STRIVING FOR NORMALITY

People in Malawi, as in neighbouring countries, are often described as being in a state of ‘denial’ with regard to the HIV epidemic. The most usual sense in which people are charged with denial concerns the dominant tendency to refuse to name AIDS in cases of illness and death. There are several reasons to be dissatisfied with this assessment. First, naming AIDS is more selective in Malawi. People do use the term edzi (a Malawianised form of the English AIDS) when referring in general to the current surge of illnesses and deaths in the country. Whereas in 1986 in the research villages, there was absolutely no talk of AIDS, and in 1990 only the beginnings of a realisation about a ‘new disease’, by the late 1990s, and even more by 2006, adults knew a great deal about AIDS. But extremely few use the term edzi when referring to the illness and death of someone they know well, particularly relatives. More use the more common ‘roundabout’ terms for AIDSFootnote ⁴ in describing deaths of others but, again, far fewer use even these indirect terms to describe the illness and deaths of close relatives. At most funerals, the public statements about the deceased carefully avoid any reference to AIDS, even though some of those attending may whisper suspicions. This practice follows the well-established tradition of not speaking ill of the recently deceased. The common association of AIDS with inappropriate sexual behaviour results in this avoidance (see below).

A more fundamental objection to describing the situation as one of denial is that it is superficial. Fassin (Reference Fassin2007: 120), for example, says that ‘denial, in its deepest sense, signifies the intolerable’. In such stressful times, the term ‘denial’ glosses over the ways in which people are torn between acknowledging the way in which the disease has felled them, and refusing to recognise the ‘unbearable’ knowledge (ibid.). Recently, Alex de Waal (Reference de Waal2006: 18) used the concept of ‘normalisation’ to refer to the ways in which people ‘adjust … reality to take account of the miseries of AIDS’. Much of what the Zomba villagers do can be seen as striving for normality. Rather than sunk in denial, they are better seen as trying to control the abnormal circumstances of the rising toll of HIV-related illness and death, making huge efforts to channel them into the normal and normative ways of their society. A very similar conclusion has been drawn by Chimwaza and Watkins (Reference Chimwaza and Watkins2004: 805) about caregivers to people sick with AIDS in Malawi, when they say that the ‘reluctance’ shown by many in speaking openly of AIDS ‘is not a denial but something more existential … [in that the caregivers] turn away from acknowledging, even to themselves, that their child, their mother, or their aunt will not survive despite their devoted care’.

Such efforts are seen in the ways in which the matrilineal family and its constituent households seek to deal with the increases in illness, death and orphans, by moving the sick and orphans around the constituent households. Enormous efforts are made to recreate a ‘normal’ family. One example among many is Mai C. Her divorced elder sister, Anna, died in 2004. Mai C took in four of the children, the two youngest going to another ‘sister’ (mother's sister's daughter) nearby. In addition, Mai C was looking after two other teenage boys, orphaned by the death of another sister. By 2006, two of Anna's sons had married, though one continued to live in Mai C's compound with his wife. During 2006, the very sick younger brother of deceased Anna and Mai C came to be looked after by Mai C,Footnote ⁵ a decision taken by herself and her two other older brothers, one of whom was well-off and provided for most of the medical costs. The sick brother had all the symptoms of AIDS though Mai C mentioned only meningitis. In no conversation did Mai C suggest that her expanded and changing household was unusual.

Another major dimension of the process of ‘normalisation’ is to assess cause (why this person died). People draw on both pre-existing sets of explanatory principles and on newer suggestions emanating from many sources, including churches, government, medical experts, and media discussions. There has been a veritable blitz of information about HIV/AIDS and most people associate the epidemic with too many sexual encounters,Footnote ⁶ but there are also competing theories about the epidemic and about any particular person's illness and death.

People in the research site do not distinguish linguistically between HIV and AIDS; the former is hardly referred to at all, while the many indirect terms used for edzi all refer to illnesses. The partial exception is that some people, including some local healers, refer to tizirombo, ‘little animals’ or organisms in the blood which they see as causing the illness.Footnote ⁷ Many knew that a person could be infected with the HIV virus but appear perfectly well,Footnote ⁸ but most attention is paid to overt illnesses attributable to AIDS.

The sexual transmission of HIV is well known in the area and is the overwhelming focus of public discourse. While the level of voluntary testing is still extremely low, there have been increases in the availability of testing in the rural areas; some people are definitely paying more attention to the known behaviour of their spouses, friends and neighbours, and making changes in their own practices, such as some youngsters ‘holding their hearts’ (that is, not engaging in sex too early or too often), some adults holding each other to stricter standards than before (a woman sending off her husband because she learned he was taking lovers on his trips to town, a chief who had himself and his new wife tested before marriage, a man who, suspecting his wife of taking lovers on her trips to market, got himself tested).

These echo the detailed information emerging from the University of Pennsylvania study sites from the late 1990s to the present (Watkins Reference Watkins2004; Smith & Watkins Reference Smith and Watkins2005). Again, this is not to say that there is radical transformation in the sexual and related practices of Malawian villagers in Zomba. To expect such would be to misrepresent the capacity of people living in severely stressed situations to ‘choose’ to completely change well-established practices. For example, as we have found in the sample area in earlier years, a woman with a decent amount of land can attract a husband more easily than others, and everyone knows that two adults can produce more crops and more income from such land than can one; so it is not surprising that widows, especially those under about fifty, tend to remarry quickly.Footnote ⁹ Many people continue to have multiple partners, some sequentially (as in remarrying after death of a spouse or after divorce), some simultaneously. Multiple partnering is not confined to the rural areas, and may even be more common in towns and among those with higher incomes.Footnote ¹⁰

For some, the sexual transmission of HIV produces blame, such as the mother who told her dying daughter that she had brought the illness on herself by her careless behaviour, or the mother who blamed the girlfriend of her deceased son for passing an illness to him. Some repeat the common stereotypes currently deployed in discussion about AIDS such as the claim that ‘poverty’ leads women to sell sex, or a tendency to blame women for affairs outside marriage even though such love relationships (chibwenzi) are common throughout Malawi and reportedly more indulged in by men than women. As Fassin (Reference Fassin2007: 32) points out, the HIV epidemic uncovers ‘pre-existing stigmas’ and prejudices, thus revealing as much or more about the society as about the disease itself.

Other interpretations given to the rise in illness and deaths resemble those identified by other researchers in Malawi and in neighbouring countries. These are: AIDS is a more virulent form of a well-known illness (kanyera) caused by infringing certain sexual rules;Footnote ¹¹ the illnesses some define as AIDS are sent by malevolent people through witchcraft; AIDS is a disease that has been deliberately inflicted on Malawians through infected condoms, family planning pills and injections. The agents assumed to be behind the purposive infliction of AIDS are God (who sent it as a punishment for immoral or sinful behaviour), as well as Whites (Europeans or Americans) and the government of Malawi (who distribute infected condoms and other contraceptives in order to reduce the population).Footnote ¹² The agents assumed to be behind the bewitching of people are seen to have a wide range of motives, depending on the relationship between them and their victims, though they generally turn on envy and jealousy.

People often hold several different ideas at once or at different times or for different circumstances. Well before the era of HIV, the interpretation of a person's illness often varied according to circumstances. Depending on the trajectory of the illness, the diagnosis of what is causing the illness may change over time, particularly if the illness fails to respond to particular treatments and the patient fails to improve. AIDS has intensified this pattern of multiple interpretations that draw on a repertoire of available and changing ideas. HIV infection has produced an unprecedented explosion of serious (chronic) illness in people who are normally the most healthy (those from the late teens to around fifty), a situation which, along with the incurable nature of this ‘new disease’, is deeply worrying. It is no wonder, then, that people are involved in an agonising search for explanations on which they can base their own actions.

THE PAST IN THE PRESENT: 1986–2006

The past twenty years in Malawi have seen palpable change: a shift from the highly controlled, autocratic regime of Dr Banda to a ‘multi-party’ regime in 1994 and subsequent political volatility; structural adjustment, increased economic instability due to both national and international conditions, and deepening dependence on foreign aid; and periodic droughts and floods suggesting an increase in climatic volatility with severe effects for agriculture. Within this overall context, the study households have been marked by increasing heterogeneity and inequality over the period (see Peters Reference Peters2006). The average landholding of study households has changed very little over the period,Footnote ¹³ but the deepening inequality can be attributed to declining overall returns to agricultural production for the majority of the smallholder cultivators after a boom in burley tobacco earnings in the early to mid 1990s, and to similar difficulties in earning money off-farm, whether in the formal wage sector or, for most, in the localised, ‘informal’ activities of casual labour, and small-scale business.

The critical changes for the study households between 1986 and 1997 were the effects of structural adjustment and market liberalisation of the late 1980s into the early 1990s, including the opening of burley tobacco production to smallholders (previously limited to estates), and the opportunity to sell it on the Auction Floors at world prices. Average sample household income rose by just under 60% between 1986 and 1997, due largely to the sharp rise in burley tobacco production. However, the aggregate sample figures masked an increasing gap between the poorest and the richest: in 1986 the top quartile households had incomes four times bigger than the bottom quartile, a ratio that increased to eleven times by 1997, remaining much the same, at 10:1, in 2006.

All households have followed diversified livelihood strategies throughout the period, but the income shifts mirror changes in these strategies. In 1986, the sample as a whole earned approximately 30% of household income from the value of retained own-produced maize (the staple food), 30% from sales of agricultural produce, 24% from non-farm sources (wages, small business, etc), and 15% from transfers (from absent working husbands and adult children). The mix of activities has not changed over the entire period but the ratio has altered. Already in 1986, the households in the poorest quartile earned more from off-farm sources and less from sale of agricultural produce than the better-off. This pattern intensified by 1997. By then, the most important source of cash income for the richest quartile households came from sales of burley tobacco, with important contributions from sales of maize, legumes and vegetables. Most of these better-off households also continued to produce all the staple maize they needed as well as surplus maize, part of which they sold, and part of which they kept to hire casual labourers in the peak cultivation season.

In contrast, the poorest households were finding it difficult to obtain as much staple maize as they needed, and so reduced their already small sales of agricultural produce while increasing the proportion of earnings (in kind as well as cash) from casual labour for their better-off neighbours and small estates in the area. Their falling further and further behind the gains of the better-off households, and the increasing proportion of their meagre expenditures spent on obtaining basic foodstuffs, revealed the increasing pressure they were under by the end of the 1990s. Some were increasingly caught in the vicious cycle whereby they were able to allocate less labour and other inputs to their crops because of lack of cash and their need to work for others in the deficit food season (which is also the peak cultivation season),Footnote ¹⁴ thereby reducing their own production.

While a major factor in increased socio-economic differentiation for the sample has been burley tobacco, it is not the only one (and in many areas of Malawi it is inconsequential). Other contributing factors are the reduction in sources of institutional credit for fertiliser and seeds, the increasing importance of access to sufficient land in face of the volatility of crop prices in less regulated markets, the great reduction in options for migrant labour, and a flood of people into petty trading with low returns (cf. Dorward & Kydd Reference Dorward and Kydd2004: 346; Ellis et al. Reference Ellis, Kutengule and Nyasulu2003).

The summarised data on income shifts present a flattened picture of people's lives, whereas the study household members were active players throughout this period of considerable political economic change. Studying and living among the study families, we were struck by how hard most people worked, whether in their fields or elsewhere, and how quickly most took up new options presented to them. As new crops and new markets appeared to open, many responded by growing and selling the crops. New varieties of maize and pigeon peas, new markets for chili peppers, sunflower, sweet potatoes and cassava all led farmers to adapt and modify their planting patterns. Not all these ‘opportunities’ proved sustainable – the boom in chili peppers of the early 1990s collapsed as supply outran the demand, sunflower prices oscillated wildly between good and bad, and some of the producer organisations set up to take advantage of bulk selling fell apart. Some of the best farmers in the sample proved to be true experimenters, testing out different seeds and different techniques to see what worked best. Many kept to tried and true systems (of plant intercropping and plant sequencing, types of seeds, and so on), and, inevitably, some were just poor farmers and equally inept in efforts to run small businesses. Overall, many of the sample households deployed considerable effort, discipline and judgment in negotiating these difficult years. Most remained poor and continued to struggle to make a living, some became even poorer, and a minority benefited.

Throughout these years, too, we learned about the cultural and moral principles according to which people judged their own and others' lives. First, there is a deep-lying and usually unarticulated value attached to the matrilineal system of kinship (a child belongs to his mother's lineage), inheritance (the main heirs and successors of a man are his sisters' children, and daughters not sons inherit matrilineal land), and the matrilocal pattern of residence (whereby on marriage a husband moves to the village/compound of his wife).Footnote ¹⁵ As described in detail elsewhere, people in this area have maintained these patterns ‘against all odds’, sloughing off the prejudices against matriliny by missionaries, colonial officers, and others (Peters Reference Peters1997, Reference Peters, Barrett, Place and Aboud2002). Documentation of study households in the agriculturally vibrant research area over the twenty-year period provides no support for the still present stereotype that men living and working on their wives' land do not have the ‘incentive’ to invest their efforts in that land. The matrilineal mode of land transfer also provides considerable flexibility to people who juggle diverse livelihoods. For example, a sister who obtains a job in town will cede her field to another relative but will have it returned or replaced when she returns; or a son returning to his natal home between marriages will be loaned a field by his mother or sister. The way in which people define kinship, inheritance and the proper mode of post-marital residence is a critical foundation to understanding their lives and their efforts to construct livelihoods. This shapes the patterns of inheritance, the disposition of responsibility for land, children and property, and the course of defining and treating misfortune, including illness.

This brief sketch of the socio-cultural and economic circumstances of the Zomba sample households up to the late 1990s now allows us to ask how these circumstances influenced the ways in which rural people in Zomba district sought to live under the onslaught of HIV/AIDS, as perceived in 2006.

HIV-RELATED DEATHS AND PATTERNS OF STRESS

The study findings echo those that document considerable heterogeneity across households in their ability to deal with the HIV epidemic, and a very low rate of household dissolution (Mather et al. Reference Mather2004). This is partly due to pre-existing levels of resources, with the better-off more able to support the increased costs of care and of death, and partly to the fact that households are not solitary units but part of extended matrilineal families that adapt and reshape themselves in relation to loss and crisis. About a quarter of adult deaths reportedly attributable to HIV/AIDS were of the principal couple – the key woman and husband. The immediate impact of such deaths is often acute, with loss of cultivation time and hence of harvests, loss of other sources of income, and rising costs, financial and other, in caring for seriously sick people and in organising funerals. However, in the large majority of cases, the households did not dissolve and were maintained by the surviving wives who often remarried within a few years, or were taken over by adult daughters, sister's daughters and a few sons. Orphans were shared out among relatives and land was taken over by heirs (daughters, sisters or sisters' daughters). We learned of no cases of ‘land-grabbing’. Rather, the land of deceased persons followed the customary routes to appropriate heirs. Widows of men known or suspected of dying of AIDS were treated no differently from other widows, the same being true of widowers. There was certainly ambivalence and tension – what commentators refer to as ‘stigma’ – around deaths attributable to AIDS, as mentioned above, but the study found no evidence that ‘HIV widows’ faced special discrimination. Doubtless, the strongly matrilineal and matrilocal practices of the area provide a key reason for this lack of special suspicion towards widows, who are normally surrounded by matrilineal relatives. In the minority of cases where wives were living in their husbands' villages, after the death of a husband the widow returned to her own natal village with her children. Such households thus ‘disappeared’ from our sample, but are likely to have followed paths similar to those of the sample matrilocal households.

The most obvious change in the sample was the shift in age distribution between 1986 and 2006 (see Figures 1 and 2). The samples are small for demographic data, but the decline in numbers of babies and infants and of adults between twenty and forty years of age can only be explained by the deaths from AIDS of young adults and of babies, presumably through infection by their mothers.

Figure 1

Age distribution in sample households, 1986.

Source: authors' fieldwork.

Figure 2

Age distribution in sample households, 2006.

Source: authors' fieldwork.

It has been more difficult to know with certainty whether or not a household has experienced a death attributable to HIV. Testing was not an option for the research. Thus, we had to rely on reports by sample households on deaths.Footnote ¹⁶ Since most people in the sample area do not care to talk of the dead, it is difficult for outsiders to pose questions about the number and causes of deaths. It has become even more difficult with the onset of HIV and the rarity with which family members attribute a death to AIDS. The information we have on deaths and their causes has been collected over the entire twenty-year period of study, and not only in 2006. The main source of information has been the ‘updates’ section on the periodicFootnote ¹⁷ household membership surveys which recorded members' deaths, sometimes with a listed cause or, more usually, a description of symptoms, with supplementary information obtained from neighbours and others. Despite our efforts, for many deaths that occurred prior to 2006 there is too little or no information on which to base an assessment of the cause of death.

With these caveats in mind, we found that of the 174 households who had been in the sample in 1986 and in 2006, 18% had adult deaths due ‘certainly to almost certainly’ to HIV, 21% had deaths ‘very likely’ due to HIV, and a further 10% were ‘likely’ due to HIV. Evidence for the first category of deaths attributed to HIV includes, first, actual statements by family or other closely connected families using the various terms for HIV; and second, sufficiently detailed descriptions of the typical sequence of symptoms and illness attributable to the disease. Evidence for the second category ‘very likely due to HIV’ is based on descriptions of symptoms and length of illness, etc., but at a lesser level of detail. The category ‘likely due to HIV’ had an even lesser level of detail. The proportion of households who had a death about which there was no or very little information was 30%, and 40% of households had one or more deaths not attributable to HIV.Footnote ¹⁸ Because some households had deaths placed in more than one category, the percentages overlap. But overall, 43% of the 174 core households had at least one death in one or more of the three HIV categories.Footnote ¹⁹ And 28% of households had no adult deaths at all. Before presenting ethnographic detail on these families, we will present some quantitative results.

THE CRITICAL ROLE OF THE EXTENDED MATRILINEAL FAMILY

A central conclusion of the Zomba research is that the matrilineal family is doing a remarkable job of providing support to bereaved households. Mobilising the wider family stems some of the consequences of HIV illness and death. While a few households dissolved in the wake of HIV-related deaths, the children, assets and land were absorbed into other, related, households. No young children (those under fifteen) were living alone. As described above, some households suffered acute economic stress after HIV-related deaths but these were in the minority, and most affected households had managed, with the help of matrilineal relatives, to take on extra burdens and responsibilities and to weather the difficulties.

The under-researched importance of the extended family has been pointed out by others – by Mtika (Reference Mtika2003) for Malawi, and for Zimbabwe by Grant and Palmiere (Reference Grant and Palmiere2003: 233), who conclude: ‘The extended family as a safety net is still by far the most effective community response to the AIDS crisis’. The relative silence about the extended family probably derives from the dominant methods employed in research. In virtually all surveys and in most other investigations, the ‘household’ is the sole unit of analysis; it is also the unit used in all the quantitative data deployed by government, donors, NGOs and research centres in the country. The household or banja is indeed a key unit in the social organisation of Malawian villages (and towns), but it is not the only one. We have been able, over the course of many years, to document how the sample households interact with other households, especially with those related to them. Households are embedded in broader networks, connected by relations of kinship, marriage, neighbourhood, village and friendship, as well as through churches and mosques. The matrilineal–matrilocal pattern dominating in the research area means that the key woman in a household is living near sisters, mother's sisters' daughters and mother's mother's sisters' daughters (cousins of various degrees according to the European system but referred to in Zomba as sisters). Male siblings and cousins usually leave for their wives' village on marriage though some marry within the village, and most have regular contact with their sisters, visiting the natal compounds. They also return home to live between marriages or, as is seen particularly now with the rise of AIDS, during serious sickness. Throughout life, matrilineal kin provide the surest support and the closest ties to a person though, being the main rivals for land, wealth and status, they may also be suspected of witchcraft.

Children move across households for play and for food, as well as to give help (a child may be sent to accompany his/her grandmother to the market or to draw water, for example). There is both a positive value given to having a viable, independent household (visually marked by a separate granary, kitchen, basic tools and utensils), and a great deal of cooperation between these independent units. The cooperation takes place in routine tasks such as women relatives and friends taking turns to brew and sell beer (so as not to compete for the same clientele) or to process crops, in exchange of usually small quantities of food, utensils and tools, and particularly in times of crisis or need, such as births, illness and death. Individual households are also involved in collective activities such as the yearly graveyard cleaning, funerals, memorial feasts, projects of road building or boreholes, and cases in the chief's court. In short, the ability of a household to be ‘independent’ relies not only on the key members (key woman, husband, adult children), but also on the supportive relations with other households. While many of these are nearby, some are the households of adult children or siblings who live and work in towns. People, goods and information move between these connected households and may be seen as the ‘lifeblood’ of social organisation in Malawian villages.

The effects of AIDS cannot be understood without recognising that households are part of larger networks on which they depend and which they, in turn, sustain.Footnote ²⁸ While one cannot underestimate the material losses, let alone the grief, of those directly affected by AIDS, the wider family can be seen as the key provider of care for the sick, the dying and the living survivors, and a bulwark against massive deprivation in this research site. The vast majority of care for the sick, bereaved and orphaned is given by the wider family; existing households within a matrilineal family reorganise themselves in response to the sickness and death of an adult member and to the needs of orphaned children; and land is not left unused because its ‘owner’ dies but it is given to another member of the matrilineage, first to the heirs of the deceased, but if they are too young then to a close relative to use until the heirs are of an age to need the land.

But this diffused responsibility exerts often considerable cost on individuals and families. Grandmothers find themselves with fewer adult children to rely on, with some adult children returning home sick and dying, and with the task of becoming mothers to their orphaned grandchildren. Sisters and daughters have to take on extra responsibility in providing care to sick relatives (sisters, brothers, parents and others), and/or to take in orphaned children of the deceased. In households where the mother, father or both fall sick, school-going daughters and sons miss days of school, some even being forced to drop out entirely, because they have to provide care to their sick parents and/or their parents are unable to raise the fees and other costs for school. Teenage orphans are more likely to drop out of school, even when there is no financial obstacle, and are more likely to marry young.Footnote ²⁹

The way in which this population has responded to and is living with the HIV epidemic has to be understood in the context of a densely populated area, where land is short and in huge demand. Therefore, our finding that the extended matrilineal family is able in most cases, so far, to absorb most of the very high costs, material and otherwise, of the epidemic should not be assumed to be the case for all of Malawi, still less for Africa. The central role of women in matrilineal–matrilocal areas (especially in family and land matters) may be most significant, but there is a need for more longitudinal studies of other family systems in Malawi and elsewhere. It is essential, too, to place these findings within their real timeframe. The phrase ‘so far’ just used refers to the fact that many extended matrilineal families are already very stretched, and, without improvements in the services to help them increase their income level, their capacity to care for increasing numbers of sick and orphans, and to gain more equitable access to medical care, some may find it difficult to maintain their roles as primary caretakers.

First, there is the obvious fact that chronic illness and deaths continue to rise (even if the HIV infection rate, currently 12% nationally and 18% in Zomba district, were to level off), which means that individuals, households and families are facing a rising burden in terms of costs in resources and time for caring for the sick, preparing for and attending funerals, taking in orphans, and so forth. Second, many of the caretakers are people, especially women, in their late fifties and older. As these caretakers age, the burdens will also increase for the next age cohort, which has a much higher rate of HIV infection. It is possible, therefore, that the current remarkable ability of matrilineal families to manage much of the impact will be further stretched and undermined.

Moreover, there are cross-household or intra-societal effects at play. The better-off sample households (those with higher levels of resources and income) have, on average, been more able to absorb the effects of AIDS illness and death. But if an increasing proportion of these households lose people, cash and time to AIDS, then not only are the households themselves and their close relatives affected: their losses have broader implications because these better-off households play important roles in local relations of production, distribution and consumption. It is precisely the better-off households who are the producers of ‘surplus’ maize which they store and sell in much bigger quantities than the majority of sample households; they are also the larger-scale producers of burley tobacco and other cash crops. Declines in their ability to produce large surpluses will reduce local supplies of maize, and the likelihood of their hiring people for food in the deficit season (which, while not equitable, nevertheless serves as a resource for some of the poorer). Additionally, their reduced overall income level will necessarily cut their expenditures, many of which have been for local products (furniture, beer, mats, baskets, pots) and services (building, well-digging, tailoring, carpentry, bike repair). In short, the socio-economic links between the better-off households and others mean that major losses among the better-off minority do not augur well for the poorer.

The obvious conclusion is that the continuing ability of the extended family to take on responsibility for caring for increasing numbers of sick and dying people and for orphaned children is not guaranteed, and requires more effective socio-economic development to improve their access to food, income and public services.