“An exercise in careful diplomacy”: talking about alcohol, drugs and family violence

Abstract The role of alcohol and drugs (AOD) in domestic and family violence is increasingly difficult to ignore, yet challenging to reconcile with dominant feminist analyses of the problem. The landmark 2015–16 Victorian Royal Commission into Family Violence recommended service integration between the AOD and domestic and family violence (DFV) sectors, but this may be difficult to achieve due to differences in history, language use and treatment philosophy between the two sectors. A central point of contention about the role of AOD in DFV can be represented by the question: “are alcohol and other drugs a cause of DFV?” Unpacking this question is essential to understanding contested problem framing in this area, and requires attention to differences in language use and research traditions; varying acceptance of gender inequity as a central causal factor; cultural attitudes about alcohol and disinhibition; and notions of accountability across the different sectors. I argue that substance abuse affects DFV in gendered ways, and that more attention should be paid to how AOD and gender intersect to affect the perpetration and experience of DFV. This article will be useful to practitioners seeking to understand the sensitivities surrounding discussion of AOD in relation to DFV, and will assist them in navigating these sensitivities to improve service coordination and thus deliver better outcomes for those affected by DFV.


Introduction
Domestic and family violence (DFV) is a problem that governments are trying to tackle the world over. In the Australian state of Victoria, the government recently held a Royal Commission into Family Violence, with the purpose of examining the family violence prevention and response system and providing suggestions to improve it. The evidence given to this Commission provides a useful case study of how people from different professional backgrounds approach the relationship between substance abuse and DFV. This can help us to understand where the differences lie and thus how to improve communication between people from different workforces who must increasingly work together to respond to DFV. 1 Due to under-reporting and inconsistency in definitions of DFV, the scale of the problem 2 is not easy to accurately measurebut it is difficult to deny the seriousness of the statistics we do have. The World Health Organization (WHO 2017) reports that worldwide, 30% of women who have been in a relationship have experienced physical or sexual violence from their partners in their lifetime. In Australia, the results of the 2016 Australian Personal Safety Survey showed that approximately one in four Australian women and one in 13 men had experienced physical violence from a current or former intimate partner. One in four women and one in six men had experienced emotional abuse from a current or former partner. Men were more likely to be assaulted in a public place by a male stranger, and women were more likely to be assaulted by a man they knew, in their home (ABS 2017). The Domestic Violence Death Review Team (2018) found that four out of five intimate partner homicides in Australia involved a man killing his current or former female partner. Of the few female-perpetrated homicides, over half were committed by women who had been the primary abuse victim in that relationship. In the state of Victoria (where the Royal Commission was held), police responded to 56,575 family violence incidents in the year ending March 2019, which constituted 14.6% of all criminal incidents in the state (Crime Statistics Agency 2019).
The use of alcohol and other drugs by perpetrators and victims of DFV is one of the most fraught issues in the domestic and family violence field. It is increasingly difficult to ignore, and yet can be challenging to incorporate into traditional feminist power and control analyses of DFV. For reasons I will explore in this article, a central point of contention about the role of alcohol and other drugs (AOD) in DFV can be boiled down to a single question: "are alcohol and other drugs a cause of DFV?" Unpacking this question is essential to understanding why the intersection of DFV and AOD is so controversial, and requires attention to differences in language use and research traditions; varying acceptance of gender as a central causal factor; cultural attitudes about alcohol and disinhibition; and notions of accountability across the different sectors.
As I show in this article, long-standing problem framing debates relating to AOD and DFV played out in the microcosm of the Victorian Royal Commission into Family Violence. To demonstrate this, I provide an overview of the literature on the links between substance abuse and DFV and explore the potential reasons for disagreement about how to characterize this relationship. I then draw on the witness statements and public testimony of several expert witnesses who gave evidence before the Commission in its AOD module, using a "frame analysis" approach to describe the similarities and differences between their positions. 3 I also refer to data from five anonymous interviews with experts who had given evidence to the Commission on this topic. Lastly, I urge practitioners to move beyond disagreements about whether gender or substance abuse is more important in DFV: I recommend consideration of the way they intersect to affect the perpetration and experience of DFV, as well as opportunities for help-seeking.

The relationship between AOD and DFV
That alcohol and other drugs have a role to play in the perpetration and experience of DFV is beyond doubt (Humphreys et al. 2005;Bennett and Bland 2008;Braaf 2012;Wilson et al. 2014). A literature review by DFV researchers (Noonan et al. 2017) suggests three ways in which AOD consumption is linked to DFV: first, it relates to the perpetration of violence; second, to the experience and severity of victimization; third, victims may abuse AOD as a coping strategy.

Perpetration
In the 2-year period 2014-2015, 28% of 139,148 family violence incidents recorded by Victoria Police involved offenders who were either definitely or possibly affected by alcohol (Sutherland et al. 2016, 3). For 13% of incidents, the offenders were definitely or possibly affected by drugs (Sutherland et al. 2016, 4). 4 Graham et al. (2011) review the link between intimate partner violence severity and alcohol consumption. They suggest that, across many different cultures, violence is more severe when one or both partners (most often the male partner) have been drinking. This is supported by Sutherland et al.'s (2016) analysis of the Victorian data mentioned above, which concluded that the presence of alcohol was associated with increased frequency and severity of violence. Despite this, men's behavior change programs in Victoria, which are the primary non-custodial intervention option for perpetrators of family violence, have not historically integrated AOD treatment (Brown et al. 2016). Gilchrist et al. (2019) note that intoxication is not the only risk factor related to AOD use: the irritability and frustration associated with withdrawal can also play a role.

Victims' substance use
Substance use by victims has also been linked to the level of violence that is perpetrated against them, though increasing their likelihood of victimization or compounding problems associated with victimization (Braaf 2012;Nicholas et al. 2012). Research suggests that alcohol use or dependency can impair a victim's judgment, making them less able to de-escalate situations of conflict; reduce their capacity to implement safety strategies; increase their dependence on a violent partner; and decrease their credibility with service providers (Nicholas et al. 2012). Evidence suggests that intoxicated victims are much less likely to have their DFV reports taken seriously by policein Victoria, DFV incidents where only the victim was recorded as using alcohol were least likely to lead to an offence being recorded (compared to no alcohol use by either party, perpetrator alcohol use, or alcohol use by both parties) (Sutherland et al. 2016).
AOD as a coping mechanism A third relationship identified in the literature concerns AOD use as a result of DFV victimization: considerable evidence suggests that victims of DFV can develop problematic relationships with drugs or alcohol as a coping mechanism (Humphreys et al. 2005;Galvani 2006;Braaf 2012;Devries et al. 2014;Noonan et al. 2017). Perpetrators may encourage this behavior, to increase their control of victims (Stella Project 2007). Victims' AOD use can have flow-on effects that increase vulnerability to further violence and reduce effective engagement with recovery supports. For example, many family violence refuges will not admit women with active AOD problems (Humphreys et al. 2005;Braaf 2012;Macy and Goodbourn 2012), and women who experience cooccurring DFV and substance abuse issues are less likely to complete treatment programs (Noonan et al. 2017). A further barrier is that many AOD services are not childfriendlyif they admit children at alland may even see children as impediments to a woman's recovery (Breckenridge et al. 2012;Salter and Breckenridge 2014). However, there may be nowhere else for the children to go; women are understandably reluctant to leave their children with violent partners while they attend detox or rehabilitation services, and may fear having their children removed by statutory authorities if they disclose substance abuse issues to government agencies (Bennett and Bland 2008). Salter and Breckenridge (2014, 169) found that under the "medical model" of addiction emphasising choice and autonomy, children were "a complicating factor in women's progression into a new, autonomous personhood free from dependency." The great causation debate Despite the statistics given above, discussion of alcohol, drugs and DFV is fraught with difficulty. I now turn to an exploration of why this is so controversialwhy, as Braaf (2012, 1) suggests, it is "the elephant in the room." As I elaborate below, concerns about AOD and causation can be traced partly to deep cultural and historical divisions between the male-dominated AOD sector and the female-dominated DFV sector, and partly to concerns about public attitudes to drunkenness and responsibility. These are consistently reinforced by population surveys such as the National Community Attitudes to Violence Against Women Survey (Webster et al. 2018). A third tension stems from different understandings of the notion of causation.

Differences between the alcohol and other drugs and family violence sectors
In Victoria, specialist DFV services began as a network of community-based women's refuges set up in the 1970s. Today these services are still delivered through community organizations, and have expanded to include a range of support and counseling services (State of Victoria 2016, vol. II, p. 2). They are funded by government contracts. AOD services in Victoria are delivered by a range of public health services, NGOs and private organizations, some of which are funded by the state and Commonwealth governments (State of Victoria 2016, vol. IV, p. 19).
The AOD and DFV sectors have a "difference in service and treatment philosophies" (Macy and Goodbourn 2012, 248), which can lead to suspicion and collaboration barriers between the two sectors. One of my interview participants (P07, an AOD practitioner) reflected that the two sectors had grown up on "opposite sides of the fence" and that "we're all carrying baggage from our history": So the drug and alcohol sector's baggage is that we sprung up out of a group of disenfranchised people who'd had histories of their own addiction, who'd come through the other end, who had picked people up off the streets who were like them, and it was advocating for them and fighting with them against the world. … my understanding of much of the family violence sector, again it's come from a group of people who've been through their own journeys and histories in same way as [the AOD group], but they'd often been on the other side, working with women fleeing domestic violence, and their kids.
In the AOD sector, which Humphreys et al. (2005) note was historically developed to work with men, many of the workers and 67% of the clients are male (Nicholas et al. 2012; Australian Institute of Health and Welfare 2016), and a gender neutral or individualist analysis of the link between alcohol and violence predominates (Humphreys et al. 2005). By contrast, in the DFV sectorat least, the part of it that deals with victim/survivorsthe vast majority of the workers are female, the clients are almost entirely women and children (Family Safety Victoria 2017; Cortis et al. 2018), and a feminist power and control analysis of DFV prevails (explained further in the following section). This is true despite Victorian use of the term "family violence," which elsewhere in the world is usually associated with a non-gender-sensitive family conflict perspective on DFV. 5 Perpetrator programs have a higher proportion of male workersin Victoria, about 50% (Family Safety Victoria 2017) and nationally about 30% (Cortis et al. 2018). But these programs still usually come from a feminist perspective: in Victoria, perpetrator programs are largely delivered as men's behavior change programs strongly influenced by the gender-informed psycho-educational Duluth model, where programs are largely delivered in group settings and participants are encouraged to think critically about the power relationships and beliefs about gender roles that underpin their violence (State of Victoria 2016, vol. III, p. 256). They are accredited by the feminist-oriented NGO No To Violence. 6 The DFV sector tends to work from a philosophy of empowering victims (Macy and Goodbourn 2012) and increasing perpetrator accountabilityit aims for men to take responsibility for their violence. Its objective has been to support women to understand that the violence is not their fault, but rather stems from men's sense of entitlement to control women, and attitudes that support or enable the use of violence to do so (Our Watch et al., 2015).
The AOD sector has employed a more "medical" approach (Humphreys et al. 2005), focused on the individual (rather than broader societal factors), and seeking to reduce the stigma of addiction by framing it as a disease or disorder. One study examining the discourse of Victorian AOD treatment providers found that they tended to alleviate the guilt and shame of substance abusers by referring to the "diseased" or "hijacked" brain (Barnett et al. 2018). This "medical" model of addiction as disease or disorder can be seen as allowing men to shift responsibility for violence. This approach sits uncomfortably with the DFV sector, which has fought for the ability to name men as violent, and for men to take responsibility for this violence (Pease 2011), regardless of their relationship to AOD. In this context and with these perceptions of the way the AOD sector works, many DFV sector workers and researchers feel that their advocacy work could be undone by allowing drugs and alcohol to be seen as a "cause" of violence.
Community attitudes to alcohol, drugs, and violence A significant minority of the Australian population continues to hold views that drug or alcohol consumption can excuse violence or diminish the responsibility of the perpetrator. The three most recent iterations of the nationally representative National Community Attitudes Towards Violence Against Women survey asked participants whether "domestic violence can be excused if the offender is heavily affected by alcohol." In 2009, 8% of participants agreed that it could, and in 2013, 9%. Encouragingly, this dropped to 5% in the 2017 survey (Webster et al. 2018). Graham et al. (2011) note that in some cultures, people may consume alcohol before engaging in violent behavior in the belief that this behavior will be excused due to the effects of alcohol (see also Humphreys et al. 2005;Bennett and Bland 2008;Rothman et al. 2011). Related to this issue is the concept of control and its relationship to violence. The notion of "power and control" has been central to the domestic violence movement's attempts to reframe the cultural understanding of domestic violence from an apolitical, individualized problem to a social problem with roots in structural systems of gender inequality (Lehrner and Allen 2008, 220). This framing interprets DFV as "an intentional pattern of abusive behaviors" by one person over another, resulting in "the establishment and maintenance of the abuser's power and control over the other (Lehrner and Allen 2008, 226).
Crucially, "power and control" framing is indexed not to individual men's pathology or addiction, but to "macro-level entitlement beliefs and attitudes, the expectation that men are powerful and have control over 'their' women" (Lehrner and Allen 2008, 226). While most feminist theorists do not deny that some violence is linked to psychopathology or other individual differences, they generally seek to connect psychological analyses to an understanding of the unequal distribution of power and socially structured patterns of male-female relations (Wendt and Zannettino 2015). Violence is seen to result from socially and structurally supported choice rather than only reactive anger. Thus, attributing causation to drug and alcohol addiction not only implies a lack of control on the part of abusers, it also moves the analysis from structural to individual in a way that sits uneasily with the DFV sector.

Causation in different research traditions
In another language-related tension, different research and professional traditions have different understandings of the word "cause." In epidemiological and public health research traditions, it can be acceptable to say that AOD is a cause of violence, viewed as part of a multicausality framework that identifies "component causes" or "contributing causes" of the disease or public health problem. A component cause may not be necessary or sufficient to cause every case of the problem, but a substantial amount of cases may still be prevented if that factor is blocked or removed (Rothman and Greenland 2005). Moreover, the strength of a causal factor can be measured by the change of the problem frequency when the factor is introduced or removed (Rothman and Greenland 2005). Addiction researcher Leonard (2005, 423) argues that since no single type of evidence is sufficient to definitively demonstrate a causal association between heavy drinking and intimate partner violence, the convergence of evidence from varied sources (e.g. longitudinal and cross-sectional studies, treatment studies, experimental studies) should allow us to conclude that heavy drinking is a "contributing cause" of intimate partner violence (see also Leonard and Quigley 2017).
In the feminist domestic violence research tradition, many actors argue that because not all men who misuse drugs and alcohol are violent and not all violence is associated with drug and alcohol use, these substances should not be construed as causal factors (Humphreys et al. 2005;Braaf 2012;Noonan et al. 2017). Because gender inequality and violence-supportive attitudes 7 are seen in the DFV field as more ubiquitous factors than AOD abuse, the latter is framed as "contributing" to or "reinforcing" the violence, or "co-occurring" with the violence (Braaf 2012;Our Watch 2014;Our Watch et al. 2015). In other words, it is seen to enable or exacerbate violence that is already there, and increase its frequency and severity, but not to cause the violence in the first place. DFV workers can be uncomfortable with implications that substance use can cause violence, fearing "letting any suggestion through that treating the issue of substance use would cure the problem of violence" (Humphreys et al. 2005(Humphreys et al. , 1312. As I have outlined, there are several issues to unpack in understanding why the question "can alcohol and drugs be called a cause of DFV?" is so contentious. Many DFV sector workers and researchers are concerned that labeling AOD as causal will focus attention on that and away from gender issues (Braaf 2012). The "medical" model of addiction as disease or disorder favored by the AOD sector can be seen as allowing men to shift responsibility for violence to substance abuse and addiction, while the DFV sector has long been focused on naming men as violent. On a community level, intoxication is often seen as reducing abusers' control of their actions, and thus their culpability for violent acts. In this context if alcohol is seen as a cause, this shifts responsibility from the violent person to an external agent, while simultaneously moving the problem analysis from a societal-level recognition of men's power over women to an individual-level focus on substance abuse. Finally, different research traditions use the word "cause" in different ways, leading to clashes and misunderstanding between public health/epidemiology policy actors and DFV policy actors.

Alcohol and other drugs at the Royal Commission into Family Violence
Given these complex sensitivities about the role of drugs and alcohol in DFV, it is interesting to examine how these matters were treated in the RCFV's public hearings on the subject. For my work on gender and family violence at the Commission (see e.g. Yates 2018a, 2018b), I interviewed 20 people who had participated in the Commission process, mainly as expert witnesses. These interviews were semi-structured (Ercan and Marsh 2016) and anonymous due to the sensitive nature of the material. Five of these were called for their expertise in the role of alcohol and other drugs in DFV. I also interviewed the Commissioner and one of the two Deputy Commissioners, who agreed to be identified.
Consistent with the literature, interview data indicated that AOD and causation was an issue at the Commission. Deputy Commissioner Faulkner, when talking about the dominant themes that she had heard throughout the Commission process, recalled: [The role of AOD] was a very contested view, because the gender equity argument says we shouldn't take too much notice of the drug and alcohol factor because there are many people who use drugs and alcohol who are not violent.
The extent to which actors are careful to say that AOD is not causal formed a 'boundary marker', according to family violence researcher P06: … these sectors aren't siloed for no reason. … And one of them is are you really clear that domestic violence isn't caused by drug and alcohol abuse. That's a boundary marker. You say the wrong thing in that area, you lose your credibility.
In the experience of P06, actors on the wrong side of that boundary marker are at risk of antagonizing or not being taken seriously by the DFV service sector and others who specialize in gender-based violence. P07, an AOD practitioner, had also come across these sensitivities when working with representatives of the DFV sector on a cross-sector practice document: So when we came to this whole issue of causality, I didn't really realise it was an issue until we started talking to some family violence organisations who said "we don't want family violence to be confused with drug and alcohol, and there's a lot of people who try and make out that it's drug and alcohol issues that cause family violence, and that's wrong".

Public health versus feminist framing in expert witnesses
The Commission was not directed in its terms of reference to investigate responses to AOD-related family violence. However, it devoted a day of public hearings to the matter (one of 23 topic-related "modules" covered in the 25 days of hearings), and specifically sought consultation with several high-profile researchers and advocates working in this area. The first panel of witnesses appeared for approximately a third of the time allocated for the AOD hearing, and comprised drug and alcohol researchers and advocates from different organizations and research traditions. Their points of agreement and disagreement formed the most interesting problem framing debate of the day, so I focus on the first panel session in this discussion. What I observed was a difference in framing between those who seemed to take a public health or epidemiological approach to the problem on the one hand, and feminist researchers and advocates on the other.
Australia has several formal and well-run policy advocacy coalitions that work to address the harms of alcohol abuse ("it's public health," remarked interviewee P19, "they're great organizers"). A strong anti-alcohol lobby saw the RCFV as a potential vehicle for the enactment of a broader policy aim to restrict the supply of alcohol in Australia. They were quite successful in getting their issue on the Commission's agenda. This was aided, according to interviewee P14, by the voices of female victim/ survivors from the Commission's consultation stage combined with a perceived reluctance from the family violence sector to address alcohol issues: … they said to me "we've just spoken to hundreds of women, and so many of them talked about the role of alcohol. And yet we talk to the violence against women sector, and they say it's got nothing to do with alcohol." Five alcohol policy coalitions or research centers made submissions to the Commission. Most of these emphasized as a top priority the prescription of (degendered) population-level interventions to reduce the physical and economic availability of alcohol. They made these recommendations on the basis that problematic alcohol use is "one policy factor amenable to change, with a robust body of evidence supporting interventions that can make a decisive impact on reducing alcohol-related harms" (National Alliance for Action on Alcohol submission, p. 16). While gender was often mentioned as a factor by these groups, it tended to be seen as one category variable among many, rather than an organizing principle that structures society and affects the experiences and opportunities of different groups of people. It is not a focus of their analysis. Expert witness Michael Thorn, CEO of an anti-alcohol advocacy organization, reported that for his organization FARE, talking about the role of alcohol in DFV "has been an exercise in careful diplomacy … because the characterisation of family violence through the lens of gender equity is very sensitive." While in Thorn's view gender was "first and foremost" in everyone's considerations about how to respond to the problem, we should not ignore, "for political reasons or whatever reason," what the evidence says about the contribution of alcohol. He reported the strategy of his organization to be "just looking at what the data says," which implied firstly that evidence can be politically neutral, and secondly that that a neutral evidence-based position allowed his organization to side-step a political issue.
While feminist researchers such as expert witnesses Professor Cathy Humphreys and Ingrid Wilson agreed with the public health-type advocates on many points, there were some clear differences of emphasis in their framing. These actors did not shy away from discussion of the relationship between AOD and DFV, but they were clear that gender is an important part of the equation. Thus, while they advocated for the role of AOD to be recognized and responded to in this policy area, they did so in a way that demanded a continuing focus on gender and men's violence against women. In other words, they tried to address the problem while staying on the right side of the "boundary marker." Their recommendations for addressing the problem were also rooted in gender awareness. For example, they aimed to increase the AOD sector's understanding of the gendered dynamics of DFV, and they wanted service providers in both sectors to recognize that many women who are abused turn to AOD as a coping mechanism.
When giving evidence to the Commission, Humphreys was keen for gender to be seen as a key causal factor, and AOD as a contributing factor: the Our Watch analysis … does look at two causal factors being gender inequity and violence supportive attitudes. So they have got that very clearly, and then they have a range of contributing factors as well, and alcohol and drugs being one of the contributing factors. So I think that there's potential … that we can be on the same page and that there is a common language and some common understandings there that we can sign up to or that we could champion (RCFV public hearings transcript, p. 611).
Fellow panelist A/Prof Peter Miller (an addiction researcher) disagreed, responding that "it is more than just attitudes and gender inequity": I think we have really strong evidence from a big body of longitudinal evidence to show that child abuse, experience of child abuse, growing up in adverse surroundings, in bad family settings, having peers -these are major predictors that go beyond just attitudes. We also have to talk about genetics (RCFV transcript, This exchange is a good example of the tension between DFV violence as having primarily societal causes (such as gender inequality) or having primarily individual causes (such as AOD or family of origin factors).

The Royal Commission's treatment of AOD and family violence
The Commissioners were determined to incorporate factors other than gender into their investigation of the Victorian family violence service system. According to Commissioner Neave, their terms of reference required them to look beyond men's violence towards women, and they very much treated the terms of reference as "ground rules." They wanted to operate innovatively and "explore things that added to our knowledge, rather than repeating what had been said in so many other reports." For these reasons, coupled with submissions and community consultations that repeatedly referenced the role of AOD, the Commission decided to focus some of its attention on this issuedespite push-back from the family violence sector, who were concerned that this would dilute the message about gender. The Commission argued that a focus on alcohol consumption did not excuse violent behavior: rather, "more extensive engagement with all of the risk factors that contribute to family violence is required to appropriately respond to violence, to support victims, and to hold perpetrators to account" (State of Victoria 2016, vol. III, p. 300).
This willingness to consider AOD led anti-alcohol advocates to hope that the Commission would make "courageous recommendations" (as P19 put it) such as reducing the density of liquor licenses, reducing trading hours, or regulating alcohol advertising. However, possibly due to the tight 14-month timeframe, the RCFV decided to leave the technical and politically controversial aspects of alcohol supply and regulation to the Government's forthcoming review of the Liquor Control Reform Act 1998. Rather than recommend some of the "harder" policy options such as alcohol regulation, the Commission decided on "softer" options around workforce training and service availability and integration. While these measures were certainly called for by researchers and advocates, they were not the top priority of the loudest public health/ epidemiology voices. On the other hand, the Commission did appear to attend to the concerns of feminist actors who wanted to make sure that increasing attention to drugs and alcohol did not diminish perpetrator accountability or water down the focus on gender as a key driver of DFV. Many of the recommendations relating to alcohol and drugs, when inspected closely, contained hints or safeguards relating to gender and accountability (for example recommending input from feminist-oriented organization ANROWS on new AOD/DFV sector coordinated perpetrator interventions, or recommending that family violence advisors be located in AOD servicesthus bringing family violence expertise into the AOD sectorbut not the reverse).
The Commission's discussion of causal factors positioned AOD as an individual risk factor that "reinforce[s] the gendered drivers of family violence" (State of Victoria 2016, vol. III, p. 248)along with other factors like mental health, exposure to violence, and socioeconomic inequality. It emphasized that much family violence is not linked to AOD misuse. This framing mirrors very closely the input of feminist actors. The report positioned intoxicated perpetrators, particularly men, as responsible for their own actions, and suggested that cultural norms rather than any effect of the alcohol itself are to blame for any disinhibition and subsequent violence. However, it did not explore ways that gender, substance use and DFV might interrelate, an analysis gap also noted by Moore et al. (2017) in their review of research on alcohol and violence that has been influential in shaping Australian policy debates.

Discussion
As outlined above, the main framing controversy of the alcohol and other drugs module related to whether AOD can be called a "cause" of domestic and family violence. Public health advocates tended to focus on the harms of addiction and talked about gender as a categorical variable (i.e. men and women), rather than a social construct related to the distribution of power and resources. In relation to the causation "boundary marker" described earlier, these actors placed themselves on the non-feminist side because they did not strongly emphasize that AOD does not cause DFV. In fact, one addiction researcher argued that it is logical, with reference to epidemiological analytical techniques, to say that alcohol is a cause of DFV. They also called for regulatory reforms to limit the availability of alcohol and control the behavior of problem drinkers. As the Commission's report noted, the submissions of anti-alcohol organizations emphasized as a top priority the prescription of (degendered) population-level interventions to reduce the physical and economic availability of alcohol. Indeed, Deputy Commission Faulker reflected: Hart and Moore (2014, 410) note that alcohol availability strategies are the primary tool of public health advocates more broadly (i.e. not just in relation to DFV policy), an approach that materializes alcohol as a "powerful (somewhat malign) agent capable of 'causing' unwanted outcomes".
In the view of Kantola (2010), degendered public health framing is concerning for DFV gender equality advocates, because it positions DFV as part of a bigger public health problem rather than part of a bigger gender equality problem, with a commensurate focus on eradicating health problems instead of gender inequality or the gendered intersections between problematic substance use and DFV (see also Flood 2015). Public health framing also lends itself to market-based and economic arguments, as a challenge to public health can endanger government priorities such as jobs, economic growth, and reducing the public burden of the health system (Kantola 2010). This does not sit well with feminist arguments from the violence against women movement about the right of women to safety and personal integrity.
Gender equality advocates at the Commission's AOD hearing framed the problem in gender equality terms and were careful to stay on the feminist side of the boundary marker by arguing that AOD does not cause DFV. Humphreys in particular argued for consistent language use in this area; gender inequality and violence supportive attitudes should be described as "causes," and AOD a "contributing factor." As interviewee P16 reported, some feminist policy actors are interested in addressing AOD as well as gender, and are frustrated that discussion of the intersection between AOD and DFV seems to have stagnated in the Victorian DFV sector. According to P16, many people in the Victorian DFV sector tend to be so concerned that AOD will be seen as causal, and that this will take attention away from gender inequality and diminish perpetrator accountability, that they are hostile to opening up this discussion. P16 found that their conversations with feminist DFV practitioners tended to stop at "alcohol and drugs do not cause family violence," with no apparent appetite for addressing the intersection between the two issues. The gender equality advocates appearing before the Commission's AOD hearing did want to explore the relationship between gender, AOD abuse and DFVwithout letting substance abuse serve as a distraction from issues of gender inequality or allowing a shift of responsibility from perpetrators to an external factor. For example, Australia's sporting culture and its association with masculinity and heavy drinking formed part of their analysis, as did norms about drinking allowing men "time out" from the normal rules of social behavior.
A way forward: focusing on how gender and substance abuse interrelate One way forward is to leave behind sensitivities about whether gender or individual factors such as substance abuse are most responsible for the harms of DFV, and instead to understand how AOD and gender combine to influence the perpetration and experience of DFV, as well as throwing up barriers to help-seeking. In other words: substance abuse affects the problem of DFV in gendered ways. Feminist DFV researchers and practitioners should take heart from the fact that it is possible to take substance abuse into account from an explicitly feminist standpoint that does not ignore the role of gender hierarchies in causing DFV. Work by scholars such as Moore et al. (2017), Hart (2016), Gilchrist et al. (2019), Lindsay (2012), Towns et al. (2011), andMahalik et al. (2007) has engaged fruitfully with the way that substance abuse and performance of masculinity, or adherence to traditional notions of masculinity, are linked to violence.
For example, Hart's (2016) case study of an Australian suburban football club showed how different modes of masculinity that held sway inside and outside its clubrooms influenced the drinking behavior and aggression of its members. Inside the club, reforms to services and infrastructure had encouraged masculine norms of drinking moderately, in a manner that did not alienate or inspire fear in women and children. Outside in the carpark and on the football ground itself, "the drinking began at the ground, then moved to a private home or licenced venue" (Hart 2016, 308). This drinking behavior was associated with performances of aggressive masculinity. Towns et al. (2011) reviewed over 50 TV advertisements for beer and other alcohol products commonly available in New Zealand, analyzing their constructions of masculinities and intimacy and assessing their relevance to DFV prevention. A recent meta-ethnography of substance use in qualitative intimate partner violence studies found that AOD abuse is intertwined with intimate partner violence in very gendered ways related to structural power differences and expectations about men as providers and protectors who control the relationship, and women as mothers and subservient partners (Gilchrist et al. 2019) Gender is also relevant to the experiences of victim/survivors who use alcohol and other drugs. As many scholars have pointed out, gender processes place women as primary caregivers of children, solely responsible for their care and wellbeing (Wendt and Zannettino 2015;Hester 2011;Buchanan 2018). This intersects uneasily with the medicalized understanding of substance abuse recovery as an individualistic journey, as discussed earlier. This means that female DFV victims who use AOD as a coping strategy are likely to have primary care of their children and see the care of children as crucial to their responsibilities and identities as women. However, AOD treatment services may view their children as a distraction or impediment for their recovery from substance abuse. Thus treatment centers are often not child-friendly, which is a barrier for these women to access treatment (Salter and Breckenridge 2014). An understanding of both gender and AOD treatment services is important for addressing this complex intersection of structures and expectations.

Conclusion
In this article, I have explored the controversy about whether it is acceptable to say that AOD is a cause of DFV. Understanding the reasons behind this will be useful to practitioners working in policy and service delivery areas related to alcohol, drugs and family violence. Because of this controversy, discussion of gender in the AOD module at the Royal Commission into Family Violence tended to focus on whether or not AOD is a cause, neglecting the finer points of how gender and AOD interrelate. In other words, my analysis suggests that a hurdle has to be overcome before any nuanced discussion of gender and AOD can take place. Two feminist witnesses at the RCFV hinted at ways in which Australia's drinking culture (particularly as it applies to men) is related to family violence, but these analyses were not a focus of the public hearings and were not reflected in the Commission's report. There was no discussion of how substance abuse (particularly drinking) and constructions of masculinity might relate to DFV, although such explorations do exist in the literature (e.g. Galvani 2004;Peralta et al. 2010;Towns et al. 2011Towns et al. , 2012Lindsay 2012;Hart 2016). I suggest that AOD should be incorporated into gendered frameworks of DFV by paying attention to the gendered ways that the abuse of alcohol and other drugs affects the perpetration and experience of DFV. Encouragingly, Victoria's 2017 DFV prevention strategy notes an intention to "[a]ddress the intersections between social norms about alcohol and gender" (Victorian Government 2017, 33). Based on the literature above, this should extend to structural issues as well as social norms, and to other drugs in addition to alcohol. 2. I use the word 'problem' here in the same sense as frame theorist Carol Bacchi (2009, x-xi), who describes it as a condition in society that is seen as needing to change, and that government policies aim to address. 3. I used critical frame analysis, which holds that a policy text will always contain a frame: an implicit or explicit representation of a problem (the problem "diagnosis"), connected to an implicit or explicit solution (the "prescription") (Verloo 2005(Verloo , 2007. 4. However, the two categories are not mutually exclusivesome offenders may be counted twice as they appear to be under the influence of both alcohol and drugs. Victoria Police members are required to use subjective tests at family violence incidents, thus the terminology "definitely" or "possibly." There is greater uncertainty regarding the presence of drugs than alcohol (more of the drug-related incidents are categorized "possibly affected") (Sutherland et al. 2016). 5. For a discussion of terminology and different approaches to DFV (including those concerned with gender, inequality, individual psychopathology, and intergenerational causes), see Yates (2018b, chaps . 1 and 3). 6. See http://www.ntv.org.au/about-us/our-vision/. 7. These are described in VicHealth (2014, 3) as those that justify, excuse, trivialize, and minimize violence, or shift blame from the perpetrator to the victim. Examples given include the idea that partner violence is justifiable if a woman is unfaithful, or that rape is only rape if the woman physically resisted.

Disclosure statement
No potential conflict of interest was reported by the authors.