Women as subjects of risk in Bangladesh’s coastal riskscapes

In this article, I examine the assumptions underlying the idea of women as vulnerable and at risk, and how this understanding contributes to shaping practices of climate change adaptation. The article is based on ethnographic fieldwork in the expert community in Dhaka and in a climate change adaptation field site in coastal Bangladesh. Following Ananya Roy’s work, I understand coastal Bangladesh to be a riskscape, a geographical space suffused with imaginations of anticipated risks that must be managed through disaster risk reduction and climate change adaptation. Roy further argues that riskscapes create new subjects of risk, and that such subjects are highly gendered because development has tended to focus on poor women ‘in very specific ways’. In this article, I build on Roy’s insights to explore how women are constructed as ‘subjects of risk’ in climate change adaptation. I also show how climate change adaptation, in becoming the new buzzword for development, continues to focus on poor women in very specific ways, which are applied to fit with the climate change ‘metacode’. While well-intended, the understanding of women as subjects of risk is imbued with ambivalence, because it may contribute to supporting structures that make women vulnerable by normalizing relations of risk.


Introduction
Coastal Bangladesh occupies a central ideological space in imagined climate change futures (Paprocki, 2019).The region is often referred to as one of the most vulnerable to climate change in the world, situated as it is in the world's largest river delta, crisscrossed by the Ganges, Meghna and Brahmaputra rivers and bordering the Bay of Bengal.The government of Bangladesh has been measured against its ability to protect its population from the risk of disasters since the country saw independence in 1971 -a time during which the country has experienced a chain of disastrous events (Aase, 2022).In response to the current anticipated calamitous climate futures, coastal Bangladesh is now seeing an influx of climate change adaptation projects aimed at aiding the people who live there to cope with and prepare for the negative consequences of climate change.
There has been a slow but steady growth in calls for more sociological research into the social effects of climate change (Beck, 2010;Lever-Tracy, 2008;MacGregor, 2009;Urry, 2009).In her contribution to this conversation, Sherilyn MacGregor ( 2009) highlights one research gap where she argues that sociologists should intervene, which is the issue of gender and climate change.Most of the research in this area, MacGregor notes, has been conducted from a development perspective, investigating how women and men experience the consequences of climate change differently.Sociologists can supplement this work, MacGregor suggests, by examining the ways in which 'discourses frame and shape dominant understanding of the issue.Gender here is not just an empirical category (i.e.men/women), it is also a discursive construction that shapes social life ' (2009, p. 127).This is important because, as sociologist Ulrich Beck (2010) has reminded us, not only does climate change itself affect already existing inequalities, but so do responses to climate change, which often rely on discursive ideas about how to understand the social challenges that climate change poses (Swyngedouw, 2010).In this article, I respond to MacGregor's call by examining the assumptions underlying the idea of women as vulnerable and at risk, and how this understanding contributes to shaping practices of climate change adaptation.I show how a particular adaptation project in coastal Bangladesh, because it was embedded in the geographical and social fabric of the village where it was implemented, became part of the social production of riskscapes and gendered subjects of risk.

Riskscapes
Ananya Roy refers to riskscapes as 'territories' which are 'owned, distributed, mapped, calculated, bordered, andcontrolled' (Elden, 2007, p. 578, cited in Roy, 2012, p. 138).The idea of riskscapes brings several concepts together.Firstly, Roy builds on Arjun Appadurai's idea of 'scapes' as 'situated imaginations' (Appadurai, 1990, pp. 296-297).Riskscapes, then, are territories of imagined risks.This does not mean that the risks are not real.However, as Ulrich Beck has noted, risks can be understood as 'threats and catastrophes which are anticipated but which often remain invisible and therefore depend on how they become defined and contested in "knowledge"' (Beck, 2010, p. 261).Since risks are about threats that may occur in the future, they can only ever be imagined.Beck continues: 'In images such as these . . .geographically remote spaces become literally perceptible, "knowable" places of possible concern and action' (Beck, 2010, p. 261).
Riskscapes then, as Roy indicates by referring to them as territories, are actual geographical places, but are infused with multiple layers of meaning by different sets of actors.Moreover, and importantly, in contrast to places, territories are subject to governing (Roy, 2015).Coastal Bangladesh, due to being understood as 'a zone of climate crisis' (Paprocki, 2019), can be understood as such a riskscape (Eilenberg & Cons, 2019).Based on the idea of risk as anticipatory, Kasia Paprocki (2019) has shown that coastal Bangladesh is one 'perceptible space', as Beck dubbed it, that is in line for 'anticipated ruination': anticipated because the area will disappear underwater or otherwise be destroyed by climatic forces.In this sense, coastal Bangladesh represents a territory where risk must be managed through disaster risk reduction and climate change adaptation (Cons, 2016).Paprocki (2019) has illustrated how the anticipated ruination of coastal Bangladesh becomes a self-fulfilling prophecy when the transformation of agricultural land into shrimp-exporting zones under the label of climate change adaptation leaves large areas uninhabitable.Hence, the imagination has material consequences.

Subjects of risk
A central aspect of the creation of riskscapes, according to Roy, 'is the making of subjects: subjects of risk' (Roy, 2012, p. 139).Such subjects of risk, she continues, 'are deeply gendered' (Roy, 2012, p. 140).Roy (2010) builds on Chandra Mohanty's (1984) critique of the idea of the 'Average Third World Woman' which captured how women in the Global South have tended to be understood as a homogeneous group with similar challenges and desires, often with reference to their status as universal victims.Roy has noted that there is 'a very specific way' in which development after the turn of the twenty-first century 'has come to be focused on poor women' (Roy, 2010, p. 79).In this way, she argues, development continues to 'hinge' on the idea of the 'Third World Woman', first as the primary victim, then as 'the icon of indefatigable efficiency and altruism' (Roy, 2010, p. 76).In both cases, women in the Global South are seen as 'different' from the norm of what constitutes 'women', and thus have to be dealt with separately (Trinh, 1989).Therefore, they must be identified, mapped and described.Such 'dividing practices', as Michel Foucault has pointed out, are useful guides to understanding processes of normalization; one should look for who is being treated as different (2000b, pp. 326-329).As sociologist Loïc Wacquant (2008) has shown, and as I will return to later in this article, such dividing practices are also closely associated with territory.
In the context of climate change adaptation and disaster risk reduction, the central question is often: who is at risk?The idea that some people are more at risk from disasters than others is closely related to the idea of vulnerability, a well-known concept in disaster and development studies.Vulnerability was launched in disaster and development thinking as a counterpoint to the view of disasters as exclusively caused by external forces.Rather, political ecologists who advocated the idea of vulnerability argued that disasters are the outcome of external hazards meeting social vulnerabilities (Hewitt, 1983;Wisner et al., 2004).However, by becoming a mainstream concept, vulnerability has also taken on a multitude of meanings and uses.As most development initiatives must now be translated into the 'metacode' of climate change language (Dewan, 2021), established ideas of vulnerability from development and disaster studies are now being translated into climate change adaptation efforts.Here, old meanings are mixed with new ones.The outcomes of such translations are important, because the ways in which concepts are understood 'influence the questions asked, the knowledge produced, and the policies and responses that are prioritized' (O'Brien et al., 2007, p. 74).
An influential definition of vulnerability is to be found in the book At Risk, by Ben Wisner Piers Blakie, Terry Cannon and Ian Davis.Here, vulnerability is defined as 'the characteristics of a person or group and their situation that influence their capacity to anticipate, cope with, resist and recover from the impact of a natural hazard ' (2004, p. 11, my emphasis).While the original idea of vulnerability thinking was to bring in social context, which was also the point made by Wisner and his colleagues, definitions such as this one led to a search for the most important social characteristics of vulnerable individuals, groups and communities, often emphasizing factors such as gender, age, race and socioeconomic status (Cutter et al., 2003, p. 246).While the need to identify factors that contribute to vulnerability is self-evident, a focus on the characteristics of individuals and groups has often led to the unintended essentialization and subsequent victimization of individuals who tend to be highlighted as vulnerable (Cannon, 2008;Fordham, 1999).The problem is that such characteristics are often applied a priori, foreclosing the need to investigate the social context in any given case.This subsequently shifts the focus from circumstances to individuals (Resurrección, 2017).
In short, in order for people who are vulnerable to risks to be aided and protected, they need to be identified and studied.Such knowledge production about population groups has an association with Foucault's (2000a) concept of 'governmentality', which describes how, in order for governments to govern their population, they need knowledge about it.This collection of information, Foucault points out, does more than just produce knowledge about someone.By making people 'knowable', knowledge produces new groups of human subjects who can be targeted for state interventions through the 'objectification' of the subjects in question (Rouse, 2005, p. 100).
Partha Chatterjee (2004) has brought the idea of governmentality into the context of earlier colonial states.In these states, Chatterjee argues, 'populations had the status of subjects, not citizens ' (2004, p. 37).In the transition from colonial to postcolonial states, a dual political system was created, Chatterjee continues, in which a minority, the elite, was granted full citizenship in line with the newly written constitutions, while the rest of the population remained subjects in the sense that they were never able to claim the full rights of citizenship.
Yet, at the same time, because these people also reside 'within the territorial jurisdiction of the state, they have to be both looked after and controlled by various governmental agencies ' (2004, p. 38), or 'by nongovernmental -if necessary, international -agencies' (2004, p. 47).This has given rise, argues Chatterjee, to a parallel 'political society' which 'secures legitimacy not by the participation of citizens in matters of state but by claiming to provide for the well-being of the population ' (2004, p. 34).This is achieved by means of 'an elaborate network of surveillance through which information is collected on every aspect of the life of the population that is to be looked after ' (2004, p. 34).Hence, after independence, the colonial states were 'overtaken by the developmental state which promised to end poverty and backwardness' by means of 'the latest governmental technologies ' (2004, p. 37).Earlier mappings of the population used in the colonial project now 'often entered the field of knowledge about populations -as convenient descriptive categories for classifying groups of people into suitable targets for administrative, legal, economic, or electoral policy ' (2004, p. 37).In short, protection and control are connected.
Below, I unpack such linkages between protection and control by drawing upon Roy's ideas of riskscapes and the female-gendered subject of risk.Firstly, I show how assumptions about women's vulnerabilities construct women as subjects of risk who must be protected by means of climate change adaptation projects.Secondly, I use Roy's concepts to explore how ideas of women as subjects of risk are translated into a concrete climate change adaptation project in coastal Bangladesh.I depict how gender relations and the territories in which the women at risk live intersect and mutually shape each other, producing riskscapes and female-gendered subjects of risk.Before I turn to this analysis, however, I will describe the data material upon which this article is based.

Methods
This article draws upon ethnographic fieldwork conducted in two phases.Firstly, in Dhaka from October 2019 until February 2020, and thereafter in the district of Barguna in coastal Bangladesh from February to March 2020.The material from Dhaka consists of 18 semi-structured interviews with experts and aid practitioners who, in various ways, work with questions of climate change, gender and development in Dhaka.They are hired by donors, universities, think tanks and national, international and intergovernmental development organizations.Some are foreigners and others are Bangladeshi nationals.While the interview guide was adapted for each interview based on the participant's expertise, all the interviews included questions about how gender considerations are understood and included in climate change adaptation in Bangladesh.
In addition to the interviews, I carried out participant observation during 12 different consultations, conferences, seminars, internal training sessions and meetings related to the same topics as the interviews.I attended the events within the same period as the interviews, and many of the interviewees were also present.Such events can be understood as communicative gatherings (Nyqvist et al., 2017, p. 5), which make visible 'what is usually taken for granted' (Hammersley & Atkinson, 2019, p. 112).With a few exceptions, all the interviews were recorded, while I relied on notetaking during observations.
The purpose for the first phase of fieldwork in Dhaka was two-fold.First, I wanted to understand how ideas of gender were understood and incorporated into climate change adaptation conversations among experts in Dhaka.A second goal was to identify a project to follow that would allow me to further understand how the meaning produced in Dhaka translated into practices on the ground, and how gender considerations were included in this work.As such, the two phases of fieldwork are interlinked.
My stay in Dhaka gave me important insight into who the main actors were, where they were mostly working, and the main focuses of the projects that were being carried out in Bangladesh at that time.After a while, a few projects that I came to know through this time stood out as especially relevant, in that they addressed climate change adaptation and claimed to have a clear focus on gender.The project manager of one of these projects, whom I had both interviewed and met at a conference in Dhaka, agreed to accommodate me in visiting one of his project's field sites.
Thus, after the months spent in Dhaka, I also conducted ethnographic fieldwork in an area located in coastal Bangladesh, to follow the climate change adaptation project that I had come to know about in Dhaka.While the interviews in Dhaka were conducted in English, I relied on a translator while in Barguna.I'll call him Ahmed.Ahmed and I spent three weeks in this community before the coronavirus pandemic cut the fieldwork short.
We travelled around the area and spoke to people mostly in the informal settings of village teashops or people's homes, but I also conducted recorded interviews, primarily with development workers and government officials.For most of our conversations in Barguna, I jotted down notes as we spoke and wrote out my field notes upon my return to the guesthouse in the evenings.The few recorded interviews were translated and transcribed by Ahmed upon our return home.
The material has been coded thematically in several rounds, where a theme 'represents some level of patterned response of meaning within the data set' (Braun & Clarke, 2006, p. 82).The first round identified 'vulnerability' as a central category, which overlapped with talk about 'gender' and 'women'.During the second round of thematic coding, I sought to unpack this category, in order to better understand how the ideas of gender, vulnerability and risk were interconnected in the data material.
Turning now to the analysis, I start out by describing the climate change adaptation project that I visited, before discussing how ideas of gender and vulnerability played out and interacted in the implementation of the project and in its identification of beneficiaries.

The climate change adaptation project
The project field site for what I will here term the Coast Project, which Ahmed and I visited, is located at the southernmost tip of the coastal Barguna district, in an upazila 1 situated between three rivers and facing the Bay of Bengal.This location exposes the area to cyclones, tidal flooding and soil erosion (Torikul et al., 2017).The area was 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.The incident also put the area on the map of relief organizations, many of which came to help the community rebuild.The aftermath of the disaster is still present here, with people missing family members, recalling lands lost, and still working to reconstruct the life that was.This is the riskscape within which the Coast Project was situated.When Ahmed and I arrived in Barguna, we were met by Feisal, a middle-aged man living in one of the villages in this project area who was working as a local consultant for the Coast Project.Feisal introduced us to the project's activities and took us around the area for much of the time while we were visiting.He explained that he had been hired as a consultant on this project due to his earlier work with other organizations.Hence, this project must be understood as a continuation of earlier relief and development work in the area, and several of its beneficiaries had also been beneficiaries of other projects with which Feisal and other development workers had been involved.As Feisal noted: 'The NGOs are doing their independent work, but the poor people, they are the same' (Barguna, 1 March 2020).This indicates that the project interacted with already existing relations in the project area, a point to which I will return.
Like many other climate change adaptation projects currently underway in Bangladesh, the Coast Project was overseen by the Ministry of Forest, Environment and Climate Change and the underlying Forest Department had the coordinating role and overall responsibility for overseeing the project's implementation.The project involved multiple stakeholders, including several state departments and the local government.Additionally, an intergovernmental organization, the one for which Feisal was working, was responsible for channelling the project's funds from the donor, the Green Environmental Facility (GEF), and offering technical assistance to the government.Feisal explained that, in addition to him, two other people from the same organization were involved in the project.The district coordinator, Tasneem, was permanently employed and was the regional project lead on behalf of the organization.She was currently away.In addition, Naomi, a high-school student, was volunteering for the project.She was responsible for carrying out the survey for beneficiary selection.
The Coast Project was seeking to protect and strengthen the coastal greenbelt that stretches along the southwest coast of Bangladesh.The mangrove forest functions as a natural protection against cyclones and high tides, and its preservation is 'seen as an important strategy for reducing the vulnerability of coastal populations to climate-related hazards in Bangladesh' (Inception Report). 2 Central policy documents have thus identified strengthening this greenbelt as the 'number one adaptation priority for coastal areas of the country' (Inception Report).However, the inception report reads, the mangrove forest is under constant threat of deforestation because many people depend on it for their livelihoods.Thus, in addition to protecting people through protecting the greenbelt, the Coast Project also aimed to create synergies by involving the local population who depend on the forest for their livelihood.The project's inception report states that 'risk reduction in coastal areas of Bangladesh can only be achieved if the maintenance of protective greenbelts is connected to tangible livelihood support' for the communities in question.In other words, the idea was to give local people alternative livelihoods so that they would not cut down the forest.'Thus', the inception report reads: the programme will reduce the vulnerability of extremely poor and highly exposed coastal communities to existing climate variability and future climate-change-related risks through a combination of measures that will strengthen both people's adaptive capacity and the long-term resilience of coastal greenbelts in Bangladesh.(Inception Report, p. 4)

'The vulnerable'
Echoing the above statement from the project's inception report, the project manager of the Coast Project explained: 'The main objective of the project is to reduce the vulnerability of the coastal climate-vulnerable people' (Dhaka, 28 October 2019).We were sitting in his office in Dhaka, where he gave me a presentation of the project he was leading.There is, however, a slight difference between the statement in the inception report and the project manager's explanation, which is worth noting: He translated 'vulnerability' into 'the vulnerable'.This is a common way of talking about potential beneficiaries among the expert community in Dhaka and deserves closer attention.
In a similar manner to the manager of the Coast Project, several of the experts whom I interviewed in Dhaka referred to 'the vulnerable people' as a given group, often without clarifying who, exactly, they meant: 'The most vulnerable will be affected the most' (Dhaka, 1 December 2019) were the much-rehearsed opening words of a consultation workshop held in Dhaka concerning how to channel funds towards climate change efforts.Further clarifications were not necessary, because everyone knows who the vulnerable people are.Yet, how vulnerability is understood is important because its given meaning influences how vulnerability is sought out and dealt with (O'Brien et al., 2007).As noted in the first half of this article, vulnerability is quickly translated into an identifiable group with a given set of characteristics: 'the vulnerable'.Who then, are 'the vulnerable' in the context of climate change adaptation efforts in Bangladesh?

Geographical vulnerabilities
You know one of our [project]  Firstly, as the quote above illustrates, vulnerability to climate change was strongly associated with place, which meant that 'the vulnerable' must be sought out in particular areas, most often in coastal Bangladesh.In relation to climate change, this takes on a new meaning, where 'territories of poverty' (Roy & Crane, 2015) -the traditional development target -have to be matched with territories of climate risk.Hence, territory, vulnerability and risk are interlinked.Coastal Bangladesh represents a riskscape that is 'in need of intervention' (Eilenberg & Cons, 2019, p. 9) -a territory that needs to be mapped and managed.Such mapping takes the form of a risk assessment.A development practitioner in Dhaka explained the process in the following way: 'So we identify the geographically vulnerable people and then identify the vulnerable pockets of that particular location.And also we implement the . . .vulnerability checklist' (Dhaka, 5 December 2019) to identify vulnerable individuals.I will now turn to who these individuals are.

Gendered vulnerabilities
So, this strategy . . .[is] providing assistance to the people who are really in need . . . to protect them from certain risks.Especially those who are most vulnerable.So here, we're looking at children, the elderly, people with disabilities, and I don't like to say that women are vulnerable, but they are considered vulnerable in the context of Bangladesh.(Dhaka, 20 February 2020) This is a common summary of beneficiaries in climate change conversations in Bangladesh.What is interesting for our purpose here is how 'women' tended to be referred to.Despite the sweeping categorizations of beneficiaries, this research participant did not of course have in mind half of Bangladesh's population when referring to 'women'.While the term 'women' tended to be used this way, without any specificity, it in fact implied a very particular group of women.Just as 'everyone' knew who the vulnerable were, 'everyone' knew which women were being referred to.It thus comes as little surprise that the two categories often overlap.It is of interest, then, to take a closer look at how these women were described in order to gain a better sense of exactly which group of women the research participant had in mind.
'If we consider the women and children, and especially the girl child and women are much more vulnerable, and those are at risk' (Dhaka, 12 December 2019), I was told.'So, you will see the divorcees, deserted, who have children' (Dhaka, 14 November 2019), one research participant summarized.'Women who have been abandoned by their husband, or who are in a really precarious situation' (Dhaka, 20 February 2020), another research participant explained.A third noted, 'abandoned by their husband with their children, or divorced.But mostly abandoned by their husband' (Dhaka, 12 November 2019).In other words, a recurring category throughout the interviews in Dhaka was women who were living without a male head of household.These types of descriptions must be seen in relation to purdah norms, where women's honour is exchanged for men's provision and protection.This forms an agreement dubbed 'the patriarchal bargain' (Kandiyoti, 1988), which characterizes gender norms in Bangladesh.In practice, however, this bargain has never included all women -not all families have been able to afford to keep their women at home (Hossain, 2017).In other words, how purdah is performed in Bangladesh varies with class, and women from low-income classes working outside have often been met with more acceptance.However, such transgressing of purdah is often described as unfortunate and hence associated with stigma (Banks, 2013;Jennings et al., 2022;Lata et al., 2021).
In short, then, the women characterized as vulnerable by the research participants in Dhaka were described as facing both economic hardship and insecurity of their person.As one research participant summarized it: 'The main challenge is safety.The main challenge is actually that they're not safe' (Dhaka, 27 October 2019).Therefore, as another research participant concluded: '[We] have to ensure their protection' (Dhaka, 14 November 2019).
The perception of women's vulnerabilities to risk in climate conversations in Dhaka mirrors messages from the emerging gender and climate change literature, which again builds on older literature from its sister fields of gender and development and gender and disasters.Within these sister fields, women have been identified as more vulnerable than men due to their greater impoverishment and dependence on natural resources.These points are closely related, because women's lower income forces them to be more dependent on natural resources than men, who are more likely to earn a wage.Furthermore, women's impoverishment is often linked to their social status, with femaleheaded households assumed to be worse off than male-headed ones.In short, women are identified as the primary victims of climate change in terms of both increased poverty and reduced physical security because climate change is expected to affect impoverished individuals more due to their dependence on natural resources (e.g.Ahmed & Fajber, 2009;Cannon, 2002;Dankelman, 2010;Denton, 2002;Terry, 2009).
It is worth recalling Chatterjee's point here, that a central aspect of the developmental postcolonial state is the focus on protecting particular population groups.The vulnerable are often associated with a particular group of women; women who are not under the protection of the patriarchal bargain and who live in places that are prone to environmental hazards.
Recalling the project manager of the Coast Project's conflation of 'vulnerability' into 'the vulnerable', we can see here that a translation occurs, in which social vulnerability is transformed into a personal characteristic.Thus, a process of what Foucault called objectification takes place, whereby 'the vulnerable' become subjects, or cases, for climate change adaptation interventions and, in the same process, are brought together as a group with similar characteristics.Women become subjects of risk.
I will now turn to the gendered riskscape of Barguna to look more closely at where some of these vulnerable women are to be found, and to explore how gender and place mutually shape riskscapes and subjects of risk in coastal Bangladesh.

Gendered riskscapes
In coastal Bangladesh, the embankment, which runs through the landscape alongside the riverbanks and coastal shores, is often an integrated part of villages.While it protects the villages against high tides and flooding (although this can be debated, see Dewan, 2021Dewan, , 2022)), it often also functions as the main road and thus constitutes a central space in how villages are structured.Embankments are commonly built at a certain distance inland to avoid erosion, which means that there is often a strip of land between the embankment and the shore.As people are not supposed to live on the outside of the embankment, it is not possible to buy land there, and this land belongs to the government.Such government land is called khas land, sometimes translated as 'public land'.What this indicates is that khas land is often utilized for various purposes by people living in a given area, and it is often the case that the strip of land outside the embankment becomes an area of available land where the landless can live.Thus, the area outside the embankment being populated by people who have no land to live on is a common sight in Bangladeshi villages.Hence, the embankment becomes both a physical and symbolic class barrier, a marker for inside and outside.It is also a constant reminder of the anticipated risks of erosion, cyclones and high tides, thus becoming perhaps the clearest reminder of the riskscape within which these villages are situated.In this way, the division manifested by the embankment is also important for the distribution of risk.Confirming this division, Feisal explained that the Coast Project made two types of interventions: one set for the inside of the embankment and one set for the outside.Here, I focus on the outside.

Identifying the 'helpless'
As with most other development and adaptation projects, in the Coast Project beneficiaries were selected on the basis of a risk assessment.Such risk assessments are carried out through detailed questionnaires, based on given criteria.The meaning of 'vulnerability', the project manager noted, was not defined in the documents, but was reflected through the combination of criteria by which beneficiaries were selected.Since this was primarily a reforestation project, criteria were primarily related to forest dependency and landlessness, which are both closely related to poverty.He noted: 'the forest-dependent people, the community people, they are poor, and also they are vulnerable'.'Actually,' he continued, 'in the coastal areas you will see that these forest dependent households are mainly women. . . .In the coastal context we are addressing mainly the women.'Explaining how the project was emphasizing women among its beneficiaries, the project manager continued: [W]e are trying [to target] women headed households first.You see?In the criteria it is clearly mentioned.So, we are trying to find out the women headed households . . .suppose in one upazila one hundred households will be involved in our project. . . .So, out of one hundred, how many women headed households are there?So, in that way women are getting more priority.(Dhaka, 28 October 2019) These women, the project manager noted, were 'the vulnerable, women's households, disabled women's households' (Dhaka, 28 October 2019).Similarly, the elected government official for the area, who functioned as the leader of the local committee that had been established to oversee the project at the local level, listed 'low income, women, divorced, distressed, and beggars' as important indicators for beneficiary selection (Barguna, 9 February 2020).When I asked him what he meant by 'distressed', he outlined the example of a woman abandoned by her husband.Capturing this sentiment, Naomi, the volunteer who was carrying out the assessment of potential beneficiaries, explained her task in the following way: 'I need to select the ones who are the most helpless' (Barguna, 8 March 2020, my emphasis).Then she recalled the last beneficiary evaluation she had conducted, in which a woman and mother of three had been abandoned by her husband, and his parents were too poor to take care of her and her children.'She was more helpless', Naomi concluded.
It was not only Naomi and the Coast Project taking part in the identification and evaluation of who met the criteria to be beneficiaries of the project.Villagers also evaluated each other and made judgements about who should receive project assistance.Naomi explained that sometimes villagers would come up to her and suggest people she should consider.Whether the 'right people' were receiving aid was also a topic that was often raised in our conversations with people in the village.Many had opinions about which organizations were fair and which were corrupt, opinions that were based on whether people in 'real' need were being assisted or whether aid went to people with connections to the holders of power.
My point here is not to engage in this evaluation of whether the right people were being included on the list.Rather, I want to illustrate how aid is an integrated part of village relationships, and hence discussions about who was more vulnerable and at risk was a topic that engaged everyone.'Target-group terminology' interacts with, and co-constructs, daily life in the social context in which it is put to use (Wood, 1985).Like Naomi, the villagers I spoke to also used the word 'helpless' to describe women who were in a vulnerable position (Evertsen, 2021).As one man put it during a conversation in a teashop, 'Women who are helpless are the ones who've lost their husband and have to go to the river to fish' (Barguna, 3 March 2020).Hence, in line with gendered norms of purdah, it then became the Coast Project's responsibility to take over the role of the male provider and protector.In the following, I elaborate on this point.

Women in masculine spaces
A central aspect of purdah norms is the physical separation of men and women in society.While women are responsible for the home and children, men are expected to provide for their household (Kabeer et al., 2011).This creates a spatial divide under which public space is seen as masculine and private space as feminine.As feminist political ecologist Farhana Sultana has noted, purdah is socio-spatial ( 2009).This socio-spatiality became visible in the division between the inside and outside of the embankment.
Feisal pointed to a makeshift settlement that stretched along the outside slope of the embankment.'These are the landless', he said, before leading the way down on the outside (Barguna, 10 February, 2020).During Sidr, all the houses on the outside of the embankment, he explained, had been 'washed away by the water' and most of the land had disappeared as well, such as the land on which the school used to stand.After the cyclone, multiple aid agencies came to the area to provide relief for the people who had lost their homes.After some time, the organizations also began to distribute materials to rebuild houses.However, there were no other places for the landless to rebuild their homes than on the outside of the embankment, on the khas land.In other words, the relief efforts contributed to setting up houses on the outside of the embankment, and people who had lost their homes were again exposed to the next cyclone.A previous colleague of Feisal, who worked for another aid organization that was introducing climate change adaptation projects into the area, explained that the government could not approve of people living on the outside because it would be held responsible if anyone were to lose their lives.However, the government did not have any alternatives to provide for these people.Since 'the government won't do anything to evict them, but also won't support people living on the outside of the embankment', he concluded, 'what we do as development workers is to meet their basic needs as far as possible' (Barguna, 8 March 2020).
Feisal introduced us to two different households on the outside of the embankment.They were beneficiaries of both the current and previous projects with which Feisal had been involved.Much in line with the explanations of beneficiary selection laid out above, both households were female-headed, in the sense that women were the main income earner of the household and were receiving livelihood support from the Coast Project.In the first house lived Ayisha with her disabled husband -a fisherman who had been stung by a poisonous fish and was now unable to provide for his family: their children, Ayisha's mother and her husband's mother.In the other house lived Sumaiya, an 'abandoned' woman, with her son and his wife, and her grandson, whose mother lived in Saudi Arabia.While Ayisha's family had been given the materials for their house immediately after Sidr struck, Sumaiya had received her house from a project for which Feisal had been working in the aftermath of Cyclone Aila, which hit two years later.Sumaiya explained that she had been living in other people's houses for the previous 17 years.Since the families she had been living with had lost their house to Cyclone Sidr in 2007 and again to Cyclone Aila in 2009, she had occupied this land on the outside of the embankment together with a group of other families and was now determined to stay there, despite it being a 'risky' place to live (Barguna, February 10 2020).
It was explained to me on several occasions that most of the families residing on the outside of the embankment were fishermen's families who found it more practical to live close to their livelihoods -the water.Ayisha and Sumaiya also fished for their livelihood.However, fishing is a man's responsibility.Women who live on the outside of the embankment are thus inhabiting masculine spaces.Furthermore, women who live in this masculine space are not under the protection of the patriarchal bargain.Thus, the fact that women lived here without a man who could provide for or protect them was often referred to as an exception, and the 'helpless' woman living on the outside of the embankment is thus on the outside of society in both a symbolic and a highly material sense.Confirming the break from purdah symbolized by the women living on the outside, Feisal noted that, in most cases, if possible, the women's families will bring them back inside.Hence, the outside of the embankment can be understood as a territory of stigma (Wacquant, 2008) and women living on the outside represent 'female bodies out of place' (Sultana, 2009, p. 431; see also Cresswell, 1996).'Female bodies that are seen to be "out of place" . . .are often thought to be in need of greater control' (Domosh & Seager, 2001cited in Sultana, 2009, p. 431).They must be governed.Here, this is achieved through the intervention of the Coast Project taking on the role of the male protector and provider.What such interventions should look like is, however, a much-debated question, as I will now turn to.

Keeping women in their place
If we can train her.To give her suitable livelihood options.This way we can ensure her security, ensure her [security] and make her resilient.I think this is the only, this is the better option.We are talking about empowerment, but empowerment doesn't mean to destroy her dignity or something like this.Empowerment, she can stand alone.If I was a policymaker, I would definitely do something in her place, where she can stand. . . .So, she should be there, because this is her comfort zone.So, we should do something so that we can keep her in her place.(Dhaka, 2 December 2019, my emphases) In the interview from which the above excerpt is taken, we were discussing how climate change adaptation projects can benefit women without adding to their already heavy workloads.The excerpt articulates some of the dilemmas with which feminist development researchers have struggled for decades.Women who do not have a male breadwinner and protector in their household are often identified as the most vulnerable, because the lack of a male breadwinner correlates with low income and hazardous living conditions.As women's hazardous living conditions are often associated with low income, a frequently proposed solution is to train women in income-generating activities (Evertsen, 2022).However, as illustrated by the explanation about how the most 'helpless' women are the ones who have no husband to provide for and protect them, and who have to go to the river to fish, women working outside is controversial.The crux of the problem is how to provide women with an income without simultaneously contributing to their stigmatization.On the one hand, providing women with income opportunities that adhere to gender norms can contribute to supporting patriarchal systems under which women suffer.On the other hand, aiming to give women 'work' that requires them to leave the home may turn out to be equally violent, because they often face a combination of stigma and dire working conditions (again made possible by the stigma and low status of these jobs) (Evertsen, & van der Geest, 2020;Jennings et al., 2022;Siddiqi, 2000;Wright, 2006).This is what the expert is describing when she suggests that climate change adaptation projects 'should do something so that we can keep her in her place', rather than 'destroying her dignity'.
As described above, the Coast Project also had a central goal of creating 'sustainable livelihoods' for its beneficiaries -one must give forest-dependent women alternative livelihoods so that they do not cut down the forest.When I asked Feisal what income opportunities were available to women in the area, he noted: 'In this area women actually don't have many opportunities to work (Barguna, February 10 2020).'There were some government initiatives, he reflected, where women could dig mud when the embankment needed fixing.But this was only for short periods and did not require many people, he explained.Besides, women would not want to do this work, because it meant they would have to work outside, which would give them a bad reputation.At the same time, substantial efforts were being made to find the most vulnerable women.Hence, the 'empowerment of women' for which many projects aim, is seldom about structural change, but rather about keeping women who occupy the margins of society afloat.As Feisal put it: 'We're giving her assistance so she can stand on her [own] two feet' (Barguna, 1 March 2020).

Concluding discussion: Normalizing relations of risk
Coming to the conclusion of this article, I will consider what the assumptions underlying the idea of women as vulnerable and at risk, and the assistance given to women to help them 'stand' based on this idea, do.I want to return here to Roy's notion of riskscapes as a 'territory'.'Territory', Roy argues, is what Foucault termed 'a "political technology" through which the "spatial order of things" comes to be normalized and perpetuated' (Roy, 2015, p. 3).Looking again at the main criteria for beneficiary selection, which is meant to capture ideas of vulnerability in this project area, being landlessness, forest dependency -which are both closely associated with class -and gender, it becomes visible how beneficiaries were selected on the grounds of their gender, as well as on the territory they inhabited and used for their livelihoods.
The land on which these people live is thus a place to be governed, which by extension also involves a governing of the people living there.The territory with which I have been concerned in this article is the land on the outside of the embankment.We have seen how the village in Barguna was divided between the inside and outside of the embankment, where the outside can be regarded as a territory of stigma.Territories of stigma, Wacquant has noted, are already 'branded spaces' or 'problem areas' in which the perception of place 'contributes powerfully to fabricating reality' (Wacquant, 2008, p. 171).There are two 'problems' with women living on the outside of the embankment: they are at risk of being 'washed away' by environmental hazards, and they are without the protection and support of a man.
The 'branding' of the outside of the embankment is important because it not only brands the territory itself, but also the people living there.This branding interacts with the social structures of the village, and confirms the idea of certain women as 'vulnerable', as the experts in Dhaka put it, or 'helpless', as Naomi and other villagers put it.Hence, it contributes to normalizing the gendered relations of risk that contribute to making women vulnerable in the first place (Wood, 1985).As initiatives are taken to aid women who are living on the outside, both in social and spatial terms, a normalization process takes place during which women who are not living in accordance with gendered norms are helped to do so.In this process, projects like the Coast Project take over the role of male protector and provider, further confirming gendered norms.As Chatterjee has argued, protection and control are interlinked -the need, expressed by both the expert in Dhaka and Feisal in Barguna, to help women to 'stand', or 'to keep her in her place', has undertones of control.Here, we see how ideas of gendered vulnerabilities shape climate change adaptation practices.
Another normalization process that is taking place concerns the risk of place.Beck (1992) noted how risks tend to be socially distributed within given limits of the tolerable; that is, the socially accepted levels of risk.Thus, he asked, where are risks channelled and where is the line of tolerance drawn?Could it be that, by helping women to 'stand', the project contributed to keeping the risks they face by living on the outside of the embankment within the given limits of the tolerable?As such, to 'keep women in their place' took on a quite literal meaning in Barguna, where, interestingly, the project aided people to remain at risk from the elements.By providing the minimum necessities to survive there, a fine balance was established whereby people could survive on the outside, thus also keeping these women's situation within acceptable levels of risk.
In Barguna, the idea of protection brings together ideas of environmental hazards, poverty, gender norms, and places of risk and stigma.As Roy has noted, it is clear that riskscapes are deeply gendered and that riskscapes and female-gendered subjects of risk mutually construct each other.
At the beginning of this article, I argued -relying on the work of Sherilyn MacGregor (2009) and Ulrich Beck (2010) -how sociologists can contribute to better understandings of the social effects of climate change by investigating discursive ideas as these are important in shaping social practice, including responses to climate change.MacGregor further highlights gender as one such discursive idea to be investigated in relation to climate change.In this article I have shown how discursive work to identify the most vulnerable women also constructs them as subjects to be targeted for climate action.This subsequently also contributes to shape climate change adaptation projects in coastal Bangladesh in ways that contribute to confirm existing ideas about gender.These insights can hopefully contribute to expand the current debates on gender and climate change to include not only how women are more vulnerable than men when faced with negative effects of climate change but also how gendered ideas of women as vulnerable are produced and acted out in efforts to combat these very effects, with material consequences to the lives of people targeted by such projects.

Funding
The author received no financial support for the research, authorship, and/or publication of this article.

Notes
1.An upizila is an administrative region in Bangladesh.Bangladesh is divided into eight divisions, which are divided into several districts or zilas, which are again divided into multiple sub-districts or upizilas.2. To ensure the privacy of research participants, the full details of the project's inception report have not been included in the reference list.
areas, it can already be under water. . . .It's a coastal island . . .I visited there last year. . . .This year I visited again, and I did not find [it], because it goes under water. . . .It's only for climate change reasons.(Dhaka, 2 December 2019)