Rethinking Male Peer Support Theory: Social Network Responses to Young Men’s Violence Against Women

Male Peer Support Theory (MPST) is one of the few principal theories about masculinity and men’s violence against women. The theory foregrounds the role of social networks in encouraging violence. This article offers a critical discussion of MPST, particularly the assumption that social networks primarily support violence. Drawing on a qualitative study of young men perpetrators in Sweden, we suggest that the concept of response is better suited than support in capturing the diversity of social network responses to violence. In our data, there were few stories about unmitigated pro-abuse support. Instead, we found responses that unequivocally condemned violence, as well as ambiguous and transformative responses. We suggest that such responses be understood in relation to changing attitudes concerning violence.

Men's violence against women remains a significant problem in many societies. Feminist scholars have highlighted how violence is part of the power and control strategies that individual men exercise over individual women, as well as how such violence both derives from and reproduces gender inequality throughout society (e.g., Dobash & Dobash, 1978).
Research on masculinity in particular sheds light on the social processes surrounding men's violence against women. This includes the dynamics that incite sexual harassment and abuse in the first place (e.g., Messerschmidt, 2004), as well as how men and their social networks manage and respond to the violence after it has occurred (e.g., Gottzén, 2017). The perhaps most explicit theory about the connections between men's relations to other men and men's violence against women partners is the Male Peer Support Theory (MPST) that has been developed by criminologists Walter DeKeseredy and Martin Schwartz over the last few decades. The theory emphasizes that while men's violence against women takes place in patriarchal society, it is also shaped by the specific social and relational settings in which these men move. This focus names a crucial factor in understanding men's violence against women: male peer support.
Despite the publication of MPST research during the last three decades, there has been little theoretical debate about the merits and drawbacks of the theory, with a few exceptions (Hearn, 1998a;Morris & Ratajczak, 2019). In this article, we offer a critical discussion of MPST and suggest that the role of social networks in relation to violence needs to be understood in a more diversified way. In particular, we propose that "response" is a better concept than support as it is a more comprehensive concept and may include a variety of at times ambiguous and contradictory ways in which friends, family members and others deal with men's violence against women. Drawing on qualitative interviews with young men perpetrators in Sweden, we suggest that social networks may not only offer pro-abuse support, which is the focus of MPST. Based on our data, we also identify three other kinds of responses: unequivocal condemnation, ambiguous responses, and transformative responses. Such responses need to be understood in relation to current cultural changes where support for intimate partner violence is declining around the world (Pierotti, 2013). However, before elaborating these arguments further, we will first provide a brief review of masculinity theorizing on violence against women, as well as present the main features of MPST and its development.

Masculinity Studies and Violence Against Women
The contribution of MPST is best understood in the context of research on masculinity and men's violence against women (DeKeseredy & Schwartz, 2005;Berggren et al., 2021). One influential strand of this research-primarily developed by Messerchmidt (1993) in his Structured Action Theory-has been inspired by socialist feminism and argues that violence constitutes a resource to "accomplish" masculinity (Messerschmidt, 1993), particularly in communities where physical aggression is admired (Connell, 2002). Messerschmidt (1993) also emphasizes how men relate differently to violence depending on historical transformations as well as class and race relations. Another prominent approach, at times referred to as "discourses of violence" (cf. Gadd, 2000), has focused on how men talk about and justify their use of violence. This research often has its roots in radical feminism, but is also influenced by accounts research (cf. Hearn, 1998b). The focus here is on the fact that individuals need to give culturally acceptable explanations of their problematic actions, for instance by providing "vocabularies of adjustment" (Cressey, 1953). Through such explanations, the perpetrator not only minimizes his violence, but also produces himself as a "moral self" (Hearn, 1998b;Mullaney, 2007). A third approach, psychosocial criminology, has attempted to bridge the gap between individual and social factors by combining poststructuralism and psychoanalytical approaches in exploring violent men's psychosocial trajectories and ambivalent investments in masculinity and violence (e.g., Gadd, 2000;Jefferson, 1994).
A fourth, recurring theme in the field emphasizes the role of social networks, and can be found across different masculinity studies traditions. This scholarship emphasizes that abusers are not necessarily neither deviant nor "loners," but that violence is carried out in a relational setting, where family and friends often learn about the abuse (Gottzén, 2016). Hearn and Whitehead (2006) argue that in order to understand violence against women, we need to explore men's homosocial relations. Similarly, Kimmel (2008) stresses how boys and men are influenced by male peers in how they relate to sexuality and violence, and Messerschmidt (2004) has shown that young men may receive pro-violence advice from friends and family members. These arguments resemble Connell's (2005) notion of complicit masculinity, which highlights how men may help sustain gender inequality without necessarily "being the frontline troops of patriarchy" themselves (p. 79). Others have maintained that (young) men may feel pressured to live up to masculinity norms emphasizing sexual prowess and, due to that, may feel motivated to use violence (e.g., Kanin, 1967;Messner, 2005;Reidy et al., 2014). Male Peer Support Theory is to date probably the most systematic attempt to conceptualize the role of the social network and thus constitutes an important contribution to theorizing masculinity and violence. In the following, we present the theory in more detail.

The Development of Male Peer Support Theory
DeKeseredy and Schwartz (DeKeseredy, 1990;DeKeseredy & Schwartz, 1993;2009;2013;Schwartz & DeKeseredy, 1997) present a theory about the link between men's peer relations and violence against women. They draw on radical feminist theory but focus on how patriarchal ideologies operate in specific male contexts, in particular on how the perpetrator's friends and social network support abusive and violent behavior against women. They also draw on empirical findings showing that peers provide learning environments for inflicting and sustaining partner violence (e.g., Gwartney-Gibbs et al., 1987). Male Peer Support Theory hence proposes that violence against women is ultimately caused by societal patriarchy, and argues that men adhering to patriarchal values and practices are more likely to be abusive towards intimate partners (Schwartz & DeKeseredy, 2008). Crucially, patriarchy is understood to be mediated through male friends. As DeKeseredy (2011) has pointed out, "many wife beaters, date rapists, stalkers, and so on are […] 'companions in crime.' In other words, they are heavily influenced to engage in woman abuse by male peer support" (p. 69). Male peer support is defined as attachments to male peers and the ways in which they encourage and legitimate woman abuse (DeKeseredy, 1990).
In the original version of male peer support theory, DeKeseredy (1988;1990) argued that some men turn to others when facing "relationship stress." In response to such stress, DeKeseredy suggested that male friends may under certain conditions encourage and justify physical, psychological and sexual abuse of women. Some years later, DeKeseredy and Schwartz (1993) developed a modified and more complex peer support theory. This framework took into account not only the individual factors of (relationship) stress and (peer) support, but also a number of issues, including patriarchal norms and values as well as heavy use of alcohol and membership in all-male groups, primarily fraternities and athletic teams (Schwartz & DeKeseredy, 1997). In these groups, men learn a narrow conception of masculinity and the sexual objectification of women. Such groups often also maintain group secrecy, which help men to avoid disclosing their friends' problematic behavior to others, and in this way group secrecy may sanction violence (Schwartz & DeKeseredy, 1997). Moreover, Schwartz and DeKeseredy (1997) take social and cultural forces into account, emphasizing the role of a sexist and patriarchal culture. The modified version of peer support theory thus incorporates a range of social forces that may contribute to men's violence against women. The idea of male peer support, however, remains central to the theory. In this way, MPST foregrounds the importance of homosociality and men's social bonds. This is particularly evident in a theoretical paper, in which Godenzi et al. (2001) combine MPST with criminological social bond theory. While gender-blind social bond theory posits that strong social bonds decrease criminal activity, the authors argue that it may be the other way around when it comes to violence against women. Since gender inequality is an unacknowledged norm in many institutions, bonds to conventional peers and institutions may in fact increase men's use of violence against women.
The empirical research exploring the connections between male peer support and violence against women has primarily been quantitative (e.g., DeKeseredy & Kelly, 1995) and has conceptualized peer support in two ways: as (1) attachments to male peers who physically, sexually and psychologically abuse their dating partners; and as (2) having friends who verbally encourage the physical, sexual and psychological abuse of dates or girlfriends in certain situations (such as challenges to patriarchal authority). These notions form the basis of two independent variables: attachment to abusive peers and (pro-abuse) informational support (e.g., DeKeseredy et al., 2018). Attachment to abusive peers is measured by an index of three items, requesting participants to disclose how many of their friends they know have been sexually, physically or emotionally abusive towards a woman date or partner. Informational support is measured by an index of seven items, where participants are asked whether any of their friends told them, for instance, that "[i]t is alright for someone to hit a date in certain situations" and "[y]our dates should have sex with you whenever you want." Male Peer Support Theory has been used to study sexual assault on college campuses (Schwartz & DeKeseredy, 1997), men's violence against women in rural areas (DeKeseredy & Schwartz, 2009), and young men's violence in intimate relationships (DeKeseredy, 1990). Summarizing the findings, DeKeseredy et al. (2017) argue that "having friends who offer such advice is one of the most powerful determinants of whether a male engages in physical, sexual, or psychological assaults on intimate female partners" (p. 170). While the empirical evidence regarding male peer support and violence against women is based on the two measurements derived from the original male peer support theory, the modified peer support framework, in contrast, is a complex multifactor theory that is much harder to measure or test. In fact, according to Schwartz and DeKeseredy (1997), it has more value as a heuristic or teaching model than as a predictive model; it is primarily meant to summarize complex literature.
Recently, MPST has been developed in a somewhat new direction (e.g., DeKeseredy et al., 2017;2018;. DeKeseredy et al. (2019) now contend that almost all peer support research has focused exclusively on men's violence against women. Thus, it has only looked at relations between male peers and ignored how female members of peer groups may also legitimate violence. Hence, this new research uses the more specific concepts of "negative peer support" and "pro-abuse peer support" (used interchangeably) rather than "male peer support." The new research also conceptualizes peer groups as mixed-gender rather than as strictly homosocial, and explores victimization among heterosexual, gay, lesbian and bisexual college students. Moreover, while MPST research has previously explored connections between perpetrators and their peer support, the newer studies instead explore the relation between victims and pro-abuse peer support (DeKeseredy et al., 2017;2018;. Male Peer Support Theory has also been used by other researchers focusing on violence (e.g., Rosen et al., 2003;Williamson & Silverman, 2001) as well as on a wider range of topics, including anti-feminist fathers' rights groups (Dragiewicz, 2008), and resistance to women in policing (Franklin, 2007). Yet, there has still been little theoretical discussion of MPST. One exception to this is (Hearn, 1998a), who has called for more attention to the ambiguity of men's support for men who have been violent to known women. We agree with DeKeseredy and Schwartz that social networks are crucial in relation to men's intimate partner violence against women. However, we contend that MPST in some respects has articulated these issues too narrowly, and following (Hearn, 1998a), we argue that more attention needs to go into thinking through the ambiguities concerning how social networks deal with issues about violence and abuse. We argue that one way to take ambiguity seriously is through focusing on response rather than support to violence. It is increasingly recognized that the response of social networks is a crucial factor in determining whether violence will continue, cease or escalate further (Hydén et al., 2016;Klein, 2012). This is also evident in the contemporary language of violence prevention, where "secondary prevention" refers to interventions that take place immediately after a violent event has occurred (Flood, 2018), and where some prevention programs aim at engaging "bystanders," particularly boys and men, to intervene against harassment and abuse and "challenge violence-supportive norms in their networks" (Bruno et al., 2020, p. 142;, cf. Katz et al., 2011).
We suggest that social network responses may be much more diverse than just providing unmitigated support for violence. In this paper, we define response as actions and reactions aimed at the violence and the perpetrator. It is a broader concept than support as it considers behavior that is not necessarily approving, helping or encouraging woman abuse. Social network responses depend, among other things, on gendered norms of violence, the knowledge individuals have about the violence and their perceived relationship to, and responsibility for victims and perpetrators. Response could thus include not only pro-abuse support, but also condemnation of violence, measures to end the abuse as well as practices with equivocal and contradictory intentions and outcomes.

Method
This paper draws on data from a qualitative interview study of young men perpetrators in Sweden, conducted 2016-2020. The study was motivated by a relative lack of data on youth intimate partner violence in Sweden (cf. Korkmaz &Överlien, 2020). The focus was therefore on young men's experiences of using violence or abuse against intimate female partners, and on the responses from their social networks. Participants were recruited through treatment centers in different parts of Sweden, as well as through advertisements on social media and on university campuses. In total, we interviewed 14 heterosexual men (17-45 years of age) who had used physical or sexual violence in intimate relations under the age of 25. The participants were predominantly white and came from working-class and middle-class backgrounds.
The interviews took place in locations such as at a treatment center or on a university campus. Each of the interviews were conducted by one of the members of the project group, which consisted of two male and one female researchers. The interviews were qualitative and teller-focused (Hydén, 2014), that is, they were focused on enabling the participants to tell their story about their use of violence and about the responses from their social networks. The interviews were 1-3 hours in length, and have been transcribed with a focus on content. For the sake of anonymity, all names have been changed, and we have removed or replaced some details.
Inspired by MPST, we expected to primarily find stories where the men experienced pro-abuse support, but it turned out that the responses to their violence were more diversified. In fact, we found no distinct examples of unmitigated, pro-abuse support in our data. As we coded the data thematically, we instead identified three new categories that could describe how the social networks tended to deal with the violence and the perpetrator: unequivocal condemnation of violence, ambiguous responses, and transformative responses to violence. In the following, we discuss these three forms of responses, as well as why we did not find any unmitigated support. The excerpts and cases presented here are representative of the sample and illustrate the three types of responses identified in the data.

Unmitigated Support
Following Kanin (1967), Schwartz and DeKeseredy (1997) argue that peer groups may help reduce perpetrators' conflicted feelings after committing violence. This can be done by providing what Kanin, following Cressey (1953), refers to as "vocabularies of adjustment," that is, discourses that normalize the use of violence. Similarly, in (Hearn's, 1998a) study of violent men's social networks in the UK, few of the men's families and friends helped to bring the violence to an end; rather, violence was fueled by a direct, one-sided defense of the man's actions, or indirectly through ignorance. Klein (2012) refers to such responses in terms of informal third-party collusion, which includes "condoning or encouraging the perpetrator's abuse, protecting him from being held accountable, dismissing the seriousness of the abuse, or 'staying out of it'" (p. 100). In contrast to MPST and these studies, we found no clear-cut examples of unmitigated pro-abuse support in our interview data. This is surprising given that MPST research tends to present such "negative" support as the only form of response that peers provide perpetrators. This discrepancy might be due to our relatively small sample, but similar findings also emerged in a previous (and considerably larger) interview study with adult partner-violent men from various social backgrounds in Sweden (Gottzén, 2016(Gottzén, , 2017(Gottzén, , 2019a. A more plausible explanation would be that in these cultural and relational settings, it is not currently possible to unambiguously endorse woman abuse. For instance, 25-year-old Björn has avoided mentioning his violence to his friends since he "knows" that they denounce woman abuse. He argues that, "It's so damn nasty and vile to use force against a woman. It's so damn nasty, so you can't imagine, so why would I want to tell my friends?" Such non-disclosure of one's violence could be argued to help peers ignore signs of abuse, "staying out of it" (Klein, 2012, p. 100), and avoid confronting their friends with uncomfortable questions (cf. Hearn, 1998a). We do not argue that pro-abuse attitudes are completely absent in Sweden, nor that violence against women is not supported at all (as discussed below), but our data does seem to indicate that it is controversial to explicitly condone such behavior in these young men's relational settings. Some young men in our sample suggested that cultural debates and feminist media campaigns about sexual violence had made them aware of their behavior as violence or abuse. This is in line with other contemporary research of young men's stories about sexual violence perpetration (Gottzén, 2019b). A plausible explanation for the difference between our data and (Hearn, 1998a) study from the 1990s would then be that violence against women has been discussed more widely in public in the last decades and that we are currently seeing some, albeit slow, cultural change in regards to norms around violence against women (Pierotti, 2013). An account of how social networks manage men's violence against women must therefore consider wider societal processes, including changes in gendered norms and practices.

Unequivocal Condemnation
A somewhat opposite response from unmitigated support of violence is an unequivocal condemnation of violence. This kind of reaction is present in our data, such as in 23year-old Christopher's case. While he argues that he repeatedly pressured his girlfriend into having sex, he also tells us that she contends that he raped her. When some of his friends found out, they did not offer vocabularies of adjustment but rather condemned the violence quite unequivocally.
Well, it was really tough! It was hard on me. I was fired from my band and lost at least about ten to fifteen friends, who jointly decided to, well, exclude me from their circle of friends.
Christopher was also threatened on social media.
They wrote a number of [public social media posts] saying that I was a woman hater and that if I read this I should watch my back.
[…] That they were going to keep track of which events I was going to, and would threaten me in some way so I wouldn't be able to make new friends. This kind of response made Christopher feel unsafe in certain locations as he thought that some acquaintances would assault him due to his abuse. Some of the young men in our data have indeed experienced other men threatening them when rumors about their abuse spread. This happened to 17-year-old Oskar, who assaulted his girlfriend in school. He recounts a party where he was threatened.
We were at this party, and some guys from [the village] where she lives came, and they came up to me and started arguing and said that I assaulted [my girlfriend], several times, and they sort of wanted to kill me because I'd done it, and then I said that, "that's not what happened; it was just a slap." And they just said, "No." So I don't know, it's hard when they come up like that. You don't know what to say because they don't believe you really.
Since the young men in our study expected to be threatened or condemned by others, they were careful about disclosing their use of violence, as Björn above. Gustav, 24 years of age, explains why he avoids the topic: Many would probably turn their backs on me if they found out that I have aggression issues. Beating a woman, intimidating a woman is also not okay. It's something that's seen, not unusual to be seen, [as] the worst thing you could do, you know. Morally, it's the worst thing to do, somehow. [It's] more OK to hit someone in the pub or commit a theft or something like that, but never against a woman, a physically weaker person and the one you love. "How low can you go?" many would think. I can imagine. Or [that's] my fear.
This example illustrates how the men in our study were reluctant to disclose their abuse to others. As Gustav argues, this was primarily due to that they anticipated critique and condemnation since women abuse is "the worst thing to do." They did not necessarily receive actual condemnation on every occasion, but always expected that family and friends would denounce their violence (cf. Gottzén, 2016). In this sense, the young men could be said to be indirectly influenced by peers and others in their social network to not use violence against women. This could be contrasted to how MPST claims that when a man experiences relationship issues he may turn to his male friends for support, and they will provide guidance that reinforces male domination and entitlement. In contrast, masculinity scholars have often emphasized that men's adherence to gender norms may take place under the forces of peer pressure or gender policing (e.g., Messner, 2005), which suggests that relations between men are often saturated with both positive and negative affect. For example, Kimmel (1994) has argued that men's construction of masculinity is a homosocial enactment intimately connected to fear and shame of not living up to normative ideals of masculinity and thus of facing negative sanctions from other men. While such pressure may motivate men to use violence, in our data, the masculinity norms the young men primarily felt they did not succeed to live up to and, consequently, caused shame was rather that they had abused their partners physically or sexually. As Gustav puts it, "Mostly I'm ashamed of it. Yeah, ashamed to tell, sort of fearing what they would think of me if I tell them." He goes on to discuss how he would react if somebody abused any of his sisters, arguing that he would want the abuser to get out of her life and asserts that "I wouldn't have any compassion." As we have argued elsewhere , partner-violent men are not necessarily esteemed in contemporary Western societies where gender equality discourses have gained influence. Instead, they may be positioned as deviant men who fail to live up to masculinity norms associated with intimate relationships. Condemnations of men's violence against women can also be embedded in other social hierarchies and therefore they need to be understood intersectionally, in relation to inequalities and subjectivities based on, for instance, class, race and ethnicity. While violence may be criticized on a general level, abuse and rape may simultaneously be "mobilized" to differentiate between groups of men (Pascoe & Hollander, 2015;cf. Gottzén & Berggren, 2021). Unequivocal condemnation may at first glance therefore be perceived as a progressive response to violence against women, but it could also generate new hierarchies between men and justify violence against perpetrators.

Ambiguous Response
In contrast to MPST's focus on how peers support violence against women, our data suggests that responses to men's violence and abuse may be contradictory and vague.
We call these "ambiguous responses" as they consist of reactions that neither support violence unmitigatedly, nor unequivocally condemn abuse. Ambiguous responses may, firstly, differentiate between abstract rules regarding violence and the specific case, as when 24-year-old Martin describes how his friends reacted when he assaulted his girlfriend in their presence.
They didn't think it was a good idea. But at the same time, they said that she's difficult to deal with. They understood me, which is sick. They understood me, everybody, everybody I've talked to has understood me.
What is interesting about this example is that it first introduces a condemnation of violence against women on a general level, that "people shouldn't fight," as Martin puts it. After this abstract disclaimer, his friends instead support the use of violence under the specific circumstances. This kind of response may refer to the man's overall difficult and stressful situation, to the woman's general character, or to that she is to blame for something specific. Thus, while individuals may argue that men generally are not supposed to abuse their partners, they could see violence as more or less acceptable in a particular case. They could also provide emotional support to the abuser and, in that way, implicitly reinforce violence against women. Responses may also be ambivalent on the local, specific level, involving a mix of support for, and condemnation, of violence. Violence may, for instance, be condemned initially, but then allowed to recede into the background and not discussed any further. The perpetrator may also be supported initially, but be condemned if the violence continues. The response could be ambivalent already from the start, as in the following example where Martin narrates how his girlfriend's father has supported him in their fights.
Her dad has also been very understanding, actually. Since he has mediated [between us], he often knows what it's all about and he usually has to talk to her and convince her about the things that I haven't managed to convince her of. So he knows that it hasn't been easy for me.
Martin argues that his girlfriend's father has also pointed out that: "The both of you need help. You need help with your aggression and you need help with your low selfesteem, or whatever." The response seems to be negative to violence, as the girlfriend's father points out that Martin needs help with his "aggression," and the father is also said to serve in the role of mediator when the couple has been quarrelling. But the reaction does not include an outright condemnation of violence, as the girlfriend's father "knows" that it has not been "easy" for Martin. He also encourages his daughter to deal with her issues, which may be interpreted as placing some of the cause of the violence on her part. His response is thus ambiguous, as it appears to be focused on making the violence cease, or at least not escalate further, while simultaneously providing Martin "vocabularies of adjustment" (Cressey, 1953;, cf. Schwartz & DeKeseredy, 1997).
Ambiguous responses could be compared with the MPST argument that men turn to peers for support when experiencing stress. In our study, the men experienced relationship or economic stress, or other forms of social or psychological problems, and at times asked for help. Whereas DeKeseredy and Schwarz (2013) focus on how networks are supportive of violence against women, our data rather suggests that responses are ambivalent and that individual network members may simultaneously condone and condemn violence. Our data also shows that responses often are dependent on the social distance and proximity between the perpetrator and individuals in the social network. It seemed easier to unequivocally condemn the violence when the perpetrator was an acquaintance or a friend's friend, while close friends and immediate family members tended to be more supportive. The latter did not necessarily endorse the violence, but according to the interviewed young men, their friends and families often emphasized the specific social, psychological or relational circumstances, which made it possible to disregard or downplay the seriousness of the abuse.

Transformative Response
The fourth category that we suggest is transformative response, which refers to an explicit orientation toward the cessation of violence. Response in the form of unmitigated support of violence is not likely to contribute to desistance since it constructs woman abuse as reasonable under certain circumstances. Unequivocal condemnation of violence, in contrast, could lead to perpetrators reevaluating their behavior and realizing that violence against women is morally wrong and not socially accepted. However, it could also lead to social exclusion and reduced well-being, which could be less beneficial in fostering behavioral change. Ambiguous responses contain elements that may lead to persistence as well as desistance of violent behavior. Transformative responses are those that unquestionably condemn violence while at the same time support the individual man, for instance through sharing information about, and encouraging help-seeking. In our interviews, such transformative response was often provided by the young man's parents. This was also the case for 22-year-old Andreas. During quarrels with his girlfriend, Andreas started to phone his mother for advice, who encouraged him to try to calm down by counting to ten. While he thinks "it's a myth that it helps," he appreciates his mother's message about finding ways to deescalate conflicts and deal with his rage. Andreas has also communicated with his stepfather about his violence, who got crossed and asked, "What the hell are you doing? If you love her as much as you say you do, then why do you behave like that?" When asking Andreas how he experienced the stepfathers' harsh critique, he argued that he understood his reaction. "He said it because he cares about me. And [he] doesn't want anything to happen to me, and not to [my girlfriend] either." Since the stepfather is experienced as both condemning violence and supporting Andreas as a person, we consider his response as transformative.
The role of parents in responses to violence against women also has to be understood in relation to our study's focus on young men. While youth is a life phase characterized by peers becoming increasingly important and young people distancing themselves somewhat from their parents (Pahl & Pevalin, 2005), young people often continue to live at home or have close ties with their parents as young adults, which may turn parents and in-laws into bystanders and at times eye-witnesses of the abuse (cf. Korkmaz &Överlien, 2020). In addition, parents may continue to feel responsible for their children's well-being also when they have moved from home (Holdsworth, 2004). When learning about the violence, parents may provide ambiguous responses (as the case of Martin above) or more transformative responses as the case with Andreas. While parents were pivotal, transformative responses could at times also come from male peers. Gustav describes how he advised one of his friends to seek treatment, mentioning his own positive experiences of help-seeking: He [friend] has big aggression issues and doesn't feel well, so I recommended that he consult [anti-violence organization]. That it is something he ought to try. And in relation to that, I talked a little bit about my own problems too.
In this example, male peers can be not only "companions in crime" as DeKeseredy (2011) argues, but also partners in change. Similarly, while Oskar was unequivocally condemned and threatened by some boys from another village, his friends-who witnessed the assault-seem to have made it clear that they currently support him, but on the condition that he changes his behavior: "Everybody around me […] they don't want to be with me in the end if I keep on [using violence]." Transformative responses are thus reactions that clearly challenge the use of violence while also in some way help the perpetrator, such as in encouraging him to seek treatment or advising how to calm down during relationship conflicts. This finding is in stark contrast to MPST scholarship that primarily identifies male peers as only providing pro-abuse support (e.g., DeKeseredy & Schwartz, 2013). (Hearn, 1998a), however, cautions against a toooptimistic view of the influence of social networks on the cessation of violence. While its influence should not be overstated, the role of the social network is nevertheless important in responding to men's violence. This insight is increasingly recognized in violence-prevention efforts to involve the bystander, that is, to educate members of social networks to play an active, interventionist role in relation to violence against women (e.g., Bruno et al., 2020;Katz et al., 2011).

Conclusions
Men's violence against women remains an immense problem in society, yet the theoretical interest in the issue has been relatively limited among masculinity scholars (cf. Berggren et al., 2021). DeKeseredy and Schwartz' work on MPST constitutes an important contribution to theorizing men's use of violence against women by foregrounding the role of the social network. In this article, we have argued for rethinking or expanding MPST somewhat, in order to account for the variety of ways in which social networks may deal with violence. Drawing on a qualitative interview study of young men perpetrators in Sweden, we have identified four different kinds of reactions they have faced. Although MPST and related research emphasize how social networks support and encourage violence, we found no clear-cut examples of unmitigated support for violence in our data. Instead, there were cases of unequivocal condemnation of violence, as well as of ambiguous response and transformative response.
While there is a theoretical possibility within MPST that support may not be "proabusive," the theory and its application exclusively focus on "negative" support, that is, only on support that-explicitly or implicitly-causes, maintains or justifies abusive behavior. In other words, the assumption is that the social network will exclusively attempt to maintain or defend the perpetrator's violence against women. While our data is somewhat limited, it suggests that the ways that social networks deal with men's violence against women are far more varied than MPST proposes. Thus, we argue that response is a better concept than support since it includes actions and reactions to men's violence that may both reproduce and oppose violence. Evidently, there is much more nuance and ambiguity to the forms of support that the young men in our study faced than the unmitigated pro-abuse support MPST presents.
Our findings also suggest that MPST needs to acknowledge that responses to men's violence against women are processes that occur in time and place. This implies being sensitive and open to change in different cultural contexts. We have suggested that the difference between previous studies highlighting unmitigated support for violence and the more varied responses in our study may in part be connected to changes in masculinity norms. Public discourse in Sweden has in recent years been influenced by a feminist understanding of violence, and violence against women is increasingly being condemned publicly and among many men . Similar processes are taking place in other societies as well (Pierotti 2013), such as in the U.S. (Pascoe & Hollander, 2015). Our point is not that violence against women is completely deviant, but rather that in contexts where gender equality discourses are becoming increasingly influential, men's relationship to violence against women is more intricate than it was previously, and it has become much more problematic to explicitly support violence. Thus, social networks need to be understood within a larger context that is not only characterized by patterns of gender inequality, but also by the possibilities of cultural change. However, it should be noted that condemnation of violence against women is not necessarily an expression of gender equality discourses as it aligns with patriarchal ideals about men and violence, as for instance articulated within the "convict code," which denounces violence against women as women are seen as weaker and more vulnerable and in need of male protection. Being sensitive to change also implies that support is not only given at one time, but over time, and that due to this the nature of support is contingent. This calls for the need of ethnographic, qualitative as well as longitudinal approaches in order to grasp the variety of network responses and support processes to men's violence.

Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Swedish Research Council for Health, Working Life and Welfare (2014-0222).