normalising power and engaged narrative methodology: refugee women, the forgotten category in the public discourse

Since the beginning of the 21st century, the discourse of othering of non-Western migrants has been growing in many European societies. And since 2015, refugees have become a quite visible component in this discourse. Although, for decades, the dominant image of refugees has been constructed as people ‘at risk’, new competing images of refugee men ‘as risk’ have recently gained ground. For refugee women, however, the image of being victims and ‘at risk’ still prevails. This shows a strong underlying gendered logic of feminine vulnerability and masculine threat. In this article, I show how these images are situated within the dominant Dutch discourse of migration with taken-for-granted taxonomies of the self and the other. Specific in this normalised discourse for refugee women is that their agency is either ignored or their possible position as activists is not acknowledged to exist. Using examples from two studies in which my research team engaged with the method of narrative engaged research, I show the importance of this particular narrative method in unsettling the normalising power of othering. The theoretical argument of this article engages with ongoing discussions on power and agency. It argues that, when the power of exclusion works through repetition and is manifested in the daily normalisation of actions, agency needs to provide an alternative in the same fluid manner. Narratives in dialogue provide an illuminating angle for discussing this specific kind of agency, as I will show through some examples from research.


introduction
Although the condition of refugees has been a subject of research for decades, the 2015 'refugee crisis' created a new impetus for studying possibilities and challenges regarding refugees' short-and longterm positions in their countries of settlement. Many of these studies have focused on the dire conditions of refugee camps both globally and inside the borders of Western countries. These centres limit the mobility of asylum seekers once they are inside Western societies, resulting in asylum seekers' marginalisation and lack of control over their daily lives (O'Reilly, 2018). Bülent Diken (2004, p. 1) refers to these centres as 'non-places' 'in which they [asylum seekers] lead a life in a permanent state of exception and [in] detention centres into which they are forced without trial'. In the Netherlands, some researchers have even used Erving Goffman's (2007Goffman's ( [1961) concept of 'total institution' to describe the condition of these centres (Geuijen, 1998;Larruina and Ghorashi, 2016). Thus, restrictive policies and the poor living conditions of asylum seekers in the country of arrival have been a major focus of most of the academic literature (e.g. Van Dijk et al., 2001;Kohlmann, 2003;Dupont et al., 2005). But it was only after decades of its existence that these phenomena were met with public outrage at a global level and in the Netherlands.
Attention on the human and societal disaster that had taken place for years within the Dutch asylum seeker centres began to grow over the past decade. The lengthy asylum procedures and the marginalisation of asylum seekers for years within often secluded centres were argued to have strong negative impacts on their agency and their ability to integrate into Dutch society at later stages (Ghorashi, 2005;Ten Holder, 2012;ACVZ, 2013;Engberson et al., 2015;Ponzoni et al., 2017). Therefore, (local) policy approaches have recently become more focused on asylum seekers' and refugees' early inclusion and participation in society. Many different actors, including governmental organisations, municipalities, NGOs and bottom-up organisations/initiatives, started experimenting with innovative approaches for the inclusion and participation of refugees (Boersma et al., 2018;Larruina, Boersma and Ponzoni, 2019;Rast et al., 2020). In spite of this growing attention to the early integration of refugees, a real inclusion in practice will not take place if the discursive foundation of dominant images of refugees in the past and present is not challenged. Through presenting findings from two earlier projects with first-generation refugee women, this article critically discusses the implications of dominant discourses and policies related to refugees in the past and present to enable reflection for the future. In addition, I show the importance of narrative engaged research as a method for de-normalising the taken-for-granted taxonomies of the self and the refugee other that are informed by the Dutch discourse on migration.
The narrative approach used in the earlier studies discussed in this article has a biographical angle that enabled my research team to approach individual positioning within discursive spaces in a dynamic and multilayered manner. The narratives gave us the opportunity to discover what was important to the narrators about their lives, the people and places they identified with and the meanings they attached to them (Kohler-Riessman, 2008). Many feminist scholars have shown the importance of such methods for marginalised groups, such as women, because it allows them time and space to express their feelings in the form of a dialogue (Ardener, 1975;Anderson and Jack, 1991). I have previously argued for the importance of such methods for refugees specifically and even more so for refugee women, who are not used to taking up space and at times are forced into silence because of their past experiences (Ghorashi, 2008). Furthermore, in the studies presented here, this narrative approach is combined with a specific form of participatory research that I refer to as engaged research. The inspiration for this type of research came from a co-creative form of research in which research subjects play an active role through engagement at the boundaries of theory and practice to enable 'a critical understanding of historical and political contexts within which participants act' (Freire, cited in Greenwood and Levin, 2007, p. 253). The methodological assumption is that by connecting academic, professional and personal knowledge, various levels of reflection will emerge to unsettle the normalised images of the self and the other. By including multiple perspectives and positionalities in the process, engaged research aims to unravel 'the power effects of particular discursive formations' (Ybema et al., 2009, p. 315). Connecting the multiplicity of non-academic positions through dialogical narratives (or narratives in deliberation) with academic knowledge enabled us to co-create knowledge within which the processes of data collection and analysis were partially merged. In the first study, which focused on refugees' inclusion, we engaged with the narratives of refugees and refugee-oriented professionals. In the second study, the focus was on the life stories of women with different backgrounds. In both studies, we started by sharing narrations, and we later analysed the narratives together with the participants. In the second study, however, we made more space for participants' engagement in the process of analysis of the narratives (both during and after the interactive sessions). the normalising power of the Dutch discourse Harriet Gray and Anja K. Franck (2019) identify several shifts in British media concerning the representation of refugees after 2015. Refugees were initially seen as vulnerable, a representation shaped by the shocking picture of Alan Kurdi's body on a shore in Turkey. This representation was in line with images of refugees as victims and receivers of help in the 1980s and 1990s (Harrell-Bond, 1999;Ghorashi, 2005). However, this image of refugees 'at risk' soon changed to one of refugees 'as risk', with a strong underlying gendered and racialised logic (Gray and Franck, 2019, p. 276). This logic refers to an intertwined presence of feminine vulnerability and masculine threat. In these images, refugee men are considered aggressive and potential dangers to society, while refugee women's agency is erased (ibid., p. 270). A strong manifestation of this image was the representation of refugee men as potential and actual perpetrators in German and Dutch media after the 2015 New Year's Eve sexual attacks on women in several German public spaces (Brenner and Ohlendorf, 2016;Vieten, 2018). Furthermore, Ulrike Vieten and Fiona Murphy (2019) argue for the dominance of the intersection of gender blindness and general mistrust towards asylum seekers and refugees in the case of Northern Ireland. In line with this European trend, in the Netherlands, refugee women are either entirely absent from discussions and policies on migrants and refugees, or they are suffering from subtle but consistent forms of exclusion based on negative images of not being competent or emancipated enough (for the same discussion on migrant women, see Ghorashi, 2010a).
The images of asylum seekers and refugees in the Netherlands need to be situated within the broader Dutch discourse on migration. Since the turn of the century, there has been an undeniable escalation of the negative othering discourse concerning migrants and refugees from Islamic countries, which many connect to the events of 9/11 and the subsequent violence in various European cities. Yet the othering discourse of migrants is not new; it has been present in the West for decades. Anshuman Prasad and Pushkala Prasad (2002, p. 61) argued that the contemporary discourse of othering is informed by 'the social and cultural construction of a fundamental ontological distinction between "the West" and "the non-West"'. The ways that migrants of colour are approached in the present, the authors argued, can be traced to the colonial legacy. Within the constructed binaries of difference, the ethnic other has been considered not only as absolutely different from but also inferior to the norm of the ethnic Western self.
The discourse on migrants in the Netherlands, however, has probably been most dominated by the arrival of 'guest workers' in the late 1950s and of refugees in the 1980s. 'Guest workers' were invited when cheap labour was needed. They had little education and came from the most traditional parts of their homelands, thus strengthening the superior feelings of the 'Dutch' versus the 'migrant other'. The notion of 'guest' added another strong connotation to the migration discourse, as it implied a temporary stay rather than permanent settlement, even though the status of the guest workers gained a permanent character in the 1980s (Lutz, 1997). Likewise, 'refugees' also began their stay in the Netherlands under the notion of impermanence. The humanitarian aspect of the condition of refugees before 2015 had identified them as 'victims', which in turn strengthened the image of their incompetence (Ghorashi, 2005). After 2015, there emerged a competing image of refugee men in particular as potential dangers to society (Brenner and Ohlendorf, 2016). What connects these images is the view of refugees as either helpless or violent, so their agency (if any) is solely ascribed in a negative sense. In addition, the underlying assumption in both images is the need for intervention to either help or constrain refugees. In both cases, refugees' talents, ambitions and willingness to contribute to their new societies are completely marginalised. In the case of refugee women, however, the image of victimhood and of being in need of help still prevails. I have previously shown how this notion of assistance creates a hierarchical relationship between the giver and the receiver, leading to the expectation that refugees should be 'grateful' (Ghorashi, 2014b). This relationship has added to the sense of Dutch superiority in the migration discourse which, together with a persistent image of migrants and refugees as temporary settlers, is an important component of the Dutch migration discourse.
The normalising power of othering is present in the prominence of hegemonic norms constituted by gendered, racial and cultural hierarchies of difference that reproduce structures of inequality in everyday practices (Young, 2007). In these normalised practices, members of non-privileged groups are depicted as not only absolutely different but also inferior and, increasingly, dangerous and unwanted. What is specific about refugee women in the Dutch normalised otherings discourse is how their lack of visibility in the dominant discourse adds an extra layer regarding their normalisation: their agency, needs and challenges are completely overlooked. Ada Ruis (2019), for example, shows how a highly educated refugee woman with a PhD and a background as assistant professor at a Syrian university becomes completely dependent on her husband after her arrival as a refugee to the Netherlands because of the male-oriented Dutch refugee policy (see also Razenberg, Kahmann and de Gruijter, 2018). Short-term policies focused solely on refugee men's integration trajectory contribute to women's marginalisation as being responsible for the household. Female refugees as activists are absolutely absent in the discourses and policies concerning refugees. In the general, normalised image of refugees as 'people out of place' (Malkki, 1992)-people living inside the nation who are merely tolerated as long as they behave and who are never fully accepted as belongingrefugee women are either forgotten or, as activists, are not acknowledged to exist.
The consequence of these images for refugee women is that they are often blamed for having a lack of emancipation, assertiveness or the required competencies to find proper jobs or other positions in the Netherlands. In an earlier study (Ghorashi and van Tilburg, 2006) on the narratives of highly educated first-generation refugee women in the Netherlands, my team found that a fixation on refugee women's shortcomings (for example, their accent in Dutch) overshadowed their achievements (earning a Dutch university degree as a refugee with very limited resources and a relatively older age), making it impossible for them to find a proper job. To explain this fixation on language perfection, we showed the prominence of the so-called 'deficit approach' in organisations (Glastra, 1999). In this approach, migrants' assumed deficiencies (regarding language, education or culturally expected behaviour) were considered the primary reason for their exclusion. To contextualise this fixation on refugees' shortcomings instead of their talents and investments, Evangelina Holvino and Annette Kamp's (2009, p. 397) comparison between how the United States and northern Europe implement the concept of diversity is revealing. These authors show that when the more talent-and business-driven approach of diversity in the United States travels to the northern European welfare states (including the Netherlands), its focus changes towards a moral approach of needing to help weak groups. Diversity loses it individual, talent-oriented focus (becoming something for ethnic-minority groups rather than individuals) as well as its businesscase rationale (being not about profit but about moral obligation). This comparison is important mainly because it shows the contextual embeddedness of concepts such as diversity when they travel. Yet, it is essential to keep in mind that both approaches to diversity have their shortcomings. The business-case approach ignores social justice arguments and lacks a commitment to overturning structural inequalities. And the moral approach within a welfare context, as I will argue in the next section, does not address structural power imbalances either, because it mainly focuses on assimilating 'weak minorities' into the mainstream without questioning the status quo.
The prominence of this line of thinking, earlier referred to as the deficit approach, is strongly situated within the founding elements of the Dutch welfare state (Ghorashi et al., 2015). The welfare state developed with an increasing tendency towards the principle of equality, which resulted in discontent about existing inequalities (Lucassen, 2006). The essence of the welfare state became caring for disadvantaged groups and changing their status. This process had two unintended consequences: 1) it led to a fixation on certain groups as weak in policy and debate, and 2) it resulted in a strong tendency to assimilate difference, to help the other become equal. This approach of equality as sameness differs from other equality approaches that include difference. Thus, despite the welfare state's positive effects on increasing individual autonomy and on the struggle against social division, it has also been a breeding ground for categorical thinking about migrants and refugees as weak groups in socially disadvantaged positions (Ghorashi, 2010b). The societal embeddedness of the deficit approach makes it an influential discourse that reaches beyond the organisational context. Although the dominance of the deficit approach is applicable to a broader range of migrants than just refugee women, what is particular about the latter is that the erased agency of refugee women in the dominant discourse, as presented earlier, makes the impact of such an approach even greater for them.
To unsettle the normalising power of this taken-for-granted othering discourse with the deficit angle, one needs distance from the hierarchical orders (fixed categories of self and other), which are contextspecific. To understand the specificities of the process of othering, we need to unravel its contextual particularities explained above. In this article, I will discuss two studies in which we have engaged with first-generation refugee women using the earlier explained method of narrative engaged research. I will show the importance of this particular narrative research in de-normalising the taken-for-granted taxonomies of the self and the other informed by the Dutch discourse. The theoretical argument of this article engages with ongoing discussions on power and agency. It argues that, when the power of exclusion works through repetition and is manifested in the daily normalisation of actions, agency needs to provide an alternative in the same fluid manner. As I will argue below, dialogical narratives provide an illuminating angle for discussing this specific kind of agency.

normalising power and academic engagement
Despite the visibility of a variety of othering practices in Europe and beyond, the immense power of these practices lies in their invisibility, their normalising capacity and their ability to become part of the taken-for-granted practices of everyday life. In late modern societies, power is more present in invisible forms of normalisation than in visible forms of domination (Bauman, 2000). This kind of power is neither coercive nor tangible but works through the routinisation and normalisation of everyday practices. In his book Discipline and Punish, Michel Foucault (1977) elaborates on the concept of normalisation through his description of disciplinary power. Disciplinary power is not about domination or resistance but about routinisation and regulation of everyday practice. In this characterisation, power is no longer a manipulable source but a web of relations capturing advantaged and disadvantaged alike (Hardy and Leiba-O'Sullivan, 1998). Power, in this sense, is located in the discourses and processes that are taken for granted and thus reproduced. When power is seen as domination, assuming that there are dominators and dominated parties, strategies to resist its impact seem more tangible and observable. A normalised discursive approach of power implies a different approach of resistance or agency. What are the bases of resistance when power is everywhere? How can one still discern power structures? To say it differently, what is left of agency? This point has been raised mainly by feminist scholars (Ghorashi and Sabelis, 2013). Some scholars argue that, 'because there are competing discourses, socialization into any one discourse is never complete, and resistance to specific discursive regimes is thus possible' (Foucault, cited in Clarke et al., 2009, p. 325). Although the normalising power of discourse lies particularly in its tacit impact regarding positioning through the often taken-for-granted disciplining of individuals' actions and interactions, its impact is not absolute since individual positioning involves 'a number of intersecting discourses' (Hall, 1991, p. 10).
Dialogical narratives (or narratives in deliberation) have the potential to challenge individual discursive positionings and, by doing so, to create a reflection of positions that are often taken for granted. 'Strong reflexivity' (Harding, 1993, pp. 70-71) resulting from intersubjective negotiations of contrasting positionalities provides a particularly interesting angle to the notion of agency (see also Ghorashi and Ponzoni, 2014;Ghorashi, 2017). Strong reflexivity is about seeing individuals (including the researcher) in relation to each other from their situatedness in particular histories and communities. This means a 'critical praxis' and engagement with the implications of the intersection of multiple locations and positionalities (Anthias, 2006). Intersubjective dialogical narratives enable contrasting positionalities that in turn could unravel 'the power effects of particular discursive formations' (Ybema et al., 2009, p. 315). These so-called 'positioning practices' (see Alvesson, Hardy and Harley, 2008) create a potential to defer/differ from the dominant hierarchical categorisations of self and other. Différance (to differ without having the self as the norm), as proposed by Jacques Derrida (cited in Hall, 1990, p. 229), serves as a marker that upsets normalised positionalities within discursive spaces. This specific form of agency refers to the 'capacity [of individuals] to be reflexive about their situation-their "discursive consciousness"-and to act on it to "make a difference"' (Zanoni andJanssens, 2007, p. 1376).
In the studies presented below, an engaged form of narrative methodology was adapted to both investigate and unsettle the dominant discourse's effects on the daily realities of first-generation refugees in relation to relevant stakeholders (first study) or in engagement with women with other backgrounds (migrant, black and native Dutch women in the second study). By embracing the logic of strong reflexivity (Harding, 1993, pp. 70-71), these projects aimed to provide a ground for creating reflective zones within discursive spaces to challenge the normalised images of othering that are reproduced. The approach of engaged research in these projects departs from 'an epistemology that challenges traditional notions of scholarly expertise, knowledge as objective and knowledge as decontextualized' (Kajner, 2013, p. 14). It considers academic engagement in a more critical, transformative sense (see also Bhattacharyya and Murji, 2013;Essed, 2013). Inspired by feminist epistemology, I argue that a combination of engaged and narrative methodologies has a strong capacity to approach the comprehensive reality of individuals' lives, which are far more layered than what categories can represent. The stories shared in these studies contradict a singular, essentialised and static presentation of the other as deviant from the self and thus provide access to the multiple, dynamic and in-between images of both the self and the other, creating possibilities for connections. Even in cases when difference had the potential to create distance, the sharing of stories provided necessary sources of identification based on shared experiences. The balancing act of sameness and difference that was facilitated through sharing stories often created unusual connections through confusion, surprise, imagination and self-reflection. This is in line with Zygmunt Bauman's (2000) plea for engaged sociology. Bauman calls for sociology to come as close as possible to yet-hidden human possibilities, which is a call to war against the obvious and self-evident of our time. The task of critical sociology is to question the normalising power of exclusion at work through engagement, knowing that engagement means also becoming complicit in the processes we try to criticise and change. For Bauman, there is no choice between engagement or being neutral, since academia that wants to be influential needs to be engaged. As early as 1959, C. Wright Mills (2000Mills ( [1959, p. 6) argued for sociology to live up to its promise and moral imperative as a social analysis that is of direct relevance to urgent public issues and insistent human troubles. He called on sociologists to use what he referred to as 'the Sociological Imagination' that 'enables us to grasp history and biography and the relations between the two within society'. The sociological imagination is then a quality of mind needed to connect everyday personal realities with larger social realities (ibid., p. 15); in a world overloaded with information, Mills argues, a sociology that only provides information is unable to perform this challenging task. Narratives and sociological imagination are needed to make a difference in the quality of the human condition (Jacobsen and Tester, 2014, p. 11). In this view, the academic and intellectual task of social scientists will be to engage with the narratives of uneasiness, indifference and exclusion by connecting them to historical and societal contexts.

identities, discursive positioning and narratives of change
As argued above, images of refugees are socially constructed and context specific. This means that all individuals are positioned within the discursive spaces they are part of and their positioning in the discourse is shaped and reshaped through a constant negotiation of their self-image and their attributed image from others (Wekker, 1998). Ybema et al. (2009, p. 301) describe this dual and processual character of identity as 'the discursive articulation of an ongoing iteration between social and self-definition'. For groups such as refugees, with a fixed and somewhat negative attributed image, the space for negotiating one's identity can be quite limited. This limited space, however, is not always experienced on the conscious level because of the subtle and gradual workings of the normalising power of the dominant discourses. In the following two examples, I show how engaged narrative research can be helpful in unsettling the normalising power of certain internalised attributed images and, by doing so, can create space for various forms of agency.

reviving individual agency
The first example involves an engaged research project focused on the inclusion of refugees, in which various stakeholders (refugees, policymakers, HR managers, members of NGOs working with refugees) were brought together to debate issues of diversity, power and inclusion (Ghorashi and Ponzoni, 2014). The research included refugees engaged in searching for work and connected them to Dutch professionals engaged with diversity issues. We assumed that, through narratives, the contrasting discursive positionings of participants with different backgrounds would create alternative spaces for reflection. In line with the epistemological choice of strong reflexivity mentioned earlier, the aim of this project was to facilitate opposing mirrors of reality through narratives and by doing so make the 'unreflective discursive positioning' reflective. Phil Johnson and Joanne Duberley (cited in Ybema et al., 2009, p. 315) refer to this as the deconstructive form of reflexivity, 'where the focus is on offering an alternative view of the same reality'. By including multiple voices in the process, we aimed to unravel 'positioning practices' informed by 'particular discursive formations' (Alvesson, Hardy and Harley, 2008). Through contrasting the various positionings in relation to certain discourses, the research enabled participants (some more than others) to go beyond what they considered common sense, which created spaces for agency.
The most profound example in this project was that of Sarah. Sarah was a participant with a refugee background who participated in the mixed group (including refugees and professionals) that was tasked with preparing stories on emancipation and identity for the plenary session. What follows is an excerpt from the field notes of the first meeting.
When Sarah, one of the refugee women, is asked to tell the group something about herself, her answer is: 'I don't think I have such a story'. After encouragement from the group to think about which aspects of her narrative she would consider as powerful or what would represent manifestations of power in her life, she remains silent. 'I don't know', she confusedly answers after a while. The group faces the unexpected inability of all three refugee women to tell their stories. At first, this brings perplexity and confusion to the group. […] The puzzlement and even pain that the question 'What is your strength?' seems to arouse in the refugee women suggests that there is something more at stake. […] Cautiously, the women become more and more engaged in reclaiming their stories. Sarah ends up telling an astonishing story of herself as a young woman fighting for her freedom and that of other women in an oppressive, male-ruled environment in Eritrea, eventually joining the armed fight for the freedom of her country, leaving her family, social position and daily certainties behind.
How can one account for the fact that refugee women who had once been political activists-who could recall a past in which they fought for ideals such as women's emancipation and the freedom of their country, who at one time believed in ideals and in the potential of their actions and personal sacrifices to bring about significant change-appeared incapable of narrating their life stories? This question together with Sarah's story was presented in the plenary session and received a great deal of attention throughout. Sarah shared that, during her stay in the Netherlands, she had only heard negative responses to all of her questions: 'No, you are not good enough'; 'No, your language needs to be improved'; 'No, you do not have the proper papers'. Several years of constant repetition of these words had caused Sarah to lose her self-confidence and, most importantly, to lose her story. We could not think of any stronger case to show how the power of discourse relates to one's positioning. This case helped us to identify the normalised sources of exclusion (the structures behind all the 'Nos') and to discover the turning point from silence towards a surge of agency. This turning point became visible when Sarah recovered her self-definition as a political activist and a 'fighter'. This transformative narrative served as content for the involved stakeholders' ongoing reflections on the normalised images and practices regarding refugees in organisations (earlier mentioned deficit approach), in policy (weak groups who need help) and in daily interactions (being either weak or a threat to society). For example, one of the HR managers said that he often gives similar feedback to people whose language is not perfect. He said he had been doing that with the best of intentions, hoping that they would improve their skills. But he had never before realised that it could have such a negative impact. One of the policymakers commented that she also had this image of refugees and migrants as people in need of help and had thought that policies focused on this would be a solution. But after hearing Sarah's story, she realised that the policies actually led to a fixation on imperfections. Thus, through this process, the taken-for-granted images and practices became visible and questioned.
What happened was much more painful than we could have imagined, yet the change of attitude we perceived in Sarah and in other participants seemed almost magical. The agency that palpably broke free in Sarah during subsequent meetings, because of her recovered self-definition as a political activist and a 'fighter', was one of the most overwhelming impacts of our whole project. By highlighting Sarah's discursive positioning in the group, a mirror was created that allowed the native Dutch participants to see how different minority positions could be compared to their own positions as part of a majority group. The switch in perspective helped them to understand the experiences of those constantly addressed as passive and in need of help, as well as the painful consequences of forgetting self-narratives.

polyphonic agency
The second example involves an intensive life history project wherein a colleague and I invited a group of women with a variety of backgrounds (refugee women, first-and second-generation migrant women and native Dutch women) to come together to share and write their life stories in conversation with each other (Ghorashi and Brinkgreve, 2010). The project took two years and resulted in a published collection of stories from the fifteen women participants. Our ambition was to create space for polyphonic writing, assuming that '[m]ulti-voiced polyphony demands that various voices or perspectives have agency' (Letiche, 2010, p. 262). The achievements and the challenges that arose from this experiment made three levels of polyphonic agency possible: 1) producing relational multiple narratives, 2) sharing the power of authorship and 3) providing agency and alternative voices to counter the dominant discourse (Ghorashi, 2014a).
The project aimed to produce writings based on the narratives of women with various backgrounds. These writings, developed through intensive writing sessions, would be about their positionings and sense of belonging. Choosing this particular form meant creating space for sharing and constructing the narratives in a relational manner. Yet this methodology also meant giving away control of the project as a researcher, at least partially, and being willing to go with the flow of the process as much as possible. This included many ups and downs in the research process, clashes of expectations (both between the initiators and with the participants), tension between willingness and opportunities to share and, most of all, an intensive time and financial investment during a period when there is less and less space for this kind of intensive research within academia.
A central assumption in this project was that, in order to counter the (negative) impact of dominant discourses on one's positioning, it is essential to create spaces in the margins for producing narratives of identity. Maddy Janssens and Chris Steyaert's (2001) condition of 'stepping back' as a group meant stepping into the margins of discursive power to create space for voices from the position of difference, rather than conforming to the dominant norm (in terms of either compliance or resistance). The main condition for this space was that it should be safe, and safety was provided through 'stepping aside'. In the context of recent criticisms about the concept of safe spaces, some communities have opted for 'daring spaces' instead. It takes some bravery to make a space as empty as possible from judgments, thereby enabling participants to share their stories in freedom. In this context, participants are not judged because of their backgrounds or the choices they have made. Stepping aside meant creating space and having patience for listening and producing stories close to one's own experiences instead of producing opinions that are often informed through unreflective discursive positioning. This safe yet daring space in the margins enabled the production of counter narratives to the dominant discourses of hierarchical categorisations. The experience of African American third-wave feminists served as an important example for us. We were inspired by their use of a variety of means to create safe spaces, using outlets such as oral self-narration of the past as well as literature and music, particularly jazz and blues (Collins, 1991;Janssens and Steyaert, 2001).
The safe yet daring space created for connecting stories in our project allowed what Johnson andDuberley (2003, p. 1290) refer to as the 'epistemic reflexivity' through which 'social constructions in-use could arrive democratically rather than as the outcome of power play'. Feeling the safety of the space made it possible for some to share heartbreaking stories in which we heard that 'the system'-as they called ittook away not only their agency but also their hope and dignity. Parwin's story is an example of this. Parwin is an Iranian who came to the Netherlands as a refugee in the 1990s. The first part of her story is about the political background of her family, especially the women, as politically conscious and as activists who created the path for her to become active in a leftist political struggle against the previous regime in Iran. She also writes about her experience of the 1979 Revolution and the harsh suppression she faced after 1981. When listening to the part of her story that takes place before she left Iran, participants were amazed by the power of generations of women (including herself) in her story. It was inspiring to hear about how she survived difficulties and of her motivation to contribute to making the country she called her home a more just society. Once in the Netherlands, after living in an asylum seeker centre for some time, she felt absolutely powerless for the first time in her life. In Iran, she remained a passionate activist until the last minutes of her stay. She knew the enemy and was sure of the cause. She expected better treatment from a democratic society like the Netherlands. In her view, this was supposed to be the country that would give her-as a freedom fighter and refugee-a new chance to live. Instead, she was treated as untrustworthy. Her spirit was broken by living for years in isolation from the rest of society, waiting in limbo, denied of her agency as a female refugee and activist. She did not know how to fight this invisible enemy. She could not deal with this unexpected situation and became depressed, losing her passion for life. After describing her experiences in asylum seeker centres, she asked herself: 'Is this the freedom and democracy I was fighting for all my life?'.
Later, Kiki (a native Dutch woman of around 65 years old), who was also an activist when younger, narrated her story by connecting it to Parwin's. She wrote: During one of the sessions, the story of one of the participants who came as a refugee to the Netherlands touched me deeply. She wrote about all the things she left behind, all the things she struggled and suffered for and all the family and friends who died for those causes. All of that for freedom and democracy, something that one would expect to find here. She wrote: 'In all those asylum seeker centres you think to yourself: Is this it?'. This line cuts through my soul. In my own country I have not felt at home over the last couple of years.
This deep level of connection between Parwin's and Kiki's stories as female activists led to profound discussion on the meaning of freedom and democracy and their relation to visible and less visible forms of exclusion and othering in the Netherlands. Feeling the safety of the space and taking time to share one's life story from the past to the present created 'unusual connections' that were quite intense. They also enabled engaged discussions on deeper notions of activism and democracy. This type of engagement includes situated inclusive conversations and collaborations in which the personal, political and contextual meet (Young, 2002(Young, [2000). Molly Andrews (2014) argues that political narratives do not have to be explicitly about politics and that most of them are not. Individual stories reveal how individuals are positioned within the communities they live in and how they negotiate their positioning within a variety of power structures. The shared stories of this project naturally entailed many differences and similarities, but what seemed to be the most prominent commonality in all the narratives was the struggle for recognition and the desire to be treated with dignity as women and as activists. Feeling the safety of the space, placing the experiences of exclusion and pain in a broader context and rethinking the potentials of democracy and inclusion brought dignity and a passion for activism back to women who too often felt helpless and insignificant because of the prevailing negativity in the dominant discourse about refugees and also because of the denial of women's agency within it. In this way, the women produced a polyphonic agency that illustrated 'the possibilities of significant change from below' as 'a challenge to a top-down view of power and politics that has ignored or trivialized them' (Stall, 2010, p. 547). The engaged narrative approach of this project enabled a meaningful contribution in the process of the 'invention of the political' (Beck and Beck-Gersheim, 2002, p. 18) by creating a climate in which an in-depth form of engagement was enacted to resist visible and invisible structures of inequality. This is an example of how an engaged form of narrative scholarship can contribute to creating spaces for reflexivity and a form of connectedness that is vibrant and dynamic. This space brought back the passion of activism for justice and societal change for refugee women such as Parwin, a passion that is erased from the dominant discourse on refugees. This example shows that while sharing narratives of the past can be painful, finding commonalities in the stories of others can bring satisfaction, and even agency (Andrews, 2014).

conclusion
When the power of exclusion works through repetition and is manifested in the daily normalisation of our actions, agency needs to provide an alternative in the same fluid manner. This means thinking about agency in terms of small changes that are taken up by individuals, groups, communities, academia and so on, in their daily reflective actions. In this way, late modern agency is about microemancipation, which 'is less grandiose and more focused than in orthodox Marxist conceptualizations. It is rather partial, temporal movements breaking away from diverse forms of oppression [or I would say, normalisation], rather than moving towards [something as a utopia or a telos]' (Zanoni andJanssens, 2007, p. 1377). Repetition of these small reflective actions in an individual's daily encounters is one of the ways to subvert the subtle and ungraspable power of normalisation. By presenting examples from the two research projects discussed above, I have shown how dialogical narratives helped refugee women find their voices, their passions and thus their agency, which was forgotten or marginalised because of the normalising power of the dominant Dutch discourse that erased their agency by portraying them as weak, lacking quality and being non-emancipated. In the first project, mirroring the perspectives of refugee women like Sarah and those of other stakeholders resulted in reflection on the power of normalisation despite one's good intentions. This disruption of normalised images and actions created a 'partial and temporal' possibility to 'break away' from the dominant fixed image of refugee women as people with shortcomings, which in turn created a potential for acting differently. The second example showed how sharing narratives with women from different backgrounds enabled the revitalising of passions for activism and social change in Kiki, a native Dutch activist woman, and Parwin, a refugee activist woman. Their shared story as women and activists fighting for justice was an inspiration for other women in the project to also reclaim their agency as women, as migrants and as refugees.
Hannah Arendt (1958, cited in Andrews, 2014 argues that storytelling is 'the bridge by which we transform that which is private and individual into that which is public, and in this capacity, it is one of the key components of social life'. And what makes society so difficult to bear, according to Arendt, is that it has lost its capacity to gather people together or, as Bauman (2000) would say, to imagine a collectivity in a vital manner. Beck and Beck-Gersheim's (2002, p. 18) invention of the political is a call for a new kind of thinking in terms of connectedness. It means new vibrant spaces in a variety of locations (but also in academia) for reflexivity that can produce counter narratives of inclusion. Academics have a task, through their critical engagement with individuals and communities, to contribute in the revival of narratives that are often marginalised and suppressed because of the negative impact of normalising discourse. As Mike Marqusee (2012, cited in Andrews, 2014, p. 86) writes: We need utopian thinking if we are to engage successfully in the critical battle over what is or is not possible, if we are to challenge what are presented as immutable 'economic realities'. Without a clear alternative-the outlines of a just and sustainable society-we are forced to accept our opponent's parameters.
In this article, I have shown how the revival of refugees' agency through sharing narratives, together with engaged forms of scholarship, provides novel ways to think about power and agency. This production of interconnected voices of transformation from the margins captures the 'invisible and local-based individual actions and collective modest struggles which are at the root of all larger social change efforts' (Stall, 2010, p. 51). Engaging with such narratives has inspired me to see the significance of the 'elusive process of resistance that is taking place beneath the surface', which has helped me 'to illustrate the possibilities of significant change from below' (ibid., p. 547). Working with narratives in research and observing the revival of agency and passion for societal engagement they produce is essential for enlarging academics' imagination to re-approach the normalising power they face on a daily basis.

author biography
Halleh Ghorashi is full professor of Diversity and Integration in the Department of Sociology at the Faculty of Social Sciences, VU Amsterdam, the Netherlands. She is the author and co-author of several books and has published many articles on topics such as migrant and refugee identity and belonging, diasporic positioning and cultural diversity both inside and outside organisations. Her most recent international publication is the edited volume Contested Belonging: Spaces, Practices, Biographies