The concurrency debate
One of the most influential recent explanations for high HIV prevalence in sub-Saharan Africa is the existence of sexual networks characterised by concurrent partners. The strength of the main argument has been well rehearsed: the number of overall sexual partners is not the key driving force of HIV, instead concurrent relationships rather than serial monogamy provide for the virus's rapid spread. One important reason for this is that those who are newly infected with HIV are more infectious themselves and thus able to pass on the virus to a concurrent sexual partner (for instance Epstein and Morris 2011; Halperin and Epstein 2004; Kenyon and Zondo 2011; Kretzschmar and Caraël 2012; Morris et al. 2013; Morris and Kretzschmar 1997; Thornton 2008).
Recently, however, a growing number of scholars have challenged the evidential basis for the concurrent partners thesis (Lurie and Rosenthal 2010; Sawers and Stillwaggon 2010; Tanser et al. 2011). The principal arguments in favour of the concurrency thesis, this critique goes, are based on mathematical models and not the actual demonstration of high rates of concurrency in sub-Saharan Africa or the relationships between concurrency and HIV prevalence. In response, proponents have argued that traditional epidemiological studies do not capture the empirical signature of concurrency which is revealed at the societal and not the individual level (Morris 2010, 31; see also Epstein 2010; Epstein and Morris 2011; Mah and Halperin 2010). Questions of method have therefore become pivotal to this debate: indeed, UNAIDS (2010) is now actively promoting methodological consistency through a series of ‘consensus indicators' on concurrent partners (see also Boily, Alary, and Baggaley 2011).
Without doubt, collecting more accurate quantitative data on concurrency will provide a better evidential basis for investigating the importance of sexual networking to HIV/AIDS. Qualitative work that develops typologies of concurrency is also useful in highlighting different forms of concurrency, for instance polygamy or relationships outside of marriage. At the same time, we need to interpret this data, and its limitations, in contextualised ways. We need to understand the social processes that might lead to, or not lead to, concurrency, as well as those that affect the duration and nature of overlapping relations. In pursuing this task, this paper argues for the importance of historical and ethnographic methods, anchored in political economy, in showing the different social contexts – class and gender relations, labour market and migration dynamics, marital patterns etc. – through which concurrency is shaped and operates.
Studying the political economy of sexuality: from variables to processes
This article's working assumption about sexuality is as follows: sexuality is not simply a set of biological acts centred on male/female genitals – acts that can be turned into variables – but a complex social phenomenon embedded in dynamic and not always easily measurable historical processes, from the very construction of a discrete domain of ‘sex' (Foucault 1978), to the way that sex is inflected by gender, the labour market, migration and much more (e.g. Parker and Aggleton 1999). This starting point yields the view that sexual attitudes and practices are always in flux, and must always be studied as they interact with other social structures and practices.
The ethnographic and historical approach rooted in political economy and outlined here is reasonably well represented in sociology, geography and anthropology (for instance Burawoy 2000; Farmer 1999; Hart 2002 – see Farmer especially on health issues). But, as Schoepf (2004, 17) notes, there is a hierarchy within AIDS research that tends to advantage epidemiological methods:
[to be recognised as real,] facts must be put about by those who are socially authorized to do so. In the domain of epidemic disease, these persons are epidemiologists and specialists in public health, not social scientists, and above all, not ethnographers who use qualitative methods to examine culture: social relations, meanings, and their contexts.
Although links between ‘transactional sex' and concurrent partners have long been recognised, this association has faced increased questioning in recent years. Lurie and Rosenthal (2010, 21), for instance, note in their critical appraisal of the concurrency thesis that ‘what tends to get reported in qualitative studies [on transactional sex] are the “interesting cases” which are often not representative and say little about the distribution of local social norms.' Others have taken this point further by suggesting that the emphasis on sex–gift links in Africa is unwarranted since courting gifts are also common in the West. In their trenchant rebuttal of the concurrency thesis, Sawers and Stillwaggon (2010, 17) ridicule the literature on transactional sex in Africa:
All over the world, people who have sex with each other also have other dimensions to their partnership, and some of those dimensions involve exchanges of services, goods and love, not just sex … Picture the reaction if The Lancet were to publish an article that said, ‘About 80% of US women reported receiving flowers, poetry, candy or jewellery for Valentine's Day, and such transactions in sexual relationships are the norm in the population … .’
But even if Sawers and Stillwaggon's comments correctly point to the need for caution in approaching the concurrency thesis, it is unhelpful to write off ‘transactional sex' in parts of Africa by suggesting, even in jest, that it is analogous to everyday relationships in the US. The approach outlined here, which stresses political economy analysis at the national and subnational scales, forefronts the different context in which the materiality of sex and concurrency might occur. Following this line of thinking, I argue that research on concurrency must not start from a large unit of analysis – the ‘why is HIV/AIDS so prevalent in Africa?' question – and then work down in scale, but begin from detailed research in different areas. Doing so leaves an openness not only for considering the multiple (and multiply spatialised) causal processes that affect sexuality but the uneven geography of HIV cofactors from male circumcision, to nutrition rates, to the prevalence of existing sexually transmitted diseases. This is global health, but from the bottom up and not the top down.
What follows therefore is not an attempt to prove one way or another whether concurrency is the central cause of high HIV prevalence in South Africa. Though the attention to ‘transactional sex' does yield the view that concurrency is one of a number of important factors driving HIV infection, it shows that concurrency cannot be seen as a single practice that a technical agenda can simply intervene to change. The rush to acronymise health issues (e.g. CP for concurrent partners or TS for transactional sex) serves to simplify by bounding complex, interconnected, processes. Yet the different ways in which sexuality intersects with other aspects of life, from gender, to the labour market, to condom use, are absolutely vital to its constitution.
To underline the importance of historicising sexuality, and develop a critical comparative lens, this paper begins with a brief overview of changing relationship patterns in the United States. There is a relatively large amount of published work on sexuality in the US that justifies and assists this task. Using this approach, I hope to give a sense of the importance of historical change to sexuality and, at the same time, rebut the suggestion by Sawers and Stillwaggon (2010) that Valentine's gifts in the West are somehow akin to ‘transactional sex' in parts of Africa.
Empirically, the shorter section on the US derives from secondary data and the South Africa data derives mostly from a long-term historical-ethnographic study that explores transformations in intimacy over the last century to better understand the contemporary AIDS pandemic (this is outlined in detail in Hunter 2010). The central question that frames South African research is how HIV prevalence can remain so high – at around 30% for most of the 2000s – despite a large amount of awareness about HIV transmission among the general population. The bulk of the South African research was conducted between 2000 and 2006 and involved the author living extensively in a predominantly isiZulu-speaking area in the province of KwaZulu-Natal. More than 95% of the residents of this area would have been classified as ‘African' in the apartheid era (at the end of the apartheid era there were four widely used ‘racial' categories: African, White, Indian and Coloured). This article draws very selectively on this and other research to make some general points; it cannot do justice to the diversity of views and the politics of a white man undertaking this research (though see Hunter 2010).
Valentine's gifts as transactional sex? Sex, love and gifts in the US
Frederick Engels’ (1971 [1884]) Origin of the Family, Private Property and the State first inspired debates about the relationship between capitalism, gender inequality, sex and love (see Sayers, Evans, and Redclift 1987). In recent years, attention has turned from the materiality of marriage to courting. In her influential account, Beth Bailey identifies a significant rise in the materiality of sex in the early twentieth century, as, she writes, ‘dating moved courtship into the world of the economy. Money – men's money – was at the centre of the dating system … men became the hosts and assumed the control that came with that position' (Bailey 1988, 21). This account hinges on a change from when men ‘called' on women at their homes to when they took women out on ‘dates' at movies, restaurants and dances. This led to an expectation that, in return, men would have sexual access to women, even if relations often stopped short of penetrative sex. Supporting her argument, Bailey notes that in its early days the term ‘date' had a close association with prostitution (22).
Drawing attention to a slightly earlier period, the turn of the twentieth century, Elizabeth Clement (2006) discusses the phenomenon of ‘charity girls’, who were quite explicitly ‘treated' by men in order to access new consumer worlds. This form of ‘sexual barter', Clement argues, was an ‘ingenious compromise' between poor women's need for respectability and their wish to access the new benefits of consumption (3, 48). Clement is particularly attentive to the dynamics of race and class, arguing that it was working-class practices of treating that led to the materiality of dating noted by Bailey: ‘Treating provided the model for the economic and sexual exchange that became a hallmark of dating for young people of both classes' (227). Hence, despite differences, both Clement and Bailey illustrate a growing materiality to everyday relationships: indeed, according to these accounts, ‘going Dutch' (splitting the bill) was considered as an embarrassment by most men and women in the first half of the twentieth century.
The post-war era, notes Bailey, witnessed the rise of more long-term ‘going steady' relations. However, a series of forces rooted in the 1960s and 70s transformed courting in ways that still reverberate today. First was the ‘sexual revolution' itself, partly fuelled by the invention of the contraceptive pill, a technology which facilitated the separation of sex from childbirth. Second was the rise of the feminist movement that promoted greater equality between men and women. These changes are sometimes seen to have led to a scenario whereby ‘[s]ex was uncoupled from romance and love' (Seidman 1991, 121). Indeed, discussing Western society, Anthony Giddens (1992) suggests that the skewed notion of romantic love – which led to women's domestic subjection – gave way to a more equal ‘confluent love' based on a ‘rolling contract' centred on mutual benefits, including sexual satisfaction. In turn, popular media programmes like Sex and the City reflect (and help to produce) a greater acceptance that women are entitled to sexual pleasure outside of marriage (see Akass and McCabe 2004).
One can see in all the trends noted from the 1960s the potential for sex to be separated from romantic love and, to take this further, become more instrumentally exchanged for gifts. Yet, on the whole, most US women do not today enter into relations premised on sex–gift exchanges. In Sex and the City, the four female stars all wish for their men to spoil them, but it is women's independence that shines through the series – their ability to say no to men and attain sex on their own terms. A somewhat similar theme emerges from accounts of relationships among college-goers. In her book Hooking Up, Kathleen Bogle (2008) shows that women often initiate casual sex on college campuses, though compared to men they are more likely to prefer longer-term ‘relationships'. Bogle also shows that participants in a ‘hooking up' encounter can retain contact and at a later time entertain a ‘booty call' (a late-night call for sex). Depending on how concurrency is defined (i.e. whether such an encounter would count as an enduring relationship that overlaps with others), this example could provide evidence that concurrency relationships are indeed quite common among certain groups of Americans.
At the same time, Bogle argues that hook-ups often do not involve penetrative sex but kissing and oral sex – the latter being a practice that young Americans today typically do not see as ‘sex’. Moreover, it is striking how gifts are barely mentioned by Bogle's informants, and this suggests that male–female gifts play only a small role in encouraging lovers to hook up: college women, after all, are embarking on careers that will lead them to experience financial independence. This sense of greater equality in male–female relations, including economic equality, is clearly also important if we consider possible discussions/negotiations over condom use, though this is also not a topic addressed in any detail in the book.
College students are, of course, a relatively privileged segment of US society, and Bogle also spoke mainly with white students. A contrasting literature suggests that everyday sex–gift connections are quite common among poorer Americans. One of the most graphic studies is Phillipe Bourgois’ (2009) ethnography of drug addiction in San Francisco. Here he shows how poor (largely black) men and women, marginalised from the labour market, can form affective bonds in relations that entangle love, gifts and addiction. Other studies – some themselves influenced by the ‘transactional sex' literature in Africa – have found links in the US context between race (and, by association, class) and the materiality of sex (Dunkle et al. 2010).
What this admittedly very brief reading of the US literature seems to suggest, therefore, is that rather than taking concurrent partners as a single category of analysis, it needs to be handled with care. Valentine's gifts between university students, even in a ‘hooking up' culture, are quite different from gifts between men and women when large discrepancies in wealth are apparent. Moreover, at least for middle-class women – and women in the US are, of course, on average considerably richer than those in Africa – sex has become arguably less material outside of marriage. While, at the beginning of the twentieth century, the new dating culture was formed very strongly in relation to men's role of providing for women, hooking up, in contrast, is partly an expression of women's relative economic equality with men.
Sex, love and gifts in South Africa
South Africa has a very different history of sex–love–gift connections to the US.1 In the early part of the twentieth century, economic hardship meant that a significant number of poor black and white women were pushed into prostitution (Freed 1949). However, the state's preferential policies for whites in respect of the labour market, housing and social services, helped pull many white women out of this sexual economy. In contrast, racialised state policies discriminated violently against black South African women, especially after apartheid's introduction in 1948. For this group, prostitution – or more informal relationships that might today be called ‘transactional sex' – continued to be a means of survival in urban areas (Bonner 1990; Jochelson 2001).
However, while ‘transactional sex' has a long history, in the first half of the twentieth century the numbers of men in cities greatly outweighed the numbers of women and marriage rates remained high; moreover, the building of urban townships was a deliberate attempt to limit the number of single women in cities by allowing them to stay in urban areas only as wives. For courting couples separated by men's temporary migrancy from rural areas, ilobolo (bridewealth) constituted the most important means through which men expressed love. Ilobolo – usually of 11 cattle in the KwaZulu-Natal region – signified commitment in part because it took several years for a man to pay in full. Men, therefore, did not typically buy their favourite girlfriends dinner to express love but gave ilobolo to their fathers.
How did ilobolo and love become so intertwined? In the nineteenth century, cattle from a father's rural homestead allowed a man to secure ilobolo to marry his girlfriend. By roughly the 1930s and 40s, however, state taxes and associated rural failure had left unmarried rural men with effectively no choice but to work in order to lobola (pay bridewealth for) their girlfriends. As millions of men became migrant labourers, money became intertwined with love and marriage but in a very different way to the US. Notwithstanding tensions wrought by men's long absences, their roles as providers engendered something of a ‘patriarchal bargain' between men and women (Kandiyoti 1988). A young man was expected to find work, lobola a woman and financially support a home. In turn, women raised children, performed domestic work and engaged in homestead agriculture – the latter which declined in importance as rural areas became more overcrowded and depleted. What this arrangement meant is that, by and large, monetary transactions were not channelled through men's gifts to women in pre-marital relations but through the more meaningful payments of ilobolo to a woman's family, usually over several years, and remittances sent to a wife residing in a distant homestead.
Certainly, evidence in the 1960s suggests that gifts from boyfriends to girlfriends did not play an important role in driving rural courting relations. In one part of KwaZulu-Natal, the anthropologist Vilakazi (1962, 49) reported great shock when asking about courting gifts because ‘the boy would be accused of trying to gwaza (bribe) the girl to love him.' The use of the word gwaza suggests that courting was a domain where gifts (to influence the outcome) would be seen as improper, even immoral, interventions. In rural areas a principal aim of a young wage-earner was to channel his earnings into ilobolo payments for a future wife.
Even in urban areas where liaisons could be easier, marriage was still the expected path and, related, sex–gift exchanges did not appear to be prevalent. For instance Levin (1947, 22), discussing Langa township in Cape Town, reported mainly reciprocal gifts between young men and women: ‘During courtship men try to win the favour of girls by giving them presents such as slabs of chocolate, jewellery, and scarves. Women, in turn, are said to give their boyfriends presents, such as ties and socks.' In sum, throughout much of the twentieth century marriage rates remained high and a large amount of unmarried men's wages were channelled into ilobolo.
It is possible to argue, however, that since roughly the 1970s a set of historical processes can help to explain the coming together of love, sex and gifts in a new configuration in South Africa. Rather than assuming similarity across Africa, detailed historical analysis is necessary because social processes interact in different ways and with a complexity that regression analysis or other quantitative tools cannot easily measure. Consequently, some of the themes relevant to the case of South Africa are applicable to other parts of the continent while others are not. For instance, unemployment and de-industrialisation are common across the continent after World Bank/International Monetary Fund ‘structural adjustment' programmes in the 1980s; some scholars have shown links between these changes and social inequalities that result in sex–gift links (e.g Schoepf et al. 2000 in then Zaire). In Nigeria and Uganda, scholars argue that love leads to condomless sex within marriage, which is still the norm for young people (Parikh 2007; Smith 2006). However, marriage rates are particularly low in southern Africa (Bongaarts 2007). Social inequalities are also higher in southern Africa because the large mining and manufacturing industries were powerful forces generating wealth and poverty in the region.
Specific to South Africa, what follows are some important historical trends relevant to understanding contemporary sex–love–gift links and concurrency. These changes have left a large number of increasingly mobile, and rarely married, women dependent on men – though typically not in prostitute but girlfriend/boyfriend relations that can endure for some time.
First, the rise of chronic unemployment
Absolutely central to the materiality of sex is the tremendous rise of unemployment from the late 1970s, a trend accentuated because of trade liberalisation that followed democratic elections in 1994. Today, only 6.6 million people are in full-time work, around 3.1 million are in outsourced work, 2.2 million are in informal work and 8.4 million are unemployed (Von Holdt and Webster 2005, 28). For young people especially, unemployment has reached terrifying levels.
Second, an increase in social inequalities that is highly gendered
Linked to rising unemployment is a widening gap between the rich and the poor. Since class-based differences work through and exacerbate gender inequalities, many single women today are dependent on men. Though a gendered wage gap has long existed, what accentuated this situation is the continued movement of largely unmarried women into the labour force, especially through their migration from languishing rural areas (Posel 2006). There is therefore a vast chasm between those working (disproportionately men) and those not working (men and women).
Third, stark reductions in marriage rates
Marriage rates have halved since the 1960s: less than 30% of black South Africans (and disproportionately the oldest) are now in wedlock (Hunter 2010). Unemployment levels among black South Africans serve as a key contributor to reduced marriage rates. The decline in marriage for poor people has created a profound sense of distrust between men and women, and helped to structure love as being more entangled with men's support of women through individual gifts – most men, after all, are neither saving money to pay ilobolo nor supporting a woman as a wife. The fact that most people's lives do not revolve around a marital home has also contributed to the greater movement of women (below). What needs to be understood therefore is that transactional sex is not just non-marital sex in the sense that it occurs outside of marriage. Rather, just as marriage configured sex and money in a particular way, the reduction in marriage rates today – and yet persistence of the institution in terms of gendered meanings – creates new sex–love–gift connections.
Fourth, geographical shifts
The pattern that characterised much of the twentieth century – migrant men working in urban areas and supporting a rural wife – has diminished. Now, both men and women are very mobile (Posel 2006), and this has important consequences for sexual networks that can stretch between rural and urban areas, with a new intensity. Women arriving in towns face particular difficulties in finding work and can rely on men for material support. These trends are manifested geographically in the tremendous growth of informal settlements since the 1970s, areas that are often the first home for new migrants to towns (Hindson and McCarthy 1994). Significantly, in these marginalised areas of towns, HIV rates are twice as high as in formal urban or rural areas (HSRC 2002, 2005, 2012).
Contemporary relations in South Africa: what has love got to do with sex and gifts?
In the US, we have seen that a culture of ‘hooking up' might involve several overlapping partners (depending on how one defines overlapping). However, these encounters’ central organising principle is mutual pleasure rather than romantic love – love in this context suggests long-term monogamous relationships with some possibility of eventual marriage. Indeed, some observers argue that sex became progressively delinked from romantic love among the US middle class in the latter part of the twentieth century. We also saw that courting gifts, for instance men paying for a meal, do not appear to play as large a role today as they did in the early twentieth century; though some expectations of male chivalry endure, ‘going Dutch' is now more accepted, at least for the middle class.
In some respects, South Africa's multi-racial middle class bears some resemblance to the US middle class in terms of being likely to marry; moreover, women's greater ability to demand sexual satisfaction in non-marital relations is tied up with – though in very complex ways – some women's greater economic independence from men (Hunter 2010). What the quite extensive ethnographic literature on poorer (mostly ‘African') South Africans makes clear, however, is that sex, love and gifts are now very closely connected (Dunkle et al. 2004; Hunter 2002; LeClerc-Madlala 2003; Selikow, Zulu, and Cedras 2002).
These material boyfriend/girlfriend relations are different from most forms of prostitution because they can endure for some time and embody feelings of ‘love'. Of course, the presence of ‘love' does not signal equality – almost always it is men who give gifts to women. And neither is there a single meaning of love in existence, especially across a diverse range of South Africans. Love can be represented by a woman's commitment to a lover who is not able to support her, a sense that ‘love conquers all' evident in common notions of romantic love (Hunter 2010); yet love can also be expressed in men's gifts to girlfriends, a scenario that might be called provider love (Ibid.). These different meanings of love overlap in everyday life, and this is one reason why quantitative data on ‘transactional sex' is so difficult to gather: whether a person will say that a relationship is based on gifts or love depends on the context of the conversation.
One unfortunate consequence of the recent prominence of the concurrency thesis is that condom use has been somewhat neglected in the AIDS field. But the ethnographic record is clear: ‘love' is often an important reason for the non-use of condoms. The most instrumental money/sex exchanges, for instance prostitution, are most likely to lead to condom use (Hunter 2010; Preston-Whyte et al. 2000). Indeed, today both monetary gifts from unmarried men to women and condomless sex can signal love in important (and dangerous) ways. These meanings are not of course wholly new; they echo long-standing emotional and physical bonds that rested on husbands supporting wives. But while love has always had a material dimension, today it is typically instituted outside of marriage, since economic circumstances make marriage so difficult.
Of course, male–female relations often unfold in unpredictable ways and their trajectory depends on a host of factors from whether a couple are cohabiting to the level of physical attraction. Yet, the way in which different concurrent partners are typically differentiated is a good example of the importance of understanding the history of relationships. In KwaZulu-Natal, a woman's main lover, described by various terms – istraight or iqonda (straight), ‘number one' or umkhwenyana (fiancé, if he has begun to pay ilobolo) – is the one with whom the relationship is most serious and might one day lead to marriage. In the rare cases when ilobolo payments have been initiated, the status of a woman's main lover is raised substantially. Such payments are the most decisive symbol of commitment and, most residents would say, obligate a woman to be faithful to that man. Importantly, it is in these istraight relationships that condoms are least likely to be used.
Secondary lovers – that is, lovers who are not istraight – can, at times, be casual partners with whom relationships are brief. Yet although surveys of sexual behaviour commonly distinguish between ‘main lovers' and ‘casual lovers' – the latter term resonating with a number of English concepts such as ‘fling’, ‘hook-up' and ‘one-night stand' – secondary lovers are rarely the same thing as casual lovers; relationships with them can persist for some time. A relationship's secondary status is typically determined not by how long it is expected to last, but by the lesser obligations and expectations it creates and by its more secret nature; for instance, a secondary lover can be called an ishende (secret lover) or umakhwapheni (also hidden lover, lit. under the armpit). The primary determinant of a relationship's status is not its duration, but the nature of its bond. These secondary relations are more likely to involve condom use but, because of the way that reciprocities change over time, this is by no means certain.
Readers might ask how the boyfriend/girlfriend relationships described above differ from marriage. One difference is that the sense of mutual fidelity is much weaker. Of course, extra-marital relations have long been practised in South Africa as elsewhere (especially by men) and there is certainly no clear association between marriage and HIV in the continent (UNAIDS 2009, 23, 24). Nevertheless, marriage does structure relationships in important ways. As we saw in the case of ilobolo, it leads men to save for many years for a single wife. It structures a household as being, to some extent, a shared project geographically located in a physical house. In contrast, unmarried people tend to be more geographically mobile (Posel 2006).
As intimated earlier, while this paper focuses on the ‘transactional sex'/concurrency nexus, it is important to note that there are other important processes in which concurrency is entangled, for example decisive recent shifts in masculinities and femininities at the time of chronic unemployment. The fact that there is a high status attached to men having multiple partners talks to both continuities and shifts in masculinities over the twentieth century (Hunter 2005). Similarly, women's ability to have multiple partners has increased significantly in recent decades as men fail to marry and support them.
Nevertheless, for the purposes of this paper, two points need to be emphasised. First, in contrast to accounts of middle-class Americans, sex, love and gifts have become more and not less connected among poorer South Africans. This is not to say many relationships in South Africa are not fraught with great tensions – they are – but relationships are still usually framed in terms of love, even the most material and violent ones. Indeed, sexual violence, which I don't consider here, is quite widely reported in South Africa, but cannot be seen as conceptually separate from histories of love. Second, the particular history of relationships means that differentiation tends to take place between lovers in terms of the nature and not simply the duration of the bond (i.e. primary/secondary lover rather than boyfriend/one-night stand).
Conclusion
A certain confidence has surrounded claims over the last decade that concurrency patterns explain high HIV prevalence in Africa. At last, it was thought, an explanation made sense that did not position Africans as being more promiscuous than Westerners (i.e. having more lifetime partners), just partaking in sexual relations that happened to be concurrent. However, in the last few years, this thesis has faced something of a backlash. The evidence for more concurrency taking place in Africa than elsewhere is unproven, critics argue.
The article's contribution is to widen the debate. This means, first, situating concurrency not as a discrete ‘cultural' practice that can be studied in isolation. Rather concurrency is embedded in wider social processes, and one process explored here is the growing materiality of sex: gifts from boyfriends to girlfriends.
Instead of there being rigid geographically based differences in sexual behaviour (Africa and the West), some aspects of concurrency in the South Africa and the US, to use this example, are similar and some are very different. We must recognise the ways that race and class engender concurrency, sexuality's constant flux and the fine-grained meanings of intimate relationships. It really does matter if a man is giving several girlfriends Valentine's gifts in a US college, or if a man is supporting one woman with housing and another with food, in southern Africa. In the latter case, relationships might endure for some time, be structured with great inequality and yet have a definitive emotional aspect.
One consequence of viewing concurrency as a series of practices always entangled with wider social structures is that resultant concurrency interventions shouldn't be seen as separate and competing to others, especially condom promotion. Concurrency and condom use are not practices that operate in separate social silos: critically, condoms are much more likely to be used in short-term prostitute relations than in longer-term relations underpinned by gifts. Indeed, the long duration of some concurrent partners clearly has important implications for HIV infection. So too does the history of concurrency raise questions about the social policies that might reduce the dependence of some women on boyfriends, for instance the huge recent increase in state grants in South Africa which now benefit over 16 million South Africans (Khan 2013).
More generally, if AIDS prevalence is ever to be significantly addressed, stronger links must be made between concurrency and underlying structural issues that in South Africa include high unemployment, the growth of shack settlements and reduced marriage rates. Recognising how the political economy of concurrent partners varies in different settings and at different times will enable a more nuanced and contextualised understanding of the relationship between social structure and HIV infection.