Village Gone Viral
Understanding the Spread of Policy Models in a Digital Age
Marit Tolo Østebø



I AM STANDING IN a clean and sparsely furnished kitchen in a rural village in Ethiopia. The room is crammed with local and expatriate development experts who all have come to see the fruits of a gender-focused development project undertaken by a local, faith-based non-governmental organization (NGO). As I take in the aromatic, smoky blend of coffee and spices, I turn my attention to the kitchen layout. A locally made, wooden, working table stands in the middle of the room. Its rough surface is a stark contrast to American shiny countertops. There are no fancy advanced electrical appliances in sight, but a relatively new invention—an energy efficient clay stove—has replaced the traditional three-stone fireplace commonly found in rural kitchens. The stove, which the Ethiopian government has singled out as a key indicator of a healthy home environment, is a testimony of a so-called model family: a household that has successfully implemented the comprehensive package of health interventions that are at the core of the country’s highly acclaimed health extension program, and one from which other households are expected to learn.

As in many other places in the world, kitchens in Ethiopia are female-dominated spaces, places where men barely set foot. Yet, this morning and this kitchen are different. As our attention is drawn to the far-right corner of the room, to a tall, slim man in his forties, we are presented with a story that challenges the traditional gendered ordering of an Ethiopian kitchen. Dressed in his finest suit, slightly bent over an empty cutting board and with a knife in his right hand, the man proudly declares, “After my wife came back from a visit to the Awra Amba community where she witnessed how they practiced gender equality, things have changed. I now help her, and I work in the kitchen.” To demonstrate the change in the intra-household division of labor, he starts moving the knife, explaining that he will assist his wife with chopping onions. But there are no onions on the cutting board, and there is no actual chopping. Only clumsy, pretend movements. “I also help my wife baking injera,” he continues as he turns to the stove. Made from teff, a gluten-free grain indigenous to Ethiopia, injera is a large, sour, spongy pancake that is a staple food in many parts of the country. To bake it requires carefully crafted skills. The pouring of the fermented batter onto a big, clay plate positioned over open fire demands precise, controlled circular movements. Holding an empty jug over the cold plate, the man continues his performance. There are no circular movements. There is no fire burning.

I witnessed this episode while visiting a rural community in West Wellega Zone of the Oromia National Regional State (hereafter, Oromia) of Ethiopia in 2010. With the aim of exploring how Norwegian gender equality policies travel and intersect with different actors in the development field,1 I had joined a delegation of international and national development experts involved in Women Empowerment and Gender Equality (WEGE), a multi-country program implemented by Digni, a Norwegian faith-based umbrella organization.2 Representatives from Digni, two of their member organizations, and their respective African partners traveled together to visit one of the community development projects involved in the program. The four-day trip included an experience-sharing workshop3 and a visit to one of the project’s target communities. Built into the visit was a meeting with a local WEGE committee—formed as part of the project—and a home visit to one of its members. This is where the incident described above took place.

Awra Amba, a small, rural village located close to the city of Bahir Dar in the northern part of Ethiopia, was hundreds of miles away. Yet during our visit with the local WEGE committee, Awra Amba emerged as a central theme; an inspiration for the radical changes the project allegedly had brought to this rural community. In the meeting that followed the demonstration in the kitchen, members of the WEGE committee who, as part of the project’s activities, had been given the chance to visit Awra Amba described it as “a place where there is no gender division of labor.” As a researcher interested in exploring gender relations and conceptions of gender equality, this depiction surely caught my attention. In fact, after hearing about Awra Amba, I considered including that village in the research I was conducting at the time. But as I struggled to manage a rather complex research project where, in addition to following the organizational chain of command of two Norwegian development organizations, I was also conducting research in two rural communities in Oromia, I quickly dropped the idea. To add yet another research site, in a region located more than three days’ travel from the areas where I already was working, did not make sense then.

As I researched how different actors who were involved in or impacted by Norwegian-funded gender-focused projects in Ethiopia conceptually and practically translated the concept of gender equality, references to Awra Amba occasionally emerged. Yet, it was not until five years later that I had the opportunity to visit Awra Amba and conduct research as part of a larger project on women’s rights and plural legal systems.4 When I finally arrived in the village in 2015, I quickly realized the preeminent position the community has gained as a model for gender equality and sustainable development in Ethiopia.

Awra Amba became publicly known when Ethiopian Television (ETV) aired a documentary about the village in 2001. The program told the story of a self-sustaining and gender-equal community, where women plowed, men worked in the kitchen, and so-called harmful traditional practices did not exist. The narrative radically challenged prevailing images of Ethiopia as a gender-conservative and aid-dependent place, capturing the attention of numerous governmental officials, gender and development experts, human rights activists, tourists, and educators. Within a short time, Awra Amba became a national model village. Known as the place “where gender equality is real,”5 it attracted policy makers and development experts from NGOs and the Ethiopian government. With an interest in identifying best practices that could be replicated and scaled up in other places, they flocked to the village, eager to learn from its success.

While part of the story I tell in this book takes place in Awra Amba, this is not an ethnography of life in an African village. Reflecting the increasingly globalized and digitalized world we live in, this is a story about how a small, rural village, founded and led by an illiterate man, has become a policy model and gone viral. After ETV reported on Awra Amba and the community then became an official tourist destination, its story has been retold countless times. The main narrators are a few select members of the community who, on a daily basis, share the community’s history and way of life with visitors to the village. But a wide range of national and international actors has also contributed to the spread of the Awra Amba story. In Ethiopia, Awra Amba is ever present in the national media, even giving name to a newspaper, Awramba Times, and inspiring Ethiopian popular music (Ahmed Teshome, 2006).6 When I tell people I meet that I do research in Awra Amba, whether taxi drivers in the capital of Addis Ababa or camel herders in the eastern lowlands of the country, I seldom have to explain what kind of place it is. Everyone has heard about the village where women plow and men make injera.

But Awra Amba is also well known far beyond Ethiopia’s borders. The community has been featured on international news channels such as BBC World News and France 24, in various European magazines, on travel blogs, and on websites of feminist activists, NGOs, and human rights organizations. Several documentaries about the community are available on YouTube and Vimeo, and Awra Amba features in a photo essay in an African studies textbook (Salem Mekuria 2017). The Awra Amba story is, moreover, the flagship of Lyfta, an award-winning educational technology (EdTech) company with offices in Finland and the UK. Lyfta is a commercial learning platform that contains 360-degree interactive digital stories and short documentary films aimed at students in elementary and secondary schools in the United States and Europe.

In this book, I use the Awra Amba case as a point of departure to engage in a broader empirical and theoretical exploration of the role and politics of models and model making in an increasingly digital and transnational policy world. While models in various forms and throughout history have been used to explain or predict “real-world” phenomena and to inform policy and govern human behavior, the bourgeoning of models and modeling practices in our contemporary world suggests that we live in a modeloscene.7 With an empirical and theoretical focus on traveling models—policy models that become “viral” and spread widely across different localities through various vectors, or carriers, ranging from NGOs and multilateral organizations to the internet—Village Gone Viral addresses three sets of questions. First, I am interested in exploring policy models from an ontological perspective. What constitutes models within the policy world and how do they come into being? The second set of questions pertains to the traveling nature of policy models. What characterizes the models that go viral? Why do some models gain followers while others do not? In other words, what facilitates and fuels a model’s virality? Third, I explore ethnographically the hidden dimensions of policy circulation and the effects of the models’ status on the models themselves. What happens to the original policy model—ideological or place-based—once it becomes a traveling model? Who benefits from the model’s popularity? To what extent do models and modeling practices rely on, produce, and exacerbate unequal power relations?

Some of the questions I ask have been discussed by other scholars. There are ontological discussions about models in anthropology (Geertz 1973, 2007; Handelman 1998; Clarke 1972), philosophy of science (Toon 2010; Arnon 2012), and economics (Morgan 2012; Morgan and Morrison 1999). These theoretical conversations have, however, entered the field of policy and development studies to a limited extent. Anthropologists and scholars within the interdisciplinary field of critical policy studies have also sought to make sense of the increased transnational circulation of policy models and ideas by introducing analytical concepts such as policy mobility (Peck and Theodore 2010; Temenos and McCann 2012) and traveling models (Olivier de Sardan, Diarra, and Moha 2017; Rottenburg 2009). The effects of the models’ status on the models themselves—a topic I explore in-depth in Chapter 8—have received limited scholarly attention.

The empirical material I present and the theoretical approach I adopt provide new and valuable insights that are crucial to an understanding of why certain policies, ideas, or innovations spread, while others do not. Inspired by assemblage thinking derived from the works of Gilles Deleuze and Félix Guattari (1987), Tony D. Sampson’s (2012) resuscitation of Gabriel Tarde’s social epidemiology,8 and lessons from virology, my overarching argument is as follows: A traveling model can best be understood as a viral assemblage—here, dually defined as a messy, fluid, socio-technical process and a constellation of actors, things, unpredictable events, and relations that have contagious and affective qualities. I approach the traveling model and policy mobility through the theoretical lens of viral assemblage to highlight three important lessons that can be learned from the Awra Amba case. First, as an assemblage consisting of heterogeneous elements, a model is a unique, historic entity. A model is, in other words, not a neutral, universal, or static self-contained entity that exists independently of its historical, political, and economic context. Nor is it, as often assumed in the policy world, an example of success that can be scaled up and implemented in new contexts. A model is always in process of becoming, constantly being deterritorialized and reterritorialized in time and space. Since a model is the result of a chaotic process where heterogeneous elements come together—a cultural construction—it demands contextualization.

Second, just as for a virus, the model’s travel and its contagious capacity are conditioned on vectors, hospitable environments, and receptive host cells. This means that the model does not spread unconditionally, but that the model’s virality, to a certain extent, is restricted. For a model to spread, it has to find the right vector. It needs to “click” with a cell that has the right receptor. There needs to be an association or element of recognition and interaction between the model and the entities that may connect with it. These associations tend to be affective in character.

Third, similar to viruses, models have both destructive and constructive capacities. While many viruses are bad and generate much fear, others may be beneficial and even essential for the survival of their hosts. The designation of the assemblage as viral allows for a greater recognition of the model’s affective qualities—of its ability to produce affect and be affected; to hold promises and threats; to set into motion fear and joy; to be emancipatory and oppressive. Looking at models through the lens of viral assemblage implies attention to how affect is both constitutive of the model and derived from it. The affective dimensions of model making and circulation are a crucial aspect that the existing literature on policy mobility and traveling models has overlooked.

In the next section, I give a brief synopsis of how models, in various forms, are increasingly used as policy instruments and in efforts aimed at generating social and behavioral change. I then detail my overall theoretical framework. Finally, I move to a methodological section where, in addition to situating myself in the field, I discuss how the concept of viral assemblage can also help us make sense of the methods, processes, and products that we, as anthropologists, engage in and create.

Models in a Transnational Policy World

Within an increasingly transnational policy world, much energy is devoted to the production and circulation of “the right policy models” (Mosse 2004, 640). A wide range of models exists: models for microfinance, disease prevention, health care delivery, maternal health, disaster management, and participatory development, just to mention a few. Many of these are pen-and-paper models that are rather methodological and prescriptive in nature. As programmatic schemes, they serve as road maps for a desired process or behavior, offering an explanation or procedure for how to implement a particular intervention successfully. While pen-and-paper models appear to be most prevalent within the policy world, other forms of models also attract considerable attention, such as epidemiological models, human role models, scale models, model cities, or, as in the Awra Amba case—model villages.

The increased use and circulation of models within the policy world and, more generally, as a tool for social and behavioral change, can be explained with reference to both technology and ideology. New computing technologies and access to big data have, for example, led to an upsurge in mathematical models used to predict, explain, and influence human behavior and social systems.9 The COVID-19 pandemic has clearly and dramatically revealed this model reality. Models are no longer confined to policy experts or scientists but have become unprecedentedly public and popularized. Graphic models depicting and comparing the spread and consequences of COVID-19 in one geographical area versus another are on the front page of major news outlets, and numerous newspaper articles and op-eds have been written about predictive models. Models increasingly govern our lives, informing us of what the future may look like if we fail to wash our hands or comply with physical distancing rules.

Inspired by economic psychology, behavioral economics, and business studies, many of these models assume that human behavior is a matter of rational choice; that a rational person will make decisions based on calculations of cost-effectiveness (Kleinman 2012). The quest for models, particularly those commonly referred to as best practices, can also be linked to the rise of new public management and the business-oriented, results-driven, and evidence-based ethos that have come to influence contemporary politics and policy making (Biehl and Petryna 2013; Storeng and Behague 2014; Shore 2008). Within the global policy world, this has perhaps most clearly been crystalized in the establishment of the United Nations (UN) Millennium Development Goals (MDGs) in 2000, which now has transitioned into the Sustainable Development Goals (SDGs). Underpinned by a “linear cause-effect model of aid practice” (Eyben 2010, 2), the MDGs implied a strong focus on reaching preset measurable goals or benchmarks and was followed by the adaptation of results-based management by a number of major policy actors, including the UN (Bester 2012).

A widely held assumption about models in the policy field is that they are examples of success or best practices that can be transported, emulated, scaled up, and implemented to bring about a desired change across global spaces. A growing body of literature in anthropology, geography, and policy studies has challenged these presumptions (Behrends, Park, and Rottenburg 2014b; Gonzalez 2010; McCann 2011; Olivier de Sardan, Diarra, and Moha 2017; Rottenburg 2009). Emphasizing models as dynamic and context-dependent entities, this recent scholarship sheds light on how a wide range of actors—who are differently situated in power structures and whose social, cultural, political, and economic interests often diverge—negotiates, contests, appropriates, translates, and changes the models they encounter. In his book Far-Fetched Facts: A Parable of Development Aid, the German anthropologist and Africanist Richard Rottenburg (2009) introduced the concept of traveling models to analyze these processes. The concept has been further developed and applied by Jean-Pierre Olivier de Sardan, Aïssa Diarra, and Mahaman Moha (2017, 74). They define a traveling model as

any standardized institutional intervention, whatever the scale or field (a public policy, a programme, a reform, a project, a protocol), with a view to producing any social change, through changes in the behavior of one or more categories of actors, and based on a ‘mechanism’ and ‘devices’ supposed to have intrinsic properties allowing this change to be induced in various implementation contexts.

The way these scholars apply the concept of the model is, however, skewed toward intentionally constructed prescriptive models—what we could term “models for” (Geertz 1973, 93ff.). This is reflected in Olivier de Sardan’s, Diarra’s, and Moha’s focus on models as standardized, institutionalized interventions, exemplified in their analysis of widely used health models such as the partogram, focused antenatal care, and performance-based financing. Behrends, Park, and Rottenburg (2014b, 1) reflect a similar focus, defining a traveling model as

an analytical representation of particular aspects of reality created as an apparatus or protocol for interventions in order to shape this reality for certain purposes. Models—and the ideas about reality inscribed into them—always come objectified and combined with material technologies to put them into practice and to transfer them as blueprints to new sites.

While I am inspired by and partly draw on the works of Rottenburg and Olivier de Sardan, Aïssa, and Moha, they neglect two important aspects. First, many models within the policy world circulate because they are perceived to be ideal examples of a particular phenomenon, policy, or desired behavior. While these models also have a prescriptive purpose, they tend to serve, to a much larger degree, as representative models, that is, as “models of” (Geertz 1973, 93 ff.). In addition to the human role model, I am thinking of projects, villages, districts, or cities that have become anointed, sanctioned, or promoted as models. Second, a one-sided focus on traveling models as standardized interventions, such as protocols and pen-and-paper models, fails to take into account that many models exist in or are closely associated with a particular place, or that they are constituted of and embodied in human actors. These models are much more than things or symbolic representations of the lived-in world. They are part of a living social world, made up of human actors who, as sentient beings, have the capacity to feel and experience, to be affected, and to produce affect. This has both empirical and theoretical implications. The effects that becoming and being a model have on these models are quite different from those of standardized interventions.

Viral Assemblage

My approach to exploring how models and policy ideas travel is best captured in the term going viral. While this expression is commonly associated with the rapid, global circulation of a story, image, or video via social media or the internet, my use of and reference to going viral moves beyond this narrow application. I have here found it useful to go back to the virus as a biological phenomenon. To think with virology can help us understand how social theorists such as Deleuze and Guattari use the term assemblage.

To use epidemiological metaphors and biological analogies to analyze social phenomena is a “delicate and ambiguous position” (Serres 2007, 43). As Sampson (2012) has argued, such an approach is potentially limiting and can block our conceptual flexibility. This is why Sampson (3) has developed an approach to social contagion “intended to probe outside the generality of metaphors and analogies.” I contend however, that he too quickly dismisses epidemiological metaphors and analogies as analytical tools, and that his dismissiveness reflects unfounded bias toward and assumptions of viruses as solely pathological, destructive, lethal, and memetic entities. As I hope to show, perspectives from epidemiology and virology can be used fruitfully to make sense of social phenomena. Such an approach does not imply that I “valorize(s) scientific practice as the source of rationality,” as Steven Brown (2002, 9) has argued. I do not treat the virus as a blueprint or model for social processes, nor do I suggest that virality in the social world can be reduced to what happens in the biological world.10 Inspired by Michel Serres (2007), who emphasizes the importance of not favoring one form of knowledge over another and insists on the middle position as a space for transformation, innovation, and meaning creation, I see the value of crossing disciplinary boundaries. In this case, this means that I draw, eclectically but reflexively, on epidemiological analogies from virology to make sense of the social.

It is chiefly the destructive capacity of a virus and its potential to spread rapidly and uncontrollably that have been used to describe the flow of information and ideas in our digital media age. This reflects a simplified and reductionist adoption of the virus metaphor. First, a biological virus is not a self-contained, evolutionary, imitating unit that spreads unconditionally. A virus is typically quite simple, with a single strand of ribonucleic acid, a protein coat that protects this genetic material, and a fatty envelope that enables it to bind to its host. It does not have a capacity to replicate on its own, with just its own components. In order for a virion—the non-living phase of the virus—to become an infectious virus, it must establish a set of relations with a susceptible and permissive cell. It has to find the right cell—a cell that has a functional and matching receptor that allows the virus to enter. Once the virus has entered the cell, it has to assemble using preformed, heterogeneous components from the cell and its interior, in effect hijacking the host’s more complex and abundant cellular machinery and materials. It is a virus’s ability to establish relations that allows it to reproduce, making it into a living entity. The fact that the virus is dependent on a living host has led biologists to conclude that it is the “ultimate parasite” (Claverie and Abergel 2012, 187).

Second, a virus’s ability to proliferate and replicate is dependent on a variety of social, political, and biological factors. Some environments are, for example, more welcoming than others; some hosts are more susceptible to a particular virus than others. Let us take the human immunodeficiency virus (HIV) as an example. The World Health Organization (WHO) estimates that since the start of the HIV pandemic in the 1980s, more than 70 million people have been infected with the virus, and more than 35 million people have died. The burden of the epidemic varies substantially, however. Some regions of the world, such as Sub-Saharan Africa, have much higher infection rates than others. There are numerous social, political, biological, and behavioral explanations for these variations. The presence of another untreated sexually transmitted disease increases the likelihood of HIV infection, for example, as does high-risk sexual behavior such as having unprotected sex with multiple sexual partners. More importantly, structural violence—such as extreme poverty and poorly functioning health systems—impacts the spread of the disease. The spread of the virus is moreover conditioned on the mobility and on the networks of actors and vectors that serve as vehicles for its transmission.11 The role mobility plays in terms of fueling and facilitating virus transmission explains why quarantine, isolation, and restriction on movement are extremely important measures during epidemics that are caused by lethal viruses. The spread of a virus is, in other words, a highly social, material, and relational process.

Finally, while viruses, due to their association with illness and deadly epidemics, often have a bad reputation and generate fear, it is important to keep in mind that some viruses are beneficial and mutualistic. For example, some viruses “are essential for the survival of their hosts, others give their hosts a fighting edge in the competitive world of nature and some have been associated with their hosts for so long that the line between host and virus has become blurred” (Roossinck 2011, 99). Infection with one type of virus may also reduce the mortality associated with another type of virus (Tillmann et al. 2001). In other words, not all viruses are bad.

While I have found it useful to draw lessons from virology in my analysis of Awra Amba as a traveling model, my theoretical approach is, as already indicated, best captured through the notion of viral assemblage. This notion offers a flexible, critical, and reflexive way of understanding the virus-like behavior of the Awra Amba model, addressing some of the potential dangers of using a biological metaphor. To avoid these dangers, Deleuze and Guattari use the concept of assemblage in multiple ways, making it—perhaps on purpose—a “traveling” concept, difficult to grasp once and for all. Manuel DeLanda (2006, 2016) has made an effort to bring together, in a cogent theoretical framework, the many ideas and definitions that Deleuze and Guattari introduce. His work provides the most comprehensive account of assemblage theory. In what follows, I draw on DeLanda’s work to clarify Deleuze’s and Guattari’s assemblage thinking, especially as it applies to Awra Amba.

First, since an assemblage is a historic, unique, individual entity—“an individual person, an individual community, an individual organization, an individual city” (DeLanda 2016, 19)—it always has a signifying, proper name. For example, while a model—which is a general category, a noun that can be applied to a range of entities—is not an assemblage, the Awra Amba model village is one.12 By approaching Awra Amba as an assemblage, I make it clear that I do not treat models as purely abstract, disembodied constructs. Instead, my approach implies a recognition of models as embodied and material entities and an analytical emphasis on the importance of contextualizing policy models in the particular political, economic, and cultural context in which they are situated.

Second, because “assemblages are always composed of heterogonous components” (DeLanda 2016, 20), they are not uniform in origin or in nature. They can include persons, material and symbolic artifacts, architecture, norms, natural resources, infrastructure such as roads and electricity, and tools and machines. This heterogeneity calls for careful analysis of the various components—covert and overt—that emerge in the Awra Amba assemblage.

Third, assemblages are often components of larger assemblages. For example, we can think of Awra Amba—the village—as an assemblage. However, there are additional assemblages within Awra Amba, such as the Awra Amba High School or the energy efficient stove used and promoted by the village as the Awra Amba Stove. This complexity requires that we carefully explore the interconnectivity between the different scales, or levels, in larger assemblages.

Fourth, the interactions between the different parts are what give rise to a particular assemblage. These relations are the emergent capacities of the assemblage. As with a virus, an assemblage is not a given, static whole; it changes and mutates. It is a multiplicity, where “what counts are not the terms or the elements, but what is ‘between’ them, the in-between, a set of relations that are inseparable from each other” (Deleuze and Parnet 2002, viii). In other words, adopting a Deleuzian-Guattarian approach to the study of models going viral implies a focus on relationality and mobility and on how models as assemblages are always becoming. It requires that we carefully examine the nature of the lines and flows that makes up the multiplicity, and how they “become entangled, connect, bifurcate, avoid or fail to avoid the foci” (viii).

By combining the concept of virality with assemblage thinking, I underscore the fruitful connection between Tarde’s way of making sense of the social and Deleuze’s and Guattari’s philosophy.13 Tarde’s approach is rightly called social epidemiology because he drew inspiration from medicine and psychology, borrowing the notion of contagion to characterize the micro-dynamics of diffusion and relationality through which the social is constituted. Influencing not only Deleuze and Guattari but also Bruno Latour’s actor-network theory, Tarde challenges, on the one hand, the idea of the “self-contained individual” (Sampson 2012, 9) whose agency is determined by conscious reasoning. On the other hand, Tarde rejects Émile Durkheim’s idea that social institutions are fixed, autonomous entities that produce “powerful downward pressures” (18) and determine the actions of the individual. For Tarde, biological, psychosocial, and social phenomena are inseparable and mutually constitutive. Characterized by a clear break with the nature-culture divide, his account of the social is “not concerned with the individual person or its collective representation, but rather with the networks or relational flows that spread out and connect everything to everything else” (7).

My use of viral assemblage as an analytical term does more than acknowledge lines and connections between theories, however. First, while the concept of assemblage often is used to describe the coming together of heterogeneous elements into a unique historical and individual entity, the designation of the assemblage as viral implies a simultaneous emphasis on outward mobility—on spread, dispersion, and contagion. This is not an alternative or new interpretation of assemblage. Rather, it is a way of recognizing and making clear that the instability, fluidity, and complex interplay of multiple variables that characterize epidemiological and viral transmission processes are integral to the assemblage. While viral assemblage may be particularly applicable and useful for analyzing and understanding entities or phenomena that are traveling in nature, the concept is not meant to denote particular kinds of assemblages. Rather, viral assemblage allows for a more precise description of the fluid, unpredictable, and traveling character of all assemblages. As DeLanda (2016, 7) concludes, “Assemblages are everywhere, multiplying in every direction, some more viscous and changing at slower speeds, some more fluid and impermanent, coming into being almost as fast as they disappear.” The simultaneous intrinsic and extrinsic nature of assemblage that the term viral assemblage denotes is also evident in one of the first definitions that Deleuze (Deleuze and Parnet 2002, 69) proffers in Dialogues II.

[An assemblage is] a multiplicity which is made up of many heterogeneous terms and which establishes liaisons, relations between them across ages, sexes and reigns—different natures. Thus, the assemblage’s only unity is that of co-functioning: it is a symbiosis, a ‘sympathy’. It is never filiations which are important but alliances, alloys: these are not successions, lines of descent, but contagions, epidemics, the wind.

There is a clear play with epidemiological language in this definition. As Martin Müller and Carolin Schurr (2016) have argued, Deleuze’s use of the terms “contagions, epidemics, the wind” indicates the transient, unstable, and unpredictable nature of the assemblage. To approach Awra Amba as a traveling model through the lens of viral assemblage thus reminds us that models and model-making processes are as much about coming together—being territorialized, as they are about being pulled or drifting apart—being deterritorialized (McCann and Ward 2012). As Anna Tsing (2015, 23) has suggested, we may even think of them as “open-ended gatherings”—assemblages that “don’t just gather lifeways; they make them.” In other words, models are not totalized, given entities that spread in the way policy experts assume or plan.

Second, and most importantly, my designation of the assemblage as viral allows for a greater recognition of the role emotions and desire play in circulating ideas, models, and policies. As Deleuze and Guattari (1987, 399) attest, “Assemblages are passional, they are compositions of desire. . . . The rationality, the efficiency, of an assemblage does not exist without the passions the assemblage brings into play, without the desires that constitute it as much as it constitutes them.” A reading of the traveling model as a viral assemblage—defined as a fluid, socio-technical process and a constellation of human and non-human actors, things, unpredictable moments, and relations that have contagious and affective qualities14—allows us to recognize the messy, material, discursive, and emotional relationships that have facilitated and fueled Awra Amba’s virality.

By paying attention to the role that emotions and desire play in fueling a model’s virality, my theoretical approach speaks to the affective turn in anthropology. While affect theorists often draw a clear distinction between affect and emotions—defining affect as the largely unconscious and unintentional forces and intensities that activate and deactivate human bodies and emotions as the subjective responses—my approach is inspired by scholars who have problematized this distinction. More specifically, I am influenced by Sara Ahmed (2004a, 117), who challenges the idea of emotions as coming from within a located, bound subject and which “then move outward towards others.” Emotions are not “a private matter, that [ . . . ] simply belong to individuals” but “a way of being directed towards particular things” (Schmitz and Ahmed 2014, 99). As a form of capital, emotions circulate between bodies, “affecting bodily surfaces or even how bodies surface” (Ahmed 2004a, 120). Emotions do, in other words, have affective capacities—the power to move things. This is partly why the distinction between affect and emotions is problematic and the reason I use these terms interchangeably.


1. For a discussion of Norwegian gender equality policies and models and their translation into practice in development projects in Ethiopia, see Hilde Selbervik and Marit Tolo Østebø (2013), “Gender Equality in International Aid: What has Norwegian Gender Politics Got to Do With It?” Gender, Technology and Development 17(2): 205–228; Marit Tolo Østebø, Haldis Haukanes, and Astrid Blystad (2012), “Strong State Policies on Gender and Aid: Threats and Opportunities for Norwegian Faith-Based Organisations,” Forum for Development Studies 40(2): 193–216; and Marit Tolo Østebø (2013), “Translations of Gender Equality in International Aid: Perspectives from Norway and Ethiopia,” (PhD diss., University of Bergen).

2. For an overview of the WEGE program see Heidi Holt Zachariassen, 2012, “From the Bottom Up: Lessons About Gender Mainstreaming in the Andes from Digni’s Women Empowerment and Gender Equality (WEGE) Programme,” Gender and Development 20 (3): 481–90.

3. The workshop was a two-day, closed event, where the members of the delegation presented and discussed their respective women’s empowerment and gender equality initiatives.

4. This research project, Protection of Women’s Rights in the Justice Systems of Ethiopia, was carried out by the ILPI in collaboration with six universities in Ethiopia and funded by the Royal Norwegian Embassy in Ethiopia.

5. See Minna Salami, “Awra Amba, an Ethiopian Village where Gender Equality is Real,”; and “Awra Amba: Where Gender Equality is Real,”

6. For the music video see Ahmed Teshome, “Awra Amba,”

7. In an earlier version of this manuscript, I used the term simuloscene, borrowed from Jeremy Trombley (2017). Since all simulations are models, but not all models are simulations, the term modeloscene is far more comprehensive, capturing the many forms and ways that models emerge and are put into use in our contemporary world. I am thankful to one of my former students, Laurin Baumgardt, for suggesting this as an alternative concept.

8. Gabriel Tarde (1843–1904) was a French sociologist, criminologist, and social psychologist known for his theory of imitation and innovation. Sampson (2012, 6–7) emphasizes that his book, Virality: Contagion Theory in the Age of Networks, is not a restoration or revival of Tarde’s social epidemiology. Instead, he claims to offer “a resuscitation of his approach. This involves a carrying forward of an interpretation of Tardean ideas so that they can be linked, transversally, to contemporary notions, breathing life into social theory, and contagion theory, in particular.”

9. See for example Lisa A. Boden and Iain J. McKendrick (2017), “Model-Based Policymaking: A Framework to Promote Ethical ‘Good Practice’ in Mathematical Modeling for Public Health Policymaking,” Frontiers in Public Health 5 (68): 1–7; Robert Christley et al. (2013), “‘Wrong, but Useful’: Negotiating Uncertainty in Infectious Disease Modelling,” PLOS ONE 8 (10): 1–13; Catherine Grant et al. (2016), “Moving Interdisciplinary Science Forward: Integrating Participatory Modelling with Mathematical Modelling of Zoonotic Disease in Africa,” Infectious Diseases of Poverty 5 (1): 17.

10. With reference to parasitology, Serres (2007, 6) makes an important point in his work, emphasizing that the vocabulary used in natural sciences “bears several traces of anthropomorphism”; it is imported from “common customs and habits that the earliest monuments of our culture tell them [ . . . ]: hospitality, table manners, hostelry, general relations with strangers.” In other words, if one assumes a historical perspective, the relationship between the social sciences and the natural sciences is not unidirectional. Just as the natural sciences may shape the ways we understand cultural and social dynamics, culture and society shape science. This is not surprising, since the natural sciences emerge from and are embedded in society and culture.

11. This is why certain mobile populations such as truck drivers and fishermen are more at risk of a pandemic than other less mobile populations. These mobile actors, in turn, may infect more sedentary populations when they go back to their villages and have contact with their wives/partners.

12. This differentiation—which may confuse the reader—is important. A model is a term we use about an unspecific, general phenomenon. When we give the model a name, such as the Awra Amba model, we acknowledge that the model is a unique, individual entity—an assemblage. Since all models are unique, historical entities that have been generated or have emerged in a particular context, all models are assemblages.

13. Sampson (2012, 6) also makes it clear that “Deleuze plays a central role in this resuscitation” of Tarde.

14. As DeLanda has argued, it is important to acknowledge that assemblage is both a process and a product.

15. I recognize that the feeling of hunger also is contingent on social and cultural factors. This does not mean that hunger is not biological. In fact, we can think of hunger as a state in which biological, emotional, and social processes are closely intertwined.