Rural Landlessness, Extended Entitlements and Inter-household Relations
in South Asia: a Bangladesh Case


‘This is an electronic version of an article of the same name published in the Journal of Peasant Studies (April 1997) Volume 24, Issue 3:25-64. This article is available online at:

A pdf version is also available.

Norman Buchignani and Doreen Marie Indra


[Acknowledgements:] We would like to thank Henry Bernstein for his insightful editorial input, and Donnan Hastings and Alan Aycock for reading an earlier version of this paper. Md. Raquibul Islam and Hasina Ahmed provided unwavering assistance in the field, supported by Md. Mostapha Kamal and Habib Rahman. We would like to acknowledge the intellectual and personal support of many individuals, especially: H.K. Arefeen and B.K. Jahangir (Centre for Social Studies, Dhaka University), K. Maudood Elahi and S.M. Nurual Alam (Jahangirnagar University), Atiur Rahman (Bangladesh Institute for Development Studies), Mabub Ahmed (Dhaka University), John Rogge (University of Manitoba), Bhagwan Dua (University of Lethbridge), Suzanne Hanchett and Md. Q. Zaman. Dennis Sheppard (Department of Geography, University of Lethbridge) produced the enclosed map. This research was supported by grants from the Social Sciences and Humanities Research Council of Canada (#42154) and the University of Lethbridge Research Fund. While actual major place names and political units are used throughout the text, pseudonyms have been assigned to all individuals in order to maintain their personal anonymity. To this end, some village names and details of personal histories also have been altered.





This paper shows how people in one part of Bangladesh rendered landless and impoverished by river bank erosion make innovative use of kinship and other ideologies legitimating reciprocity and mutual aid to re-establish themselves rent free on the land of others. It thereby addresses a larger empirical issue: where are the fully landless rural poor in South Asia living, and through what means? Theoretically, it extends Dréze and Sen’s analysis of entitlements and poverty to instances of inter-household cooperative conflict and mutual aid among extremely poor people. A culturally informed, gender disaggregated analysis of those locally called uthuli because they have settled on others’ land without monetary payment demonstrates that women’s “extended entitlements” (Dréze and Sen, 1989:10) as daughters, sisters and mothers are often critical assets in establishing residence. Women are also key agents in the establishment and maintenance of uthuli residence and in managing the benefits stemming from it. Using this approach, we show how landless women’s entitlements are pivotal in securing access to a houseplot for themselves and their families.

In this paper we show how people rendered landless and impoverished by river bank erosion in one part of rural Bangladesh make innovative use of kinship and other ideologies legitimating reciprocity and mutual aid to re-establish themselves rent free on the land of others. In doing so, we address a basic empirical issue largely ignored in both ethnographic and development oriented research in South Asia: where are the fully landless rural poor in South Asia living, especially the many millions of superpoor for whom renting or sharecropping others’ land is beyond their financial and social means? Through what claims and processes do they live where they do? We ground our discussion on Sen’s theory of entitlements and poverty and following Agarwal [1994], extend this approach in several ways. Originally a breakthrough theory of historical famine response, over the past ten years Drèze and Sen [1989: 10] and others [Sen, 1988, 1990: 141; Sharma, 1985; Agarwal, 1986; 1990: 353-4; Papanek, 1990; Kabeer, 1989; 1990; 1991] have expanded the scope of entitlement theory to address contemporary situations of rural poverty, especially gender-based intra-household distribution of food and other benefits among the poor. Agarwal [1994] recently has highlighted many ways in which household gender differentials in property rights have affected South Asian women’s power, status and well-being.

We use Drèze and Sen’s concept of culturally based “extended entitlements”, and Agarwal’s discussion of women’s access to land, to help frame local ideas about a particular form of dependent residence in rural Bangladesh and the people locally called uthuli who make use of it. This involves the application of a gender disaggregated form of entitlement analysis to inter-household relations, notably to an example of co-operative conflict and mutual aid concerning land rights among extremely poor people living on the margins of a kin-based patriarchal society. Using this approach, we show how landless women’s entitlements as daughters, sisters and mothers are pivotal in securing access to a houseplot for themselves and their families. We also show how women are often centrally active agents in the establishment and maintenance of uthuli residence as well as in managing the benefits that stem from it.

This paper has three main sections and a conclusion. In the first section we chronicle the rise of entitlement analysis and its recent application to the study of household gender relations among the rural poor in South Asia, Bangladesh in particular. We then provide a rationale for the application of entitlement analysis to the inter-household relations of poor women and men. We stress there the importance of having a village-based place to live to rural people with few other resources at their disposal. Section II briefly sketches the setting: Kazipur Union, Kazipur Thana, Sirajganj District, Bangladesh. Once a socially and environmentally stable mainland farming region, over the past twenty years Kazipur has been transformed – indeed, almost destroyed – by the relentless westward march of the Jamuna River. Washing away whole villages, the shifting Jamuna has rendered thousands of households landless and homeless.

The third section locates uthuli residence rather far down a continuum of options hypothetically available to these forced migrants, inferior to the re-establishment of an ‘independent’ household on owned land, on a rented house plot or on the land of the husband’s father or brothers. It shows how and why uthuli residence nevertheless is a valued option for some in comparison to far worse alternatives, and how those who succeed in establishing uthuli residence make use of a particular kind of ‘cultural capital’ or culture-based entitlement: banked-up obligations to help kin and close village mates in acute need. We demonstrate that most uthuli arrangements are, and are locally acknowledged to be, based on the successful, if ill-legitimated, kinship claims of household women on their kin rather than on the more conventional claims of men on theirs. While arable land rarely is involved, this is a clear instance of poor South Asian women effectively accessing vital resources in a time of great personal need. In addition, Section III illustrates a range of ways in which uthuli women thereafter help maintain beneficial relations with their (often poor) landlords through astute management of social relations and the exchange of their labour, food and social support for continuing goodwill.

We then turn to the more general issue of adequately characterizing and engendering situations of chronic rural poverty and landlessness in South Asia. Researcher bias and methodological practice have ensured the continuing hegemony of ‘the household’ as the central object of inquiry, especially in applied studies. This once largely rendered both intra- and inter-household relations invisible and muted women. Feminist South Asianists have productively inserted a measure of gender disaggregation into this discourse on households, often through the use of extended entitlement analysis. We argue that an extension of the latter strategy to the more concerted study of inter-household relations of both women and men is necessary to further highlight the place of land and residence claims, and other inter-household entitlement negotiations, among the survival strategies of the very poor.


I. Entitlements, Poverty, Gender and Landlessness in South Asia

In less than twenty years the language of entitlements has become deeply encoded in South Asian applied and historical studies discourse. The starting point was of course Amartya Sen’s [1976; 1977; 1981] initial formulation of entitlement analysis, which he applied to the study of famine. One of Sen’s central points was that in agricultural, exchange-based societies lacking extensive state social welfare systems, unequal distribution and allocation of food rather than an overall “food availability decline” (FAD) [Sen, 1976: 1273] is the usual cause of starvation during famine. In these instances, “starvation is the characteristic of some people not having enough food to eat. It is not the characteristic of there being not enough food to eat” [Sen, 1981: 1]. According to Sen, households unequally possess “endowments” making possible independent production. “Exchange entitlements” such as labour that can be traded for a range of “commodity bundles” including food are also unequally distributed. When a household cannot produce sufficient food itself and its exchange entitlements are inadequate, household members may starve. This may occur even while others prosper or when, as in Sen’s now classic analysis of the great 1943 Bengal famine, it can be shown that there is sufficient food available in the system in principle to feed everyone. Failure of exchange entitlement (FEE) during famine differentially hits certain classes and categories of people already on the margins of rural society, such as landless agricultural labourers.(1)

Critical consensus has judged entitlement analysis a major advance in famine studies, and FEE approaches to famine now predominate, largely displacing earlier FAD ones. Entitlement theory has also greatly influenced crisis food aid provision program strategies around the world. More generally, Sen has helped create greater sensitivity to diversity among “the poor". As Appadurai [1984: 483] notes of Sen’s contribution, “...we are best off dispensing with the idea of “the poor” and thinking instead of more finely discriminated categories that differ in complex ways in their relationship to the forces of production and exchange".

Certain aspects of Sen’s formulation have also received pointed criticism. Most was directed either at the definitional and empirical status of exchange entitlements or at his globalizing use of households as the central social and economic units of analysis. Sen originally [1981: 53] and retrospectively claimed [Sen, 1989: 140; Drèze and Sen, 1989: 10] that entitlements were conceived of as having a “legal” basis: as “...the legal means available in a society, including the use of productive possibilities, trade opportunities, entitlements vis-à-vis the state, and other means of acquiring food” [1981: 53]. We believe that this definition suggests an implausibly narrow reading of Sen’s case by case use of entitlements. Be this as it may, Greenough’s [1982] wholesale embrace of a more general (though rather ideal and rarefied) cultural approach to the obligatory basis of entitlements in the 1943 Bengal famine, followed by much direct criticism of Sen’s “excessively legalistic view of social relations” [Appadurai, 1984: 403; see also Beck, 1991] quickly carried the day. From the mid-1980s, entitlements have been seen primarily as socially binding, culturally-based moral claims. Drèze and Sen [1989: 10] have in part accepted this interpretation, defining an additional category of “extended entitlements”:

While the concept of entitlement focuses on a person’s legal rights of ownership, there are some social relations that take the broader form of accepted legitimacy rather than legal rights enforceable in a court. [Such a person] ...can be seen as having a claim the legitimacy of which is accepted and thus effective, even though it is not a claim that can be upheld in a court or enforced by the power of the state. Despite their legally weaker form, such socially sanctioned rights may be extremely important in determining the amount of food or health care or other commodities that different members of a family get...

The topical scope of entitlement analysis was expanded in several critical ways beginning in the mid-1980s. The primary agents here were academic feminists already deeply involved in reformulating applied studies paradigms to more adequately portray the lives of South Asian women and to better characterize the increasing “femininization of poverty” [Pryer, 1990: 125]. Theoretical insights derived initially from acute historical famine situations were quickly applied to contemporary conditions of chronic poverty and malnutrition. While many authors maintain the macrosocietal level of analysis of Sen and Greenough [for example, Kabeer, 1990; 1988; Papanek, 1990; Chowdhury and Bairagi, 1990; Jahan, 1989; Maloney, 1988: 42-8], others have asserted the need for, or have carried out micro-level, situational, culturally-grounded studies of poor households. A central innovation here has been the rejection of the long-established ethnographic and development convention of treating South Asian households as a “black box” [Whitehead, 1981: 89] or as internally undifferentiated wholes [Agarwal, 1986: 165; 1994: 51-2], because doing so mutes women and renders household gender relations invisible. One result has been extensive inquiry into gender-based intra-household patterns of consumption, distribution, production, and exchange – and by extension, into the power and authority of variously placed women and men in poor households.

This in turn has led to the development of a more flexible, local-level, culturally informed notion of entitlements articulating “ideas of distributional justice shared by members of a group in society...” even if there remains “...a relative scarcity of direct studies of expressed consensus about entitlements” [Papanek, 1990: 170]. Sen himself [1983; 1988; 1990] has actively participated in the extension of entitlement analysis to intra-household dynamics, noting with approval that the approach now gives greater acknowledgement to the primacy of “deep-seated notions of “legitimacy” operating in the distribution within families... ...supplementing the operations of entitlement relations at the levels of households, occupational groups and classes” [Sen, 1990: 139].

Greater attention is now paid also to poverty as a process, as opposed to a state; on “the means by which basic needs are met rather than [just] on the extent to which they are met” [Kabeer, 1991: 244]. Advocates now counsel the application of entitlement analysis to “bargaining” [Sen, 1990: 123; Agarwal 1994: 54] between poor household women and men that highlights both gender inequities and women’s “ “agency” as opposed to [just] their well being” [Sen, 1990: 148]. While “this often appears as what Kandiyoti [1988: 275] terms “the patriarchal bargain” which implies fundamental subjugation of women to men... also suggests fluidity and re-negotiation” [Kabeer, 1990: 138]. Received ideologies about the centrality of the family and household notwithstanding, it involves a recognition that, women may have “self interests other than the collective interest of the domestic group” [Papanek, 1990: 181]. In this regard, Agarwal’s landmark A Field of One’s Own [1994] powerfully theorizes ways in which unequally allocated rights in land may lead South Asian peasant women’s self interests to be, and to be felt to be, different than those of household men with whom they bargain. Explicitly extending Sen’s work, she [1994, 54] sees women’s land rights as powerfully affecting their “fall-back position in intra-household bargaining over consumption”, their sense of whether extant land right distributions are legitimate, and the objects for which they bargain.

Many reasons can be adduced for the immediate extension of gender-disaggregated extended entitlement analysis to relations between poor South Asian households, especially to those which are landless. The most persuasive is the current near invisibility of this important dimension of social and economic life, especially in the research literature on Bangladesh. Lacking this kind of information, it is not surprising to us that household-based economic audits of the very poor in Bangladesh are often unable to show how such people survive at all. The poorest households appear to lack sufficient income and local production even to fulfil basic physiological needs. They frequently are found to have no ability to save and erratic income patterns that regularly dip to zero despite essentially constant caloric requirements.(2)

Such unacceptably meagre evidence as can be gleaned from the full range of studies points to the high probability that millions of rural Bangladeshis are protected from absolute starvation or household dissolution by their reliance on the small-scale exchange of inter-household extended entitlements and commodities, which operate as a social and economic safety net. If this observation is correct, what makes such reciprocity possible at all are not laws or legally binding entitlements (the chief concern of Sen), but mutual trust and obligation based on notions of distributional justice and the recognition of others as legitimated social beings operating in the same moral universe.(3)

A necessary condition for poor people to be recognized as full social persons is that they be grounded in space and have a social place. Living in a world of unbounded needs, the moral weight or effective “enfranchisement” of entitlement claims in rural Bangladesh usually is dependent on physical residence in a village and at least marginal acceptance by some – typically kin and close friends – as legitimate participants in village life. Academic ideologies may still so strongly imbue ‘village Bangladeshis’ with “nativeness” [Appadurai, 1988] that there is little felt need to investigate these things further. Yet becoming some kind of socially accepted person is a critical processual challenge facing millions of Bangladeshis each year.

Establishing space and place are especially pressing for people whom fate has rendered absolutely landless, like the environmental migrants under discussion here. For the poorest among them, having a physical place upon which to erect one’s house and carry on everyday household activities is literally a life and death issue. Such people have virtually no monetary or food reserves when displaced from their prior locale.They are left with only their labour power, skills, social ties, personal effects and houses – all of which must be recombined immediately and effectively to provide “food, shelter and energy” [Sayeed, 1991: 15], the basic needs of poor people in Bangladesh. As Kabeer [1991: 258] notes, “ownership of homestead land has an independent significance to the poor, since it offers scope for productive effort and also anchors the household unit to a stable place in the community". Even control of an unowned house plot makes a critical contribution to the household economy of landless people [Jansen, 1987: 109; Indra, 1994]. Denied some long term benefits such as security of tenure and greater social legitimacy that ownership conveys, in the near term such individuals get protection from the elements and a place to muster their slender resources to quickly re-establish a critical cycle of household activities: a place to sleep in private; for women, a place around which to collect fuel and some food in comparative safety and honour and then to prepare food and cook it; perhaps the ability to make use of local networks to offer and receive small loans or contributions of food, or to find work.

Finding a place to stay in rural Bangladesh can be deeply problematic for those without assets. Many others are looking also, for at least 40 per cent of all rural households in Bangladesh own no land [Abdullah and Murshid, 1986: 102].(4) There is virtually no ‘bush’ or uncontrolled government land available in many regions.(5) Rental land is hard to access by people possessing few economic assets other than their own labour. We are convinced that many of the poorest of landless people at some point turn to other households to negotiate a place to stay and hopefully, to eventually achieve an honourable social place. Doing so must inevitably involve asserting and exchanging extended entitlement claims and benefits in complex ways that vary as a function of gender, relation and situation. Our object in what follows is to investigate some of these processes at work in uthuli residence in Kazipur.


II. The Setting

Riverbank erosion is one of the most significant forces changing the physical landscape of Bangladesh, and with it the lives of millions of people. Virtually the whole of Bangladesh is within the deltas of the Jamuna (Brahmaputra), Padma (Ganges) and Meghna Rivers. These rivers drain a 625,000 square mile area of South Asia [Rob, 1990: 366] that includes much of the Himalayas. During the monsoon season the runoff flow is immense. Running across a flat plain composed of unresisting alluvial soils, these rivers frequently change their course, creating new channels or shifting extant ones through progressive riverbank erosion. In a country where rural population densities often exceed 2,000 per square mile, the social impact of the resulting erosion is incalculable. Roughly a million people a year shift residence either as their homes and land are washed away, or to try to take advantage of newly accreted land [Wiest, 1988: 3].(6)

Located 75 miles Northwest of Dhaka, the 138 square mile area encompassed by Kazipur Thana in Sirajganj District is recognized as one of the most erosion‑prone in the country [Haque, 1988: 189; Elahi, 1991: 106-7].(7) Considered poor by Bangladesh standards, Kazipur’s economy is based on highly intensive, small scale rice agriculture. Aid is economically significant, coming chiefly as grain allocated for road and flood embankment construction and to support some extremely poor women. Thana officials and Wiest [1988: Table 5] estimate that by the mid-1980s over three‑quarters of the population was functionally landless. One-fourth of all households had no land at all [Halli, 1991: 129].


Schematic map of Kazipur, BangladeshOf Kazipur Thana’s eleven unions, or administrative subdivisions, Kazipur Union (official 1981 population: 27,735; fewer than 10,000 now actually in residence) is among the worst erosion‑affected.(8) Once this was not so. On regional cartographic surveys done in the early 1920s the most westward branch of the multiply‑braided Jamuna was an insignificant one four miles east of its present course. Informants claim that the first major channel was then a five hour walk away. Kazipur was a stable, orderly, village-based mainland area (khiar), contrasting strongly with the rapidly shifting depositional char islands to the east and erosion-threatened (bir) lands immediately alongside the river. Since then, the nearest branch of the river has grown and shifted westward at an annual rate of over 200 meters a year [Mafizuddin, 1991: 89], such that long ago it destroyed the town of Kazipur, the region’s major marketing and administrative centre. The Union lost about 40 per cent of its agricultural land base (estimated loss, 2,000 acres), the Thana administrative headquarters and four villages during 1968‑81 alone [Haque, 1988: 189].(9) Floods during 1984 and 1988 resulted in further land loss and population displacement.(10) During our initial fieldwork (December 1989‑August 1990) erosion coincident with a flood covering nearly all the arable land in the Union took away several hundred additional acres and important schools and government buildings in the Union’s sole remaining town of Meghai.(11) Severe erosion continued during the 1991 and 1992 rainy seasons, destroying more of Meghai town and parts of two more villages. Although the 125 mile long Brahmaputra Right Bank Flood Control Embankment passes through the Union, its perhaps misguided purpose [Boyce, 1990] is to stop annual flooding, not erosion. Breached repeatedly by the river, periodic embankment reconstruction each time destroys and marginalizes still more local land. Although the Union formally encompasses roughly eleven square miles, its current total land base is less than a third of this – a rectangle two and a half miles north to south by a mile and a half wide. More than half of the union’s people have been uprooted due to riverbank erosion during the past fifteen years, many several times.(12)

A central objective of our ongoing research in Kazipur Union is to find out how variously situated individuals, women in particular, address the calamity of losing their land, livelihood and much of their way of life as previously experienced in long‑standing village communities.(13) Our research program in Kazipur began in 1989‑90 with an initial focus upon 550 households then living directly on top of flood control embankments alongside the river.(14) Both local people and researchers then considered these landless embankment dwellers to be the Union’s poorest environmentally‑forced migrants.(15)

Soon after initiating fieldwork, we found that many equally poor river-eroded people were distributed throughout the villages and homesteads of the Union. As field research proceeded, we found that by loose reputational or by self‑ascriptive criteria as many as 10 per cent of Kazipur Union’s village households were environmental migrants locally distinguished as uthuli because they lived ‘by charity’ on other people’s land. We already knew that Zaman [1988: 88] and Wiest [1988: Table 1] claimed the proportion was 25 per cent on parts of some very nearby riverine island chars in the Thana such as Naturapara, where erosion, land scarcity and extreme social stratification are chronic. It also appeared that most uthuli‑patron relations in Kazipur Union were much more egalitarian and reciprocal than those on Naturapara [Zaman, 1991; 1988: 88‑127]. Landless uthuli on parts of Naturapara in the early 1980s were in a strongly inferior client relationship to landlord/patrons, who could demand from uthuli agricultural labour, household service, political support and service as client lathiyals (part time clubmen) in their battles to control more land. This did not seem to be the case on the mainland.

Studies done in other parts of Bangladesh on peoples’ responses to flood, erosion and landlessness do not mention that displaced people made extensive use of this resettlement option.(16) As our intuitive understanding of uthuli increased, it appeared that in this part of rural Bangladesh the establishment and maintenance of uthuli relations activated women’s kinship resources in interesting and very significant ways, not anticipated by conventional patriarchal characterizations of Bangladeshi family and kinship. Critically, residence – a key land privilege if not a right – appeared to be the outcome of inter-household bargaining in which women played key roles.

These observations made the need for a study of uthuli in Kazipur Union compelling. Besides carrying out many extended informal interviews with and about uthuli, we administered two formal surveys (n = 199 and n = 85) in Bangla to members of uthuli households in every village in the Union during the rainy season (May‑July) of 1990.(17) Subsequent data collection involved mapping uthuli residence patterns (1992), doing extended semi‑formal interviews with ten sets of women and men uthuli household heads and their women and men patrons (1993), securing additional interviews about local perceptions of uthulis (1993), and tracking the situations of those interviewed in 1993 (1994-5).


III. Uthuli Residence and Extended Entitlements

Conceptualizing uthuli women and men as entitled social persons

Establishing the social legitimacy of any personal claim on others is a complex, often difficult and frustrating process for those with few resources at their disposal. Indeed, the state of poverty itself can be seen as the cumulative result of the repeated failure of individual entitlement claims to gain sufficient legitimacy. Unable to remake the world around them, resource-poor women and men in Bangladesh inevitably must make extensive use of legitimation strategies based on the selective, self-interested interpretation and presentation of received moral values and talk about fair and moral responses to particular situations. For displaced people seeking a new place to reside, other Kazipur peoples’ general notions about uthuli status and the moral universe surrounding it are simultaneously a major constraint and opportunity structure. Extensive interviewing and data analysis shows that when not linked to a particular person, these ideas are very uniform.(18)

Uthuli women and men are first and foremost conceived of as poor people who reside on the land of others. Degree of perceived independence and class are mitigating variables. Village people owning solely their own house plots are not considered uthuli because they are conceptually rendered ‘independent’ by their fully entitled residence. However, they are seen to share many personal qualities with uthuli associated with poverty, and for this reason may at times metaphorically be called uthuli. They may even claim to be so when it is strategically useful.(19) Likewise, previously rich people impoverished by erosion and now dependent on others are deemed uthuli, while still well-off landowners, teachers and others residing with others rent free are neither called nor claim to be uthuli. Those living on flood control embankments or along the one pucca (paved) road in the Union are never said to be uthuli, as no one considers residence on government land as being based on the charity of others.

Uthuli have arrived at their present dire circumstances by becoming river eroded people (bhanga elakar lok). As Abul Alim, large landowner and local leader states: “many [once] became uthulis because of litigation or multiple marriage. In earlier days many people would also become uthulis because of their inability to pay back loans taken from the mahajans [money lenders]. Nowadays, of course, these [things] do not happen”. The biographies of uthuli people inevitably show several factors leading to their progressive impoverishment, such as the estrangement of husbands from their fathers and brothers, a chronology of poor economic decisions, or a chronic, debilitating illness.(20) Even so, today such things play only a minor definitional role in characterizing uthuli. Having been river eroded is now the archetypal Kazipur tale of personal misfortune. Hossain Ali, an uthuli man, follows many in linking this misfortune to a traditional proverb: “as you do not hear the cries of birds when their nests are destroyed, you will also not understand the cries, the pains, the sufferings of the people whose houses are destroyed due to river erosion”. In a region where most people have experienced it, river erosion and the consequent loss of homes, arable land and social position imbue displaced peoples’ residence claims with much moral weight.

Uthuli do not pay those who control the land where they reside for the privilege of living there. Kazipur people sharply distinguish uthuli relations from all locally recognized forms of tenancy in which specific payments in cash or kind are made. Uthuli Habibar Rahman nicely summarizes this: “people will not call a person uthuli if he can take land on khai khalashi [rent] in exchange for money or he can purchase land. If one takes land on khai khalashi he can claim some right on the land and from that he gets mental strength. No one will be called uthuli after his/her purchase of land. In cases of bandhak [mortgage] and khai khalashi the land will ultimately go to the owner and that is why people not living under such arrangement may be called uthuli. It is not right to call a person uthuli if he takes land on bandhak but still, people call him uthuli if he quarrels with others”. Even token payments or minor site improvements are seen to confer greater social legitimacy and hence some residence rights. As male patron Nasiruddin noted with some dismay of those living on his household’s land, “they have spent their own money to fill up the pagar [ditch, in which they live]. So they cannot be removed easily”.

Uthuli have been allowed to live on others’ land in recognition of a moral tie signifying a degree of closeness and mutual obligation. That tie usually is one of kinship. Ironically, women and men who can demonstrate sufficiently strong kinship-based entitlements to fully legitimate their residence with kin are not considered uthuli. Most Kazipur people are dependent upon others, often on kin. What distinguishes uthuli status is the unconventional and insufficiently legitimated nature of this dependency. Uthuli lack sufficiently strong personal entitlements to have an undisputed right to live where they are. In this regard, Muslim Bangladeshi kinship ideologies well recognized in Kazipur identify three differentially entitled options available to men in desperate circumstances seeking to re-establish residence with kin. Though subsequently we show it to be untrue in practice, these patriarchal ideologies assert that women do not exercise significant residence choices, as at every point in their lives they are supposedly subject to the residence decisions of men in immediate authority over them. Uthuli Siddique Hossain’s account evokes all three: “Those people are called uthuli who have no house, no land, no cows, no goats and sheep of their own. The son will not be called uthuli if he has rights concerning the land of his father. He will be called uthuli if he does not have a right in his brother’s land. If a woman comes back to her father or brother and if she has a right over their properties then she will not be called uthuli. If any woman has any right in her father’s property then she will not be called uthuli even if she comes back with her husband and sons and daughters. A daughter may live in her father’s property along with husband even if she takes her share [of her paternal inheritance], but socially she will be known as uthuli”.

The best legitimated choice potentially available to a man is to rejoin his father or brothers at the site of his natal household or elsewhere, if erosion or another factor has led to its loss. Such residence arrangements are not considered to be uthuli, as the responses tabulated in Table 1 indicate.21. If this option is precluded, a man might make use of his moral credit with more distant agnates such as his father’s brother or their sons. Households of this sort are often said to be uthuli. Failing in these, a man might consider making a much more marginal claim on his wife’s brother or father. Here, he might call upon long-standing obligations arising from two sources: first, that created by his wife not having taken her share of her fathers’ estate entitled to her by Muslim and Bangladesh government law;(22) second, goodwill generated by his support of his brother-in-law’s sister or father-in-law’s daughter, and fear that if this is not reciprocated it could lead to divorce.(23) While data provided (see below) demonstrate that important women’s entitlements are in fact activated here, a woman in need returning to her father’s house along with her husband is the paradigmatic cited instance of people who are uthuli.(24) In contrast, divorced or abandoned women returning to their father’s or brother’s home are seen as being entitled to protection, and are not thought of as uthuli except perhaps when they earlier took their share of their father’s inheritance.

As Table 1 shows, these patrilineal kinship ideologies contribute to uthuli status frequently being associated with residence with a wife’s close kin rather than a husband’s.

Relationship of patron granting residence to uthuli

Informant considers uthuli

Informant claims others may consider uthuli

Informant does not consider uthuli

Husband’s father




share in property taken




share not taken




Husband’s brother




share in property taken




share not taken




Husband’s father’s brother




Husband’s father’s sister




Husband’s sister








Wife’s father




share in property taken




share not taken




without husband




share in property taken




share not taken




with husband




share in property taken




share not taken








Wife’s Brother




Wife’s father’s brother




Wife’s sister








Table 1. Kinship relations and ascribed uthuli status(25)

Uthuli therefore are people partially out of place. Uthuli status is associated with marginality, impermanence, shame, disempowerment, wildness, impropriety and lack of voice.

It is a matter of shame to live as uthuli because many people say many things about them. The village people express annoyance with us because we have increased the number of people in the village, which they do not like. The villagers rebuke us because we rear goats, hens and ducks. We have to live here keeping our heads always down. We have to do this more [than others] as we rear goats and cows. For this, people say, ‘How many khadas of land you have here that you rear cows and goats?’(26) We teach our children to keep their heads down. It is more shameful for women, as we do not have any land here even to build a latrine. (Aktar Ali, uthuli man)

As a cultural construct, uthuli relations occupy – and are dependent upon manipulating – the ambiguous moral middle ground. They involve neither individuals so socially close and well entitled that they can demand and honourably claim the right to live on someone else’s land (something rarely automatic in practice), nor those so socially distant that only the calculus of classical exchange entitlements and contract applies. To call someone an uthuli is derogatory, and it is sometimes used as a general term of abuse. The shame and stigma associated with the term largely arise from loss of economic independence and (especially for men) key agnatic extended entitlements; from the association of uthuli with dire poverty, lack of control and lack of privacy; and from the widespread sense that they have no right to voice their opinions publicly or to participate in village decision-making. They must always “keep their heads down”. As Chan Banu indicates, “we live here in such a way so that nobody can find any occasion to say anything. ‘The bird sits on the branch which can withstand its weight’ (je dale pakhi bhar sai, se dalei pakhi base)”. Potent regional culture/nature symbolism contrasting mainland people and island char dwellers also distinguishes ‘permanent’ villagers from uthuli. Uthuli are seen to be on the margins of society and often to be “wild,” though not so much so as those residing on embankments and alongside public roads. Sometimes, the natural symbolism applied here is very direct. As a government official comments, “the very fact of living as a uthuli is matter of shame. The village people do not look on them with good eyes. Many address them as ‘chika’ [a local kind of rodent]. They think that the uthulis spread bad smells like the chika, destroy the beauty as well as cleanliness of the houses, and create an undue lack of privacy”.

Uthuli residence is something to be avoided if possible, but it is widely recognized that often there are no better alternatives. While uthuli status is shameful and marginalizing, it is also linked to many positive qualities associated with everyday life in settled villages: comparative security and privacy, some access to kin, association with people of honour, a claim on a respectable collective identity, and the potential to develop reciprocal economic exchanges with patrons and others. Despite its stigmatic associations, uthuli status is seen by many in Kazipur as much more orderly and respectable than ‘squatting’ on government land: “[In contrast to village uthuli life,] in the C&B [pucca road ‘squatter’ settlement] there live ‘thirteen people of twelve villages’ (baro gayer taro jon) resulting in the absence of harmony among people there” (Bisha Sheik, uthuli man);“It is more shameful to live on the embankment because there is no baobera [fences] or jhup jangal [bushes]. There is no purdah” (Abul Hossain, uthuli man). Moreover, there is a strong local sense that uthuli are doing the best they can under very difficult circumstances. Two aphorisms often repeated in reference to uthuli are:

The cultivators have no shame, the fisherman have no shame (Laj nai hailar, laj nai jailar) [for doing what cannot be otherwise]. (Alea Khatun, uthuli woman)

 We do not have branches that fall, we do not have wings with which we can fly (Parar moto dal nai, urar moto paha nai). (Hasina Khatun, uthuli woman)


Establishing Uthuli Relations:
when ‘marginal’ women’s entitlements critically matter

Thousands of women and men in Kazipur Thana have lost critical resources because of erosion and consequent uprooting: land, if they had any; animals and other capital goods sold during uprooting; long term labour relations (especially if they were maintained in a now eroded region); some of the symbolic attributes of normal family life for both women and men. They face dismal economic prospects. There is little land available in Kazipur for sale, sharecropping or rent, and most men join a much under‑employed pool of kumla (general, unskilled) labour.(27) Men’s social resources (especially their kin‑based extended entitlements) usually are nearly exhausted.(28)

People prefer to re-establish themselves within the Thana, even if over one‑half of displacees currently may leave.(29) Those who were already functionally landless have practical reasons for resettling near their samaj, past employers and patrons, most often giving the economic and social benefits these confer as the reasons why they have re‑established themselves nearby (Zaman, 1988: 115).(30) Those who own significant amounts of land now under the Jamuna try to remain nearby in the hope that they might regain their endowments should this land subsequently re-emerge.(31) Those who do stay usually re-establish themselves within two miles of their last residence [Haque, 1988: 217]. Our analysis of the migration patterns and natal villages of uthuli shows that they are local people also.(32)

The residence options of poor eroded people wishing to remain in the area nevertheless are very limited. While the kin ideologies outlined previously assert that men in desperate circumstances have certain ‘rights of return’ to their father’s or brother’s households, the latter often are economically marginal and many are landless too. Ideologies of assistance and legal entitlements notwithstanding, landless environmentally‑forced male migrants have little to offer their agnates. We collected many accounts of how their return posed an economic burden, a threat to the entitlements and statuses of those already there and a potent source of familial disharmony. We do not know what proportion of environmentally forced migrants in this region establish residence with or adjacent to the landed close kin of family men.(33) However, to use the terminology of Appadurai [1984], we are certain that this entitlement is not consistently “enfranchised”, and that there is considerable resistance among settled villagers to such arrangements.(34)

And yet there are few alternative ways for Kazipur people without means to establish independent residence. People who possessed farm land immediately prior to being eroded often are able to rent a house plot or even a little agricultural land. Instances of already landless people renting more than a house plot are exceptional, due primarily to high initial rent charges at a time when household financial resources are in disarray.(35) The typical current rent for a small house plot of an eighth of an acre for four years is Tk. 350 to 500, equivalent to 20 to 50 days income for a male agricultural labourer.(36) Government land in the Union upon which people can (illegally) erect houses is limited to the top of a network of several miles of active and abandoned embankments and alongside Kazipur’s one pukka road. These now are occupied almost fully. Under such restrictive circumstances many poor people have considered other options, such as becoming uthuli.

Almost all uthuli we contacted throughout 1990-3 certainly were very poor by local standards, and most had been so before becoming uthuli. Often they reported a dramatic deterioration in their economic situation over time.(37) For example, Abu (70) and Sharifa (57) both were born into well-off farming and business families in Manikpotol and Maizbari, east of Kazipur. Both their fathers owned well over 100 bighas of land (33 acres), placing them among the local élite. After marrying, they lived almost twenty years in Abu’s natal household, eventually leaving because of conflicts between their children and those of Abu’s four brothers. They then farmed 16 bighas (roughly five acres) independently. In 1981 their home in Maizbari was eroded (eventually along with the whole village), and they moved to Meghai, still farming 7-8 bighas of their own land. In 1990 this land and their Meghai home were eroded also. The river’s force had also been directed to Manikpotol, dispersing Sharifa’s brothers’ families to the embankment and roadside ‘squatter’ communities. Abu and Sharifa secured a place to stay from Zafar Ali, a very poor distant relative who ironically once was their long-time household servant. Though they paid nominal khai khalashi rent for their house plot, local people consider them uthuli. They are extremely poor, living primarily on remittances from two sons who work in Dhaka garment factories and gifts from their daughter and son-in-law. They still hope to live to see some of their land re-emerge from the Jamuna.

However, poverty alone does not lead to uthuli status. In good accord with local folk expectations, what is necessary is the demonstration of immediate dire circumstances. Most frequently this is displacement by the Jamuna; for women-headed households it may be widowhood, divorce or abandonment.(38) 94 per cent of uthuli interviewed in 1990 and all of those surveyed in 1993 had been eroded, one‑quarter of the former more than once. Of these, almost every household either moved directly from an erosion‑threatened house plot to become uthuli, or else stayed a short time on the embankments in-between. Indeed, the proportion of environmentally forced migrants among Kazipur uthuli is as high as it is in nearby embankment communities stereotypically associated with this phenomenon. Clearly, uthuli residence has become an important local adaptation to persistently trying environmental conditions and loss of land and houseplots in the absence of more viable residence alternatives. It is also an instance of how ecologically-driven redistribution of property control and access can change local ideological constructions [Agarwal, 1994: 15-16].

What makes some potential patrons respond positively to this acute dilemma that annually faces so many? When the population pressure was less a generation ago, mainland Kazipur landlords occasionally allowed unrelated people to place houses on their land for long periods of time. Uthuli men were either already household labourers or servants, or else were thereafter virtually obligated to provide labour on demand. Uthuli women provided a variety of unpaid household-based services. A variant of this system is said to be still in place on nearby Naturapara char [Zaman, 1991; 1988], where some landlords had as many as two hundred uthuli client households in the early 1980s (Zaman, personal communication). Today, few patrons in mainland Kazipur have even two resident uthuli households. Not one has the kind of neo-feudal power and wealth characteristic of the big talukdars and jotedars of Naturapara.(39)

Another current local distinction is at least as striking. In Kazipur what is increasingly critical to becoming uthuli is the ability of individuals to exercise kinship-based entitlements. Only 23 per cent of uthuli households polled in 1990 secured homesteads from individuals lacking a kinship tie to them.(40) Moreover, our data shows an increase over time in the proportion of uthuli relations established between kin. As a case in point, over 90 per cent of those interviewed in 1990 who had been uthuli for three years or less claimed a recognized kinship relationship to their patrons; the rest were between people who were essentially fictive kin, chiefly very close friends and village mates.

A host of recognized kinship relations could potentially be activated here, but only some are with any frequency. An illustrative case is that of Fatema Khatun and her husband, Habibar Rahman: after one year of their separation from Habibar’s father’s household their houseplot was destroyed by river erosion. They then built their house elsewhere in Maizbari village on land given to them by his father. Five years later, that houseplot also was destroyed by river erosion and they left for Meghai, where they were again given land by Habibar’s father. When that was submerged his father and others went to the house of Habibar’s mother’s father in Manikpotol, where Habibar’s mother had 2.5 bighas of land. Habibar and Fatema did not go there since they thought that they would not get along. They stayed for a few days on the embankment after their houseplot in Meghai was destroyed. When the embankment was eroded they then took shelter in the house of the brother of Fatema’s aunt in the centre of Meghai. They lived in this place for one year, but when that house started eroding the owner of the house requested that they leave. At this stage, Fatema Khatun’s sister’s husband, Minhazuddin, took them in as uthuli.

Three more general points are relevant here. First, this example illustrates how continuous social disruption brought on by incessant erosion of settled villages has progressively exhausted poor men’s patriarchal entitlements, particularly with respect to more distant agnates. Only 27 per cent of those uthuli resident with kin in 1990 were living with an agnate of the husband or with the affines of such kin (see Table 2). Second, this correlates with an increase in the visibility of women’s entitlements and dependence on women’s kin in times of need. Remarkably, two‑thirds of kin-based uthuli relations were formed with the wife’s (or widow’s) agnates or their affines. The most common arrangement was to live with the wife’s father (30 per cent of all kin‑connected instances, and 45 per cent of those related through wives).(41) Seventeen percent lived with the wife’s brother and 6 per cent resided with daughters.


Relationship of Patron                      Percentage


I. Related to uthuli wife/widow

           wife's father or mother                     30%

           wife's brother                                   17%

           wife's sister                                        8%

           wife's mother's brother                      5%

           wife's father's brother                         2%

           wife's mother's sister                          1%

           other                                                  3%

                 subtotal                                                 66%        

II. Related to uthuli husband                                         

           husband's sister                                  6%

           husband's brother                              5%

           husband's mother's brother                5%

           husband's father's brother                  4%

           husband's mother's sister                   3%

           husband's father or mother                2%

           other                                                  2%

                 subtotal                                                27%

III. Daughters of uthuli                                        6%


           Total                                                             99%

Table 2. Kin relations between those identified as uthuli and their patrons(42)

Our third point bears directly on the centrality of women’s inter-household entitlements. Frequent dependency upon women’s relatives here is not simply the result of residual male kinship entitlements, as Bangladeshi patriarchal ideologies would suggest. It is certainly true that uthuli relations do reflect a heavy use of women‑mediated connections between men who could not otherwise make a successful claim on each other. Fully 80 per cent of uthuli relations based on kinship ties involve women in this fashion.(43) However, we have persuasive evidence that establishing uthuli relations activate women’s entitlements much more directly than this. For one thing, all uthuli and their patrons questioned on this point in 1993 explicitly acknowledged a daughter’s limited right to return to her father’s, brother’s or son’s household. This included all cases where women actually had established this kind of residence. No one ever spoke of the rights of these women’s husbands. In fact, widowed, abandoned or divorced women often are able to enfranchise such claims so strongly against their sons or fathers that (in analogy with a son returning to his father) they are not consistently called uthuli at all.(44)

Women’s’ entitlement claims are weaker if women are married and their husbands intend to take up residence. As in other parts of Bengal [Dyson and Moore, 1983: 46-53] couples typically stay in close contact with the wife’s parents. The historical ideal is that relationships between husband and father‑in‑law and between father and daughter should be warm and mutually supporting. When this is so, both husband and wife may then go to her father and ask for a place to stay “for our children” with a good chance of not being denied.(45) Even so, in close accord with general notions about what makes someone uthuli, poor couples residing with the wife’s father actually are routinely classed as uthuli, while women returning alone usually are not. A widower loses this claim on his father-in-law.(46) In contrast, a widow may assert residence claims on both her own and her husband’s kin as well as on her children.

Uthuli residence is also dependent on active, effective and overt negotiation. We came across many instances where women directly initiated and presented such requests, and were the key players bringing negotiations to a successful conclusion. Here are two representative verbatim accounts:

The house of my husband was in Muslimpara. I feared the destruction of river erosion and that is why I have come here [to Beara]. If the river comes here I will go anywhere possible with others. [Before we came here] I told my brother, “how can we stay here?” My brother said, “if you cannot stay there, come here. But I will not be able to give you food. You will have to earn your own living. I do not have the ability to provide you with food". [So] I talked to my brother before I came to this house. One day I suddenly came to this house, consulting with nobody beforehand. I convinced my husband to come here. I did not say anything about staying with anybody before we came to stay here. My brother’s wife [now] wants to evict us. However, my brother does not say anything, considering that I am his sister. I do not know how long we will stay in my brother’s house. Either we will go to our own house or to the roadside. We would like to go to my sister in the C&B [road], but there is no land there. Allah knows what will happen to us. I wanted to stay with my brother. I am a girl of this village. Moreover, my brother wanted to give shelter. I also am entitled to come to my brother. So we came. (Zarina, uthuli woman)

We started living in others’ houses when in 1988 my husband sold his homestead land after the flood. After two months of living as uthuli [in the household of her husband’s adopted father] we wanted to stay as uthuli in my father’s house. Then I did not know anybody except my father and therefore, we came to his house. I myself talked to my father in this respect. [Thus] my father found that his daughter had no place to go. So he brought us. We did not speak of any condition [of residence]. We also do not know how long we will be here. We will stay here if we do not get any better chance anywhere else.(Paruli, uthuli woman)

Finally, we have some evidence that in some situations achieving uthuli residence may thereafter further empower women and enfranchise certain women’s entitlements, especially by strengthening their fall-back positions [Agarwal, 1994: 54, 261]. Men divorcing and abandoning wives in situations of extreme poverty is common in Kazipur, as elsewhere in rural Bangladesh. Most such instances can be read as the disenfranchisement of women’s extended entitlements by household men. Yet the only limiting instances we found of husbands abandoning wives while they were uthuli were where husbands had found permanent work far away and rarely returned. This may suggest that the very dependency typified by becoming uthuli constrains or re-priorizes these male options and so temporarily enfranchises wives’ entitlements to ‘support’ and honourable status within their own households. As people out of place, uthuli households are under intense ‘surveillance,’ particularly by patrons and others living nearby. Accordingly, uthuli divorce or abandonment would be shameful for patron kin, and would be a recipe for future conflict between uthuli men and others. Interview data also demonstrates a general pattern of greater confidence, energy and assertiveness among uthuli women than men – though we cannot say whether this can be attributed to being uthuli.

We believe that these findings and other supporting evidence that follows underscores the point made more generally by Agarwal [1990], White [1988] and Kabeer [1990] that while Bangladeshi men typically command much more significant resources than women, this does not make women passive. They are active, and though comparatively strongly disadvantaged, may exercise a range of important entitlements that are rendered invisible by focusing exclusively on normative gender ideologies, men, statistical averages summarizing complex and varying behaviours, or on what happens around the hearths of particular households.(47) They also are capable of measurably affecting what is considered acceptable – even doxic – and are far from complete captives of doxa or ‘tradition’ [Agarwal, 1994: 54-60].


Endowments, extended exchange entitlements and
the benefits of being uthuli

As Drèze and Sen [1989: 71] note, “rural communities faced with a precarious environment often develop sophisticated institutions and strategies to reduce or cope with the insecurity of their lives". Uthuli-patron relations clearly exemplify what Drèze and Sen refer to in this context as “informal security systems". Speaking globally, they characterize such informal security systems as typically exploitative, and “neither ‘costless’ nor even particularly attractive for poor households” [Drèze and Sen, 1989: 74]. Becoming an uthuli does engender economic, social and psychological costs about which people are clearly aware and as illustrated above, individuals locale uthuli residence far from their ideal, both perceptually and in the choices they make. Even so, it would be inappropriate to characterize uthuli residence as a straightforward example of the exploitation of the super poor.

Consider economic relations between uthuli and patrons. While over 90 per cent of patrons own, rent or sharecrop some agricultural land, typically this is insufficient to support the patron family itself. Nearly all patrons are poor, and a few even are uthuli themselves. In most instances, patrons are incapable of providing extensively for uthuli or of exploiting them greatly: they lack the means for the first, and the resource base and motives required for the second. A quantitative assessment of economic values exchanged between patrons and uthuli, so narrowly-framed that it neglects to factor in thecritical benefits to uthuli of houseplots as sites of household production and reproduction, probably would show a net benefit accruing to patrons.(48) A more all-encompassing accounting may not.

In any case, over dependence on net balance tallies can mask ongoing reciprocity from which both parties may benefit, even if unequally. Houseplots allocated to uthuli are always small, and usually are threatened by future erosion, flood-prone, or of marginal quality. Patrons lack alternative uses for them that would materially affect their economic condition. Even so, patrons claim that by allowing families to settle there they forgo more modest options such as planting the plot with vegetables. As patrons usually lack means to easily remove uthuis once they have become established, they also place themselves at a small risk of losing control over the land in question. Gifts of food, clothes, and money to uthuli likewise may be very modest, but patrons and uthuli often are so poor that they are both an important contribution to uthuli household economies and a definite economic cost to patrons.

In exchange for the use of a houseplot and such gifts, women and men uthuli frequently contribute unpaid and underpaid labour to patrons, often at tasks not associated stereotypically with male, paid agricultural labour (such as harvesting and weeding). When patron women are overtaxed, uthuli women may help clean their houses, cook, carry water, help in food preparation, prepare grain for storage, or tend patrons’ children. Almost always this receives no direct compensation or only meals and rice. As Jamila Khatun declared while the Jamuna raged only thirty feet from her and her patron’s houses, “If the patrons have a great deal of work, the work like bringing of bundles of paddy at home from the field, husking them or tending cows is done by the members of my family. I myself winnow rice and paddy, and clean, wet, dry, and husk paddy. I also clean houses and hold their children on my lap. Wages are sometimes given in kind [i.e. rice or food]. Cash money is never paid. But after the seasonal work is over they give us some paddy to prepare special cakes. This is a part of culture and social custom here". So it had been for her during the twenty five years she had been an uthuli in that place.

Roughly 40 per cent of uthuli men do occasional work for their patrons: marketing, tending animals, moving agricultural products from place to place, and sometimes harvesting, moving soil and weeding. For lighter, more episodic work men are rarely paid a wage; this is an ironic instance of the ‘feminization of poverty’, as their dependency on others has led men to accept working conditions that are more conventionally associated with poor, disempowered women. Many uthuli men are ill or otherwise incapable of sustained hard, day-long agricultural labour. When they do conventional men’s agricultural work for patrons it is usually paid at the market rate (about Tk. 12-15 and meals throughout 1990-3), though sometimes specifically at about 30 per cent less.

Few men who work for their patrons see this arrangement as wholly exploitative, even if they express ambiguity about it. Jamila Khatun’s husband is fairly representative. Ill and no longer capable of sustained earth-cutting work, Nurul Islam readily acknowledges that he and his son get standard wages when doing ‘men’s work’ for their patron. Yet he is resentful about contributing unpaid labour:

Even flies take advantage of the dead (moira gele machhio jot paye). I have worked many times for our patron when they never pay me any money. Uthulis live like dead men. The land they get to live on [in fact] is duly paid. Patrons make them work sometimes without any wage. They take benefits from them. I have given labour many times when I have not got any wages due. Many times I worked for the whole of the day but did not get anything to eat and ultimately passed the day without food. People do nothing these days without profit. Most of the land of the patron has now gone under water and so the volume of work has declined. The amount of work has declined, but still I have to do most of the work of tending cows. For this I do not get anything.

As Shafiqur Rahman, the Chairman of a char union parishad observes:

uthulis say that they get food and money in exchange of their labour. This is right. But when there is a cloud in the sky and the patron has 10 maunds of paddy spread on the courtyard the uthuli help to take the paddy to cover without being requested. They also do the same with the jute that is drying. They bring the things from the market to the patron’s house and also go call anybody and everybody at the instruction of the patron at any time. They get nothing for these petty jobs, but the patrons get their work done. The value and extent of this service is well realized by the patrons.

And yet uthulis also secure significant economic benefits, many of which involve women’s agency directly. As noted previously, uthuli residence makes it possible for women to carry on a wide range of productive and reproductive activities vital to these households’ very existence. It does not do this to the same extent for uthuli men. In both the initial stages of becoming uthuli and always thereafter, uthuli women’s steady, low risk household work lays the economic foundation that makes it possible for healthy men to explore a range of paid labour options in a shifting, episodic and very uncertain market. Women cook, process food, collect fuel, water and edible plants. When allowed by their patrons, they grow vegetables around the house, and maintain chickens and goats.(49)

Women’s economic contributions extend much further than this. Few uthuli nuclear families could survive on a wife’s household labour and a husband’s often-erratic income alone.(50) In any case, 16 per cent of uthuli households lack adult men and so are entirely women-supported. In many others illness or old age have rendered men incapable of steady work. Still others are women-managed for long periods of time while household men work elsewhere.(51) Purdah conventions in Kazipur are less restrictive than in many other parts of Bangladesh, but they nonetheless profoundly limit poor women’s prospects for finding steady work outside their immediate environs. In only 6 per cent of uthuli households surveyed in 1990 did women work consistently outside their own and patrons’ households.

Economic benefits secured as a consequence of their relations with patrons often make up part of the overall shortfall. Meals and contributions of food secured by women through their work for patrons and other nearby households constitute significant household economic inputs. Like other poor women living in villages, uthuli women soon develop complex inter-household patterns of reciprocated small scale loans of food and money. These help families get through times when family monetary and food reserves are exhausted because men go away to work, local men’s work is unavailable or household men are ill. As one uthuli man noted, “people are afraid of extending loans to the poor people,” especially if the people in question are socially distant or unobligated. Here, virtually all uthuli women negotiate and secure such loans at least from their patrons – who also typically are their kin.(52) Women also try to maintain a jola (hidden personal savings), and they are the primary conduit through which occasional gifts of food flow between patrons and uthuli. If left with no alternative, some will beg food from their neighbours and others; men will not.

Uthuli men benefit from a general tendency among farmers in need of labour to hire kin first [BRAC, 1986: 44] and for patrons to hire uthuli men. As long as they do not become estranged from their patrons, uthuli men (who all have a surplus of available labour to contribute) can often count on securing a certain amount of income from this source.

Still, uthuli residence is not just a matter of economics. Social and psychological tradeoffs are also involved – tradeoffs that were assessed quite differently by one person than another considering their resettlement options. Both uthuli and embankment dwellers frequently mentioned the many social and psychological costs entailed by uthuli residence: being burdened with ongoing feelings of marginality, shame and disempowerment; having to face name-calling and a range of muted and more direct assertions that they were inappropriately dependent people out of place; being excluded from some social groups and activities participated in by those with more conventional village statuses; and especially, having always to carefully manage their relations with patrons in order to diffuse tensions and head off inter-personal conflicts that might otherwise lead to demands that they go elsewhere. Even so, two-thirds of uthuli women and men interviewed in 1990 preferred their current arrangementsto ‘squatting’ on embankments, identifying several personal, social, and political net benefits. For example, in contrast to embankment life, those who were uthuli said that it provided more honour-maintaining privacy and a greater sense of personal safety. This was particularly important to women, who claimed to benefit socially and psychologically from the greater physical seclusion that village life provides: screened cooking areas, wells located in or near household areas, more substantial latrines, more conventional places to wash, and less transient human traffic. To even nominally control a village house plot is also to control a home: a geographically‑defined, socially-honourable protected space that allows women a base of operations, and a range of choices as to when and why to go ‘out’.

Over time, uthuli women and men become increasingly active in everyday village affairs. Should they stay for years, uthuli may become part of a village samaj, and eventually may participate as much in local rituals and social affairs as other similarly poor and disempowered village people.(53) Such village-based activities are strongly kin-linked, and so long term uthuli residence may lead such families to become closely tied socially to the wife’s relatives. Uthuli may also be able to use patrons and influential village people as brokers to find work or secure food aid. As most patrons themselves are poor and disempowered, typically they can neither provide much political protection nor use uthuli political support themselves. Again, this is in sharp contrast to how Zaman [1988: 88‑91; 1991] has characterized politics on Naturapara char, where a few powerful landlords solicit many households to become uthuli in order to exploit their consequent dependency as “cheap wage labourers and sharecroppers” and political clients.


Women’s Agency and the Provisionality of Uthuli Relations

We have not yet tracked many households over their whole time as uthuli. However, length of residence data collected so far show that uthuli arrangements typically are quite transitory. Environmentally‑forced migration, the death of husbands and other factors lead to the establishment of many new uthuli relations every year. Even so, 56 per cent of the Union’s uthuli in 1990 had been so only for two years or less, and three‑quarters for seven years or less. Only a handful of adult uthuli had been born and raised as such. Erosion-generated population displacement over the last decade and a half has been particularly high, and without longer term data it would be unwise to assume that the rates of formation and dissolution of uthuli households are in balance. If this were so, however, a quarter of all resident uthuli would abandon this status every year, while another one out of fifty other Kazipur households would become uthuli.

Data collected in 1993 point to two broad reasons why people quit being uthuli. One ‘pulls’ many out of this status: the desire to escape the shame and inter-household discord they face makes many families try hard to make the transition to ‘independent’ residence in villages – even when the economic benefits that accrue are unclear. What is minimally needed to make this transition is the willingness of someone to rent one a house plot, the Tk. 500 or so needed for rent, and the moral authority to get others to recognize that such a rental arrangement signifies a significant status transition. People call upon a wide variety of sources to secure the needed money – savings, remittances from daughters and sons working in Dhaka, selling remaining capital goods or house materials, etc. Others decide to resettle on embankments rather than to face the shame, social conflict and dependency entailed by continuing as uthulis. This transition is facilitated by evolving embankment and roadside community values that highlight the notion that people achieve greater freedom and symbolic autonomy of residence there and downplay local assertions of its wildness and unconventionality.

Equally powerful ongoing forces ‘push’ uthuli out. Conflict over residence rights is now endemic in rural Bangladesh, even when prevailing kinship ideologies appear to strongly legitimate personal entitlements, such as to the continued residence of sons in their natal households after marriage [Aziz, 1979: 51]. In situations where entitlements are more marginal and personal power less, such as when widowed women attempt to control family land without the support of adult sons, others may eventually drive the individuals in question away. This kind of struggle shows up frequently in uthuli‑patron relations, which often exhibit many indications of instability and low grade discord. Considering those with whom uthuli typically reside (Table 2), in almost every case there is a high probability that someone with greater “assets and personal force” [Maloney, 1988: 54] will find some aspects of the relationship in conflict with their interests.(54) Individual patron self-interest and narrow familialism cultivated in situations of chronic want and insecurity often threaten to undermine the moral force of uthuli claims.

Uthuli women play an important role in heading off and defusing conflicts of this sort for as long as possible – conflicts which often are initiated by patron women. Consider situations of uthuli residence with a woman’s father or brother. Uthuli women do what they can on an ongoing basis to shore up the legitimacy of their claims. Key strategies here are to emphasize the closeness of this natal family connection, their own family’s chronic need, and the latter’s efforts to regain a more independent footing. They will also do much day to day emotional work aimed at their sisters-in-law and other key patron household women, and frequently back this up with significant contributions of labour that lighten the latter’s workload in the expectation that this will be reciprocated by greater acceptance.



Research on the poor in South Asia exhibits a great irony. Most of this research has some kind of a development connection. This ensures at least tacit subscription to prevailing ‘grassroots’ development axioms, emphasizing the importance of understanding and engaging individual agency and everyday community-based participation in the alleviation of poverty. And yet researchers historically have not been centrally concerned with either of these dimensions of social life. Particularly in Bangladesh applied studies, the focus instead has remained strongly on poor rural ‘households,’ which are still often treated as isolated, corporate, seamless wholes disconnected from social space.

It is not hard in retrospect to see how this came to be. Up to 1980, research characterized Bangladesh family and kinship quite similarly to ideal élite folk ideologies: family relations among Muslim Bangladeshis were said to routinely empower men at the expense of women and those in their productive years at the expense of the rest. Kinship relations were typified by the strong solidarity of close, co‑resident, male agnates. Women were portrayed as having little power in family affairs and less in inter‑familial kinship relations. Perhaps too imbued with older notions of peasant familialism, researchers wrote easily about the autonomy of interests of specific ‘households’ and the solidarity of those within them rendered necessary by the hostile world outside. A “...long-standing assumption in economic theory and development policy... [was then] ...that the household is a unit of congruent interests” [Agarwal, 1994: 3].

Secondarily, it was a case of methodological determinism. Operating under a governmental and sponsor panopticon of ‘accountability’ in countries with large populations and few investigative resources, even ethnographically-oriented researchers came to depend strongly on survey methods to produce ‘reliable’ data on rural poverty. The intrinsic strengths and weaknesses of survey methods aside, practical matters involved in their actual use consistently highlighted ‘the household’ as the central unit of study – and in a highly androcentric way.(55) As Agarwal notes throughout A Field of One’s Own, this practice continues to be rendered plausible today by the hegemonic weight of ‘the family’ and ‘the household’ in research discourse, and how it has affected the evolving course of entitlement theory.

Fortunately, over the past decade and a half the increasingly obvious limitations of using ideal models to explain empirically‑observed Bengali families, households and kinship have led to their progressive elaboration, particularly in studies of families and individuals coping with economic calamity.(56) Case studies showed that even ideal folk notions of how family and kinship relations ‘ought’ to be configured varied across Bangladesh as a function of gender, religion, region, class and circumstance. Hartman and Boyce [1988] and others brought forward instances where ideal conceptions of family and kinship and actual behaviour diverged dramatically. It was also shown that in everyday life the ideal conceptions that critically mattered were in essence locally applied family ideologies: ideas selected and emphasized by specific, socially‑contextualized individuals and configured for and through strategic practice in particular situations. What happened in actual families, households and kinship networks thus differed dramatically as a function of personal interpretations, resources, strategies and goals.

An important element being introduced here was already central to family studies elsewhere [Yanagisako, 1979: 190‑200; Collier, Rosaldo and Yanagisako, 1982; Lamphere, 1974]: greater attention to individual initiative, “agency” [Sen, 1990: 127], decision‑making and “bargaining” [Agarwal, 1994: 54-72]. As mentioned, so far this has chiefly focused on how individual women and men in households seek to achieve (culturally‑constrained, but also idiosyncratic) personal and familial goals through the strategic use of entitlement claims. This approach reduces the tendency to reify household and family. It also suggests the means by which a ‘grammar’ of family and household arrangements might be produced to derive a range of empirically‑observed situations out of a somewhat more uniform and invariant set of cultural ideals. By introducing notions of struggle, negotiation and diversity of interests into the study of the everyday fabric of family and household (including what nominally appear to be ‘stable’ intra‑household arrangements) it sensitizes one to the necessity to analyze role‑specific access to power, privilege and prestige more flexibly and contextually. Clearly, it is now unjustifiable to ignore women’s power in ‘patriarchal’ families simply because culture-based family and kinship ideologies assert that men monopolize power, or because in many instances men in fact nearly do. Women are active, if unequal, participants in such patriarchal systems, who often try in both conventional and innovative ways to advance what their perceived interests while operating under severe constraints.

The further incorporation of struggle into the analysis of contemporary Bangladeshi family, household and kinship relations is particularly important in the light of the radical economic, political, cultural, social and ecological transformations that poor Bangladeshis face today. To date, though, the full potential of this approach has not been developed. Indeed, one still sees throughout the literature many themes predicated on earlier, more monolithic and ideal notions of family and household, for example, of “nuclearization” [M. Islam, 1985: 44], the “decline of patrilineal groups” [Arens and van Beurden, 1977: 102], “breakdown” [Kabeer, 1989: 34; 1991: 257] and individualization [Maloney, 1988]. Such processes as poverty, exploitation, functional landlessness, underemployment and forced migration have supposedly reduced adherence to received dictates about how family, household and kinship ought to be, broken down and weakened inter‑familial kinship ties, and increased what Maloney [1988: 52] terms “Bengali pragmatic individualism”. Poor women are said to be disproportionately the victims of these linked changes, losing what protections the historical patriarchal system might have once conferred without gaining compensatory benefits. Framed differently, this is a thesis of the massive, progressive disenfranchisement of rural people’s key extended entitlements, women’s in particular.

While much case-by-case evidence supports this very global thesis, taken as a general truth it poses deep risks of muting diversity.(57) It may blind researchers to the many ways in which the poor may come to new ways of thinking about family, household and kinship, particularly in regard to relations between individuals in different households. As we have said before, for Bangladesh as for South Asia, we have as yet no extended entitlement theory (gendered or otherwise) applicable to inter‑household relations to complement those at an intra‑familial and macro level – or even as yet many calls for its development.(58) Moreover, as Agarwal has stressed, entitlement analyses continue to be too dependent on notions of rigidly delimited entitlements supported by law or extended entitlements by fixed, highly legitimated custom. Innovation so far enters entitlement theory almost exclusively either in discussion of the disenfranchisement of entitlement claims, or their strategic use. Yet even ‘legal’ entitlements are themselves contentious. At the heart of such struggle is the potential for individuals to frame extant entitlements in a novel way, and to generate ones that are entirely new.

Uthuli residence in Kazipur is an illustrative case in point. People on the mainland in Kazipur have had to deal with erosion and consequent displacement for almost two generations. This has led to many local cultural adaptations – adaptations evolved through ongoing interpersonal struggle and negotiation. We have addressed one set of such adaptations here, showing how some Kazipur people have taken a ‘traditional’ concept and through conflict and co-operation modified it to better fit their new circumstances and interests as poor, environmentally forced migrants. Many Kazipur conceptions of uthuli residence and notions of family, household and kinship that envelop it are novel, and cannot simply be placed on a continuum at some distance from the ideal. In addition, uthuli and their patrons are far from atomized. In dealing with each other, they demonstrate a kind of mutually beneficial reciprocity and mutual aid seen in other contexts among all categories of displaced poor of Kazipur. Indeed, though often unequal, uthuli‑patron relations are sufficiently balanced and reciprocal to force a reconsideration of the widely‑held thesis that in Bangladesh “in the event of crisis, a household must fend for itself. Relief from neighbours 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]. This kind of mutual aid still largely falls between the cracks of contemporary paradigms, with their strong focus on men, on intra‑household relations, and on relations between households, markets and the political system.

Moreover, it is inappropriate to view ideas and behaviours relating to uthuli as pale reflections of the ‘true essence’ of Bangladeshi culture or “impoverished emulations” [Kolenda, 1987: 295] of what the well-off do. Forced migration has radically altered peoples’ individual resources, as progressive impoverishment does more generally. This has required local people to reconsider their personal strategies and goals and to evolve new ways of thinking and acting. This is not just an instance of the weakening of men’s patriarchal entitlements, which of course can be read as the ‘loss of the traditional.’ Neither are the women's entitlements that are being activated in their place just ones that have ‘always been there’ waiting to be used. Such a view again assumes without evidence that the overall ‘system’ largely remains unchanged. Rather, while women and men in Kazipur often still evoke ideal obligations and responsibilities in forming and maintaining uthuli relations, they are also actively changing the practical nature of kinship-based extended entitlements and with them, the everyday realities of family, household and kinship.


Abdullah, A.A. and K.A.S. Murshid, 1986, ‘Inter‑district Changes and Variations in Landlessness in Bangladesh’, Bangladesh Development Studies, Vol. 14, No. 3: 97‑108.

Agarwal, Bina, 1986, ‘Women, Poverty and Agricultural Growth in India’, Journal of Peasant Studies, Vol. 13, No. 4: 165‑219.

Agarwal, Bina, 1990, ‘Social Security and the Family: Coping with Seasonality and Calamity in Rural India’, Journal of Peasant Studies, Vol. 17, No. 3: 341‑412.

Agarwal, Bina, 1994, A Field of One’s Own: Gender and Land Rights in South Asia, Cambridge: Cambridge University Press.

Ahmed, Kazi S., 1991, ‘Distances between Two Populations: Displaced and Non-displaced in Flood and Erosion Hazard Areas’, in K. Maudood Elahi, K. Saleh Ahmed, and M. Mafizuddin (eds.), Riverbank Erosion, Flood and Population Displacement in Bangladesh, Savar, Bangladesh: Riverbank Erosion Impact Study.

Ahmed, Rahnuma and Milu Shamsun Naher, 1987, Brides and the Demand System in Bangladesh: A Study, Dhaka: Centre for Social Studies.

Alam, S.M. Nurul, 1989, Socioeconomic and Political Dynamics of Char Land Settlement and Its Implications for Poverty, paper presented at the UNDP/UNFPA seminar on People and Environment in Bangladesh: Interdependence of Demographic Forces, Natural Resources and Deterioration in Environmental Conditions, Dhaka, Bangladesh.

Alamgir, Mohuddin and Sadiq Ahmed, 1988, ‘Poverty and Income Distribution in Bangladesh’, in T.N. Srinivasan, and Pranab Bardhan (eds.), 1988, Rural Poverty in South Asia, New York: Columbia University Press, pp. 11-38.

Ali, Shawkat, 1981, ‘Land Reform Measures and Their Implementation in Bangladesh’, in Muhiddin K. Alamgir (ed.), 1981 Land Reform in Bangladesh, Dhaka: Centre for Social Studies, pp. 141‑194.

Appadurai, Arjun, 1984, ‘How Moral is South Asia’s Economy: A Review Article’, Journal of Asian Studies,Vol. 43, No. 3: 481‑97.

Appadurai, Arjun, 1988, ‘Putting Hierarchy in Its Place’, Cultural Anthropology, Vol. 1, No. 3: 36-49.

Arens, Jenneke and Jos van Beurden, 1977, Jhagrapur: Poor Peasants and Women in a Village in Bangladesh, New Delhi: Orient Longman Limited.

Aziz, K.M. Ashraful, 1979, Kinship in Bangladesh (Monograph Series No. 1.), Dhaka: International Centre for Diarrhoeal Disease Research, Bangladesh.

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

Barman, Dalem, 1988, Emerging Leadership Patterns in Rural Bangladesh, Dhaka: Centre for Social Studies.

Beck, Tony, 1991, ‘Review of Drèze, Jean and Amartya Sen (1989) Hunger and Public Action’, Oxford: Clarendon Press, Disasters, Vol. 15, No. 4: 389‑392.

Boyce, James K., 1990, ‘Birth of a Megaproject: Political Economy of Flood Control in Bangladesh’, Environmental Management, Vol. 14, No. 4: 419-428.

BRAC, 1986, Who Gets What and Why: Resource Allocation in a Bangladesh Village (Rural Study Series Vol. 1.), Dhaka: BRAC.

Cain, Mead, 1978, ‘The household life cycle and economic mobility in rural Bangladesh’, Population and Development Review, Vol. 4: 421‑438.

Chowdhury, Mridul and Radheshyam Bairagi, 1990, ‘Son Preference and Fertility in Bangladesh’, Population and Development Review 16, No. 4:749-757.

Collier, Jane F., Michelle Z. Rosaldo, and Sylvia Yanagisako, 1982, ‘Is there a Family? New Anthropological Views’, in Barrie Thorne (ed.) with Marilyn Yalom, Rethinking the Family: Some Feminist Questions, New York: Longman.

Drèze, Jean and Amartya Sen, 1989, Hunger and Public Action, Oxford: Clarendon Press.

Dyson, T. and M. Moore, 1983, ‘On Kinship Structure, Female Autonomy and Demographic Behavior in India’, Population and Development Review, Vol. 9, No. 1: 35-60.

Elahi, K. Maudood, 1991, ‘Riverbank Erosion, Flood Hazard and Population Displacement in Bangladesh: An Overview’, in K. Maudood Elahi, K. Saleh Ahmed, and M. Mafizuddin (eds.), Riverbank Erosion, Flood and Population Displacement in Bangladesh, Savar, Bangladesh: Riverbank Erosion Impact Study.

Elahi, K. Maudood, K. Saleh Ahmed, and M. Mafizuddin (eds.), 1991, Riverbank Erosion, Flood and Population Displacement in Bangladesh, Savar, Bangladesh: Riverbank Erosion Impact Study.

Greenough, Paul R., 1982, Prosperity and Misery in Modern Bengal: The Famine of 1943‑44, Oxford: Oxford University Press.

Greenough, Paul R., 1983, ‘Indulgence and Abundance as Asian Peasant Values: A Bengali Case in Point’, Journal of Asian Studies, Vol. 42, No. 4: 831-850.

Halli, S. S. , 1991, ‘Economic Impact of Riverbank Erosion in Kazipur Upazila’, in K. Maudood Elahi, K. Saleh Ahmed, and M. Mafizuddin (eds.), Riverbank Erosion, Flood and Population Displacement in Bangladesh, Savar, Bangladesh: Riverbank Erosion Impact Study.

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, University of Manitoba.

Hartman, B and J. Boyce, 1988, A Quiet Violence: View from a Bangladesh Village, London: Zed Press.

Indra, Doreen Marie, 1994, ‘Living With the Jamuna: Representing River Eroded People in Bangladesh’, in S.M. Nurul Alam (ed.), Natural Disasters in Bangladesh: Views and Issues for the Social Sciences, Dhaka: University Press.

Islam, A.K.M. Aminul, 1974, A Bangladesh Village: Conflict and Cohesion: An Anthropological Study of Politics, Cambridge, Mass.: Schenkman.

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, Dhaka: Ford Foundation.

Jansen, Erik G., 1987, Rural Bangladesh: Competition for Scarce Resources, Dhaka: University Press Limited.

Jiggins, Janice, 1986, ‘Women and Seasonality: Coping with Crisis and Calamity’, IDS Bulletin, Vol. 17, No. 3: 9-18.

Kabeer, Naila, 1988, ‘Subordination and Struggle: Women in Bangladesh’, New Left Review, No. 168: 95-121.

Kabeer, Naila, 1989, Monitoring Poverty as if Gender Mattered: a Methodology for Rural Bangladesh (Discussion paper 255), Institute of Development Studies, University of Sussex.

Kabeer, Naila, 1990, ‘Poverty, Purdah and Women’s Survival Strategies in Rural Bangladesh’ in Henry Bernstein, Ben Crow, Maureen Macintosh and Charlotte Martin (eds.), 1990, The Food Question: Profits versus People, London: Earthscan, pp. 134‑148.

Kabeer, Naila, 1991, ‘Gender Dimensions of Rural Poverty: Analysis from Bangladesh’, Journal of Peasant Studies Vol. 18, No. 2: 241‑262.

Kandiyoti, Deniz, 1988, ‘Bargaining with Patriarchy’, Gender and Society, Vol. 2, No. 3.

Kolenda, Pauline, 1987, Regional Differences in Family Structure in India, Jaipur: Rawat Publishers.

Lamphere, Louise, 1974, ‘Strategies, Cooperation, and Conflict Among Women in Domestic Groups’, in Michele Z. Rosaldo and Louise Lamphere (eds.), 1974, Women, Culture and Society, Palo Alto: Stanford University Press.

Mafizuddin, Mirza, 1991, ‘The Characteristics of Riverbank Erosion: Kazipur Case Study’, in K. Maudood Elahi, K. Saleh Ahmed, and M. Mafizuddin (eds.), Riverbank Erosion, Flood and Population Displacement in Bangladesh, Savar, Bangladesh: Riverbank Erosion Impact Study.

Maloney, Clarence, 1988, Behavior and Poverty in Bangladesh, Dhaka: University Press Limited.

Mannan, M.A., 1989, ‘Family, Society, Economy and Fertility in Bangladesh’, Bangladesh Development Studies, Vol. 17, No. 3: 67‑99.

Martius‑von Harder, G., 1981, Women in Rural Bangladesh: An Empirical Study in Four Villages of Comilla District, Fort Lauderdale: Breitenbach Publishers.

Mukherjee, Ramkrishna, 1971, Six Villages of Bengal, Bombay: Popular Prakashan.

Papanek, Hanna, 1990, ‘To Each Less Than She Needs, From Each More Than She Can Do: Allocations, Entitlements, and Value’, in Irene Tinker (ed.), 1990, Persistent Inequalities: Women and World Development, New York/Oxford: Oxford University Press, pp. 162‑184.

Pryer, Jane, 1989, ‘When Breadwinners fall Ill: Preliminary Findings from a Case Study in Bangaldesh’, IDS Bulletin, Vol. 20, No. 2: 49-57.

Pryer, Jane, 1990, ‘Hunger and Women’s Survival in a Bangladesh Slum’, inHenry Bernstein, Ben Corw, Maureen Mackintosh and Charlotte Martin (eds.), The Food Question: Profits versus People? New York: Monthly Review Press, pp. 125-133.

Rob, Md. Abdur, 1990, ‘Flood Hazards in Bangladesh: Nature, Causes and Control’, Asian Profile, Vol. 18, No. 4: 365‑378.

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

Sayeed, Sabiha, 1991, ‘The Sources and Supply of Basic Food Stuffs in Dhaka City’, European Network of Bangladesh Studies, Vol 4: 14‑15.

Sen, Amartya, 1976, ‘Famines as Failures of Exchange Entitlements’, Economic and Political Weekly, Vol 11: 1273-1280.

Sen, Amartya, 1977, ‘Starvation and Exchange Entitlement: A General Approach and Its Application to the Great Bengal Famine’, Cambridge Journal of Economics, Vol. 1: 33-59.

Sen, Amartya, 1981, Poverty and Famines: An Essay on Entitlement and Deprivation, Oxford: Clarendon Press.

Sen, Amartya, 1983, ‘Economics and the Family’, Asian Development Review, Vol. 1, No. 2: 14‑21.

Sen, Amartya, 1988, ‘Family and Food: Sex Bias in Poverty’, in T.N. Srinivasan and P.K. Bardhan (eds.), 1988, Rural Poverty in South Asia, New York: Columbia University Press, pp. 453‑472.

Sen, Amartya, 1990, ‘Gender and Cooperative Conflicts’, in I. Tinker (ed.), Persistent Inequalities: Women and World Development, Oxford: Oxford University Press, pp. 123‑149.

Sharma, U., 1985, Women, Work, Class and the Urban Household: A Study of Shimla, North India. London: Tavistock Publications.

Westergaard, Kirsten, 1983, Pauperization and Rural Women in Bangladesh: A Case Study, Comilla, Bangladesh: Bangladesh Academy for Rural Development.

White, Sara C., 1988, In the Teeth of the Crocodile: Class and Gender in Rural Bangladesh, Ph.D. Dissertation, University of Bath, 473 pp.

Whitehead, A., 1981, ‘ ‘I’m Hungry Mum’: The Politics of Domestic Budgeting’, in K. Young, C. Wolkowitz and R. McCullogh (eds.), 1981, Of Marriage and the Market. London: CSE Books.

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. 23pp.

Yanagisako, Sylvia, 1979, ‘Family and Household: The Analysis of Domestic Groups’, Annual Review of Anthropology, Vol. 8: 161‑205.

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.

Zaman. M.Q, 1989, ‘The Social and Political Context of Adjustment to Riverbank Erosion Hazard and Population Resettlement in Bangladesh’, Human Organization, Vol. 48, No. 3: 196-205.

Zaman. M.Q, 1991, ‘Social Structure and Process in Char Land Settlement in the Brahmaputra‑Jamuna Floodplain’, Man, Vol. 26, No. 4: 549‑566.

1. Greenough’s [1982] parallel study of this famine asserts that, unencumbered by binding legal or social constraints, patrons abandoned their clients and men abandoned their “dependants” (wives and children) in attempts to survive themselves.

2. See Jiggins [1986] for a general discussion of poor women’s responses to seasonal variations in income and food availability.

3. Agarwal [1994] repeatedly calls for greater attention to be given to the role of informal social ideologies in securing, maintaining and legitimating peasant women’s land rights.

4. Landless agricultural labourers were virtually unknown in Bengal at the turn of the century [Greenough, 1983: 884] and only comprised 14 per cent of the population in 1951 [Alamgir and Ahmed, 1988: 21].

5. Ever-shifting depositional islands (chars) in river systems and the Bay of Bengal are a marked exception.

6. A clear distinction must be made between permanent displacement caused by riverbank erosion and that temporarily caused by annual monsoon flooding. In some other regions annual flooding is severe but causes no significant erosion or long term displacement.

7. Until the end of the Ershad regime, upazilas were the smallest national administrative regions in Bangladesh. The national government presence was marked by a complex of administrative offices, a local court and a police thana. Each upazila also had an elected chairman with powers distinct from those of locally‑based state bureaucrats. During 1991-2 some of these administrative powers were shifted to the District level, and upazilas were renamed thanas.

8. For a combination of political and legal reasons this official population estimate from the Bangladesh Bureau of Statistics [1988: Table 50] follows national census practice in including villages long ago destroyed by the Jamuna, as well as people who have ‘temporarily’ moved away to gain work. Using this method of accounting, Thana officials estimated the 1987 Union population to be 31,734.

9. Over 18,000 people were displaced from affected unions in the upazila during 1974-81 [Elahi, 1991: 106].

10. In many erosion‑prone unions loss of land is at least partially compensated for by the accretion of new (albeit typically infertile and flood‑prone) land elsewhere. The mainland part of Kazipur Union gained virtually no such land during 1980-93, though lower rates of erosion in 1994 and 1995 may signal a phase of increased land deposition.

11. Our analysis of Bangladesh Water Development Board maps shows that at some points the bankline shifted westward a quarter mile during 1989‑90, and over a full mile during 1979‑93.

12. Ahmed [1991: 144] reports that 67 per cent of Kazipur residents interviewed in 1985 had been displaced, 40 per cent more than three times.

13. We note that if this paper were able to address ‘eroded people’ in more detail the term would require much qualification and elaboration. For example, a few are far from poor: landowners with unthreatened land elsewhere, some in government service, teachers, local labour contractors, those with substantial businesses, a few politicians and the like. One major landlord and one‑time powerful politician claimed to have been displaced six times over seventy years. He was forced to move yet again during 1990 when flood reached his house in Meghai.

14. Doreen Indra did initial fieldwork in Kazipur for nine months during 1989-90, during which time she was joined by Norman Buchignani for three months. Two research assistants did follow-up work for extended periods each year since then. Doreen Indra returned to the field once more during Fall 1993.

15. Our research was initially inspired by that done by teams from the University of Manitoba (Winnipeg, Canada) and Jahangirnagar University (Savar, Bangladesh) as part of the IDRC‑funded Riverbank Erosion Study [Rogge and Elahi, 1989; Elahi, Ahmed and Mafuziddin, 1991]. As part of this study a large formal survey of some of the households on this section of the embankment and in nearby areas was carried out in 1985.

16. Following our initial fieldwork, Suzanne Hanchett (personal communication, 1993) has interviewed uthuli residents of Sarishjabari Upazila as part of the Flood Response Study (FAP-14) of the Flood Action Plan.

17. The first of these surveys covered a wide range of issues: household composition, kin relationships to landlords, economic and physical conditions of life and migration history. The second more extensively considered household composition and the density of uthuli kin networks. While it was our intention to interview one person in every uthuli household in the Union, a serious breach in the flood control embankment near the end of this phase of the study put nearly all farm land under water and isolated many homesteads and paras (village sections) from each other. Getting around by boat and by foot after that was difficult and communication poor. We are nevertheless confident that few households generally recognized as uthuli were missed.

18. Conclusions made about uthuli as a cultural concept derive primarily from a discourse analysis of interviews carried out during 1993. All interviews were tape recorded, literally translated, and incorporated into a hypertext database using the program Folio Views 3.1. The criteria listed for uthuli status identified in this paper interpenetrate each other, but space constraints do not allow all these cross-connections to be elaborated here.

19. When going house to house asking “who are uthuli around here” in 1990 the response would sometimes be that "we are all uthuli around here, so take our name". The anticipation was that through being interviewed they might receive government assistance, and the implication was that they were all needy.

20. As Pryer [1989] demonstrates, illness among “breadwinners” can have a devastating effect on Bangladeshi household economies.

21. A rarely voiced exception was when the men in question have already claimed their share of their father’s inheritance and thus charitably are ‘given’ a house plot by their brothers.

22. By law, rights to ancestral land are unequally inherited upon the death of the father, daughters receiving rights to one-half the portion of sons and surviving widows with children one-eighth. It is however uncommon for daughters in Kazipur (as elsewhere: Arens and van Beurden, 1977: 53; Mannan, 1989: 85; Aziz, 1979: 69; Martius von Harder, 1981: 79; Westergaard, 1983: 69; White, 1988: 236] to actually exercise independent land rights, as they have little power to enforce them. Sometimes even locally-resident sons do not have sufficient power to hold onto their father's land after his death [Arens and van Beurden, 1977]. Rural Bangladeshi women who do claim their share may not be welcome in their natal homes thereafter [Arens and van Beurden, 1977: 53; Aziz, 1979: 69; Westergaard, 1983: 71-2; Hartman and Boyce, 1988: 92-3].

23. As A. Islam [1974: 82] notes, “since [normally] the husband is supporting his wife, she does not have to claim her brother’s assistance. By supporting his wife, a husband is doing a service to his wife’s brothers who otherwise would be obliged to help their sister. This naturally obliges the wife’s brothers to be on good terms with their brother‑in‑law who could otherwise demand that his wife claim the portion of her father's property that she would have inherited legally".

24. If her father lacks sons and asks his daughter’s husband to reside with him, in Kazipur that son-in-law usually is considered ghor-jamai, not uthuli. Ghor‑jamai (ghor = roughly “nuclear family in a household”, lit. “room”; jamai = son‑in‑law) is a relationship of uxorilocal residence (Aziz, 1979: 57; Mukherjee, 1971: 253, 271; Jansen, 1987: 82) in which the wife may inherit land from her father, and her children remain part of her patriline.

25. This table summarizes all statements associating kinship and uthuli status found in informal interviews done in 1993. Those residing with all kin more distant than those indicated in this table or with non-kin are considered uthuli.

26. A khada or kata is a traditional Bengali unit of land measure that is locally considered to be one-sixteenth of a bigha, which itself is equivalent to one-third of an acre..

27. We talked extensively with men from fourteen adjacent households in a long‑established area near Meghai town while they were preparing to shift for the first time just before the 1990 rainy season. Almost all had been well off, well‑connected, land-owning farmers, and, yet they complained of their complete inability to secure rental land in the Union.

28. Most of those with whom we spoke were already physically and socially separated from the households of their brothers and fathers before being eroded. This usually was because they had already been environmentally forced migrants earlier or because of disputes occurring soon after their marriages. Many facing erosion cannot return to their natal homes, paras or villages in any case, as these places have been destroyed by the river.

29. For those with money, cheaper land prices elsewhere have been an incentive to move outside the Thana, particularly to nearby Sherpur (since the early 1960s) and other parts of Bogra District, and to Rangpur District seventy‑five miles to the north (since the early 1950s). There is now even a para named Kazipur in Sherpur. We note here a lacuna shared by anthropological research on environmentally forced migrants and political refugees: comparative lack of interest in those actually in the process of making migration decisions. Almost all such research is on those who have already migrated.

30. A samaj locally operates as “the largest intra- and inter-village social unit that brings together people who share common religious beliefs and practices in accordance with certain Islamic values interpreted by the samaj leadership” [Zaman 1991: 680] as well as a source of reciprocity and patronage.

31. By Bangladesh law and local custom individuals do not automatically lose ownership of eroded land; it is but ‘temporarily’ submerged. As long as they maintain the registration of the land and pay the necessary taxes they have at least the formal right to reclaim it if it re-emerges within twenty years [Ali, 1981: 180]. Actual practice is not so straightforward, since the submergence of land opens up opportunities for politicians, big landowners and strongmen to attempt to appropriate it through force, litigation or bribery [Zaman, 1991; 1988; Alam, 1989]. Any person who moves far away from their land (eroded or otherwise) stands a good chance of quickly losing control of it.

32. At least two‑thirds of uthuli husbands and 60 per cent of wives surveyed in 1990 were born in Kazipur Thana.

33. Our full census of local embankment households in 1990 shows that where government‑owned embankment land upon which to place one's house was once freely (though not legally) available, families connected by male agnatic kin links often did re‑establish contiguous residence; 239 of 553 households (43 per cent) on the embankment were living immediately adjacent to household men’s agnatic kin, while 18 (3 per cent) were living next to women’s kin.

34. For example, only a handful of those forced to give up living on the embankments because of erosion during the 1990 flood were able to make use of this settlement option. Out of 14 households about to be eroded from a long‑established area near Meghai town, no men claimed to be going to reside with their kin, even though in several instances kin‑linked adjacent households were planning to move together.

35. House plots and agricultural land in Kazipur are usually rented through a khai khalashi arrangement (see Jansen, 1987: 128‑130 for details and regional variations), in which the full rent is pre‑paid for a fixed period – now four years or less.

36. Local daily wage rates for male agricultural labour could go as high as Tk. 20 and food during harvests, dropping to as low as Tk. 10 and no food during off seasons. Tk. 12-15 and food was typical.

37. For example, in 1990 about four out of ten claimed to have owned at least a homestead immediately before, while after becoming uthuli the ratio of even nominally landed to landless fell to 1:100. Six per cent sharecropped some land after becoming uthuli; six per cent owned a cow (as opposed to 35 per cent previously) and eight per cent (as opposed to 19 per cent) a goat. Over half had houses, clothing and daily diets that were below average for the region; only one‑third had any children in school, and two‑thirds had at least one daughter or son die as a child. Despite their impoverished situation, only eight per cent interviewed in 1990 said that they had ever received any government aid, either when moving or thereafter; these were primarily women registered for free wheat distribution under the Vulnerable Group Feeding program.

38. One further indicator of acute need is that 80 per cent of uthuli became so during the rainy season, the time of greatest erosion. People generally stay put until erosion of their house plot is imminent. On a small time scale the speed of erosion is not very predictable, and while some families make elaborate preparations to salvage everything and shift residence over time, most are forced to move with no more than a few days’ warning. As a case in point, at the start of major erosion during the 1990 rainy season we saw about thirty families decide that erosion was imminent, disassemble their houses, move these and their other possessions (including small banana trees) to an unoccupied part of the embankments, reassemble their houses and start up a cycle of household activities again – all in less than 48 hours.

39. It is not possible here to do more than list some of the root causes of these remarkable differences between mainland and nearby char practice: Kazipur Union was at one time an socially and environmentally stable region and its village life and kin relations still partially reflect this, whereas life on the chars could never be so characterized; uthuli migrants in Kazipur are chiefly kin moving a mile or two east to west from previously safe villages related to the places to which they are moving, but no such pattern clearly shows up on the chars; there is virtually no newly accreted land in the Union and therefore (unlike Naturapara) little incentive for landowners to make use of uthuli clients to help secure it; the wholesale use of force sometimes employed by char landlords in their battles with each other for control over land is precluded on the mainland by the nearness of the Thana headquarters and police station; the use of force on the mainland is less necessary, as there are many effective non‑violent means through which landowners, politicians and touts can get the government bureaucracy and the courts to serve their interests.

40. This figure is probably overstated. It includes one anomalous instance where seven households (themselves complexly related) had established themselves on an unrelated person’s land over the course of twenty years. It also includes many instances of fictive kin ties.

41. The incidence of the residence of married couples on the land of wife’s father in other parts of Bangladesh cannot easily be extracted from the family and kinship literature, due to the tendency of researchers to concentrate on households rather than relations between them; household-based survey methodology would only pick up such couples if they shared the same cooking unit or house as others. Barman [1988: 39] reports that nine of 171 families in his natal village outside Dhaka were resident with the wife’s father or mother. Three of these instances appear very similar to uthuli living with a wife’s father in Kazipur.

42. The total does not equal 100 per cent due to rounding. This table excludes cases where there were no kin relations between patrons and uthul.

43. 80 per cent were either with patrons who were the wife’s agnatic relatives or the affines of the latter, or else were the husband’s female kin or someone linked by marriage to them.

44. 16 per cent of uthuli households surveyed in 1990 were nevertheless composed of women alone or else were women‑supported. Of these, over 90 per cent claimed to be actually widowed rather than divorced or abandoned.

45. Over the past twenty years, though, an exploitative, pervasive dowry system has established itself in Kazipur. It has been one of several things making asymmetrical what once were more balanced relations between a wife’s and a husband’s immediate relatives. Wife‑receiving families are now at a strong advantage over wife givers [Ahmed and Naher, 1987]. Sons‑in-law are the primary beneficiaries, while wives’ fathers shoulder the main burdens. By making the marriage of daughters an onerous financial responsibility, the dowry system seems also to have weakened ties between fathers and daughters. Today, men who actively refuse a large dowry that otherwise could have been secured, place a moral obligation on their fathers‑in‑law that may later be converted into uthuli residence. Even those men who take every advantage of the dowry system can try to claim uthuli residence by stressing the burden they have taken on in marrying a man’s daughter with a forcefulness impossible a generation ago. A subtext here is that there is a surplus of marriageable women, and that a man can always divorce his wife and find another.

46. We discovered no instances where a widowed or divorced man lived with his father-in-law as uthuli except when his wife had died after the couple had gone to live with her father.

47. As a pertinent case in point, data collected from male ‘household heads’ exclusively about intra‑household structure and process [Wiest, 1988: 3‑4] in Kazipur Thana earlier led Wiest to predict that “the incidence of conjugal pairs living with the wife’s natal family in rural Bangladesh would likely reflect severe hardship and few options, since it would deviate from the culturally expected alignment of the patriline [a hypothesis based on normative kinship ideologies]”, as well as to the finding [1988: 6] that among 247 households there were only two daughter’s husbands, nine brothers‑in‑law and 22 sisters‑in‑law to be found [a finding indicating a lack of attention to ties between adjacent households]. The conclusion: “given the cultural expectations of the patriarchal ideology and the strong patrilateral bias in the control of rural resources, we can easily understand the low incidence of son‑in‑law or daughter’s husband [residence]". Our data on uthuli‑patron kin ties established between separate but adjacent households gives neither this empirical conclusion nor the underlying generalities concerning gender and kinship much support.

48. This is most marked where uthuli feel a strong obligation to their patrons. This obligation is lowest among those living with the wife’s father or who have paid something to establish residence. In the latter case, fathers-in-law are reluctant to appear to be exploiting their sons-in-law (and hence, their own daughters).

49. 16 per cent of women and six per cent of men questioned in 1990 made use of small plots of land on which to grow vegetables for consumption and at times for sale. Some patrons do not allow uthuli to keep animals or grow things, animals because of the nuisance they cause, growing vegetables because patrons wish to maintain these rights themselves.

50. Even if a man consistently were to secure the highest local daily wage – Tk. 15 and meals – he rarely would be able to get more than twenty days of work a month. The cost of rice was Tk 11 (1992), Tk. 7-8 (1993) and Tk. 10 (1994) a kilo.

51. Jahan [1989] estimates that about 15 per cent of all rural and 25 per cent of all landless households in Bangladesh 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. Agarwal [1994: 30] estimates that 20 per cent of Bangladesh households are women-headed. Many of these women are thus “living like servants or must beg for their support” [Aziz, 1979: 74].

52. As Jansen [1987: 100] notes more generally, such loans are frequently secured from women’s relatives.

53. Individuals often pointed to their participation in meat distribution at Eid as symbolic of their social incorporation into the local scene.

54. As Maloney [1988: 54‑5] notes, the actual influence and power of an individual within their own circle of kin depends greatly on that person’s personal resources, and is not simply a function of the ideal status assigned to that person in the hierarchy of kin relations: “As in so much of Bangladeshi life, the system and the institutions are weak, but the relational aspect is strong. So the individual must strive throughout life within the complex of interpersonal relations and competing personalities...” by the exercise of “Bengali pragmatic individualism” [Maloney, 1988: 52].

55. To summarize: to select households rather than individuals as the unit of study renders both the identification of a sample and the delivery of the survey instrument easier. It may also reduce the sample size (and hence the cost), and allow one to represent results as pertaining to a larger overall population. Most urban-based South Asian researchers willing to work in villages are men, and patriarchal cultural conventions make it far easier for them to interview men than women. Male ‘heads’ are thus routinely the ones interviewed, and their responses subsequently identified with the interests of the whole household.

56. Agarwal [1994: 3-5] outlines the processes through which the assumption of a unitary household has been challenged in South Asian studies.

57. It certainly has reinforced a view that all Bangladeshi families, households and kinship relations can be put on a continuum ranging from a ‘traditional’ ideal (now seen as either historical, the preserve of the well‑off, or unrealizable goals of others) to the isolated nuclear family and the individual. Paradoxically, this may have given the ‘traditional’ ideal a new lease of life.

58. Even Agarwal’s [1994: 73-77] programmatic discussion of women’s bargaining with “the community” is very restrictively (and somewhat pessimistically) framed in terms of a general expectation that women would have greatly limited options, as compared with men.