Not Just Dis-placed and Poor:

How Environmentally Forced Migrants in Rural Bangladesh Recreate Space and Place under Trying Conditions


Doreen Indra


[The current text here was published as a book chapter of the same name in Elzbieta M. Gozdziak and Dianna Shandy, eds. (2000) Rethinking Refuge and Displacement. Selected papers of Refugees and Immigrants Volume VIII. Washington: American Anthropological Association, pp. 163-191. This volume is available from the American Anthropological Association.]

A .pdf version is available for download.



While development, aid, and allied research discourses on flood and riverbank erosion in Bangladesh highlight the poverty and powerlessness of those affected, this is not the only way in which their lives can be usefully characterized. Most people displaced by riverbank erosion in Kazipur Upazilla, Sirajganj District, Bangladesh, are already poor and disempowered before being uprooted by the shifting local channel of the Jamuna River. Thereafter, a significant proportion face even bleaker prospects. Nevertheless, those who have resettled locally have done much to maximize their chances within the limits of powerful natural, economic, social, and political constraints. Having experienced high rates of both general social change and riverbank erosion for more than a generation, they are adept at reconfiguring “tradition” to better address new circumstances. As forced migrants, they have been resilient, resourceful and innovative in ways that their conventional representation as passive victims of natural and social forces outside their control would not suggest.


 In this paper I present findings from my ongoing study of people displaced by riverbank erosion in Kazipur Upazilla, Sirajganj District, Bangladesh.(1) I address several issues concerning how social science research has represented environmentally displaced migrants in Bangladesh. First, I identify several factors that have ensured that discourse on people displaced by riverbank erosion identifies them almost exclusively with crisis periods, characterizes them as far more homogeneous than they are, and portrays them as passive victims of natural and social forces outside their control. The symbolic impact of these conventional images of helpless, powerless forced migrants is a primary determinant of how scarce aid and development resources are allocated. However, I suggest that this characterization is neither “natural” nor more “accurate” than alternative representations that do not fit this particular mold. In the following section, I describe the setting. Drawing on my fieldwork in Sirajganj District, I highlight dimensions of river-eroded people’s lives that are dissonant, but not necessarily contradictory, with their conventional portrayal as powerless, passive and homogeneous. Next, I identify many individual and collective “pre-adaptations” Kazipur people have made to rapid change and to erosion. Finally, I illustrate how people exercise individual enterprise and strategic choice concerning resettlement, and demonstrate a richly varied range of adaptations concerning production and social reproduction under restrictive circumstances. I also show that there is a high degree of subsequent social and cultural connectedness within this highly variable population. In overview, I argue that while most forced migrants in Kazipur are profoundly disadvantaged, their lives are also imbued with the diversity, spontaneity, agency and power that is always characteristic of daily lived experience (Simone 1990:161).

The natural disaster and research discourses on flood and riverbank erosion

Flood and riverbank shifting have played a major role in shaping the physical, biotic and demographic landscape of Bangladesh for a very long time.(2) Because of the endemic and widespread nature of flood and riverbank shifting in the region, people all across what is now Bangladesh have a long history of adaptation to flood (Muhkerjee 1938; Rogers et al 1989). In some areas, the destruction of land due to riverbank erosion has been a critical part of daily experience for centuries, particularly on some riverine and delta chars (islands of depositional land) and adjacent areas. In many other locations, uprooting people from their home villages is a comparatively new phenomenon. Riverbank erosion has progressively affected more and more people as population density has increased over time. Current rural population density often exceeds 2,000 per square mile, and the human cost of riverbank erosion is incalculable. Roughly a million people a year are believed to shift residence, either as their homes and land are washed away or to take advantage of newly accreted land (Wiest 1988:3). While they are interconnected physical phenomena, riverbank erosion and flood are distinct, and each has unique consequences for those affected.(3) Local people in affected areas make sharp distinctions between the two. As one person said, “flood is terrible but it goes away and the land is still there. With erosion, nothing is left.”

Over the past twenty years, severe flood has been constructed within Bangladesh and internationally as an immense problem posing the potential for calamitous natural disaster. So thoroughly has flood been framed in this way by government, aid and financial organizations, and the media, that it now shares the iconographic spotlight in Bangladesh’s international image with older, more established images of extreme poverty and general disaster.(4) This representation of flood as a potential natural disaster has been powerfully and effectively advanced. In response, massive amounts of aid have been forthcoming over the past thirty years to develop protective infrastructure, or to deal with the physical consequences of flood. Such development projects date back to Pakistan times, when an extensive set of embankments were built along the Jamuna River. Thereafter, further government and foreign aid resources have been allocated for the construction of more embankments, flood-proof roads and emergency shelters. Periods of extreme flooding now quickly mobilize concern and emergency assistance.(5) Chiefly as a result of extensive world-wide coverage of the ravaging 1988 flood, an international and national effort was organized under the aegis of the Flood Action Plan to evaluate massively expensive engineering options to limit future floods.

As Adnan (1991) forcefully noted, research in support of the subsequent Flood Action Plan was obsessed with finding engineering fixes for flood and, secondarily, for erosion. Most socio-economic dimensions of flood and riverbank erosion were initially to be considered only inferentially and tangentially (see Hanchett, Akhter, and Akhter 1998:210-215), as in the Flood Response Study (Government of Bangladesh 1992) as it was initially framed, and in the Eastern Waters Study (Rogers et al 1989). Criticism of this highly technical approach eventually led to greater consideration of social factors in the Flood Response Study, but here too, concern with flood predominated (Hanchett, Akhter, and Akhter 1998).

Kazipur, Sirajganj, Bangladesh at flood
Dramatic scenes like this help make flood figure strongly in disaster discourse

Much national legislation, local government and folk practice has evolved to deal with land erosion over the years, particularly concerning the ownership and control of submerged and accreted land. However, riverbank erosion has only recently received significant recognition in media, development and research circles. Even today, development and research discourse remains dominated by engineering and physical science issues. The joint University of Manitoba (Winnipeg, Canada) and Jahangirnagar University (Savar, Bangladesh) Riverbank Erosion Impact Study (Rogge and Elahi 1990) was the first substantial research focusing on the sociocultural dimensions of riverbank erosion. Although the five year project formally ended in 1989, researchers continue to analyze project data.(6) Representations of flood have such great iconographic power that it is unlikely that riverbank erosion will ever attain a comparable level of salience and attention in development circles, despite its massive human impact.(7) There are, however, some signs that the human dimensions of erosion-generated displacement will gain increased significance in the natural disaster discourse and research.

Research discourse is of course an issue here. Shapiro (1987), Foucault (1972), Bourdieu (1984) and many others have drawn attention to how our research is embedded and formed within larger discourses, cultural constructs and practices.(8) This orientation provides us with effective ways of thinking about our situated roles as social scientists in both producing and politicizing knowledge. It also emphasizes that as individuals and networks of researchers, we possess agency and power, and consequently, responsibility in forming particular, often symbolically and politically potent images in our research (Fraser 1989). We have responsibility, because, as Bourdieu notes (1991:375-6), “the fact that he or she produces representations of the social world endowed with the authority of science [means that] the sociologist is engaged willy-nilly, whether he or she knows it or not, in the symbolic struggles for the imposition of legitimate principles of vision and division of the social world involving other specialists in symbolic production.” It is particularly important to keep this point in mind when doing research in a country like Bangladesh that is so heavily dependent on development and aid. Social science knowledge production in Bangladesh is usually initiated in the context of development or aid activities and their powerful symbolic metaphors, ideological associations, and influential “other specialists in symbolic production”—many of whom are not social scientists. Even those studies that are not specifically development or aid oriented are highly dependent on a research literature that is.

Rebuilding a flood control embankment, Kazipur, Bangladesh
Enormous resources are devoted to maintaining flood control embankments

These influences can be seen clearly in the small but growing body of research on riverbank erosion. Riverbank erosion results in enormous destruction of property and puts many of those affected in a precarious state. Simply to observe it evokes dramatic images of uprooting and loss resulting from an unequal contest between already disempowered people and an inexorable natural process. Such images are often amplified when framed by social research and development discourses concerning flood and disaster. Because the lion’s share of the pertinent development discourse on flood is generated by engineering, geology, hydrology and allied fields, research tends to objectify the people affected, reducing them to one of a large set of “variables”, to be considered in a search for a “solution” to the “problem” of flood (Adnan 1991; Zaman 1991b:16). Moreover, many elements of the natural disaster conceptual grid, present in flood discourse, have been transferred into riverbank erosion research. Human crisis in the face of massive geographical change underlies the latter. This has affected what is researched and what is not, and the way in which findings are clothed in associated meanings and symbols. The most powerful of such tendencies is too broad to consider adequately here: to see “the problem” almost immediately from a technical point of view, thereby marginalizing social, cultural and individual factors at the onset (Adnan 1991:11-14). Within the somewhat narrower context of sociocultural research these include:

1. A concentration on the problematic and immediate, but not in the sense of having “been there” when erosion occurred. Directly observed responses to immanent erosion are rarely observed, and longitudinal data is sparse.(9) Most research findings are based on after-the-fact reconstructions generated out of peoples’ recollections of what happened to them.

2. People being portrayed as relatively uniform in initial condition and post-erosion response—chiefly as landless, desperately poor and powerless. This deflects concern from peoples’ perceptions of their own situations and their complex and diverse survival strategies (Alam 1990b:7-8; Scudder and Colson 1982).

3. A strong suggestion that flood and riverbank erosion detach people from their timeless traditional roots (Muinul Islam 1992). Rather than making this a question for study, it is assumed that displacement has this effect and that people actually were “traditional” and “rooted” before being displaced (Malkki 1992:26-7).

4. A tendency to portray people chiefly as reactive rather than actively involved in processes of change.

5. An implication that these forced migrants are increasingly “nuclearized” and “individualized.

If these preconceptions are unquestioned at the onset of sociocultural research on riverbank erosion, the consequences are obvious. As seen in the evolution of many other “social problems” discourses, these and other images, concepts and anticipatory expectations will become naturalized and established as facts without ever becoming problematized (Bourdieu 1991). Once in place they will become axiomatic, and research built upon them will circularly confirm their validity. Little of the authorial voice of those people whose lives have been so characterized will be heard unless it fits this grid of expectation and the development of more multidimensional representations of their world and their lives will prove difficult.

In order to avoid contributing to this possible scenario here, I have taken the reactive tack of briefly describing one environmentally displaced population differently. I stress again that this is not to suggest that most of the people I consider are not desperately poor and politically weak. Stories could be told, and statistics provided, to follow the canonical grid and reasonably represent the majority as victims in an enormous tragedy, and as some of the most disadvantaged individuals in this comparatively poor part of Bangladesh. But without doing violence to the harsh facts of their situation, a different representation can be drawn that broadens one’s understanding of their circumstances in ways that allows one to more clearly see their fluid adaptability, courage and resourcefulness, their effective maintenance and elaboration of social, cultural and geographical connectedness, and their ability to transform cultural norms to better suit their changed circumstances and requirements for daily survival.

The locale

Located 75 miles northwest of Dhaka, the 138 square mile area encompassed by Kazipur Upazilla in Sirajganj District is one of the most erosion-prone in the country (Haque 1988:189).(10) It is a region that is considered poor by Bangladesh standards. The local economy is based on intensive, small-scale rice agriculture. Aid, chiefly as grain and cash paid for road and embankment work, also plays an important economic role. Upazilla officials and Wiest (1988: Table 5) estimate that over three-quarters of the local population is functionally landless; many of these people at some point lost land to the river.

Kazipur Union, one of the Upazilla’s eleven unions, is among the worst erosion-affected.(11) Official statistics estimate the population at 28,000 to 32,000, but fewer than 10,000 are currently in residence. Once, this was not so. In the early 1920s, the most westward branch of the multiply braided Jamuna river was four miles to the east. Since then, it has grown and inexorably shifted westward. Long ago it destroyed Kazipur Town, the region’s major marketing and administrative center. Rapid erosion continued along the west bank of this seasonally mile-wide channel. Between 1968 and 1981, the Union lost about 40 percent (roughly 2,000 acres) of its agricultural land base as well as the Upazilla headquarters and four villages (Haque 1988:189). The floods of 1984, 1988, 1990 and 1991 caused further devastation.(12) During my initial fieldwork (December 1989-August 1990), erosion, coincident with a flood covering nearly all arable land in the Union, took away several hundred additional acres and important public buildings in the Union’s sole remaining town. Severe erosion continued during the 1991 and 1992 rainy seasons, destroying more of the town and significant portions of two previously unaffected villages. My analysis of Bangladesh Water Development Board maps shows that at some points the bank line shifted westward a quarter mile during 1989-90, and a full mile during 1979-92. Although the 125 mile long Brahmaputra Right Bank Flood Control Embankment passes through the Union, its purpose is to stop annual flooding, not erosion. It periodically has been breached by the river only to be rebuilt further west, thereby destroying and marginalizing still more local land. Altogether, more than half of the union’s current residents have been forced to move during the past fifteen years, and less than four of the Union’s original eleven square miles of land remain. People in this much reduced mainland area now look across the Jamuna to a world of ever-shifting chars where adaptation to erosion has been chronic (Zaman 1988, 1991).


Displacees settled on a flood control embankment, Kazipur, Bangladesh
Displacees settled on a flood control embankment

Evidence of riverbank erosion is immediately visible in Kazipur, even to a casual observer. When I first did fieldwork there in 1989 the most obvious were more than eight hundred dwellings of river eroded people strung along the top of flood control embankments and the single pukka (paved) road that passed through the Union. Locally, these people were considered to be the Union’s poorest environmental migrants. This was also the view of REIS researchers.(13) People that were settling the embankments and margins of the main road were occupying highly visible and disputed space. They attracted much attention. Those on embankments lived physically separate from, and literally above, everyone else.(14)

Upon arrival in Kazipur, I quickly confirmed an earlier observation (Elahi and Rogge 1990:5) that only a minority of displaced people resided on these embankments. The rest were in one way or another “hidden” from view and possessed little iconographic value. Most displacees were sprinkled throughout local villages, paras (village sections) and clusters of households. Some better off households already owned, or subsequently bought or leased house plots. Hundreds of destitute others were living as uthuli—landless, partially dependent poor people who had been allowed to place their houses on the land of others rent-free (Indra and Buchignani 1997).(15) A majority of displacees over the last 15 years may have left the Upazilla altogether.(16)

Change as the normal state of affairs in Kazipur

The central underlying organizational “story line” about environmentally displaced migrants in Bangladesh depends heavily on a combination of two images: an image of temporally limited “natural disaster”, and an image of independent “peasants”, sharecroppers, and landless laborers who are its victims. The story is, of course, a tragedy. A strong suggestion that people were operating in a relatively timeless, quasi-traditional state to which they had adapted over centuries highlights this tragedy.(17) They are confronted with an unfamiliar agent of enormous change—a destructive natural phenomenon completely outside their control. Afterwards, they are left impoverished and disorganized, in socio-economic situations essentially out of their control, and with few cultural or historical insights about what to do. While this (paternalistic) representational grid may be symbolically effective in eliciting concern among providers of acute aid, it does not contribute much to a deeper understanding of the lives of people who are living under very difficult conditions. In particular, it severely discounts the ability of people with very limited material and power resources to actively respond to natural disaster.

In fact, I found that even very poor men and women often have some cultural and social resources that partially compensate for their comparative lack of economic and political assets. Critically, all Kazipur people—even those who never have been directly affected by erosion—are used to radical changes, and have reconfigured much of local cultural and social “tradition” to address and effect these changes. This whole region has been transformed in the last generation, and people have learned how to deal effectively with complex change.

1. Demographic, spatial and ecological change. Kazipur has seen an enormous population increase in the past fifty years. Early population growth was a result of the in-migration of agriculturalists. More recent population increase has stemmed from a combination of natural growth and nearby forced migrants moving into the area. Nowadays, virtually all land is effectively controlled and productively used. A generation ago, some forest and bush land still existed and played an important economic role. Now, it has all been transformed into farmland. Fish stocks in the Jamuna have decreased precipitously. In the past 15 years, more than half of the union’s population has been uprooted; many families several times. Despite these challenges, people have found ways to survive, and some, to prosper.


Kazipur, Bangladesh connected to the outside even during flood
Connected to the outside even during flood

2. Infrastructural change. A pukka road and subsidiary earth feeder roads now connect the Union with Sirajganj City to the South and Bogra to the North. Over the past fifteen years, local mechanics and boatwrights have adapted small diesel irrigation engines to power twenty to thirty foot boats, which have replaced sail powered craft almost entirely. Local people did this with no outside guidance, and in spite of an initial government ban on the use of such engines for other than irrigation purposes. They revolutionized regional travel, which has had significant economic and political consequences not yet evident when the REIS data was collected. There are no privately owned cars or trucks in the Union. Frequent bus service and a growing number of pedal-powered vans and rickshaws provide transportation for people who earlier would have walked, packed goods on horses or donkeys (both now entirely gone), or used sail-powered boats—or most likely would not have traveled at all. Chars that ten years ago were a risky day’s sail away are now reached routinely in an hour or two. Rapid transportation, the development of local Upazilla and police infrastructure, the growing centrality of government–primarily in determining who gets what, and in the provision of aid—exerts unprecedented political authority, control and surveillance in outlying areas. A complex brokerage system based on local touts now plays a key role in determining access to government and agency-controlled resources.(18) Mosques that have become linked to Dhaka, and by extension to Arab countries, currently function as sources of metropole religious and cultural orthodoxy. Operating in conjunction with schools, radio and television, they limit local cultural variability and introduce elements of “national” and “international” culture.

3. Economic change. Multiple cropping, new strains of rice (the traditional strains are rarely grown except on chars), fertilizer, and (among those better off and well connected) irrigation, and powered threshing mills have become commonplace during the past fifteen years. Shallow tube wells for drinking water were uncommon twenty years ago, but are universal today. “Traditional” farming and subsistence strategies that are not viable have been abandoned, leading to a severe reduction in jute acreage, large scale vegetable production, and fishing. New and varied marketing and purchasing strategies have developed to take advantage of easier transportation. Thousands of men routinely travel to seek short-term agricultural work elsewhere, establishing a pattern that did not exist a generation ago. Some labor contractors take local work gangs to distant parts of the country, while others have developed a highly complex local system for providing kumla (unskilled, paid) labor for aid-funded earth work construction and maintenance. Some families have sent laboring foot soldiers of a more permanent nature to cities. Rural-urban chain migration resulted in several hundred Kazipur residents working in a single Dhaka clothing factory. People continue to marry locally, but many now have kinship, friendship, village-based, economic, and other networks that extend to a number of other places where important opportunities and resources circulate. Today, they and their identities are powerfully connected in real and symbolic ways to the outside world.


House wall and roof sections in Kazipur, Bangladesh are 'modular' and can be quickly disassembled and moved
House wall and roof sections are 'modular' and can be quickly disassembled and moved

Innovation, adaptability, connectiveness and variation among displaced people

The space in which we live, which draws us out of ourselves, in which the erosion of our lives, our time and our history occurs, the space that claws and gnaws at us, is also, in itself, a heterogeneous space…we live inside a set of relations. Michel Foucault (1986:23)

No Kazipur people are traditional, “conservative” peasants, even though many icons of their traditional life still abound. Major forces other than riverbank erosion impact them in ways that are, in the largest sense, fundamentally outside their control. However, even within these maximal constraints, local people have become extremely innovative in maximizing their personal responses to change. Well used to change in general, people have developed a wide variety of innovative and systematic strategies to deal with changes stemming from riverbank erosion.

Housing and resettlement strategies. On first inspection, housing among Kazipur embankment dwellers appears to be “very poor” (Elahi 1989). Considered more carefully, the typical embankment house may not be all that inferior to the regional norm. Local houses may look less substantial than elsewhere in Bangladesh, but this at least partially reflects its functional adaptability in a region where erosion is endemic. Throughout Kazipur, houses of poorer people are built on a frame of wooden support poles, to which are lashed discrete woven wall and ceiling sections, typically made of catkin (a tall, coarse grass). Better-off people build larger houses using the same construction, but with galvanized metal roof panels and sometimes wooden doors symbolic of past or present prosperity. The still better off add galvanized wall panels. There are very few houses made of sun dried brick; when erosion comes such structures are entirely at its mercy, as was the fate of many local brick and concrete public buildings.

Each of these one-story dwellings is made of modules that can be disassembled quickly, and with little waste of materials, when faced with immanent erosion. Each of the modules can be transported rapidly by no more than six to eight men. A typical house can be re-erected in a day or two, using only household labor, particularly if extensive preparation of the new site is not required. If necessary, the reconstructed house can be made smaller and the surplus materials sold.


In Kazipur, Bangladesh, a basic house of a displaced person can be reassembled in a day
A basic house can be reassembled in a day

Most other chattel property is easily transportable and can be quickly set up in a new place. Property that is not transportable, such as the baked clay cooking stoves of the poor, are soon remade at the new site. The proficiency with which people can move is nothing short of remarkable. I have seen people threatened by the river disassemble their houses on eroding parts of the embankments, move houses, household goods, animals, and even banana plants and small trees to another portion of the embankments, reassemble them, and reestablish some semblance of household order in two days.(19) Those under inevitable but less acute threat of erosion so methodically remove everything of utility and value from their land that the abandoned homesteads look like moonscapes.

Over the years, a wide range of differentially preferred and entitled local strategies have been developed concerning where to move. Obviously, these strategies depend on individual and familial circumstance. People who remain well off or have non-farming occupations have the most options, and typically move to previously owned or purchased land, either within a few miles of their prior homes or outside the region.(20) Most people lack such substantial resources, but have developed complex strategies to rent very small pieces of land for house plots and agricultural use. I believe the majority of displaced people who remain in the area eventually make use of some kind of rental or lease option. Others will try to exercise kin-based residence entitlements, either temporarily or permanently. All across Bangladesh powerful patriarchal ideologies confer greater status on families that can maintain patrilocal residence. Settlement with the husband’s father or brother is therefore the paradigmatic ideal, but often such relatives will have been displaced as well and cannot help much.

One consequence is that residence with wife’s and more distant husband’s maternal kin is increasingly common. Reportedly, approximately 10 percent of Kazipur Union’s resident households currently may be uthuli (Indra and Buchignani 1997), and most of them are resident with such kin. Elsewhere in Bangladesh uthuli often have a strongly inferior client relationship to their landlord/patrons. Landlords demand agricultural labor, household service, and political support in exchange for a house plot. In contrast, over the last generation, uthuli patron-client relationships in mainland Kazipur have become both more common and more reciprocal. They are now extensively used by the poor to provide mutual aid in times of deep personal crisis. More than three-quarters of uthuli-patron relations in Kazipur are between kin (patrons often being little better off than their clients). In contrast to the prevailing patrilineal bias, they are usually dependent on the kinship claims and entitlements of displaced women, rather than those of men (Indra and Buchignani 1997).

Since the late 1970s, other displaced people have decided to settle on the embankments. When the first displacees moved there, virtually all had alternative residence possibilities. Most chose embankment residence because they lacked sufficient resources to settle in another place where their status, independence, security and economic benefits would be higher. More recent displacees have had fewer advantageous residence options, leading to a dramatic increase in the embankment population in the late 1980s and early 1990s.


Embankment communities in Kazipur, Bangladesh are highly ordered
Embankment communities are highly ordered

Production and reproduction. Once on the embankments, people have generated a rich variety of orderly adaptations to life in restricted circumstances. The suggestion that these are “rural bastes (slums)” (Elahi 1989) is only accurate in the deeper sense that studies of poor urban communities in Bangladesh have shown them to be highly organized places full of innovation. As in urban “squatter settlements,” many such adaptations are at least partially spatial, as space on the embankments is at a premium. Rising roughly four to five meters above the surrounding land, embankments are at most 15 meters wide, including usable side slopes. Until 1991, people settled densely some parts of the local embankment system, leaving other parts unoccupied. They have done so for security reasons, to maintain a sense of community, and to have access to jobs and hats (weekly markets).There were never any scattered settlements or isolated dwellings. Houses and other structures are immediately adjacent to each other, lining both sides of a central roadway. The physical distribution of houses, kitchens, work areas resemble those in local, linearly arranged char paras. The use of the open roadway—changes seasonally; it is used as a collective space for drying fuel, fodder, and spices. Shallow tube wells have been sunk by some households for drinking water, and are extensively shared. Occupancy conveys strong use rights. It is clear to all which household “owns” what land, and where the boundaries are. Such downslope boundaries are not accepted by adjacent landowners, who also claim these small areas, and many vigorous disputes arise out of their use.  Household and roadway spaces are intensively used by women for household production.(21)

People develop innovative economic options for themselves and their families. In this regard, the work of women is critical to the existence of all households, including both “conventional” households having a married couple at its core and “unconventional” ones that do not. Indeed, analyses of embankment households concentrating solely on the economic inputs of males indicate that they cannot “technically” survive.(22) But they do survive, in part because of economic activities of women, and sometimes children. Purdah restrictions in Kazipur are not as constraining as in the southern part of the country; they are even less so on the embankments, where women have innovated important changes in how they access public space. While only a few poor women occasionally do field work, most women have limited freedom to carry on a range of economic activities in and around their and adjacent households. Every available patch of ground is used by women to grow spices and vegetables, particularly in the areas of the embankment that have been occupied for some time. Many vegetable plants are trained to grow up and over the houses.(23) Catkin planted around houses provides building material and household privacy. Women often keep a chicken or two, or more rarely a goat (often owned by share or acquired through sharecropping). Having an established place gives people the physical means to extend women’s activities outside the household perimeter. This is critical, as household production alone cannot entirely support any of these families. Some poor and older women and girls range the embankments and riverbanks to collect roots, grass, sticks for fuel, and wild plants used for food and animal fodder.

Women's gathering activities in Kazipur, Bangladesh are economically important
Women's gathering activities are economically important

Women and their daughters commonly do household work for other families off the embankment in return for food. A few beg. Other women husk paddy (rice) for better off farm families. Here and there, women sell things from their houses, while a few have successfully negotiated the local power structure to secure paid, aid-supported work maintaining the Union’s earth roads. Some extremely poor, widowed, divorced, abandoned, or very old women with good political connections get monthly rations of wheat provided through international aid. All across the embankments, complex patterns of reciprocal economic exchange link women from different households. Relatives and village mates who still have land sometimes provide women with gifts of food and clothing, if they can afford it. Men and women remitting wages to their families from afar are a critical resource.

Men’s options are no less varied, and are better remunerated. Some young men and boys permanently work in households off the embankments, receive wages rather than simply food, and contribute a portion to their families. A few families still own small pieces of land and use male household labor to do the field work. Others lease, rent or sharecrop some nearby land. However, most men use diverse strategies and personal connections involving greater access than women to public space to get agricultural, construction and other laboring work. Many are absent for long periods during which women manage the households. In fact, at certain times of the year, over a quarter of embankment households are headed by women.(24) There is also considerable male entrepreneurial activity on the embankments: several dozen shops have been established, and others support themselves as oil pressers, blacksmiths, scrap dealers, rice brokers, labor contractors, healers (kobiraz) or religious experts. Some partially support themselves through home manufacture of mats, house wall units and fishing traps. Until erosion displaced them yet again in 1991, a whole localized embankment Hindu community specialized in fishing.

Connectedness. Like the term “political refugee,” that of “environmentally forced migrant” conveys an image of fundamental detachment from “roots” (Malkki 1992:27-9). In Bangladesh, this has a well-established analogue in the more general thesis of “nuclearization” (Mahmuda Islam 1985:44), the “decline of patrilineal groups” (Arens and van Beurden 1977:102), family “breakdown” (Kabeer 1989:34; 1991:257), and individualization (Maloney 1988) among the rural poor. Poverty, exploitation, functional landlessness, underemployment, and forced migration are claimed to have lessened adherence to family and kinship norms, broken down and weakened inter-familial kinship ties, and increased what Maloney (1988:52) terms “Bengali pragmatic individualism.” As with other elements of the conventional story line, I could provide considerable support for this thesis. Forced migration certainly has put enormous strains on kin and social systems. Most have seen not just their homes and land, but also whole paras and villages, washed away. The latter can never be fully replaced. Most are much poorer than before. To other people in the Union, living on the embankments or as uthuli is a mark of loss of independence and honor, and some women and men feel this stigma particularly acutely. Rates of marriage discord and breakup do appear to be markedly higher among displacees.

At the same time, embankment dwellers and other displacees have put an enormous amount of energy into maintaining both behavioral and symbolic connections with their prior lives. This effort must occur elsewhere in Bangladesh where displacees have settled, but appears to have been under-emphasized in problem-oriented research. In part, this may be an unintended consequence of over-dependence on quickly collected survey data solicited solely from the “heads” of individual households. Such data used in isolation inevitably emphasizes the distinctiveness, independence, and isolation of households rather than connections among them. Indeed, in many respects, environmentally displaced people in Kazipur frequently contradict the widely held thesis that in Bangladesh “In the event of [environmental and other kinds of] crisis, a household must fend for itself. Relief from neighbors and kin cannot be counted on, whether because they too are affected or simply because they are caught up in the struggle for their own survival (Cain 1978:426).”

Let me use embankment communities as a case in point. Because they were formed anew and often very rapidly, they seem to be among the most “untraditional” mode of living in the region. However, these communities exhibit a very strong pattern of kin-based chain migration and settlement. During 1990-91, 239 of 553 households (43 percent) were living immediately adjacent to household men’s agnatic kin, while eighteen (three percent) were living next to women’s kin. Moreover, village-based resettlement patterns have been strong. In 1991, 85 percent of embankment households claimed to be from the same village as at least one of their immediate neighbors.(25) Superficial appearances aside, embankment settlement was far from random. Many people were able to activate a wide range of kinship, friendship, village-based and other links immediately upon establishing themselves there; this pattern was even stronger among individuals resettling in villages.


In Kazipur, Bangladesh, this respected elder played an important role in his eroded village's Kurbanir Eid (Eid-ul Azha) celebrations
This respected elder played an important role in his eroded village's Kurbanir Eid (Eid-ul Azha) celebrations

A good example is Monikpotol, a nearby “traditional” mainland village that was completely destroyed by the Jamuna in the past few years. Typically, the majority of people either dispersed throughout the Upazilla or left it entirely. But a significant core moved to the embankments, bringing with them the village school and mosque, as well as a host of personal statuses, relationships, and strong informal community networks. They also transplanted a range of community institutions. These people continue to define their part of the embankment as Monikpotol, not only to support identity and honor, but also to maintain a legitimate claim to land that may re-emerge in the future, to vote and have local political representatives, to secure government funding for schools, and to secure aid. It is crucial to recognize that such gestures are both significant individual empowering strategies and factors that make a difference in the political and economic power arenas of the Upazilla.

Embankment people are highly variable. So far, there has been a tendency to focus research on the poorest of rural environmentally displaced people. Although some attention has been paid to powerful char landowners (Zaman 1988), this contrasts with a rather “flat” and normative characterization of everyone else. But as Appadurai (1984:483) and Kabeer (1990:135) note, “the” poor in South Asia are most adequately represented as a number of types of people who are situated differently with respect to power, production and prestige. This certainly is true for displaced people in Kazipur, who are in fact quite diverse.

Consider those who moved to embankments. The generic image of an environmentally forced migrant in Bangladesh is of a person who was previously a marginal, usually male, farmer, and whose land was destroyed by riverbank erosion. However, a fair number of embankment dwellers are not forced migrants at all, or if they are, were not previously farmers. A few others were local farmers who moved temporarily to the higher land of the embankment during a flood, and stayed. Others were entrepreneurs, small scale manufacturers, fishermen, carpenters, hand sawyers, blacksmiths. There were also a couple of teachers and salaried government workers who found it a convenient, rent-free locale. Still others, after migrating from afar, chose to build their first home in Kazipur on the embankment. None were traditional village-based, land-owning peasants. Most were already functionally landless, and many had been displaced several times before. Some were already quite disconnected from “traditional” kin and community, while others were not. A handful–such as the two elected Union politicians and the head of the Hindu Haldar fishing community— were, and remained, relatively powerful. Those forced migrants who still had land typically did not settle on the embankments. Nonetheless, the class statuses of embankment dwellers varied considerably. Though landless now, some of the latter migrants still possess “symbolic capital” (Bourdieu 1984) which they can convert into economic and status benefit(26).


By 1990 some embankment areas in Kazipur, Bangladesh looked like villages
By 1990 some embankment areas looked like villages
A long settled embankment area in Kazipur, Bangladesh in 2007
A long settled embankment area in 2007

People’s post-migration circumstances also varied greatly, particularly in a temporal sense. Socio-cultural research on forced migration has so far concentrated primarily on sketching out the current socio-economic situation of recently displaced people, and comparing it with an ex post facto reconstruction of what their lives were like before displacement.(27) But in such chronically unstable regions as Kazipur, longer term, ongoing processes of personal, familial, and community adaptation are also important. Indeed, for each person displaced in a given year, many others displaced earlier are still dealing with the consequences. Taking the embankment population as a whole, at the time of my initial fieldwork, the “typical” person had lived there for only two years. However, such averages mean little in terms of understanding displacement. For example, parts of the embankment system located near the sole remaining town were being actively eroded during the 1990 flood, and so were quickly abandoned by those living there. This space was totally re-occupied within three days by thirty desperate households eroded out elsewhere, who needed a short term place to stay. In contrast, some people had been living on more stable parts of the embankments less than a half mile away for over eleven years. Indeed, save for the obvious spatial constraints, the stable, first-settled parts of the embankments did not look much different from a “traditional” local para. And yet, even these nominally stable regions are potentially susceptible to radical change. One such area on an older, retired embankment section near this town was itself eroded during the 1991 rainy season, displacing the whole Haldar community. People also had radically different resources available to them after being displaced. Some women were so poor that they could barely provide themselves with basic protection from the elements or a minimal level of honorable privacy. A few households were able to bring substantial material (and social) assets with them to the embankments.

As an illustration of this heterogeneity, Table 1 lists the main claimed occupations, prior locality and length of residence of men and one woman in some immediately adjacent households living on a previously secure branch of the embankments that was beginning to erode away in early 1992.



Prior Occupation and Locality


Length of Residence
local primary school teacher
long term resident
agricultural worker from the chars
four months
agricultural worker from the chars
two years
carpenter from Rajshahi
six months
agricultural and sometime earth worker from Bogra
six months
local (Muslim) fisherman
six months
local Hindu cane weaver
seven months
local Hindu potter
six months
local Hindu fisherman
four months
local Hindu community leader
six months
local Hindu fisherman
six months
agricultural worker from Bogra
six months
local widow, soon to move to her daughter’s place in Dhaka, where the latter is a garment worker
long term resident
agricultural worker from the chars
one year
local primary school teacher
long term resident


Table 1. Prior Occupation and Locality, and Current Length of Embankment Residence

Of course this Table does not adequately convey the range of economic and situational diversity, inasmuch as most of these people would in fact secure food and income from a number of different economic strategies before and after their arrival on the embankments. Every short term resident had come immediately from another part of the embankment system.

V. Conclusion

[Social science] has this privilege—that it can take as an object its own functioning as a relatively autonomous social world and that it is thereby able to bring to rational consciousness at least some of the socio-political constraints that bear on scientific practice. Consequently, it can utilize the awareness and the knowledge it has of its own functioning, and the negative or positive epistemological effects that ensue, to try to overcome some of the obstacles that stand in the way of scientific progress. Bourdieu (1991:374-5)

I have written this paper in the spirit of the above quote, which I take optimistically. I certainly have no wish to imply that those who research environmentally displaced people in Bangladesh or elsewhere in a highly problem-oriented, “victimological” mode are doing anything improper or that their findings are less objective or sound than mine. Neither do I want to suggest that the conceptual hold and power to allocate salience that the flood and erosion disaster discourse have over research is anything particularly aberrant. One sees similar patterns in such diverse things as development, migration, refugee, ethnic relations and “family problems” research—indeed, in just about every social science research domain that is linked to a powerful social issue of the day.

What I do wish to assert is the need for forced migration researchers to be more aware of these externally imposed conceptual constraints and, for several reasons, to resist them. I have put forward some of these reasons here: so that research does more than legitimate the worldview of allied social problems discourses; so that through improved methodologies and writing styles, the people under scrutiny may be given greater scope to present their concerns, situations and ways of seeing more forcefully in what we write; so that we, as researchers, recapture a bit more autonomy ourselves, so that when we speak we add to the discourse our own unique social science ways of seeing; to be aware as researchers of the highly leveraged power potentially available to those offering up rigorous but dissonant and challenging views to that which is otherwise taken for granted or considered mere matters of “common sense.”

Today these points are well appreciated by some theoreticians of natural disasters (see Zaman 1991b:1, 3-5). However, regarding erosion and forced migration in Bangladesh, we are at present far from achieving a diverse, polyphonic discourse. If this does not change, then not only the theoretical, but the practical consequences may be very significant. In particular, nominally well researched flood and erosion protection decisions involving millions of people will be made with little effective input from those affected.

References Cited

Adnan, Shapan

1991 Floods, People and the Environment, Institutional Aspects of Flood Protection Programs in Bangladesh, 1990. Dhaka, Bangladesh: Research and Advisory Services. 127 pp.

Ahmed, Mohiuddin and Andrew Jenkins

1988 Traditional Land-Grabbing and Settlement Patterns in the South Eastern Delta. Paper presented at the International Symposium on the Impacts of Riverbank Erosion, Flood Hazard and Population Displacement, Dhaka, April.

Akhter, Golam Morshed and Nasimul Gani

1988 Riverbank Erosion Impact Study - Response Strategies to the Problem of Landlessness, Rural Impoverisation and Marginalisation. Paper presented at the International Symposium on the Impacts of Riverbank Erosion, Flood Hazards, and Population Displacement, Dhaka, April.

Alam, S. M. Nurul

1990a Annotation of Social Science Literature on Natural Disasters in Bangladesh. Produced by PACT Bangladesh/PRIP in collaboration with ADAB, September, 1990.

1990b Perceptions of Flood Among Bangladeshi Villagers. Disasters 14(4):354-357.

Amin, A. T. M.

1988 Settlement Strategy for Riverbank Erosion Displacees in Bangladesh: A Human Resource Development Approach. Paper presented at the International Symposium on the Impacts of Riverbank Erosion, Flood Hazards, and Population Displacement, Dhaka, April.

Appadurai, Arjun

1984 How Moral is South Asia’s Economy: A Review Article. Journal of Asian Studies 43(3):481-497.

Arens, Jenneke and Jos van Beurden

1977 Jhagrapur: Poor Peasants and Women in a Village in Bangladesh. New Delhi: Orient Longman Limited.

Bangladesh Bureau of Statistics

1988 Small Area Atlas of Bangladesh: Mauzas and Mahallahs of Pabna District. Dhaka: Bangladesh Bureau of Statistics.

Bourdieu, Pierre

1991 Epilogue: On the Possibility of a Field of World Sociology. In Social Theory for a Changing Society, P. Bourdieu and James Coleman, eds. Pp. 373-388. Boulder: Westview Press.

1984 Distinction: A Social Critiques of the Judgement of Taste. Cambridge, MA: Harvard University Press.

Cain, Mead

1978 The household life cycle and economic mobility in rural Bangladesh. Population and Development Review 4:421-438.

Chen, Lincoln C., ed.

1973 Disaster in Bangladesh. New York: Oxford University Press.

Elahi, K. Maudood

1989 Population Displacement due to Riverbank Erosion of the Jamuna in Bangladesh. In Population and Disaster. James I. Clarke, ed.Oxford: Basil Blackwell.

Elahi, K. Maudood and John R. Rogge

1990 Riverbank Erosion, Flood and Population Displacement in Bangladesh. A Report on the Riverbank Erosion Impact Study. Savar, Bangladesh: Jahangirnagar University, 98 pp.

Foucault, Michel

1972 The Archeology of Knowledge. New York: Random House.

1979 Discipline and Punish: The Birth of a Prison. New York: Vintage.

1986 Of Other Spaces. Diacritics 16:22-27.

Fraser, N.

1989 Unruly Practices: Power, Discourse and Gender in Contemporary Social Theory. Minneapolis: University of Minnesota Press

Government of the People’s Republic of Bangladesh, Ministry of Irrigation, Water Development, and Flood Control

1992 FAP-14. Flood Response Study Conference Briefing Notes. Irrigation Support Project for Asia and the Near East. February, Dhaka.

Halcrow, Sir William and Partners Limited in association with Danish Hydraulic Institute, Engineering & Planning Consultants Ltd. Design Innovations Group

1992 River Training Studies of the Brahmaputra River. Individual Flood Action Plan Team Handouts from the Second Conference on the Flood Action Plan. Dhaka, March 1-5.

Hanchett, Suzanne, Jesmin Akhter and Kazi Rozana Akhter

1998 Gender and Society in Bangladesh’s Flood Action Plan. In Water, Culture and Power: Local Struggles in a Global Context. John Donahue and Barbara Johnston, eds. Pp. 209-234. Washington, DC: Island Press.

Haque, Chowdhury Emdadul

1988 Impact of River-Bank Erosion Hazard in the Brahmaputra-Jamuna Floodplain: A Study of Population Displacement and Response Strategies. Unpublished Doctoral Dissertation, Department of Geography. Winnipeg: University of Manitoba.

Indra, Doreen Marie and Norman Buchignani

1997 Rural Landlessness, Extended Entitlements and Inter-Household Relations in South Asia: A Bangladesh Case. Journal of Peasant Studies 23(3):25-64.

ISPAN Irrigation Support Project for Asia and the Near East

1992 Newsletter No. 25.

Islam, Muinul

1992 Natural Calamities and Environmental Refugees in Bangladesh. Refuge 12(1).

Islam, M. Aminul and Howard Kunreuther

1973 The Challenge of Long Term Recovery from Natural Disaster: Implications for Bangladesh. The Oriental Geographer (Dacca) 17(2):51-62.

Islam, Mahmuda

1985 Women, Health and Culture: A Study of Beliefs and Practices with Female Diseases in a Bangladesh Village. Dhaka: Women for Women.

Jahan, Rounaq

1989 Women and Development in Bangladesh: Challenges and Opportunities. Ford Foundation. 22 pp. with tables.

Kabeer, Naila

1989 Monitoring Poverty as if Gender Mattered: A Methodology for Rural Bangladesh. Discussion Paper 255, Institute of Development Studies, University of Sussex.

1990 Poverty, Purdah and Women’s Survival Strategies in Rural Bangladesh. In The Food Question: Profits versus People. Henry Bernstein, Ben Crow, Maureen Macintosh and Charlotte Martin, eds. London: Earthscan.

1991 Gender Dimensions of Rural Poverty: Analysis from Bangladesh. Journal of Peasant Studies 18(2):241-262.

McCarthy, F.

1984 The Target Group: Women in Rural Bangladesh. In Room For Manoeuvre: An Exploration of Public Policy Planning in Agricultural and Rural Development. E. J. Clay and B. Schaffer, eds. Pp. 49-58. London: Heinemann Educational Books.

Malkki, L.

1992 National Geographic: The Rooting of Peoples and the Territorialization of National Identity Among Scholars and Refugees. Cultural Anthropology 7(1):24–44.

Maloney, Clarence

1988 Behaviour and Poverty in Bangladesh. Dhaka: University Press Limited.

Mukherjee, Radhakamal

1938 The Changing Face of Bengal, a Study in Riverine Economy. Calcutta: University of Calcutta.

Rahman, Atiur

1988 Rural Touts and Power Structure in Bangladesh: Some Indicative Results of a Perception Survey on the Patronage Network. ADAB News, July-August 15(4).

Rogers, Peter, Peter Lydon and David Seckler

1989 Eastern Waters Study: Strategies to Manage Flood and Drought in the Ganges-Brahmaputra Basin. Irrigation Support Project for Asia and the Near East. April. Washington, DC

Rogge, John R. and K. Maudood Elahi

1989 The Riverbank Erosion Impact Study Bangladesh. Final Report to the International Development Research Centre, Ottawa. Prepared by Dr. John R. Rogge, University of Manitoba and Dr. K. Maudood Elahi, Jahangirnagar University.

Scudder, T. and Colson, E.

1982 From Welfare to Development: A Conceptual Framework for the Analysis of Dislocated People. In A. Hansen and Anthony Oliver-Smith, eds. Pp. 267-287. Involuntary Migration and Resettlement: The Problems and Responses of Dislocated People. Boulder: Westview Press.

Simone, Timothy

1990 Metropolitan Africans: Reading Incapacity, and the Incapacity of Reading. Cultural Anthropology 5(2):160-172.

Shapiro, Michael J.

1987 The Rhetoric of Social Science: The Political Responsibilities of the Scholar. In The Rhetoric of the Human Sciences: Language and Argument in Scholarship and Public Affairs, John S. Nelson, Allan Megill and Donald N. McCloskey, eds. Madison: University of Wisconsin Press.

Wiest, Raymond E.

1988 Domestic Group Dynamics in the Resettlement Process Related to Riverbank Erosion in Bangladesh. Paper prepared for presentation at the International Symposium on the Impact of Riverbank Erosion, Flood Hazard and the Problem of Population Displacement, Dhaka, Bangladesh.

Zaman, M.Q.

1988 The Socioeconomic and Political Dynamics of Adjustment to Riverbank Erosion Hazard and Population Resettlement in the Brahmaputra-Jamuna Floodplain. D. Phil. Dissertation, University of Manitoba.

1991a Social Structure and Process in Char Land Settlement in the Brahmaputra-Jamuna Floodplain. Man 26(4).

           1991b The Displaced Poor and Resettlement Policies in Bangladesh. Disasters 15(2):117‑125.


1. I would like to thank John Rogge and Maudood Elahi, who assisted me in the development of my project, and Norman Buchignani, Hasina Ahmed and Raquibal Islam, who worked with me in the field. Norman Buchignani and three anonymous reviewers made important editorial comments on earlier drafts of this paper, and Dr. Bhagwan Dua has supported my research in many ways. Grants from the University of Lethbridge Research Fund and the Social Science and Humanities Research Council of Canada (SSHRC) supported part of this work. I did nine months of continuous fieldwork in Kazipur during 1989-90. Research assistants continued to collect data during 1990-5, and I did a two month follow up period of fieldwork in 1993. Riverbank erosion continued to cause further changes in the region after 1995, but this paper does not address them. There were 553 households established on top of Union embankments when I did a full census during the 1989-90 dry season; in early 1992 there were 531 households residing there on a much erosion-reduced set of embankments.

2. The physical forces involved are monumental. The west bank of the Brahmaputra River, for example, seems to have been moving westward 40-50 meters a year over a period of centuries (Halcrow and Partners 1992:3).

3. A clear distinction must be made between permanent displacement caused by riverbank erosion and that temporarily caused by annual monsoon flooding. In some regions annual flooding is severe, but causes no significant erosion or long term displacement. Erosion is usually at its highest during the times of peak river flow, which is when flood is most likely. Short term flood can develop permanent new river channels, which in turn generate enduring new sources of erosion.

4. Conceptually, disaster and Bangladesh have been connected intimately since the country’s foundation (Chen 1973; M. Aminal Islam and Kunreuther 1973).

5. Here, the image of people captive to an immense natural force destroying their crops, livelihoods and dwellings operates powerfully to mobilize national and international resources. It is interesting that during the massive erosion, breach in the local embankment, and consequent flooding faced by Kazipur Union in 1990, international newspaper reports of the flood brought then President Ershad for a visit—not chronic erosion, which had been going on at a high rate for months.

6. See in particular the work of Zaman (1988, 1991a, 1991b) and the project summary by Rogge and Elahi (1989). For research not connected to REIS, see Ahmed and Jenkins (1988), Akhter and Gani (1988), Alam (1990a) and Indra and Buchignani (1997).

7.To this point, public, media and development discussion of riverbank erosion has been incomparable with that concerning flood. The amount of sociocultural research done on the subject recently is about one-third that on flood, and the vast majority of the former are REIS studies. See S. M. Nural Alam’s (1990b) annotated bibliography of the Bangladesh natural disaster literature.

8. Following Foucault (1979) and Shapiro (1987:363), “discourse” here refers to “any systematic or disciplined way of constituting subjects, objects, and relationships within a linguistic practice.”

9. Alam (1990b:7-8) advocates studies with longer time depth, of which there are presently none. My own work in Kazipur eventually will span at least 1989-2000.

10. Upazillas are the smallest national administrative regions in Bangladesh. The national government presence is marked by a complex of administrative offices and a police post. Upazillas are in turn subdivided into Unions.

11. For a combination of political and legal reasons official population estimates the Bangladesh Bureau of Statistics (see 1988: Table 50) include villages long ago destroyed by the Jamuna, as well as people who have “temporarily” moved away to gain work.

12. In many erosion-prone areas loss of land due to erosion is at least partially compensated for by the accretion of new (albeit typically unfertile and flood-prone) land. Kazipur Union gained no such land until 1992.

13. See Elahi and Rogge (1990), and Elahi (1989) for further details. As part of this study a formal survey of some embankment and nearby households was carried out in 1985.

14.They are also strategically placed symbolically. For instance, when foreign teams (say, from the World Bank) or government officials from Dhaka visit the area to “see the effects of flood and erosion first-hand,” they are most often taken by local officials to embankment settlements. Here they dramatically encounter crowds of embankment people, who often surround them, begging for their stories to be heard so that some kind of solution to their plight might be implemented.

15. I was able to expand my research accordingly, chiefly because my long term residence allowed me the flexibility and time to make this adjustment. This kind of discovery and readjustment process rarely can be done using the mode of survey-based “socioeconomic” data collection typically used in large scale studies in Bangladesh.

16. Several authors using REIS data have suggested otherwise, but they have misinterpreted their data. What they found was that interviewed displacees still living in the Upazilla typically had lived in the Upazilla previously (Wiest 1988). No REIS data was collected locally from those about to be displaced.

17. Zaman’s rich characterizations of char life are a partial exception, inasmuch as he stresses long term local adaptations to river erosion (1988, 1991a). His studies, however, do not centrally address the current high rate of structural change on the chars caused by other factors.

18. Rahman (1988) earlier noted the rise of this tout system more generally.

19. As on erosion-prone chars (ISPAN Newsletter 25, 1992), Kazipur people have developed strategies to mobilize help and to physically move when erosion threatens. Unpaid cooperation is not extensive (save sometimes by unaffected relatives), but paid laborers can quickly be drafted to help one move.

20. As Amin (1988) notes, the consequences of erosion can be quite different for farming and non-farming displacees. My data on first time about-to-be-eroded people show this to be so for Kazipur.

21. As McCarthy (1984:51) notes, the use of such land by women for productive purposes is critical to household economies, yet has largely been unrecognized by researchers and planners.

22. A failure to symmetrically consider women and men in economic and other domains can lead to research discourse which identifies “the people” almost exclusively with the stated concerns of men. In such discourse women are framed (as in the “idealized” middle class conventional research paradigm) as their dependents rather than showing how women and men have reciprocal ties and interwoven dependencies and obligations.

23. Government officials responsible for the construction and maintenance of the embankments at first strongly objected to these settlements, arguing that the embankments are weakened when people cut away their sides to erect buildings, and that their residence encourages rain-generated erosion and networks of rat holes. At this point there is little evidence in support of this thesis, and the opposite argument is equally plausible: that people strengthen the embankment by compacting the soil and rooting it with plants and trees.

24. Jahan (1989) estimates that about 15 percent of all rural and 25 percent of all landless Bangladesh households may now be “female-managed” or “female-supported”, to make Kabeer’s (1989:34) important distinction between those situations in which a woman’s husband may be away at work and those where a woman has no husband and thus herself provides the main economic support for the household.

25. The migration history of many of these individuals is very complex and not adequately summarized by a simple statistic like this. One “man of Monikpotol” may have been born and raised in Char Kazipur, married into Monikpotol, and when the latter was eroded, settled for a while on the local main road before coming to the embankments. Another “man of Monikpotol” could have an altogether different settlement history.

26. Zaman’s char research is a notable exception.