For researchers endowed with the abstraction habitus of techno-scientific practice, it did not take much to see the benefits of no-till even if most of them would not become salient until after a few cropping seasons. The agronomists and other researchers I met seemed indeed well convinced by it. “But the farmer”, some of them told me, “whatever we say it’s no use. The farmer needs to see it”. This is a refrain Brazilian researchers would promptly recognize. Indeed, many of the demonstration techniques deployed by the project were visual. The main one was the pit , a hole usually around 4-5 feet deep dug by a test plot in order to expose a vertical slice of soil . This technique was deployed in the no-till fields in all participating institutes, to display the compaction layer and the different plant roots. Indeed, within only a couple of years, the difference in length between the cotton roots in the no-till field and in its control counterpart was remarkable. The roots of the different cover crops were also a common target of demonstration through this technique; the brachiaria pits in particular unveiled a whole underground root ecology that is normally not visible, and therefore not taken into account, by farmers and even researchers. The pit displayed elements that were not immediately visible without the mediation of scientific artifacts: even when comparative variables measured in the experiments had not yet shown significant differences between no-till and the control situation, for instance, the underlying problem could be quite readily visualized in the materiality of the compaction layer.Panels were another visualization device deployed at the parcel, to address a key problem found in the milieu paysan: inappropriate fertilizing. Given local farmers’ little access to soil analysis, visual identification came to the fore as an alternative, though much rougher,grow lights way of estimating nutritional deficiencies in cotton and other crops . “Some of them are easier; if the cotton leaf is yellow, the peasant knows he should add urea [i.e., nitrogen]”, one of the local researchers explained.
The reason why it was easier to identify nitrogen deficiency as opposed to others was more practical than cognitive: it had less to do with the clarity of the symptom itself than with the fact that this was the only nutrient for which peasants counted with a disaggregated fertilizer. For other basic agronomic nutrients, most notably potassium and phosphorous , fertilizers were provided by the cotton companies in a single formula, the complexe coton, making it harder to disaggregate what symptoms were linked to which nutrient deficiency. There were multiple other visualization devices, such as pictures of insect pests and natural enemies in catalogues brought from Brazil or already available locally , removal of the cotton plant’s leaves so that its architecture could be examined, collection of in-depth soil samples to show in detail the composition and texture of the compaction layer. Other sensual channels for demonstration like smell or touch also boiled down to a minimum the need for scales, thermometers and other artifacts; these non-human mediators were not always available in the recipient context – most obviously among peasants, but sometimes even in the research institutes. Finally, some pieces of equipment were the object of demonstration at the parcel. The main attraction was the wheeled pulverizing machine, a device I also came across during the CECAT trainings. It had been developed in Embrapa for small-scale agriculture, and was able to carry more liquid and cover more lines than the common back sprayer used by West African farmers, while sparing the laborer of a quite heavy load. Also commonly demonstrated was a plastic bottle device for applying granulated fertilizer; it was cheap – in fact, virtually free – and easy to make. It also reduced the harshness of labor for, when tied to a stick, it could be used standing up straight rather than bending down , besides guaranteeing homogeneity in the distribution of fertilizer dose per hole – an important benefit from the researchers’ point of view.
These were all potential technologies in the waiting to be one day transferred to farmers – their deployment in the project parcel was therefore also experimental. Peasants who visited the parcel did seem to show interest in them, but without the provision of support to their local production and sale, dissemination of brand new artifacts seemed unlikely. The technologies displayed at the parcelle were therefore aimed principally at researchers, technicians, extension agents and farmers – that is, those directly involved in growing cotton. But the project parcel also targeted another audience that was almost as important: government officials, politicians, diplomats, journalists from television and other media, local schools, and occasionally, ethnographers and other academics. As remarked in Chapter 4, visits of this kind were an important project activity, as they sought to assemble a public around it that was essential to sustain the overall network. This lay audience was generally incapable of judging technical matters; it was able however to make a basic aesthetic judgment. Therefore, besides appropriately following all technical parameters, the parcel had to be, above all, beautiful. This aesthetic aspect was also emphasized in other project activities such as the construction of the lab buildings, for which the Brazilians were commended for their care with “presentation and cleanliness”, as the first project coordinator put it. Similarly, in the project fields, plants had to be uniformly tall, green, and loaded with – depending on the time of the year – yellowish flowers or snow white cotton capsules. They had to look alike all across the field, aligned very straight to fill up perfect rectangles and squares. The alleys between the fields had to be free of weeds, with the grass cut short. The whole area had to be clean, with no random objects or garbage thrown around. Panels had to be unsoiled, visible and placed on the right spots.
Two chiwaras were tied up to either side of the vitrine gate, providing a charming finishing touch to the whole scene . Finally, these systematic methods were compounded by another kind of demonstration that happened at the parcel, of a more subtle and practical kind, which was routinely embodied by the only researcher fully dedicated to it, the project coordinator. This was the demonstration of a certain work ethics, which, even if involving much discipline and some degree of ascetism, was not the Calvinist ethics made famous by Weber. It consisted above all in continuous presence and care,led grow lights from very early morning, as farmers themselves do, until later in the day : supervising the work of technicians, doing hands-on work, running multiple research and administrative errands that inevitably appear on a daily basis. The reason for this deep commitment was, above all, practical – as remarked in Chapter 4, among some of the Brazilian cooperantes there was a sense that this project could not afford to fail, and therefore that the parcel, as its most visible face, could not be allowed to “go bad”. This is something that could easily happen if the controls introduced by the project were removed. The beauty and exemplarity of the project parcel was sustained by a delicate composition of controls that were both technical and social: from the rigorous execution of experimental protocols to continuous vigilance against animals, crop looters and other external hazards; from constantly touching base with UNDP and other financial channels so that resources would keep flowing at the pace and magnitude necessary, to making sure that partners in the local institutes were up to date and committed to the project’s routine work. Moreover, this can also be seen as a strategy of enrollment in the Latourian sense, addressing other front liners through display of exceptional commitment and dedication, and some personal challenging. As both an experimental and a demonstration field, therefore, the parcelle was a highly controlled environment, both in a technical and in a social sense. On the one hand, it was protected from external disturbance: it was fenced, guarded, continuously surveilled; it was constantly supervised and acted upon so it would remain aesthetically and technically appropriate; there was a continuous effort to keep local front liners from being diverged from it to other tasks and interests. On the other hand, the parcel was also controlled in a positive sense: those who worked on it enjoyed resources that were not always available to those outside, most notably the peasants but also some of the local researchers and technicians; plants, soil and other non-humans were continuously nourished by vitalities and protections that they would not otherwise enjoy; its aesthetic and technical qualities resonated farther than its boundaries through a selected public that was being formed as the experiments unfolded. Multiple elements of context were therefore brought in in a controlled manner, and this was true not just of material but conceptual entities. Most notably, the milieu paysan could not be brought into the parcel but in an abstract and standardized form; actual peasants would go there to compare the new technologies with something other than what most of them did in their own fields. In the parcel, this abstracted milieu paysan met the new technologies that came from Brazil, who had also been disembedded from their original context.
In the experiments, these were brought together into yet another interface, at the most micro of the scales to be approached in this dissertation. There, they were compared by making non-human actants relate to each other in certain ways and by nourishing them differentially in certain directions; the plan according to which this happened was the experimental dispositif, to which we now turn.The C-4 researchers deployed the term dispositif to refer to the experimental design orienting the tests in the parcels. It is what made the parcel more than just a demonstration site, but, as fundamentally, an experimental one, aimed at testing the behavior of travelling artifacts and techniques in a new environment. It is also that which allowed for scaling down broader contexts – the filière and the milieu paysan – into the parcel’s bounded time-space. As noted, this was done by converting the cotton sector in the C-4 countries into the so-called témoin, or control situation. This control situation worked as yardstick for evaluating the performance of another, which added new elements to it: Brazilian cotton varieties and cover crops, and the three pillars of no-till. This was done by measuring and comparing the behavior of plants and other actants in the two situations according to common factors. This comparison between local system and the local system including the travelling technologies was not one-to-one but multi-factorial, that is, the experiments involved statistically mediated comparisons of multiple factors at the same time. This was done following a common design in agricultural experiments, the so-called split-plot: the experimental field was delimited and successively sub-divided into units of equivalent area, forming a nested configuration in which each sub-unit was “treated” according to a factor that varied quantitatively or qualitatively, so that the effects of differential treatments in each of them could be compared both within the same and across different cropping seasons. This way, the project’s three components were simultaneously brought together in the overlapping treatments performed in each sub-unit. While a set of treatments defined for instance the cereals to be grown as main crop , another prescribed association with different cover crops for each of them. The following year in the same plot, a set of treatments would define the cotton varieties to be sowed, and another would prescribe no-tillage or tillage. Other than the variations in treatment prescribed by the dispositif, the remaining factors were kept constant: environmental conditions such as rainfall and presence of insects, fungi, and other small forms of life, and basic crop management operations such as sowing, fertilizing, thinning, weeding and harvesting. Each treatment was repeated in as many sub-units as possible for the results to be considered statistically reliable – or, as one of the researchers put it, “scientific”. The technicians and researchers then measured, at certain dates and for each sub-unit, the effects of these controlled interventions on variables established beforehand as relevant: crop yields, plant biomass, physic-chemical characteristics of the soil, plant density and height, appearance of first flowers and leaves, or, in the case of cotton, both agronomic and technological indicators .