Enander 2011 Leaving Jekyll and Hyde

Feminism & Psychology http://fap.sagepub.com/ Leaving Jekyll and Hyde: Emotion work in the context of intimate partner violence Viveka Enander Feminism & Psychology 2011 21: 29 originally published online 15 November 2010 DOI: 10.1177/0959353510384831 The online version of this article can be found at: http://fap.sagepub.com/content/21/1/29 Published by: http://www.sagepublications.com Additional services and information for Feminism & Psychology can be found at: Email Alerts: http://fap.sagepub.com/cgi/alerts Subscriptions: http://fap.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://fap.sagepub.com/content/21/1/29.refs.html Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 F eminism & Psychology Article Leaving Jekyll and Hyde: Emotion work in the context of intimate partner violence Feminism & Psychology 21(1) 29–48 ! The Author(s) 2011 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0959353510384831 fap.sagepub.com Viveka Enander Institutionen för socialt arbete/Department of Social Work Göteborg SE-405 30, Sweden Abstract The aim of this qualitative study was to investigate battered women’s emotion work in the context of male-to-female intimate partner violence and, more specifically, in the context of leaving violent men. A total of 22 informants were interviewed and the material consists of 47 interviews. The results suggest a process in which victims initially conceptualize abusers as good, but subjection to violence leads to a cognitive-emotive dissonance that is responded to by emotion work. Over time, conceptualizations of the abuser shift from good to bad and efforts are made to change emotions from warm to cold. Connections between this process and previously described leaving processes are discussed. Keywords battered women, domestic violence, emotion work, leaving processes Introduction Male-to-female intimate partner violence is a common social phenomenon with severe consequences for the victims (Campbell, 2002; Samelius, 2007; Smith, et al., 2002). A pattern of abuse has been described in which abusers may adopt several different controlling strategies such as isolation, verbal harassment, control of family finances, destruction of important things for the victim (such as photos and letters), threats/intimidation and physical and sexual violence (Kirkwood, 1993; Pence and Paymar, 1993; Strauchler et al., 2004). Several researchers have emphasized that abusers also display what seems to be love and care; they analyse Corresponding author: Viveka Enander, Institutionen för socialt arbete/Department of Social Work Göteborg SE-405 30, Sweden Email: [email protected] Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 30 Feminism & Psychology 21(1) this as part of the abusive pattern (Graham et al., 1994; Lundgren, 2004; Yoshihama, 2005). The reported effects on victims are confusion, altered perception and increased bonding to abusers (Barnett, 2001; Dutton and Painter, 1981; Herman, 1992). Lundgren (2004), for example, describes ‘alternation between violence and warmth’ as a key feature in abusers’ gaining and maintaining control. Similarly, Graham et al. (1994) point out that the abusive pattern includes ‘kindness’ as well as cruelty, with the following consequences for the victims: With the perception of kindness and hope the victim denies any feelings of danger, terror, and rage that the abuser creates in her or him. The denial occurs because the terror, and thus danger, is experienced as overwhelming . . . , and the rage, if expressed, invites retaliation by the abuser. Such denial allows the victim to commence bonding to the positive side of the abuser’ (1994: 38) This ‘bonding to the positive side of the abuser’ can also be interpreted in terms of emotion work. Sociologist Arlie Russel Hochschild (1979, 2003[1983]) introduced the concepts of emotion work and feeling rules to cover how emotions are ‘managed’ in different ways to align with cultural prescriptions. Trying to align one’s feelings with what one wants or thinks it is right to feel toward an intimate partner would be to follow the feeling rules of love and commitment. Emotion work would then entail focusing on the abuser’s positive sides. If one was, contrarily, trying to fall out of love, emotion work would entail focusing on the abuser’s negative sides. The aim of this study is to investigate battered women’s emotion work in the context of male-to-female intimate partner violence, especially in the context of leaving abusive men. What kind of emotion work is discernible in the informants’ accounts of living with and leaving abusive men? How is emotion work connected to leaving? Previous research Within the field of domestic violence, considerable research has been conducted on violent relationship dynamics and much effort has been devoted to explorations and explanations of ‘why battered women stay’ (for reviews, see Barnett, 2000, 2001). The issue of what finally motivates battered women to leave abusive men has also been studied, including qualitative research conceptualizing leaving as a process (Anderson and Saunders, 2003). Researchers with this approach suggest that leaving an abusive partner is not a single event but a process extending over time, involving temporary break-ups and preparatory strategies; they conceptualized it as a process of disentanglement occurring in different stages or phases (e.g. Enander and Holmberg; 2008; Holmberg and Enander, 2004; Khaw and Hardesty, 2007; Kirkwood, 1993; Landenburger, 1989; Merrit-Gray and Wuest, 1995; Moss et al., 1997; Rosen and Stith, 1997). Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 31 Enander Many leaving as a process studies have found that when women begin ‘to redefine the relationship as abusive and label themselves as victims’ (Anderson and Saunders, 2003: 175–6), it leads to a termination of the abusive relationship. Enander and Holmberg (2008), however, envisaged this redefinition and labelling of abusive experiences as a separate process predominantly taking place after the woman has left the abuser. Although the women in our study did convey an increasing awareness of negative experiences during the course of the relationship, the pattern we observed was reluctance to label these experiences in terms of violence and victimisation until after the break-up. The ‘cognitive shift’ described in several studies (Ferraro and Johnson, 1983; Goetting, 1999; Khaw and Hardesty, 2007; Landenburger, 1989; Patzel, 2001; Rosen and Stith, 1997) may thus entail reconceptualization, but not necessarily in such clear terms. When it comes to emotion, some researchers have described altered emotions, such as losing hope of change (Enander and Holmberg, 2008; Ferraro and Johnson, 1983; Holmberg and Enander, 2004; Hydén, 1994; Landenburger, 1989; NiCarthy, 1987) or mobilizing anger, (Ferraro and Johnson, 1983; Kirkwood, 1993; Landenburger, 1989) as an impetus for leaving. Enander and Holmberg (2008) described the emotions of love, fear, hate, compassion, guilt and hope as ties creating a traumatic bond binding women to their abusers. Leaving – in its widest sense – was seen as disentanglement from this bond. In this study, attention is devoted primarily to emotion and emotion work. Emotion is regarded as intrinsically connected to cognition, which will thus also be discussed. While emotion work in intimate heterosexual relationships has been the focus of several studies (e.g. Duncombe and Marsden, 1993; Minotte et al., 2007, Strazdins and Broom, 2004), emotion work in relation to violence against women, or rather, women leaving violent men, has not been extensively studied; this is the explicit focal point of this paper. Material and method This article is based on qualitative interview material with women who have left abusive men. A total of 22 informants, in two different groups, were interviewed; the material consists of 47 interviews. All interviews except one were tape-recorded and transcribed verbatim. One informant declined tape-recording; extensive notes were taken instead. Confidentiality and secure treatment of data were guaranteed to all informants. Informant Group I consists of 10 women, located via women’s shelters throughout Sweden. These informants were interviewed three times, the first two times in person and the third time by telephone. Three informants declined further participation after the first interview and one could not be located for the closing telephone interview, yielding 23 completed interviews. Informant Group II consists of 12 women recruited through public notice boards in Göteborg. Each informant was interviewed twice in person, generally Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 32 Feminism & Psychology 21(1) within a period of one to three weeks, except in three cases in which months elapsed between the first and second interviews. All informants participated on both interview occasions, yielding a total of 24 interviews. Some differences between the two informant groups deserve attention; the most important are shelter contact and length of study period. Group I was originally recruited for a study on battered women’s leaving processes (Enander and Holmberg 2008; Holmberg and Enander 2004). For this study, women with shelter contact were interviewed three times during three years, at one-year intervals, to enable an analysis of whether the informants’ conceptualizations of their reasons for leaving were constant or changed. To further investigate some questions evoked by and themes discovered in these interviews, new informants were recruited and interviewed. A longitudinal approach did not seem warranted, but only women without shelter contact were recruited to this second informant group, in order to broaden the material. Informants in Group I had thus received assistance from shelters and had presumably discussed their experiences with shelter staff and volunteers. In contrast, Group II informants had rarely shared their experiences – or discussed them – in this way; one informant stated that participating in the study entailed telling someone in detail for the first time ever about her experiences of abuse. However, while some informants in Group I had had very brief contact with the shelter, some informants in Group II had actively participated in support groups in which discourses on violence were similar to those found in shelter settings. Similarities and differences regarding earlier disclosure and discussion of violence therefore exist both between and within the informant groups. Furthermore, the two groups differ somewhat when it comes to class and urbanity/rurality. The class profile in Group I is more diverse, while informants in Group II have a narrower middle-/upper-middle-class profile: all but one had some university-level education, five had graduated and one had moved on to postgraduate studies. Furthermore, Group II is more urban because these interviews were conducted in the same large city. Another difference between the two groups is the interval between break-up and participation in the study. While all informants were interviewed post-separation, informants in Group I could have left quite recently. In the case of Group II, however, one of the participation criteria was that informants should have left the abusers at least one year prior to participation. This criterion was ethically motivated because informants without shelter contact were sought. It was considered crucial that they not be actively involved in abusive relationships or in acute crisis, in order to avoid endangering or unsettling women who might have no other supportive contacts. Other important demographic information on the 22 informants in both groups is as follows: participants ranged in age from 24 to 61 and had different levels of education, with (nine-year) compulsory school as the lowest. All were white, 18 were ethnic Swedes and four had immigrant backgrounds. All informants had been cohabiting with, and 13 had children with, the respective abusive partner. Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 33 Enander Two participants had survived two abusive relationships. Relationship duration varied from under 1 to 25 years and the time elapsed from the point of final separation ranged from quite recently to as much as 20 years ago. Interviews were semi-structured and focused, in the case of Group I, on informants’ interpretations of why they had left their abusers. This yielded rich material covering extensive experiences of both living with and leaving abuse. Subsequently, the Group II interviews focused on the informants’ interpretations of similar kinds of experiences and of themselves and their abusers. The study is both explorative and inductive and research questions have surfaced and been included throughout the whole research process. My approach was hermeneutic, both in general methodology as well as in the analysis. Interpretation, theoretically based in feminist theory on violence against women and the sociology of emotions, has thus been a key feature throughout the process. Several analytic modes were applied in an intuitive ad hoc manner, as described by Kvale (1996). The choice of emotion work as the main analytical tool resulted from the first analyses of the material and was not originally part of the research design. Although the informant groups differed as described above, emotion work was evident and displayed similar features in both sources of material, which were analysed jointly. The analysis was begun by listening to the interview tapes and reading the transcribed interviews, pen in hand, and a modified thematic analysis (Aronson, 1994) was performed in the search for recurrent themes. The themes of relevance for this study touched on abusive ex-partners being described as Jekyll and/or Hyde, on rationales for staying and leaving, and on emotions and cognitions described in connection to this. In accordance with the hermeneutical tradition of interpretation, ‘parts’ have been related to the ‘whole’ and vice versa in a reflexive process (Kvale, 1996). In practice this means that passages touching on similar themes have been compared within each interview, between interviews with the same informants, between interviews with other informants and in relation to the material as a whole. Thus I have, in line with Haavind, (2000) tried to maintain a ‘vertical’ – (within interview analysis) as well as ‘horizontal’ (between interviews analysis) mode of approaching the material. Further, the themes have been investigated in light of feminist theory on violence against women and the emotion sociological framework presented below. Again, this has been a highly reflexive process, where the findings have been compared, contrasted and challenged by the literature and vice versa. Finally, some important limitations to this study must be noted. No women actively involved in abusive relationships were interviewed (for the reasons given above). If women currently involved with the abuser had been included, it may have been possible to obtain a clearer picture of emotion work women engage in before leaving. Furthermore, the role of significant others in (re)conceptualizations and emotion work is absent; covering this topic would have expanded and deepened the analysis, but it was unfortunately beyond the scope of this study. Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 34 Feminism & Psychology 21(1) Analytical framework Sociologist Arlie Russell Hochschild’s (2003[1983]) writing on emotion work is the point of departure for this analysis. Hochschild uses the concept to cover the different ways in which we try to manage our feelings: when we try not to be sad on a happy occasion or try to show sincere grief at a funeral, when we rein in our anger, swallow our pride or put a damper on our love. Hochschild envisages two kinds of emotion work: surface acting and deep acting. In surface acting we merely play the expected part; i.e. we aim to display a certain feeling we believe is required without actually feeling it, as when we feign interest or display unfelt cordiality. Deep acting, however, requires more than feigning and occurs, according to Hochschild, in two ways: ‘by directly exhorting feeling’ and ‘by making indirect use of a trained imagination’ (2003[1983]: 38). In the first case, we ‘address’ our feelings directly; we ‘order’ ourselves to feel less, more, at all or differently. In the second case, we use our imagination to achieve this; by picturing to ourselves how desperate a friend in need must be, for example, we may overcome our ambivalence to devoting the time and energy to give help and care. Or – relevant to this study – by focusing only on the bad memories of a person we want not to love, we may eventually manage to fall out of love, a change that is difficult to achieve by only giving ourselves ‘orders’ not to love. In deep acting we aim not only at the emotional display we consider to be ‘right’, we try to align our feelings with this ‘right’. Thus, the purpose of deep acting is to make feigning unnecessary; as a result of deep acting, feigned interest may become real and unfelt cordiality become heartfelt. What, then, is this ‘right’ way of feeling and how do we know what is ‘right’? In response, Hochschild turns from the psychological to the sociological realm, by introducing the concept of feeling rules. Drawing on Goffman (1956, 1967), who pointed out that emotions – or, more specifically, embarrassment – was not a disturbing outcome but rather an important part of social interaction, Hochschild pictures social systems as entailing feeling rules that govern who owes what feelings to whom. Being a ‘subterranean’ part of social systems, feeling rules are most easily discernible when broken; we glimpse a feeling rule in statements such as ‘at least you could have shown some gratitude’ or ‘you might have tried to be happy for my sake’. Feeling rules are shared within social groups; we can aim at breaking or surpassing them, but they are generally ‘known’ and internally acknowledged. Dissonance is a central concept to this study. Social psychologist Leon Festinger (1957) formulated the classical theory of cognitive dissonance. According to Festinger, people holding contradictory cognitions will experience a strain that they will wish to reduce, generally by changing one of them, thereby achieving consistency. Hochschild, concurring with Festinger, proposes that this also applies to emotion: A principle of emotive dissonance, analogous to the principle of cognitive dissonance, is at work. Maintaining a difference between feeling and feigning over the long run Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 35 Enander leads to strain. We try to reduce this strain by pulling the two closer together either by changing what we feel or changing what we feign. (Hochschild, 2003[1983]: 90) I propose that dissonance is active in processes of staying and processes of leaving, which will be illustrated in the analysis. However, I will not distinguish between cognitive and emotive aspects of this phenomenon since I view them as interwoven and separable only at the theoretical level. Hochschild makes a distinction between emotion work and emotional labour: the former relates to the private and the latter to the public sphere, where emotions may be managed to sell a product, or the managed emotions may be the product itself. Thus, the art of managing emotion may be public or private and more or less situational. But, as Hochschild underlines: (Yet) the achievements of the heart are all the more remarkable within roles that last longer and go deeper. Parents and children, husbands and wives, lovers and best friends expect to have more freedom from feeling rules and less need for emotion work; in reality however, the subterranean work of placing an acceptable inner face on ambivalence is actually more crucial for them. In fact, the deeper the bond, the more emotion work, and the more unconscious we are of it. (Hochschild, 2003[1983]: 68) While Hochschild’s focus is primarily on emotional labour, my focus is solely on emotion work. Private emotion work within the bonds of intimate relationships is, however, a difficult object of study, mainly due to two circumstances. First, as Hochschild points out, this work is not consciously noticed and may therefore be hard to discern (for both informant and researcher). Second, that retrospective interviewing itself may be part of emotion work and contain its own feeling rules. The feelings the informants describe may be interwoven with the feelings they think they are expected to describe. This poses a delicate dilemma for the researcher and requires a critical eye in the reader (cf. Frith and Kitzinger, 1998). Results Robert Louis Stevenson’s novel Strange Case of Dr. Jekyll and Mr. Hyde was published in 1886 and was a major success (Stevenson, 2003[1886]). The story of the well-reputed Dr. Jekyll, occasionally transformed into the vile and violent Mr. Hyde, has since been reprinted numerous times and portrayed in several different ways on the stage and in films. ‘Jekyll and Hyde’ has become a metaphor to describe a duality of, or a transformation from, good to bad. It has been noted that this metaphor is commonly used by abused women when describing the abusers (e.g. Enander, 2010; Goetting, 1999; Zink et al., 2006). In the following, I will illustrate a process in which women’s conceptualizations of their abusers shifts from good to bad: or from Jekyll to Hyde. First I will present a more overarching process, illustrated by two empirical cases, then I will exemplify parts of this process more in detail. Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 36 Feminism & Psychology 21(1) From exclusively Jekyll to Hyde as well: Shifting conceptualizations and emotion work Kind, warm, considerate, charming, humorous, empathic and exciting were words commonly used by the informants to describe the man they first met. The description of having been ‘seen’ by this man, in the sense of being perceived and appreciated just as one ‘is’, was also recurrent. It is a good man the informants describe falling in love with, and this initially encountered good person is perceived not just as the original, but also as the genuine man for a long time (c.f. Jeffner, 1998). In accordance with the Jekyll/Hyde metaphor, informants talked about falling in love with Jekyll, not Hyde, and that the duality of the abuser only became apparent later. As Malin put it: I mean, it’s not like you meet a violent man and say ‘Yes, now everything’ll be wonderful!’. You meet a wonderful man who turns out to be something entirely different later. (Malin (II),1 second interview) However, after inflicting painful experiences of psychological, physical and/or sexual abuse, the perpetrator was perceived as ‘‘two people’’– or Jekyll and Hyde – in recurrent descriptions (c.f. Enander, 2010; Goetting, 1999; Zink et al., 2006). An example from Erika (I): Cause he’s like two people, this man, the most proper, kind, sweet, and oh, so charming fellow. He’s Dr Jekyll and Mr Hyde . . . Amazing, completely different people, you can’t even imagine how they can wear the same clothes. (first interview) Informants described confusion, inability to reconcile the differences and selfblame as a consequence of the introduction of Hyde alongside the good Jekyll: You end up feeling you’re crazy, like. Because he’s so kind and sweet, on the one hand, and then all of a sudden he can change into someone completely different. At last you almost go bonkers yourself, or something like that, don’t you? Because you wonder: ‘How can such a nice, kind bloke turn into someone like that? Well, perhaps it’s because of me then.’ (Malin (II), second interview) In the extract from Malin above, she expresses feelings of incomprehension on the basis of the assumption that her abuser was, essentially, ‘a nice, kind bloke’. Like Malin, informants perceived the ‘good part’ as more genuine than the bad, which is no wonder since it was this good man they first encountered and got to know. However, when violence and degradation continued, the conceptualization of the man as basically caring and good ‘didn’t add up’, as phrased by Elisabeth, with his behaviour: It didn’t add up. How he . . . If he now says he loves me so much and at the same time does like this . . . (I, first interview) Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 37 Enander When bad turns to worse, ‘Hyde’ eventually takes precedence over ‘Jekyll’; at some point, which may vary greatly, women’s conceptualizations of their abusers shift from good to bad. I am not suggesting that this is a clear and linear process. Women make distinctions of different kinds, such as between the person and the act, in their conceptualizations of their abusive partners. However, in the interviews I have analysed, women’s conceptualizations of their abusers did tend to shift ‘‘from Jekyll to Hyde’’. The time-point at which such a shift occurs is probably related to a multitude of external and internal factors. Interestingly, these shifting conceptualizations were described by some informants as connected to leaving: I didn’t choose him when he was bad, did I? No, I felt that I fell in love with him for his good sides. And in the beginning, the good sides outweighed the bad all the time but then the bad sides took over more and more and there was less good left . . . and then it was easier to get out, you know. (Jenny (II), first interview) Active emotion work is discernible in these shifting conceptualizations. Initially, when the abuser is still perceived as mainly good, informants seem to tackle being hurt, scared or disappointed by focusing on his good aspects, as exemplified by his caring, considerate and repentant behaviours; that is, they try to align their feelings with the conceptualization of being in a loving relationship with a basically good partner. More degradation and violence bring increasing dissonance that makes this conceptualization untenable. When the abuser is finally conceptualized as more bad than good, opposing emotion work is visible in informants’ descriptions. Thus, with shifting conceptualizations, from good to bad, efforts are made to turn emotions from warm to cold. The cases of Liljana (I) and Lena (II) are presented as an illustration of my analysis. Liljana. Liljana, a mother of two who described herself as having fought hard for her family, recounted that she experienced the first three years of the relationship with her ex-husband as happy. But after that, said Liljana, he ‘changed a lot’: he started to hang out with friends, stayed away from home longer periods of time and battered her. Liljana described being shattered by this and she reported the abuse to the police a number of times. Yet she did not, apparently, initially want to completely lose trust in her husband. She described him as convincing and manipulative, and said that she also wanted to believe him, for example, when he stated that he had renounced violence for ever. Thus, the first time police reports resulted in a trial Liljana, not wanting to harm her husband, took back her accusation of abuse, believing that he would never abuse her again. But she also recounted: Do you know what occurred to me at that trial? I’ll be honest. The thought hit me, just at the end. It had been going on for three or four hours and it hit me, ‘What if he does it again, and here you are, lying?’ That just occurred to me . . . Then there was the Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 38 Feminism & Psychology 21(1) second part, ‘No, he won’t. We’ll go get some help. There’s family counselling and that sort of thing’. I was wrong [whispers]. (first interview) As I see it, Liljana still predominantly conceptualized her ex-husband as ‘Jekyll’, i.e. good and trustworthy, at the time of this trial. However, his hurtful behaviour seems to have created a dissonance that suddenly resonated at the end of the trial. However, the dissonance, as I interpret it, was not strong enough at this point to lead to a reconceptualization of the abuser yielding the interpretation that he would, most probably, be violent again. Instead, Liljana sought to resolve the dissonance by enforcing her hope that her husband would indeed stop his violence, with help if necessary. Retrospectively – and in a whisper – Liljana acknowledged that she had been wrong. She discussed, too, why she wanted to believe in her exhusband’s recurrent assurances that he would not hit her again: I can’t completely explain why you want to believe him. First, you have to be 100 percent brainwashed into believing you can’t manage living without him. You can’t do anything without him. You can’t go to the loo without him. You can’t live without him. You have no future without him. You’re meant to be scared when he just shows his face, and not talk about when he hits you. You’re to have that fear 99 percent, 24 hours a day. I call that brainwashing. And you still think up things like that. You feel sorry, not for yourself, you feel sorry for him the whole time. How’ll he manage if I leave him? How will he do the laundry, how will he cook, how will he . . . ? And the third part is about that you’ve been happy with this man at some point in your life. You had a great relationship. I had a great relationship that most people just dream about. So it’s difficult for me to understand that it’s that man. And that’s why it went so far. That’s why it took time until I realized that he’d changed 100 percent. I kept trying to find the man I’d met in 1994. But I couldn’t. I was fooling myself. (first interview) In the quote above, Liljana brings up several effects of abuse described in earlier research; e.g. being ‘brainwashed’ (e.g. Mega et al., 2000) or subjected to ‘distortion of subjective reality’ (Kirkwood, 1993), internalizing violence (Lempert, 1996, Lundgren, 2004) and developing emotional ties of fear, compassion and hope (Enander and Holmberg, 2008; Holmberg and Enander, 2004). But emotion work is also discernible. ‘I kept trying to find this man I met in 1994’, said Liljana, which can be interpreted as her trying to counter the dissonance created by abuse with active emotion work; the man she met in 1994 was perceived as good and loving and worth living with. I would call trying to ‘find’ him ‘emotion work towards Jekyll’, that is, towards conceptualizations of the abuser as basically good. More specifically, I would argue that Liljana describes an attempt at what Hochschild calls deep acting: when faced by violence she does not try to feign positive feelings for the man to counter dissonance, as in surface acting, but to evoke them through finding their source. Cognitive sociologist Morris Rosenberg Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 39 Enander (2006[1991]) describes how such emotional management is actually achieved with cognitive means. It is through different modes of selective attention, perspectival selectivity and selective interpretation that we manage our emotions, claims Rosenberg. He thus points to how very complex the emotion–cognition relationship is. Liljana’s emotion work toward Jekyll seems to have been successful up to a point; Liljana stated that ‘that last bit in my heart, a little bit of love left there’ was one reason she kept hoping for change. However, she was eventually subjected to very serious violence by her husband, who took her to a desolate place where he nearly killed her and probably – though not expressed explicitly in the interview – also raped her. I presume that this horrific event led to her finally changing conceptualizations – and losing hope. In earlier research, losing hope has been found to be an important part of leaving (Enander and Holmberg, 2008; Ferraro and Johnson, 1983; Holmberg and Enander, 2004; Hydén, 1994; Landenburger, 1989; NiCarthy, 1987), and it was indeed after this event that Liljana left. Lena. Lena, who had married a man she thought she would live with for the rest of her life, shifted conceptualizations of her husband almost immediately after violence entered the relationship; at least that was how she described it retrospectively. Lena recounted that the first time she was physically abused – late in pregnancy – she was utterly shocked, and understood that she had ‘drawn the short straw’ in her choice of husband, but saw no way out of the relationship due to the expected child. The second time her husband beat her she urged him to seek professional help to change his behaviour and decided – privately – that she would leave if it happened a third time. And the third time she did indeed leave. This is her description of what happened to her emotionally with the onset of violence: And I guess that’s what I felt when he had beaten me the first time, then you know . . . even though . . . well, it’s hard to say but the love, I mean my . . . love is based on trust and my trust disappeared, you know. And so did lo . . . true love. The other thing was, you know, more of . . . well . . . a formality to do with our son having a mum and a dad . . . But when something breaks, it breaks. It’s not as easy as it just . . . well . . . being there just like before. It was completely . . . for me it was a completely different world from one day to the next I was living in different worlds. (second interview) Emotionally, something ‘broke’, said Lena, which I interpret as also leading to shifting conceptualizations of her whole living situation to ‘a completely different world’ (c.f. Mellberg, 2002). When Lena was asked about her situation between these three abusive events – that is, from the first to the third and leaving – she replied: Well . . . well, now that I look back on it it was just like trying to keep up appearances and . . . it was very important that no one should find out, you know. So it . . . I think it was [inaudible], that everything should look all right, you’re to . . . try to look happy Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 40 Feminism & Psychology 21(1) and . . . try to pretend that everything was all right. And you felt horrible . . . horrible inside. (first interview) Keeping up appearances while feeling horrible inside is an everyday way of describing what Hochschild calls surface acting. Lena describes trying to look happy rather than trying to be happy, which would entail deep acting. In a totally changed world, Lena appears to have managed surface acting, but not deep acting actually affecting her emotions. For her, it seems, the dissonance was too large. We will continue with Lena, because she unfortunately experienced – and left – two abusive relationships. However, they differed, in her description, leading to different kinds of conceptualizations of the abuser and to different emotional effects, as I interpret it. While Lena’s husband was physically and sexually abusive in a way she described as ‘explosive’, the partner she met approximately ten years later was not physically or sexually violent, but psychologically abusive. Lena described being controlled and threatened to death by this man, and stalked after leaving him. Comparing this to the violence she experienced from her exhusband she concluded that: I guess his way of terrorizing me was more intelligent because it went much deeper. It scared me much, much more than physical violence or sexual violence did. Because that had a beginning and an end, didn’t it, but this didn’t. No, this terrorizing was there all the time, the fear was there all the time, wasn’t it? (second interview) Apart from creating a more pervasive fear, the psychological abuse Lena suffered from her ex-partner may have differed from the physical and sexual abuse she had suffered from her ex-husband in another way. Although the psychological abuse was ‘there all the time’, it may have been less tangible than physically and sexually violent acts that had ‘a beginning and an end’, thus also less likely to noticeably collide with former conceptualizations and lead to undeniable dissonance. Actually, while Lena managed an almost immediate breakup from her husband, this was not the case with the psychologically abusive partner. She described leaving the latter because of ‘pure exhaustion’ (second interview), when she was worn out and down and seems to have, as described in earlier research, ‘hit rock bottom’ (see Enander and Holmberg, 2008; Holmberg and Enander, 2004; Ferraro and Johnson, 1983; NiCarthy, 1987). Comparing Lena’s accounts of the two violent relationships she was involved in, I would argue that the blatant physical violence that Lena was subjected to by her husband, and which she describes as very surprising and shocking, was strongly and clearly dissonant with perceptions of the man as caring and good. The more gradually increasing psychological control and intimidation she was subjected to by the other man may, however, not have protruded in equally stark and obvious contrast to positive perceptions of the abuser. In the first case Lena left as soon as she had the means to. In the second case it was a long and painful process, as Lena describes it, to extricate herself from the relationship. Although there may of course Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 41 Enander be other factors involved here, I suggest that the different kinds of violence Lena was subjected to mattered. In the first case the shocking experience of sudden physical violence would have led to an immediate dissonance, and Lena left almost immediately. In the second case the violence was less sudden and not physical; the dissonance would have been less immediate, as was Lena’s decision to leave. This leads to the interpretation that dissonance is significant to leaving. Dissonance exemplified I have proposed that the onset of violence leads to a dissonance that increases with more violence and that may lead to changed conceptualizations of the abuser/ relationships. Some specific examples that support this are called for at this point. Below, Cecilia (II) describes primarily emotive (aspects of) dissonance that – as she expresses it – ‘did not reach her brain’: when we, like, were getting along, it was fantastic, wasn’t it, like I told you. Everything was so incredibly good. But I always had a knot in my stomach because I didn’t know when or where I would do something wrong so that things would go bad. It was like walking in a minefield all the time. I never knew when he would explode. When I got home, when I’d been in school or whatever, I got to the bus stop and would think ‘Oh, how lovely, the sun’s shining and I feel marvellous and the weather is brilliant’ and as I got closer I got that knot in my stomach. And when I finally was outside our house, ‘I wonder what mood he’s in today’ . . . then it was like . . . No. My whole body felt like I wasn’t happy. And it was just that we didn’t reach . . . it didn’t reach my brain fast enough. (first interview) Cecilia was obviously experiencing discomfort: as she expressed it, her body felt that she was not happy. Thus the dissonance was there, but not fully acknowledged in the sense that she let it affect how she perceived the relationship; it did not ‘reach her brain’. According to my interpretation and analysis, Cecilia letting the dissonance reach her brain would imply that she would change her conceptualizations of the abuser and/or the relationship. The two quotes from Rebecka (I) chosen to illustrate dissonance end with the same declaration: ‘it didn’t work’: Every time he was that nice I could talk to him, he understood, he really got it. But when he was aggressive, you couldn’t say a word. Even if you just said ‘Hello’, you could practically be thrown out of the house. So it was very strange for me since I got to the point where I didn’t know where I stood or what I felt, or, I didn’t know anything. Half of me knew what it wanted and the other half said absolutely not. I mean, 50-50 and then I have to fight with myself. It didn’t work. (first interview) No, part of me wanted to stay because I was trying to see that good part of him that he had all the same. And trying to get rid of that other part. But it didn’t work. Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 42 Feminism & Psychology 21(1) Cause I was suffering from it the whole time. And I tried to serve dinner when he got home, you know, to do everything as perfectly as possible. And to try to stop these rows and all. But it didn’t work. (first interview) As indicated by the first quote, it was ‘fighting with herself’ that did not work for Rebecka; i.e. having a dissonant cognitive and emotive comprehension of the situation, according to my interpretation. Previous research has described internalization and adopting the abuser’s worldview as effects of abuse (e.g. Graham et al., 1994; Lempert, 1996; Lundgren, 2004). Rebecka was one of the informants who most vividly described what could be understood as internalization; she talked about becoming ‘strapped into the other person who is a person instead of you’ (first interview). However, as noted elsewhere (Holmberg and Enander, 2004), this internalization is not complete but leads instead to a split – or dissonant – sense of reality. In the second quote from Rebecka, what did not work was the attempt to see the good part of the abuser and get rid of the other. This can be read as an attempt at reducing dissonance by deep acting. Hochschild describes two ways of managing emotion through deep acting: by ‘directly exhorting feeling’ and by ‘making indirect use of a trained imagination’. Rebecka had previously described the latter kind of deep acting: picturing to herself how in need of love her depressed partner must be and how he was suffering; here, she also touches on the former kind. By trying to get rid of the bad part of the abuser, she also attempted to get rid of her negative feelings. However, this did not work; the dissonance was not resolved, probably because of something else that did not work according to Rebecka; putting an end to ‘the rows’ or, in other words, to her partner’s violence. What eventually did not work, then, was staying in the relationship. Emotion work toward Hyde What remains to be illustrated, in the process suggested above, is what was described as ‘opposing’ emotion work, meaning emotion work toward Hyde and away from Jekyll following shifting conceptualizations of the abuser. This emotion work emerged most visibly in accounts of post-leaving experiences. Below is an example from an interview with Eva (II): I must say that I don’t think I’ve ever had such a good relationship as when he finally moved out and started to sort of come round and date me a bit again and we were almost back in the old days and I felt all the warning signs there are, didn’t I, I mean my God, it’s . . . It’s so easy to slide . . .’ cause he was so nice it was unbelievable, you know. And he was going to stop drinking and ‘this is OK with me’, you know, to prove to me somehow that ‘we can start up from the beginning again’ and . . . it was a bit scary, actually. I thought, I’ve really got to talk to myself, so I don’t start sliding into that. (first interview) Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 43 Enander When Eva’s ex-husband, shortly after the separation, displayed kindness and an interest in reuniting, she felt a need to ‘really talk to herself’ not to ‘slide’ into the relationship again. Deep acting by directly exhorting feeling seems to be a method Eva employed to maintain the separation and not re-enter the relationship. Agnes (II) described not using mental energy to think about the abuser’s good sides, because she found it pointless. This was also connected to her not wanting contact with him: And then in hindsight it’s not productive for me to add the nice side to my image of that bastard. Because it’s . . . it’s meaningless to me’. Cause he isn’t a person I want to spend time with. So I’ve chosen to focus on one part of his personality. (second interview) Retrospectively, Agnes saw no need to focus on positive aspects of the abuser. And by deep acting she seems to have managed to avoid this; this interpretation is based on a comment she made concerning the first interview, when interviewed the second time: I think that the hardest part . . . of the last interview was . . . having to think about how I felt about him from the very beginning. Because . . . it’s difficult somehow. I know what he did to me, you know, but having to think that he was a person I had started seeing, who seemed to be a certain way at first. It was quite difficult to try to think . . . about him that way and describe him that way. By bringing up positive memories and earlier conceptualizations of the abuser, Agnes was also reminded of feelings that she had (emotion) worked her way out of. In this case it is difficult to discern by which kind of deep acting she had achieved this, exhortation or indirect imagination; perhaps both. In summary, I have presented a process in which women’s conceptualizations of their abusers shift from good to bad, and emotion work in connection to these reconceptualizations has been illustrated. The cognitive/emotive dissonance created in their victims by abusive men’s use of violence has also been illustrated, and described as playing a crucial role in this process. Discussion Two questions were asked in the introduction: What kind of emotion work is discernible in the informant’s accounts of living with and leaving abusive men? and How is emotion work connected to leaving? Starting with the first question, two kinds of emotion work have been illustrated: emotion work toward ‘Jekyll’ and emotion work toward ‘Hyde’. In the first case, informants try to align their feelings with conceptualizations of their abusers as basically good and worth living with; in the second case the tables are turned and informants struggle to not feel ‘good’ Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 44 Feminism & Psychology 21(1) feelings for partners/ex-partners that they have decided are basically bad and not worth living with. From this we can conclude that emotion work can support processes of staying as well as processes of leaving. How, then, are emotion work, dissonance and shifting conceptualizations connected to leaving? As described in the literature review, a ‘cognitive shift’ is described in several leaving process studies (Anderson and Saunders, 2003; Ferraro and Johnson, 1983; Goetting, 1999; Khaw and Hardesty, 2007; Landenburger, 1989; Patzel, 2001; Rosen and Stith, 1997). This cognitive shift was gradual, but brought about by catalysts such as increased level of violence, fewer good periods, or negative impact of the violence on the children (Anderson and Saunders, 2003). This study suggests that such catalysts will be in contrast to positive conceptualizations of the abuser and create emotive–cognitive dissonance. The cognitive shift described in previous research is further, I would argue, equal to a shift in conceptualizations of the abuser from Jekyll to Hyde. When will this shift then occur? On the basis of this study the logical answer is; when dissonance is too large to be able to counter with emotion work toward Jekyll. A turning point in the process of leaving is repeatedly mentioned in the literature (Enander and Holmberg, 2008; Haj-Yahia and Eldar-Avidan, 2001; Kearney, 2001; Landenburger, 1989; Patzel, 2001; Rosen and Stith, 1997), and is at times equated with the aforementioned cognitive shift. Perhaps it is at the turning point that Hyde simply becomes undeniable. However, I would argue, this is not equal to – as suggested in previous literature – redefining the whole relationship as abusive and labelling oneself victimized. Conceiving that one’s relationship or male partner is ‘bad’ is not the same thing as spelling out the concepts of violence, perpetrator and victim. From my previous research (Enander and Holmberg, 2008) I suggest that a more profound redefinition of the relationship is a separate process; mainly occurring after the turning point/break up. Several researchers have described leaving as a process occurring in different stages or phases (Burke et al., 2001; Cluss, et al., 2006; Enander and Holmberg, 2008; Khaw and Hardesty, 2007; Landenburger, 1989; Merrit-Gray and Wuest, 1995; Moss et al., 1997; Pilkington, 2000; Wuest and Merrit-Gray, 1999). However, the beginning and endpoint of this process and the number of defined stages vary considerably. Regardless of the number of stages included in these different models, it is, however, more probable that emotion work toward Jekyll, which upholds staying rather than leaving, will be dominant in earlier stages of the leaving process, while the emotion work towards Hyde will be more evident in latter stages, supporting or upholding a decision to leave. This study fills a gap in the literature, because an explicit focus on battered women’s emotion work in relation to leaving processes has been lacking in previous research. More research along these lines is warranted. Emotion work is further centred on feeling rules that remain to be described. What feeling rules may be involved in processes of staying and processes of leaving? When may these emotion rules support women’s right to physical and psychological integrity and their right Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 45 Enander to live without violence, and when may they serve as an obstacle? These are important questions for further research. Moreover, the issue of which discourses nourish such feeling rules is a relevant subject for further analysis. Several authors have pointed to constructions of masculinity and femininity and notions of romantic love that may serve to maintain women in abusive relationships (Bochorowitz and Eisikovits, 2002; Boonzaier and de La Rey, 2003; Jackson, 2001; Towns and Adams, 2000). Classical romantic narratives urge the beauty to love the beast and kiss the frog, because these apparitions only hide the wonderful prince or good man within. It is in line with these narratives that the good Jekyll may be interpreted as more genuine than the bad Hyde, and emotion work towards Jekyll, and staying, will seem reasonable. Connecting discourses and feeling rules is a future analytical task, and creating counter discourses to notions of women’s love as the remedy to men’s violence is an urgent political task. Another task for research, policy and practice is to analyse and challenge practices in which victimized women are urged to undertake emotion work towards Jekyll rather than towards Hyde. In the Nordic countries, shared parenthood with an ex-partner, violent or not, has become the norm and several researchers have highlighted the dangers this may impose on abused women and their children (e.g. Eriksson et al., 2005). Sociologist Maria Eriksson (2003) has described how a gap is construed between ‘fathers’ and ‘violent men’ in Swedish policy and practice concerning custody and contact. Thus, a violent man/father can still be regarded as a good parent with whom the woman/mother should cooperate. Policy and practice concerning custody and contact may thus contribute to women feeling impelled to perceive the man as Jekyll rather than Hyde. Possibly this can prolong women’s disentanglement from abusive relationships and needs to be challenged. Finally, a major contribution of this study is its emphasis on how emotion work goes hand in hand with women’s shifting conceptualizations of their abusers, shifts brought about by additional experiences of abuse. I have also claimed that victimized women use emotion work to counter dissonance, by either strengthening or by weakening positive feelings for the abusers. These suggestions need to be confirmed by further research. But lay and professional support providers may well ponder what an awareness of emotion work as an active ‘dissonance manager’ can bring to practical work with battered women. How can we empower women so they may ‘listen’ to dissonance that speaks of violence without resorting to emotion work that trivializes it? Note 1. Roman numeral I or II in parenthesis indicates informant group. References Anderson D and Saunders D (2003) Leaving an abusive partner: An empirical review of predictors, the process of leaving, and psychological well-being. Trauma, Violence & Abuse 4: 163–91. Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 46 Feminism & Psychology 21(1) Aronson J (1994) A pragmatic view of thematic analysis. The Qualitative Report, 2. Available at: www.nova.edu/ssss/QR/BackIssues/QR2–1/aronson.html. Barnett OW (2000) Why battered women do not leave, Part 1: External inhibiting factors within society. Trauma, Violence & Abuse 1: 343–72. Barnett OW (2001) Why battered women do not leave, Part 2: External inhibiting factors – social support and internal inhibiting factors. Trauma, Violence & Abuse 2: 3–35. Bochorowitz DY and Eisikovits Z (2002) To love violently: Strategies for reconciling love and violence. Violence Against Women 8: 476–94. Boonzaier F and de La Rey C (2003) ‘He’s a man, and I’m a woman’: Cultural constructions of masculinity and femininity in South African women’s narratives of violence. Violence Against Women 9: 1003–29. Burke JG, Gielen AC, McDonnell KA, ÓCampo P and Maman S (2001) The process of ending abuse in intimate relationships: A qualitative exploration of the transtheoretical model. Violence Against Women 7: 1144–63. Campbell JC (2002) Health consequences of intimate partner violence. Lancet 359: 1331–6. Cluss PA, Chang JC, Hawker L, Scholle SH, Dado D, Buranovsky R and Goldstrohm S (2006) The process of change for victims of intimate partner violence: Support for a psychosocial readiness model. Womens Health Issues 16: 262–74. Duncombe J and Marsden M (1993) Love and intimacy: The gender division of emotion and ‘emotion work’: A neglected aspect of sociological discussion of heterosexual relationships. Sociology 27: 221–41. Dutton D and Painter SL (1981) Traumatic bonding: The development of emotional attachments in battered women and other relationships of intermittent abuse. Victimology 6: 139–55. Enander V (2010) Jekyll and Hyde or ‘Who is this guy?’ – Battered women’s interpretations of their abusers as a mirror of opposite discourses. Women’s Studies International Forum 33: 81–90. Enander V and Holmberg C (2008) Why does she leave? The leaving process(es) of battered women. Health Care for Women International 29: 200–26. Eriksson M (2003) I skuggan av Pappa: Familjerätten och hanteringen av fäder våld. (In the shadow of daddy: The family law and the handling of fathers’ violence). Stehag: Gondolin. Eriksson M, Hester M, Keskinen S and Pringle K (eds) (2005) Tackling Men’s Violence in Families: Nordic Issues and Dilemmas. Bristol: Polity Press. Festinger L (1957) A Theory of Cognitive Dissonance. Evanston, IL: Row, Peterson. Ferraro KJ and Johnson JM (1983) How women experience battering: The process of victimization. Social Problems 30: 325–39. Frith H and Kitzinger C (1998) ‘Emotion work’ as participant resource: A feminist analysis of young women’s talk in interaction. Sociology 32: 299–320. Goetting A (1999) Getting Out: Life Stories of Women Who Left Abusive Men. New York: Columbia University Press. Goffman E (1956) Embarrassment and social organization. American Journal of Sociology 62: 264–71. Goffman E (1967) Interaction Ritual: Essays on Face-to-Face Behavior. New York: Pantheon Books. Graham DL, Rawling EI and Rigsby RK (1994) Loving to Survive: Sexual Terror, Men’s Violence and Women’s Lives. New York: New York University Press. Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 47 Enander Haavind H (2000) På jakt efter kjønnede betydninger (Looking for gendered meanings). In: Haavind H (ed.) Kön och tolkning. (Gender and Interpretation). Stockholm: Natur och Kultur. Haj-Yahia MM and Eldar-Avidan D (2001) Formerly battered women: A qualitative study of their experiences in making a decision to divorce and carrying it out. Journal of Divorce and Remarriage 36: 37–65. Herman JL (1992) Trauma and Recovery: From Domestic Abuse to Political Terror. New York: Basic Books. Hochschild AR (1979) Emotion work, feeling rules and social structure. American Journal of Sociology 85: 551–75. Hochschild AR (2003[1983]) The Managed Heart. Berkeley, CA: University of California Press. Holmberg C and Enander V (2004) Varför går hon? om misshandlade kvinnors uppbrottsprocesser. (Why does she leave? On the leaving processes of battered women). Ystad: Kabusa böcker. Hydén M (1994) Woman Battering as Marital Act: The Construction of a Violent Marriage. (PhD thesis). Oslo: Scandinavian University Press. Jackson S (2001) Happily never after: Young women’s stories of abuse in heterosexual love relationships. Feminsm & Psychology 11: 305–21. Jeffner S (1998) ‘Liksom våldtäkt, typ . . . ’: om ungdomars förståelse av våldtäkt (‘Like, you know – rape . . .’ : On young people’s perception of rape) (PhD thesis). Stockholm: ROKS. Kearney MH (2001) Enduring love: A grounded formal theory of women’s experiences of domestic violence. Research in Nursing & Health 24: 270–82. Khaw L and Hardesty JL (2007) Theorizing the process of leaving: Turning points and trajectories in the stages of change. Family Relations 56: 413–25. Kirkwood C (1993) Leaving Abusive Partners: From the Scars of Survival to the Wisdom for Change. London: Sage. Kvale S (1996) Interviews: An Introduction to Qualitative Research Interviewing. Thousand Oaks, CA: Sage. Landenburger K (1989) A process of entrapment in and recovery from an abusive relationship. Issues in Mental Health Nursing 10: 209–27. Lempert L (1996) Language obstacles in the narratives of abused women. Mid-American Review of Sociology 19: 15–32. Lundgren E (2004) The Process of Normalising Violence. Stockholm: National Organization of Battered Women’s Shelters in Sweden. Mega LT, Mega BT and Harris BM (2000) Brainwashing and battering fatigue: Psychological abuse in domestic violence. North Carolina Medical Journal 61: 260–5. Mellberg N (2002) När det overkliga blir verklighet: Mödrars situation när deras barn utsätts för sexuella övergrepp av fäder. (When the unreal becomes a reality: Mothers’ situation when their children are sexually abused by fathers) (PhD thesis). Umeå: Boréa Bokförlag. Merrit-Gray M and Wuest J (1995) Counteracting abuse and breaking free: The process of leaving revealed through women’s voices. Health Care for Women International 16: 399–412. Minotte KL, Stevens DP, Minotte MC and Kiger G (2007) Emotion-work performance among dual-earner couples: Testing four theoretical perspectives. Journal of Family Issues 28: 773–93. Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011 48 Feminism & Psychology 21(1) Moss VA, Pitula CR, Campbell JC and Halstead L (1997) The experience of terminating an abusive relationship from an Anglo and African American perspective: A qualitative descriptive study. Issues in Mental Health Nursing 18: 433–54. NiCarthy G (1987) The Ones Who Got Away: Women Who Left Abusive Partners. Seattle, WA: Seal. Patzel B (2001) Women’s use of resources in leaving abusive relationships: A naturalistic inquiry. Issues in Mental Health Nursing 22: 729–47. Pence E and Paymar M (1993) Education Groups for Men Who Batter: The Duluth Model. New York: Springer. Pilkington FB (2000) Persisting while wanting to change: Women’s lived experiences. Health Care For Women International 21: 501–16. Rosen KH and Stith SM (1997) Surviving abusive dating relationships: Processes of leaving, healing and moving on. In: Kantor GK and Jasinski JL (eds) Out of the Darkness: Contemporary Perspectives on Family Violence. Thousand Oaks, CA: Sage, 170–82. Rosenberg M (2006[1991]) Self-processes and Emotional Experiences. In: Howard JA and Callero PL (eds) The Self-society Dynamic: Cognition, Emotion & Action. Cambridge: Cambridge University Press, 123–42. Samelius C (2007) Abused women: Health, somatization, and posttraumatic stress. Sweden: Linköping University, PhD thesis. Smith PH, Thornton GE, DeVellis R, Earp J and Coker AL (2002) A Population-based study of the prevalence and distinctiveness of battering, physical assault, and sexual assault in intimate relationships. Violence Against Women 8: 1208–32. Stevenson RL (2003[1886]) Strange Case of Dr. Jekyll and Mr. Hyde. New York & London: W.W. Norton & Company. Strauchler O, McCloskey K, Malloy K, Sitaker M, Grigsby N and Gillig P (2004) Humiliation, manipulation, and control: Evidence of centrality in domestic violence against an adult partner. Journal of Family Violence 19: 339–54. Strazdins L and Broom DH (2004) Acts of love (and work): Gender imbalance in emotional work and women’s psychological distress. Journal of Family Issues 25: 356–78. Towns A and Adams P (2000) ‘If I really loved him enough, he would be okay’: Women’s accounts of male partner violence. Violence Against Women 6: 558–85. Wuest J and Merrit-Gray M (1999) Not going back: Sustaining the separation process of leaving abusive relationships. Violence Against Women 2: 110–33. Yoshihama M (2005) A web in the patriarchal clan system: Tactics of intimate partners in the Japanese sociocultural context. Violence Against Women 11: 1236–62. Zink T, Jacobson JC, Regan S, Fisher B and Pabst S (2006) Older women’s descriptions and understandings of their abusers. Violence Against Women 129: 851–65. Viveka Enander, PhD, is working in the Department of Social Work at University of Gothenburg and at the Västra Götaland Region Competence Centre on Intimate Partner Violence, Sweden. Her research focuses on battered women’s leaving processes. Downloaded from fap.sagepub.com at Goteborgs Univ.- bibliotek on February 9, 2011