Tackling violence in intimacy: Interacting power relations and policy change

Introduction

The core issue of this article concerns interventions against violence in intimacy in a complex social order: to understand what multiple and interacting power relations can mean for individual or collective actors’ capacity to influence the world as regards violence in intimacy and for their ability to make a difference through their actions. The issue of children exposed to intimate partner violence is used as an empirical example and policy change in Sweden is the case in point. ¹ There has been research from a range of countries and evidence from women’s refuges about children and intimate partner violence since the 1980s. However, it was not until the second half of the 1990s that the first Swedish empirical studies or reports were published on this topic (Arnell and Ekbom, 1999; Weinehall, 1997). At this time interest in this group of children grew rapidly among Swedish practitioners, policy-makers and researchers, but it was only after the turn of the millennium that significant and safety-oriented changes were made to law and policy, and that support interventions aimed at children exposed to violence started to develop on a broader scale (Eriksson et al., 2006).

Against the backdrop of Sweden’s reputation internationally for advocating children’s rights and for challenging violence against children by parents, the relative silence on children exposed to intimate partner violence may seem somewhat surprising. However, it is clear that the ban on corporal punishment in 1979 was not followed by a general recognition of children’s vulnerability and victimization in a family setting. Physical violence was recognized as child abuse, while exposure to violence against a parent was not. What makes the picture even more complex is the fact that the recognition of exposure to intimate partner violence as a form of child abuse also warrants recognition of violence in heterosexual intimate relations, especially men’s violence against women. Children exposed to violence thus constitutes an ‘intersectional’ issue, shaped by inequality tied to gender as well as age and kinship.

In the article it is argued that to understand and explain recent shifts in policy when it comes to children exposed to intimate partner violence it is important to recognize, on the one hand, the complex social order shaping policy and practice, and, on the other hand, the role of collective actors in the field as well as the opportunity structure enabling social change. The point is that social actors do not find themselves outside culture. Also, challenges to the social order are culturally intelligible. For example, a struggle to support children’s right to protection in a family setting can be enabled by well-established notions of children as objects of adult care, protection and control. Furthermore, interactions between different power relations may create tensions enabling the agency of, in the case discussed here, voluntary organizations or individuals who are trying to bring the issue of children exposed to intimate partner violence to the fore.

As regards the possibilities for creating policy change there seems to be a key difference between the dynamics of age and gender inequality. In the discussion that follows it is argued that subordinating children to adults comes across as a particularly potent way of challenging gendered inequality in the form of men’s violence against women in intimate relationships. Possibilities for social change thus seem to be greater in relation to gender, than age. An empirically sensitive and actor-centred intersectional analysis must be sophisticated enough to grasp complexities such as these if we are to be able to fully understand agency and policy change within a complex social order.

The issue of children exposed to intimate partner violence

Men’s violence against female partners is an issue of direct concern for children (see Hester et al., 2006; Kitzmann et al., 2003; Øverlien, 2007). Almost all children in these families are subjected to the emotional cruelty of seeing, hearing or being aware of violence against their mother. ² In an increasing number of countries this form of cruelty to children is defined as child abuse and in current Swedish legislation children witnessing intimate partner violence are even defined as crime victims.

Researchers have pointed out that the notion of children ‘witnessing’ violence is often too limited to really capture children’s experiences when one parent is violent against the other (Eriksson et al., 2007; Hydén, 1995). The concept of witnessing is easily associated with being a distant eye witness. This is a misrepresentation of children’s position and it can easily lead to an underestimation of the problems children may suffer from. The impact upon children due to their closeness to both the abusive and abused parent may be underestimated. Furthermore, the focus on the sense of vision may downplay the experiences. Children live in the middle of violence and are often involved in some way, for example through attempts to stop the violence (e.g. Eskonen, 2005; Mullender et al., 2002). They can be in the next room or in the same room with their eyes shut. They may not see the violence, but they hear it. They can experience violence after the fact through marks on their mother or on the furniture. Thus children experience violence from one parent against the other by seeing it, hearing it, getting involved in it and/or through the consequences of violence (Edleson, 1999). To capture all of these dimensions of experience the concept of exposure to violence is used in this article, instead of witnessing. For the sake of clarity when dealing with violence within heterosexual partnerships, the concept of intimate partner violence is preferred over other concepts often used in the field, such as, for example, domestic violence, which sometimes includes children, sometimes not.

What is of particular interest for the discussion in this article is the fact that when the issue of children exposed to intimate partner violence was finally put on the agenda, policy – and to some extent practice – developed quite quickly, especially compared to the previous policy and practice development in Sweden as regards men’s violence against women. The speed of the shift in policy – and to some extent in practice – is quite remarkable and warrants explanation. It can also be noted that the policy development as regards children started earlier and has been more far-reaching in Sweden compared to the other Nordic countries, in spite of the fact that research and practice in some ways have been more advanced in, most notably, Denmark and Finland (Eriksson, 2010). The development in Sweden thus constitutes a very interesting case when trying to understand spaces to create change regarding responses to violence in intimacy, as well as social change more broadly.

The discussion draws upon several different pieces of empirical research: (1) a study of ways of tackling violence from fathers against co-parents/mothers and children in Sweden in a family law context (Eriksson, 2003, 2008); (2) a study of approaches to the issue of children exposed to violence among Nordic women’s shelters/crisis centres and children’s rights organizations (Eriksson, 2010); and (3) two studies mapping support services for children exposed to intimate partner violence (Eriksson, 2006; Eriksson and Wychichowska, 2010).

The discussion starts with an outline of the theoretical points of departure. Thereafter follows, first, an outline of recent developments in policy and practice in Sweden regarding children exposed to intimate partner violence, and second, an outline of the context for policy change regarding children exposed to this form of violence. Next comes a section discussing how voluntary organizations and the discourses they draw upon may account for the change seen in Sweden in recent years. The article is finished with some concluding remarks about policy change in complex social orders.

Challenging interacting power relations

As mentioned above, children exposed to intimate partner violence constitute an ‘intersectional’ issue, shaped by inequality tied to gender as well as age and kinship. In the typical case, it is the children’s father who is violent to their mother. In societies like Sweden, father, mother, daughter and son are social positions tied to several social structures. ³ Empirically the positions of man and woman (gender), adult (age) and parent (kinship) are ‘inhabited’ simultaneously. Notions of gender give meaning to notions of age and vice versa. Different power relations construct and are constructed in relation to each other. Systems of meaning that can be separated analytically are intertwined empirically.

The idea that gendered inequalities must be analysed in relation to other social inequalities is well established within social research today. Already in the 1970s and 1980s the relationship between gender and, for example, class (e.g. Hartmann, 1979; Walby, 1986), sexuality (e.g. Rich, 1980) and race (Lorde, 1984) was explored. Later on, the concept of intersectionality was established (e.g. Crenshaw, 1991; Hill Collins, 1998; West and Fenstermaker, 1995; Young, 1997) and entered into social theory more broadly (see Walby, 2009). However, age as in child–adult relations seems to have been quite absent in these debates and developments. Barrie Thorne pointed out already in the late 1980s that on the whole, children have not been included in the discussions about intersecting power relations:

I want to add to that complexity [of theorizing intersecting patterns of domination and their effects on knowledge] by pursuing a relatively simple observation: Both feminist and traditional knowledge remain deeply and unreflectively centered around the experiences of adults. (Thorne, 1987: 86)

She also asks how we can bring children more fully into our understanding of social life, including processes of social change. Almost 20 years later, the situation seems to be quite similar when it comes to intersectionality and very young age (Krekula et al., 2005). One of the aims of this article is thus to bring child–adult relations a bit more clearly into the picture. Furthermore, since one contested question in the debates on intersectionality seems to be how the theoretical insight about interacting power relations should be put into research practice (see e.g. de los Reyes et al., 2003; Lykke, 2003; Walby, 2009), this article aims to contribute to such a discussion as well.

Doing difference, inequality regimes and the societal order of inequality

The discussion draws upon two key concepts: doing difference (West and Fenstermaker, 1995) and inequality regimes (Acker, 2006). As Candace West and Sarah Fenstermaker point out, different power relations are both experienced and produced simultaneously: the very same actions may produce several different forms of inequality (West and Fenstermaker, 1995; see also Messerschmidt, 1998). To make the framework of intersectionality useful for empirical analyses it is vital to recognize this ongoing simultaneity, they argue. However, intersectionality has often tended to be framed in a way that makes it both too static and too overarching to be useful in an actor and practice-centred analysis. This theoretical framework clearly needs to be translated into more concrete analytical tools.

One such tool used here is inequality regimes. According to Joan Acker (2006), inequality regimes are loosely interlocked practices and processes that result in systematic disparities between the organizations’ members in terms of power. They are shaped by inequalities in the surrounding society and the organizational structure as well as a number of different organizational practices. When outlining the concept, Acker focuses on class, gender and race. However, she claims that other forms of inequality may be relevant as well (Acker, 2006). ⁴ As will be shown, inequality tied to age is a key to the case under scrutiny in this article (cf. Krekula et al., 2005; Thorne, 2004). Another aspect of particular interest for the discussion below is Acker’s claim that the level of visibility and awareness within the organization regarding forms of inequality will vary. A lack of awareness may be unintentional and linked to privilege: men tend not to see their gender privilege, and white people tend not to see their race privilege. Another important issue is the relative legitimacy of inequalities. According to Acker, legitimacy is partly linked to the organizational ideology and she points to the differences between organizations with democratic goals and bureaucratic ones. She also claims that gender and race inequality tend generally to be less legitimate than class (Acker, 2006: 452). I would argue that the relative legitimacy of different forms of inequality is an empirical question. In the discussion below on how to understand the rapid change in policy regarding children exposed to intimate partner violence, I will return to the question of the relative visibility and invisibility, legitimacy and illegitimacy, respectively, of inequality tied to age and gender.

I want to add that sometimes it is necessary to make a distinction between the organizational level discussed by Acker, and more overarching societal patterns. Connell (1987) argues that gendering practices may also be analysed as a social order: at any given moment in time it is possible to ‘freeze’ the picture and explore the pattern of social relations and positions, as well as the system of meaning produced through everyday practices. According to Connell, such explorations may concern the overall societal gender order as well as the gender regimes of specific organizations (Connell, 1987: 98). Combining ideas from Connell and Acker, it is possible to make a distinction between organizational inequality regimes and the societal order of inequality.

Policy and practice developments

Moving to the empirical example for the discussion of policy change, I want to start by outlining some key developments over the last decades. Already in the 1980s there had been a few early attempts made to bring the issue of children exposed to violence to the fore in Sweden (e.g. Svensson, 1988). However, during most of the 1980s children exposed to intimate partner violence were more or less invisible in policy and practice as well as research. The lack of attention created a number of problems in legal and welfare practice. For example, the first national inspection of the practices of the local government’s social services as regards violence against women and children ‘witnessing’ violence carried out in 2008 and 2009, shows that all reviewed agencies had some problem in their responses to cases involving intimate partner violence (National Board of Health and Welfare, 2009; see also Children’s Ombudsman, 2005).

However, in recent years things have changed considerably. In the middle of November 2006 reforms to the Swedish law came into force that give children exposed to intimate partner violence a right to crime victim compensation from the state. The 2006 reform package also meant that these children were named ‘crime victims’ in the Social Services Act (SFS 2001:453), granting them special attention from the local governments’ social services/child protection agencies. Over approximately a decade the often serious consequences for children thus moved from a policy ‘non-problem’ to one of formal recognition. Children exposed to intimate partner violence are now considered crime victims in their own right, and the number of agencies supporting children who have experienced violence continues to grow.

The societal order of inequality

As an intersectional issue, policy change regarding children exposed to intimate partner violence implies not just a challenge to inequality tied to age, but gender as well. When analysing the established democratic order in Sweden and the political-cultural code on the accepted and ‘forbidden’ in terms of gender, power and democracy, Maud Eduards argues that naming men as a political category, as an interested party in a power relation with women, is ‘the most forbidden’ in the Swedish democratic order (Eduards, 2002). She claims that in the established democratic order, the essence of ‘woman’ is to be available for men and not to act independently from men. Furthermore, when women reconstruct the category of women by acting collectively, they also name men as a group and as a political category. Thereby gender is made visible as a political dimension and the Swedish hetero-normative discourse on gender equality, harmony and cooperation is challenged, according to Eduards.

To substantiate her claims Eduards draws, among other things, upon an analysis of two public campaigns in the late 1990s and early 2000s against men’s violence towards women and the responses that followed. One campaign concerned men’s violence against women in general and one focused upon children whose father abuses their mother. Both campaigns were heavily criticized in the media for being ‘men-hating’ and ‘anti- daddies’. However, the debate on the campaign on children was more conflicted and ‘it is obviously more problematic to be portrayed as a bad father, than as a bad partner’ (Eduards, 2002: 117). Eduards concludes that it seems problematic to point to men as responsible for the victimization of women and children. But the question as to why it is more problematic for a man to be portrayed as a bad father, than as a bad partner, she leaves unanswered.

What happens when men are named as a political category in relation to more than one political dimension? One interpretation of the outrage that followed in the Swedish media after the campaign on children with violent fathers is that it challenged a ‘natural attitude’ to age and kinship as well (Eriksson, 2003). It pointed out differences of interests between women and men, adults and children and parents and children – simultaneously. It named fathers, adult men-parents with minor children, as a specific something and held them accountable in terms of gender, age, kinship and power. Therefore a stronger response was to be expected.

Drawing upon Eduards’ argument and elaborating with the help of an intersectional perspective it is possible to outline a particular social order that needs to be recognized and problematized when explaining the rapid policy change in Sweden when it comes to children exposed to intimate partner violence. If Eduards is right, the recognition of violence from men who are also fathers will imply a particular challenge for those trying to achieve policy change. This challenge is linked to inequality tied to age and kinship, in addition to gender. The social position of father ⁶ is partly, but not entirely, overlapping with the position of man. It is also overlapping with the positions of adult and parent respectively. It seems that the established democratic order in Sweden is shaped by a societal order of inequality which is patriarchal in a very literal sense: the core is father-power based upon ties of blood to not yet adult children. Here, the use of the concept of blood-ties is quite deliberate. Both policy and law take heterosexual, biological parenthood as the point of departure. Swedish family policy may thus be understood as a specific kinship-policy: biological ties – ties of blood – between children and their parents both form the point of departure and are reproduced (Eriksson, 2008). ⁷ The reproduction of this order of inequality can at least partly explain a number of empirical phenomena in Sweden in the field of intimate partner violence, such as the problematic practices and lack of recognition of violence from fathers against mothers/co-parents and children among legal and social work professionals. But how does this picture of a patriarchal order of inequality ‘fit’ with the rapid change in policy regarding children exposed to violence?

Policy change and voluntary organizations

The central role played by the women’s movement in general and the women’s shelter movement in particular is well documented when it comes to the formulation of men’s violence against women as a policy problem (e.g. Dobash and Dobash, 1992; Eduards, 1997; Hagemann-White, 1998; Jonassen, 2005). During the last two decades, women’s shelters in many parts of the world have started to pay more and more attention to the situation of children. An important aspect of the developments in the Swedish context is the fact that parallel to the increased attention paid to the child’s situation by women’s shelters, some children’s rights organizations have also started to focus upon children who experience men’s violence against women. This is in common with other European countries, including the United Kingdom.

The definition of problems and problem solutions is an important part of political struggles (Bacchi, 1999). The problem definitions and strategies of women’s shelters and children’s rights organizations can be expected to differ at least to some extent since they form a part of separate social movements that place different social positions and different forms of inequality centrally (‘woman’ and inequality linked to gender, and ‘child’ and inequality linked to age respectively). However, on this particular issue, the women’s shelters and some children’s rights organizations now seem to share a common understanding (Eriksson, 2010). ⁸ Empirical support for this observation comes from a study of 10 women’s shelters and children’s rights organizations at a national level in Denmark, Finland, Norway and Sweden. ⁹ The study shows that when it comes to the ways in which the issue of children exposed to violence is defined and talked about, the organizations in Sweden come across as being shaped by a similar approach, seen through both the written material and interviews with key respondents in the respective organizations. Notably, the notion of children as ‘crime victims’ is most clearly visible in the material from Save the Children Sweden (Eriksson, 2010). Furthermore, different areas of lobbying for legal change can be seen in the material from each organization. One topic presented as at the centre of lobbying activities is children’s rights as crime victims. Again, this is most clearly emphasized in the material from Save the Children Sweden. When it comes to the women’s shelter organizations, they tend to frame the issue of children exposed as a part of men’s violence to women and the lobbying for change in policy and law talked about in the interviews, and also seen in web documents, primarily concerns family law, which of course is of key importance to abused women who are mothers. One interpretation of the development in Sweden is that it is at least partly enabled through the convergence of two social movements focusing on the ‘intersectional’ issue of children exposed to men’s violence against women. Together they lobby for policies that will simultaneously tackle the vulnerability of women (who are mothers) and children.

As regards voluntary organizations in the violence field, both women’s shelter and children’s rights organizations have been innovative as regards help and support to abused women and to children exposed to violence (Eriksson, 2006, 2010a). Most notably, it was through a project at Save the Children that Ami Arnell and Inger Ekbom (1999) developed the model for individual crisis and trauma intervention called Trappan (‘the stairs’), which today is the most widely disseminated intervention model in direct work with children exposed to intimate partner violence in Sweden (Eriksson, 2006; Eriksson and Wychichowska, 2010). It is thus possible to outline a link between, on the one hand, voluntary social work within these organizations and, on the other hand, the measures targeting children in crisis and traumatized children that in recent years have started to develop within the child health and welfare system in Sweden. This means that the impact from these voluntary organizations may be seen both at the level of legal change, following on from their lobbying activities, and at the level of everyday welfare practice, following efforts to improve help and support services.

The discursive opportunity structure

In spite of the differences when it comes to strategies to achieve change in law and policy, it can be argued that all of the Nordic organizations studied display a high degree of similarity in the outcomes they are aiming for through lobbying activities (Eriksson, 2010). Overall, the aim of their strategies seems to be to bring the issue of children exposed to violence into the public domain and to put it onto the political agenda. The collective solutions to tackle this issue that are described in the web material and interviews tend to be presented as primarily about protection and support, not about holding the perpetrator accountable and granting children rights to financial compensation (Eriksson, 2010). Although it is possible to argue that a ‘justice discourse’ (constructing children as victims with rights to compensation) to some extent shapes the organizational practices, a ‘welfare discourse’ (constructing violence as an issue for the welfare system rather than for criminal justice) thus seems to have a much stronger impact upon these organizations’ approaches to children exposed to intimate partner violence (Eriksson, 2010).

An explanation for this relative dominance of ‘welfare’ over ‘justice’ may be the ways in which discourses on children exposed to violence are linked to discourses on children more broadly. In the study of social movements, it has been argued that the cultural resonance of different framing strategies is a key to movement success (Benford and Snow, 2000: 619ff.). Frames, such as ‘children exposed to violence as crime victims’, exist in a context that is enabling in certain ways, yet restraining in others: discursive opportunities are structured, both in the sense of having pattern and form, and in the sense of being anchored in key political institutions (Sewell, 1992, in Ferree, 2003). This context thus both precedes and is influenced by the framing strategies developed by movement activists. ¹⁰ Thus, the dialogic nature of the interaction between ‘authorities’ and ‘challengers’ may be lost. Resonance does not simply reflect properties of the frame itself, but is – at least partly – created through interventions by individual and/or collective actors.

Conclusions: Policy change in a complex social order

Connell argues that gender is shaped by several structures (1987, 2002). ¹¹ Furthermore, that a society’s gender order is inherently unstable and marked by historical ‘unevenness’ (gender constructs from different historical periods exist side by side), contradictions and internal differentiation (Connell, 1987: 96). A similar argument can be made when it comes to age and kinship orders. ¹² As is illustrated in the section above, an empirically sensitive and actor-centred intersectional analysis cannot build upon a homogenization of age, gender, kinship, etc. Historical ‘unevenness’, contradictions and internal differentiation do in themselves open up spaces for challenging practices in relation to the gender order. However, as other forms of inequality are added to the picture the possibilities for change become even more apparent.

In the debates on intersectionality the focus is often upon how different forms of inequality work ‘in the same direction’. For example, in their original formulation of the doing difference perspective West and Fenstermaker (1995) primarily focus upon how multiple forms of difference may be produced simultaneously. However, it is clear that the organizational practices regarding children exposed to violence discussed in the previous section entail both dichotomization and neutralization combined in complex ways. Most notably, age dichotomization is used to neutralize gender: when challenging gender inequality, age is reproduced as difference through notions of responsible adults and subordinated, dependent children. An empirically sensitive and actor-centred intersectional analysis must be sophisticated enough to grasp complexities such as these if we are to be able to fully understand agency and policy change within a complex social order (Eriksson, 2008).

As regards the possibilities for creating policy change I want to emphasize that there seems to be a key difference between the dynamics of age and gender inequality. Formally – and also in many instances informally – men’s dominance over women is perceived as something illegitimate in contemporary Sweden. Already in the early 1980s, Hanne Haavind argued that it is because men’s power is perceived as something illegitimate that it has to be covered up: it becomes women’s task to make their own subordination to come across as a voluntary and individual choice (Haavind, 1982). However, although children’s rights have been placed upon the policy agenda it is not possible to argue that adult power over children – for example, parents’ and other adult rights to make decisions regarding minor children, or the adult monopoly on voting – is perceived as something illegitimate which should be formally abolished. On the contrary, adult dominance is often taken for granted.

What follows from power is responsibility, and I would argue that the legitimacy of adults’ power over children (age inequality) contributes to the fact that it is often easier to make visible adults’ responsibility for children’s situation, compared to men’s responsibility for women’s situation. If men ‘are not supposed to’ be dominant, how can they be responsible for women’s subordination? While it is possible to find a number of examples of media debates in Sweden about the problem of ascribing men ‘collective guilt’ (see Eduards, 2002), no similar debates about ascribing adults ‘collective guilt’ can be found in relation to the issue of adults’ violence to children. Furthermore, in consultation papers and other public documents, a recurring theme is that parents and other adults ‘should’ place children’s needs before their own, which can be interpreted as an expression of the legitimacy of adult dominance (cf. Ribbens McCarthy et al., 2000). Indeed it would be political and social ‘suicide’ to suggest that adults – and especially parents – should not be responsible for children. It thus seems to be much easier to hold adults as a collective responsible for children’s vulnerability and victimization, than to hold men as a collective responsible for women’s vulnerability and victimization. The moral order where adults place children’s needs first is linked to a social order where adults – not least parents – have power over (young) children.