The gendered reconfiguration of low-income urban household space in response to AIDS: Perspectives from Botswana

Abstract

The study focused on three main actors within low-income urban households in Gaborone, the capital of Botswana. These were primary care givers, persons living with AIDS (PLWA) and AIDS orphans. The survey investigated the extent to which the traditional family system was coping with provisioning for its PLWA and AIDS orphans, the level to which the main actors reorganize household space in response to evolving demands, to cater for PLWA and AIDS needs, and the degree to which such spatial reorganization resonates with the gendered utilization of household space. Although the methodology of this case study was mainly participatory, it also relied on documented literature from archival sources and the Internet for its conceptual, theoretical and statistical information. The results suggest that the traditional family system is failing to cope with care provision for the PLWA and AIDS-orphaned children and that adjustments to their increasing presence within the households influence the gendered reconfiguring of household space. The traditional gendered utilization of space is thus being constantly negotiated in response to the requirements of HIV patients and those of HIV/AIDS-orphaned children. The proliferating female HIV/AIDS patient care giving and orphan children guardianship appear to be threatening the central and pivotal position of the father figure within the patriarchal household setting.

Keywords:

• Botswana
• gender
• HIV/AIDS
• space
• urban poor

Introduction

In spite of the fact that HIV/AIDS affects almost every community and household in sub-Saharan Africa, only a few geographers have published research on HIV/AIDS in the continent (Aase 2005, 1; Mayer 2005). The geographic perspective on the AIDS pandemic integrates knowledge from diverse academic disciplines into spatio-temporal aspects that characterize the disease pattern and trends (Kalipeni et al. 2004; Aase 2005). Recently reported results have brought fresh insights into our space-time understanding of this epidemic in sub-Saharan Africa. Those investigations dealing with its spatial essentials have described and explained how and why HIV transmission rates vary significantly between countries (Abebe 2005; Mayer 2005), whereas those that focus on its place dimensions have explained variations in HIV vulnerability and intensity in terms of peculiar economic, socio-cultural and spiritual factors associated with specific localities (Iversen 2005; Teye 2005).

The gender perspective on AIDS has explored and explained vulnerability differentials between men and women under various settings (Agyei-Mensah 2005; Iversen 2005; Teye 2005). In an attempt to formulate meaningful intervention strategies for influencing changes in sexual behaviour, studies on knowledge, attitude and practice have also been conducted. Research on the discrepancy between knowledge and practice has underscored the role of socio-cultural, spiritual and economic factors (Iversen 2005; Kalipeni & Mbugua 2005; Teye 2005). Furthermore, an understanding of trends and possible impacts of HIV/AIDS is critical for the formulation of strategic policy interventions. Whilst the possible demographic impacts have been articulated (Agyei-Mensah 2005), a time-space model with a time lag between the increase in HIV infections and the full burden of orphanage has been established (Abebe 2005). Abebe (2005) has also explored ecological metaphors in orphan fostering.

The analytical scale of the geographical perspective of AIDS has nonetheless mostly remained at the macro- and meso-levels, with less attention being paid to the spatial dynamics and the attendant power relations that they generate between men and women at the micro-scale household level (Gwebu 2005). Qualitative geographical studies have, however, begun to be conducted on the subjective spatial activity responses of persons living with AIDS (PLWA), in relation to HIV/AIDS at the microlevel. Wilton's (1999) ethnographic study of nine men with symptomatic HIV and AIDS in Los Angeles shows how the respondents’ lives were fundamentally altered by the disease. The illness caused a series of both positive and negative effects on well-being which translated into expansion as well as decline of people's daily worlds.

Brown (2003) has explored a set of consequences of terminal hospice home care in Seattle. The findings of his study demonstrate that hospice care creates certain paradoxes: a normative paradox of home being a good and bad place to die; a territorial paradox of control itself changing the home; and a relational paradox between autonomy and dependency for the PLWA. While the Wilton and Brown studies highlight the spatial response manifestations of persons living with HIV/AIDS, they remain silent on the gendered aspects of such responses within the household setting. In Southern Africa, a series of household level studies on AIDS orphans have begun to shed light on the spatial geographies of AIDS at the household level (Davids et al. 2006; Mahati et al. 2006; Rusakaniko et al. 2006). However, the prevalence, impacts and evolving coping strategies, at both the community and household levels, need to be analysed and interpreted within specific gender and spatial contexts. The case for focusing a gender-sensitive lens on the ramifications of HIV/AIDS has been made (Amaratunga 2002).

The aim of the present research is hence to determine the resultant gender roles and spatial relations within those households having PLWA and the AIDS-orphaned children. The study focuses on low-income urban households in Gaborone, the capital of Botswana (Fig. 1). The article is divided into five main sections. The introductory section on the geography of HIV/AIDS is followed by a statement on care giving for PLWA and AIDS-orphaned children. The analytical framework of space and gender that constitutes the conceptual base of the article is then articulated. Next, the characteristics of the study area and research methodology are presented. Finally, the research findings are discussed in terms of gendered confinement to home space, gendered alienation from the home, renegotiation of gendered domestic space, and the emerging marginalization of women to impoverished environments.

Fig. 1.  Botswana: Location of Gabarone.

Care giving for people living with AIDS and orphan children

In those countries most heavily affected by HIV/AIDS, the public sector costs of providing care threaten to overwhelm their health systems. This is leading many organizations and governments to explore the traditional but more cost-effective ways of providing care to people with HIV/AIDS. Much of the burden of HIV/AIDS care in developing countries is consequently now falling onto communities and households. This hospice care is called home-based care (HBC), where families and friends, together with professional health visitors, work together to provide care for patients at home. Ogden et al. (2006) have argued that a strategy of simply downloading responsibility for care onto women, families and communities can no longer be a viable, appropriate or sustainable response. To date, public structures available to support home-based care include health institutions, the church, non-government or community-based programmes and networks of PLWA.

In Botswana, the hospice programme was initiated in 1996 and refers to the care given to PLWA in their own natural environment, which is their home, by their families, and supported by skilled social welfare officers and communities, to meet spiritual, material and psycho-social needs (Republic of Botswana 1996).This is consistent with the goals of Vision 2016 of creating a compassionate, just and caring nation (Republic of Botswana 2004).

Apart from providing care for the terminally ill, households have to cater for AIDS-orphaned children. There appears to be a fairly long incubation period between the spread of the virus among adults and the manifestation of AIDS orphans because the orphan epidemic curve seems to continue rising even when the HIV epidemic curve is declining (Abebe 2005). The number of orphan children, or of children up to 18 years of age who have lost one or both biological or legal parents in Botswana, is expected to rise from c.65,000 to 159,000–214,000 in 2010 (UNDP 2000). According to Abebe (2005), children experience a consistent change in the roles they play and the social positions they occupy within their own households and in their community, as a result of changing social and economic circumstances of the deceased and/or surviving parent. Care giving for the AIDS orphans and PLWA has generated a lively debate between those who believe that the traditional system of orphan care is overstretched and eroded beyond endurance by epidemic loss and those who claim that the system is resilient enough to absorb the shocks (Kaleeba et al. 1991; Chirwa 2002; Bray 2003; Abebe 2005, 44–45).

The ecological rupture metaphor

Kaleeba et al. (1991) note that AIDS has depleted the traditional social safety net system to breaking point, reducing the number of adults in their prime and piling fresh responsibilities on the elderly, who themselves will soon die. Nyambedha et al. (2003) have also come out strongly in support of the view that the absorptive capacity of the social safety net has reached maximum and its complete breakdown is imminent. According to Abebe (2005), parental death due to HIV/AIDS clearly marks an ecological transition within the family or microsystem, the smallest organization in human interaction. The ecological contexts of extended families or meso-systems, kinship ties and communities or exosystems, as well as larger national or macro system and international level structures and processes, are also important in the lives and experiences of AIDS orphans. These environments which affect the contexts for the (mal-)functioning of ecological environments to cope with AIDS epidemic also affect orphans’ livelihood possibilities.

Several social scientists have referred to the system of wealth flows between generations as the ‘intergenerational bargain’ (Carmichael & Charles 1999; Barnett & Whiteside 2002, 196). They regard this process as one of the core and fundamental bargains made and maintained between generations and the basis on which social order is constructed. Throughout Africa, the intergenerational bargain is becoming progressively harder to maintain as a result of the HIV/AIDS pandemic and the stress that it imposes on traditional social security safety nets (Ackroyd 1997; Kabeer 2000). The role of HIV/AIDS and its impacts on intergenerational bargaining are most noticeable in relation to, first, AIDS orphans, where transfers should normally proceed from parents to children and second, the economically active, where transmissions are usually expected to be from children to their parents (Sarandon 2001). The HIV prevalence rate in Botswana overall is 17.3%, and among the economically active 25.3% (CSO 2004). AIDS disrupts social roles, rights and obligations. As parents die, the children and the elderly are forced to take over the reigns of impoverished households which often lack the labour and capital to maintain the livelihood system.

Child-headed households are a new socio-demographic phenomenon in the era of AIDS. Many orphans are struggling to survive on their own, in their late parents’ house. Traditional patterns of care in the extended family are either no longer easily available or are being denied. Within the contested domain of urban space, especially in low-income neighbourhoods, widows and their children often become displaced and disinherited by opportunistic and greedy family members.

While some orphans struggle to survive in child-headed households, caring for siblings younger than themselves, others will find their way onto the streets, living by begging, stealing or, particularly in the case of girls, as commercial sex workers (Campbell & Ntsabane 1995). They are far more at risk of rape than children with homes to sleep in and they tend to start having sexual relations at a vulnerably young age. Orphans, especially girls who assume productive and reproductive roles, are more likely to attend school less frequently, have improper schooling or drop out of school altogether. A study in Botswana revealed that, in many instances, AIDS-orphaned children in sibling family units are frequently hungry, fail their school exams and show behavioural problems.1

For the orphaned child, there is often premature entrance to burdens of adulthood, without the rights and privileges or strengths associated with adulthood status. Such children lose the joy of their childhood and the skills that childhood develops in children. Their childhood is effectively sacrificed. It is therefore important to realize that they could become part of a socio-economically and socio-culturally disenfranchised future generation. High levels of voluntary or enforced exclusion lead to their ostracization and marginalization. If nobody cares for them, they may end up being nomadic, irresponsible, aggressive, dirty and unhealthy adults (Sarandon 2001, 11).

While caring for the vulnerable has always been an assumed part and parcel of the cultural tradition of the Batswana, the increasing number of PLWA, AIDS orphans and other vulnerable children has begun severely to stress the traditional safety nets, especially the traditional extended family coping mechanisms (Tshitswana 2003). Faced with a situation where they seem to be failing to fulfil their roles, families often find themselves in great despair and disarray as extended family safety nets have been stretched beyond their capacity and therefore no longer cope. As the numbers of PLWA and AIDS-orphaned children in communities increase, uncles and aunts – the first choice as substitute caregivers – become unavailable and grandparents become recruited into childcare (Foster 2000; Gwebu 2005). Older and weaker grandparents have always been the last resort as caregivers but today they take care of AIDS orphans because other relatives refuse to (Tshitswana 2003). An information officer working for Emang Basadi, a women's NGO, recently admitted, ‘Within our traditional society, there is a certain standard of care that is expected but people are no longer willing to do this’. Members of the extended family are thus becoming increasingly fatigued, incapable and/or unwilling to provide what had been taken for granted in the past (Gwebu 2005). Faced with these problems, how do the primary caregivers optimize the use of available household space to accommodate the increasing numbers of PLWA and AIDS orphans?

The ecological resilience metaphor

The ecological rupture perspective has been challenged by another ecological metaphor that suggests that the adaptive capacities and strengths of the informal traditional child-care system can still support a larger number of AIDS orphans and critically questions the notion of social breakdown (Chirwa 2002; Bray 2003; Kalipeni et al. 2004). Orphans not only negotiate their livelihoods and assume new social and economic responsibilities but also autonomously redefine their social and geographical position in their own society (Abebe 2005, 43).

Bray (2003) has also examined and questioned the predictions by academic and policy literature of social breakdown in Southern Africa, in the wake of anticipated high rates of orphanhood caused by the AIDS. She argues that such alarmist and apocalyptic predictions are unfounded and ill-considered (Bray 2003, 1).The argument is that AIDS orphans are unlikely, compared with other orphans, to be more predisposed to an upbringing in deprived environments that could precipitate social breakdown. Several points are worth noting about Bray's perspective on the situation.

First, Bray's article (2003, 40, 45) calls for empirical evidence in this debate and treats anecdotal and qualitative evidence as unreliable. This is in spite of the fact that such phenomenological/existential studies are normally participatory and in-depth and try to capture the lived experiences of the affected. Abebe (2005, 46) has acknowledged that geographers have not paid adequate attention to orphans’ own testimonies and their experiences of lived orphanhood while sociologists have ignored the importance and implications of spatiality in their livelihoods. Second, Bray's argument does not take account of the rapidly changing socio-cultural, socio-economic, socio-political and environmental dynamics that impinge on household resilience at the household level in this sub-region. With increasing globalization, urbanization and Westernization, the traditional extended family and its safety nets are giving way to the less elastic nuclear one. Moreover, support to the elderly caregivers, in the form of remittances from the migrants, is rapidly diminishing as the economically active become victims and succumb to AIDS themselves, thus exacerbating the dependency load incurred by the elderly.

In the case of Botswana, the introduction of cost-recovery measures, retrenchments to streamline the economy, depreciation of the local currency, and rising inflation all threaten the ability of the elderly and the working population to support themselves, quite apart from the AIDS orphans and PLWA. Furthermore, traditional modes of livelihood, such as arable and livestock farming, are increasingly becoming vulnerable to climatic change. Finally, Bray's conclusions appear to be contradicted by current and detailed empirical regional studies, conducted under the auspices of the Human Sciences Research Council with funding from the Kellogg Foundation, that confirm the threat of social rupture due to PLWA and AIDS orphans (Davids et al. 2006; Mahati et al. 2006; Rusakaniko et al. 2006).

Implications of the ecological metaphors for household spatial organization

Evidence from empirical studies in Southern Africa, including the present case study, demonstrate clearly that the traditional care giving system for AIDS orphans and PLWA is under strain. Consequently, households find themselves having to make fundamental adjustments to their livelihoods and living environments. Under such stressful situations traditional gender norms governing the use of domestic space are becoming sacrificed to exigent pragmatism.

The space and gender analytical framework

Haggett (2001, 796) defines space as a continuous unlimited isotropic or anisotropic area that may either be a reference grid or a forced field, as revealed by multidimensional mapping. Time-space geography lays particular emphasis on the role of time constraints on the shaping of human spatial activity. Massey (1994, 1) has characterized such spatial conceptualizations as synchronic systems of structuralists, who regard space as consisting of some absolute independent dimension. She criticizes the views that regard space as a singular, immobilized and unproblematic surface devoid of politics. According to her, space is constituted out of social relations and ‘the spatial’ is social relations ‘stretched out’.

Activity space is conventionally regarded as a geographical area within which individuals conduct their normal cultural and socio-economic activities. Besides its spatio-temporal dimension, activity space has an emotional value dimension added to it. This is the essence of topophilia: the affective bond between people and place (Tuan 1974; 1990[1974]). Activity space is therefore not simply a neutral and objective entity but also both an emotional and emotive one. For example, the tendency towards territoriality is a clear indication of the emotional attachment to activity space by primates, including humans. Foucault (1980, 149) remarked, ‘A whole history remains to be written of spaces which would at the same time be the history of powers … from the great strategies of geopolitics to the little tactics of the habitat’.

To what extent, then, has HIV/AIDS impacted on this problematic configuration of space at the household level? One useful mode of analysis to respond to this question is through gender analysis. Hirdman (1991, 191) was among the first scholars to introduce the useful concepts of gender contracts and gender system into feminist literature. Gender is the social construction of male and female relations. It is the cultural definition of behaviour accepted as appropriate to men and women in a given society, at any given time. The gender contracts are socially binding obligations regulating the relations between men and women at all levels of the society. Gender contracts at the personal, household, community, and national levels therefore create socially prescribed norms, roles and behaviours expected of men and women. Hirdman perceives the gender system as having two basic attributes: first, the clear-cut separation between what is regarded as female and male spheres, and second, the subordination of women and female-related issues.

Spatiality is socially produced and, like society itself, exists in both substantial/concrete forms and as a set of relations between individuals and groups, an ‘embodiment’ and mediation of social life itself (Soja 1989, 120). Space is therefore not gender neutral because it consists of a mosaic of domains that are perceived and utilized differently by men and women (Umiker-Sebeok 1985; Fortuijn et al. 2004). Space and gender are defined by sets of relationships between men and women negotiated within certain frames of reference or contracts. Societies generate their culturally-determined ‘ground rules’ for demarcating ‘social maps’ that structure given modes of perception and social interaction (Ardner 1981). Hall (1969, 83), in a study of proxemics, or the study of humankind's perception and use of space, claimed that ‘like facial expressions, gestures, and postures, space speaks’. Thus the prime directive of proxemic space is that we may not just come and go everywhere as we please. The seeming personification of space implies a reciprocal relationship between it and individuals who occupy it; they influence it and it, in turn, influences them. Thus the analytical focus of spatiality should be on the intricate linkages between space and gender and the construction of gender relations.

Spain (1992) has focused on the gendered characteristics of space. Taking a cross-sectional and historical perspective, she has examined how the organization of space determines gender relations and has come to the conclusion that gender-biased spatial segregation reinforces status differences between women and men, to men's disadvantage.

The gendered character of space is related to power relations between men and women. Massey (1994, 3) argues that since social relations are inevitably and everywhere imbued with power, meaning and symbolism, the view of the spatial is a dynamic social geometry of power and signification. Thus its analytical focus should be on the intricate linkages between space and gender and the construction of gender relations.

According to Massey (1994, 179), the limitation of women's mobility, in terms both of identity and space, has been a crucial means of subordination in some cultural contexts. Moreover, the limitation on mobility in space and the attempted consignment/confinement to particular places on the one hand, and the limitation on identity on the other, have been crucially related. In the West, one of the most evident aspects of the joint control of spatiality and identity has been related to the culturally specific distinction between public and private spheres. The attempt to confine women to the domestic sphere was both a specifically spatial control and, through that, a social control of identity (Massey 1994, 179).The fact of women having access to an independent income was itself a source of anxiety in that it might subvert the willingness of women to perform their domestic roles and could provide them entry into another public world, where life was not defined by family and husband.

Gender dynamics within the household setting can be further explored at the ideological level on the basis of the domestic ideal which assumes that the family:

• exists for pooling, controlling and equitably sharing common resources

• commands equitable control over resources, power and decision making among the adult members in matters affecting their livelihood

• pursues altruism in which the respective members subordinate their individual tastes and preferences, in the interest of common goals, for the benefit of the individual family members (Chan 1997).

In the context of the domestic ideal, gender division of labour within the household is taken as socio-culturally defined and is relatively immutable. The mother is expected to stay at home to fulfil productive and reproductive roles while the father, assumed to be the breadwinner, is expected to work and finance family needs. Since individual roles are clearly defined and taken for granted, conflicts and inequality are assumed seldom to exist, but where they do exist, they may be resolved quite easily if each member accepts and assumes their expected role(s). As a result of the repressive aspects of the domestic model, many women have remained confined to the ‘home’ where they are discriminated, marginalized, subordinated, or even excluded from work. Because of this form of repressive socialization, some women have come to define their identity in terms of providing services to their husband and children (Gittins 1985; Leornard & Hood-Williams 1988; Chan 1997).

The pre-colonial patriarchal structures and institutions in Botswana were characterized by male leadership and dominance (Dow & Kidd 1994). The socio-cultural milieu was based on a skewed socio-political system with dominant male leadership in the public sphere and subservient female roles confined to the private sphere (Dow & Kidd 1994, 1–2; Larsson 1999, 72; Kalabamu 2005, 251). The public geopolitical sphere of politics, government, defence, hunting, and livestock management was meant exclusively for men whereas a subordinated domestic sphere was reserved for women (Dow & Kidd 1994; Larsson 1999). Marriage was the dominant gender contract which defined women's positions and roles in society (Kalabamu 2005, 251).

Larsson & Schlyter (1993) have analysed space and time differences in gender roles and described the evolving power relationships between men and women under changing socio-economic circumstances through the use of gender contracts and gender-systems concepts.

The patriarchal system and gender contracts have, however, become increasingly challenged and transformed by colonial and post-colonial Western influences (Schapera 1953; Dow & Kidd 1994; Datta 1996). Over the past century, Botswana has experienced dramatic socio-cultural changes resulting from labour migration and rapid urban industrialization, and the increasing tendency for women to head households. Post-independence urbanization and rural-urban female migration in search of cash employment and independence of marital and familial subordination (Datta 1996; Larsson 1999) and democratic governance through a constitution that sought to uphold and guarantee equality for all, have been the main socio-cultural transformation agencies (Kalabamu 2005). As a result of women's widespread participation in economic life, a high proportion of women in Botswana are no longer just ‘home managers’ but ‘breadwinners’ as well (Kalabamu 2005, 253). According to Kalabamu (2005, 252), modernization influences have, however, tended to augment rather than replace indigenous socio-cultural structures, although women continue to participate in the public sphere and make substantial contributions to household wealth, regardless of their marital status (Larsson 1999, 77–78). The traditional patriarchal system has been weakened but not totally replaced (Kalabamu 2005, 253–254).

The gender system, including its spatial manifestations, is therefore dynamic, being continually constructed and reconstructed through negotiations, whenever deviations from societal norms and expectations become apparent (Massey 1994; Niranjana 2001, 34–35). Such negotiations may take place between couples, within households and within communities. In the present article, the crucial question to be resolved is whether construction, deconstruction and reconstruction of gender roles is reflected in the response at the household spatial level in Botswana to the both modernization and the HIV/AIDS epidemic.

The questions may be asked whether HIV/AIDS is a catalyst for change in power and role relationships within households in terms of gender and age, and how the resultant spatial dynamics are manifested.

Study area

Old Naledi is an upgraded Self-Help Housing Agency (SHHA) site-and-service residential neighbourhood, located in the south-western part of Gaborone (Fig. 2). With a current estimated population of 46,000 people, and an area of 114.1 ha, its population density is highest in the city at 403 persons per hectare (Department of Town and Regional Planning 1997). On average, there are 27 persons per plot.

Fig. 2.  Location of the study area Old Naledi in Gaborone.

A recent study conducted by HAAS-Consult (2002) shows that 36% of the housing structures are built of mud, plastic, cardboard boxes, and other non-durable materials; 61% of the households are served by one standpipe; and 93% of the households use pit latrines, while 2% of the households or 918 people have no toilets at all. There is one refuse skip for every 36 plots. The fact that the skips are usually overflowing with refuse points to their inadequacy. Only 18% of the population is in formal employment; 23% of the households are self-employed, whereas 39% are unemployed. The rest subsist on pensions, rent collection or farming (Gwebu 2003).

The study area was chosen for three main reasons. First, it has a substantial poor population. Poverty is a major cofactor in the occurrence of HIV/AIDS. The area therefore merits research and policy attention. Second, it is a transitional socio-cultural area for the incoming urban migrants, some of whom end up becoming victims of the AIDS scourge. Gendered spatial changes under the impact of HIV/AIDS are therefore most likely to be found there. Finally, the area is accessible and facilitated in-depth interviews at convenient times for the respondents, even late in the evenings.

Research design and methodology

A literature review was conducted on the conceptual and theoretical ideas that underpin the study. Further, documentary sources on the topic of HIV/AIDS home-based care and AIDS-orphaned children were consulted. A qualitative research design was adopted in order to understand the dynamic relationship of women and men in the social context of evolving gender roles and power relations within HIV/AIDS-affected households. It was participatory and based mainly on anecdotes, narratives of concerns, subjective lived experiences (genre de vie), and perceptions of the respondents (Lofland & Lofland 1984).

Selection of informants

Several participatory methods for collecting qualitative data were utilized. Five key informants were purposively selected. These were individuals based in the community, representing either religious denominations, Non-Governmental Organizations (NGOs) or the City Council. They provided a descriptive account of the socio-economic conditions prevailing in Old Naledi. The identification and selection of the types of households and the recruitment of the potential respondents were partly based on the information obtained from the key informants. This information was supplemented by the snowballing technique in order to ensure that it was up to date and accurate. Snowballing involved identifying households with PLWA and AIDS orphans from the leads supplied by key informants and already interviewed households. A total of 11 households were selected in this way. From each household, appropriate and willing discussants were identified in order to facilitate the administration of various types of in-depth interviews aimed at obtaining the profile and activities of each member. A total of 16 individuals were involved in the interviews; 10 of these were females. In order to satisfy research ethics, permission to use audio tapes was sought from each interviewee prior to conducting the interview. Also the reporting format of results ensured the preservation of anonymity of each respondent. Pseudonyms and aggregation at the report presentation phase assured confidentiality.

Fieldwork

A guided interview schedule, with open thematic items and open-ended questions, was used to guide the discussion. The first part of the questionnaire focused on the socio-demographic and socio-economic attributes of the household members. The second consisted of themes including who was ill with HIV/AIDS, the history of illness, the care giving events, impacts of care giving within the household, guardianship of any AIDS orphan children, threats to household livelihoods, and the role of other stakeholders in PLWA care giving and guardianship for AIDS orphan children.

The interviews were conducted with the help of a Setswana-English speaking research assistant. Audio tapes and notebooks were used to record both verbal and non-verbal responses to the questions. This exercise was carried out after normal working hours. My research assistant, who was a social worker at Old Naledi, and especially informative on nuances of local idiom or terminology and sub-culture, proved to be very helpful in providing additional information because she was in contact with the affected households on an almost daily basis. Close observations and notes were made on household assets, home-based activities and the various supportive roles played by government, NGOs and Community-Based Organizations (CBOs).

Data analysis

The raw field data were in the form of field notes, tape-recorded interviews and questionnaire responses. Tape-recorded interviews were transcribed, translated from Setswana to English and organized according to research objectives, ideas, categories, and themes, for reporting purposes. The preliminary findings were presented at three different seminars, in Gaborone, Maseru and Windhoek, whose audiences consisted of fellow researchers, invited researcher, and representatives of community-based organizations. The present article incorporates the suggestions from those deliberations.

Discussion

This section of the study will discuss gender and space in the context of the research findings. These are dealt with in terms of gendered confinement to home space, renegotiation of gendered domestic space and the gendered disruption of topophilia, exclusion and consignment of women to impoverished environments.

Care giving and the persistent confinement of women to the domestic sphere

Grandmothers, mothers, aunts, and sisters are at the forefront of home-based caring for PLWA and guardianship for the AIDS orphan children. Spatially, this gendered role occurs within the private sphere of the home. The comments from a guardian aunt about her ailing niece are reflective of such socio-cultural expectations of the care giving contract:

I strongly feel that my role as caregiver to my niece cannot be taken over by anybody else because, in our culture, I am the real surrogate mother to this child. I had to give up self-employment to look after this child. Before the child started taking medication she used to become very ill, especially at night: vomiting, having diarrhoea, and fever. Only my late sister could have had the patience and enough care to assist the child during such trying times. Now I make sure that the child is kept clean, eats recommended food, takes her prescribed medication at the proper times and goes for regular medical check-ups. My mother, who stays in the rural areas, had wanted to raise this child after the death of my sister, but she is now too old and too far from the necessary material and health care support institutions. My brothers and uncles cannot easily provide the required motherly emotional and immediate material and maternal needs of the child. Even when they do visit here occasionally, they only provide moral support. Most of the time they remain distant.

The aunt assumes that her maternal role is not substitutable. In her absence, her elderly mother would have willingly performed the care giving role, as dictated by a maternal contract. Brothers and uncles, whose activities occur mainly in the public sphere, expect the women to play a leading role in providing for the reproductive needs of the sick child, within the confines of the private sphere.

A home-based caregiver wife also underscored the significance of the care giving contract by highlighting that she was ‘duty bound’ to stop work and nurse her ailing husband:

Before my husband became very ill, I used to do piecework jobs as a washerwoman, to supplement family income, but now I have stopped working so as to give full time attention to my husband. Since he is now always bedridden, I provide all the day-to-day care and collect the required medication for him. I feel duty bound, as his wife, to provide all the requisite care. He used to be a successful carpenter but our savings have now been almost depleted by this disease.

Through questioning my research assistant as to the communal interpretation of duty bound, a number of issues became clear. Perhaps the most important was how the community perceives the cause of the disease. HIV/AIDS is commonly associated with witchcraft. This was confirmed through an interview with another HIV/AIDS patient:

I do not want to go for an HIV/AIDS test because the hospital doctor will wrongly diagnose me as being HIV positive. I know that this illness in my body has been caused by someone with ill-intentions against me, those people who hate me. See, even my daughter has been bewitched. She has completed Form Five, but even with her high level of education she cannot get any job. We are spending most of my husband's salary on financing the services of the traditional doctors who will soon sort out our problem.

If the duty-bound wife of the PLWA were to abdicate her culturally defined role as the primary caregiver, aspersions would be cast about her culpability. What is clear here is that the women's public production activity space shrinks significantly as wives and other female relatives assume reproductive roles for the sick and dying. Remarks made by the Social Welfare Officer in Old Naledi also confirm the existence of a care giving contract for PLWA and AIDS orphans: ‘adult females generally provide primary home-based care to ill members of the family because socio-culturally it is the assumed duty of females (mothers, sisters, aunts) to provide such care to the ill members of the family. Mothers automatically become the main care providers where there are no other family members around to assume the role of the mother’.

Such observations highlight the productive and reproductive roles which are ‘assumed duties of females’ by the patriarchal social system in Old Naledi. From the conversations, it is clear that not only is it expected of women by society to act in a particular maternal manner, women themselves also demand to fulfil their expected care giving role as dictated by the gender contract. Spatially, their reproductive role occurs mainly within the confines of the home.

Renegotiation of gendered spatiality within the domestic sphere

The notion that there exists a static and clear-cut separation between what is regarded as female and male spheres is, however, becoming challenged by the diverse roles that women are increasingly performing as the HIV/AIDS scourge worsens. Also, depending on specific conditions, gender roles have been shown to be flexible and reversible, with men, boys, women, and girls fulfilling both productive and reproductive roles. Gender contracts at the household level are therefore continually being negotiated. A sister who had moved in with her HIV/AIDS-positive elder sister made the following pertinent observation in this context:

Although I was then nursing my two month old infant, I managed to provide the essential day-to-day care for my sick sister, except for cooking because our culture does not allow a nursing mother to cook … My 18-year-old brother would, however, come in during the day and help with the cooking, cleaning, washing utensils, and running simple errands for us. For a teenage brother, he was very responsible and helpful. I guess that was because of the way our mother brought us up.

An important factor which accounts for such contractual negotiations is socialization, ‘the way our mother brought us up’. This implies the existence of a socialization contract by which boys and girls are being brought up, which is being renegotiated at the household level, permitting males to enter the domestic space that is normally reserved for women. Much as the respondent attributes the young man's actions to the way their mother brought them up, it is doubtful if this was not forced by circumstances. Probably, no female relative was available to cook, clean and wash the utensils. On the other hand, this might reflect an ongoing reconstruction of males’ identities. Datta's study (2004) revealed that male behaviour change could be greatly influenced by the gender of the head of household. Indeed, one of her male discussants confessed, ‘Where a mother is the head of a family there is not going to be much difference between the upbringing of sons and daughters. Married women come from the context of a man … In my case there is no distinctive difference between my sisters and myself. If I compare myself with boys with fathers, I find myself more disciplined’. According to Datta (2004), this discipline translated to doing more housework and sharing domestic chores, such as cooking with his sister. Additional evidence from her focus group discussions with men confirmed greater participation of men in domestic work as more and more women work in the public sphere.

Another example of the dynamism in the reproductive contract is provided by a young widower who was a mechanic. His plot has been partitioned into residential and workshop space. The latter space is dominated by him and his apprentice ‘spanner boys’. He nonetheless has to provide for the reproductive needs of his five year old son when the sisters were at school:

My wife died of AIDS about a year ago. My two daughters attend primary and community junior schools, respectively. My boy is only five and remains under my constant care when his sisters are away at school. All my children are not eligible for City Council assistance because they have not been registered with City Council. Also because I am working there is no need to get such assistance. … Before she got seriously ill, my wife used to oversee and perform all the basic household chores, with the assistance of my daughters. However, when she became very ill, she was taken away by her relatives to the rural area where she spent the last days of her life trying to recuperate. Since her illness, deterioration of her health and eventual death, my two daughters, especially the elder, have assumed more responsibility for some of their mother's reproductive roles. However, since they both attend school, they can neither prepare breakfast, afternoon meals nor do any laundry during the week. My busy schedule permits me to only improvise certain meals for myself and my little son, usually in the form of maheu (a beverage). All our laundry is now done by a hired laundress.

Although the father operates mainly in the public sphere in his workshop outside the house, the death of his wife now compels him to spend some time indoors improvising meals for himself and his male child.

Outside the female-headed households, the productive contract within the patriarchal setting at Old Naledi expects most husbands to be the major breadwinners, whose activities are predominantly conducted in the public sphere. The wives, on the other hand, should be confined to the private domain as a petty-trader/‘housewife’ around the house. The HIV/AIDS pandemic appears to be challenging this mindset, as previously shown in this article and as demonstrated by an excerpt from the following case recorded by a resident Community-Based Organization representative:

A 59-year-old male caregiver is looking after for his 46-year-old spouse who has all the symptoms of HIV/AIDS, such as chronic coughing, vomiting, diarrhoea, and wasting. He is employed as a watchman, and he draws a regular monthly salary of approximately P 400 (USD 80). Before she became very ill, the wife used to supplement the family income by taking on piecework jobs. They live with their 5 children – 3 males and 2 females whose ages range from 2 to 18 years. The husband cooks for and nurses the wife when the children are away at school. Normally, the eldest son and daughter share and rotate in the performance of household chores, but only the male child washes his father's clothing. This happens because the eldest girl is a stepchild of the household head.

Several points worth noting emerge from this case. First, it is the husband who is fulfilling the reproductive contract expected of the wife within the private sphere. His formal employment takes him to the public sphere at night but his daytime activities confine him to the private sphere in and around the house. Second, gender roles between the son and the stepdaughter are performed on a rotational basis. The occupational contract which differentiates between what men and women should do appears to have broken down. Third, perceptions, values, beliefs, and norms have overridden the stipulations of certain reproduction contracts. For instance, the cited stepfather-stepdaughter relationship prevents the stepdaughter from performing certain reproductive activities, expected of the female child, for the stepfather; instead, his biological son assumes these roles. The HIV/AIDS pandemic therefore appears to account for the construction, deconstruction and reconstruction of reproductive and productive contracts at household level along gender lines within this low-income urban setting. Whereas in the past there used to be a clear separation of where men and women could operate, now both boys and both girls venture into and cross over what was previously and socio-culturally regarded as women's space, and vice versa.

Under normal circumstances, only the married or cohabiting partners could share ‘their room’. Girls and boys would sleep in separate rooms, whilst grandparents could sleep with the small grandchildren of either sex. The kitchen was regarded as the women's domain.

Available space is negotiated over either as a result of the increase in the number of orphans or in response to the deterioration of the condition of the PLWA. The study has identified houses and rooms which provide internal environments for households. Among over 75% of the households, available indoor space consisted of a small structure, usually in the form of a single 20m² room shared by at least five people. Typically, the main room served multiple and often also incompatible uses such as a bedroom, storeroom, pantry, etc. Adaptive strategies have evolved in response to the increase in household size, caused by periodic visitors and adopted AIDS orphans further worsening congestion. One female respondent stated: ‘We only had a small rented room which we shared with seven young male and female dependents’. Under normal situations, the respondent would have preferred to have her room to herself. However, in order to cater for AIDS-orphaned relatives this was no longer possible. Adult females normally share the only available bed, whilst children (both sexes) usually sleep on the floor under such circumstances. The gendered use of private space is thus overlooked under the circumstances. Alternatively, additional space (mekhukhu) or shacks may be improvised to accommodate the burgeoning numbers of residents in restricted household space. The following excerpt illustrates this point:

The AIDS-orphaned brothers had to sleep outside under some old corrugated iron which leant dangerously against the wall of the house. Meanwhile, a boy of much the same age, their cousin, slept in a comfortable bed indoors. That year there were heavy rains; a helper built the orphaned boys a small room of corrugated iron roof and walls (mekhukhu), in the yard – not the best thing in the heat.

Although, culturally, young boys are expected to be stoical anyway because survival under such harsh conditions is regarded as a sign of ‘developing into a real man’, these were very extreme conditions indeed.

Conflict and stress normally worsen indoors as a result of changes in family numbers and members. Extreme intergenerational tension sometimes prompts the departure of some adolescent orphans, particularly the boys, who then live on the streets. The latest survey of street children in Gabarone found 95% of them to be boys ranging from 14 to 20 years in age (Campbell & Ntsabane 1995, 31). According to the researchers, ‘A plausible reason for this gender distribution is that boys are involved in more visible economic activities which take place on the street. Females, on the other hand, are less visible because they are subject to stricter, less public conditions at home’. Where an affected household has access to additional rooms, the use of private space appears to have been renegotiated. The caregiver spouse to a seriously ill husband disclosed:

Naturally, I used to share a bedroom with my husband before his physical condition deteriorated. I have, however, now relocated to share a room with our children to give him more room because it has now become very difficult for him to sleep throughout the night in certain normal positions on the bed.

Such renegotiations over the use of private space violate intergenerational respect towards the mother, who has to share premises with her teenage children. Also the expected gendered allocation of private space for teenage boys and girls is violated.

The presence of a PLWA within the household could have an impact on the renegotiation and subsequent use of indoor space. As the disease progresses, the PLWA may start using the main room as a convalescence sanctuary, bathroom and toilet. On certain occasions, the PLWA's change in behaviour may regulate access to their private space, strictly along gendered lines, as demonstrated in the following case:

Unfortunately, with the progression of the disease, my sister's condition deteriorated. Confined to her room, she began exhibiting abnormal behavioural changes. She became depressed, abusive and deeply suspicious of and repulsive to almost everybody. She developed an overwhelming hatred for men, including our own male relatives. The result was that she no longer entertained any male visitors. Even our younger brother, who used to assist us, ended up having to do the household chores stealthily out of her sight, outside the house.

Normally, when an adult individual is very ill, any close relative or friend may be permitted to enter their room, regardless of gender. However, in this case, the PLWA permitted only close female relatives, preferred female neighbours and selected female friends into her private space, due her paranoia, neurosis and xenophobia. She seems to have blamed all males, presumably due to their adventurous sex exploits, for her untoward fate.

Gendered disruption of topophilia and consignment of women to impoverished environments

Home-based care support personnel and the Social Welfare Officer from Gabarone City Countil disclosed that the traditional inheritance systems seem to work against their efforts of assisting widows to benefit from their late partners’ estate. This poses a serious threat to maintaining the geographical and social link with what was, all along, regarded as home by the widow and her children. After the death of a male PLWA partner, the prevailing property inheritance contract disadvantages the surviving female spouses who may be compelled to vacate the comfort and security offered by the familiar home environment. The wife of a gravely ill HIV/AIDS husband expressed her fears as follows: ‘There exists a real threat that after the death of my husband I could be sent away from this house by his relatives. It is difficult to imagine where I would go and how I would continue looking after all these children. I am now unemployed and my people live far away from here’. A feasible explanation for the underlying causes of this crisis was presented as follows:

Under living-together arrangements in Old Naledi, male spouses do not normally entitle their partners as next-of-kin, preferring rather their own male relations or mothers. Batswana are culturally averse to making wills, in the belief that this practice represents fatalistic thinking which could translate into drawing the signatory closer to the grave. This has deplorable consequences for the property rights and vulnerability of the orphans and widows.

Such problems reflect the existence of numerous laws which disempower women's access to productive resources (Kalabamu 1997, 7). The majority of women are married according to customary law which largely governs family relations. Through marriage laws and inheritance customs, males become the principal beneficiaries of family property. The legal systems and cultural norms reinforce gender inequality by giving men control over productive resources, such as land. These need to be renegotiated at household and state levels in order to ensure and consolidate widows’ property rights. Otherwise, mothers and their children are relegated to alien and sometimes hostile environments.

Conclusions

Whereas previous studies on HIV/AIDS and home-care provision have placed emphasis on macro-level, quantitative, supply-driven approaches, this article has adopted a microlevel, qualitative and demand-driven research design. It thus contributes to the emerging literature on a paradigm shift in medical geography (Kearns 1993; Cummins & Milligan 2000; Brown 2003). Methodologically, the article has demonstrated how spatiality and HIV/AIDS can serve as focal points in the analysis of gender dynamics between men and women within households, by using gender contracts and gender systems. Care giving is conducted within the private domestic sphere. The gender contract of home caring continues to be reinforced by men's and women's stereotypes in role ascription. Moreover, the fact that formal home care institutions and personnel are feminized simply strengthens these perceptions. However, there is ample evidence to show that the home caring contract is not immutable. There is renegotiation of gender roles and spatiality within the domestic sphere. Socialization is a strong catalyst towards promoting change. In low-income neighbourhoods, such as Old Naledi, male spouses cannot afford to hire maids to provide home care and the onus therefore falls directly on themselves.

The HIV/AIDS pandemic has prompted negotiations over private space. With the progression of the disease, the PLWA normally require additional private space. Also, pressure mounts as more and more orphaned children require fostering. Progressively, available space shrinks and the normally gendered private space becomes an unaffordable luxury. Boys, girls, aunts, and parents find themselves having to share the limited available space. Beyond a certain pressure threshold, male children may opt for street life (Campbell & Ntsabane 1995).

One decade ago, Townsend (1997) acknowledged the vital importance of paternity, within the cultural context of Botswana, in which men contributed to the children's well-being as both biological and social fathers. The central pivotal role of the father in the male-headed households is being challenged rapidly as women, comprising daughters, aunts, grandmother, sisters, and female-dominated formal support institutions, move from their socio-culturally marginalized peripheries of domestic decision making to the very core.

There is gendered disruption of topophilia which threatens to consign and confine women-impoverished environments. Widows face eviction from houses previously shared with spouses whereas the female children are threatened with relocation to remote and marginalized environments similar to those noted elsewhere in this region of the sub-continent (Ansell & van Blerk 2004).

The author acknowledges the financial support of SAREC for fieldwork, through the Institute of Southern African Studies, as well as the suggestions from two anonymous referees and cartographic assistance by G. Koorutwe.

Notes

1. Daniel, M. 2003. Children without Parents in Botswana: The Safety Net and Beyond. Paper presented at a scientific meeting on empirical evidence for the socio-economic impact of AIDS, 26–28 March 2003. Health Economics and HIV/AIDS Research Division (HEARD), University of Natal, Durban.