The Emergence of the African Sexual Hypothesis
In the early years of the AIDS epidemic, the registration of cases by American epidemiological surveillance centers was based on the identification of characteristics shared by people afflicted with the disease, making it possible to identify a number of “high-risk groups.” At first, AIDS was thus known as the “four H’s disease,” for homosexuals, hemophiliacs, Haitians, and heroin addicts. As researchers sought to identify the pathogenic agent, its forms of transmission had already been well established. Because it could be transmitted from mothers to their children, between heroin addicts who exchanged needles, and between hemophiliacs using blood derivative products, it was clear that an infectious agent was transmitted through blood. And because homosexuals contaminated one another and women whose partners had the disease were infected in turn, it was clear that it could also be transmitted sexually, both homo- and heterosexually.
Heterosexual transmission, which at first went unnoticed and was even denied,  was invoked in particular to explain the situation in Africa. It became official in 1986 and 1987, when the World Health Organization proposed its typology of epidemiological profiles, which distinguished between, on the one hand, North American and Western European countries in which homo- and bisexual transmission predominate (type I), and, on the other hand, the countries of Sub-Saharan Africa, where heterosexual transmission was assumed to prevail (type II). Heterosexual transmission, which by the late 1980s was seen as the primary characteristic of “African AIDS,” was the central concept on which epidemiological data collected in Africa was based and analyzed.  Other forms of exposure to the risk of infection were at times downplayed, at others dismissed. Attention was focused on African “sexual activity,” which, it was believed, had to be characterized, measured, and, ultimately, modified.
- Affiche de campagne anti-sida
An immense body of literature has emerged from these concerns. While many questions were raised and different approaches proliferated, it must be emphasized that this work was based upon a shared conception of sexuality. A brief look back at history sheds light on its emergence. In the early years of the epidemic, the scientific community sounded a cry of alarm: nothing—or next to nothing—was known about human sexuality. While this anxiety echoed the confusion many felt before this new epidemic, there is much that this assertion overlooked. Numerous anthropological studies made it possible to document various forms of union and representations relating to fertility. Sociological studies flourished in the 1970s, exploring topics such as the sexual lives of couples and changing norms concerning the onset of young people’s sexual activity. Finally, in the field of biology, the term “sexual behavior,” coined by Kinsey, appeared in the mid-twentieth century. With the latter’s oeuvre, which gave birth to clinical sexology, the scale of sexual research changed: the study of particular sub-groups gave way to the study of the general population. Though sexuality had already been the subject of research in specific disciplines, what changed with studies undertaken as part of struggle against AIDS was the type of questions that were asked. Turning away from married couples and questions relating to masturbation, orgasms, and procreation, researchers began to examine oral and anal intercourse and partner networks, which were conceptualized analytically as “high-risk” behaviors or practices. Sexuality was now conceived as an epidemiological problem or, more precisely, as a health risk. This return to the association between sexuality and health, which harked back to nineteenth-century campaigns against venereal disease, brought an end to the reverse trend—the de-medicalization of sexuality—that had emerged following the Second World War.
Over time, a number of sexual explanations of the African epidemic were formulated. I propose to consider three of them. The first relates to one “high-risk behavior” that has existed in other times and places: prostitution. The two others refer to behaviors for which specific categories had to be elaborated: “transactional sex” in the 1990s, and “multiple concurrent partnership” in the 2000s.
The Changing Face of Prostitution
Ever since its first decade, prostitution was considered one of the epidemic’s inroads into the broader African population.  In “northern” countries, sex workers (SW) were identified as a group that had been particularly affected by the epidemic since its earliest beginnings. Some authors, drawing, on the one hand, on studies undertaken in various African contexts showing this group’s high seroprevalence rate, and, on the other, on the claim concerning the primacy of heterosexual transmission in Africa, put forth the idea that it was prostitutes and their clients who first spread the epidemic to the general population. These analyses, which at times legitimated the implementation of repressive legislation authorizing obligatory testing and the quarantining of afflicted individuals, were nonetheless contested in the 1990s. It was pointed out that SW could only be vectors of HIV transmission in conditions that had accumulated in Africa, but not in most Western countries. It was shown that vaginal or oral sexual relationships—which are by far the most practiced by women in prostitution—did not constitute high-risk activities as such, either for acquiring or transmitting HIV. Thus HIV does not behave like a traditional sexually transmittable infection (STI): among SW, infection was less tied to “sexual promiscuity” than to other risk factors relating to drug injection, such as having a drug-injecting partner or having a preexisting sexually transmittable infection.
The scientific debate over the role played by SW in the development of the African epidemic was subsequently placed on the back burner, and only in the 2000s did it reemerge, to be considered, this time, from a different standpoint. This was part of the broader project of placing high-risk groups back on the political and scientific agenda. They were now called “target groups,” and, more recently, “key groups” (so as not to reproduce the stigmatizing logic of other formulations). With the support of the growing power of the Global Fund to Fight AIDS, Tuberculosis and Malaria,  the return of the concept of key groups drew on studies showing that not only were the epidemic’s effects on these groups poorly known—despite the fact that they were overexposed in the public sphere and massively afflicted—but also that the share of funds allocated to them in the fight against AIDS was completely inadequate.  In this context, a new preoccupation emerged: what impact would such actions have on the epidemic’s overall dynamic? A number of researchers began to examine this question, elaborating mathematical models (called goal models) that allowed them to estimate the impact. By way of an example, whereas some studies consider the impact of improving prevention and treatment coverage on a key group, others assess the impact of reducing violence on the number of new infections among SW and, thus, on the general population.
- Peinture murale : Campagne au Congo
In this research paradigm, it is claimed that when determining how key groups impact the epidemic’s dynamic, “size matters.” Thus in 2007, a study estimated that the number of SW in a country is a robust causal factor for explaining the HIV prevalence level on a national scale. For these authors, this explains why Africa finds itself in this situation: as the number of female sex workers is proportionally higher, the epidemic is greater.  Yet this assertion seems rather hazardous, for at least two reasons. First and most commonly, because the categories of prostitution and sex work necessarily have multiple meanings. Since their definitions vary according to period and context, any attempt to estimate the number of SW necessarily depends on a problematic choice of definitions, which is rarely made explicit. Finally, because the numerical estimation of a social reality that is the object of moral condemnation, in addition to eluding routine governmental record-keeping, is dangerous. To hold that that there are proportionally more SW in Africa than elsewhere is an assertion more than a demonstration. While this mathematic model may seem conclusive, its premises, however, are entirely questionable.
Are African Sexual Relations More “Transactional”?
In addition to analyses that refer to prostitution to explain the scale of the epidemic in Africa, a new kind of sexual behavior began, in the early 1990s, to appear in the literature: transactional sex. Two Australian demographers, John and Pat Caldwell, have been the most frequent advocates of this concept. In an article from 1989 that was vigorously denounced, they argued that in Africa, sexual relations have a transactional character—in other words, they are accompanied by an exchange of money or goods. In their view, there exists an “African system,” which is distinct and endowed with its own internal coherence. One of its peculiarities consists in lower moral and institutional barriers than those found in the “Eurasian system.”  Their claim is twofold: Africans are more sexually active than Eurasians for reasons that could be described as cultural (tied to the “African system”); and sexual relations in African are characterized by a transactional dimension. This approach, which in its patent culturalism verges on racism, was abundantly criticized. And while the idea of an “African system” was largely discredited, the transactional thesis has, for its part, been considered an established social fact in Africa. In political and scientific circles, moreover, it was seen until the mid-1990s as a major factor in HIV transmission in sub-Saharan Africa. 
In the 1990s, studies addressing this issue flourished. Some authors set out to analyze sexual transactions from the standpoint of structural rather than cultural factors, showing how the connection between material exchange related to sexuality and the major socio-economic transformations underway in African society. Within this paradigm, some placed the emphasis on economic need (with some even proposing the idea of “survival sex,” when exchange is related to immediate survival), while others tended to stress the relationship between these factors and forms of consumption seen as modern. Finally, others still called attention to the way in which transactional sex participated in the broader framework of patron-client relations, which are very important in many African societies.
In addition to the diversity of analytical approaches to transactional sex that we have briefly discussed, there is the problem, as is often the case, of definitional vagueness. An examination of the literature reveals a range of significantly different meanings, even if they are all defined in relation to prostitution. For some, the category refers to the exchange of gifts rather than money. For others, what matters is the rationale on which the exchange is based: at times the transaction can be a gift, at others it can be described as informal, non-professional, and even non-commercial. For others still, the purpose of the category is to spare actors the stigma of prostitution and its derogatory connotations. Within this paradigm, there are even those who see transactional sex as a form of prostitution that social actors have not labeled as such. Thus one reads in a United Nations Population Fund document that “there is a widespread view that occasional engagement in transactional sex … constitutes ‘sex work’” ” or, in an article proposing an estimate of the number of sex workers in several countries, that “[p]opulation surveys were scanned for proportions of adult women having sex in exchange for money or goods.”  Thus it is apparent how the problem of defining prostitution, which we considered above, pervades and haunts the different meanings of sexual transaction.
Whatever debates there may have been about the definition of sexual transactions and the reasons people engage in them—whether they have more to do with “culture,” economic necessity, or social change—it is very striking to observe that until the mid-2000s, these studies dealt exclusively with Africa. Transactional sex was considered peculiarly African, and it was on this basis that is was presumed to explain the epidemic’s exceptional character on this continent. Not until the mid-2000s did the concept begin to travel, first to “poor” countries, where we again see the idea that the decisive reason for engaging in such transactions was economic necessity, then, even more recently, in research undertaken in the “North,” which at last did justice to studies that show that material and economic exchange exists in all sexual, romantic, and intimate relations. 
Are “Concurrent” Relationships More Common in Africa?
In the mid-2000s, a new category seeking to explain the scale of the epidemic in Africa emerged: concurrent sexual partnership. While several definitions of this term have been formulated, the most consensual version, which was proposed by UNAIDS in 2009, defines it as any situation in which sexual intercourse with one partner occurs between two acts of intercourse with another partner. The term refers to the existence, for a given individual at a given time, of several sexual partners, and is to be distinguished in this way from serial partnerships, which refers to sexual relations with one partner at any given moment, but with successive partners over time. A number of different situations are thus designated by the same term. Having a regular partner while having a one-time sexual encounter with another partner; having two regular partners over the long term; and having an irregular but recurrent partner whom one sees while visiting a particular geographical, while also having a long-term partner—all of these are instances of concurrent sexual partnership.
For some scientists and institutions, including UNAIDS, concurrent sexual partnerships have been and remain one of the major forces driving the HIV epidemic in Africa.  Their case rests upon a series of arguments that, when juxtaposed, are presumed to have proven their case. The first argument concerns the statistical correlation between the degree of concurrent partnerships and the HIV rate in various African contexts. The second argument rests on mathematical models developed by researchers, which maintain that concurrent partnerships cause the epidemic to spread more rapidly than other forms of heterosexual partnership. The third and final kind of argument relies on secondary sources—specifically, quantitative and qualitative studies showing that concurrent partnerships are more common in Africa. Their opponents  reply point by point that if correlations can at times be seen between concurrent partnerships and significant levels of the epidemic, they do not as such prove that a causal relationship exists.
They also emphasize the fact that longitudinal studies, measuring both sexual behavior and the HIV rate, are best suited for testing this correlation, and recall that the only two studies of this kind were unable to establish such an association. They also criticize the epidemiological data included in these mathematical models, emphasizing, on the one hand, that their focus on individuals does not allow them to really test the concurrent partnership hypothesis (to do so, data relating both to the sexual behavior of a given individual as well as to that of all his sexual partners would have had to be gathered), and, on the other, that the basic data does not clearly distinguish between concurrent partnership and other high-risk behavior. The mathematical models employed are also subject to debate: their opponents contest the methods used, criticizing them notably for overestimating the number of concurrent partnerships. Finally, the two parties disagree over whether concurrent partnerships are uniquely African. The proponents of this new explanation believe that concurrent partnerships over the long term are indeed a characteristic of sub-Saharan Africa. Yet as their critics rightly object, the few comparative studies that had been conducted on a large scale show that there is no significant difference between sexual behavior observed in Africa and behavior documented on other continents. 
Beyond the Sexual Hypothesis and African Peculiarities
While this essay has confined itself to considering explanations formulated by the “sex science” of AIDS, the major political consequences of these theories deserve to be mentioned. As Guillaume Lachenal writes,  the landscape of the anti-AIDS struggle in Africa would have been very different if others forms and circumstances of HIV transmission had been taken into account. The primacy of the sexual hypothesis has resulted, practically speaking, in a myriad of large-scale prevention programs that have primarily targeted sexual behavior. Africa has witnessed wave after wave of large-scale prevention programs, often designated by the acronym IEC (Information Education Communication), which have had a lasting impact on the African mind and public spaces. These programs willingly embrace individualistic approaches, along the lines of the ABC strategy (“Abstinence, Be Faithful, Use a Condom”). The latest scientific debates, notably the recognition, beginning in the 1990s, of the need for an approach that transcends the individualistic focus and emphasizes the cultural and structural factors shaping sexual behavior, have had little impact on the ground.
This brief overview of the scientific community’s efforts to explain the AIDS epidemic in Africa illustrates the quasi-hegemonic character of the sexual hypothesis. In scientific as well as political circles, it has been maintained, since the 1980s, that the reasons for Africa’s particularly high HIV rates are to be found in African sexual behavior. Whether this situation is attributed to prostitution, sexual transactions, or concurrent partnerships, it is the same idea of Africa’s sexual peculiarities, though presented in different forms, that is invoked. This is troubling when one recalls the history of medicine in Africa and the great interest that medical discourse of the colonial period took in African sexuality.  The current issues are, of course, very different from what they once were: previously, the main concern was to explain racial and sexual difference, whereas now, it is with accounting for the scale of an epidemic caused by a sexually transmittable virus. And while some studies have made their own the idea of Africa’s radical otherness that borders on racism, the idea that Africa exhibits distinctive sexual behavior seems primarily to function as a ready-made explanation, which can be invoked at different times and as a function of the different realities Africans face.
The sexual hypothesis does, of course, have its detractors. Some studies, though they are in the minority among the scientific community and are not heeded by the main anti-AIDS organizations, directly contest it.  These authors call for a return to epidemiological approaches, which are considered traditional outside the realm of HIV/AIDS research, focusing on evaluating factors related to the host (such as the presence of sexually transmittable infections, suppressed immune systems and or malnutrition) in the transmission of infectious agents, and giving greater consideration to other forms of transmission (notably the reuse of needles used for medical care). Among these authors, some go so far as to question the appropriateness of “sexuality” as an object of research for understanding the HIV/AIDS epidemic, and call for a moratorium on research relating to sexuality in Africa