Introduction

Since the 1990s, with the development of ‘gender mainstreaming’ (i.e. approach that seeks to institutionalize equality by integrating gender-sensitive practices and norms into public policy structures and processes (Daly, 2005) within the United Nations, domestic violence and the role of the patriarchal environment that propagates this violence (Mélan, 2019) have received significant public attention in Europe. Since then, researchers as Sardinha and colleagues (2022) have shown that intimate partner violence against women is still widespread. Worldwide, nearly a third of women aged 15–49 who have ever been in a relationship have experienced physical or sexual violence, or both, at the hands of their partner during their lifetime (Sardinha et al., 2022). It is therefore important to understand how some individuals or couples manage to desist from violence (Merchant & Whiting, 2018). In Belgium, the law of November 24, 1997, and subsequent so-called ‘zero tolerance’ directives have confirmed the societal nature of partner violence while calling for a retributive justice model, i.e. a unilateral imposition of a penalty for repair of justice (Wenzel et al., 2008), a model particularly discussed by recent research (Vanneste, 2017). The effectiveness of this type of policy has been brought into question by research that has highlighted the limited effectiveness of punishment and batterers interventions, as well as the low degree of self-responsibility perpetrators of domestic violence feel. In particular, this can be seen in the high rates of recidivism (Devaney, 2014), nearly 38% in Belgium (Vanneste, 2017). Studies that have focused on programs aiming to change men’s behaviour and batterer interventions have also questioned the effectiveness of this type of intervention, particularly with regard to recidivism and long-term outcomes (Travers et al., 2021). Identifying the desistance process is therefore primordial but still under development in the field of intimate partner violence (Walker et al., 2017), could provide insights into strategies for the psycho-judicial management of perpetrators. Through discourse analyses, this article examines the processes of change for 13 perpetrators of intimate partner violence (IPV, i.e. any behavior within an intimate relationship that causes psychological, physical, or sexual harm or suffering to those who are part of it (World Health Organization & Pan American Health Organization, 2012)).

The desistance process

Since the second half of the twentieth century, research in psychology and criminology has seen the development of the concept of desistance. Desistance can be summarised as ‘long-term abstinence from crime in individuals who had previously participated in persistent criminal patterns’ and then the maintenance of ‘crime-free behavior in the face of life’s obstacles and frustrations’ (Plesnicar, 2015, p.192). Early studies on desistance highlighted maturation as the process associated with desistance but ontogenic and sociogenic paradigms viewed crime as declining with age alongside social factors that enable change (Glowacz & Born, 2017). However, age alone does not explain desistance. Laub and Sampson have demonstrated the influence of conventional ties on the individual – such as marriage, military service or stable employment – in the desistance process, which they have termed ‘exogenous turning points’ (Laub & Sampson, 2001). These exogenous turning points support a reorientation of the subjects’ trajectory. Building on the concept of desistance, Maruna and colleagues have distinguished between the cessation of violent acts, designated as primary desistance, from secondary desistance, which is a process during which the subject no longer defines himself as an ‘offender’ (Farrall & Maruna, 2004; Walker et al., 2017). The gaze of another who measures and then forgives the offender’s actions constitutes tertiary desitance (McNeill, 2016) or a form of redemption. Desistance is then only possible through a combination of social, contextual and cognitive elements (Maruna et al., 2004; Maruna & Lebel, 2012; Nugent & Schinkel, 2016). Nugent and Schinkel (2016) propose a classification in which ‘act-desistance’ stands is non-offending, ‘identity desistance’ is the internalization of a non-offending identity and ‘relational desistance’ is the recognition of change by others. Thus, it is now generally accepted that desistance is a non-linear process, rather than an endpoint in and of itself and this makes the phenomenon difficult to measure (Maruna et al., 2004; Nugent & Schinkel, 2016; Weaver, 2019).

The desistance process and intimate partner violence

On the whole, it is difficult to define a ‘time frame’ after which it is possible to speak of true desistance (Walker et al., 2013). In the context of intimate partner violence (IPV) it is even more difficult to conceptualise desistance. Indeed, domestic violence is often cyclical in nature as it alternates between phases of varying length, frequency and intensity. In other words, desistance, as a cessation or reduction of violent acts, appears to be an integral element of the IPV dynamic.

Some studies have addressed the subjective experience of perpetrators of partner violence and observed individual factors that are involved in change processes, such as empathy, isolation from affect, and planning a nonviolent life associated with increased responsibility (Chen et al., 2020). However, while individual change remains the primary reference point for assessing desistance among perpetrators of domestic violence (Whitaker et al., 2010), desistance can be approached from the perspective of marital dynamics and relational changes associated with partner behaviours (Walker et al., 2013; Giordano et al., 2015). The process of desistance and self (re)definition in the context of IPV raises questions of self-identity or relationship identity. In this vein, Giordano and colleagues (2015) have highlighted that the theme of ‘the relationship’ is at the heart of the changes and motivations of perpetrators who successfully desisted (Giordano et al., 2015).

For the couple, desistance means focusing on oneself and changing the dynamics which lead to conflict in order to build trust, commitment, and satisfaction. In Merchant and Whiting’s (2018) model, desistance within relationship begins with a turning point, an event associated with a reset that leads to a conscious decision to change. With this initial recognition then come the initial changes that reinforce the decision to change. For many couples, this means focusing on themselves and changing the dynamics that lead to conflict. This study further supports the notion that desistance is both an individual and a dyadic process that includes communication and a perception of change, such as evolving conflict dynamics. More individual mechanisms, like empowerment, are also dependent on partner support to succeed (Merchant & Whiting, 2018). Learning new communication and conflict management styles, for both the perpetrator and victim, and perceived changes can lead to lasting change within the relationship (Walker et al., 2013). Thus, while interventions involving both partners are controversial, not least due to legitimate concerns about victim safety, there is a consensus that care must consider not only individual processes, but dyadic processes, and the systemic relationship between the two (Whitaker et al., 2010; Merchant & Whiting, 2018).

The objective of this study is to better understand the process of change at work in IPV male perpetrator desistance. The research project has been approved by the Ethics Committee of the Faculty of Psychology, Logopaedics and Educational Sciences of the University of Liège, Belgium. Its collection method involves a comprehensive approach based on a qualitative and inductive ‘life courses’ methodology which was defined in consensus with multidisciplinary professionals from a Belgian research group working within the framework of the Federal IPV-PRO&POL research project (Belspo, 2017–2021).

Method

Participants

The sample consists of 13 male participants who are or have been perpetrators of violence between heterosexual partners. Nine participants were between 30 and 49 years old at the time of data collection; one participant was under 30 years old and three were between 50 and 60 years old. All participants were Belgian residents, but two were not Belgian nationals.

Seven men reported having been violent in a single relationship and six in several conjugal relationships. The relationships in which they acknowledged having been perpetrators of violence varied in length from one year to 34 years, with an average of seven years. Six participants were in the violent relationships at the time of the interview. Out of this group of six, one said that he had not been violent towards his partner for at least three years (JL) and one was in the process of separating from his partner (FM).

The pathways are diverse. All of these men had been in contact with the police following a complaint to the police or call for help from their partner. Seven of them said that they had been convicted for acts of violence committed against a partner. Out of all the subjects, six were involved in a therapeutic treatment program for batterers at the time of the interview. Of these six men, one was participating voluntarily (JL). Of the other five who received a court order to attend this program, one was serving a sentence for acts not related to IPV (GB); one was just beginning the procedure at the time of the interview (VM) and another was convicted for acts of IPV while undergoing the batterer program he had voluntarily enrolled in (FM). In addition to these six men, two declared that they had taken steps voluntarily to initiate this type of treatment (JS and FP). The other five subjects were currently in prison but only two (FS and NP) had been incarcerated for violence towards their partner. In the case of NP, his acts of violence had led to the death of his partner.

Procedure

Participants were recruited (2019–2020) through newsletters and posters placed in various medico-psycho-social and judicial offices. People who were part of a care system for violence between partners were reached via social media posts and word of mouth. In order to collect the testimonies of all persons who considered that they had experienced a violent intimate relationship, we included no specific definition of IPV in the recruitment material. The composition of the sample was not based on judicial variables, such as a formal sentence or on the actual cessation of violence. The purpose of this study was to examine the dynamics of violence and, more specifically, the processes of change and exit from violence among people who recognise themselves as perpetrators of violence in an intimate relationship. Meeting subjects with different trajectory profiles allowed us to understand the dynamics at the root of the change processes.

We only included participants over eighteen years of age who had the ability to understand and express themselves independently in French and who had experienced violence in Belgium or had been taken into the care of a Belgian institution. The semi-structured interviews were recorded using an audio recorder. Before each interview, participants were reminded of the rules regarding confidentiality and anonymity verbally and via a consent form. All participants gave their free and informed consent. Two interviews were not recorded and note-taking by hand was done in these cases. The interviews lasted between one and four hours, averaging two and a half hours. Interviews were conducted face-to-face in a private and quiet room. As we were aware that recalling painful events can induce feelings of anxiety, we gave the participants telephone numbers or addresses of organisations that could support them and help them if they felt the need after the interview.

Interviews

We developed a semi-structured interview guide based on up-to-date IPV literature and a ‘Life Courses’ perspective. This narrative strategy allowed us to trace life trajectories that reflected the meaning subjects gave to their life story (Rosenthal, 1993; Kacen, 2002). Through this semi-structured interview, participants could discuss, at their own rhythm, their representations and experiences of violence (‘Can you tell me about the acts you committed?’); how they perceived changes in the dynamics of violence and the process of change and/or of stopping the violence (‘What would stop the violence?’; ‘What would you need to stop the violence?’; ‘If the violence stopped, can you tell me how it stopped?’; ‘What led to the decision to change?’; ‘Have you received support?’; ‘How have you maintained this change?’). The interview guide permitted the analyse of the processeses of change at different stages of violence trajectories. The interview made it possible to consider, by adapting questions to the subject’s situation, the experience of change through a narrative approach of retrospection or projection. We also used a qualitative Life History Calendar (LHC; Nelson, 2010), which is an instrument for exploring life trajectories. By using a large sheet of paper and different coloured markers, the temporal domains and markers were affixed to the timeline to reflect the subject’s narrative. The LHC allowed for a dynamic analysis of violence by capturing the events and their sequencing, as well as the context in which they occurred.

Data analysis procedure

Researchers carried out a thematic analysis of the interviews based on the method presented by Paillé and Mucchielli (2016). This approach focuses on what is fundamental in a narrative (i.e. ‘the story told’) to understand an issue. After word-for-word transcription and an initial reading of the interview, the discourse is divided into units of meaning. Codes, or ‘themes’, are then assigned to these sequences to describe their content. In order to minimise the risk of interpretation bias, the themes are defined in such a way as to respect the subjects’ narrative (words or expressions) as much as possible. Then a thematic tree is created. In this study, for each case, a thematic tree was carried out alongside the individual life calendar in order to highlight temporality and trajectory. The trajectory analysis was first carried out by the lead researcher then reviewed by another researcher who works in the field of violence between partners, and finally discussed by both. The combination of semi-structured interviews and LHC method permitted us to analyse the dynamics and experiences of intimate partner violence with a particular focus on the process of desistance. Furthermore, NVivo software was used to allow for more precise coding and sub-coding of each unit of meaning and the emergence of new categories, broader categories or finer trees (Daigneault & Pétry, 2017). The absence of any new themes emerging from the thematic analysis after 13 interviews led us to conclude that empirical saturation has been reached.

Maintaining the intergrity of a qualitative methodology

The thematic analysis was carried out from a phenomenological perspective that expects the researcher to focus on the meaning the subject gives to their narrative and encounter the subjectivity of the subject. Thus, the stages and subjective changes presented are the ones described by the perpetrators. However, we worked with a comprehensive approach by considering the dynamics of subjects’ interpretation of the world in given contexts and the intervention. The triangulation of perspectives of researcher, theory, and field has been an important contribution through circular work in which the field is read through the researcher’s concepts and knowledge, and then this same knowledge is refined through field experience (Charmillot & Dayer, 2012).

Findings

Our analyses allow us to understand the IPV perpetrators’ exit processes as part of the continuation of relational dynamics anchored in a life history. Studying the 13 perpetrators’ discourse highlighted a common trajectory in which two types of violence dynamics could be distinguished. The first dynamic of violence, the most present in our sample (Type 1, 9/13), appeared as a means to maintain the domination over others – and the victim. The second dynamic (Type 2, 4/13) is singularised by the posture of powerlessness of the subject.

The results of the thematic analysis revealed eight main themes, which can be grouped into three main categories – engagement, dynamics, and exit from violence – constituting an exit trajectory from violence. Under the first category, “engagement in violence dynamics”, we find the themes (1) history of violence and development of relational scripts, and (2) divergent expectations between partners. The second category, “the dynamics of violence” comprises of (3) understanding the violent event; (4) the attribution of responsibility; and (5) the escalation of violence. The final category, “the engagement in a process of change”, includes the themes (6) oneself and couple reassessment; (7) intrasubjective changes; and (8) the development of strategies to end the violence. These main categories are presented in this article, section by section, differentiating between the two types of dynamics mentioned above.

The process of exiting from violence

Engagement in violence dynamics

(1) History of Violence and Development of Relational Scripts. In both dynamics of violence, subjects spontaneously referred to experiences of family abuse, during the interview. In the type 1 (T1) dynamic, the subjects added that they had been perpetrators of multiple acts of violence or delinquency sometimes since their early adolescence: ‘There is a kind of competition that is already starting to take place, each of the two wanted to be the boss. I see that in society it’s like that too…when you want to be the boss you do things to become the boss’ (FS). The thematic analysis also highlights another recurring theme in the perpetrators’ discourse, that is the reinforcement of masculinity that is imposed for some in male/female relationships: ‘Yeah I’m a misogynist. If I’ve already told her to shut the fuck up, I do what I want. I have no limits. She has nothing to say to me’ (FS). Concerning the type 2 (T2) dynamic, the relationship with the mother emerged as a central theme: ‘[I had] an experience where things occurred with my mother that were effectively incestual. Where I was […] castrated! […] There was a whole process of psychological castration that took place’ (VM). At the same time, the interviewees expressed feelings of victimization in a society that they perceived as unfavorable to men: ‘It is very complicated for a man […] to recognize that his power has been devastated to that extent’ (VM); ‘So yes [women] ha[ve] [their] place in the job market but that has made social changes that are extremely important at that level. And now we are also in an age of the “useful object”, the disposable object, and we are not far from it [becoming a disposable object] ourselves’ (GB). The theme of masculinity also appeared in this type of dynamic. Some subjects explained that they questioned the notion of masculinity, their own in particular, and felt confronted by gender stereotypes.

(2) Divergent Expectations Between Partners. The interviewees spoke about divergences between the partners’ expectations that they have perceived. Participants spoke of discrepancies between the expectations they had of their partner and the response given by the latter to these expectations. In the type 1 dynamic, this moment was associated with themes related to ‘power games’ between partners: ‘In fact, I had an image, the image I saw… it was the one I had at home… it was the guy who comes to a [clean home with dinner on the table]’ (FS); ‘Always a question of control, always a question of wanting control, […] a question of having the last word’ (JS). In type 2, these discrepancies were linked, according to the subjects, to a feeling of being infantilised by their partner or of not existing in the couple: ‘I remember at the beginning, she was so attentive, I said to her : listen you are harming my autonomy’ (FM). They spoke of feelings of insecurity, powerlessness and even fear: ‘I have the impression that there was something psychotic about her that sometimes frightened me’ (GB).

The dynamics of violence

(3) Understanding the Violent Event. The analysis of the discourses highlights different forms of violence within one or more relationships. The two types of dynamics manifested principally in psychological, verbal and physical forms. There were also cases of sexual violence, but to a lesser extent. In type 1, the subjects associated the act of violence with jealousy or considered it as a reaction to a presumed infidelity. Alcohol and drug use were omnipresent themes and were perceived as triggers for the violence: ‘We were drunk together, we drank together. We’d start early in the morning, we’d go back to the rosé at 9:00 a.m.’ (NP); ‘The alcohol wasn’t the cause, but it was the trigger, I think’ (RM). In T1, men associated violence with a form of impulse or an intention to hurt. Some spoke of a need for control or domination and equated it with a perceived abusive attitude of the partner: ‘[Whether I] wanted it or not, I hurt the person, when she started to hurt me or I started to hurt her. Then, what makes the other person sad leads them to hurt the other’ (JS); ‘The first reason was, I told you, to hurt’ (JL); ‘It is the dominant male’ (FS). In the type 2 dynamic, subjects linked the act with depression, anxiety or panic attacks: ‘Because I thought for a long time that I was having nervous breakdowns, but, in fact, they were panic attacks. I’m pretty sure of that’ (FM). The results underline that violence is linked to themes such as ‘powerlessness’: ‘In any case what I remember is that she was standing on the bed because she was not very tall, it is precisely because she wanted to take over, domination was possible. And there was an accidental punch’ (GB) or ‘rebirth’: ‘It’s kind of like a baby punching or kicking to get out’ (VM).

(4) The Attribution of Responsibility. In both types of dynamics, the perpetrators referred to individual responsibilities. However, there are nuances between the two categories. In the type 1 dynamic, interviewees attributed responsibility for the violence to the partner: ‘I’m sick of women who let themselves be manipulated, who let themselves be robbed, who let themselves be beaten, who let themselves be insulted, treated badly. I would love to meet a woman who doesn’t let it happen’ (FP). Their discourse revealed forms of minimisation or normalisation of acts of violence: ‘So here is violence that could be termed sexual violence…that for me was normal, in fact’ (FS); ‘I grabbed her and I threw her in the boat like that. She fell like that, her heel broke and she hurt her ribs because she fell on the table. She had big bruises on her arms but she was 100 pounds’ (NP). In the type 2 dynamic, violence appeared to be minimised or even denied: ‘I think that I have very few memories in fact. So, either it’s that there’s no reason to remember anything because it didn’t, there wasn’t much that happened’ (VM). As with type 1, responsibility for the violence was largely attributed to the partner: ‘Because I know that on a lot of things, probably unconsciously, she also tests me a lot’ (FM). However, the mother was also designated, by some perpetrators, as one of the causes of their relational problems: ‘By being initiated, if I may say so, to a non-relationship, that is to say, fundamentally, to not having been able to develop a minimum of trust in the other, in the authenticity of the other’s love’ (VM).

(5) The Escalation of Violence. All participants’ mentioned an escalation of violence. They associated this escalation with an intensification of verbal violence, verbal violence that became physical, and/or an increase in the frequency of violence by both partners: ‘From the moment that the nastiness sets in […] you get into this cycle where there is no more dialogue and then you get into this spiral again until the moment you come to blows’ (NP). In one of the T1 cases, the escalation of violence led to the death of the spouse: ‘When I slammed her in the doorway, I broke her thyroid cartilage with my thumb, so that […] 4 hours later she was dead in the chair’ (NP). In some cases, the escalation of violence was linked to the separation of the partners. The interviews highlighted new forms of post-separation violence, such as harassment, mostly in T1: ‘I did not accept the separation and… because you have such strong love… you end up harassing them’ (NP). Some men also associated the post-separation escalation with a role reversal, where the victim became, according to them, the perpetrator, and they became the victims. This mainly concerned type 2: ‘Because the problem with the violence is that it didn’t stop, now it’s continuing, there’s still violence between her and me. But slowly it is becoming more like she is violent towards me’ (FM).

The engagement in a process of change

(6) Oneself and Couple Reassessment. The majority of the interviewees noted a shift in the dynamics of violence following a break-up or an ultimatum imposed by the partner. Some of the participants emphasised that they had been confronted with the impact of the violence. They associated it with this tipping point which promoted a reassessment process. In type 1 dynamics, perpetrators of violence associated the initiation of a process of change with the partner because the latter confronted them or gave them an ultimatum: ‘She told me to stop: either you get treated for your violence, or you get the hell out’ (JL); ‘And I had a big emotional shock when my partner took everything away, in two hours, all the furniture was gone. She left’ (JS). Others stressed that their partner was responsible because they could leave if they could not stand the situation: ‘And then I didn’t understand either, [she] just [had] to leave’ (FS). Some T1 men grew conscious of the impact of their violent acts on their close family, and on their children in particular: ‘Somehow it allowed me to act on it, to think. I don’t have the right to do this. I don’t want that for my son, nor for my ex’ (FS). Imposed distance, such as incarceration, could also lead the perpetrator to engage in a process of reassessment: ‘Afterwards, when I went back to prison, I understood that she was not a woman for me’ (JJ). In the type 2 dynamic, the analysis highlighted themes of victimization and powerlessness. In the perpetrators’ discourse, the re-evaluation of the relationship dynamics occurred in paralled with a feeling of becoming the victim of their partner developed : ‘She is so into pathos, she told me : it’s odious the suffering you’re inflicting on me. Magnificent. And in everything she had to say, in the end, it was only negative and that I had not brought her anything, that is not correct’ (GB). This re-evaluation also gives rise to the feeling that they have been a victim of society: ‘But the problem is that, that is why, I wanted to say my mother on one side, society on the other, which is also a matrix. And, in fact, there is castration and dispossession’ (VM).

(7) Intrasubjective Changes. Following or occurring at the same time as the re-evaluation of the self and the couple, the participants referred to a moment of self-reflection and internal changes. In the type 1 dynamic, the main themes raised focused on the awareness of the abusive nature of the dynamic, as well as on individual changes involving the management of emotions, anger and possessiveness: ‘With hindsight and experience you learn that it doesn’t help. Because by being possessive you close a lot of doors’ (NP); ‘But somewhere I also think there must be a problem, because it doesn’t make sense to [be] like that’ (AS). In this context, some participants discussed in detail the image they had of themselves, the construction of a new identity and on the importance of differentiating ‘being violent’ and ‘doing violence’: ‘The image of myself. The image of myself’ (JS); ‘It bothered me a little to discover myself like that, I don’t want to be someone like that. […] Well, someone who does violence, who produces violence’ (RM); ‘From then on, I started to be someone else’ (JL). At this point, the discourse analysis highlighted the centrality of formal or institutional interventions that accompanied the change process. Subjects in the T2 dynamic focused their discourse on the introspective approach they had taken towards their history and their childhood. They questioned an ‘abandoned’ aspect of their personality and/or looked for an explanation in the medical sphere: ‘I don’t want to be relieved of responsibility for my actions. What I hope is that if there is a pathology, well, there will be a treatment and maybe a solution’ (FM); ‘The history that I have in relation to my mother that I will make the current women pay for’ (GB). Once again, the theme of masculinity emerged in the interviews. The participants explained that had questioned, or were still questioning, their sexuality and/or their masculinity: ‘I hope it’s not latent homosexuality […] maybe it’s complicated to be with a woman, but basically I love women’ (GB); ‘I wondered if I wasn’t a repressed homosexual sometimes. Because I really asked myself questions!’ (VM). Participants linked this introspective work with their involvement in psychological follow-ups. They referred to judicial and mandated resources with terms that were focused on distrust and fear: ‘The first time […] I felt [the consellor] judging me. The second time I felt she was much more benevolent […] But that doesn’t necessarily mean that she’s always right’ (FM).

(8) The Development of Strategies To End the Violence. Finally, the interviewees referred to strategies they had developed to avoid violence. Certain themes were common to both dynamics of violence, such as distancing themselves from their partner, or from women more generally; managing impulsivity or anger; changing their lifestyle by stopping or decreasing alcohol and drug consumption and investing their time in various activities: ‘I love sport, it allows me to evacuate a lot, to have confidence in myself, to evacuate but also to be able to keep a healthy lifestyle, and always in this set framework or structure’ (JS). The support of the partner was also central and appeared in the interviews as an element supporting the process of bringing an end to the violence within the couple. In the type 1 dynamic, the participants discussed other themes, such as communication, the establishment of limits – for which the responsibility remained attributed to the spouse – or the beginning of new relationships more in line with their perception of the couple: ‘We have our disagreements like all couples, sometimes we don’t agree etc., but they are not arguments’ (NP); ‘When I come back I’m going to find a woman, a good one. So that she can take care of my children and me too’ (JJ). In type 2, men emphasised the management of empowerment: ‘The problem of dependence is not to live without dependence. Because there are a lot of people who are gradually becoming aware of this issue of emotional dependence’ (VM). Moreover, they stressed the need for social recognition: ‘I found a part-time job as a night nurse in a nursing home, so it does me a lot of good, in terms of my autonomy, to regain my financial health, and also in terms of my social recognition’ (GB).

Discussion

This paper aims to comprehend the desistance process in the case of 13 male IPV perpetrators. The interviews we conducted with men at different stages of their violence trajectory allowed us to develop a progressive schema of exiting violence. The thematic analysis of the narratives revealed eight key stages leading to an exit from violence : (1) the life history and development of relational patterns; (2) a confrontation with the ideal couple and reality; (3) acting out; (4) disempowerment and expectation of external change; (5) an escalation of violence; (6) reassessment of self and couple, (7) the initiation of intrasubjective changes and (8) the development of strategies to end violence (See Fig. 1). Although presented in a linear format, the process is not fixed. Individuals may move through a stage or back and forth between them. These stages are not mutually exclusive or sealed-off from one another. A person may find themselves going through two stages; for example, between an escalation in the violence and a ressassement of themselves and the couple. An end to the abusive relationship may be considered at each stage too.

Studies in IPV exit processes have considered change as an incremental mechanism, primarily emotional and cognitive, according to which the exit process begins within the relationship and may extend beyond the physical separation (Cluss et al., 2006). In this study, our results suggest that the dynamics underpinning the process of ending violence are a continuation of the relational dynamics developed during the couple’s relationship, as well as also being link to earlier life experiences (see Fig. 1, stage 1). The participants’ trajectories appear to be rooted in childhood histories and family life characterised by strict boundaries in the case of the T1 dynamic and more diffuse in the case of the T2 dynamic. It is now well-established that the first experiences of parental care have a central role in the development of attachment in children, but also later in adolescence and adulthood (Ørke et al., 2021). The development of an insecure attachment pattern in childhood can affect couple relationships through bonding strategies that are associated with anger and aggression (Ørke et al., 2021). Type 1 dynamic appears to be closer to an avoidant attachment mode, where the person appears rather disengaged. Individuals with an avoidant attachment orientation are more at risk of perpetrating violence, including violence between partners, because of their hostile behaviour and poor conflict resolution skills (Spencer et al., 2021). In this study, violence between partners appeared in a context of multiple acts of violence, delinquency and rejection of authority for a number of participants. Of the nine T1 men, five were in prison for various illegal acts (assault, theft, drug trafficking, murder) at the time of their interview. This type of dynamic closely mirrors one of the major types of male perpetrators described by Babcock et al. (2007): ‘characterological’ perpetrators for whom violence is not necessarily limited to the family and is employed as part of an effort to dominate a partner. The subjects displaying type 2 dynamics appear to have an anxious attachment orientation and seek closeness with the others. Their speech shows a search for a ‘structure’, represented by the partner, and for containment that would allow them to become autonomous.

Fig. 1
figure 1

Schematic of the change processes undergone by male IPV perpetrators

Full size image

Subsequently, the establishment of an intimate relationship with a partner could, in some cases, brought about a confrontation between the reality and an individual’s idealised view of such a relationship (see Fig. 1, stage 2). The relationship will either undermine their sense of control (T1) or give them the impression of being taken over, ‘swallowed up’ by the partner or the relationship (T2).

Then (see Fig. 1, stage 3), the former (T1) associate the violent act with a need for control or domination and assimilate with a perception that the partner is violent. The second group (T2) focus on their experiences of anguish, panic and powerlessness. In both cases, the responsibility for the dynamics of violence is most often directed towards the external.

Denial of responsibility, victim blaming or other external attributions, such as economic and professional difficulties, are indeed common to perpetrators of violence (Lila et al., 2014). While most of our subjects recognised their acts of violence, a posteriori, they were often minimised. Few took responsibility for the violent dynamics and acts (see Fig. 1, stage 4). The acts of violence are not denied but they are not recognised as such. In other words, they were able to acknowledge having slapped their partner, for example, but would minimise the act to the point of not considering ‘the slap’ as an act of violence. This finding supports the notion of ‘deliberate cognitive distortion’, proposed by Maruna and colleagues (2004). The act is acknowledged, but not their individual responsibility. According to our results, this way of disclaiming responsibility appears in the type 1 dynamic as a refusal to be associated with the identity of perpetrator. In type 2, it is the individual’s profound identification with the status of victim that prevents them from taking responsibility. At this stage, we cannot speak of a process of ending the violence as such. Indeed, even when violence may lead to a separation of partners, we can observe, at this stage of the processus, a repetition of patterns in the following relationships. As many studies have already shown, the breakup of a violent relationship does not always initiate a process of desistance (Halpern-Meekin & Turney, 2018). Moreover, external attribution of responsibility for the violence appears to be a risk factor for the escalation of violence within the couple.

The fifth stage (see Fig. 1, stage 5) not only sees an escalation in the violence but also, and more importantly, a perceived change in the dynamics of violence that now includes the partner leading to bidirectional or mutual violence. Perpetrators met may then associate their violence with their partner’s behaviours that they consider violent, which leads to an escalation of violence (Giordano et al., 2015). That consolidates the external attribution mechanisms of the violent dynamic and victim blaming.

However, escalation may also allow them to move to another stage through a re-evaluation of the situation (see Fig. 1, stage 6). Our results demonstrate that, in the case of T1, the dynamic changes as a result of separation – break-up or incarceration – or an ultimatum imposed by the partner that most often follows an escalation of violence. In the case of T2, the relationship dynamic reassessement overlaps with the feeling that they have become a victim of their spouse. The change factor is mainly external again, but it also appears to be strongly dependent on the couple’s relationship. Resources are mainly the family and, above all, the partner and children. Indeed, experiences within or outside the intimate relationship can lead to the cessation of abusive behaviours (Walker et al., 2017). Some of the resistance strategies commonly implemented by victims of violence, such as help-seeking, active opposition or exiting appear, in this study, to have played a role in the perpetrators’ shifting dynamics of violence and/or disengagement. This dyadic aspect opens up a new line of analysis for desistance (Walker et al., 2013; Giordano et al., 2015; Chen et al., 2020). In our sample, strategies enacted by the partner can lead the perpetrator to re-evaluate the couple’s dynamics. This is especially the case when the partner supports the perpetrator in their search for help or care. On the other hand, judicial interventions are, at this stage and for both types of dynamics, associated with feelings of anger, injustice and/or insecurity and they also increase the risk of escalation. Judicial institutions, and the spouse who has called upon the judicial system, appear at this point to be a hindrance to the individual’s goals, which can provoke a state of negative emotionality and lead to violence (Olson, Martin & Connell, 2020; Slep et al., 2021). Desistance thus begins with events that are associated with a reset and then lead to a conscious decision to change (Merchant & Whiting, 2018). Turning points can trigger change in isolation, but they are not enough, they must be perceived as sufficiently relevant. If so, they facilitate change when the person reaches the decision point, i.e. the autonomous decision to change (Walker et al., 2017).

What happens next is primarily intrasubjective (see Fig. 1, stage 7). For our participants, an internal change occurred, particularly in terms of responsibility and identity. The view of others and social interactions are central factors in change processes, including the identity change process (Farrall & Maruna, 2004; Maruna & Lebel, 2012; Copp et al., 2020). Chen’s study further demonstrates how the expression of an opinion regarding one’s own ‘bad behaviours’ is related to intrasubjective change (Chen et al., 2020). In these cases, the regard of a partner and children, as well as that of various intervenors – as representatives of society – appears to be central to the implementation of a process of change. Some men in the T1 dynamic emphasised that they experienced, or had experienced, difficulty in shedding the label of ‘violent man’. Our analysis, as other studies, emphasises that it is through formal and especially informal exchange and through social support, that perpetrators make these identity transitions; change is not only spontaneous, it’s mostly assisted (Dufour & Villeneuve, 2020; Maruna, 2020). Our results support the idea that these transitions must be integrated into a life history by anchoring the perpetrators’ journey in their childhood experiences and then allowing them to project a future. Moreover, it is interesting to note that fatherhood, as has already been highlighted for women (Rodermond et al., 2016), is an important supporting factor. How perpetrators perceive antecedents to their violence helps to initiate a transition towards change, towards a ‘new way of being’. In other words, these men actively participate in stopping the violence by learning to manage the antecedents and triggers of the violence (Walker et al., 2017). However, these changes may cause fear and anxiety about the unknown (Gålnander, 2020). Typically, in the case of Type 2 dynamics, some men prefer to emphasise elements beyond their control, such as behavioural or mental disorders. Therapeutic follow-up, therefore, essential during this stage.

Finaly, perpetrators met will put in place strategies to bring an end to the violence (see Fig. 1, stage 8), which are globally similar between the two types of dynamics, such as managing drug and/or alcohol consumption, anger and impulsivity. In both cases, relational and community support factors are fundamental. Among the resources upon which perpetrators might draw, we note a commitment to therapeutic work, as well as elements such as spirituality and the establishment of new relational bases. In the type 1 dynamic, the perceived resources mainly concern the setting of limits for oneself but also within the couple and emphasises, with the exception of a subject who said he had been desisting for 3 years, the rejection of the label of perpetrator by speaking of non-judgment. In the type 2 dynamic, the work of empowerment, the need for social valorisation and the recognition of one’s experience of victimhood are central.

Limits and Perspectives

The sample on which this research is based comprises 13 interviews. Two interviews were recorded by note-taking, which may have limited their in-depth analysis. Moreover, the sample consists of heterosexual male perpetrators with a history of violence who were met during a single meeting and through a psychosocial or judicial institution. As a consequence, our results represent the experience of the perpetrators’ exit from violence at a very specific moment. Considering that the desistance process is a long and non-linear process, longitudinal approaches would allow us to broaden our findings and highlight the internal and external strategies of self-regulation adopted over time (Villman, 2021). Other studies could also compare our results with other populations, such as women who have committed violence against their partners or homosexual couples. Moreover, while no interview – with the exception one subject – described intentional gender-based violence, the theme of masculinity was redundant in our analysis. In both of the dynamics perpetrators’ concept of themselves as ‘men’ was undermined by the dynamics of violence or in the process of change. Indeed, the way the individuals negotiate the belief systems embedded in their society can have an impact on criminality as well as on the success of desistance (Bersani & Doherty, 2018). Accordingly, it is important to focus some of our attention on child abuse and the unequal dynamics that govern male–female relationships from the earliest age too for the development of preventative and intervention policies in the domain of intimate partner violence. According to Rollero (2020), primary prevention perspectives addressing gender norms and stereotypes would indeed reduce the risk of partner violence. Also, for the treatment of perpetrators of violence between partners, the construction of a conception of a “pro-social self”, through the recognition of patriarchal norms and the learning of non-violent behavior is essential (McGinn et al., 2020).

Moreover, while previous research points to a desistance process that involves an identity shift where the subject no longer defines themselves as ‘an offender’ (Farrall & Maruna, 2004; Walker et al., 2017), there was no identity shift associated with the ‘partner violence perpetrator’ identity in our results, with the exception of one participant who said he had been desisting for 3 years. The desistance process is a long process, so it is possible that the participants we met had not yet made this identity transition. We are also aware that some of our participants were in prison at the time of the interview and that this may have had an impact on the desistance process. However, recent studies have shown that change processes can be initiated in prison. While there are many social and structural challenges, by recognising and supporting the motivations for desistance, prisons programmes may be able to encourage it (Villman, 2021). This nevertheless raises the question to how important it is that those responsible for the prison system recognise the desistance process. Our study has mainly highlighted the role of the partner and the recognition of the impact of the actions for the engagement in a process of change. Our findings reaffirm that couple-level dynamics with individual partner characteristics come into play in the disengagement process (Walker et al., 2013; Giordano et al., 2015; Dziewa & Glowacz, 2021). Thus, desistance in the case of domestic violence needs to be thought of it in terms of self-identity, the identity of the couple’s relationship and, more broadly, the relationship with the others.

Implications

There is already a large body of work that has explored the effects of correctional or penal interventions on desistance and reintegration processes. However, the extent to which this work has adequately influenced criminal justice policy and practice is debatable (Weaver, 2019). The results of this study provide interesting data for the treatment of perpetrators of partner violence. Even today, Western strategies to address domestic violence are primarily based on three pillars: prevention of occurrence and recidivism, victim care, and perpetrator punishment (Devaney, 2014). In Belgium, as in much of Europe, criminal policies reaffirm the unacceptable and criminally reprehensible character of intimate partner violence. Viewed from this perspective, the sooner a perpetrator is confronted with the law and sanctioned, the more effective intervention is a barrier against such violence (Vanneste, 2017). These measures make the penal system a strong warning signal wanting to force the aggressor to recognise his behaviour as problematic and make him accountable for his actions (Silvergleid & Mankowski, 2006). These practices may be counterproductive, however, as studies have criticised their effectiveness in cases of partner violence; criminal sanctions most often result in a sense of injustice and anger (Devaney, 2014). Despite this, the maintenance of such a policy is linked to the important symbolism of the sanction, which clearly denounces IPV as criminal (Vanneste, 2017). This policy includes the detection and prosecution of perpetrators of violence, as well as court-ordered participation in therapeutic and accountability programs for offenders. In Europe, the cognitive-behavioural and profeminist approaches of the ‘Duluth-model’ are the most widely used in the treatment of perpetrators of violence (Hamilton et al., 2013). Both of these models, however, have been criticised by desistance studies (Patton & Farall, 2021).

Rehabilitation work must be able to emphasise the possibility and the benefits of change and provide resources that allow participants to perceive and make progress. This involves teaching cognitive skills as well as values that they can integrate into their identity and behavioural patterns (Hamilton et al., 2013). Taking responsibility is one of the particularly important factors that can be addressed in follow-up work with perpetrators of partner violence (Rollero, 2020). Nevertheless, such work must be able to recognise and take into consideration different relationship patterns, different patterns of disempowerment, and the different therapeutic needs that arise from these. One method for adjusting intervention formats is the implementation of interventions according to the risk-need-response (RNR) framework (Andrews & Bonta, 2010). Recent work by Travers and colleagues (2021) has shown that intervention formats based on the RNR framework can have significant effects. This RNR model involves tailoring treatment intensity to an individual’s risk by focusing on risk factors that have been empirically associated with general criminal delinquency, such as, but not limited to, a history of antisocial behaviour, family/marital status, leisure problems, or substance abuse. It also takes into account the individual’s ‘criminogenic needs’, in other words, the needs potentially related to delinquent behaviour (Lila et al., 2014; Travers et al., 2021). Moreover, motivations for change appear to be important variables for desistance. Beyond accountability, integrating the notions of needs, hope and well-being in work with perpetrators would favour their involvement in a process of change (Glowacz, Puglia & Devillers, 2020; Olson et al., 2020; Patton & Farall, 2021).

Through narrative identity, i.e., how people interpret their lives, empowerment work can trigger corresponding changes in one’s way of thinking (Pals, 2006), probation or rehabilitation work should support an individual’s capacity to plan for and control their future, as well as foster their engagement with new identities, all the while allowing the participant in such work to make sense of their previous life experiences (Maruna et al., 2004). In order to achieve this, confrontation with different levels of reflection and understanding by means of individual therapeutic follow-up and/or talk groups that include participants at different stages of change appears especially relevant (Di Piazza et al., 2020). In this study, the role of the victim was central to the perpetrators’ discourses in both dynamics of and disengagement from violence. By symbolising the existence of other moral universes, Ward and colleagues propose that the presence of the victim – or a victim, while controversial, could be helpful in working with perpetrators of domestic violence. Recognising the relational dynamic and considering the perpetrator as part of a relationship would ensure greater investment in the desistance process (Ward et al., 2014). Along with the work carried out within a formal support network, the process of recognising and changing behaviours must be supported by the informal network (Walker et al., 2017).

Conclusion

The thematic analyses discussed above have revealed two types of violence dynamics that influence the desistance process among perpetrators of partner violence. Anchored in life histories characterised by violence, type 1 dynamics of violence appear through the reinforcement of masculinity, the search for control and/or a framework. In this dynamic, perpetrators feel that the relationship has undermined their dominant position and their violent responses serve to keep control. For type 1 perpetrators, the process of desistance requires them to internally reorient the attribution of responsibility for the acts they committed. Perpetrators in the type 2 dynamic, on the other hand, question their masculinity. In this case, couple life confronts the individual with inability to achieve autonomy and violence manifests itself as a search for individualisation. For type 2 perpetrators, the desistance process requires retro-introspective work that is focused on their relationship with the other. Beyond legal sanctions, stakeholders, in the legal and therapeutic fields, must be able to support the processes of desistance by adjusting intervention strategies to the dynamics of violence and the exit trajectories from violence in order to promote the rehabilitation of perpetrators of intimate partner violence.