Looking for Women in the Field: Epistemic Ignorance and the Process of Othering

Abstract In this article, I ask how my challenges in the field can shed light on dynamics that contribute to excluding women as research participants in (climate) migration studies, and investigate the relationship between the absence of women as migrants in literature and challenges of accessing women in the field. Multiple studies have established that women – like men – migrate, and consequently have called for greater inclusion of women into international labour migration literature. Nevertheless, women still often disappear as research participants within this field. Through an investigation of literature I find that this is because women as migrants are Othered in academia. I build on this insight to investigate the dynamics that may also exclude women as research participants from other subfields of migration research, more specifically the field of ‘climate migration’. Next, I reflect on my unexpected challenges in accessing the information about women migrants in coastal Bangladesh. Men in the village, who were gatekeepers in my access to women, were not immediately willing to talk about women’s mobility. Rather, they tended to distance themselves from knowing these women, effectively turning migrant women into the ‘Other’. Bringing together insights about the Othering of migrant women in both academia and the field, I show how both contribute to upholding systems of epistemic ignorance. I argue that an awareness of such ignorance can be utilized as a tool in fieldwork by lending sensitivity to whom we include and exclude as knowers in our research.


Introduction: challenges in fieldwork
To 'enrich our understanding of the production of knowledge in a particular field' writes the feminist philosopher Nancy Tuana (2006), we need to investigate ways of 'not knowing' (2006, p. 3). In this article I investigate two different but interrelated ways of not knowing about women migrants.
I have been curious why women as migrants are not more visible in climate migration studies. Motivated by feminist researchers, who have questioned what they regard as a masculine positivist tradition and discussed various more suitable ways to research women and other overlooked groups, I wanted to contribute to filling this knowledge gap (Gioli and Milan, 2018). Thus, in early 2020, my translator Ahmed 1 and I travelled to the southernmost part of Barguna district, an area which has been identified as one of many climate change 'hotspots' in the country. The debate about who 'climate migrants' are and what the consequences of increased human mobility will be is currently very active in Bangladesh. The combination of being 'a country made for disasters' (Poncelet et al., 2010) and the high mobility of people living here has also made Bangladesh an ideal case study for researchers interested in the relationship between migration and climate change. Given that migration is so high on the political agenda in the country, one should expect it to be a simple task to research gender and migration there. However, this was not my experience.
In this article, I ask how my challenges in the field can shed light on dynamics that contribute to excluding women as research participants in (climate) migration studies and investigate the relationship between the absence of women as migrants in literature and challenges of accessing women in the field.
Feminist epistemologies typically bestow great importance on the perspectives of research participants (Coddington, 2017;Geiger, 1986; or 'herstory' (Jenkins et al., 2019, p. 417), and a substantial number of texts focus on techniques for creating safe environments and building trust when researching 'vulnerable' and 'hard-toreach' groups (Bhopal, 2010;Devault, 1990;Marsh et al., 2017;Newton, 2017). Intentions and techniques were of little help, however, when I was not granted access to the women I wished to speak to. In spite of earlier experience of researching women's mobility in Bangladesh, I encountered unexpected challenges in accessing the information I was looking for. Several of the men to whom Ahmed and I spoke, who were gatekeepers in our access to women, were not immediately willing to talk about women's mobility or introduce us to returned migrant women. Rather, the men tended to distance themselves from knowing these women, effectively turning migrant women into the 'Other'. It eventually became clear, however, that both women and men do move away from this southernmost part of Barguna to seek income-generating opportunities elsewhere.
In a typical narrative of accessing the field and recruiting research participants, Hortense Powdermaker (1966, pp. 51-59) reflects upon her first fieldwork experience on an island in today's Papua New Guinea, and describes how, during her first night alone in the field, she was struck with panic and loneliness, unsure of what to do next. Yet, after only a couple of hours and to her great relief, she was contacted by curious villagersand she was never lonely again. Powdermaker conducted her fieldwork during the 'traditional era ' spanning 1900-1950, which has influenced much later fieldwork in anthropology and sociology. Ideas about fieldwork during this period were built on 'myths of the heroic' (Atkinson et al., 1999, p. 461), or perhaps the lonely, fieldworker.
Multiple contributions to the challenges of fieldwork have been published since, including new ideas about what constitute 'the field'; the role, positionality and reflexivity of the researcher; and the (power) relationship between the researcher and the researched. Feminist research has made important contributions to these discussions (Sultana, 2007;Sundberg, 2003;Wolf, 1996). It is widely acknowledged that fieldwork is gendered, and several accounts describe how women are able to access field sites which are 'not open to men' or how 'women may find themselves restricted to the domestic world of fellow women' (Hammersley and Atkinson, 2019, p. 73). Less has been written about challenges women fieldworkers may face in accessing other women in the field (Grünenfelder, 2014;Waldrop and Egden, 2018;Williams and Drew, 2020). Moreover, while approaches to fieldwork have changed and multiplied since Powdermaker's time, the idea of the fieldwork 'hero' who overcomes (or doesn't face any) hurdles is still influential, and challengesespecially those which are not overcomeoften remain untold (Contreras, 2019).
In an article in this journal, Anne Waldrop and Sissel Egden address the classic idea of fieldwork in their description of how, when they were students, a lecturer described how he gained access to the field by 'just sitting down on a tree-stump' and people would come to him (2018, p. 240). The authors then illustrate why this was not viable in an urban Indian setting, in which both authors conducted fieldwork for their PhD projects, and where upper-class women are hidden behind high walls. Although I did my research in a rural context perhaps more similar to classic ideas of 'the field', the passive 'tree-stump' approach whereby the researcher waits to be contacted is arguably biased towards men's perspectives, as will become clear below.
Challenges in gaining access to the field are frustrating. Yet, 'this process is an important part of data production in its own right, which is often overlooked' (Waldrop and Egden, 2018, p. 242; See also Schramm, 2005) because it can feel irrelevant (Katz, 1994) or because uneasy moments can be uncomfortable to share as fieldwork is 'hauntingly personal' (Van Maanen, 1988, p. xiv). Yet, it is important to reflect upon the challenging sides of fieldwork, write Waldrop and Egden, because it can provide new insights of which we are not immediately aware. In this article, I continue this conversation by returning to the challenges I faced when trying to access information about women's mobility in the field.
In the first part of the article, I investigate the first way of not knowing about women migrants. Borrowing insights from a more established field than climate migration studies, I discuss how studies of women migrants have historically been Othered in literature on international labour migration where the default assumption is that the migrant is a man. Then, in the second part of the article, I look at the second way of not knowing and describe how women migrants were Othered in the community I visited. It is productive to look at these two types of Othering together because both, albeit in different ways, arguably contribute to the epistemic ignorance about women migrants through creating 'disadvantaged epistemic identities' (Tuana, 2006, p. 13). While the process of Othering in the literature creates the image of women as 'implausible' knowers about migration, the Othering of women migrants in the field through local norms confirms this image. This process is well captured by Edward Said (2003), who showed that researchers enter the 'field' with ideas about the empirical world they are about to encounterideas which they often look to confirm. Epistemic ignorance is thus produced through a recurring pattern of devaluing women's role as potential knowers.

Women as 'the Other' in migration studies
What is often termed 'climate migration' has been added as a subfield to the large, complex and interdisciplinary field that constitutes migration research. This emerging epistemic community of climate migration studies encompasses all the issues related to the idea that 'climate change, whether in the form of sea level rise, extreme weather events or drought, will "induce" a complex pattern of human mobilities, including migration, displacement and resettlement' (Baldwin and Bettini, 2017, p. 1). Here, I use the term 'climate migration studies' as a reference to this epistemic community, because there has been an increasing trend towards use this term in recent years (Nature Climate's November 2019 special issue entitled 'Climate Migration' is illustrative of this trend). It is not without hesitation that I use this term, however, because it has received well-founded criticism from political ecologists for being mono-causal, emphasizing external climate risks at the expense of the underlying political, economic, social and demographic processes and inequalities that cause people's vulnerabilities and influence human mobility (Baldwin and Bettini, 2017;Paprocki, 2015). I therefore want to emphasize that I use this term in reference to this epistemic community, not as a characteristic of the individuals who move.
Aside from a few notable exceptions (Massey et al., 2010;Tacoli and Mabala, 2010;Thiede and Gray, 2017), literature investigating the link between human mobility and climate change has thus far paid little attention to women as migrants, more often focusing on how women are 'left behind' when men move (Eastin, 2018;Ingham et al., 2019;Jacobson et al., 2019). This one-sided focus on women as immobile within this literature is curious, given a large body of literature highlighting the 'feminization of migration' within other branches of migration research (Hunter and David, 2009).
Lori Hunter and Emmanuel David (2009) suggest that 'early thoughts on the potential for gendered climate-migration connections can be informed by bringing together literature on the feminization of global migration streams' (p.10) for income purposes, as climate change is believed to negatively affect people's livelihoods where they live (Warner and Afifi, 2014). Following this suggestion, I draw insights from two special issues of journals which examine the inclusion of gender in international labour migration. The first special issue was published in the International Migration Review in 2006. In their contribution to this issue, Sara Curran, Steven Shafer and Katharine Donato review the inclusion of gender in research on international labour migration, mainly within North American sociology. They show how the assumption of the migrant as a man came under challenge from feminist scholars from the 1970s onwards, who pointed out that women also take part in the migratory process (Chant, 1992;Houstoun et al., 1984;Zlotnik, 1995). One of the reasons why women were not included as active in the migratory process, it was argued, is because women's work is seldomly counted (Morokvasic, 1984). Women were therefore not seen as important to this field of research, as they were believed to passively follow their husbands when they moved (Morokvasic, 1984). Challenging this presentation of women, feminist researchers focused on women's specific roles and contributions during migration processes and highlighted that women were independent economic actors either when part of a moving household or when migrating on their own, while having experiences that differed from those of men (Boyd, 1984;Lauby and Stark, 1988;Pedraza, 1991;Tienda and Booth, 1991). It is important to note that, although women's migration has increased over time in some parts of the world, this 'feminization of migration' reflects an increase in the visibility of women in the literature as much as an actual increase in the number of women in migration flows (Green, 2013).
In line with broader trends within feminist research, since the 1980s and 1990s gender and migration studies have moved away from essentializing stereotypes towards approaches that understand gender as socially constructed. Firstly, there has been a shift away from 'women' and migration towards 'gender' and migration and, secondly, there has been a shift from a focus on differences between men and women to also considering differences among women. Such intersectional perspectives incorporate other social characteristics and power dynamics into the analysis, alongside gender. Intersectional approaches build upon the work of postcolonial feminist researchers, who argued that feminist studies have favoured the worldview of white Western women and that other perspectives, hereunder different methods, need to be included (hooks, 1982;Mohanty, 1984;Yuval-Davis et al., 2005). Chandra Mohanty, for example, illustrated that white Western feminists (and others) tend to characterize women as: a singular group on the basis of a shared oppression. What binds women together is a sociological notion of the 'sameness' of their oppression. It is at this point that an elision takes place between 'women' as a discursively constructed group and 'women' as material subjects of their own history. Thus, the discursively consensual homogeneity of 'women' as a group is mistaken for the historically specific material reality of groups of women. (1984, pp. 337-338) The concept of 'epistemic ignorance' (Code, 2014;Sullivan and Tuana, 2007;Townley, 2011), which addresses what we do and do not know and why this is, is relevant here. While ignorance can appear as simple gaps in knowledge that can be corrected once discovered, feminist and postcolonial scholars have pointed out that such ignorance is not a mere accident, but often the consequence of a more complex set of power dynamics. It is therefore of interest to look into how ignorance is 'produced and sustained' (Sullivan and Tuana, 2007, p. 1). Kristie Dotson links such epistemic ignorance to who we understand as being 'knowers ' (2011, p. 243). We believe that only those whom we regard as knowers can tell us anything of relevance. Thus, those who are not believed to be knowers about the topic at hand will not be asked, or heard. Epistemic ignorance is thus a practice of silencing though a recurring pattern whereby certain groups are left out as knowers through 'the construction of epistemically disadvantaged identities' (Tuana, 2006, p. 13).
The concept of Otheringthe process of constructing Others as different from, or in opposition to, the norm (Språkrådet, 2017) explains how such disadvantaged identities are created. Othering is a fundamentally relational concept which relies on comparison with, and distancing from, the Other. We look to others to decide what we are not and, thus, the Other tends to be devalued. Edward Said (2003), while not using this particular phrase, showed how such a process can also occur in a collective manner when certain groups are constructed as 'Others' through exclusion in order to construct a collective self-image. In order to construct a collective Other, an essentializing process is necessary, in which Others are made into a homogeneous group (Said, 2003, p. 97), as also illustrated by Mohanty's quote above with regard to women. Said furthermore linked the process of collective Othering to practices of knowledge production, building on Michel Foucault's (1972) ideas about the power inherent in knowledge. By highlighting how essentializing 'truths' in the colonial project were based on empirical research, Said illustrated how the role of research can be intertwined with the process of Othering. Discourses, argued Said, are composed of a mixture of empirical 'reality' and the ideas that the researcher already holds. Thus, discourses are 'not "truth" but representations' (Said, 2003, p. 21). Such representations contribute to the production of epistemic ignorance, where the 'normal' is systematically researched, heard and represented, while the Other is not, and is thus silenced.
In 2017, a new special issue was published about the situation of gender in the international labour migration literature in the journal American Behavioural Scientists, following up on the special issue from 2006. Despite the developments in inclusion and visibility provided by feminist research, as described above, Katharine Donatoone of the editors of the 2006 issueand her colleagues write: 'As we near the close the second decade of the 21st century, the gender content of migration scholarship in sociology remains, at best, frozen and stalled' (Donato et al., 2017(Donato et al., , p. 1079. They find several explanations for why few sociologists integrate gender considerations into their studies on international labour migration. Firstly, while including gender already complicates research because it adds an additional component, the expectation that researchers will also use intersectional approaches to include additional components that intersect with gender, like class and race, may cause them to hesitate. Secondly, the authors argue, only women research women, but both women and men study men. This is in line with Monica Boyd and Elisabeth Grieco's (2003) argument from more than a decade earlier, upon which Curran, Shafer and Donato drew for their review in 2006, that little effort has been made to include gender considerations into international labour migration studies. One challenge to incorporating gender in such studies, Boyd and Gricero argued, is that there are many subfields of migration studies, and gender and migration studies have effectively become one such subfield, dealing with 'Other' questions than 'normal' migration research. Because this field is mostly made up of researchers, gender and migration research is viewed as a 'woman's issue', with the consequence that 'gender' is often collapsed to mean 'women'. This means that male migrants are not understood as gendered 'men' but as the normal migrant. This is in line with the explanation for women's absence from migration studies put forward by Jørgen Carling (2005). He argues that women are often overlooked because, when moving for work, they are perceived as acting like men. When women migrants are treated as a deviation from the 'normal' image of the migrant as a man, the image of the migrant as a man remains intact. The migrant man is the norm while the migrant woman is the Other. The consequence of this is that researchers who are not committed to the subfield of gender and migration are not likely to look for gendered differences in their research. In effect, constructing subfields legitimizes ignorance.
The image of who the migrant is and who we look for when we are researching human mobility is important, because it will guide the questions we ask, which, again 'guide our research and writing' (Green, 2013, p. 783). One of the main reasons why women and gender considerations disappear from view in studies on international labour migration studies, it seems, is that women are not believed to be knowers. A consequence of this is that, as Donato and her colleagues point out, most researchers studying migration do not ask questions about gender relations. This is an important acknowledgement, because, to use Nancy Green's words again, new knowledge is 'as much the result of our shifting ways of knowing than of anything inherent in the subject itself' (Catignani and Basham, 2021).
In this section, I have shown how women migrants have been constructed as the 'Other' in international labour migration research and how this has manifested as epistemic ignorance about women as potential knowers about migration experience. I have thus investigated one way of not knowing, and which I initially set out to challenge. In the following, after laying out the context of my fieldwork, I will turn to a different kind of Othering, which took place in the field, and led to anotherand more unexpectedway of not knowing. Thereafter, I will return to my question of how my challenges in the field shed light on the dynamics that contribute to excluding women as research participants, and how the two types of Othering are interrelated.

Women's mobility in Bangladesh
Today's Bangladesh is best known for its labour migration to other countries (Siddiqui et al., 2018). Between 1976 and 2017, a total of 11.9 million Bangladeshis migrated abroad. From the 1990s, women also started to move abroad to take jobs as domestic workers, and in 2016, women constituted 19 per cent of this workforce, numbering 118,000 (Siddiqui et al., 2018, pp. 21-26). For many years, internal migration has also been an integral part of income diversification in the country. Day labourers have moved to other districts for harvest during parts of the year, and the introduction of ready-made garment factories in the 1980s led hundreds of thousands of people, especially women, to move to the cities for work (Afsar, 2002;Huq-Hussain, 1995;Kabeer, 1991). Although less researched, women have also moved in large numbers to work as domestic workers for Dhaka's upper-class households (Bork-Hüffer et al., 2016). A volatile environment due to the country's location on the world's largest river delta has also been an integral part of life and a major reason for human mobility in Bangladesh. The combination of exposure to environmental risks and the high mobility of people living there has, as mentioned earlier, made Bangladesh an ideal case study for researchers interested in the relationship between migration and climate change. The debate is currently very active about who 'climate migrants' are, how many there are, and what the consequences will be for both rural and urban areas. Looking to the knowledge presented here about women's migration and the great attention given to both internal and international migration on the political agenda of the country, one would expect it to be a simple task to research gender and migration in Bangladesh. Other dynamics complicate this picture, however.
In Bangladesh, gender roles are mediated through the concept of 'purdah', which, directly translated, means 'curtain', and refers to the separation of men and women in society. Tools for maintaining purdah can be split into the physical segregation of men and women and the covering of the female body (Papanek, 1973). While purdah norms are usually linked to women's morality, purdah itself is a relational concept which guides expected behaviour for both men and women. While women are responsible for the home and children, men are expected to provide for their household (Kabeer et al., 2011). This has two different but interrelated meanings. Firstly, the man is responsible for earning an income to meet his household's material needs. Secondly, he is required to protect the women and children in his household. When a woman takes up waged work outside the home, it indicates that her guardian has failed in providing for her in one or both of these ways, signalling that he has failed to fulfil his responsibilities as a man. Men's disapproval of women's participation in waged work is often due to this stigma, rather than a lack of belief that women can contribute to their family's well-being (Evertsen and van der Geest, 2020). This division of labour, where women's honour is exchanged for men's provision and protection, has been termed 'the patriarchal bargain' (Kandiyoti, 1988), and captures the dynamics of purdah in Bangladesh (Hossain, 2017, pp. 75-90). In practice, however, this bargain has never included all womennot all families have been able to afford to keep their women at home. Moreover, it is important to note that purdah norms are contested and changing, and hence more flexible than is often believed by outsiders (Hossain, 2017;White, 1992;. While the reasons for why women migrants slip out of focus in research are clearly many, it is possible that purdah norms play a role in creating an image of Bangladeshi women as immobile, especially for foreigners unfamiliar with the local context. This leads us to my project.

Arriving in Barguna
After having spent several months in Dhaka, the capital of Bangladesh, my translator Ahmed and I travelled to the coastal district of Barguna. Bargun means big rope, referring to the ropes used by wood traders from the north to tie up their boats to secure them from the strong currents of the Khagdum River (BBS, 2013). We went to the southernmost part of the district, an upizila situated between three rivers and facing the Bay of Bengal. This location exposes the area to cyclones, tidal flooding and increased salinity levels (Torikul et al., 2017). The upizila is home to almost 164,000 people, who make a living mainly from farming, fishing, fish farming and fish businesses (BBS, 2018, p. 21). Although economic insecurity persists, the area is doing relatively well, with a lower poverty rate than the country average (BBS, 2018, pp. 59-71). Having received a positive response from an aid organization working there to come and learn about their work helping villagers adapt to climate change via their climate change adaptation programme, I went there to conduct my research.
I went to Barguna with the expectation that this was an area from which a substantial number of people would be moving elsewhere for work, either to cities in Bangladesh or to the Middle East, for shorter or longer periods. Several experts I had spoken to in Dhaka had pointed to Barguna as a region that women migrate from, so I expected women to move as well as men. However, the cultural norms of purdah imply that many community leaders will be men. Moreover, because women are not supposed to talk to men outside of the family without a man from their household present, one will meet few women in public spaces in rural Bangladesh. Men, however, are very accessible because they will be working outside or taking a break in the tea-shops, which can perhaps be said to be the equivalent of the 'tree-stump' referred to in the introduction to this article. Because public spaces are occupied by men, while one must have access to people's homes in order to speak to women, men effectively become gatekeepers in the field, and were thus controlling my access to women.
As a white, Norwegian woman who does not speak Bangla, I faced a dilemma because I was dependent on men as gatekeepers for access to the field, while travelling to coastal Bangladesh primarily to study the women living there. Should I, for example, work with a woman or man as my translator? This was an important choice, because fieldwork in a cross-cultural and cross-language setting creates a triple subjectivity made up of the researcher, the translator and the research participants (Caretta, 2015). Having a man as my translator would allow access to community leaders, who tend to be men, as well as men's spaces, while it would hamper my communication with women. In this case, I first employed a man as my translator, which allowed me access to men's spaces. Then, after we had identified a number of women participants, I planned to go back with a woman translator to conduct interviews with women and spend time in women's spaces.
Over a period of two months, my translator Ahmed and I spent three weeks in this community before the coronavirus pandemic cut the fieldwork short. We travelled around the area, speaking to people mostly in the informal setting of the village teashopsthe social meeting point for men across Bangladesh. Feisal, who worked for the aid organization I had contacted beforehand, also showed us around for much of the time, especially during our first week. Apart from a few formal interviews with key informants, the conversations were not recorded. I jotted down notes as we spoke and wrote out my field notes upon my return to the guesthouse in the evenings. From these jottings, I was left with close to a hundred pages of field notes. Karin Widerberg (2001, pp. 125-126) has described how the process of writing and analysing are often intertwined. I started writing up ideas immediately after the fieldwork before I returned to a closer reading of my field notes. After a thematic round of coding (Johannessen et al., 2018) to ensure that the analysis was grounded in my field notes, I realized that I had more material concerning the migration of both men and women than my gut feeling had told me. In a second round of coding, I looked specifically for places in my material where I had encountered challenges or avoidance. While the image of women migrants as the Other in literature creates an image of the migrant as a man and thus undermines questions about women's mobility, it became clear from my material that a different type of Othering was going on in the field. In what follows, I will share some of my experiences from my weeks in Barguna, and how researching women's mobility there turned out to be more difficult than I had anticipated. I will show how women migrants were Othered by the villagers I spoke to, who consistently distanced themselves from knowing these women.

Talking about women's mobility
During our first week in the field, we heard several stories about how the area had been severely affected by Cyclone Sidr, which hit India and Bangladesh in 2007. Many people lost family members or had to move because they lost their land and belongings to the sea. On several occasions, we were told that, after Sidr, many development projects had been introduced into the area. During the last couple of years, the area has also seen investment in improved infrastructure, and several of the main roads were being asphalted during our stay. We were told that many people were now investing in income-generating opportunities other than fishing and farming, such as driving a car or motorbike, or, as will become clear, moving elsewhere for waged work for various periods of time. Often, several income-generating activities were combined.
According to several villagers we spoke to, these interventions had helped the economy in the area so that fewer people now needed to move. Gradually, people had returned. It was difficult to get a sense of how many people had moved in the past, and how common it was for people to move elsewhere for work now. Migration was often expressed as an exception. During one tea-shop conversation, for example, a man said that 'people used to go to the cities for work … but this is not necessary anymore'. Then he added, 'although some people go'. From the outset, I wanted to avoid using the word 'migrant', because this is a term with which people may have very clear associations, and hence only relate it to one type of human mobility. The word 'migrant' (obhibashi) is also a 'city-word', according to my translator, and it was not a word that the people we talked to were using. Although people did not normally use the word 'migrant', they were clearly dividing people into different categories according to their reasons for leaving, their destination, and the length of their stay. Avoiding the word 'migrant', we would ask about people who move away and/or come back (chole gechen / phire eshechen), because this was also the wording used by most people there. By using this wording, however, we seemed to tap straight into the post-Sidr narrative of people who moved away for a period of time before returning.
Several people referred us to talk to the village police officer, as he was said to have an overview of people who came and went from the village and could provide us with a 'list'. While we were at his house: … it struck me that people may think of different things when we ask about people who move or migrate. I ask him what he thinks of when he thinks of migration. He answers that people who have moved to earn money elsewhere, but who are eventually coming back he does not really consider migrants. I'm realizing that the reason why some people would say that many people move while others are not saying so is that they may have different ideas about who is a migrant. If most people think like the policeman, then people who move to Oman for a period of time but come back twice a year and have as their ultimate goal to return when enough money is saved, are not considered migrants but rather someone from the village who is just not there right now. The people who have migrated are the people who left. The ones who are not coming back. (Fieldnotes, Barguna, 3 March 2020) It turned out that, by people who had migrated (although they did not use this word per se), people in this area were often referring to those who were unable to return after Sidr. To leave the village, at least on a more permanent basis, was associated with loss and failure. As one man said when I asked whether most people returned: 'You know, this is their home. They wouldn't go there to live'.
The village police officer also raised another interesting point. He said that he could easily give us a list of men who had migrated off the top of his head. The women were more difficult, he explained, because they were not really migrants: they would always return. From this starting point, it was difficult to direct the conversation towards women's mobility within and away from the area.
In one of our tea-shop conversations towards the end of the first week, this difficulty became apparent. About 10 men were standing outside the tea-shop talking to us. I started the conversation by asking about life in the village. Much in line with what people had been saying throughout the week, the shop owner said: 'Other people may not like it here, some people want to live in the town, but we love it here, we can't live anywhere else'. Do some people from here go to live in towns? I asked. 'Yes', they said, 'this man here used to live in Dhaka'. One of the men, I'll call him Hasan, explained that he had lived and worked in Dhaka for eleven years. He came back in 2011 because the wages dropped. Having a man in front of me who has himself lived in Dhaka, I asked whether women also go. 'Yes', the men confirmed. Hasan explained that he is married to a woman he met in Dhaka who worked in garments. She is from another district.
'How is migration by women perceived here?' I wanted to ask. My translator Ahmed looked at me and said, 'you cannot ask that, this man here is married to a garment worker, he might be offended'. To take the role as a naïve outsider can enable the researcher to ask questions which may be considered 'off limits' by locals (Grünenfelder, 2014). Ahmed, however, although living in Dhaka and thus occupying a middle position between insider and outsider, was uncomfortable asking this question. His uncomfortableness highlights the role of the translator in the interaction, or the 'triple subjectivity' (Caretta, 2015): Although it was my question, and I was the one to blame for any negative reaction, Ahmed was the one who had to utter the words, and as such the question also became his. Ahmed's intuitive reflex to avoid this question was also informative in the sense that it illustrated how controversial it is to bring the (broken) patriarchal bargain into view.
We changed the question and asked instead: 'Are women encouraged to go?' Hasan gave us a little smile and shook his head. 'No', he said. 'Actually, the men will not ask their women to go so it is the women themselves who take the initiative.' 'Does this create conflict in the families?', I asked. 'Yes', he confirmed. He continued by explaining that a man will not let his daughter go, because he thinks that she will meet someone in the city and then she will move to his district instead of marrying someone from here. If the daughter marries without the consent of her family, the family's honour is violated, he explained. And then the husband will eventually leave her, he added, and she will have to come home to her parents again. 'They will think that if she hadn't moved in the first place, none of this would have happened. ' At the beginning of the conversation, when we talked about Sidr, the men in the tea-shop were very engaged in the conversation. They spoke loudly, interrupting each other to tell their Sidr story and how the area has been doing since. They told me about the livestock they lost, the conditions for fishing in the area, and their concern about the state of the embankment. When I began to ask about women leaving the area to work elsewhere, they lost interest and started leaving the teashop. At the end of our conversation, we were only talking with Hasan. I was left with the feeling that women moving away for work was not something they wanted to talk about.
Hasan's explanation mirrors the presentation of purdah above. When a woman moves for work, the patriarchal bargain breaks down. In my previous research, women justified this by pointing to the fact that the men in their families were unable to provide for them (Evertsen and van der Geest, 2020). In other words, according to these women, it was the men, not the women, who were the first to break the bargain. Thus, when a woman moves for work, it indicates a failure of the men in her family. This might explain men's unwillingness to talk about women's mobility. While I did eventually get some more information about women's mobility in this area, speaking to men in a group about this turned out to be especially tricky, as they were reluctant to speak about women's mobility in front of other men. My field notes reflect my frustration with this, and how I tried to change strategies: During breakfast, I fill [Ahmed] in on my plans and what I have been thinking. … I tell him … that instead of asking 'whether' people move from here, we will ask 'who' move and where we can find them. If nobody are moving from here, this will nevertheless be the answer people will give, even if we ask the migration questions more directly. (Fieldnotes, Barguna, 2 March 2020) The day after I had hatched this new strategy, we went back to a tea-shop we had visited several times. There were a couple of familiar faces there by now. After chatting a bit, I once again explained that I was trying to learn about people who leave the area, and asked which women go. It is mostly women who are 'helpless' (osohay) who go, they said. 'The helpless who have lost their husbands, they go to Dhaka to work in the garments.' 'Are more women going now than before, or were there more women going before?' I asked. After Sidr, there were more helpless women, they explained, but now there are fewer, referring to the improvements in the area described above. I continued by asking whether there are any differences between the women who go to Dhaka and the women who go to the Middle East. One man said that the women who go abroad are more helpless. Another young man, Hafiz, disputed this and said that women who go abroad usually have more skills than the women who go to Dhaka because they have often worked in Dhaka for some time and acquired skills before they go abroad. I asked them if they knew women who had returned and whether they could introduce me. None of the men said that they personally knew any women who had moved away to work.
The older man says that a relative in Dhaka offered to help him send women from this area abroad. But he did not want to be involved in this type of work, he says with a stern face. This relative has a training centre for women who are going abroad, he explains. I ask him if we can have his relative's contact information so that we can talk to him when we go back to Dhaka. He shakes his head, saying that he had his card but he does not have it anymore, and he does not have his phone number. (Fieldnotes, Barguna, 3 March 2020) As if to compensate me, the old man finally pointed towards the embankment. 'I know two women are living nearby who just returned', he said. 'If you call me, I can take you to them.' The conversation soon turned to other topics.
By referring to women who move away for work as 'helpless', the men were indirectly saying several things. Firstly, 'helpless' is a negative description in the sense that it is an unwanted phenomenon. By saying that it is only the helpless women who go, they were saying that this is not only an unwanted phenomenon, but also an uncommon one. An exception to the rule. Importantly, this is primarily meant to refer to the fact that women who have to move outside the home to work will face hardship, and therefore it is a negative experience for them. In this sense, describing them as helpless is meant to display empathy with their hardship. This framing also implies that women who move have no choice, which makes it permissible, but that it would be better for them if they did not have to go. In this way, the men were displacing the blame onto structural factors, rather than individual choice or bad morality.
At the same time, portraying the women in question as helpless contributes to upholding the rationale of purdah, by indicating that women are dependent on men as their providers and protectors. Interestingly, the men said that it is the women whose men have 'abandoned' them who leave. They were not saying that it is the women whose men cannot provide for them who go, but the women who have no men around at all. Here, purdah norms are again reflected through reference to the idea that only women whose men cannot provide for them will break the patriarchal bargain and move for work. Poor households cannot afford to keep their women at home. While the men wanted to express empathy with the helpless women, they were also indirectly criticizing the men in those households, which made it difficult for them to include themselves in this story without jeopardizing their own honour, as Hasan had explained to us earlier. In hindsight, it is perhaps not surprising that they only referred to 'other' women when I asked if they knew any returned migrant women. As we shall see, however, several of these men turned out to have women family members who worked in other cities.
On our way back, one of the men from the conversation, Reaj, hitched a ride with us. As he was getting out of the car and was about to leave us, Reaj stopped and said: 'My sister used to live in Jordan'. He explained that she had just returned and was not planning to go abroad again. She lives with her husband in a neighbouring village, Reaj said. 'If she comes here to visit, I will let you know.' Then he left us. Ahmed and I looked at each other. It seemed clear that he had not wanted to speak about his sister in front of the other men. This would have put his sister in the group of 'helpless' women and would have been an indirect criticism of his brother-in-law, as well as reflecting negatively on himself.
The next day, Ahmed and I joined Feisal for a training session the local government was holding for women's saving groups in the area. It was held in a back yard on the outskirts of the village. Afterwards, we were all walking back towards the bazaar. As we walked, I chatted with a woman, Rohima, who invited us back to her house. Before parting, we exchanged some words with Feisal as we were standing at the crossroads between the main road and the small road leading to Rohima's house. 'On the opposite side of the main road here lives a woman who has been going back and forth to Dhaka many times', Feisal said, and pointed. 'She has also been abroad', he added. I made a mental note of this before we followed Rohima alongside the pond in front of her house.
As we were sitting outside Rohima's house, Hafiz from the previous day's tea-shop conversation entered the courtyard and joined us. Rohima explained to us that her and Hafiz's fathers were cousins and that the woman Feisal had just mentioned as going to Dhaka is Hafiz's aunt. I recalled the discussion at the tea-shop the day before, where Hafiz had disagreed with the other men about which women had most skills. I realized that, without telling us so, he had been referring to his aunt. The day before, however, Hafiz had not mentioned to us that he knew one of these women himself but had talked about 'other women'. Now, however, he spoke to us about her. Hafiz explained that his aunt and her husband met in Dhaka before coming here and that they have been moving back and forth ever since. I asked whether they were both in the village now. She was in Dhaka, he explained, but hearing that her husband had remarried, she came back here. When we left Rohima's house, Hafiz walked back with us. He invited us to come to visit him and pointed out his house. I asked him where his aunt was living, and whether he could introduce us to her as well. 'Yes', he said, 'she lives in the same house as me'. Like the men who had left us alone with Hasan in the tea-shop, it seemed clear that Reaj, like Hafiz, was reluctant to speak about his aunt in front of the other men because this would reflect negatively on himand hisfamily.
Towards the end of our stay, we ate lunch at Feisal's house. After lunch, he asked a neighbour to take us to a woman's house nearby. 'She's an old, single woman who has returned here from another village. I thought you might want to talk to her since she has returned here', Feisal explained. After lunch, the neighbour, Iqbal, took us to her house. It was just a couple of houses away. After we had been talking to the old woman for a bit, she said that she has three daughters living in Dhaka. I explained to her that I was very interested in talking to women who live in Dhaka, and she gave us the name and number of one of her daughters. A couple of days later, Feisal told us that Iqbal himself has a daughter in Dhaka. 'This man is very poor', Feisal explained. 'He has to cross my land to get to the road.' I was surprised to hear that Iqbal had a daughter in Dhaka. When we were visiting the old woman, and I explained that I wished to get in contact with women who had moved to Dhaka, he did not say a word about having a daughter who lived there. 'This is the third time this has happened this week', I noted. Feisal said that he is very worried about Iqbal's younger daughter because her father cannot afford to buy the books required at school. 'If he can't afford this she will have to drop out. And then, she will have to be married off at a young age.' 'How old is she?' I asked. 'Thirteen, fourteen.' Feisal sighed and said: 'If a girl's education stops, she will have to be married or go to garments'.

Concluding remarks: epistemic ignorance and the process of Othering
How can the challenges I experienced in the field shed light on the dynamics that contribute to excluding women as research participants? Through the examples described above, I have followed Waldrop and Egden (2018) in arguing that it can be productive to reflect upon challenges experienced in the field because this may provide new insights that are not immediately visible. More particularly, I have showcased how, despite entering the field armed with feminist epistemologies, previous experience from similar contexts and several months of preparations to find the 'right' place to go, there were still dynamics in the field that created challenges for my research.
The 'tree-stump'or 'tea-shop'approach gave me access to men, but not to women. Furthermore, women's mobility was not something that men wanted to talk about unless asked directly about it. While I was aware that travelling with a man as my translator would make it difficult for me to speak to women, I had not anticipated that it would be this challenging to speak with men about women's mobility.
First, in the tea-shop conversation with Hasan, most of the men left when I introduced the topic of women's mobility. Then, in the latter group conversation with the men in another tea-shop, the group painted the image for me of the 'helpless' women who have to move away for work because they are unfortunate and abandoned. However, they all talked in general terms and did not let me know until I was with them alone that they themselves had women relatives who had moved for work. Finally, Faisal's neighbour Iqbal kept hidden form us that he, too, had a daughter in Dhaka.
One reason why I received so little information about women's mobility was linked to initial misunderstandings about how villagers perceived the division between 'migrants' and 'non-migrants' in the aftermath of Cyclone Sidr. Women did not count as migrants, because, as the policeman said, 'they would always return'. However, the examples above clearly illustrate that, even when I had sorted the wording out and asked the men more directly about women, men would rather talk in general terms than acknowledge in front of others that one of their own women relatives has moved elsewhere for work. When re-reading my material, I noted that when I asked who typically moves during group conversations, I got answers like: 'those who cannot fish or work have to go to garments' and 'the ones who go to Dhaka go mainly to work in the garment factories'. My impression from these conversations is that talking about 'garment workers'who are often womenallowed the men to talk about women without talking about any particular women. This way, they could politely answer my questions, while also avoiding articulating something that might reflect negatively on themselves or be offensive to other men in the group who might have women relatives living elsewhere. While purdah is often used to explain women's situation in South Asia, it is important to remember that this is a relational concept which also dictates what constitutes a 'good man'. A good man can provide for the women in his household. Othering is also a relational concept, where we look to others to decide what we are not. The discussion of my field experience has described how men view women, through which it becomes visible how they view themselves and how Othering becomes a self-protective tool.
My positionality as an outsider made it harder for me to grasp the implicit meanings in my conservations with the men. While the men knew each other and knew what was kept unsaid in the group conversations, I did not. Moreover, it is possible that, by adhering to the norms by being careful not to ask 'offensive' questions, Ahmed and I contributed to this silencing.
As discussed in the first half of this article, studies of international labour migration are still based on assumptions that the migrant is a man and the woman is the Other, whereby the process of Othering creates the image of women as 'implausible' knowers about migration. This image makes it less likely that researchers will ask questions about women as migrants. Here we see how the absence of women as migrants in literature and challenges of accessing women in the field are brought into relation with each other. When information about women or gender considerations is not readily available to the researcher, as was the case in my fieldwork, this contributes to confirm the image painted in much of the literature that women are not of interest to migration researchers. My experiences thus illustrate how quickly women can be excluded as research participants, and how difficult it might be to include gender or sex compositions once the data has been collectedby this stage it may be too late because women have already become invisible in the material and hence are automatically excluded as research participants.
One consequence of this is that, in situations where men move more often than women, the story of the majority quickly erases the stories of minorities. The story becomes not that men move more often than women, but that only men move and women never do.
While the process of Othering helps us to understand how the construction of epistemically disadvantaged identities takes place, the concept of epistemic ignorance illustrates that positioning as the Other is systemic and thus reproduces the ignorance of the Other as a knower. As such, Othering is the outcome but also the precondition for epistemic ignorance as a system. Returning to Sullivan and Tuana's call to investigate how epistemic ignorance is 'produced and sustained' (2007, p. 1), the concrete examples in this article highlight how the image of women migrants as the Other in literature and the Othering of migrant women due to normative expectations in the community where I did my research and can reinforce each other and contribute to upholding a system of epistemic ignorance. Epistemic ignorance is produced and sustained by multiple yet interrelated processes of not knowing, and here I have investigated and highlighted two such processes.
Thinking about areas of potential ignorance can be productive (Townley, 2011) because it allows us to better understand both what we do not know and what this means for what we think we do know, which will again influence how we go about our fieldwork. Awareness of epistemic ignorance can be used as a tool in the field, lending sensitivity to the fact that the researcher cannot expect to come into contact with all the relevant research participants by following passive strategies, as the tree-stump approach suggests. Who do we think of as knowers and who do we exclude from our research? While not every researcher will have the time or interest to look for groups that may not represent the majority or the primary trend, this should be explicitly acknowledged in order to avoid the erasure of other existing stories. As Said pointed out, unless what is understood as 'normal' is spelled out, it becomes impossible to question, and what is understood as 'different' cannot be questioned either (Said, 2003, p. 106). As Bangladesh is becoming the case study for researchers interested in the relationship between climate change and migration, the idea that women can be knowers of interest to (climate) migration researchers should not have to be reinvented.