Robert Thornton never wanted to study AIDS. But, as he tells us in the preface of his book on the disease, as an anthropologist working in sub-Saharan Africa he could not avoid the subject. Indeed, from an anthropological perspective, HIV/AIDS is all-encompassing. Leaving a generation of orphans in its wake, it has realigned the structures of kinship. Where children-headed households become the norm, cultural traditions suffer the fate of collective amnesia. AIDS has delimited the sphere of politics; access to treatment takes centre stage in national and international debate. AIDS has devastated economies and introduced new sources of labour, capital and value through new forms of philanthropic, commercial and scientific endeavour. Here AIDS is the substrate of culture, society and the self. To study it requires, therefore, a conceptual framework capable of describing sexual relations on the scale of the collective.

By providing the inductive steps necessary to link personal experience and population status, Unimagined community points us toward such a heuristic. Thornton begins with a straightforward statement of purpose. His intent, he tells us, is ‘to offer an anthropological or “ecological” approach to the study of acquired immunodefinicency syndrome (AIDS)’ (p. 1). More than a register of everyday life in a time of crisis, Thornton’s ethnography re-conceptualizes AIDS transmission, and, in so doing, shows that the public health community has gotten the epidemic wrong.

To make this point, Unimagined community takes as its case studies Uganda and South Africa; two countries whose experiences of the epidemic are oppositional. The site of some of the earliest HIV/AIDS cases in Africa, and at one point considered the most affected place in the world, Uganda is currently seen as model for HIV containment. On the other hand, South Africa, with the strongest public health system on the continent, remains crippled under the weight of an epidemic which shows no sign of waning.

Why this difference? Reports released by US Agency for International Development (USAID, 1998, 2002, 2006) take Uganda as proof of the effectiveness of ABC measures, a public health campaign that promotes Abstinence, Being Faithful and (as a last resort) Condoms. A number of studies have complicated, and in some cases, contradicted these findings, demonstrating that Ugandans have not, in fact, changed their sexual practices—at least to any significant degree (e.g. Epstein, Reference Epstein2005; Parkhurst and Lush, Reference Nesje, Millao and Mpaka2004). The debate turns on the importance of condoms, and the political stakes are clear. Foreign aid under the Bush administration had an explicit ideological bent; over the past decade, the USA’s international public health policies have been strongly inflected by evangelical religious groups.Footnote 1 Celebrating the success of ABC campaigns, then, is as much a political act as it is an outcome assessment. That refraining from sexual activity will prevent the spread of AIDS cannot be disputed. However, for its critics, the moral message of ABC loses sight of the socioeconomic realities that shape peoples choices. Funds should remain focused, they argue, on condom distribution and on strategies that encourage local responses to AIDS, such as comprehensive sexual education.

Though he shares these critical views of ABC, Thornton’s take on the Uganda story reframes the terms of the debate. Media campaigns, education programmes, and Knowledge Attitude Practice (KAP) surveys have been the primary tools of public health in tackling the epidemic.Footnote 2 As the case of South Africa proves, these measures are insufficient. Put simply, individual health-seeking behaviours tell us very little about the epidemiology of the virus. ‘Changing prevalence’, Thornton argues, ‘is a property of the social whole’ (p. 79). To bring the ‘social whole’ into view, Unimagined community investigates the ‘sexual network’, the links between groups of people that either promote or inhibit transmission. In the context of world health policy directed toward risk and awareness, Thornton’s focus on the social structure of sex is quite radical. Rather than a decision between two individuals, the sexual activity described in Unimagined community is a relational dynamic enmeshed in other forms of cultural, political and economic connection. AIDS prevention must, therefore, be approached systemically, in the way that ecologists address climate change or economists fiscal crises.

To map the distinct configurations of the Ugandan and South African networks, Thornton draws from a range of empirical sources. Much of the book is devoted to providing a dense political history of the two regions. We learn that Uganda was a British protectorate and never a colony. Kingdoms, tribal associations, marital practices and land tenure remained stable for centuries. These isolated yet strongly linked groups were reinforced by the political unrest of post-independence. Idi Amin’s dictatorship had the effect of bringing development to a halt; limited job opportunities kept people away from urban centres, and violent attacks by the National Liberation Army (LRA) precluded most cross-country travel. After Amin’s exile, the Museveni government, in an effort to stabilize the political landscape, restored kingships. In contrast, Thornton argues, over the last two centuries, South African society has been relatively integrated. Despite the segregation enforced during Apartheid (and indeed, as Thornton points out, because of it), most South Africans are highly mobile and live in multi-ethnic communities. The struggle against Apartheid, moreover, fostered a commitment to equality and autonomy: collective identity in South Africa is a product of political action and not of geography. Economic developments and improvements in infrastructure have enhanced the dense and interconnected nature of that social landscape.

There are, then, some obvious explanations for the difference in the efficiency of disease transmission between the two countries: clearly, labour migrations, trucking routes and urban sprawl amplify the number and diversity of sexual encounters. Thornton’s analysis, however, involves more than merely calculating the number of connections in a sexual network. His first two chapters deal with epidemiological trends in quantitative detail. Much of epidemiological research is based on the Gaussian distribution curve, a statistical model which assumes that infection is spread randomly through a population and will, after reaching a peak incidence, drop to zero. But unlike most viral and bacterial epidemics, HIV does not progress exponentially. Transmission is irregularly distributed through social structures and is highly periodic (an individual’s viral load decreases over time); rates of HIV prevalence and incidence depend, therefore, not only on the number of connections but also on the ways in which those connections are made. Because South Africa’s networks involve many and multiple points of connection across wide spaces, transmission functions the way the Internet does, becoming denser as the network grows. Alternatively, in Uganda, connections within groups are greater than those between them. Thus, while infection at the ‘nodes’ can be quite high, the cross-links in Uganda’s sexual network are minimal. This ‘lumpy’ or ‘fractal’ character means that while incidence will shoot up during the start of the epidemic, clusters of sexual activity are easier to isolate. The upshot is that while changes in individual behaviour would do little to reduce transmission in South Africa, such modification among Ugandans would impact the spread of the AIDS virus.

There is a reason why Claude Lévi-Strauss insisted that observable data and structural analysis be treated separately: belonging to different orders of description, individual attitudes can neither prove nor predict collective experience (Lévi-Strauss, 1953). Thornton demonstrates the problems of that elision by describing some of the explanatory acrobatics scientists have used to connect observed data with standard epidemiological models. These conjectures can have disastrous consequences: it was the ill-fit between the Gaussian curve and disease prevalence that ‘proved’ that AIDS was a syndrome caused by lifestyle (like obesity) and had nothing to do with the HIV virus. But failures in understanding epidemics are not always the fault of the model. Tim Allen has, for example, drawn on ethnography to argue against using Uganda’s experience with HIV/AIDS to produce insights about the virus in general (2006: 1–22). Drawing on his work among internally displaced persons, Allen questions the reliability of data which shows that AIDS in Uganda is, indeed, declining. He reminds us that, for the most part, public health data across Africa is derived from antenatal surveillance. This method of data collection poses several problems: HIV positive women are less likely to become pregnant, surveillance is more concentrated on peri-urban and urban areas and it excludes those infected by non-sexual practices. Myths about decline reinforce those about how HIV/AIDS is spread. Prevention and treatment policies based on selective statistical surveys can only prove ineffective and potentially counterproductive.

But, regardless of whether Uganda is a success story in the international struggle against AIDS, Thornton’s efforts to focus global public health on sexual networks has profound policy implications which, I believe, anthropologists such as Allen would support. Thornton’s attention to sex as a social relation embedded in other forms of exchange and connection privileges the everyday relevance of local norms rather than the moral message of ABC campaigns. The central chapters of Unimagined community look at the meanings of sex in South African and Ugandan contexts. Thornton describes how, in South Africa, the significance of sex is best understood in terms of vital flows of substance that link both partners and generations. Blood, heated by sexual activity, provides the physical substance that connects families, one’s past to one’s present. From this perspective, condoms create blockage; the risks of sexual contact do not override its vital effects. In Uganda, sexual activity is fundamental to the social order, but its significance is less cosmological and more mundane. Sex and sexuality are part of everyday conversation; daily magazines are devoted exclusively to personal sex stories and new sex games. Before marriage, girls are traditionally taught by older women how best to please their husbands. Many Ugandans, Thornton tells us, say that life without sex is scarcely worth living. Both cases prove that the drivers of sex lives are far more complex than questions of behaviour. ‘Sex’, Thornton reminds us, is more than an activity between partners, it ‘is an act through which males become men and females become women’ (p. 217). Like any other social relation, sexual activity is shaped by conventional mores (the exchange of fluids), modern ideals (gender equality), material resources (kinship and property) and institutional alignments (religious movements, government legislation and NGO interventions).

Here let me return to the question of method. No matter how rich its ethnographic data, the power of Unimagined communities resides not in the subtleness of its ethnography, but in the way it conceptualizes sex. Beyond its public health implications, this theoretical turn presents an interesting challenge for anthropologists. Thornton’s approach runs against the grain of contemporary anthropological research, in particular the sort typically conducted in healthcare settings. The corroborative potential of ethnographic data for public policy has generally been understood in terms of the context it adds to health practice. Experts in local attitudes and cultural beliefs, anthropologists define the logics that connect health understandings to health-seeking behaviours. As such, they facilitate the translation between policy and effective action (e.g. Hahn, Reference Hahn1999; Porter, Reference Parkhurst and Lush2006). The anthropological context can also be scaled up. Paul Farmer’s work on HIV/AIDS in Haiti, for example, shifts public health concerns from the cultural basis of health-practices to the ways in which these are shaped by political economies. Chronicling the lives of this ‘high-risk’ population, Farmer reveals how the increasing incidence of the disease cannot be reduced to promiscuous sexual practices. Instead, the choices made by the poor Haitian women he studied—many of whom had only one partner—were determined by their marginalized circumstances (Farmer, Reference Farmer2002). Farmer terms these dynamics of exclusion and inequity ‘structural violence’. The ethnographic task is, again, one of translation. However, rather than being between public health knowledge and folk understandings of disease, here the anthropologist is positioned between economic processes and the individuals that embody them. When it comes to health, Farmer argues, ‘social context is not merely local, nor are standards of care … the context is trans-national’ (2002: 1266).

In contrast, Thornton pursues AIDS, à la Durkheim, as a social fact. At a point at which anthropologists have supposedly left the study of ‘peoples’ behind—according to George Marcus, ‘ethnography is predicated upon attention to the everyday, an intimate knowledge of face-to-face communities and groups’ (1998: 82)—Thornton returns us to the problem of aggregation and the constitution of the community. This concern represents a considerable departure from a bi-focal account of global forces and local realities. Thornton does not rely on thick descriptions of the history of the Lord’s Resistance Army’s movements through Uganda, or of the detailed accounts of the parallels between the ANC’s Freedom Charter and the HIV activist AIDS Charter in South Africa to explain the transmission of the virus. Like the roads connecting cities, WHO funds, the distribution of condoms, these contextual particulars all form part of a network of relations through which the virus moves. Thus, to contain the epidemic, health interventions require commitment that extends from the personal to the collective. Preventing AIDS is a matter of inspiring the kind of civic concern that can call a public into being.

At the heart of Unimagined community is a fascinating argument about how political action can both advance and retard those efforts. By turning to the shape of the sexual network, Thornton asks us to think about the relationship between AIDS and national identity. Because of its early emergence in Uganda, the virus was defined locally (described as siliimu or slim, referring to the wasting away of its victims) before it was reclassified as AIDS. This indigenous categorization made AIDS a part of family history and a feature of community problems: deaths occurred among those one knew, children were taken in by neighbours and graves were built within compounds. In the wake of Idi Amin, AIDS was associated with the country’s wider social, political and economic troubles. Thornton suggests that the linkage of the AIDS struggle with that of national development has located AIDS prevention within a discourse of social progress.

We are all familiar with Thabo Mbeki’s denial of AIDS and the consequences of his aetiological theories for the spread of the disease, but Thornton enhances our understanding by demonstrating that Mbeki’s failures were compounded by activists critical of his policies. The fight for AIDS recognition, he argues, was inflected with the same moral outrage expressed against Apartheid. Rather than encouraging widespread mobilization, HIV activism re-articulated familiar tensions and entrenched the exceptional status of the infected (p. 158). In an ironic turn, the price of politicization—insisting that AIDS was not a local disease but a transnational epidemic—was to dislocate the problem from the quotidian. AIDS became an issue for human rights, to be solved on a global scale not within local communities.

The task, then, according to Thornton, is to imagine a community that can bridge domestic and social space, collapse private morality and public good, and ‘unite the larger scale political order with the small scale personal experience of sex, illness and death’ (p. 192). This requires an imaginative leap, a singular vision of the ‘social’. As international research institutions, non-governmental organizations, and aid-based charities perform the vital functions formally ascribed to the state, the public loses its representational character. So what then is the locus of the collective identity of the AIDS community? Thornton’s allusion to Benedict Anderson’s famous study on the sensibilities of nationalism, Imagined communities (1983), suggests that one place to look for collective identity would be the state. This impulse, of course, runs counter to contemporary tendencies to situate social solidarity within inter- and transnational processes of governance. But Thornton understands how the state can still provide the conceptual links that yoke citizen health and social welfare. As the case of Uganda illustrates, civic engagement can have a major impact on epidemiological dynamics if it is linked to other forms of social progress. Rather than a pandemic, a dark scourge on the political consciousness, a spectre of development’s failures, AIDS became a concern of and for a Ugandan public. In Uganda AIDS was, to use Anderson’s words on nationalism, ‘imagined, because the members … will never know most of their fellow-members, meet them, or even hear of them, yet in the minds of each lives the image of their communion’ (1983: 15). Ultimately, Unimagined community shows how the effectiveness of public health interventions requires a highly nuanced idea about the political character of the public whose health they aim to improve.