Kawanda biotech laboratory, Uganda. September 2017 (S. Calkins)

This piece grapples with a big question: Should the forms of agency of living beings in an infrastructure be analytically held apart from those of other-than-living matter? While I can’t satisfyingly answer this question here, this short piece gestures towards a response by juxtaposing two mundane elements of a particular scientific infrastructure that I encountered while doing ethnographic fieldwork in Uganda and later in Australia, namely banana plants and synthetic chemicals. I follow Susan Leigh Star (1999) in understanding the assembly of humans, technologies, chemicals, plants, ideas etc. that enable the practice of plant biotechnology in Uganda as its scientific infrastructure. Chemicals are an interesting part of the infrastructure: they are at once necessary for procedures in molecular biology and, at the same time, they can leak and wander, causing anxiety, pollution, and damage. By contrast, for Ugandan biologists, who at times experience toxic exposures to chemicals due to unstable infrastructures, their experimental plants seem less problematic infrastructural elements.

This might surprise given that the plants in question here are GMOs and are produced in the lab. These banana plants and their fruits are at the core of this infrastructure and the transnational scientific work it enables. They owe their existence to a finely calibrated set of scientific procedures that aim at artificially increasing the beta-carotene contents of plants and thereby prevent vitamin-A deficiency within the Ugandan population. Yet Mr. Nayiga, a Ugandan agronomist, asserted that “bananas do what they always do: they grow”. He denied, in other words, any essential difference between conventional banana plants and the genetically transformed banana plants that he and his colleagues were testing in trial fields. Mr. Nayiga’s response was typical of how many Ugandans think about the banana – namely, as a sturdy plant that grows without much fuss. In the trial field at Kawanda research institute, researchers devote a great deal of attention to experimental plants and their particularities. It is possible to think about banana plants here as agents, pursuing their own projects in the world, but still these plants do not subvert the research infrastructure. They rely on it and grow along with it. Even conventionally farmed banana plants rely on human assistance for their reproduction (Calkins 2019). Wild banana plants are full of hard, inedible seed. The domesticated banana plant was bred over hundreds of years to contain no seed. That means it reproduces asexually. New plants are always clones of the mother plant and are produced through tissue culture techniques in commercial banana farming; in subsistence settings farmers dig up young shoots from the root and replant them elsewhere in the field. A multispecies approach might be inclined to try to tease out how the banana plant nonetheless exceeds human design and planning (which it surely does too). The more salient points, however, are that banana plants collaborate with humans, and that their current form as high-yielding, fruit-bearing crop hinges on human-built infrastructures; this of course is all the more so true for GM bananas.

Fieldwork in the banana field trial, January 2016 (S. Calkins)

Banana plants, buildings, power supply, technologies, scientific protocols, chemicals, and skilled personnel are all elements of the scientific infrastructure that facilitate the practice of molecular biology at this Ugandan lab. Classical work like Star (1999) mobilized the notion of infrastructure to capture forms of agency that are distributed between humans, concepts and things, and which cannot neatly be ascribed to a singular entity. This recalls ANT approaches that, though working with the trope of the network and not infrastructure, undertook a symmetrical treatment of the non-human without a priori decisions about who can be considered an actor or not. ANT wasn’t particularly attuned to the forms of expression of non-human living beings, but we still find examples of living non-humans. Think, for example, of the well-known mutineering scallops of St. Brieuc Bay who refuse being enrolled by scientists (Callon 1986). Moreover, the larger concern ANT put on the table lingers: when writing about infrastructures do we need to decide a priori whose agency – whether human, animal, plant or thing – needs attending to? 

In recent years scholarly interest in multispecies relationships has grown rapidly (e.g., Tsing 2015). At times, it may seem that living nonhumans like animals, plants or fungi, due to their liveliness as well as their ability to affect humans, deserve more attention than non-living nonhumans. Comparing the roles of plants and chemicals in the scientific infrastructure at this Ugandan lab lets us at least question this assumption. Chemicals here are infrastructural dissidents that affect humans physically and instill fear. Industrially produced chemicals are used at the molecular biology lab in Kawanda to work on plant genomes, to genetically modify and test them. These synthetic substances seem to be less compliant elements of the research infrastructure than banana plants. Unlike experimental plants that remain rooted in a place, chemicals percolate and diffuse through infrastructural gaps. The agentive ways in which researchers talk of various chemicals underscores this: chemicals do things like disrupting, fusing, clarifying, opening, staining, congealing, forging. However, scientists also describe them as seeping, leaking, smelling, sticking, or expiring. They also see them as causative forces behind a range of health concerns, from headaches, eyesores and runny noses, to miscarriage and cancer.  

Registering bodily effects of infrastructural neglect is particularly common across settings in the global south, where infrastructures were planted by colonial regimes and  later often only renewed in a piecemeal fashion (Geissler et al. 2016). To give one example for how the dangerous material and metaphorical lives of chemicals are enabled by the frequent breakdowns and shortages of infrastructure at Kawanda: benchwork in molecular biology that involves toxic chemicals is done under so-called fume hoods – a small chamber that sucks up the fumes. The scientist’s head stays outside and only the hands are moved inside. John Okemi, a postdoctoral researcher, was busy pipetting one day when the light bulb inside of the fume hood burned out and the lab would unable to replace it for works. Without that light, Dr. Okemi could not see the fine layer of plant DNA in the test tubes that he had to remove. Under pressure to produce results for project deadlines with oversea partners and to prevent sensitive materials from degrading, he began to move the tubes out of the fume hood, holding them up in the room to be able to tell the DNA apart from plant debris while pipetting. The protocol involves several rounds of adding chemicals, centrifuging and then pipetting off the upper phases. This means significant exposure to chemicals over three or four hours, particularly to the highly toxic beta-mercaptoethanol. As the room was filling with the stench of this chemical – a stinky mixture  of burnt plastic and rotten eggs that lingers for hours, sticks to clothing and hair – Dr. Okemi, his assistants and I began experiencing unpleasant physical sensations like dizziness, burning teary eyes and respiratory irritations. Some of us had to step outside for a few minutes.

Work at the fume hood, safely (S. Calkins)
Work at the fume hood, unsafely (S. Calkins)

The whole lab in Kawanda is perceived to be unsafe and overexposed to toxic chemicals that have seeped into its equipment, walls and furniture; that stick to desks, cupboards and door handles. Chemicals leak into the environment, forming bonds with organic and inorganic matter, confounding the boundary between life and non-life. What adds to their spectral quality is that evidence of chemical exposure is hard to establish without a costly apparatus that would allow testing for specific toxic substances, a fact that plays into the hands of power – in this case, overseas partner labs in the Global North – that can be strategically ignorant of exposures (Murphy 2006: 9, 15). 

Infrastructure is uneven and creates uneven geographies of toxic exposure. This case at hand where chemicals are leaky and vagrant in a Ugandan lab at least cautions against reinscribing an a priori asymmetry between living beings as actors, able to carry out some form of unreckoned or purposive action, and materials that presumably lack this capacity.  The question I opened with was whether we need to differentiate between different nonhuman agencies assembled in an infrastructure, and more specifically whether we have to think other-than-living agencies like chemicals differently from animal or plant agencies. Although a number of authors have for a while sought to flatten such distinctions, treating various agencies in more symmetrical ways (Callon 1986; Bennett 2010), these questions are still unsettled, especially as we theorize the multitude of complicated agencies assembled in expansive infrastructures. Elisabeth Povinelli (2016) recently challenged that we need not only rethink the symmetry between human and nonhuman as well as between living and non-living entities but rather dissolve the distinction between life and non-life itself because this lies at the heart of the Western anthropocentric ontology. Similarly, Chen (2012) also with reference to toxic chemicals had suggested thinking about their animacy, their life-like ability to migrate, affect and alter their surroundings. In line with such authors, the task for researchers of infrastructure then is not to give priority to living or non-living agencies but to rather unknow these distinctions a priori and explore what animacies, substances and sentiments matter in a particular case at hand. 

At Kawanda, banana plants collaborate with the research infrastructure put in place and they themselves remain rooted in a place as well; being a “vegetal dissident” was the exception, not the rule. Synthetic chemicals, by contrast, were insubordinate and unruly, moving into spaces and bodies where they shouldn’t have moved. Leaking chemicals confound the material and conceptual boundaries of infrastructure as well as of body and environment, by forging new molecular bonds as they seep outside, and it is this boundary crossing that demands more conceptual engagement and accounting for politically. Of course, speaking of chemicals as a class of things might be as misleading as the category plant insofar as it subsumes a great diversity, and scholarly engagement with specific chemical substances is on the rise. However, Max Liboiron, Nerea Calvillo and Manuel Tironi (2018) recently cautioned against a tendency especially among scholars of science and technology to get overly excited about specific toxic chemicals and foregrounding them in their analysis. Such fetishizing of toxic substances, they argue, risks backgrounding the larger political issues – or for the sake of my argument the infrastructural inequalities of toxic exposure – of why and how such substances are encountered and by whom. While my heart also beats for multispecies relationality, in places like Kawanda the local biochemical relations that are produced in unequal transnational scientific infrastructures deserve as much or more ethnographic scrutiny as they shape where, how and who bonds with toxic chemicals.



Bennett J (2010) Vibrant matter: A political ecology of things. Durham: Duke.
Calkins S (2019) Health as growth. Bananas, humanitarian biotech, and human-plant histories in Uganda. Medicine Anthropology Theory 6(3): 29–53.
Callon M (1986) Some elements of a sociology of translation: Domestication of the scallops and the fishermen of St. Brieuc Bay. In: Law J (ed) Power, action and belief: A new sociology of knowledge? London: Routledge, pp.196-233.
Chen M (2012) Animacies: Biopolitics, racial mattering, and queer affect. Durham: Duke University Press.
Geissler WP, Guillaume L, Manton J and Tousignant N (2016) Traces of the Future: An Archaeology of Medical Science in Twenty First Century Africa. Chicago: University of Chicago Press. 
Liboiron M, Tironi M and Calvillo N(2018) Toxic Politics: Acting in a Permanently Polluted World. Social Studies of Science 48(3): 331-349.
Murphy M (2006) Sick Building Syndrome and the Problem of Uncertainty. Durham: Duke. 
Povinelli E (2016) Geontologies. A requiem to late liberalism. Durham: Duke.
Star SL (1999) The Ethnography of Infrastructure. American Behavioral Scientist 43(3): 377-391.
Tsing A (2015) The mushroom at the end of the world: On the possibility of life in capitalist ruins. Princeton University Press.

Sandra Calkins is Assistant Professor for Social and Cultural Anthropology at the Free University of Berlin. Her current research deals with postcolonial science, plants and humanitarian biotechnology in a Ugandan-Australian research collaboration.