Why HIV/AIDS prevention strategies fail in Botswana: considering discourses of sexuality

Abstract

A phenomenological study that was carried out among five ethnic groups of Botswana revealed the importance of taking into account culturally situated sexual realities when prevention policies for HIV/AIDS are considered and implemented. Furthermore the study threw light on the ineffectiveness of the current national HIV/AIDS prevention strategy of ‘Abstain, Be faithful, or use a Condom’ (ABC), a strategy that has been externally imposed on communities, without sufficiently engaging the behavioural practices and values of the communities themselves. This paper therefore advocates educational strategies for HIV/AIDS prevention that take into consideration localised social relations and value systems. Devising policies that engage with the discourses that are dominant in each ethnic group can make a difference in a country that has been hard-hit by the HIV/AIDS epidemic.

1. Introduction

In spite of the country's substantial investment in HIV/AIDS prevention strategies, Botswana has one of the most highly infected populations in the world (Brigaldino, 2002). We argue that one way of addressing this situation lies in understanding more deeply the cultural practices of sexuality that are embedded in different ethnic communities. There is a substantial literature which uses epidemiological and biomedical analyses to explain the dynamics of HIV transmission in different national contexts (for example, Girardi, 2003; Low-Beer & Stoneburner, 2004). Epidemiology has been criticised, however, for explaining insufficiently why transmission persists in certain contexts, primarily because of ‘sexual culture’ in different communities. Dowsett (1999: 223), for instance, argues that: ‘We need three sources of research knowledge to help us produce effective prevention: epidemiology, social-behavioural monitoring; and cultural and educational research’. It is this last aspect that this paper addresses. We argue that one way of understanding cultures of sexuality for HIV transmission in Botswana is to use a Foucauldian discourse analysis as an explanatory framework. This approach has been used for explaining unequal gender power sex relations in South Africa (Shefer & Foster, 2001) and in exploring attitudes to people with HIV/AIDS in Mali (Castle, 2004). The understanding of why and how people behave in the way that they do is seen as important for enabling the behaviour change that is necessary to reduce infection:

Ascertaining the sexual culture of any such community is a central research task for effective health education, because all communities are differently structured. (Dowsett, 1999: 228)

Supporting this argument, Low-Beer & Stoneburner argue that communications from within communities, in addition to formal HIV interventions, are key to ensuring behavioural change—in ways that are ‘engaging rather than disrupting its values and structures’ (2004: 10). This entails first identifying the meanings embedded in different cultural contexts (Dowsett, 1999).

This paper provides an in-depth analysis of two communities in Botswana, taken from a larger study of socio-cultural practices, particularly addressing the way in which women interpret and reproduce their unequal gender power relations. The study focuses deliberately on women, rather than both sexes, as women have been targeted in Botswana as the potential resource for addressing behavioural change. The use of women's voices to explain gendered social practices demonstrates the degree to which women's hegemonic understanding is embedded in social interactions and the extent to which space for alternative discourses has to be found.

We start with an explanation of our theoretical position, followed by a brief description of the country and research design. The second half of the paper then discusses the findings in relation to the theory, followed by recommendations for action.

2. Discourse and power

The concept of discourse represents the beliefs, attitudes, values, language and behaviours in societies. Policies, rules and regulations of any society follow the dominant discourses of those societies. Such discourses are held in place by a complex network of power relations, which are always vulnerable to resistance and change. The tensions that govern these power relations are explained primarily by using Foucault's rendering of power, discourse and knowledge. These concepts are outlined here because of their relevance to the way policies for behavioural change, in relation to HIV/AIDS, can influence, or not influence, sexual practices in their recipient countries. They also demonstrate how ‘heterosexual relations are a central site for the reproduction of unequal gender power relations and male dominance’ (Shefer & Foster, 2001: 375). The study shows how the repression of women's individual agency hinders women from challenging the risky practices of their male partners and puts them in a hegemonic position of ‘desire’ to maintain the status quo of their sexual behaviour.

Most people are so embedded in their societal belief systems that they neither question their society's dominant values nor realise how much they themselves are naturalised into them. Certain behaviours therefore become entirely predictable and unquestioned in their own social environment. Their behaviours are ‘normalised’ (Foucault, 1980; Fairclough, 1989). Centralised policies, values and beliefs seem normal to the instigators, but not necessarily so to the people they are trying to influence.

People's positions within power relationships are multiple, of course. They play different roles according to the social composition of the participants in any interaction, but the way societies operate often militates against the success of individual efforts to resist the dominant views in close proximity to them. Foucault sought to explain how people internalise and normalise their unchallenging behaviour or attitudes as a form of self-regulation. He called this ‘disciplinary power’. This is a means of self-control where people survey their own behaviours as if they were being watched from an imagined, all-seeing gaze. By policing themselves in this way people are taking away their own will to resist by internalising as ‘common sense’ certain rules and norms. They become players in their own ideology: ‘Discourses define what is normal and what is normal is then seen as in need of normalisation or conformity to the norm’ (Ramazanoglu, 1993: 22). The self-surveillance is held in place through institutional structures, conditions and hierarchies where individuals oversee each other. People believe they must conform because they believe in the expectations of their conformity, so that the system is: ‘taking away their wish to commit wrong … mak[ing] people unable and unwilling to do so’ (Ramazanoglu, 1993: 154). The family, for example, is an institution through which sexuality discourses are enacted for economic or other socially relevant purposes. The body and sexuality are tools for an expression of power relationships.

The ability to change discourses relies on our understanding of how they are set up in the first place and how they function on a daily basis. The potentially unpredictable combinations of power relations and discourse interactions (as the mechanisms for power relations) allow for the possibility of resistant forms of discourse and the possibility of changing power relationships. Agents of power and their discourses are therefore constantly under threat – hence the intricate networks of agencies, institutions and discursive practices such as rules and internalised rationalities, for sustaining the status quo (Foucault, 1980). Marginalised cultural groups that are close-knit and have little exposure to other, external values, attitudes and beliefs, are likely to have closely guarded disciplinary power mechanisms in place to maintain their society's status quo. Therefore, in order to change existing discourses for sexual behaviour in particular communities it is necessary to find a way of changing those discourses from within those communities. As Dowsett argues:

The meanings embedded in those aspects of community life and mobilized in interventions most effectively drive behaviour change by making it a shared commitment to survival and security. (1999: 229)

This position has also been supported by research in Uganda and other countries, where it is argued that successful behavioural change is always preceded by community responses where:

They were able to translate HIV prevention into their own country contexts: they each set rather than followed international best practice. (Low-Beer & Stoneburner, 2004: 10)

For countries such as Botswana, cultural or ethnic groups and tribes are still relatively homogeneous and behavioural ties are strong and bound by their discourses. External interventions and their assumed position of neutrality in these contexts can therefore be simply perceived as a foreign discourse to sustain the power of certain viewpoints – to serve the interest of the foreigners. This external knowledge is sustained through ‘experts’ who perform, say and do, based on their own status and internalised understanding of the world. However, what may appear true (knowledge) for one society may not be so for other social groups.

Foucault (1980) and other writers (Hill Collins, 1990; Lynch & O'Riordan, 1998) argue that it is possible to unearth forms of knowledge other than the dominant worldview. This is the kind of knowledge which rarely receives political status. It is often subjugated and disregarded by those with authority to know. Subjugated knowledge is, as Foucault says, delegitimated knowledge, local and without authority. The subjugated knowledge and discourses are what this paper is unearthing. It describes these discourses from the rationales and contexts of different cultural groups in Botswana in relation to sex and HIV/AIDS. In so doing we reveal why external policies (discourses) for behavioural change are often resisted or simply not heard by certain groups.

3. The Botswana study

Botswana is a landlocked country in Southern Africa, estimated at 540 000 square kilometres (Central Statistics Office, 2001). Although most citizens are members of the Setswana-speaking tribes, there are other ethnic groups. The last census (2001) showed that Botswana has a population of 1.7 million, of whom 52 per cent are female. Botswana is a patriarchal society in which there is sexual division of labour (Griffiths, 1997; Dow 2003). Marriage is valued, and those who marry receive special status and privileges.

Of Botswana's population aged 15 to 49 years, 38.5 per cent is HIV positive. According to Esilaba et al. (2003), the epidemic in Botswana and other sub-Saharan countries is mainly sexually driven through two modes: heterosexual and mother-to-child transmissions. The prevention strategy in Botswana has been a centrally managed, and largely externally funded, campaign. It has focused on the distribution of condoms, establishment of testing centres, anti-retroviral therapy and capacity building (Ministry of Health, 1997). These strategies are all externally imposed without deference to specific cultural behaviours or attitudes. Research conducted on HIV/AIDS has focused on bio-medical factors (types of virus and sexually transmitted infections: Girardi, 2003); impact assessments of HIV/AIDS [on education, health, economy: National AIDS Coordinating Agency (NACA), 2004] and on behavioural studies (for example, attitudes, behaviours and practices: Castle, 2004). National prevention strategies do not use targeted and culturally relevant materials for behavioural change. Furthermore, traditional medicine, spirituality and other influential cultural factors have not been addressed. Neither have other influential people in local society been given the opportunity to play a role in encouraging positive cultural practices (NACA, 2004). Against this background, between July and September 2003 a qualitative research study was conducted in the rural areas of Botswana to address one of the gaps, namely, cultural dimensions of sexuality and HIV/AIDS.

3.1 Purpose of the study

The purpose of this study was to understand the cultural dimensions of sexuality within five ethnic groups in Botswana and their perception of the relationship between HIV/AIDS infection and sexual behaviour in their unique patriarchal cultural context. In-depth conversational discussions focused on four research objectives:

1.	A description of sexuality within each ethnic group (to include sex education, sex rituals and their rationale, role of the different genders, role of the family or society, etc.)
2.	An analysis of the social function of sex in the culture of ethnic groups
3.	A description of cultural health education practices and their historical experiences with incurable diseases
4.	An examination of the perceived relationship between sexuality and HIV/AIDS infection and prevention

3.2 Methodology

Although some of the social functions of sex (e.g. procreation) cut across all ethnic groups, the two ethnic groups are identified here because of their unique culturally based sexual practice and discursive rationales for retention of cultural identity based on a specific mode of economic mode of production. For example, the Bakalanga ethnic group has the practice of nkazana or the need for a man to have a ‘small house’ – another sex partner from the wife's family to help provide farm labour, and the Bangwato ethnic group's sexual practice of mantsala or playful sex with a blood or ethnic cousin and setlogolo ntsha dithogo or young girls' sex play with ‘mother's brothers’ was necessary to regulate and control sexual behaviour because of seasonal migration between different family settlements: the lands, the cattle post and village.

For each ethnic group, ten knowledgeable and articulate females aged between 15 and 90 were selected using a purposeful sampling technique. Five interviews per site were analysed in depth as all ten interviews revealed very similar data. The age spread was chosen in case sexual behaviour was experienced differently depending on age. In-depth conversational interviews provided an opportunity for participants to discuss openly their ethnic sexual experiences and HIV/AIDS issues and constraints with regard to power, authority/control and other cultural expectations.

Small, qualitative samples are not meant to be generalisable across broader populations. Their value is in their relevance to the community to which they belonged. Data saturation suggested that the findings reflected common community values. The potential of bias in interpretation was addressed through seminar presentations and discussions across the academic community in Botswana and by specifically discussing findings with other member of the ethnic groups in question.

4. Findings

From across all the ethnic groups their sexual behaviour and experience revealed the following findings with unwavering consistency:

1.	For all ethnic groups, sex has a social function, including procreation, pleasure, family property, exchange, personal interaction, healing/cleansing, religion/spirituality interrelationships and control/oppression.
2.	Sex is culturally regulated, and accepted types of sexual behaviour are learnt through socialisation.
3.	Each ethnic group has access to national HIV/AIDS education processes but they felt the message has ignored the cultural sex and health education with which most people identify.

The next section highlights for two groups the situated knowledges and realities, with their own regimes of truth and hegemonic beliefs, that sustain gender power relations and identify particular practices within each culture as normative. This is then followed by an analysis of the social functions and their implications for a new approach to HIV/AIDS prevention.

4.1 Cultural dimensions of sexuality

4.1.1 Kalanga: Nkazana

Nkazana is a heterosexual behaviour practised by the Kalanga ethnic group in the northern part of Botswana. Nkazana (literally translated as ‘small house’) is a cultural practice where a new husband is given authority to ask for sexual favours from a younger female sibling of the wife. This young girl is identified and introduced to the husband and the community by the family of the wife during the wedding ceremony. A 15-year-old respondent said:

Nkazana exists in our culture. When a girl gets married, her youngest sister will be nkazana. She will be sitting at the table eating with her sister's husband during the wedding ceremony.

When the wife is taken to the in-law's home this young girl also accompanies them so that the husband's family can also get to know her. An elderly women rationalised the practice in this way:

The practice of nkazana is important in our culture because marriage is for procreation and there is no divorce. So if this woman cannot conceive for whatever health reason nkazana's children are officially hers. When she gets older or does not satisfy the husband sexually nkazana is already there.

It is important to mention that all elderly women from this culture strongly believed that the health of their son-in-law was the responsibility of the wife's family. So giving this man nkazana ensures that he does not sleep with other women outside the family and end up with sexually transmitted diseases. This ideology is therefore rationalised as simply common sense within the situated knowledge of the culture.

Data from interviews with both the younger and older women from this ethnic group revealed that sex and sexuality in their context are social constructs. Both the family and societal norms are in place to socialise, regulate and control the gendered sexual behaviour of members (disciplinary power). For instance, grandmothers start engaging in sexual conversations with girls from the age of 15 while grandfathers do the same with boys. Discussions usually centre on topics such as potential sex partners, sex and health issues, and preparation for sex in terms of what to expect from the male partner, how to handle the opposite sex partner and when to say no to sex. This is how one teenager explained her situated reality:

When the girl gets her first menstruation period she has to tell her grandmother or her mother, elder sister, cousin or aunt. Then she will be confined to the hut or room to be taken care of by her grandmother for the whole week or till the period stops. This is the time when you are told that sex with boys will result with pregnancy or sexually transmitted diseases such as gonohorrea and syphilis.

In addition to sex education, participants also mentioned that the girl's first menstruation in this culture is also used to introduce her to traditional health prevention and curative herbal remedies for sexually related health problems or ailments.

My grandmother smeared cow dung on my waist and back when I had my first menses so that I would not have period pains. She also told me that in our culture women do not engage in sex when they are menstruating because the blood is too hot as it is being shed to purify the female reproductive system.

So we see that subjugated knowledges have their own regimes of truth and institutional practices (discourses) that reinforce and sustain those truths.

As a further example, marriage itself marks another passage in the sexual life of a woman in this culture. As part of the institutional practice of marriage socialisation, the couple go through further sex training. For instance, societal norms for pregnancy prevention are in place to ensure spacing of children. This is because, culturally, breast-feeding is continued for two years and during this time the lactating mother is not supposed to engage in sex with the husband. If for some reason the woman is not able to breast-feed for two years then they both drink traditional medicine to prevent pregnancy. This is how one woman described her experience:

A thread is put around the woman's waist with a piece of traditional herb as a reminder to the man that it is not time to have another baby. It is also believed that this herb magically discourages the woman's sexual activity because it will be an embarrassment to the family if the couple can have a baby prematurely.

Asked who the recommended sex partners were in this culture all participants mentioned that girls were socialised to accept sex with sons of their mother's brother and their sister's husband is officially assigned the responsibility of nkazana. However, we can see that, as with all discourses, there is the potential to resist, or appropriate, such behaviours to suit new rationalities. It is here that there is potential for exploring how alternative discourses might engage with existing ones, while promoting a new understanding of risky sex practices in relation to HIV/AIDS. Appropriations, nevertheless, are still influenced by the power relations of dominant discourses for that culture. For instance, in addition to these cultural sex partners, the younger participants and middle-aged women confessed that because of economic hardships and interaction with people from other cultures some women and men now engage in sex with other people than the traditionally expected norm. One of the elderly grandmothers said:

This is where our young people get AIDS, they really have to stop sleeping with outsiders. The cemetery is full.

But beyond a certain age if a girl is still not married, parents, especially grandparents, expect their unmarried daughters to have children. This puts pressure on girls and young women to engage in unprotected sex. A grandmother would say:

At least give me a grandchild. I have no work now. I want to see my great grandchild before I die.

It appeared that resistant discourses were rare in these contexts. Even if new regimes of truth are introduced, power relations are likely to be reproduced through the extended family's institutional discourses that protect male interests. Data on the relationship between these ethnic sexual practices and HIV/AIDS revealed that women are aware that they are at risk of HIV infection because most men do not agree to use condoms. The following statements are testimony:

As a wife or nkazana how do you even begin to dream of mentioning a condom unless you want to be beaten.

But disciplinary power also encourages women to rationalise their existing power relationships:

It is hard because men play their part, i.e. feed the family, take care of his in-laws, so all you have do to is satisfy him sexually not a condom.

And even when women have accepted the new, externally introduced discourses about HIV/AIDS, they only argue within the dominant discourses of their culture and menfolk:

Knowing about HIV/AIDS is not a problem. We all know and are actually tired of the radio messages about HIV prevention but in this culture women have no power when it comes to sexual decision matters. Not only that, most men in this culture do not believe in using condoms. They say they develop rash around the male organs when they use condoms. Government should try condoms with boys not their fathers, who knows maybe the young ones are different.

Other findings related to HIV/AIDS were the cultural beliefs and internalised regimes of truth surrounding the epidemic. Because AIDS symptoms are family ailments people believe that traditional curative and cleansing herbs such as nshashanyama, mzeze and mphalola should be able to cure AIDS. If they do not cure it means God is punishing them for something with this epidemic and it will pass just like other epidemics in the past such as measles, leprosy, etc. Therefore, although participants in this study realise that women are most vulnerable to HIV infection because of the nkazana sexual practice and male power in sexual decision-making, they nevertheless respect their culture and expect the solution for HIV infection to come from within the culture.

The Bangwato ethnic group and its matsala and dithogo sexual practices is discussed in the next section.

4.1.2 Ngwato: Mantsala

Mantsala means playful sex with a blood or ethnic cousin. Although this cultural sexual behaviour is practiced by the major Tswana-speaking tribes or ethnic groups in Botswana, the Ngwato ethnic group was selected to represent this cultural practice because in their case, sex with sons of mother's brothers or cousins extends to tribal cousins, for example the Bakgatla. Furthermore, this ethnic group is also associated with sexual favours from nieces to their uncles or setlogolo ntsha dithogo. Data analysis revealed that cultural sex education for this ethnic group includes preparing girls to expect requests for sexual favours called mantsala from three different categories of men. These are blood male cousins, i.e. sons of mother's brothers, tribal cousins from the Bakgatla ethnic group and uncles (i.e. their mother's male siblings) in the form of dithogo. This statement was echoed by all participants from this ethnic group:

My uncles' boys are my God-given husbands. So there is no need for them to ask, they simply say: cousin it's time, let's go.

Yet another said:

I always joke with wives of the sons of my mother's brothers by reminding them that I am the real wife, but they are mothers of my children.

This statement is not surprising, because both the socialisation and sex education encourages marriage between these families or ethnic relations. Proverbs provide a strong indicator of how norms and regimes of truth are constantly reproduced as a gender power relation through language. For example, ngwana wa ga malome nnyale kgomo di boele sakeng means ‘my mother's brother's son marries me so that cows get back into the family kraal’ and is a way of reinforcing these discursive sexual relations. In Botswana in marriage the husband's family pays bogadi or bride price of a minimum of eight heads of cattle to the wife's family as a token of appreciation for giving a woman who will in turn give them children – another self-preserving, interdependent function of the mechanisms for ensuring that gender power ‘is a persistent registration of truth’. The extent to which hegemony is truly embedded in such power relations can be seen from the following example. Respondents mentioned that sex with one's uncle seldom happens and it is the best gift a niece can give to her uncle, hence, it could not be protected sex. This is how one middle-aged woman put it:

It is very rare for an uncle to ask for sex as dithogo so you really look forward to that honour. Most of the time they [uncles] ask for small things like a shirt, money, shoes, but this is the man who is vested with all the powers in your marriage and responsibility during hardships.

As with all societies, new rationalities and regimes of truth are continuously reconstructed to counter external attempts to infiltrate discursive practices that have held good for generations. In Botswana this is reinforced by people's awareness of their colonial history and distrust of western attempts to control the balance of power. So it was explained why men were not comfortable with the use of condoms:

Most men do not believe that condoms do not have the HIV virus.

In Botswana there is a widespread belief that if you put water in a condom some worms are seen floating in the water and that is the HIV virus. So condoms are not trusted because of the lubricant and the fact that they do not originate from within the culture. This is how one put it:

Men ask questions such as, where do these condoms come from? I am black or brown and just look at the colour [cream white] of the condom, it shows that it is from the West; it is another tactic of wiping Africans from the face of earth. Unless you show me a factory of condoms in Botswana I can't believe it does not have AIDS.

So even though respondents were aware of symptoms and official prevention strategies they simply referred to their naturalised ‘grooves of being’ (Bourdieu, 1993). As one put it:

What they [modern health personnel] propose cannot happen in this culture. How do you abstain from what is part of you? We will just continue using our traditional medicines and cleansing herbs.

Another said:

We can't stop having children. With or without AIDS the pressure from husband and the extended families is beyond women's control in this culture.

One of the elderly women in this culture emphasised this point:

In our culture, in marriage two people become one in sickness and in health, this means that if the other person is sick you are also sick. How can I listen to people [western medicine] who do not understand my culture and ask my husband, cousin or any potential marriage partner to use a condom? I must be sick in the head to do that.

Shefer & Foster (2001: 377), among others, have already demonstrated how men are constructed through discourses like this, as the ‘active subjects’, while women remain passive objects. If women are to have some sense of agency for behavioural change in the HIV/AIDS scourge then they have to find ways of challenging risky behaviour that is acceptable to their own communities. In keeping with the recommendations (for example, Dowsett, 1999) that communities find solutions from within their own frame of reference, respondents in this study stressed that everybody has to participate in the fight against the epidemic and not just the Government and outside health experts. Their cultural discursive practices and subjugated knowledges have to count as well. In agreement one traditional midwife said:

Our traditional doctors and spiritual healers are now realising that most of our herbs are probably too weak for the HIV virus, so they are referring some patients to the government clinics while working around the clock to identify stronger herbs for this disease.

In all cultures, disciplinary power is protective of the dominant discourse. The discourses identified in the above examples ultimately control the woman's behaviour, rather than the man's – or they are at least designed to favour the man. The men dominate the knowledge production that reinforces the power that controls the females. While these practices seem particularly alien to many external, and medical, discourses, they are presented here to demonstrate they are simply that – discourses. So external discourses are equally alien to localised cultures whose regimes of truth and self-surveillance mechanisms have their own defined rationalities and situated knowledges.

4.2 Understanding sexuality from its own cultural perspective

The cultural dimension of sexuality described in the earlier section shows that sexuality discourses have a social function and they inform gender identity. For instance, in the above section it is clear that for every ethnic group sex is regulated by the family and society because of its diverse roles that include procreation, pleasure, family property, exchange, personal interaction, healing/cleansing, religion/spirituality and control/oppression. A few examples of the social sex roles that emerged from the findings of this study are discussed in this section. It is important to analyse and understand each culture from its own perspective because ideas on HIV/AIDS intervention have to take into account cultural realities and avoid the transfer of HIV/AIDS definitions and understanding based on different experiences.

4.2.1 Sex as procreation

One social function of sex in the Botswana culture is that marriage is primarily a union for the production of children. As evidenced by the nkazana and wife inheritance, sexual behaviour is not ignored in society. In fact, because of this social function the sexual life of married people is affected not merely by individual desires and varying proficiency in the art of making love but by both the family and societal sex regulations. For instance, a man whose wife has given birth must abstain from extramarital sex until he has resumed sex with his wife. Otherwise he would have ‘crossed the poles’. This means he would have disobeyed the procreation spirits' rules. It is believed that failure to abstain in this way will cause the newborn to be stupid or mopakwane, or grow weak and deformed. It is important to mention here that these customary restrictions are based upon locally accepted ideas of disease or other misfortune caused by wrongful indulgence in sex (Schapera, 1971; Strahl, 2003; Castle 2004).

Based on the central importance of producing children, traditional education institutions like the family and initiation ceremonies prepare boys and girls for this role prior to puberty by teaching them how, when and with whom to engage or avoid sexual relations (Gupta, 2000). So very few young people enter marriage with no sexual experience at all, because once past puberty girls and boys engage freely in full sexual relations so that by the time they are married they will have acquired personal experience of intercourse. Therefore, an HIV/AIDS strategy that emphasises abstinence for this age group is definitely not culture-sensitive.

4.2.2 Sex as pleasure

Another social role of sex in these cultures is that it is a source of extended family entertainment. While it is the duty of the wife to give the husband carnal satisfaction and no attempt is made to disguise this, sex with cousins, uncles, stepfathers and boyfriends is testimony that it is also used as entertainment. It is important to understand this cultural dimension of sex for HIV/AIDS prevention, because it has a direct bearing on one message of the current ABC prevention strategy, namely, ‘be faithful to one partner’.

4.2.3 Sex as family property

Because of the role of the family in an individual's sexual practice, another social function in these cultures is sex as family asset. For instance, men or boys who impregnate girls (only with the first child) pay the girl's family eight head of cattle, or money equivalent, for what they call tshenyo or damages. The charge is less if the male is a family member. On the other hand, in marriage the family of the bride receives bogadi or bride price from the husband's family as a token of appreciation for having given them a wife (Schapera, 1971; Gupta, 2000).

4.2.4 Sex as an exchange

The data show that some ethnic groups practice wife inheritance for economic and social responsibility reasons. For instance, supporting a widow and her children is the responsibility of the deceased husband's family and this is provided in exchange for sex with the male relative. The setlogolo ntsha dithogo or ‘sex with mother's brother’ is also performed in exchange for the fatherly responsibility assigned to the uncle. Owing to economic hardships that have resulted in poverty and lack of employment opportunities, the youth and single women in particular mentioned that they engage in unprotected sex outside the family in exchange for money and other basic services. Discussions with participants of this study about their male sex partner's reluctance to use condoms revealed that the act of intercourse itself is an exchange. This is how one middle-aged woman put it:

I am a single parent with three children. I am lucky to have a boyfriend who gives me soap [i.e. money]. You know why he does that, it is because we don't use a condom and he realises that I need soap to wash his semen. Now if we use a condom there will be no need for soap because he simply throws away the condom with his semen.

4.2.5 Sex as cleansing and healing

One sexual practice represented by the different cultural dimensions of sexuality is inter-generational sex or sex between young women or girls with an older male sex partner. This happens because of the cultural belief that sex with young girls cleanses men's reproductive systems, while menstruation and birth cleanses the female reproductive system (Strebel, 1994). The withdrawal method is not only used to prevent pregnancy but it is also used to ‘spill the hot blood’. In the HIV/AIDS era this cultural practice encourages sex with multiple partners, thus increasing the spread of the virus.

The restrictions and regulation of sexual intercourse described by these ethnic groups are associated with the belief that at certain times a person's ‘blood’ (meaning semen or vaginal discharge) is ‘hot’ and until they have cooled down both are in a condition to be harmful to others. For example, a woman is hot during menstrual periods and after an abortion, widows and widowers are hot for a year after their bereavement or boswagadi; a traditional doctor is hot for three days after one of his patient dies. Hotness resulting from sexual behaviour itself is believed to indicate that if the person affected indulges in intercourse before cooling down his partner will be stricken with disease and may die (Seloilwe & Ntseane, 2000).

All ethnic groups reported that they use traditional medicine and herbs to purify the blood and to remove sexually transmitted infections. Healing is another important social function of sex described by the different ethnic groups that participated in this study. In addition to drinking herbs such as purgatives and grapple plant for cleansing the blood, sex is also denied while on treatment for sexually transmitted diseases (STDs). It is important to mention that participants reported that traditionally there was no stigma or secrecy associated with sexually transmitted diseases or any disease for that matter, compared with what they have observed with HIV and AIDS. Modern medicine (commonly understood as being derived from the West) emphasises patients' individual rights. Participants in this study therefore associated the secrecy and confidentiality of the HIV/AIDS ailment with western discourse.

The cleansing and healing social functions of sex seem to have given a lot of power to the family for enforcing all preventative and curative health aspects. The data revealed that in the past this function was carried out successfully under the supervision of the family for all ethnic groups. Unfortunately with HIV/AIDS the family's involvement was overlooked by official preventative policies until very late, with the introduction of the home-based care programme (Ministry of Health, 1997).

4.2.6 Sex as control/oppression

From the gender perspective, the sexual practices and behaviour described by the five ethnic groups fit the definition of sexuality provided by Jackson (2002: 88), who states that sexuality refers to the aspects of gender identity that relate to sex. As far as sexual behaviour is concerned, men in many societies can be proud of having multiple partners because it shows their sexual prowess; but Shefer & Foster (2001) argue that for women sex is predominantly about pleasing men, essentially her husband, and about having babies. Hence sexuality often refers to male needs and desires, while women's sexuality is looked down on, ignored or feared and repressed.

Although women in these cultures have been socialised into not seeing this as oppression or even control, some of the women's statements suggest the contrary, even if they are not yet able to resist. One woman said:

Men in this culture decide on how they want their sex.

Another said:

Having access to condoms and knowing how to use them is not helping us women because if you bring the condoms to a Motswana man he calls you a prostitute.

Nevertheless, albeit rather weakly, these indications of awareness are potential starting points for exploring alternative discourses from within the community that might help to stimulate preventative behavioural change.

4.3 Opportunities for empowerment from within the culture's own discourses

Kindervatter (1979: 62) defines empowerment as ‘people gaining an understanding and control over social, economic, and/or political forces to improve their standing in community’. We argue that understanding sexual practices from inside specific cultural perspectives creates a framework for exploring ways in which HIV/AIDS prevention policies can empower, rather than disempower communities. The following section discusses opportunities for doing this in the Botswana context by interpreting the discourses associated with sexual behaviour.

4.3.1 Collective sexual responsibility

The fact that both the family and society play a key role in sexual matters, and the respect for their culture manifested by all ethnic groups in Botswana, should be seen as advantages that might be tapped for effective HIV/AIDS prevention strategies. The importance and power of the family is not unique to the cultural dimensions of sexuality in the context of Botswana. Caulfield (1993) shows that it is actually a defining characteristic of African culture. This study has shown that in Botswana a woman's place in social life is not in any direct sense a product of the things she does, but rather the meaning her activities acquire through social interaction. Therefore, the family should be a crucial resource in facilitating women's health and empowerment.

4.3.2 Traditional sex education structures

Given that the current HIV/AIDS information, education and communication are perceived as externally imposed and culturally alienating, an understanding of traditional sex education can be analysed to identify indigenous role models. For example, teachers of menstruation rituals, annual sexual education ceremonies and religious sex education activities can be used to mainstream HIV/AIDS prevention mechanisms in ways that appropriate, rather than resist, existing discourses. For example, traditional healers can be encouraged to promote some form of HIV/AIDS testing as part of their healing and traditional counselling processes. They can also be used to test their herbs to identify culturally acceptable treatments for familiar opportunistic HIV/AIDS diseases such as diarrhoea, thrush, TB, etc. These people are very influential, trusted and accessible and their herbs are cheaper than ARV therapy. Religious sex education activities offer abstinence prior to marriage. This can be used for prevention education to address the spread of the virus among the youth. Non-penetrative sex practices as safe sex rituals can be used to introduce other sex-satisfying approaches not known in the culture, although on their own these cannot guarantee effective protection. An African (i.e. African skin colour) condom must be introduced to traditional doctors and packaged locally (in Botswana) to use as part of sex education. Finally, there must be an HIV/AIDS project administered by a non-governmental organisation (NGO)/association of traditional and modern doctors aimed specifically at addressing, orally, the ideas, myths, beliefs and misconceptions about HIV/AIDS. In an oral society such as Botswana information spreads swiftly, so wrong information about the epidemic has spread like wildfire and this has to be remedied through traditional networks.

4.3.3 Participatory approaches

Botswana is known for its effective democratic values such as collective consensus, respect for ideas and opinions shared at the village parliament or kgotla. This participatory approach is a strength that can be used to engage people in critical reflection about the impact of global development agendas on their unique indigenous health culture and curative medicines, including HIV/AIDS prevention. This will not only be in line with the cultural decision-making process but will also facilitate family and community ownership of, and commitment to, agreed HIV/AIDS prevention and care measures.

4.3.4 Involvement of traditional healers

Because of their influence in the health of families and societies, traditional healers can be used in strengthening community-based HIV/AIDS health-care systems such as that of addressing opportunistic infections and peer counselling. The fact that they have registered their concern about being excluded in HIV/AIDS matters (Seloilwe & Ntseane, 2000) means that they are ready and HIV is their problem, too.

5. Recommended focus of empowerment for HIV prevention

Participation in policy decision making has been linked to empowerment as a way of encouraging the public to see themselves and their roles as relevant and indispensable in community activities (Ministry of Finance and Development Planning, 1997). Dean (1999) argues that for a people to be active in their own development they must be given power that connotes mutual respect for, and use of, local knowledge, experiences and interest. In the absence of a cure, behavioural change is key to reducing the incidence of HIV transmission. But there is need for a new HIV prevention strategy that is informed by a relevant empowerment training programme, based on critical thinking (Brookfield, 2000) approaches. Policies need to encourage education strategies that emphasise participation and self-reflection, using the learners' own abundant experiences, as a resource for empowerment and change. We propose that such forms of education are necessary for Botswana to empower themselves to effectively redress the escalating HIV infection rates in their context. This means taking the following positions in relation to the discourses of Botswana ethnic communities.

5.1 Empower women

Women know a great deal about HIV/AIDS but that knowledge is not being used because of their powerless status in decision-making matters about sexuality. Given that HIV/AIDS prevention strategies have focused on providing HIV/AIDS information without recognition of African women's limited economic empowerment, these strategies have not been effective. It is therefore recommended that future HIV/AIDS empowerment programmes should be target-specific and context-based. There are many conferences on HIV/AIDS prevention but participants are either educated elites or in-school youth. It is recommended that a shift of focus should be towards village women in their localities, the out of school youth and the younger youth (aged 10–15). These are the most vulnerable groups in society because they are the ones most immersed in local discourses.

5.1.1 Empower men for HIV/AIDS prevention

In patriarchal societies like the ones described, men have a dual role. They are not just leaders, policy makers and decision makers in matters of sex; they are also sex clients in an environment conducive to the spread of the HIV virus. For instance, nkazana and mantsala have not been reached by HIV/ADS information and prevention strategies. It is therefore recommended that HIV/AIDS empowerment programmes should promote men's understanding of their role and responsibility with regard to protecting women's health and reducing the transmission of sexually transmitted diseases, including HIV/AIDS. This is in line with their cultural role of provider and protector of women. The push for women's rights advocated by those from the women-only gender perspective will not give the required immediate results from the prevention activities. As behaviour change theorists rightly point out, before anyone can change their behaviour they need to know and realise the need to change behaviour and this is always done in relation to one's cultural values and beliefs. It is therefore important to empower the men to dictate safe sex practices by providing them with discourses that appropriate, rather than resist, existing ones.

5.1.2 Work with traditional doctors

There is a need to work with the most influential and effective health educators, namely, traditional doctors, because they are part of the solution. For example, AIDS testing centres should take time to understand and check whether it is possible to test for HIV infection through traditional processes. One of the female traditional doctors interviewed registered her complaint thus:

We have not been fully involved in the fight against AIDS, we refer patients to the modern medical doctors but they never refer patients to us. Is it because they look down upon our practice? They better remember that they [white doctors] and their pills found us here and our patients know better. That is why they still come to us.

5.1.3 Review of condom use

Because of existing cultural beliefs and values, it may be difficult to encourage the use of condoms. It is therefore recommended that existing cultural forms of sexual intercourse should be investigated and encouraged where possible. It has to be realised that the social functions of sex such as those of procreation, and a culturally specific focus on the interconnectedness of spirits with the living and the dead (Avoseh, 2001) may be far more important than the risk of HIV infection which may or may not kill you after a few years. While condoms have been seen to be effective in some circumstances, this is usually as a secondary factor introduced after locally driven and context specific initiatives (Low-Beer & Stoneburner, 2004).

5.1.4 Research

Studies on cultural beliefs and sex education could be useful for identifying cultural norms that can support cultural identity, but further research is needed on how to adapt sex practices that encourage HIV infection. For example, the emphasis on blood and tribal cousin relationships could be retained but, if collectively acceptable, the sexual encounter could be replaced with material gifts instead or some alternatively agreed scenario. Similarly, studies on the relationship between traditional medicine and testing could identify appropriate methods of encouraging testing for HIV/AIDS. New discourses could support traditional sexual practices but in an atmosphere that encourages HIV negative people only to sleep with others who have tested negative.

6. Conclusion

The findings of this study have confirmed that sexuality and sexual behaviour are gender-laden, discursive practices embedded within particular contexts. The social function of sex demonstrated by different ethnic groups includes (among others not discussed here) sex as procreation, entertainment, family property and healing. The study has also revealed that cultures have indigenous sex and health education discourses (disciplinary power mechanisms) that are reinforced through the language of taboos and proverbs. An assessment of the current Botswana HIV/AIDS prevention strategy shows high levels of public information and messages to ‘Abstain, Be faithful, and use Condoms’ (ABC). However, the results of this study revealed that the ABC discourse alone is often ineffective when it conflicts with traditional sex education and practices. Furthermore, there are high levels of mistrust about the information being provided and its lack of synchronicity with cultural beliefs and values. Gender power relations in these contexts are not adequately addressed by the simple instruction that women should take control. Brecause men have the responsibility for providing economic and material things to women and are also decision-makers in sexual matters, they should also be targeted for HIV and AIDS prevention. It is recommended that – alongside raising women's consciousness of gender power relations – men be empowered to realise that as decision makers in sexuality issues they have to lead in the fight against HIV and AIDS.

Although behavioural change is still crucial for HIV prevention in Botswana, it is recommended that alternative policies be context- and target-specific. Further research is needed to identify relevant indigenous knowledge systems, and appropriate traditional sex education content and training methods and their relevance to empowerment activities for HIV and AIDS prevention. Empowerment strategies should focus on the family and the community instead of the individual. Furthermore, it is important to use participatory methods of planning and delivery of prevention activities because of the collective and consensus nature of society in Botswana.

Overall, the findings of this study have demonstrated that HIV/AIDS prevention policies should be based on culturally specific education principles that mobilise people to seek solutions within their own diverse cultural contexts.

Additional information

Notes on contributors

Peggy Gabo Ntseane

Peggy Gabo Ntseane, Department of Adult Education, University of Botswana, Private Bag 0022, Gaborone, Botswana, Southern Africa.

Julia Preece

Julia Preece, Department of Adult and Continuing Education, University of Glasgow, 11 Eldon Street, Glasgow G3 6NH, UK.