskforce on Domestic and Family Violence in Queensland (henceforth referred to as the Taskforce) in 2014. In a radical divergence from previous generations of DV policy, the Taskforce approached DV not as an individual problem, but as a problem grounded in community culture (Special Taskforce on Domestic and Family Violence in Queensland, 2015).
In the years that followed the establishment of the Taskforce, the Queensland Government developed and began implementing significant DV policy reforms, informed heavily by the strong cultural focus taken by the Taskforce (Queensland Government, 2016). These reforms form the basis of our policy analysis, which includes the following three policy documents: Not Now, Not Ever: Putting an End to Domestic and Family Ella Kuskoff, Andrew Clarke, and Cameron Parsell Violence in Queensland (Special Taskforce on Domestic and Family Violence in Queensland, 2015); Queensland Government Response to the Report of the Special Taskforce on Domestic and Family Violence (Queensland Government, 2015); and, Queensland Domestic and Family Violence Prevention Strategy (Queensland Government, 2016).
We began the analysis by conducting an inductive thematic analysis of each policy text, with the assistance of NVivo software to organise and manage the data. Given the large amount of data contained in the three policies, this was a necessary means of sorting the data into manageable, relevant, and focused sub-sets of data. Codes related to community, culture, and participation in policymaking were included in the CDA component of our analysis. To conduct the CDA component, the data within these codes was analysed drawing on Fairclough's (1992) approach to CDA. Rather than offering a comprehensive method, CDA can be viewed as an overarching analytical approach that encompasses an extensive 'toolkit' of strategies for analysing discourse (Fairclough, 1992). The specific analytic strategies we drew on for the purpose of this study are framing (i.e. analysing how representations of certain issues are constructed within the data) and transitivity (i.e. analysing the perception and treatment of certain people and processes within the data) (Fairclough, 1992). Together, these strategies allowed us to examine how the Queensland Government conceptualises the community through the policy discourse, and how this conceptualisation was constructed.

Interviews with members of the policy community
Augmenting the analysis of formal policy documents, we conducted and analysed interviews with members of Queensland's DV policy community. These interviews sought to understand the links between policy language and social practice by examining the processes through which the policy discourses were constructed, reconstructed, and made dominant. We purposively sampled members of the policy community who had been involved in the development or implementation of the policies included in our analysis. A total of fifteen members of the policy community were interviewed for the study. The sample included Queensland Government employees from various state departments and levels of authority (n=7); community service organisation CEOs (n=3) and employees (n=3); and academic experts from various institutions and disciplines (n=2). Prior to conducting the interviews, ethical approval was sought and granted by our institution's Human Research Ethics Committee.
All interviews were conducted on a one-on-one basis, either in person or over the phone. Interviews were semi-structured and guided by interview schedules. These schedules were developed drawing on key preliminary findings of the policy analysis, but were tailored slightly to each participant to ensure their unique positions and experiences are able to be captured. On average, interviews lasted between thirty and sixty minutes. All interviews were audio recorded and transcribed with participants' informed consent, and all participants were given the opportunity to review their transcripts prior to analysis.
To analyse the interview data, we adopted a similar method as for the policy data. Interview transcripts were thematically coded using NVivo software. The content within each code was then analysed using the previously-discussed tools of framing and transitivity. Findings from the policy and interview analyses were then integrated according to overlapping themes. Below, we present these findings, which we ground in Governing Through (An Exclusive) Community contemporary theory and literature. To support our arguments, we provide data extracts from the policies and interviews we analysed. These extracts are not exhaustive examples; rather, they are intended to exemplify the themes that emerged from the analysis. To maintain participant anonymity, excerpts of interview data are labelled based on the participants' general role in the DV sector (e.g. Government Employee, Service CEO, etc.). This labelling convention protects participants' identities while still providing an indication of the potential agendas and social power each participant holds relative to the others.

Defining community boundaries
Reflecting the complex and often contradictory nature of policy discourses (Fairclough, 1992), our analysis does not point to a single policy construction of community or its boundaries. When the policies discuss DV as a community issue, for example, the community is presented as an inclusive concept, encompassing every individual in the state: The following key principles will form the basis of how Queensland, as a community, can work together to end domestic and family violence. (Queensland Government, 2016: 19) It is beholden upon all of us-every single citizen of this diverse, vibrant state-to take a stand against domestic and family violence. (Special Taskforce on Domestic and Family Violence in Queensland, 2015: 1) When the policy begins to discuss the importance of shared community values and behaviours, however, the inclusive conceptualisation of community starts to weaken. Indeed, drawing on both our policy and interview analyses, we identified five distinct community groups that are constructed through the policy discoursenamely, Us; Potentially Us; Vulnerable People; Children; and Them. The boundaries of these groups are defined according to the members' perceived values and behaviours regarding DV, and each group is assumed to hold different roles and responsibilities towards the issue. We discuss each group in turn.

'Us' group
The first and most ideal community group we identify throughout the Queensland Government policy is the group we term 'Us'. Labelled to reflect the concept of the 'Us' group previously identified in the literature, this group represents: 'Us' members are positioned as understanding DV as 'a horrific blight on our community' and share the values of non-violence and respect (Special Taskforce on Domestic and Family Violence in Queensland, 2015: 47). Importantly, 'Us' group members not only uphold the desired values by refraining from engaging in violent and disrespectful Ella Kuskoff, Andrew Clarke, and Cameron Parsell behaviour; they are also engaged community members who actively engage in responsible bystander behaviour by calling out and intervening in violent or disrespectful situations when they witness them occurring. Throughout the policies, members of the 'Us' group are exemplified and celebrated for their engagement in such prevention activities: Some men describe themselves as compelled to action : : : [Joe] is an encouraging example of one man making the decision to take action in his daily life to challenge domestic violence. (Special Taskforce on Domestic and Family Violence in Queensland, 2015: 179) Wanting to create awareness of abuse within culturally diverse communities in Australia, the Queensland Edifest Association developed a series of posters : : : The posters were widely applauded and received significant media attention. (Queensland Government, 2016: 20) As responsible bystanders, the 'Us' group are the exemplar of self-governing community in relation to DV. They are both invested in appropriate community values and willing to champion those values in their interactions with other community members. They are therefore an important resource for the Queensland Government's efforts to govern DV 'through community' (Rose, 1999).
However, as the policies and interviews both make clear, members of the 'Us' group are currently a minority: We wanna really shift the culture where you're not the minority, you're not the only person that's gonna say or do something. (Participant 4, Service CEO) We must create a society where fewer people will find excuses for domestic violence, fewer people will blame victims, and more people will know how to take action against domestic violence. (Special Taskforce on Domestic and Family Violence in Queensland, 2015: 152) Members of the 'Us' group are therefore upheld as positive examples of good community members, and are encouraged to continue leading by example by enacting the culture that the Queensland Government wants others to emulate.

'Potentially us' group
This brings us to the next, most visible community group throughout the policies, which we term 'Potentially Us'. 'Potentially Us' group members are positioned as holding the desired cultural values, but not necessarily the desired behaviours. Like the 'Us' group, 'Potentially Us' members understand DV to be unacceptable and do not engage in violent behaviour. However, unlike the 'Us' group, 'Potentially Us' group members are seen as lacking an understanding of their own responsibility for reinforcing cultural standards and stopping others' use of violence. As a result of this lack of understanding, members of this group are seen as failing to behave as responsible bystanders. A key aspect of the Queensland Government's governing through community strategy is, therefore, to instil 'Potentially Us' group members with this understanding and to enrol them in the process of community self-governance through bystander intervention.

Governing Through (An Exclusive) Community
The 'Potentially Us' group is positioned as capable of changing their behaviour in a way that allows them to more actively enact their responsibility for violence prevention and become one of the 'Us' group. For example: There appears to be a significant gap between an individual's belief that the violence is wrong, and the willingness to talk about the violence or take action to do something about it. 'Potentially Us' members are therefore positioned as an important group that should be targeted by community education and cultural-change strategies. Where 'Us' group members are given the responsibility of demonstrating desirable cultural norms and values and encouraging members of the 'Potentially Us' group to do the same, members of the 'Potentially Us' group have the responsibility of changing their own behaviours to align with those in the 'Us' group and to become active participants in community selfgovernance. The upward mobility of 'Potentially Us' group members is a key theme in the policy, with numerous strategies suggested to help enable 'Potentially Us' group members to change their behaviours and move into the 'Us' group.

'Vulnerable people' group
A third community group constructed through the policy discourse is a group we term 'Vulnerable People'. This group consists predominantly of victims of DV and mothers with children. While the policy discourse positions this group as sharing the desired cultural values of non-violence and respect, like members of the 'Potentially Us' group, 'Vulnerable People' are not seen behaving in ways that challenge others' use of violence. However, unlike the 'Potentially Us' group, 'Vulnerable people' are not given the responsibility of changing their behaviour in a way that allows them to move into the 'Us' group and participate in community self-governance. Indeed, our interviews with members of the policy community suggest that 'Vulnerable People' group members are viewed as having a legitimate reason for not actively upholding desired bystander behaviour. For example, one service employee said: In terms of standing up and doing something about that, it could be very different scenarios for different people in the community. You know I might be a single mum pushing my pram down the street and see a woman being abused and attacked and I may not physically intervene because I need to be safe for me and my children. (Participant 10, Service Employee) A woman who's been abused isn't necessarily, I mean one woman may be there to help another woman and another woman may find that too confronting for her. (Participant 10, Service Employee) Members of the 'Vulnerable People' group are therefore absolved of their responsibility to act in the way the policies expect the 'Us' group community members to act. Instead, they are encouraged to recognise DV if and when it is perpetrated against them, and to seek help for themselves.

'Children' group
The fourth community group constructed through the policy discourse is the group we term 'Children'. 'Children' are positioned as either not yet having fully formed their cultural values and behaviours, or as developing problematic values and behaviours that can and should be changed. The policy thus foregrounds the importance of education and positive role models to help children develop desired values and behaviours. Members of the 'Us' group are given the responsibility of providing appropriate guidance and leading by example so children are able to grow up to be members of this group.
We all know that the true early intervention : : : really happens at that kind of forming ages of 3,4,5,6,7,young people,children. (Participant 11,Government Employee) Education of children and young people is key to achieving generational change in behaviour. (Queensland Government, 2016: 19) The cultivation of self-governing community therefore requires adults in the 'Us' group to educate and guide children in learning the appropriate cultural attitudes and behaviours towards DV. This implies that, should children grow up to hold violence-supportive attitudes, the responsibility lies with the 'Us' members of the community for not adequately instilling in children the appropriate values.

'Them' group
The 'Them' group is an additional group constructed by the policy discourse; crucially, however, this group is not positioned as part of the community. On the contrary, this group-a group comprised of perpetrators, potential perpetrators, and people with violence supportive attitudes-is constructed as external to the community on account of their failure to uphold desired values and behaviours. This is in contrast to 'Potentially Us' members who are seen as sharing desired values, but who fail to behave in ways that challenge others' use of violence. 'Them' group members are seen as failing to share the value that violence and disrespect are unacceptable, as evidenced by their use or condoning of violent behaviours.
Throughout the policies, a clear separation is made between 'Them' and the rest of the community, with unacceptable behaviour firmly positioned as a key factor motivating this divide: These responses are fundamentally important in providing a strong message to the community and to perpetrators of domestic violence that their behaviour is not acceptable, will not be tolerated and will result in consequences. (Special Taskforce on Domestic and Family Violence in Queensland, 2015: 154, emphasis added) Everyone in our community must make it clear that we will not tolerate the behaviour of anyone who hurts another person within a relationship of intimacy and trust. (Queensland Government, 2016: i) Further, whereas the failure of 'Potentially Us' group members to uphold desired behaviours is attributed to a lack of understanding of the cultural factors that contribute Governing Through (An Exclusive) Community to DV, 'Them' group members' failure to uphold desired behaviours is positioned as stemming either from a strong defiance of cultural norms or from innate personal characteristics. In contrast to 'Children' group members, who are viewed as being in a process of development and able to change potentially undesirable values and behaviours with appropriate guidance, 'Them' groups members appear to be assumed as largely incapable of significant and sustained change. As the Taskforce Report states: By positioning 'Them' group members as separate to the community and using strong emotive language to highlight their divergence from community norms and values, the policy constructs people who engage or are at risk of engaging in DV as deviants who irrevocably threaten community cohesion. They are thus part of the problem that selfgoverning communities must manage, rather than a segment of the community that must be responsibilised for the prevention of DV.

Disc us sion
The existing literature surrounding policy conceptualisations of the community provides insight into why the Queensland Government may have conceptualised the community in this way. The literature demonstrates that government policy tends to assume that community members share common values, identities, and trust relationships, and act out of a sense of moral solidarity (Adams and Hess, 2001;Cooper, 2008). To invoke the values and moral responsibility necessary to encourage the community to change culture, the Queensland Government excludes from the community those citizens who do not share these traits. Indeed, so-called 'good' community members are unlikely to feel a moral obligation towards their community if that community includes people without shared values and with whom they do not identify. Those who are included as community members are therefore positioned by the policies as having a responsibility to engage in bystander intervention as a means of protecting their community from the outsiders who threaten their unifying traits. Those outsiders are not themselves interpellated or addressed by the policies; they are the objects, not subjects, of governing through community.
The Queensland Government's construction of perpetrators, potential perpetrators, and people with violence-supportive attitudes as existing outside of the community reflects the 'Us/Them' binaries that have been well-explored in existing scholarship on community (Wilcox, 2006;Cooper, 2008;Taylor, 2011). This scholarship demonstrates that 'Us' groups are generally seen to be moral and constitute the powerful majority. 'Them' groups, on the other hand, are generally seen to be immoral and represent a threat to the community (Winter, 2014). While 'Us' and 'Them' are generally positioned as a dichotomy, our analysis has shown that Queensland DV policy articulates the 'Us/Them' distinction as considerably more nuanced. Winter (2014) encounters a similar Ella Kuskoff, Andrew Clarke, and Cameron Parsell phenomenon in her analysis of discourses of multiculturalism, and draws on the notion of a 'We' group to help explain this distinction. Winter (2014) argues that the 'We' group is made up of 'Us' in addition to 'Others' who stand together in opposition to 'Them'. To form the 'We' group, 'Us' and 'Others' are united through the similarities they share and which set them apart from 'Them'.
The notion of the 'We' group helps explain the Queensland Government's conceptualisation of the community that it wishes to mobilise to change culture and address DV. The Queensland Government has conceptualised a 'We' group comprised of 'Us', 'Potentially Us', 'Vulnerable People', and 'Children', who are united by their lack of approval of DV. This 'We' group stands in opposition to the 'Them' group on account of 'Them' group members' use or support of DV, and thus their failure to uphold community values and norms. Importantly, the Queensland Government perceives the possibility of movement-and indeed seeks to enable such movement-between the sub-groups within the 'We' group (e.g. 'Potentially Us' members may learn to change their responses to DV and move to the 'Us' group; 'Children' group members may grow up to hold the desired attitudes and behaviours and move to the 'Us' group). However, the policies largely fail to account for the possibility of movement between the 'We' and 'Them' groups. Although there is acknowledgement that 'Them' group members should 'receive appropriate sanctions and access to assistance to stop using violence' (Queensland Government, 2016: 17), this is not accompanied by a strong framework for designing and implementing such assistance. Indeed, as Powell and Harris (2021) argue, DV policies across Australian states often recognise the importance of working with perpetrators, yet fail to invest the necessary resources into designing, delivering, and evaluating the number and types of perpetrator interventions necessary.
In contrast, DV policies tend to place a strong focus on implementing strategies aimed at members of the 'We' group, including strategies to educate community bystanders on their responsibility to prevent DV through active intervention (Kuskoff, 2021). Not only does this foregrounding of strategies targeted at the 'We' group result in a policy that largely overlooks the significance of mobilising cultural change among perpetrators of DV ; it assumes that 'We' and 'Them' are mutually exclusive categories, thereby overlooking the potential for people who may be perceived as 'We' members to also perpetrate violence. This helps to create an erroneous boundary between 'good' community members and 'bad' perpetrators, and minimises 'good' community members' responsibility to reflect on their own problematic cultural values and behaviours (beyond failing to intervene when they witness DV) and to consider how these may be contributing to the problem (Mohamed, 2015).
Positioning violent men as deviant community outsiders can also make it harder for 'We' group members to believe that a man who is perceived as belonging to their group is capable of perpetrating violence. This can help perpetuate stereotypes linking DV to already marginalised groups or neighbourhoods, and overlooks the potential for men who engage in violence to perform multiple identities (Wilcox, 2006;Featherstone and Peckover, 2007). This in turn helps to facilitate a social environment in which violence perpetrated by men who do not fit the perception of deviant outsiders may be overlooked, minimised, or excused by the community (Mankowski et al., 2002). This is particularly true for men who hold certain social attributes, such as being white and middle class. Such men may hold enough power and respect to convince others of their community belonging. This can have important implications for judicial responses to DV in particular, Governing Through (An Exclusive) Community as it may lead people to doubt the truthfulness of victims' experiences and contribute to more lenient sanctions for perpetrators with these social attributes. This suggests that by constructing the community in a way that excludes certain groups and failing to account for the possibility of future inclusion, the Queensland Government policy may undermine its own efforts to change the problematic cultural values and behaviours that underpin DV.

Conc lusion
International evidence clearly implicates cultural values and behaviours as key driving factors of men's use of DV against women, indicating a critical need for state-led efforts to mobilise cultural change within the community (United Nations, 2006;World Health Organization, 2009;Council of Europe, 2011). This need is being increasingly recognised by states across the globe, several of which have begun introducing DV prevention policies that foreground the importance of changing values and behaviours surrounding violence and gender equality at the community level. However, given that such policy efforts are relatively new, there is currently limited evidence regarding how such policies construct the community, or how such constructions influence the policies' potential for success.
Although existing literature shows that state-led efforts to change community culture may hold merit, our findings demonstrate that such efforts can also be undermined by rigid and exclusive constructions of the community. Indeed, in the Queensland case, state-led efforts to engage citizens in cultural change are based on flawed assumptions of a unified community culture, and may thus fail to adequately speak to and appropriately mobilise diverse target populations. Of course, the assumption that community entails naturallyoccurring social bonds formed around a shared culture is precisely what makes it attractive as a resource for governing in political contexts where neoliberal rationalities prevail; for it is these qualities that are believed to endow community with its selfgoverning capacity (Rose, 1999). Yet, for culture-centred responses to DV to be effective, governments must engage much more critically with these assumptions and with the make-up of the 'communities' that they govern.
Drawing on our findings, we offer two key recommendations for policies seeking to motivate community cultural change. First, we recommend that conceptualisations of the community be expanded in the broadest sense, encapsulating all individuals in a given jurisdiction. These renewed conceptualisations must not be tokenistic, but rather fully embedded throughout the policy, underpinning its design, proposed strategies, and implementation. Crucially, renewed conceptualisations must include those who are perceived as holding shared values and behaviours (the 'We' group), as well as those who are not (the 'Them' group). As our findings suggest, how the community is conceptualised is critical, as it has significant implications for who is targeted and engaged in policy strategies, and who is given responsibility to enact change.
Expanding conceptualisations of the community to include members of the 'Them' group may thus facilitate our second recommendation. Given that it is members of the 'Them' group whose cultural attitudes and behaviours are most problematic, we recommend that policies make greater efforts to target these community members in their strategies for change. Policies must move beyond the treatment of perpetrators of DV as the objects of bystander interventions, and begin engaging them as active subjects that Ella Kuskoff, Andrew Clarke, and Cameron Parsell may themselves be targeted and engaged through policy. In doing so, perpetrators need to be held accountable as members of the community, while simultaneously having access to the community supports they require to change their behaviour. Engaging perpetrators is thus not a means of downplaying their violent behaviours, but rather an avenue for recognising that perpetrators of violence hold critical perspectives that must be accessed to inform a transformed policy agenda .
Perpetrators, potential perpetrators, and people with violence-supportive attitudes do not exist outside the community, nor are they substantively different from perceived 'good' members of the community. This may be an uncomfortable idea for governments and their constituents to grapple with. Indeed, it seems much easier to draw a line between us and them, and vigilantly maintain that line so we may remain safe and separated from their deviance. It is difficult to recognise how we on the 'right' side of the line contribute to the problem, and even more difficult to consider how we might help those on the 'wrong' side to cross back over. Taking steps to include 'Them' within the community will be an important start to dismantling this divide.