Love matters: exploring conceptions of love in Rwanda and Swaziland and relationship to HIV and intimate partner violence

Health risks such as intimate partner violence (IPV) and HIV infection often occur within intimate sexual relationships, yet the study of love and intimacy is largely absent from health research on African populations. This study explores how women and men in Rwanda and Swaziland understand and represent love in their intimate sexual partnerships. In Rwanda, 58 in-depth interviews with 15 couples, 12 interviews with activists, and 24 focus group discussions were carried out during formative and evaluative research of the Indashyikirwa programme, which aims to reduce IPV and support healthy couple relationships. In Swaziland, 117 in-depth, life-course interviews with 14 women and 14 men focused on understanding intimate sexual partnerships. We analysed these qualitative data thematically using a Grounded Theory approach. Participants described love as being foundational to their intimate sexual partnerships. Women and men emphasised that love is seen and expressed through actions and tangible evidence such as gifts and material support, acts of service, showing intentions for marriage, sexual faithfulness, and spending time together. Some participants expressed ambivalent narratives regarding love, gifts, and money, acknowledging that they desired partners who demonstrated love through material support while implying that true love should be untainted by desires for wealth. IPV characterised many relationships and was perceived as a threat to love, even as love was seen as a potential antidote to IPV. Careful scholarship of love is critical to better understand protective and risk factors for HIV and IPV and for interventions that seek to ameliorate these risks.


Introduction
While a wealth of research has investigated intimate partner violence (IPV) and HIV within sexual partnerships in Africa, few researchers have explored how love and emotion shape the intimate relations within which health risks can occur. As Thomas and Cole (2009) argue in the ground-breaking edited volume Love in Africa, "Discussion of emotion has long been absent from Africanist scholarship on intimate relations… Whereas countless studies have analyzed how sexual behavior fuels the [HIV] epidemic, few have explored how that behavior is imbedded in emotional frameworks" (p. 4). Of the handful of chapters in this volume which address love in contemporary Africa, only Mark Hunter's (2009) exploration of "provider love" in South Africa examines the intersection of love with HIV, and none substantively address IPV.
The lack of attention to love and intimacy in Africa is also problematic given evidence that conceptions of love are intimately linked to prevention of IPV and HIV. The SASA! intervention in Uganda found that as IPV and conflict in relationships decreased, love and trust increased (Starmann et al., 2016). In a study of Ugandan HIV serodiscordant couples, some reported that love motivated them to continue the relationship and adhere to pre-exposure prophylaxis (PrEP) (Ware et al., 2012). Lack of condom use within established relationships has repeatedly been linked to perceptions of love, emotional intimacy and trust (Parikh, 2007;Stern & Buikema, 2013), and condom promotion efforts have attempted to reframe condom use as an expression of mutual care and love (Rhodes & Cusick, 2000). Yet love, affection, and tenderness may coexist with power inequities and even violence (Jewkes, Levin, & Penn-Kekana, 2002;Rasool, 2013). Research from South Africa has shown that love may be constructed in ways that legitimise male dominance and IPV (Wood, Maforah, & Jewkes, 1998). Indeed, women may perceive sexual coercion as normalised masculine behaviour if perpetrated by a trusted partner in an intimate relationship (Stern, Buikema, & Cooper, 2016).
In our view, public health, medical anthropology, and allied fields are long overdue for a change in orientation which recognises that love and intimacy should occupy as central a place in research of intimate sexual relationships as they do in people's lived experience. Sexual relations in Africa, as anywhere, can take place within contexts of violence, coercion, and constrained choice. Yet we agree with anthropologist Jennifer Hirsch's critique that public health researchers have often been guilty of depicting "all sex among the poor as the product of violence, lust, or need" (2007, p. 102), unidimensionally portraying women as vulnerable victims and pathologising men as violent or lust-driven sexual aggressors. In prioritising research of violent, transactional and age-disparate sexual partnerships, research has typically problematised sexual relations in Africa and largely ignored relationships that are characterised by mutual love and support. Such positive characteristics may in fact be critical determinants of physical and mental health (Proulx, Helms, & Buehler, 2007) and provide protection against IPV and HIV risk. Love can be a powerful force in transforming unequal power dynamics and gender roles within relationships (Deutsch, 2007). Moreover, attending to love, affection, and positive ideals of intimate relationships strengthens rather than works against attempts to address gendered hierarchies and sexual violence (Bhana, 2013b). Hunter (2005) notes that Western scholars have long viewed Africa as "loveless", but such a view is at odds with the ethnographic record. In a landmark review of ethnographic data from 186 societies, Jankowiak and Fischer (1992) conclude that romantic love "constitutes a human universal, or at the least a near-universal" (p. 154). While they code nearly a quarter of societies in sub-Saharan Africa as lacking evidence of romantic love (the highest proportion of any region), they argue that this purported lack is likely due to ethnographers failing to note evidence of romantic love. Similarly, Ghanaian writer Ama Ata Aidoo (2006) observes that scholars have assumed that "what we do not know about Africa does not exist," and have therefore failed to recognise that Africa has many love stories of the most "serious" nature, in that they speak of the "enormity of the consequences of loving" (p. vii-viii). The virtual absence of scholarship on love in Africa contrasts starkly with a wealth of evidence, including in African literature (Aidoo, 2006), that romantic love is a major component of the lived experience of women and men across the African continent (Bhana, 2013b;Cole & Thomas, 2009).

Love in Africa: historical and contemporary perspectives
Yet scholars have also noted the seismic changes that took place through European colonisation of Africa, in part as missionaries introduced notions of monogamous companionate marriage, marrying for love, and freely choosing one's spouse (Vaughan, 2010). Historian Megan Vaughan argues that many "traditional" African societies set passionate love in opposition to the responsibilities and emotions required by marriage, and that romantic love as the "exclusive motivational basis for marriage" may be viewed as a "distinctively modern and imported discourse" across the post-colonial world (2010, p. 10). Hirsch and Wardlow (2006) describe the rise of "companionate marriage" as a global ideal and assert that public health research has neglected the fact that marriage can be simultaneously a site of gendered power struggle and of love and pleasure. Smith (2001) claims that for Africans, love became a more important criterion for selecting a marriage partner beginning in the latter half of the 20th century. Cole contends that young people in contemporary Africa claim modernity through "adopting the lexicon and practices associated with 'love'" (Bhana, 2013a, p. 99).
In tracing historical shifts in cultural expressions of love, one must ask the related question of what is meant by love. As Cole and Thomas (2009) propose, we approach love as an "analytic problem" rather than a "universal category" (p. 2). We further adopt Hunter's (2010) framework, derived from extensive fieldwork in KwaZulu-Natal, South Africa, of provider love as intertwined with, yet distinct from, romantic love. Provider love is "enacted through cooperation and mutual assistance" including exchange of money and resources within intimate sexual partnerships (Hunter, 2010, p. 16). Romantic love is seen in expressions of passion and desire for freely chosen, companionate partnerships (Hunter, 2010).

Love and material exchange
A rich ethnographic literature describes the complex entanglements of love, sex and material exchange in Africa, with the historical record showing that such entanglements are far from a recent phenomenon (Caldwell, Caldwell, & Quiggin, 1989;Vaughan, 2010). In a study of transactional sex in Malawi, Swidler and Watkins (2007) locate giftsex exchanges within "broader African patterns of unequal interdependence" such as patron-client relationships, which are undergirded by moral duties to support the needy (p. 148). Swidler and Watkins (2007) argue that across Africa, "economic exchange is considered integral to a wide range of sexual relationships, from marriages… to brief affairs" (p. 148). Cole similarly identifies the expression of love through the exchange of money and gifts as common across Africa (Bhana, 2013a), and Vaughan (2010) casts an even wider net in observing that both "interest" and "emotion" have long been elements of romantic love in European and African contexts.
A wealth of research has investigated how material exchange and gifts intersect with love and romance in contemporary Africa (Fielding-Miller et al., 2016;Mojola, 2014;Poulin, 2007;Samuelsen, 2006;Selikow & Mbulaheni, 2013;Stoebenau et al., 2011). Researchers have focused primarily on transactional sex among unmarried populations and particularly on "sugar daddy" relationships between older men and younger women, perhaps setting up this type of transaction as the prototypical sex-gift exchange. Yet not all sex-gift exchanges constitute transactional sex. 1 For example, Poulin (2007) presents evidence that gifts and money are "part and parcel of the courting practices" of young unmarried adults in southern Malawi and are "as much about the expression of love and commitment as they are about meeting the financial needs of women" (p. 2383). Furthermore, as argued by Stoebenau et al. (2016) based on an exhaustive review of the transactional sex literature, transactional sex itself often involves "material expressions of love" (p. 186).
Far fewer studies have investigated the relationship between love and economic exchange within marriages or marriage-like relationships, which outnumber unmarried sexual partnerships in most of Africa. Thompson (2013) records that for a young Zanzibari wife, a husband's love was manifest in how well he provided food and clothing. Cornwall (2002) investigates economic exchange within marriage in south-west Nigeria and offers a conclusion that emerges with remarkable consistency across research from various settings and populations in Africa: love is expressed by "doing something", which Nigerian wives understood to mean spending money.

Theoretical and demographic perspectives
In this paper, we describe how women and men in two study populations in Rwanda and Swaziland understand and represent love in their intimate sexual partnerships. The first two authors initially began to compare findings from these studies during informal conversations. We were struck by the many similarities in the expressions and conceptualisations of love in our respective studies, although the two populations are located in different regions of Africa and the studies were focused on different health risks (IPV and HIV) and used different research questions. We felt that the similarities (and differences) we observed warranted comparison, and that comparing data from two diverse African populations to other available ethnographic data might also inform the question -previously explored by Jankowiak and Fischer (1992), Hirsch and Wardlow (2006) and others -of whether there are commonalities across cultures in how love is experienced within intimate sexual partnerships.
In both studies, we arrived at a focus on love and particularly on expressions of love through a Grounded Theory approach (Charmaz, 2006;Glaser & Strauss, 1967). Love was not an a priori focus of either study, but emerged as a major theme in the accounts of women and men in both settings. While we allowed these accounts to determine our conceptualisations of love as much as possible, we also draw on Hunter's (2010) definitions of provider and romantic love as they adhere closely to our data.
In Rwanda, formal marriage is a critical route for young women and men to achieve societal status and adulthood (Sommers, 2012). According to the 2014-15 Demographic and Health Survey, the median age at first marriage was 22 years for women and 26 years for men (NISR, MOH, & ICF International, 2015). Among adults in their thirties, 1 in 10 had never been in a union while most were formally married (60% of women and 62% of men); an additional 18% of women and 25% of men were cohabiting (NISR et al., 2015). Rwandan women who are not married by their late twenties are often severely pressured by their families and communities to marry (Kaiser Hughes et al., 2016;Sommers, 2012). Rwandan couples who cohabitate and behave as though they are married without formal (legal) marriage are sometimes referred to as having "illegal marriages", which are not recognised or protected under current laws (Mwendwa Mechta et al., 2016;Powley, 2007). The Rwandan government promotes formal marriage as a strategy to protect women's rights to property and custody of children in the case of separation or divorce, but some men prefer to informally marry to avoid giving their partners the rights that accompany formal marriage (Brown & Uvuza, 2006;Kaiser Hughes, Ndangiza & Ikirezi, 2016;Polavarapu, 2011). The costs of bride wealth, marriage registration fees and hosting a traditional wedding have also been identified as barriers to formal marriage (Polavarapu, 2011).
IPV in Rwanda is high, with an estimated 21% of women aged 15-49 years having experienced physical or sexual violence from a husband or partner in the past 12 months, and 40% of ever-married women having ever experienced IPV (emotional, physical and/or sexual) (NISR et al., 2015). Women's acceptance of IPV has been found to inhibit them from reporting the violence or leaving a violent relationship (Mannell, Jackson, & Umutoni, 2016;Rani, Bonu, & Diop-Sidibe, 2004). HIV prevalence is stable at 4% among women and 2% among men (NISR et al., 2015).
In Swaziland, as throughout southern Africa, marriage and cohabitation occur later and are less common than in other parts of the continent, a factor which has been implicated in the region's HIV epidemic (Bongaarts, 2007). The historical and economic reasons for the decline in marriage across southern Africa are beyond the scope of this paper, but the critical point is that while marriage remains a widespread aspiration in Swaziland (Ruark et al., 2016), many Swazis are not married or in a marriage-like relationship. According to the most recent Demographic and Health Survey (2006-07), the median age at first marriage was 24 years among women and 28 years for men. Among adults in their thirties, 20% of women and 26% of men had never been in a union, while for both women and men 53% were formally married and 11% were cohabiting (CSO & Macro International Inc., 2008).
Swaziland has the highest national HIV incidence in the world (Justman et al., 2017), and HIV prevalence is 26% (CSO & Macro International Inc., 2008). There are no population-based IPV data available from Demographic and Health Surveys or similar surveys, but available data suggest a high prevalence of IPV. More than 1 in 10 Swazi women report having experienced forced sex in their lifetime (Tsai et al., 2011), and a recent study of pregnant women found that 1 in 4 reported emotional IPV in the past year, 15% reported physical IPV, and 14% reported sexual IPV (Fielding-Miller & Dunkle, 2017).

Data collection in Rwanda
This paper draws on 58 in-depth interviews with 15 couples, 12 interviews with community activists and 24 focus group discussions (FGDs) with community members conducted in Rwanda as part of the formative and evaluation research of the Indashyikirwa programme. The Indashyikirwa programme aims to reduce IPV and support equitable, healthy couple relationships. It is being implemented by CARE International Rwanda, Rwanda Women's Network and Rwanda Men's Resource Centre. The programme offered a 5-month participatory curriculum to formally and informally married heterosexual couples aged 21 to 59 years who had been in a relationship for at least 6 months. The curriculum was designed to help couples address power imbalances, identify causes and consequences of IPV, and build communication and conflict resolution skills to manage triggers of IPV. Approximately 25% of participants received further training to become activists tasked with promoting non-violent relationships in their communities. The programme also works to support an enabling environment for change by establishing safe spaces for women and by training opinion leaders to prevent and respond to IPV.
Qualitative research was conducted in three intervention sectors which represented rural, urban and peri-urban locations: Rurembo sector, Western province; Gishari sector, Eastern province; and Gacaca sector, Northern province. In each sector, five couples were interviewed separately in November 2015, before beginning the Indashyikirwa programme. Couples were purposively sampled to include formally and informally married couples. Same-gender Rwandan interviewers carried out in-depth interviews in Kinyarwanda, asking about household roles, conflict resolution, communication skills and joint decisionmaking within the relationship. In May 2016, 14 couples were re-interviewed separately immediately after completing the 5-month curriculum and were asked about their impressions of the curriculum and how it had affected their relationships. We recognise that participation in the intervention may have influenced participants' accounts in these interviews, and thus described Rwandan interviews as baseline or midline throughout this paper. In each sector, two male and two female community activists were also interviewed about their experiences as activists in November 2016, after having completed several additional weeks of activism training and implemented activism activities for over one month.
In addition, eight FGDs of six to eight participants each were conducted in each of the three sectors in November 2015. In each sector, participants were purposively sampled to form 4 groups: young unmarried women aged 18 to 25 years, older married women over age 25 years, young unmarried men aged 18 to 30 years, and older married men over age 30 years. Same-gender Rwandan facilitators conducted the FGDs in Kinyarwanda, and interviewed each group twice. For the first round of FGDs, the facilitator presented a social vignette and participants discussed attitudes, behaviours and social norms related to IPV in their communities. Follow-up FGDs explored how men and women should behave in marriage, including division of labour and household decision-making, and common causes of conflict and IPV among couples.
Before each interview and FGD, participants were given information on the aims, risks and benefits of the research and gave informed written consent. Participants were compensated 2 000 Rwandan francs (US$2.65) per interview or FGD. Interviews were conducted at locations deemed private and appropriate for participants. The study was approved by the Rwandan National Ethics Committee

Data collection in Swaziland
In Swaziland, 14 women and 14 men between the ages of 20 and 39 years participated in a total of 117 in-depth, life-course interviews focused on understanding intimate sexual partnerships. Given the goal of sampling participants who were representative of young Swazi adults, participants were recruited in shopping centres or outdoor spaces in central Mbabane, the capital of Swaziland. We recruited participants in two rounds, using purposive sampling in the second round to ensure diversity in marital status and age. The methods of this study have been described in further detail elsewhere (Ruark et al., 2014;Ruark et al., 2016). All participants reported being in heterosexual partnerships (except one woman who was not currently in a partnership) and slightly fewer than half were married or cohabiting.
Same-gender Swazi researchers interviewed participants three to five times each between July 2013 and September 2014, asking participants to describe their previous sexual partnerships and to discuss what makes a good relationship, how a man should treat a woman in a relationship (and vice versa), and whether they thought their intimate partnerships had the hallmarks of a good relationship. Participants were also asked to describe the last time they felt loved, appreciated, or respected by their partners. In addition, they were asked to rank and discuss a set of 15 relationship characteristics which included love. Interviews were conducted in siSwati or a mixture of siSwati and English and were iterative, allowing for exploration of themes in subsequent interviews based on emerging findings or insights from earlier interviews.
Participants were given information on the aims, risks and benefits of the research and gave informed written consent before the initial interview. In addition, participants were asked to give verbal consent at subsequent interviews. Participants were given 20 Swazi emalangeni (

Data analysis and interpretation
All FGDs and interviews were audio recorded, transcribed and translated verbatim into English. The transcripts were then analysed by the second author (Rwanda) and first author (Swaziland) using a thematic coding framework, using NVivo 11 (QSR International, Doncaster, Victoria, Australia). We conducted thematic analysis to uncover predominant themes and provide a rich, detailed and holistic account of the data (Braun & Clarke, 2006). Once love was identified as a major theme, we used focused coding to further explore the data. The first and second authors also regularly debriefed with the research teams after data collection to capture their initial impressions and contextual insights. These research summaries were used to inform the analysis. The overall process of data analysis was conducted collaboratively through discussions among the study teams to identify and describe themes and discuss interpretations.
In writing this paper, the first and second authors shared thematic analysis frameworks for the respective data sets and carried out joint interpretation and analysis. The third author (a Swazi) and fourth author (a Rwandan) played critical roles in verifying the analysis and interpretation presented here and in providing emic (insider) perspectives on the data. Given the long history of racism and disdain displayed by Westerners towards intimate relationships in African cultures, the first author (an American) and second author (a Canadian) endeavoured to be particularly reflexive about their identities as outsiders to Rwandan and Swazi culture, and sensitive in presenting the research findings.

Findings
As much as possible, we present findings in participants' own words. As most quotes were taken from in-depth interviews, we note the source of data only when quotes were taken from FGDs. Participants are further described as married (formally or informally), cohabiting, or partnered (neither married nor cohabiting).

Love as foundational to relationships
In Rwanda and Swaziland, women and men described love as one of the critical reasons for having an intimate sexual partnership and as being essential to their relationships. Swazis frequently characterised love as being the "foundation" or "main thing" in their relationships and as a powerful force that could overcome adversity: If you love each other then your love can stand against anything, even if you have problems… The main reason we are in this relationship is that we love each other (27-year-old woman, partnered, Swaziland).
Love should be the base of every relationship. If there is no love the relationship is as good as dead (34-year-old man, cohabiting, Swaziland). In Rwanda, women and men were more likely to discuss love as a primary but not necessarily singular reason for entering into formal or informal marriage. Other motivations such as social expectations, financial concerns and family pressure also carried weight. Several individuals noted that love was a critical pathway to marriage and that couples in love might marry even if other social expectations had not been met, such as securing family approval for the marriage: I thought that I was going to marry a man with whom I could chat, who would love me, who would not let me work alone and who would not insult me (27-year-old woman, formally married, midline, Rwanda). The first criterion for me [for a wife] is that she loves me (young unmarried man, FGD, baseline, Rwanda).

Expressions of love
In speaking of love in their intimate sexual partnerships, both Rwandan and Swazi participants repeatedly emphasised that love is shown, seen, and expressed. They described love primarily not as an emotive or affective state, but rather as a set of actions and practices, which often involved tangible evidence such as gifts or material support. Love could be shown through a variety of actions that proved a partner's devotion, honesty, commitment, and intentions to formalise the relationship through marriage. These actions and practices were strongly gendered. Not a single Rwandan or Swazi man mentioned love being expressed through gifts and money, but most women did. Men frequently discussed the importance of women showing love through practical acts such as cooking and cleaning. Although men were much less likely to perform such duties, when they did women interpreted this as an act of love. . In Swaziland, both women and men referred to being able to "see" that their partner loved them, which further illustrates the value ascribed to tangible expressions of love. One man commented that, "women are quick to see that they are still loved" (36-year-old man, married, Swaziland), while a woman said that, "I do feel and see that he loves me" (21-year-old woman, partnered, Swaziland). Conversely, some women communicated scepticism about love that was expressed through words but not actions: I think love should be accompanied by actions, because you can tell me that you love me whereas you just love my beauty (31-year-old woman, partnered, Swaziland). Some Rwandan participants related that the intimate "honeymoon" phase of marriage could fade if love was not nurtured and expressed. Practices of love could nurture love and foster reciprocal love from partners: On what thing is love based? For me, I think that there must be something from which grows love. Can love just come without reason? It depends on the care that the girl gives you (young unmarried man, FGD, baseline, Rwanda). Table 1 presents the expressions of love that were most salient in participants' accounts, along with quotes which illustrate how they were conceptualised in the two countries. These expressions range from the practical and material (such as acts of service and gifts) to more relational practices (such as forgiveness and honesty). In many cases these expressions overlap somewhat. For example, respect, honesty, and sexual faithfulness were conceptually linked for many participants. The common theme among all these practices is that they were seen as providing proof and evidence of love.
Several expressions of love were described consistently in the two study populations, such as acts of service, gifts, respect, sexual faithfulness and forgiveness. Other expressions of love seemed to be more salient in one study population. For example, multiple Swazi participants mentioned the importance of a male partner showing his love through taking steps towards marriage and proving that he was "serious". Rwandan participants may have been less likely to emphasise these actions as they were all already formally or informally married. The fact that Swazis were more likely to discuss honesty as a sign of love may signify the normative nature of concurrent sexual partnerships in Swaziland (along with the deception that often accompanies secret affairs), or the fact that even tacit acknowledgment of such affairs was more taboo in Rwanda. Swazis mentioned public displays of affection as a sign of love whereas Rwandans did not, which may be linked to differing levels of social acceptability of public affection in the two contexts. Rwandans were more likely than Swazis to discuss spending quality time with one's partner, and use of terms of endearment to express love. This may be due in part to the fact that the Indashyikirwa programme encouraged couples to spend time together and otherwise foster their relationships.

Love and IPV
Whereas most participants described intimate sexual partnerships characterised by love, many also described experiences of IPV. Most Rwandan women who had enrolled in the Indashyikirwa programme and were interviewed reported experiencing physical or sexual IPV during their lifetimes. Focus group participants were not asked to report on experience of IPV. The programme encouraged participants to reflect on triggers of IPV, and they identified factors including men's alcohol abuse, wives neglecting household duties, disputes over income generation and household spending, men's attempts to correct or "discipline" their wives, and actual or suspected extramarital affairs. Women's experiences of IPV may have fuelled their motivation to join the programme. The Indashyikirwa curriculum supported couples to eliminate IPV and build healthy, equitable relationships, and some Rwandan participants reflected on the relationship of love to IPV. Whereas IPV reduced and undermined love, love could provide a powerful antidote to IPV: When he does it [sex] by force, without preparing me, the consequence that I see is that your love progressively disappears due to misunderstandings (33-year-old woman, informally married, baseline, Rwanda). A 36-year-old female activist related that the Indashyikirwa programme had caused participants to re-evaluate IPV in relation to love: The facilitator told us, "If you really love someone, and that person you love has been in Rugarama market all day, you don't even know if she had an accident or not. Then when it is 7 p.m. you go against the wall with a stick waiting for her to come and you beat her? How can you do that to someone you love?" So we realised that there is no love in doing that. Among the 14 Swazi women interviewed, only 3 did not report previous physical or sexual IPV, and 5 reported emotional or physical IPV from their current primary sexual partner. A large proportion of IPV seemed to occur in the context of violent conflicts which erupted over suspicion or proof of infidelity and often when one or both partners were drunk. Both infidelity and physical violence were sometimes mutual, although men's infidelity was more common and women never matched in severity the sometimes extreme and even life-threatening violence perpetrated by men.
Whereas some Swazi women reported that IPV dealt a death blow to love and the potential for a healthy relationship, other women's accounts painted a more complex picture of the relationship between love and violence. One woman described her relationship by saying, "there is no love here other than being his punching bag", but reported repeatedly breaking off the relationship and then accepting her partner back when he professed his love for her. A second woman took years to separate from a husband who beat her so severely she feared for her life because she, "loved him and had the hope that he was going to change for the better". A third woman described repeatedly suffering physical violence at the hands of a former partner (and sometimes reciprocating physical blows), and repeatedly forgiving him, until she finally decided to end the partnership. She said: The fights caused my love for him to die because you can't say you love someone while you're busy hitting that person… My life changed because I had turned into an animal… He, the father of my child, caused the love to finish [end].

Ambivalent narratives of love, gifts and money
Gifts and material exchange warrant particular discussion. Both Rwandan and Swazi women discussed the provision of life's necessities as expressions of or requirements for love. Narratives about love, gifts, and money were one of the most prominent features of Swazi women's discussions of love: . Financial provision could also be interpreted as a sign of sexual faithfulness, as in the case of a Swazi woman who deduced (correctly) that the father of her child had started another sexual relationship when his financial support of her stopped.
In Rwanda, several women related how receiving gifts from their partners was an expression of genuine love, whereas not receiving gifts could cause emotional pain: Interviewer: So how do you take it when he buys you clothes? What does it show you? Participant: Love. It shows me that he loves me and he also shows me his feelings that he loves me. He doesn't pretend, he means it (34-year-old woman, formally married, midline, Rwanda).

When [my husband] didn't buy me fabric I thought that he didn't love me and that made me feel like I am useless. I even thought of committing suicide
(32-year-old female activist, Rwanda). Some Rwandan men were dismayed by their perception that women only love men with resources: There is a person who loves another person simply because that person has many possessions (older married man, FGD, baseline, Rwanda). In contrast, a notable minority (5 of 14) Swazi women expressed a powerful counter-narrative that women should love their partners regardless of the man's ability to financially provide. They upheld a view of true love that triumphed over poverty, untainted by desire for wealth. In the words of one woman (21 years old, partnered), "even if he doesn't have money or any assets you need to love him unconditionally". Another woman referred to "serious love" which could withstand any situation including poverty. Yet even as they defended an ideal notion of love as separate from economic exchange, Swazi women's statements betrayed the fact that they continued to value a partner who demonstrated love through gifts and money. One woman (37 years old, partnered) upheld the value of loving and respecting a partner in spite of poverty or changes in fortune, yet also described a relationship which involves both love and money as a "package", implying that having a wealthy partner was desirable.
You just have to love someone for the way he is and not for material things he might have and your relationship won't go wrong. You can fall in love with someone who is unemployed but you love this person [anyway], not that he has anything but if he has money then that is a package… Maybe he was working and he was given respect and when he is unemployed then you as a woman need to give him the same respect you gave him when he was employed… Keep on loving him as you know that you can't respect someone you don't love so you have given him the love and the respect he deserves, so that he can see that you don't only love him for material things. The statements of a 31-year-old partnered woman regarding love and money contained a similar ambivalence. While stating that a woman who considers a man's financial resources doesn't "really love" him, she nevertheless affirmed that love is expressed through financial support: Physical attraction and financial support, I regard them as the last things I can consider in a relationship as I feel that if you consider these things then you don't really love this person. If he loves you then he will give you his money. Although most Rwandan participants believed that a man having a house and being the main financial provider were important criteria for marriage, some men and women acknowledged that given harsh economic realities, this was not always possible. In their view, men who demonstrated ability and eagerness to work hard and had plans for economic progress could be desirable partners even if they lacked sufficient income to own or rent a home. As one woman (45 years old, formally married) said, "I wished to have a husband who has a job. You don't always get what you ask for". A Rwandan man expressed the hope that a woman could look past her lover's poverty: She may not base her decision on the house. She may see that you are poor but you have good thoughts, to the extent that she believes that you may work together and make progress together.
In that case, she can see that you can achieve everything (older married man, FGD, baseline).

Discussion
In this paper, we present the accounts of women and men in Rwanda and Swaziland regarding their intimate sexual relationships, and arrive at two conclusions. First, women and men understood love to be foundational to their intimate sexual relationships and conceptualised love primarily in terms of how it was expressed, with practical and tangible expressions being particularly valued evidence of love. Second, careful scholarship of love is critical to understanding risk and protective factors for IPV and HIV and to informing interventions which seek to ameliorate these risks, given that IPV and HIV (or the risk of HIV) often occur in intimate sexual partnerships characterised by love. As Sølbeck (2010) argues in her description of young mothers and their lovers in Mali, love can be understood as a "mode of action". Our data resonate with numerous ethnographic accounts, particularly Hunter's (2010) description of provider love in KwaZulu-Natal, but also with Rule-Groenewald's (2013) finding that adults in Western Cape, South Africa, conceptualise romantic love as something that has to be proven through displays of care, honesty, and respect. The expression of provider love through gifts and material support is arguably the prototypical form of love in action, but was only one of many ways in which participants in our studies described love being demonstrated and proved. We suggest that all such ways of showing love through practical and tangible means are quite closely linked to Hunter's (2010) concept of provider love, and might be conceptualised as part and parcel of the same set of expectations and values regarding love.
The accounts of some Swazi women both upheld and subverted the concept of provider love. They simultaneously took for granted that men should show love for women through gifts and money, and claimed that true love should endure even in the absence of such support. Similarly, women in south-west Nigeria maintained that it was not money but love that was essential to a relationship, as a woman who truly loved a man could stay with him, despite his poverty, and earn her own money (Cornwall, 2002). Cole notes that although the Malagasy word for love (fitiavana) implies material exchange, in contemporary Madagascar young people have begun to refer to love without the expectation of material support as "clean love" (Bhana, 2013a, p. 103).
As historians of love remind us, conceptions of love in Africa (as elsewhere) are not trans-historical (Bhana, 2013a), but shift over time and in response to specific historical and economic realities (Hunter, 2010;Vaughan, 2010). In parts of Africa where men can no longer muster the resources to pay bride wealth and build homesteads, women may still desire husbands who can fulfil the financial obligations of formal marriage. However, these older forms of exchange have largely given way to lesser exchanges of material support between unmarried lovers, as Hunter (2010) meticulously documents in KwaZulu-Natal, South Africa. Rwandan and Swazi women's accounts of sustaining love when provider love is absent suggest that conceptions and expectations of love may be changing yet again, as men are unable to pay not only bride wealth but even to fulfil the expected financial obligations of provider love. Of course, economic pressures are not the only forces at play. Just as globalisation inexorably spread an ideal of companionate marriage across the post-colonial world (Padilla, Hirsch, Munoz-Laboy, Sember, & Parker, 2007) global discourses of romantic and passionate love may be shaping local conceptions of "pure" or "real" love that is untainted by expectations of material support.
Although the participants in this research emphasised provider love and other practical expressions of love, we also note evidence of romantic love in their claims of love transcending material interests, the value given to verbal affirmations of love, and the terms of endearment used by Rwandan lovers. Indeed, Hirsch and Wardlow (2006) note that one component of the companionate marriage ideal has been the privileging of verbal expressions of love rather than "instrumental expressions" such as acts of service. The broad range of expressions of love which emerged in this research -from the romantic and affective to the instrumental and tangible -suggest that the companionate ideal may have become part of the "repertoire of concepts" (Hirsch & Wardlow, 2006) of Rwandans and Swazis but that long-standing patterns of structuring sexual partnerships around economic exchange have far from disappeared.
Any discussion of economic exchange within sexual relationships requires consideration of the broader economic structure within which such relationships are situated. Marriage in particular has long been viewed -in its vastly diverse forms and varied socio-cultural contexts -as an economic contract, and not one that typically benefits women. For Marx andEngels (1848/1967), marriage was a microcosm of class warfare, rooted in men's control of economic resources and women's labour, and inevitably resulting in strife as women became economically oppressed "head servants" (Engels, 1884(Engels, /1972). In the same vein, Simone de Beauvoir famously wrote that women "sell themselves through marriage" (1953, p. 614). More recently, economists have conceptualised any quest for a romantic or sexual partner as an exchange of assets for assets (Adshaday, 2013), as women and men capitalise on their education, youth, beauty, and "erotic capital" (Hakim, 2011) to snare the most desirable partner possible and maximise the benefits they derive from the relationship (Adshaday, 2013).
Our research supports the assertion that love cannot be separated from economic realities. However, it also suggests that people seek assets that are emotional as well as material, and which are perhaps not well accounted for by a marriage-as-contract or Marxist materialist paradigm. In our research, women recast gifts bestowed by lovers as being valuable not because of their economic value, but because of the emotional currency of such gifts as demonstrations of love. In fact, Fielding-Miller and Dunkle's (2017) research with Swazi women finds that there are very real emotional benefits to receiving gifts, in that women who receive more gifts are less likely to report emotional IPV.
Most women in this research were not formally married, but communicated positive feelings and strong aspirations towards marriage. While marriage may decrease women's autonomy and economic capital -and the answer to this may vary considerably between contexts and even individual marriages -many women believe that marriage will bring advantages such as social capital and benefits to their children (Ruark et al., 2014). In Rwanda, women in informal marriages have been found to be more prone to social discrimination, have less influence in household decision-making compared to married women, and have less legal protection than do married women regarding land and property rights, custody of children, and IPV (Stern & Mirembe, 2017). In contrast, Simelane (2011) and Golomski (2016) provide evidence that Swazi courts (both traditional and civil) do not protect married women against mistreatment by husbands, even in cases of extreme physical violence.
We find clues in our data that point towards love as a powerful, irrational, and even dangerous force. Rwandan participants used the phrases "love is blind" (urukundo ni impumyi) and "the eyes of the person who is in love don't see clearly" (amaso akunda ntareba neza). A Swazi participant used the similar phrase "bound by love" (kugigwa lutsandvo). Much like the English phrase "love is blind," these phrases describe how a person in love may lose the ability to accurately discern reality. Being "blinded" or "bound" by love may be one factor which leads women to ignore risks to their health and wellbeing and remain in poor quality or abusive relationships, including those which place them at risk of IPV and HIV. Indeed, some Swazi women who participated in this research reported that love was one factor which had caused them to stay with partners they knew had concurrent sexual partners (Ruark et al., 2014). We cannot fully understand the risks faced by women in such relationships until we take seriously their claims of love.
Our data also suggest how concepts of love, built on respect and trust, can be critical antidotes to IPV. Indeed, Fielding-Miller and Dunkle's (2017) research of transactional sex in Swaziland found that love was protective against IPV, in that women who reported having sex with a partner for reasons other than love or affection (such as poverty, money, hunger, fear of abandonment, or violence) had increased odds of emotional, physical, and sexual violence. Starmann et al. (2016) assessed processes of change among couples exposed to the SASA! IPV prevention programme in Uganda and suggested that the focus of SASA! on core relationship values such as love, respect and trust was essential to reducing violence and improving relationship quality. The authors conclude that this "points to the perhaps untapped potential for interventions to promote love and intimacy as a mechanism to achieve more balanced power in relationships and prevent IPV" (Starmann et al., 2016, p. 9). The Indashyikirwa programme in Rwanda draws on such valuable lessons in its goal to emphasise positive and inspiring messages for couples to aspire to and support couples in building skills to express care and foster high-quality intimate relationships.
We note several limitations. The data presented in this paper were collected by two different studies with diverging methodologies and goals, and neither study was designed with a focused exploration of love. The study methodologies undeniably shaped the data collected, including likely introducing social desirability bias through face-to-face interviews in both studies, and through exposing and equipping Rwandan participants with skills for healthy relationships through the Indashyikirwa programme. As is always the case with small qualitative studies, the findings of these studies may have limited transferability to other populations in which different socio-cultural and economic realities may cause love to be constructed in very different ways. Nevertheless, we view it as a strength of this study that similar findings regarding love emerged from study populations in two different regions of Africa.
Our goal in this paper was to identify facets of love that may extend to many different socio-cultural groups in contemporary Africa. We note the value of cross-cultural comparisons on such unexplored domains in the African context, and as researchers, the importance of attending to what participants deem as important to their lives and intimate relationships. We hope that future research will further explore which expressions or understandings of love are most salient in other African populations, how romantic love relates to provider love, and how conceptions of love are shifting across time and in response to socio-cultural, economic, and historical forces. Ultimately, we hope to see the development of a robust scholarship of love in Africa which humanises romantic and sexual bonds and recognises the many positive manifestations of love on the continent.

Conclusion
Rwandan and Swazi participants in this research, as women and men elsewhere in the world, conceptualise love as being foundational to their intimate sexual partnerships. Positive conceptions of love may provide a powerful antidote to vulnerabilities and health risks such as IPV and HIV. Interventions that seek to not only prevent IPV or HIV but support healthy, equitable relationships have promising implications for the field of IPV and HIV prevention and also for fostering positive forms and expressions of love. Love matters, and the time is ripe for a research agenda as well as IPV and HIV interventions which take into account the various ways that love in intimate sexual partnerships is conceptualised, valued, and expressed.