Beyond ‘Safeguarding’ and ‘Empowerment’ in Hong Kong: Towards a Relational Model for Supporting Women Who Have Left their Abusive Partners

This project explores the post-separation needs of Chinese women in Hong Kong who have left their abusive partners and how they might be addressed The project aims to provide insights for improving the local domestic violence service, whose main focus is on crisis intervention. Cooperative Grounded Inquiry (CGI) was developed as a novel participatory action research methodology (PAR) for fostering collaboration between social work practitioner-researchers and women service users. Its purpose is to generate useful knowledge and provide support for abused women and their children. The project involved 7 Hong Kong Chinese women as participant-researchers. The inquiry group met at least once a week for 6 months to explore the post-separation needs of the women and their children, and to implement and evaluate the practices/services developed through this project together in a participatory manner. Women participants identified the problems of doing either ‘victim’ or ‘survivor’ that respectively underpin the ‘safeguarding’ and the ‘empowerment’ models; and they developed practices for ‘doing being oneself’ beyond the victim-survivor dichotomy. This paper presents the changing self-narratives of women participants over the research project, from victimhood to survivorhood and from survivorhood to survivor-becoming. These narratives demonstrate the importance of safeguarding women’s space for undertaking symbolic action and of empowering them through using volcabulary that can help them describe themselves/their experiences differently from mainstream discourses. Women’s narratives highlight the existing ‘planetary difference’ between the safeguarding model, which treats women as helpless and vulnerable and in need of external support, and the empowerment model which treats them as powerful, resilient and with resources and solutions to problems. The study transcends the victim-survivor dichotomy and service models by proposing an alternative relational model that emphasises power sharing in making sense of abusive experiences and finding one’s own voice in a supportive community.


Introduction
The Tin Shui Wai tragedy in 2004 saw a mother, Kam Shuk Ying, and her two twin daughters stabbed to death by her abusive husband. The tragedy unfolded even though Kam Shuk Ying and her friend had routinely reported his abuse to police. Public feeling was one of shock and also alarm at how fatal domestic violence could be. The government came under immediate pressure to deal with domestic violence more effectively and in 2005 the first domestic violence prevalence study was published by Hong Kong's Social Welfare Department (Chan 2005). The study showed that more than 1 in 7 spouses have been battered by their intimate partners at some point in their lives. More than 1 in 5 households have spouses who have been battered by their partners. In Chan's study, Conflict Tactics Scale 2 (CTS2) and Parent-Child Conflict Tactics Scale (CTSPC) were used to estimate the annual and ever prevalence rates of different types of domestic violence. Around 1 in 10 of the interviewed spouses had committed or experienced physical assaults in their intimate relationship, among which about 4% would lead to physical injuries. The same study shows that over 50% of the interviewed spouses had committed or experienced psychological aggressions in the spousal relationship. Around 6% of the interviewed spouses had either committed or experienced sexual coercion. Although the gender difference is not large in Chan's study, partially due to the tendency to produce gender neutral results through CTSs (see Johnson 2006 on debates around gender symmetry and conflict tactics scale), it still shows that women are more likely to be affected by domestic violence (by spouse: male -12%; female −15.7%; p value -0.000*; by respondent: male -14.9%; female −15.3%; p value -0.746). The gender asymmetry in intimate partner violence is further confirmed if we look at the cases reported to the government systems, majorly the police, social services and hospitals. Female victims consistently account for more than 80% of the total number of intimate partner violence cases (it used to be called spousal violence in Hong Kong, and has changed to reflect violence happens between intimate partners who are not spouses of each other) in the past 10 years in Hong Kong (the Central Information System of Battered Spouse and Sexual Violence, Social Welfare Department of HKSAR).
Concerns over service effectiveness and accountability continued to draw the attention of the Hong Kong public as the inquiry into the Tin Shui Wai tragedy unfolded. Repeated calls for help, both recorded and unrecorded, by the migrant mother of two revealed there had been a lack of adherence to the guidelines on the part of social service professionals (Review Panel on Family Services in Tin Shui Wai, 2004). Research also revealed patriarchal values embedded in the Police Force, as well as the discriminatory attitudes of social service professionals towards new immigrants from China (Hong Kong Christian Service 2004;Wu 2004). The strong public demand for effective social intervention in the field of domestic violence (Review Panel on Family Service in Tin Shui Wai 2004) resulted in a series of changes such as revisions to the procedural guidelines for handling cases of domestic violence and child abuse, streamlining crossdepartmental coordination, setting up a central information system and a shorter application period for injunction orders. While procedural and system changes are critical for protecting 'victims' of abuse, these changes also seemed to focus primarily on 'crisis intervention' with little attention to the preventive and supportive work that is meant to be part of Hong Kong's three-pronged approach to tackling domestic violence (Legislative Council Secretariat 2013). A brief analysis of the working principles of the Procedural Guidelines for Handling Cases of Intimate Partner Violence (revised 2011) (Fig. 2) clearly shows the tendency to focus`on crisis intervention.
We can see that the Guidelines for facilitating multidisciplinary collaboration is majorly set around crisis intervention by targeting risk reduction and avoidance of re-victimization. The involvement of victims in assessment and action planning is restricted to 'direct communication'. Once abused women are rehoused, either in private rental housing or public housing, cases are more likely to be terminated for the reason that 'spouse battering elements have subsided'. Abused women who have left the matrimonial home and petitioned for divorce have become the most unattended in Hong Kong's intimate partner violence services. Harmony House, the first shelter in Hong Kong, was aware of the lack of support for abused women who had separated from abusers. They developed 'after shelter services' to take care of the emotional and adaptation needs of women who had left the shelter, including those who return to and permanently left the matrimonial home. These good intentions were not effectively translated into practice however due to the government's austerity measures (Harmony House 2007). While some project-based initiatives for rebuilding self-esteem and self-confidence are identified (Hong Kong Family Welfare Society 2017), government-funded long-term post-separation services for abused women and their children are still virtually absent in Hong Kong. The absence of post-separation services for abused women in Hong Kong reflects the assumption that women's post-separation lives are problem-free; that they are strong enough to handle 'their problems'; and that having left their abusive partners they are survivors rather than victims.
This study is concerned with the post-separation needs of abused women. It aimed to (1) understand the needs of abused women who have left the abusive relationship; (2) co-design a social work practice/service that could meet the identified need(s); and (3) run and evaluate the practice/service developed in this research with all the research participants. The process helped unpacking women's experiences of living with and without the perpetrators of abusea process that has been understood in association with women's conceptions of choice, gain and loss of personal agency, as well as the availability of social and financial support (Ben-Ari et al. 2003;Dobash and Dobash 1979;Kirkwood 1993). Narratives and pictures co-produced with women participants problematize the 'either victim or survivor' dichotomy evident in the design of Hong Kong's intimate partner violence service, as well as elsewhere in the world (see Dunn 2004;Thapar-Björkert et al. 2016). Drawing on these data, this paper engages with the debate about the appropriateness of 'victimhood' or 'survivorhood' in representing abused women's lived experiences. The former often underpins 'safeguarding' work and the latter is used for promoting 'empowerment' practices. This paper seeks an alternative in making sense of abused women's agency and vulnerabilities. By looking at how women cope with post-separation challenges and seek ways of 'doing' daughters, mothers, sisters, activists and citizens differently (Gueta et al. 2016;Katz 2015;Zufferey et al. 2016), we have found that women engage in identity work that 'troubles' the 'either victim or survivor dichotomy'. 'Doing being' is central to the analysis of the identity work carried out by women participants as it links identities to social behaviours and social practices performed by the actors. It sees them as products and constituents of social orders (Sacks 1985). Women's identity work that is captured in this study shows the need for supported autonomy and agency and a relational model (Smart 2007) for working with women who have experienced intimate partner violence.
The Discourse of 'Either Victim or Survivor' and its Limitations Victimhood was extensively employed by feminists in the U.S. for mobilising civil society towards bringing domestic violence to the public agenda (Dunn 2004). The 'battered women movement' in the US started out from public tolerance and silence towards the problem, and the women it aided were 'beaten women, whether at home or on the run, need much and can give little' (Tierney 1982, p. 212). Experiences of abused women in the early years of the movement supported the construction of pure victimhood, where women had low personal agency to resist the violence against them needed external supportfor remediating the problem. This particular form of 'blameless' and 'innocent' victim identity served as a 'politicized collective identity' for mobilizing public resources and brokering public sympathy for raising the profile of this emerging social problem of 'wife battering' (Dunn 2004;Nissim-Sabbat 2009;Thapar-Björkert et al. 2016;Tierney 1982).
Soon after its emergence in mid-1970s and 1980s, the pure victim identity was criticised for being too reductive and failing to capture the complexity of abused women's experiences. Dunn (2004) analysed the victimizing stories told by battered women and the media, and discovered four types of victims: 'precipitating victims', 'ideal victims', 'stigmatized victims', and 'heroic victims'. Leisenring (2006) also discussed the different victim claims made by abused women themselves, consisting of the 'pure victim' claims, victim empowerment framework, the responsibility claims, and victim-survivor claims. These studies have challenged the monopoly of the traditional weak and blameless 'ideal victim/pure victim' identity in understanding the abused women's lived experiences. Some studies carried out in the U.K. and the U.S. also found that the victim identity can cause harms to women's post-separation recovery and self-efficacy as it jarred with the self-perception of abused partners (Donovan and Hester 2010) and contributed to victim mentality (Leisenring 2006). Dunn (2004) further argued that 'victimhood', to be justified in western culture which emphasizes autonomy and agency, inevitably connoted a power differential between the sympathizers and the sympathizees (p.239). In Chinese culture, making victim claims can be stigmatising because Chinese family virtues would require moral women 'not to spread the family shame' (家醜不出外傳) including violence against them. Women can be blamed for breaking and shaming their own and their husband's families. Despite the cultural differences, it is conceivable that women in Hong Kong, the UK and the US can feel reluctant to identify themselves as victims of abuse, or to label their experiences as abuse, because of the potential damage that the victim identity can do to their own self-image (Muehlenhard and Kimes 1999;Leisenring 2006;Donovan and Hester 2010;Brosi and Rolling 2010).
To address the stigmatizing dimension of 'victim' identity, a discourse of survivorhood has been created that emphasizes women's resistance, their ability to cope, and the choices they made for surviving the abuse (Hydén 1999;Herbert et al. 1991;Davis, 2002;Johnson 1992). The focus of intimate partner violence studies has also shifted from 'staying, leaving, and returning' to 'resisting, coping, and surviving' (Leisenring 2006;Humphreys 2000). The recognition of strengths and resistance is also a sign of moving away from victimhood (Glumbíková and Gojová 2020) and the start of post-abuse recovery (Brosi and Rolling 2010). Women's strengths and capabilities are recognized and made explicit through the construction of survivorhood. Their positional knowledge has also gained more appreciation and recognition in policy and service design (Mullender and Hague 2005;Beresford and Croft 2000).
When the rise of survivorhood marginalises victimhood, it can also become stigmatizing however (Sweet 2019). While victimhood might restrain women from articulating their experiences and personhood differently from being blamelessly weak and powerless, it can enable women to explicate their needs and garner sympathy and assistance (Leisenring 2006). This enabling and restraining property of victimhood creates a paradox where survivorhood and survivor-based practice are developed as a counter narrative or practice against victimhood. Women who express their needs, sufferings and life challenges are considered inferior to the strong and fearless 'survivors'. They are considered viewed as 'lingering' to the relationship with the abusers and personally not willing to end the victimization. As such, the monopolisation of survivorhood is as problematic as the monopolisation of victimhood, revealing limitations of the 'either victim or survivor' dichotomy.

Women's Experiences of Leaving as a Critique of the Dichotomy
Women's experiences of 'leaving' the abusive partners further highlight the problematic nature of the 'either victim or survivor' dichotomy. Leaving is not a clear cut process of separation marked by moving out or divorce, but a back and forth process that involves loops of staying-leaving-returning (Kirkwood 1993). In cases of 'successful leaving', each loop of staying-leaving-returning is carried out on the basis of the strengths gained during previous loops. Leaving is therefore a continuous process of intertwined choices and entrapment, and resisting, coping and subordinating. Even for abused women, who 'successfully' leave the abusive partners, they suffer extensively in the help-seeking process such as through the bureaucratic welfare systems and insensitive police responses (Lutenbacher et al. 2003;Wolf et al. 2003). The disinterest of helpers and difficulty meeting their financial, housing and emotional needs are also factors that contribute to women's feelings of re-victimization, decisions to stop fleeing, and returning to the relationship (Wuest and Merritt-Gray 1999, p. 112). While being punched, slapped, terrorized with weapons, stalked and humiliated in public may not be physically fatal, the history of these 'traumas' remain influential to their lives and wellbeing after separation. This is shown in psychological studies of post-traumatic growth, which identify women's ongoing suffering from past abuse as posttraumatic stress (Joseph and Linley 2008).
With a better understanding of women's experiences of staying, leaving and returning to the abusive partners, 'victim-survivor' or 'victim/survivor' is now seen more often in the literature as a linguistic response to the problematic 'either victim or survivor' dichotomy and as an acknowledgement of the complexity of women's experiences of abuse. While the hyphen space employed in the existing literature helps acknowledge the uncertainty or the hybridity of abused women's experiences, we still know little about how women negotiate their identities around the dominant discourses of victimhood and survivorhood. For women who have left the perpetrators and become labelled as 'survivors' in Hong Kong, how do they negotiate their identities for expressing their needs and sufferings, and to re-discover and reappropriate their strengths? How do negotiated identities help women who have experienced intimate partner violence face their post-separation challenges? This paper presents the narratives, textual and pictorial data co-created with women who have left their abusive partners and unpacks women's identity work as negotiated in the context of participatory action research (PAR) project.

Methodology
Cooperative Grounded Inquiry (CGI) (Kong 2016), a new methodology belonging to the larger umbrella of participatory action research (PAR) (Heron and Reason 1997), was developed specifically in this project for fostering collaboration between social work practitioner-researchers and service users to build useful knowledge and services grounded in each other's positional knowledges. The project involved women who have left their abusive partners and it involved their teenage children. Participants sought to make sense of and find ways to meet their post-separation challenges. These experiences suggested that women suffer from the categorical application of 'victim' or 'survivor' identities imposed onto them by social welfare systems, their friends and colleagues. The lack of space for negotiating these identities led to a strong sense of helplessness and (professional and personal) relationship breakdown. This project identified ruptured/troubled identities of abused women in the post-separation stage, and provided space for new identities to emerge and serve as a critical voice for expanding the restrictive 'either victim or survivor' dichotomy (Gueta et al. 2016).

Data Collection, Analysis and Action
This CGI involved a social work practitioner-researcher (the author), women who have left their abusive partners, and their teenage children, in weekly meetings (around 6 h per meeting). These lasted for 6 months and involved designing, delivering and improving post-separation services. We applied the modified reflection-action-reflection cycles (Fig. 2), as part of the CGI approach to make sense of post-separation challenges (propositional knowing) and to formulate and deliver appropriate services/practices for addressing them (practical knowing). The different forms of data, including conversations, observational data and interactive data generated in the problem-solving processes (experiential knowing) were captured by photos, drawings, reflective logs, fieldnotes and videos and audio recordings (presentational knowing) produced individually and collectively. These data were presented back to the inquiry group for reflection and collective analysis (propositional knowing). The analysis of data was aided by constant comparative analysis (Glaser 1978) which is a technique borrowed from Grounded Theory Methodology to help co-inquirers compare and contrast experiences, emotions, behaviours, interactions and attitudes of each other. It allowed us to develop an understanding of the diverse experiences of women in coping with their post-separation challenges. This paper presents only the findings on how women participants' identity work was carried out in this project while details of the methodology and the other findings have been presented elsewhere (Kong 2016;Kong 2018).
CGI demands the cultivation of reflexivity among each coinquirer so that we could identify how the construction of oneself affects the construction of understanding about the 'outside world' and our interactions with 'others'. This variation of reflexivity is named as 'relational reflexivity' (D'Cruz et al. 2007). To achieve this, a reflective session was timetabled in each meeting. Through writing reflective notes and ongoing analysis of personal performances in different social occasions, co-inquirers acquired deeper awareness of themselves in relation to others. As the initiating researcher, I also kept fieldnotes and reflective notes throughout the inquiry. These notes highlighted my dual insider-outsider identity as a 'historically disenthralled sister' that shaped both my self-disclosure and those of others (Author 2015).

Women Participants
The practitioner-researcher recruited the women participants from a local survivors' group. 7 women participants (practitioner-researcher, HL, NF, PF, YY, KW and YT) were officially involved in this inquiry (Table 1). For reasons of safety, we recruited women who had separated with their abusive partner for at least a year and a half and who had no record of physical violence for at least half a year at the point they joined the group. In this project, 'separation' does not reflect the marital status but the physical distance that women participants had with the abusers by living apart from them and being safely rehoused in private/public housing estates. This recruitment strategy helped address the service gaps presented in Hong Kong's domestic violence service (the lack of postseparation services).
All women participants had experienced both physical and psychological abuse for at least 5 years, while most of them, except NF and the researcher-participant, had children aged between 12 and 17 when the inquiry began. Among women participants, one was undergoing divorce proceedings and two were still fighting for custody during the inquiry. The inclusion of both local (n = 2) and mainland Chinese (n = 5) women in this sample also enabled us to delineate the impact of migration on women's search for social recognition and their journey of recovering from the trauma of abuse (see Reconstructing the Survivor Identity). The age difference among women participants, ranging from their late twenties to mid-sixties, created unique group dynamics that resembled Chinese filial piety and respect for the seniors (Kong 2018). Ethical approval was obtained from the University of York before the enquiry commenced.

Findings
Troubled 'Pure Survivor': 'You Cannot Leave Us Uncared' The group meetings at the first phase of the inquiry were filled with complaints about the 'poorly performing' domestic vio-lence (both intimate partner violence and child abuse) service in Hong Kong. The assumption that abused women were 'problem-free' or able to 'stand on their own two feet' after leaving their abusive partners was contested by women participants' experiences in their post-separation lives. All women participants agreed that they had been living with the impacts of intimate partner violence, including poverty, isolation, traumas, sadness, and unfair treatment imposed by abusers and the social care system and family court. As women talked about their own sufferings in the inquiry group, they found these resonated with the experiences of other group members and a sense of solidarity began to develop in the group (Kong et al. 2020). Phrases like 'we need help', 'you cannot leave us uncared' and 'if I can do it myself, then I won't…' were frequently used by women participants to assert their needs for support and care, and to foster empathetic forms of solidarity that helped coordinating between one another for producing change (Banks 2014). These shared experiences of suffering created a strong rupture with the 'pure survivor' identity that had been assumed, reinforced and prompted by the domestic violence service in Hong Kong. The dissonance between the pure survivor discourse and the lived experiences of abused women triggered the reengagement with the 'victim' identity by women themselves in order to explicate their need for support and care by the state and individuals. One of the many sufferings that continued to affect women after their separation with the perpetrators is the long-term physical impact of intimate partner violence, e.g. disposition of bones, headache, dizziness and poor health. Women often talked about their emotional vulnerability, such as feeling depressed, angry or agitated, which did not cease with separation or over time.
'Our sisters [women who have experienced domestic violence] just can't be happy. Every of us were the same. When we had just left the bad guy (abusive partner), we were very unhappy. Even though people around us were celebrating for the Luna New Year we were impervious to the vibrant atmosphere. You just can't be happy.' Said YY. HL added, 'I had exactly the same experience. I met NF for the first time in a Luna New Year celebration. Sisters in the shelter took me there. I hadn't felt thankful for their kindness; instead, I found them annoying and offensive. I thought, "I am now very depressed, why are you so happy when I am so miserable?" You will feel even worse.' Emotional instability, in the form of sudden outbursts of anger and mood fluctuations, were evident in the inquiry meetings. Displays were particularly heightened when abused women were expected to be 'happy' or 'cheered up' by other 'survivors' in the group when their lived experiences did not resonate with these expectations. Women sometimes reacted to these expectations with withdrawal behaviours or microaggression directed towards other women participants, or even their own children.
'You may not know her temperament. She (one of our participants) scolds and yells at me whenever I can't perform according to her expectation. It is very difficult to stand it. It is stressful.' Said PF. 'I was so sad when I heard my daughter repeatedly calling me "useless"! I locked up myself in the toilet and she came over to check if I were good…I was nearly driven mad, so mad that I was scared of beating her up! I burst into tears and ran away from home… That was at night.' Said YT.
Women participants claimed that emotional disturbances were commonly shared by those who have left the abusive partners and that these firmly stood in the way of their 'recovery' (Abrahams 2007). Women participants also came to see their long-term exposure to coercive control had undermined their ability to control anger, especially in the face of people's comments and criticisms. Anger, as a way to create psychological distance, is a common and a relatively safer way to resist the violence and micromanagement exercised by the abusive partners (Ben-Ari et al. 2003). However, when outbursts of anger have become a conditioned response to criticisms it can affect women's abilities to rebuild social networks and intimacy in their post-separation lives. The psychosocial wellbeing of women can also be worsened through financial deprivation, which is a commonly identified consequence of coercive control (Stark 2007;Stark 2013;WHO 2013). The dearth of support and resources continued to undermine women's selfimage through shaming and blaming them for 'failing to protect' their children. For example, KW had been relying on food banks for months and expressed a great sense of remorse for her 'incapability' to provide.
'My son and I have been eating instant noodles and canned food for 2 months already. They were all preserved food, just not healthy for a boy in puberty.' KW said.
By viewing themselves as 'victims' of domestic violence in the post-separation context, women participants found the vocabulary to challenge the notion that 'leaving the abuser can cure all the problems of abused women', and to justify the need for post-separation support and care. By drawing on the linguistic resources of 'victimhood' to garner care and support for the currently under-served separated abused women, 'pure survivor' identity was troubled/problematized.

Reconstructing the Survivor Identity: 'Chungsangje' as a Discourse of Rebirth
Re-appropriating the 'victim identity' amidst the overwhelming expectation of becoming 'survivors' seemed to help women participants identify their needs for care and support. The reflection-action-reflection cycle drove the group to further reflect on what action they could take to address these concerns, and it led subsequently to the service responses developed within the inquiry group. These included 'personal problem-solving conferences', 'emotional support' and 'health boosting exercises' run by women participants to serve other abused women. When the strengths of women participants became visible in the collective problem-solving processes, more participants would recall the 'good old days' when they had lived with confidence, dignity, and pride. Most of these positive moments in life had taken place before migrating to Hong Kong. They were times when their qualifications were recognized, their jobs were secure, and their abilities appreciated. PF, YT and YY recognized how their self-worth was undermined through the process of migrating to Hong Kong. In a conversation (2nd session), This became a reference point for reclaiming the strength women participants had before entering the abusive relationship. Women participants examined the term 'survivor', which, if translated in Chinese, can carry two different meanings -'chungsangje' (重生者) which means someone who has died but returned to life with new strengths; and 'hengchuenje' (倖存者) which means someone who has survived a disastrous experience by sheer luck. The latter does not only carry a negative connotation as one of passiveness and powerlessness, it also hints to the fact that women might not be as lucky next time. Women participants rejected the term 'hengchuenje' to describe themselves and the group, and unanimously committed to the term 'chungsangje' to represent the positive personal qualities and strengths rediscovered by living through the trauma. 'Chungsangje' also captures the 'born-die-reborn' sequence of their lived experiences of going through and, most importantly, breaking away from the abusive partner. The establishment and continuous employment of 'chungsangje' as an organizing concept for actions, such as care and service delivery in the group, helped justify and drive the development of new skills for supporting other women who were still trying to break away from the abusive partners or dealing with the post-separation challenges outside the inquiry group. Chungsangje identity captured the personal and collective strengths that were grounded in women participants' own histories, cultures and experiences, and increased the linguistic stock for accessing and mobilizing these strengths and skills in making plans and devising action. The new identity supported women participants undergo a transition from women being helped to women who were helpers. A need arose in the inquiry group to redistribute responsibilities (and power) among themselves for ensuring each member enjoyed a more equal opportunity to serve and to be served.
'I think it is good that we could start redistributing responsibilities to others (sisters)… in the past I always played the role of organizer… when we were still in the association (a local survivor group), it's always me and SW who did the work and other sisters just came and enjoyed the time. Shopping for groceries was actually a lot of hard work… and they all made excuses not to participate. We should encourage them to participate more next time (when we organize events)' said PF.
In revisiting the positive past of women, the impact of migration on women's sense of entrapment and vulnerability was also revealed. Migration from mainland China to Hong Kong disrupted formal recognition and informal recognition in the lives of women participants. Formal recognition was about having a 'leading role at work', a 'professional qualification', a 'professional role at work' and an 'advanced educational qualification'. Informal recognition meant 'being trusted', 'being appreciated' and 'being included in social networks'. Migration and intimate partner violence aggravated the social disconnect experienced by women participants, depriving them the social relationships and networks in which their strengths and abilities could be recognised and appreciated. The absence of recognition also impacted women participants' self-image, making these more susceptible to destruction and manipulation by abusive partners.
'At the time I left, I still thought he was so right that I was useless. I was never good for anything. I used to truly believe in such description about myself…' I said, echoed by all other women participants. (2 nd session) Thereby, confidence boosting words, such as 'you are great', 'all thanks to you, we can successfully…', 'we will make it through', (clapping) and (thumbs up) (from WhatsApp), became one of the commonest responses to participants' commitment and achievement. Where women failed to obtain formal recognition because of new qualifying criteria (specifically English proficiency in a former British colony), women participants relied heavily on confidence boosting words as a form of psychological compensation. 'Chungsangje' identity recognises the co-existing strengths and vulnerability in women's history, memories and life practices, and creates internal contradictions in their selfunderstanding that foster reflection on how some personal qualities/ways of living had been framed as either vulnerabilities or strengths by the dominant discourses of individualism, survivorhood and victimhood. These reflections led to the recognition of heterogenous understandings and experiences that women participants had in performing their identity as 'chungsangje', including the different ways they 'related to the abuser' and 'related to society'. For those who selfidentified as chungsangje, they felt they were more ready to re-engage with and contribute to the community. Instead of personalized and particularized services, they preferred services that promote the general well-being of people (i.e. health boosting and socializing activities) and those that offer learning opportunities (i.e. community outreach, skill fostering sessions and health knowledge). However, the way a 'chungsangje' related to the abuser also shaped the extent to which they would like to 'relate to the society'. For example, those who maintained a relationship with their abusive partners were reluctant to turn themselves into public figures when advocating for the rights of abused women. YT who left the abuser physically, but not psychologically, retained a desire to reconnect with the abuser. She then wanted to conceal the socially undesirable behaviours of her ex-husband so as to pave a way for possible reconciliation in the future.
YY asked if YT still loved her ex-husband, YT defended her ex-husband and said that their relationship would not have deteriorated if he had never gambled. YT even said she would not have divorced him if he had not initiated it. The 'love', 'desire for reconnection', 'desire for care' etc. were found to be concepts representing YT's way of relating to the ex-husband. These concepts contradicted with 'anger', 'feeling indifferent', 'fear', 'desire for separation' etc. found in many other participants. (Field note, 25 May 2013) For the participant HL, who physically and psychologically left their abusive partner but remained in contact with him as a friend, she would also consider public action inappropriate because it could antagonize their friendship.

Identity Construction as an Ongoing Process
Chungsangje identity as understood and performed differently by different women is an illustration of women's identity work during their post-separation lives. By revisiting the mixed and messy experiences in relation to their different social positions, women participants creatively deployed and re-appropriated symbols embedded in victimhood and survivorhood to seek ways to describe, justify and resource their actions in order to resist coercive control and address its impacts (Fontes 2015). Chungsangje, despite diverse understandings and performances in the group, could still risk disempowering women particularly when some participants came to realize that they were 'not yet' their ideal 'chungsangje'.
The construction of 'chungsangje-becoming' arose through participants' realize that 'from victim to chungsangje' was a process in which they might not be able to constantly fulfil all the qualities of a 'chungsangje'.
'I have known a number of sisters who have been living apart from the abusers for more than ten years. However, they are still suffering…they have not yet gone through the thing. It is not a matter of time, but your psychological state…if you can break through the psychological barriers that inhibit you, it is your success. Success does not necessarily mean one in advocating a policy or making changes in services.' said NF. Replied YT, 'I think I am not there yet.' (18th session) The barriers and problems standing in women participants' way of becoming their ideal chungsangje were identified in the group conversations. For example, 'being unable to get over the experiences of being abused', 'not ready to disclose their history of abuse' and 'fluctuating psychosocial status'. Chungsangje-becoming identity helped women participants understand why they generally felt good about their situation but still suffered from occasional emotional fluctuations, depressive moods and social disengagement.
'Yes! I was just like MM…I would say yes at this moment and say no at the next. I just couldn't understand my fluctuations. Honestly, in these 3 years, I have never been back to Tuen Mun where I used to live with the bad guy.' KW said when we were exploring the persisting influence of abuse experiences on us. (19 th session) Women participants in the group, who were all living apart from their abusers, recognized that physical separation was effective for removing the cause of victimization and for allowing time for recovery and 'rebirth/chungsang (重生)'. Those who had not left physically would be considered by women participants as 'victims' as the cause of victimization was still present. Alternatively, for those who had physically left but psychologically stayed/affected, the group would locate them as 'chungsangje-becoming (重生緊)'. For those who were described by the group as chungsangje, but by themselves as chungsangje-becoming, such as PF and YY, their identity negotiation prevailed throughout the latter half of the inquiry.
'Chungsangje-becoming' was invented by women participants to maintain the strength-based undertone of their identity. At the same time it enabled them to show the need for care and help without returning to victimhood. Chungsangjebecoming perceived themselves as different from 'victims' in terms of financial and psychological stability and the frequency of their vulnerabilities on display. In terms of living conditions, chungsangje-becoming were less unsettled, for example by being permanently housed and financially secure. 'Chungsangje-becoming' therefore focused more on hands-on skills training and relevant policy learning in order to prepare them for helping others. For instance, an emotional support workshop and policy statement writing sessions were held for polishing skills and increasing knowledge in running services for 'victims' and 'chungsangje-becoming'. The creation of 'chungsangje-becoming' in the inquiry group also suggested that rebirth could be successful only when there were supportive empathetic others. These others can support individuals in seeking and performing their identities differently from those shaped by the dominant 'victim or survivor' dichotomy.

Discussion
When abused women sought ways of meeting the postseparation challenges in the inquiry group they also contested the categorical application of victim or survivor identities. Categorical application of either victim or survivor identities clearly failed to capture both women's need for support in their post-separation lives and their eagerness to support other women. Their identity work demonstrated that victim and survivor are not static identities, but clusters of symbolic resources that abused women could draw on for articulating their heterogeneous, messy and dynamic lived experiences and needs. Refuting the categorical application also further problematised the either 'victim or survivor' dichotomy when women participants started to examine the discrepancies between their life challenges. For example, mothering in the context of domestic violence (Radford and Hester 2006;Fauci and Goodman 2019) and poverty, and the identity categories imposed onto them by domestic violence services in Hong Kong. The discrepancy forced women to seek and perform alternative identities that could better reflect their life circumstances and which were more useful for re-organising the human and material resources for solving their problems.
Identity work therefore provides an adaptive function for people in a community (Fowler 2010). It challenges and revises the restrictive narratives/discourses that restrain people from understanding and solve their problems. In a similar vein, the different ways of living out the chungsangje identity in terms of how one relates to the abusive partner and the society, enabled women participants to recognize alternative possibilities of resisting violence and violent husbands other than leaving. These experiences challenge the discourse that privileges 'leaving' among other 'choices' of relating to the abusive partners. In the inquiry, the experiences of HL and YT did not conform to the formula stories of 'pure victim' and 'villain abuser' (Loseke 2001) because both chose to maintain relationships with their ex-husbands, in the form of either friendship or romance, while refusing to be abused again. Social expectations that abused women display a consistent identity, either victim or survivor, could impede both their capacity to make sense of their lived experiences and their agency for adapting or solving their everyday life problems. Narratives of abused women collected in this study clearly contested the either 'victim or survivor' dichotomy reproduced and reinforced by Hong Kong's domestic violence service framework. The identity of 'chungsangje-becoming' represents the 'victim and survivor' experiences of abused women as well as their aspirations to leave victimhood at the post-separation stage.

Safeguarding and Empowerment Revisited: Towards a Relational Model of Women Support
Safeguarding work is central to protecting 'victims' of domestic violence. It enables 'victims who were previously ignored, belittled, and blamed [to become] assisted, advised, advocated for, sheltered, and supported' (Mcdermott andGarofalo 2004: 1246). Despite the aim of victim safeguarding to remove the blame women feel for the abuse directed towards them, safeguarding also rests on the concept of 'innocent' or 'blameless' victims that are unable to resist or stop the violence against them. These underlying assumptions of pure victimhood suggest the potential for disempowerment, particularly when safeguarding work is 1) brought closer to risk management and risk reduction (Donovan 2013;Robbins et al. 2014) because of the increasingly managerial culture in social care services (see also Fig.1) and 2) based on the individualistic understanding of women's wellbeing independent of the wellbeing of their children and their mother-child relationship (Kong 2018). These disempowering elements are evidenced in the procedural guidelines for handling cases of intimate partner violence in Hong Kong (see fig.1) which puts riskreduction at the centre of safeguarding work and tends to stop the support services once the 'violence subsides'. The Guidelines clearly reveal the government's expectation that women should be able to deal with their problems once the source of oppression (the perpetrators) is removed from their lives. This expectation on women explains the lack of government support available women who have left their abusive partners, even though long-term physical, psychological, social and financial impacts are well documented in WHO's multi-country study (2013). This approach to adult safeguarding is in contradiction with women participants' experiences as 'chungsangje' and 'chungsangje-becoming': women's resistance, strengths, vulnerabilities and needs for help co-exist and shift over time in response to the emerging challenges in their post-separation lives. Identity work is therefore a continuous performance of self that helps justifying and reorganising relationships and life practices so that women can garner the support/help needed from others and utilise personal and public resources to solve their problems.
As an alternative to safeguarding, women's empowerment has been considered an important preventive and supportive model for tackling intimate partner violence. It aims at addressing the root causes of gender inequality and the physical, psychological, social and financial impact that violence could have on women (Tiwari et al. 2005;Tiwari et al. 2012;Hester, 2013). While the UK has been through waves of feminist movements to challenge patriarchy, the women's movement in Hong Kong began by rejecting the western import of gender equality that could be 'threatening [to] the integrity of the family or trampling on men' (p.104). These early years of the women's movement shaped empowerment practices, making women in Hong Kong focus on personal skill enhancement and individual capacity building (Ibid). This focus aligns very well with other empowerment models practised with Hong Kong Chinese abused pregnant women (Tiwari et al. 2005) and community-dwelling abused women (Tiwari et al. 2012). These seek to 'engage with individual abused women to empower them and link them to community services, with ongoing support, informal counselling, or both as required' (Ibid:537). The individualistic undertone of empowerment practices in Hong Kong persists and is reconfirmed in the Strategies and Measures in Tackling Domestic Violence in Selected Places published by the Hong Kong Legislative Council Secretariat (Lee 2008:35): 'Empowerment and a victim-centred approachservices must ensure that victims identify and express their needs and make decisions in a supportive and nonjudgemental environment, that victims are treated with dignity, respect and sensitivity; and promote serviceuser involvement in the development and delivery of the service' The combination of empowerment with victim-centred approach, adopted by the HKSAR government, reflects the assumption that women are autonomous individuals who can make rational decisions, act for their best interest and be themselves when they are not coerced by another person.
The Cartesian model of self that underpins the empowerment models in Hong Kong, however, contradicts abused women's experiences. Women make sense of their reality and life preferences in the context of relationships. For example, in this project, women's shifting relationships with their children, the ex-partners and other women participants create and limit the space in which women's sufferings and strengths can be recognised, validated and acted upon. To re-create a space for recognising marginalised stories and hence identities, women participants undertook identity work that problematised the categorical application of either victim or survivor onto them. They creatively constructed culturally more appropriate identities that enabled them to express both strengths and vulnerabilities. The resulting identities would often offer a revised socio-relational space in which women's marginalised experiences could be recognised, validated and acted upon as resources for solving emerging life challenges. Identity work, therefore, is not the project of an individual seeking true/authentic self but a 'relational project' (Combs and Freedman 2016) of reconstructing relationships. It can be among women themselves and between women and significant others, such as their children and abusive ex-partners. The relational approach emerging from the project is about acknowledging the fluidity and multiplicity of identities performed by abused women at different times and space. It is to see identity work a crucial practice for bringing women's marginalised stories to the surface and re-organising social relationships in ways to address power differences, such as that between sympathizees and sympathizers, dependents and independents and victims and survivors. Identity work can be empowering only when we see it as a relational project, and when we cultivate supportive, empathetic and egalitarian relationships (see Kong 2018) for it to take place. This study therefore suggests a new service direction for supporting women who have left their abusive partners. That is, investing in and cultivating social relationships that enable women to revisit, reappraise and re-appropriate their experiences to cope with their challenging post-separation lives and to live out their preferred identities as a person beyond victim or survivor (Nissim-Sabat 2009).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.