This hybrid, socio-technical character is a sensibility that is by now common sensical in science and technology studies. But it seems to be also present in the emic notion of experimental controls itself: much as in common notions of political control as one’s capacity to influence the behavior of others, what experimental controls do is precisely to direct the behavior of certain non-human entities and make them act in certain ways in relation to others.In socio-technical networks, actants are made to circulate and mutate in certain ways, and not in others; techno-science is not just about metaphysics and epistemology, but fundamentally about practical controls. In both technical and political controls, moreover, agency is not a matter of one-to-one relations, but can be potentially traced to ever-unfolding scales of context. And just as actor network theory is timid about “social” power relations that lie beyond the phenomenological reach of the myopic ethnographer, it does not have much to say either about the agency of non-humans beyond the immediate reach of the scientists’ transcription devices either; in particular, it downplays their more elusive, uncontrolled systemic interactions. But the kind of research that African and Brazilian front liners were performing in their adaptive experiments dealt not with dead mice or synthetic substances within a lab,grow bucket but precisely with a vast range of living elements freely interacting in an open-ended assemblage over which field researchers have much less control than their laboratory counterparts.
As will be suggested here, their work was less about constructing facts through successful stages of purification than about channeling vitalities that were otherwise in certain ways and to certain ends. In other words, and echoing the classic Spinozean question so beloved by neo-vitalists and affect theorists, it was less about figuring out what they were than what they could do. Just as in lab science, this task demands inscription artifacts and experimental protocols; but it relies as often in personal experience, subjective evaluation and even intuition. Vitality in this sense is the Janus face of controls, in society as in nature: it is that which controls try to break down into predictable sets of relations, and channel according to certain ends. Although these aspects were particularly evident in the kind of research observed during fieldwork, they seem to be as true to basic research in the lab as to adaptive experiments in the field. Latour sees field and applied sciences as an extension of the laboratory into the world out there; indeed, as technology transfer makes evident, a whole socio-technical apparatus has to be put in place to sustain the reproduction of a technology in the world out there. But the reverse claim can also be made: lab science, as described in Latour’s own ethnography , may be viewed as a highly “compressed” version of field research, involving many more controls in a much more artificialized environment. Again, in both cases, techno-scientific work is about figuring out what these entities are by looking at what they can do: in Hacking’s terms, to represent natural entities by intervening in them, and vice-versa.
This chapter will provide an account of the field experiments from the double point of view advanced here: the controls deployed, and the vitalities they aimed at channeling. This account will progress along with the project’s scaling moves, identified here through emic idioms encountered in the field: the filière , the milieu paysan , the parcelle , and the dispositif . At each level, the agency of certain actants and the relative controls available to them are foregrounded: the cotton companies that mediate between local production and world markets; the peasant farmers that grow cotton; the project front liners’s technical-political work of influencing human and non-human behavior; and the non-human actants in their actions to the controls introduced. What does cotton production in the C-4 countries look like? This was one of the first questions the project had to tackle, since the project’s ultimate goal, broadly advertised by diplomats and the Brazilian Cooperation Agency, was to help increase cotton production in the four countries. But cooperation discourse, as I argued early on, may go a long way on its own, even without being a faithful reflection of front line practice. In the latter, the broader context of cotton production became important for practical reasons: above all, because the experimental design required a scaled down version of it. Adaptive research involves a controlled comparison between at least two situations: business-as-usual , and the control situation including the new elements to be transferred. The aim is to evaluate how the new technologies will behave in the new environment by controlling common variables that are deemed significant across the two situations. But how to bring such a wide, heterogeneous reality – cotton production in the C-4 countries – into a couple of hectares of field in the project’s experimental plot? This section and the next three will provide an account of the concatenated scaling down operations that made this possible.
Not unlike the “rendering technical” of traditional development aid, at each step heterogeneity and complexity were replaced by an increasingly narrow set of quantifiable and controllable variables. But these context-making and scaling moves took into account projections of the differential controls exercised by the various actors in the cotton assemblage. This was done with an eye on the next stage, when the technologies and its surrounding networks would be scaled up again during transfer to cotton farmers. The broadest scale of the cotton assemblage in the C-4 countries is what I will refer to here as the filière – a rough but not perfect equivalent to the term production chain, or the cotton sector more generally. It is a technical term deployed by economists and other development experts that also appeared in common parlance in the C-4 countries I have been to.220 Internally, it encompasses not just the cotton farms, but all pre- and post-harvest links in the production chain, from basic agronomic research to commercialization,dutch bucket for tomatoes until the cotton fiber is exported. Externally, it is the key structural mediator between local actors involved in cotton production – peasant farmers, research scientists, extension agents, government officials – and processes happening at global scales. As was seen in Chapter 4, it was at the latter scale that the Brazilian cooperantes were first brought into relation with their African counterparts. And even if, by design, a technical cooperation project has to be selective in terms of the filière actors with which to engage during implementation, the front liners had to work with a projection of how the cotton sector at large functioned. In the project’s original text, as we saw, the cotton sector in the C-4 countries was described largely in terms of its failure to keep up with “the technological developments found in other producing countries in other parts of the world” – not just in the North, but in much of the global South. And, in a globalized world, this seriously jeopardizes their competitiveness. Cotton peasant agriculture and the rest of the production chain were therefore portrayed not as in need of being brought into the modern technological innovation treadmill , but as having been pushed to its bottom. A diagnosis like this – that the fundamental problem with the cotton sector in West Africa is its low competitiveness vis-à-vis other producing regions – is typical of a post-neoliberal reforms environment. According to this logic, for instance, the problem with the subsidies question is not the rules of free trade themselves, but the fact that the very countries that have championed their universalization abide by them only selectively. As was evident in the editorial brought in Chapter 4, these rules, which the WTO strives to impose on national legal frameworks worldwide through more or less coercive means, were also taken for granted – at least in discourse – by Brazil’s African partners.
It was not by chance that the cotton sector in the C-4 countries came to the center of debates on global trade: in a highly liberalized trade environment, peasant farmers’ income was directly, and hopelessly, affected by world prices distorted by American subsidies. This highly asymmetric local-global entanglement, manifested most dramatically in the low purchasing price paid each season to West African cotton farmers, has been in place since the early beginnings of cotton production for the rising European textile industry during colonial times. But judging from the available historical literature, ironically by then many cotton farmers actually seemed to have more options, since West Africa was home to a dynamic regional textile industry that,221 in the wake of decolonization and then neoliberal globalization, has been largely out competed especially by Asian textiles. African cotton farmers were left at the mercy of global processes over which not just them, but their own national governments, had little control – thus Compaoré and Touré’s appeal to the rest of the world’s sense of fairness in the pages of the New York Times. Many development analysts and even scholars would be satisfied with tracing the roots of West African farmers’ plight to this asymmetric global configuration; but although this is a key – or perhaps even the key – scale, it is far from the complete picture. The relation between local farmers and global markets is importantly mediated at a national level by the way the filière is structured. The filière mediates not just how much peasants get for the cotton they grow, but minimal industrial-technological standards for the cotton fiber they should follow, or the cost and availability of agrochemicals and other inputs to production. In all four project countries, the sociétés cotonnières, or cotton companies, are the backbone of the filière coton. They are the ones that provide peasant farmers with inputs on credit in the beginning of the crop season, and purchase their production at the end. They are also in charge of most other links in the value chain, both upstream and downstream . In countries highly reliant on cotton exports like Burkina Faso and Mali, they also wield significant political influence. The cotton sector in the C- 4 countries is also an articulation of two-directional processes: both inwards, and outwards to the global stage. The extremely verticalized character of the filière coton is quite straightforwardly a legacy of cotton’s colonial history in Sub-Saharan Africa. Like their predecessors in the colonial monopsonies,the cotton companies remain as the crop’s exclusive buyers, and where there is more than one company, they are assigned jurisdiction over specific production zones. Even if, as part of the Structural Adjustment Programs that took Africa by storm during the eighties and nineties, the national sociétés have been today fully or partly privatized,in many ways the “free peasant economy” still lies “at the interstices of markets and compulsions” of various kinds .Strictly speaking, no West African peasant is today forced to plant cotton – just as no national state is forced to keep relying fiscally on it. For some countries, however, cotton has remained one of the few feasible options. As a researcher from Burkina Faso explained it to me, “Chad has oil, Mali has gold, Benin has the coast; we have but cotton. Not even Thomas Sankara’s revolutionary government could afford to disinvest in cotton, even if it was considered a colonial crop”.Similarly, in West Africa rarely does a farmer grow exclusively cotton, but most often a mixture of this export cash crop and others aimed at local consumption or regional markets. In all four countries, however, cotton is the only crop that counts with an organized filière: in practice, this means a structured way by which farmers may access basic agricultural inputs, especially fertilizers. Outside of the filière coton, an open market in agricultural inputs is either absent or little developed. This lack of control over inputs to production is at the root of the previously mentioned gap between research’s recommendations and peasants’ practices: fertilizers provided by the companies to be applied on cotton are diverted to food crops. In spite of the centrality of the sociétés for the cotton sector in the C-4 countries, during Phase I the project worked closer with the two other major components of the filière: the research institutes and, secondarily, farmers.